Tesis Eduardo Infantes

Instituto Mediterr´ aneo de Estudios Avanzados IMEDEA (CSIC-UIB) Departamento de Ecolog´ıa y Recursos Marinos Departamento de Tecnolog´ıas Marinas, Oceanograf´ıa Operacional y Sostenibilidad

Wave hydrodynamic effects on marine macrophytes

Memoria presentada para optar al t´ıtulo de doctor en Biolog´ıa de la Facultad de Biolog´ıa, Universidad de las Islas Baleares by Eduardo Infantes Oanes

PhD supervisors: Dr. Jorge Terrados Mu˜ noz Dr. Alejandro Orfila F¨ orster Ponente: Dr. Miseric` ordia Ram´ on Juanpere Esporles, Julio de 2011

To my parents and family...

iii

iv

Abstract This doctoral Thesis was developed to explore interactions between benthic marine macrophytes, substratum type and fluid dynamics. Quantitative knowledge as well as a predictive capacity is obtained to estimate the presence of a seagrass meadow, two species of invasive macroalgae and two species of seagrass seedlings as a response of near-bottom orbital velocities, drag forces, root anchoring capacity and substratum type. Additionally, the effect of a seagrass meadow on wave propagation under natural conditions is also evaluated. In order to get a deeper knowledge on these processes, data from different sources such as numerical models, aerial photographs, field experiments and flume measurements are combined.

Light and temperature are considered as the main determinants for the spatial distribution of marine macrophytes but hydrodynamic conditions and substratum type are also key factors limiting the distribution of marine vegetation.

Wave hydrodynamic effects on macrophytes have been studied in three ways: i) The near-bottom orbital velocities that set the upper limit of a Posidonia oceanica seagrass meadow are obtained by correlating hydrodynamics and the spatial distribution of the meadow. ii) The role of hydrodynamics in the establishment P. oceanica and Cymodocea nodosa seagrass seedlings is evaluated.

Drag forces and root anchoring capacity of seedlings are studied in a biological flume while seedling survival is addressed under natural conditions. iii) Substratum type plays an important role on the spatial settlement and distribution of marine macrophytes. Substratum cover of the invasive macroalgae Caulerpa taxifolia and Caulerpa racemosa indicates that these species are more abundant in rocks with photophilic algae and in the dead matte of seagrass P. oceanica than in sand or inside the P. oceanica meadow. Correlative evidence shows that C. taxifolia and C. racemosa tolerate near-bottom orbital velocities below 15 cm s−1 and that C. taxifolia cover declines at velocities above that value. v

Wave damping induced by a seagrass meadow of P. oceanica is evaluated under natural conditions using data from bottom mounted acoustic doppler velocimeters. The bottom roughness is calculated for the meadow as 0.42 m using flow velocities above the seagrass. Wave friction factor has been related to the drag coefficient on the plant and obtained for two storms so as to compute the damping along a transect. Drag coefficient values ranged from 0.1 to 0.4 during both storms. The expected wave decay coefficient for different seagrass shoot densities and leaf lengths are also predicted.

A relation between fluid dynamics and benthic macrophytes is shown. The influence of high near-bottom orbital velocities have an effect on the spatial distribution, growth and colonisation processes of seagrass and macroalgae species.

The influence of a seagrass meadow on wave propagation is also aparent, with potential impact on sediment stability and coastal erosion. Predicted and measured quantitative results are provided suchs as near-bottom orbital velocities and drag coefficients that could be tested and compared to other benthic marine macrophytes species on other locations.

This Thesis is in the intersection between Ecology and Fluid dynamics a research area characterized by strong nonlinear interactions. To get a deep insight on the processes involved a bench of mathematical tools are used in order to apply physical principles to understand and predict the behaviour of marine macrophytes in the Mediterranean Sea. Besides, experiments to validate the theories have been developed under both controlled and natural conditions.

vi

Resumen Esta Tesis doctoral ha sido desarrollada para explorar las interacciones entre el oleaje, los macr´ ofitos bent´ onicos marinos y el tipo de sustrato. Se ha obtenido conocimiento cuantitativo y capacidad de predicci´on para estimar la presencia de una pradera de faner´ ogamas marinas, dos especies de macroalgas invasoras y dos especies de faner´ ogamas, relacion´ andolas con las velocidades orbitales en el fondo, las fuerzas de resistencia, la capacidad de anclaje de las ra´ıces y el tipo de sustrato. Adem´ as, ha sido evaluado el efecto de una pradera de faner´ogamas en la propagaci´ on del oleaje en condiciones naturales. Con el fin de obtener un conocimiento m´ as profundo de estos procesos, se han combinado datos de diferentes fuentes tales como modelos num´ericos, fotograf´ıas a´ereas, experimentos de campo y medidas en canal de ensayo.

La luz y la temperatura estan considerados como los principales determinantes de la distribuci´ on espacial de macr´ofitos marinos, pero las condiciones hidrodin´ amicas y el tipo de sustrato son tambi´en factores clave que limitan la distribuci´ on de la vegetaci´ on marina. Los efectos del oleaje en la distribuci´on espacial de macr´ ofitos se han estudiado de tres formas. i) La velocidad orbital en el fondo que establece el l´ımite superior de una pradera de la faner´ogama marina Posidonia oceanica se ha obtenido mediante la correlaci´on del hidrodinamismo y la distribuci´ on espacial de la pradera. ii) Se ha evaluado el papel del hidrodinamismo en el establecimiento de pl´antulas de P. oceanica y Cymodocea nodosa. Las fuerzas de resistencia y la capacidad de anclaje de las ra´ıces de las pl´ antulas se han medido en un canal de ensayos, mientras que la supervivencia de pl´ antulas se ha estudiado en condiciones naturales. iii) El tipo de sustrato juega un papel importante en la colonizaci´on y en la distribuci´on espacial de los macr´ ofitos. La cobertura del tipo de sustrato que colonizan las macroalgas invasoras Caulerpa taxifolia y Caulerpa racemosa indica que estas especies son m´ as abundantes en rocas con algas fot´ ofilas y en la mata muerta de P. oceanica vii

que en arena o en el interior de praderas de P. oceanica. Datos experimentales en el mar confirmaron que la cobertura de ambas macroalgas es mayor sobre mata muerta de P. oceanica que sobre arena.

La atenuaci´ on del oleaje inducida por una pradera de P. oceanica en condiciones naturales se ha estudiado con datos obtenidos por velocimetros ac´ usticos doppler fondeados en la pradera. Se ha calculado la rugosidad de la pradera utilizando velocidades de flujo por encima de la pradera. El factor de fricci´on ha sido relacionado con el coeficiente de resistencia durante dos tormentas con el fin de calcular la atenuaci´ on del oleaje a lo largo de un transecto. Los coeficientes de resistencia varian desde 0.1 hasta 0.4 en ambas tormentas. Se han predecido los coeficientes de atenuaci´ on de onda esperados para diferentes densidades de plantas y longuitudes de hojas.

La influencia de las altas velocidades orbitales en el fondo tienen un efecto sobre los procesos de distribuci´ on espacial, el crecimiento y la colonizaci´on de faner´ ogamas y macroalgas. La influencia de una pradera de P. oceanica en la propagaci´ on del oleaje es tamb´ıen evidente, con un impacto potencial sobre la estabilidad de los sedimentos y la erosi´ on costera. Las predicciones y los resultados cuantitativos obtenidos de las velocidades orbitales en el fondo y los coeficientes de resistencia, pueden ser testados y comparados con otras especies de macr´ ofitos bent´ onicos en otros lugares.

Esta Tesis se encuentra en la intersecci´ on entre la ecolog´ıa y la din´amica de fluidos, un ´ area de investigaci´ on que se caracteriza por fuertes interacciones no lineales. Para profundizar en los procesos que intervienen se utilizan un abanico de herramientas matem´ aticas con el fin de aplicar los principios f´ısicos para comprender y predecir el comportamiento de los macr´ofitos marinos en el Mar Mediterr´ aneo. Adem´ as, los experimentos para validar las teor´ıas se han desarrollado tanto en condiciones controladas y como naturales.

viii

Acknowledgements Foremost, I would like to thank my parents and family who always supported me with love and care during good and difficult moments. I also deeply thank them for always giving me the education that they though was best. I hope to do as well with my children as they did with me.

I would like to thank the support of my two Ph.D advisors Dr. Jorge Terrados Mu˜ noz and Dr. Alejandro Orfila F¨orster for motivating me from the beginning. Thanks for their always availability to discuss, comment and solve problems with a constructive feedbacks over the last years. To Jorge, for introducing me into the amazing world of seagrass ecology, showing me the importance of scientific knowledge and experimental design to understand the marine ecosystem. I will remember the conversations over hundreds of SCUBA-dives and during field trips. To Alejandro, for showing me the importance of coastal dynamics and for his patience in teaching me the numerical aspects of fluid dynamics. Thanks for the time you spent answering all my questions, and pushing me towards the end of the Thesis, thanks for your continuous support, great advice and friendship.

Thanks to the Ministerio de Educaci´on y Ciencia for the FPI grant BES200612850 linked to the project CTM2005-01434 for giving me funding for the PhD work, but specially for giving me the opportunity to visit other Universities and research institutes such as the Massachusetts Institute of Technology (USA), the Nederlands Institute voor Ecologie (Netherlands) and the Horn Point Laboratory (USA). These visits opened the opportunity for collaborations and future working perspectives.

Thanks to Professor Heidi Nepf and Mitul Luhar for hosting me at the Massachusetts Institute of Technology (MIT), Boston, US. Heidi great expertise and ix

wave flume let me look in detail at hydrodynamical effects of seagrasses on the waves and currents where I became interested in wave attenuation caused by seagrasses. It was inspiring to compare my field-base knowledge with a model-base perspective. During my visit to the MIT Mitul became interested in working with seagrass in the field and he came to Mallorca to work on wave attenuation in the field with the seagrass Posidonia oceanica.

Thanks to Dr. Tjeerd Bouma for hosting me at the Nederlands Institute voor Ecologie, (NIOO). Tjeerd ideas, endless energy and always great advice contributed to my interest in drag forces under unidirectional and oscillatory flows as well as to root anchoring capacity of seedlings and sediment erosion.

The excellent biological flume at the NIOO let me work with seedlings of Posidonia oceanica and Cymodocea nodosa.

Thanks to Professor Evamaria Koch and Dale Booth for hosting me at the Maryland Environmental Center, Horn Point Laboratory (USA). Evamaria shared with me her enthusiasm and expertise in seagrass research linking seagrass, hydrodynamic and sediment. Thanks to Dale for her great hospitality during the three months and for showing excellent logistics and organisation skills during field trips.

Thanks to my working colleagues Francisco Medina-Pons, Ines Castej´on Silvo, Fiona Tomas Nash and Marta Dom´ınguez for sharing their time helping setting field experiment using SCUBA and for the good moments shared outside work. Other people that I would like to express my gratitude and support during these years are Constanza Celed´ on Neghme, Francisco Sancha Bermejo and Pablo Fern´ andez M´endez. Thanks to my flatmates Albert Fern´andez Chac´on, Patricia Puertas, Ylva Olsen, Susana Chamorro, Lorena Basso who make my life happier after work.

I would like to thank Puertos del Estado for the HIPOCAS and WANA wave climate data. To the Marina of Cala D’Or (Santanyi), Club N´autico de S’Arenal and Club N´ autico de Cala Bona which kindly made available its harbour facilities for executing the field work.

Eduardo Infantes Oanes

x

Contents Abstract

iii

Resumen

v

Acknowledgements

vii

Contents

ix

List of Figures

xiii

List of Tables

xix

List of Symbols

xxi

1 Introduction 1.1 Upper limit distribution of macrophytes 1.2 Hydrodynamics and seedling survival . .

1.3 Substratum type and invasive species . .

1.4 Wave damping by macrophytes . . . . .

1.5 Aims and motivations . . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

2 Wave energy and the upper depth limit of P. oceanica 2.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.2 Material and methods . . . . . . . . . . . . . . . . . . . .

2.2.1 Study Area and Regional Settings . . . . . . . . .

2.2.2 HIPOCAS database (1958 - 2001) and deep water characterization . . . . . . . . . . . . . . . . . . . .

2.2.3 Shallow-water wave conditions . . . . . . . . . . .

2.2.4 Bottom typology and bathymetry . . . . . . . . . xi

. . . . .

. . . . .

. . . . .

. . . . .

1 4 5 7 8 9

13 . . . . 13 . . . . 14 . . . . 14 wave . . . . 15 . . . . 17 . . . . 18

2.3 2.4

Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3 Seedling tolerance to wave exposure 3.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . .

3.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3.2.1 Drag forces and drag coefficient . . . . . . . . . . .

3.2.2 Critical erosion depth and minimum rooting length 3.2.3 Field study . . . . . . . . . . . . . . . . . . . . . .

3.2.4 Statistical analysis . . . . . . . . . . . . . . . . . .

3.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

19 22 27 27 28 28 30 31 33 34 39

4 Substratum effect on invasive Caulerpa species 45 4.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 4.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 4.2.1 Study Sites . . . . . . . . . . . . . . . . . . . . . . . . . . 46 4.2.2 Presence of Caulerpa over different substrata . . . . . . . 46 4.2.3 Short-term experiment evaluating substratum effect on Caulerpa persistence . . . . . . . . . . . . . . . . . . . . . 47 4.2.4 Wave climate and modelling at the short-term experimental sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 4.2.5 Field wave measurements and velocity profiles . . . . . . 50 4.2.6 Statistical Analysis . . . . . . . . . . . . . . . . . . . . . . 51 4.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 4.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 5 Effect of a seagrass meadow on wave propagation 5.1 Summary . . . . . . . . . . . . . . . . . . . . . . . .

5.2 Theory . . . . . . . . . . . . . . . . . . . . . . . . . .

5.3 Material and Methods . . . . . . . . . . . . . . . . .

5.3.1 Determination of Nikuradse roughness length 5.3.2 Determination of drag coefficient . . . . . . .

5.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . .

5.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

63 63 64 67 69 71 71 75

6 General discussion and conclusions

81

7 Recomendations for future work

87

xii

A Background review on fluid A.1 Waves . . . . . . . . . . .

A.2 Wave propagation . . . .

A.3 Boundary Layers . . . . .

dynamics 89 . . . . . . . . . . . . . . . . . . . . . . 89 . . . . . . . . . . . . . . . . . . . . . . 93 . . . . . . . . . . . . . . . . . . . . . . 94

Bibliography

97

Other publications

113

xiii

xiv

List of Figures 1.1

Schematic diagram of the approach presented. . . . . . . . . . . .

10

2.1

(a) Location of Mallorca in the Mediterranean Sea. (b) Location of the study area of Cala Millor in Majorca. The asterisk (* indicates HIPOCAS node 1433. (c) Bathymetry of Cala Millor with isobaths (in meters). . . . . . . . . . . . . . . . . . . . . . .

15

Directional wave histogram for HIPOCAS node 1433 10 km from Cala Millor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

16

Bottom typology off Cala Millor (Mallorca, western Mediterranean Sea). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

20

(a) Distribution of wave heights (m) at the beach derived from mean wave conditions (11.25o , Hm = 1.53 m, Tp = 7.30 s). (b) Distribution of near-bottom orbital velocities (ub , m s−1 ) at the beach and cover of the P. oceanica meadow (gray) and dead rhizomes (dark gray). . . . . . . . . . . . . . . . . . . . . . . . . . .

22

Percent coverage of the different bottom types in each of the nearbottom orbital velocity (ub ) intervals established in Cala Millor. .

23

Mean near-bottom orbital velocity above each bottom type in Cala Millor. Different capital letters indicate significant differences between bottom types (post-hoc multiple pairwise comparison of mean ranks, p < 0.05). Error bars show 95 % confidence intervals. n-values indicate the number of points selected randomly in each bottom type. Differences in n-values between bottom types are driven by the differences in percentage covers of each bottom type in the study area. . . . . . . . . . . . . . . . .

23

2.2

2.3

2.4

2.5

2.6

xv

3.1

3.2

3.3

3.4

3.5

3.6

3.7

3.8

Sketch of the flume experimental set-up showing the critical erosion depth and minimum rooting length (Ld ) of seedlings. (a) Seedling in the flume, (b) seedling before dislodging from the sediment after the discs addition. Critical erosion depth is equivalent to the total height of discs added when the seedling is dislodged.

Not drawn to scale . . . . . . . . . . . . . . . . . . . . . . . . . .

30

(a) Location of Mallorca Island in the Mediterranean Sea. (b) Location of study area and WANA node. (c) Location of the experimental sites, Cap Enderrocat (triangles) and Cala Blava (circles). Bathymetric contours in meters. . . . . . . . . . . . . .

32

Experimental plots with a) Posidonia oceanica and b) Cymodocea nodosa seedlings at the beginning of the field experiment. . . . .

32

Drag forces acting on seedlings in (a) unidirectional flow and (b) oscillatory flow. (mean, SE, n = 5). . . . . . . . . . . . . . . . .

34

Drag forces acting on individual seedlings of different surface area in (a) unidirectional flow and (b) oscillatory flow. Flow velocity of 16 cm sâ&#x2C6;&#x2019;1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

35

Drag coefficient versus Reynolds number for Posidonia oceanica and Cymodocea nodosa under undirectional and oscillatory flow.

Experimental data under uniform flow for P. oceanica seedlings (triangles) and for C. nodosa (squares). For oscillatory flow P. oceanica seedlings (circles), and C. nodosa (crosses). Solid lines are the linear fitting for the different flow conditions. . . . . . . .

36

(a) Critical erosion depth and (b) Minimum rooting length of Posidonia oceanica and Cymodocea nodosa exposed to two orbital velocities (u = 5 and 10 cm sâ&#x2C6;&#x2019;1 ). . . . . . . . . . . . . . . . . . .

38

Seedling survival on the experimental plots in Mallorca, August 2009 to February 2010. (a) Posidonia oceanica, (b) Cymodocea nodosa, and (c) wave heights in deep water (WANA node) shown in grey line and propagated wave heights in Cala Blava at 12 m depth shown in black line. Gaps in propagated Hs corresponds to wave directions other than Southwest to Southeast no affecting the study area. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

41

xvi

4.1

4.2

Location of the study areas. (a) Mallorca Island in the Mediterranean Sea. (b) Location of the four study areas and deep water wave data (WANA nodes). (c) Bathymetry of Cala D’Or with location of WANA node. (d) Bathymetry of Sant Elm. Experimental plots, WANA nodes and ADCP location in Cala D’Or. .

47

Photographs of the experimental set up for Caulerpa taxifolia in (a) natural sandy bottom and (b) the model of dead matte of the seagrass Posidonia oceanica. (c) Illustration of Caulerpa fragments fixed to the plots using pickets. . . . . . . . . . . . . .

48

4.3

Percent cover of (a) Caulerpa taxifolia and (b) Caulerpa racemosa in the different types of substratum during two consecutive years. 53

4.4

Relative frequency of total substratum surveyed and of the presence of Caulerpa taxifolia and Caulerpa racemosa on each substratum. Low presence of rocks was recorded at the C. taxifolia site. Rocky bottom with photophilic algae is indicated as Rocks.

54

(a) Caulerpa taxifolia cover (normalised means ± SE, n = 3). (b) Water temperature at the experimental site, hours of light per day and near-bottom orbital velocities, ub (means ± SD). Cover area on each sampling date (A) in cm2 normalised by initial cover area (A0). Dotted square indicates the period used for the statistical analysis ANOVA, Table 4.3. . . . . . . . . . . . . . . . . . . . . .

57

(a) Caulerpa racemosa cover (normalised means ± SE, n 3). (b) Water temperature at the experimental site, hours of light per day and near-bottom orbital velocities, ub (means ± SD). Cover area on each sampling date (A) in cm2 normalised by initial cover area (A0). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

58

Vertical profiles of velocity (means ± 95 % CL). For convenience only the 40 cm above the bottom is shown. . . . . . . . . . . . .

59

(a) Location of Mallorca Island in the Mediterranean Sea. (b) Location of the study area. (c) Location of the transect and deployment sites in Cala Millor. (d) Bathymetric profile and distance between the deployment sites. . . . . . . . . . . . . . . .

68

Acoustic doppler velocimeter (ADV) deployed in the Posidonia oceanica seagrass meadow. . . . . . . . . . . . . . . . . . . . . . .

69

4.5

4.6

4.7

5.1

5.2

xvii

5.3

5.4

5.5

5.6

5.7

5.8

Top panel: Significant wave height at 16.5 m depth (solid), at 12.5 m depth (doted), at 10 m depth (dash-doted) and at 6.5 m depth (dashed). Grey areas indicate the three storms. Bottom panel: near bottom orbital velocities at the same locations (depths and lines are the same as the top panel). . . . . . . . . . . . . . . . .

Normalized significant wave height (by incident) along nondimensional distance (by wave length) for the first storm. The first panel corresponds to conditions on July 13th at 4am and sequence are elapsed 2 hours. The solid black line displays the computed normalized significant wave height including the dissipation due to P . oceanica meadow. The grey line corresponds to the Hs assuming no dissipation by the seagrass. Dashed lines displays the predictions for wave decay for a 15 % error in the measurement of initial wave height and wave period. . . . . . . .

Normalized significant wave height (by incident) along nondimensional distance (by wavelength) for the second storm. The first panel correspond to conditions on July 18th at 12am and sequence are elapsed 2 hours. The solid black line displays the computed normalized significant wave height including the dissipation due to P . oceanica meadow. The grey line corresponds to the Hs assuming no dissipation by the seagrass. Dashed lines displays the predictions for wave decay for a 15 % error in the measurement of initial wave height and wave period. . . . . . . .

Wave damping H/H0 along the transect for different conditions of shoot density. Initial wave height is Hs,0 = 1.56 m, ω = 1.38 s−1 and CD = 0.41. . . . . . . . . . . . . . . . . . . . . . . . . .

Decay coefficient (Kd ) versus seagrass shoot density (N) and leaf length (lv ) for an incident wave Hs,0 = 1.56 m, ω = 1.38 s−1 and CD = 0.41 propagating over constant depth h = 5 m (top), h = 10 m (middle) and h = 15 m (bottom). Units in m−2 . . . . . . .

Drag coefficient versus Keulegan-Carpenter number for experimental data -crosses-, for S´ anchez-G´onzalez et al. (2011) -black solid-, Kobayashi et al. (1993) -black dashed- and for M´endez and Losada (2004) -grey-. . . . . . . . . . . . . . . . . . . . . . .

A.1 Diagram of a shoaling wave approaching to shore. Varible names are defined in the text. . . . . . . . . . . . . . . . . . . . . . . . .

A.2 Relative depth and asymptotes to hyperbolic functions. . . . . . xviii

70

73

74

75

76

78

91 92

A.3 Vertical profile of velocity. Top, weak currents (black line) and fast currents (dashed line). Bottom, velocity reduction due to benthic macrophytes. . . . . . . . . . . . . . . . . . . . . . . . . .

xix

95

xx

List of Tables 2.1

3.1

3.2 3.3 3.4

4.1 4.2

Percent cover of bottomtype at different depths. The total area of coverage of each bottom type is presented in the last row. . . .

21

Two-factor way ANOVA testing differences between species and orbital velocities and minimum rooting length. Significant differences are expressed in bold as, ∗∗∗ p < 0.001, and ns = not significant. Cochran’s C-test no significant. df = degrees of freedom and MS = mean square. . . . . . . . . . . . . . . . . . . . .

37

Morphological characteristics of seedlings at the beginning of the field experiment (mean ± SE, n = 144). . . . . . . . . . . . . .

37

Computed near-bottom orbital velocities (cm s−1 ) at the experimental locations during the sampling periods (mean ± SD). . .

39

Results of three-factor way ANOVA of seedling survival percentage in October 2009. Significant differences are expressed in bold as ∗∗ p < 0.01 , ∗∗∗ p < 0.001, and (ns) not significant. Cochran’s C-test no significant. df = degrees of freedom and MS = mean square . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

40

Wave heights from model outputs and field measurements in Cala D’Or (means ± 95 % CL). . . . . . . . . . . . . . . . . . . . . . .

50

2

Chi-square (χ ) test of observed Caulerpa cover on different substrata. Caulerpa taxifolia has a critical χ2 p=0.001∗∗∗ , 2df = 13.81 and Caulerpa racemosa has a critical χ2 p=0.001∗∗∗ , 3df = 16.26. C. taxifolia was not found on rock substratum in enough quantities for the test. The type of contribution was classified as more than expected (M) or less than expected (L). . . . . . . . . xxi

55

4.3

Results of repeated measures of ANOVA performed to evaluate if the cover of Caulerpa taxifolia and Caulerpa racemosa was different between a model of dead matte of Posidonia oceanica and a sandy substratum. ∗ p < 0.05, ∗∗∗ p < 0.001, ns = not significant. Cochran tests were not significant. . . . . . . . . . . .

xxii

56

List of Symbols a= ab = av = a0v = cg = CD = E= Fx = fw = g= h= Hs = Hs,0 = k= ks = Kd = L= Ld = lv = lâ&#x2C6;&#x2014; = N = Re = Tp = u= ub = uc = x= Îľ=

wave amplitude orbital wave excursion plant surface area plant surface area per unit height group velocity drag coefficient wave energy force acting on the plant wave friction factor acceleration of gravity water depth significant wave height incident significant wave height wave number bottom roughness wave decay coefficient wave length minimum rooting length vegetation length characteristic length number of shoots per unit area Reynolds number wave peak period fluid velocity near-bottom orbital velocity characteristic velocity outside of the boundary layer horizontal distance relative roughness xxiii

= D = δ= Ď = Ď&#x201E;b = Î&#x2DC;= ν= νT = Ď&#x2030;=

non linear parameter rate of energy dissipation boundary layer seawater density bottom shear stress wave direction kinematic viscosity of water eddy viscosity wave angular frequency

xxiv

Chapter 1

Introduction Coastal marine ecosystems are home to a host of different species ranging from microscopic planktonic organisms that comprise the base of the marine food web (i.e., phytoplankton and zooplankton) to large marine mammals like whales. Coastal habitats alone account for approximately 1/3 of all marine biological productivity and those dominated by macrophytes (salt marshes, seagrasses, mangrove forests) are among the most productive regions on the planet.

Coastal zones provide humans with a wide variety of goods and services including foods, recreational opportunities, and transportation corridors. Despite the importance of marine ecosystems, increased human activities such as overfishing, coastal development, pollution, and the introduction of invasive species have caused significant damage and pose a serious threat to marine biodiversity.

Coastal zones are continually changing because of the dynamic interaction between the oceans and the land. Waves, currents and winds along the coast are both eroding rock and depositing sediment on a continuous basis, and rates of erosion and deposition vary considerably from day to day along such zones.

Other sediment inputs in the coastal zone are from river runoffs, glaciers and the inorganic remains of dead organisms. The energy reaching the coast can become high during storms, and such high energies make coastal zones areas of continuous morphological evolution (G´ omez-Pujol et al. 2011).

The shallow areas of the coastal zone (from 0 m to â&#x2C6;ź 50 m depth) are inhabited by benthic marine macrophytes. Macrophytes are large aquatic plants that grow in/or near water and are either emergent, submergent, or floating 1

2

Introduction

namely mangroves, marshes, seagrasses and macroalgae. These aquatic plants play an important role in coastal environments providing many ecological services (Costanza et al. 1997). Macrophytes are primary producers that supply oxygen and food, stabilize the seabed, provide nursery grounds, attenuate wave heights and current velocities preventing coastal erosion.

Since water is the medium in which macrophytes exist, any physical disturbance of it would be expected to influence them. Benthic macrophytes in nature are exposed to fluid motion where inertial forces are more important than viscous forces. Hydrodynamics affects almost all biological and chemical processes of macrophytes along their entire life cycle in many different ways such as spatial distribution (Frederiksen et al. 2004), nutrient uptake (Thomas et al. 2000, Cornelisen and Thomas 2004), pollen dispersion (Ackerman 1986), seed and seedlings transport (Koch et al. 2010), seedling survival (Rivers et al.

2011), etc. Hydrodynamics also drives sediment dynamics causing macrophytes to become buried or eroded during storm events (Fonseca et al. 1996, Cabaco et al. 2008). On the other hand, macrophytes are a source of drag which reduces flow velocities along the portion of the water column where vegetation it is present (Nepf and Vivoni 2002, Luhar et al. 2010). The presence of macrophytes reduces sediment re-suspension in the water column (Ward et al. 1984, Short and Short 1984, Terrados and Duarte 2000, Gacia and Duarte 2001).

Macrophytes have also shown to attenuate waves under laboratory (Fonseca and Cahalan 1996, M´endez and Losada 2004, Bouma et al. 2010) and natural conditions (Newell and Koch 2004, Bradley and Houser 2009).

Wave energy can have both positive and negative impacts on seagrass physiology, productivity and distribution. The periodic motion of the blades may enhance the water exchange between the seagrass meadow and overlying water column (Koch and Gust 1999). Nutrient uptake is enhanced in a seagrass meadow during wave conditions (Thomas and Cornelisen 2003). Stevens and Hurd (1997) suggest that the enhanced water flux derives from a periodic striping of the diffusive boundary layer by the wave motion. Wave-induced motion can also detach epiphytes from seagrass blades (Kendrick and Burt 1997). Waves can also have a negative effect on seagrass communities, by mobilizing sediment that can bury, erode and dislodge plant fragments (Marb`a and Duarte 1994, Paling et al. 2003, Frederiksen et al. 2004). Because of these negative influences, the spatial distribution of wave energy is believed to control, at least in

3 part, the spatial distribution of seagrasses (Koch and Beer 1996, Koch 2001, Frederiksen et al. 2004, Koch et al. 2006).

Some macrophytes populations show extensive spatial and temporal fluctuations and landscapes are under continuous transformation due to disturbances (Fonseca and Bell 1998). Physical disturbances are considered one of the main extrinsic factors controlling the spatial structure and species diversity of seagrass meadows (Clarke and Kirkman 1989, Duarte et al. 1997). Wind-generated wave dynamics and tidal currents create sediment movement, which may either bury plants, expose roots and rhizomes or during heavy storms even uproot entire plants (Kirkman and Kuo 1990, Preen et al. 1995). Depending on the degree of wave exposure, seagrass communities form characteristic landscapes ranging from highly fragmented to almost continuous meadows covering extensive areas (Frederiksen et al. 2004). In sheltered areas with calm hydrodynamic conditions, seagrass patches have been observed to be predominantly circular, while if the area is exposed to wave action and tidal currents, seagrass patches are more elongated (Den Hartog 1971). These patterns are not only present in seagrass communities but also on some coral reefs, freshwater plants and mussels (Den Hartog 1971).

When a seagrass meadow is disturbed, gaps are formed presenting new unoccupied sites that are potentially available for colonization. These new sites encompass several substrata including rock, sand and dead matte. Seagrass seedlings and invasive macroalgae are known to colonize these areas depending on the substratum type (Balestri et al. 1998, Hill et al. 1998, Rivers et al. 2011). Assuming that physical dislodgement is the primary mechanism for macrophytes failure, a favourable area would be one that reduces physical disturbance or enhances anchoring ability. Also, macrophytes morphology and rooting capacity will determine their success during physical disturbances. Despite all these evidences, to date there has been little effort to correlate wave and current energy with the spatial distribution and growth of seagrass meadows, seagrass seedlings and invasive macroalgae.

Gaps in knowledge of marine macrophytes are to date still evident. For instance, the effects of wave energy on the distribution at a landscape scale of plants in many cases are based on qualitative evidence. Besides, the tolerance of seedlings to wave energy and the capacity of seedlings to colonise a specific

4

Introduction

coastal environment are very scarce. The importance of the Mediterranean seagrass P. oceania on coastal protection as well as a reservoir of biogenic sediments has not been established quantitatively. In this Thesis, it is investigated the interaction between the wave induced hydrodynamics and some of the above mentioned points.

1.1

Upper limit distribution of macrophytes

The study of the light requirements for macrophytes growth has been a major focus of research in marine ecology and different quantitative models provide predictions of the lower depth limit of the distribution of seagrasses (Dennison 1985, Dennison 1987, Duarte 1991, Kenworthy and Fonseca 1996, Koch and Beer 1996, Greve and Krause-Jensen 2005). Seagrasses can thrive up to depths where the irradiance at the top of the leaf canopy is higher than 11 % of surface irradiance (Duarte 1991) or where the number of hours with values of irradiance higher than those that saturate the photosynthesis rate is higher than 6-8 hours (Dennison 1985). The upper depth limit of distribution of seagrasses has been related to their tolerance to desiccation at low tide (Koch and Beer 1996) and to ice scour (Robertson and Mann 1984). In seas with low tidal range and no ice formation like the Mediterranean, the upper depth limit of seagrasses will be determined mainly by their tolerance to wave energy (Koch et al. 2006).

However, quantitative estimates of the wave energy that sets the upper depth limit of seagrasses are to date still scarce.

Conceptual models have been proposed to explain the differences in shape, bottom relief, and cover of seagrass meadows (Fonseca et al. 1983, Fonseca and Kenworthy 1987) or the depth distribution of intertidal seagrass (van Katwijk et al. 2000) as a function of wave energy and/or current velocity. Local hydrodynamics (indirectly estimated from the weight loss of clod cards) has been related to depth and seagrass presence and used to identify the habitat requirements of South Australian seagrass species (Shepherd and Womersley 1981). Keddy (1984) developed a relative wave exposure index (REI) in order to quantify the degree of wave exposure by using wind speed, direction and fetch measurements.

REI values have been correlated to different attributes of seagrass meadows such as the content of silt-clay and organic matter of the sediment, seagrass cover, shoot density (Fonseca and Bell 1998, Fonseca 2002, Krause-Jensen 2003, Frederiksen et al. 2004) and biomass (Hovel 2002). Shallow seagrass popula-

1.2 Hydrodynamics and seedling survival

5

tions tend to be spatially more fragmented in wave-exposed environments than in wave-sheltered ones (Fonseca and Bell 1998, Frederiksen et al. 2004) and temporal changes of seagrass cover are also higher at the most wave-exposed sites (Frederiksen et al. 2004). Plaster clod cards and REI provide only semiquantitative idiosyncratic estimates of current and/or wave energy and, therefore, comparison between studies is difficult. Additionally, REI estimates do not consider the influence of depth on wave damping (but see van Katwijk and Hermus 2000). Direct quantitative measurements of the energy of waves and currents impinging on seagrasses are necessary to elucidate its effects on them, to identify their habitat requirements, to predict their spatial distribution and their response to both natural (i.e., storms, hurricanes, etc) and anthropogenic (i.e., dredging and beach nourishment, coastal development, etc) disturbances.

Posidonia oceanica (L.) Delile, an endemic seagrass species of the Mediterranean Sea, grows between the depths of 0.5 m and 45 m (Procaccini et al.

2003) and covers an estimated surface area between 2.5 and 5 million hectares, about 1-2 % of the 0-50 m depth zone (Pasqualini et al. 1998). P. oceanica is included in the Red List of marine threatened species of the Mediterranean (Boudouresque et al. 1990) and meadows are defined as priority natural habitats on the EC Directive 92/43/EEC on the Conservation of Natural Habitats and of Wild Fauna and Flora (EEC, 1992). Fundamental knowledge of these seagrasses species in nature is required to understand their interactions with hydrodynamics such as the maximum flow velocities that are able to stand. Some studies have look at the effect of P. oceanica on flow and the horizontal spatial distribution of plants with depth (Granata et al. 2001) but little is known about the influence of hydrodynamics on P. oceanica distribution.

1.2

Hydrodynamics and seedling survival

Seagrass seedlings are the elements that form new meadows and their survival in nature contributes to seagrass population dynamics. Reproduction strategies of seagrasses are diverse, for example buoyant fruits of P. oceanica will be dispersed by wind and currents while C. nodosa fruits will remain buried next to the parent plants (Orth et al. 2006). Waves and currents can strongly influence the spatial distribution of meadows (Fonseca and Bell 1998, Frederiksen et al.

2004), suggesting that hydrodynamics may be also important for seed dispersal

6

Introduction

and seedling establishment (Koch et al. 2010). Measurements showing higher seedling survival rates at deeper locations suggest the same (Piazzi et al. 1999).

Moreover, it has been shown that hydrodynamics can directly affect the survival of seagrass seedlings transplants (van Katwijk and Hermus 2000, Rivers et al.

2011). The effect of waves and currents on seedling establishment is expected to be through sediment movement, as this can cause burial or dislodgment of small seagrass plants.

Seagrass capacity to withstand sediment burial is strongly dependent on size and morphology (Idestam-Almquist and Kautsky 1995, Cabaco et al. 2008).

The effect of sediment erosion on survival could be determined by the root capacity to remain anchored to the substratum (Madsen et al. 2001). Hence, seedlings may be expected to be highly vulnerable to sediment movement compared to adult seagrasses, as seedling roots will penetrate less deeply into the sediment. Although the effect of sediment burial on seagrasses has been reviewed (Cabaco et al. 2008), to our knowledge, little work has been devoted to the effect of sediment dynamics, in particular the effect of sediment erosion and root anchoring capacity on seedling survival. In this Thesis this question is addressed for the two most common seagrass species in the Mediterranean Sea (Green and Short 2003) that both strongly declined during the 20th century (Boudouresque 2009): Posidonia oceanica and Cymodocea nodosa.

Both species have morphological differences that could affect their interaction with hydrodynamics and their survival capacity. P. oceanica leaves are longer and wider than C. nodosa leaves (Buia and Mazzella 1991, Guidetti 2002). These differences will affect the drag force exerted on the leaves and thus the anchoring capacity of the roots. Wicks et al. (2009) suggested that Zostera marina seedlings with the same root length but higher leaf area would have less chance to survive by being dislodged from the sediment than seedlings with lower leaf area. As P. oceanica has a higher total leaf surface area than C. nodosa (Guidetti 2002), the drag experienced on the leaves will be higher in P. oceanica than in C. nodosa. The assessment of the capacity of seedlings to remain anchored to the substrata under different hydrodynamic conditions is necessary to get a fundamental understanding of seedling survival in nature.

1.3 Substratum type and invasive species

1.3

7

Substratum type and invasive species

Consolidated substrata (i.e., rocks) offer a stable, non-mobile surface where macroalgae and some seagrass species attach effectively. Unconsolidated substrata (i.e., sand, mud) are unstable, mobile and only seagrasses and macroalgae with root-like structures can colonize them. Waves and currents can mobilise sediment particles producing sediment erosion and accretion that may affect seagrasses negatively through uprooting and burial (Fonseca and Kenworthy 1987, Williams 1988, Terrados 1997, Cabaco et al. 2008). Thresholds of maximum wave energy tolerated by some seagrasses species have been determined (Fonseca and Bell 1998, van Katwijk and Hermus 2000, Koch 2001) as well as the relationship between wave exposure and temporal variability of seagrass coverage in shallow sands (Fonseca et al. 1983, Fonseca and Bell 1998, Frederiksen et al. 2004). In contrast, information about the wave energy levels that macroalgae tolerate is limited, especially in unconsolidated substrata (D’Amours and Scheibling 2007, Scheibling and Melady 2008).

Most marine macroalgae grow over consolidated substrata such as rocks or over other macrophytes but a few Chlorophyta species of the order Caulerpales are also able to grow on unconsolidated substrata (Taylor 1960). The thallus of Caulerpa species is composed of a creeping portion, the stolon, that it is attached to the substrata by root-like structures, the rhizoids, and an erect portion, the fronds that have different shapes depending on the species (Taylor 1960, Bold and Wynne 1978). The rhizoids of Caulerpa are able to bind sediment particles (Chisholm and Moulin 2003) and anchor the plant in sandy sediments or other unconsolidated substrata, and to attach to rocks and other macrophytes (Taylor 1960, Meinesz and Hesse 1991, Klein and Verlaque 2008).

Hence Caulerpa species may be present both in wave exposed, rocky bottoms and sheltered, sandy-muddy sediments (Thibaut et al. 2004).

Macroalgae have been introduced to non-native locations around the world through human activities including aquaculture, shellfish farming, aquariums, shipping (ballast water and hull fouling), fishing gear and as lessespian inmigrants (migration through the Suez Canal) (Ru´ız et al. 2000). Decline of seagrass meadows within the Mediterranean Sea often results in the replacement of seagrasses by opportunistic green macroalgae of the Caulerpa family. Caulerpa taxifolia and Caulerpa racemosa have been introduced in the Mediterranean Sea within the last 25 years showing an invasive behaviour (Meinesz and Hesse 1991,

8

Introduction

Boudouresque and Verlaque 2002, Klein and Verlaque 2008). Both Caulerpa species are strong competitors which tend to eliminate native species forming monospecific beds (Verlaque and Fritayre 1994, Piazzi et al. 2001). C. racemosa is able to overgrow other macroalgal species, reaching virtually 100 % cover in invaded areas (Piazzi et al. 2001, Balata et al. 2004).

Initial observations of C. taxifolia invasion suggested that rocks and dead matte of P. oceanica were appropriate for the establishment of this species but sandy or muddy sediments in sheltered conditions were also colonized by it (Hill et al. 1998). A short-term experiment showed that algal turfs were more favourable for the establishment of C. taxifolia and C. racemosa than other macroalgal communities on rocky bottoms (Ceccherelli et al. 2002). Both Caulerpa species seem to tolerate a certain level of sediment deposition and burial (Glasby et al. 2005, Piazzi et al. 2005) a feature that likely facilitates their development in sandy sediments. Elucidation of the factors influencing the establishment and spread of introduced Caulerpa species is essential to identify benthic communities susceptible of being invaded and to have scientific tools to predict future invasions.

1.4

Wave damping by macrophytes

Submerged plants increase bottom roughness and extract momentum from water resulting in flow velocity reduction and less energy for sediment transport (Koch et al. 2006). In addition to the effect of leaf, canopies on water flow, seagrass rhizomes and roots extend inside sediment contribute to its stabilization (Fonseca 1996). Flume and in situ measurements have shown that water velocity is reduced inside seagrasses. In sparse canopies, turbulent stress remains elevated within the canopy while in dense canopies turbulent stress is reduced by canopy drag near the bed (Luhar et al. 2008). The reduction in velocity due to seagrass canopies is lower for wave-induced flows compared to unidirectional flows because the inertial term can be larger orcomparable to the drag term in oscillatory flow (Lowe et al. 2005, Luhar et al. 2010). Except for intertidal systems, where currents are dominant, most seagrass meadows are in wave-dominated habitats. Interaction between seagrass canopies and oscillatory flow has been, however, much less studied than that with currents. Nearbed turbulence inside seagrass canopies are lower than those on sands under wave-generated oscillatory flows (Granata et al. 2001). Wave energy and sediment resuspension are

1.5 Aims and motivations

9

also reduced by seagrasses (Terrados and Duarte 2000, Verduin and Backhaus 2000, Gacia and Duarte 2001), but water exchange between inside and outside the canopy increases (Koch and Gust 1999).

Wave attenuation by macrophytes has been studied in the laboratory (Fonseca and Cahalan 1992, Kobayashi et al. 1993, Augustin et al. 2009), in the field (Knuston et al. 1982, Newell and Koch 2004, Bradley and Houser 2009), and using analytical methods or numerical models (Kobayashi et al. 1993, M´endez et al. 1999, M´endez and Losada 2004, Chen et al. 2007). Wave attenuation by seagrass canopies has been detected only in shallow systems where canopies occupy a large fraction of the water column (Fonseca and Cahalan 1992, Koch 1996, Chen et al. 2007, Bradley and Houser 2009). Posidonia oceanica is the dominant seagrass species in the Mediterranean Sea, where it forms extensive meadows in depths up to 45 m (Procaccini et al. 2003) and canopy occupies less than 20 % of water column height. Although commonly assumed (Luque and Templado 2004, Boudouresque et al. 2006), wave attenuation by P. oceanica meadows has never been accurately assessed in the field.

1.5

Aims and motivations

The aim of this Thesis is to obtain quantitative knowledge as well as a predictive capacity about the responses of seagrasses and macroalgae to the bottom shear stress caused by waves. First, the role of near-bottom orbital velocities on the spatial distribution of a P. oceanica meadow, and on the survival of P. oceanica and C. nodosa seedlings are evaluated. Second, the importance of substratum type and hydrodynamics on the distribution and persistence of two marine invasive macroalgae, Caulerpa taxifololia and Caulerpa racemosa is explored. Finally, the wave damping effect of a seagrass meadow of P. oceanica on wave propagation under natural conditions is quantified and a relationship between the wave friction factor which integrates the seagrass landscape and drag coefficient exherted on individual plants is derived.

The wave energy impinging on seagrass meadows is assesed in a pilot area and presented in Chapter 2. Following a methodology that combines numerical modelling and geographical information systems, GIS, (Fig. 1.1) a quantitative and testable relationship between wave energy and the upper depth limit

10

Introduction

of the Mediterranean seagrass P. oceanica is obtained. Wave energy estimation is obtained by means of the analysis of forty-four years of wave data. We considered that long-term historical wave data, rather than present-day wave measurements, are more appropriate to link wave energy to the presence of P. oceanica given the low rates of vegetative growth and space occupation of this seagrass species (Marb` a and Duarte 1998). Aerial photographies and bathymetric data are used for an accurate mapping of the meadow as well as to delimitate the upper depth limit of P. oceanica.

Figure 1.1: Schematic diagram of the approach presented.

The capacity of seedlings of P. oceanica and C. nodosa to remain anchored to the substrata under different hydrodynamic conditionsis evaluated in Chapter 3. The underlying mechanisms that can affect seedling survival are studied in a biological flume. Drag measurements are carried out to obtain the effective drag forces acting on the seedlings and compute the drag coefficient under unidirectional and oscillatory flows. Further, the minimum root lengths necessary to remain anchored to the sediment are measured under different wave conditions on the flume. Additionally, a replicated short-term field experiment is carried out to compare the survival of seedlings at two depths (18 m and 12 m) to assess the effect of exposure to hydrodynamics on the survival of P. oceanica and C.

1.5 Aims and motivations

11

nodosa seedlings.

The effect of substratum on the persistence of two invasive Caulerpa species is evaluated in Chapter 4. First, the presence of C. taxifolia and C. racemosa is quantified in different types of substrata during two consecutive years testing if their presence was independent of the type of substratum. Then, a shortterm experiment is performed where fragments of C. taxifolia and C. racemosa are established in a model of dead matte of Posidonia oceanica and in sand.

The cover of both Caulerpa species is correlated to near-bottom orbital velocities and friction coefficients over sand and the model dead matte are computed.

The effect of a seagrass meadow on wave propagation under natural conditions is evaluated in Chapter 5 by a novel approach that relates individual plant parameters with meadow parameters. First, wave heights and orbital velocities in a P. oceanica meadow are measured during three storms. Data for one of the three storms measured are used to obtain the roughness length for the meadow.

The friction factor for the two other events are then calculated and the drag coefficient related with meadow parameters. Most coastal models, introduce bottom effects through the roughness length and therefore it is worthwhile to obtain such magnitude for P. oceanica. Finally, the wave decay for different seagrass meadow densities and leaf lengths is predicted.

For completeness, a background review of fluid dynamics is presented in the Appendix A, where the basic concepts of linear wave theory, wave propagation and benthic boundary layers are summarised.

12

Introduction

Chapter 2

Wave energy and the upper depth limit distribution of Posidonia oceanica Parts of this chapter has been published as an article on Botanica Marina, Special Issue on Mediterranean Seagrasses1 .

2.1

Summary

It is widely accepted that the availability of light sets the lower limit of the bathymetric distribution of seagrasses while the upper limit depends on the level of disturbance driven by currents and waves. The establishment of light requirements for seagrass growth has been a major focus of the research in marine ecology and therefore different quantitative models provide predictions of the lower depth limit of the distribution of seagrasses. In contrast, the study of the influence of energy levels on the establishment, growth and maintenance of seagrasses have received less attention and to date there are no quantitative models predicting the evolution of seagrasses as a function of the hydrodynamics 1 Infantes E, Terrados J, Orfila A, CaË&#x153; ´ nellas B, Alvarez-Ellacuaria A (2009) Wave energy and the upper depth limit distribution of Posidonia oceanica. Botanica Marina 52: 419-427.

13

14

Wave energy and the upper depth limit of P. oceanica

at large scales. Hence it is not possible to predict either the upper depth limit of the distribution of seagrasses or the effects that different energy regimes will have on them. The aim of this Chapter is to provide a comprehensible methodology to obtain quantitative knowledge as well as predictive capacity to estimate the upper depth limit of the bathymetric distribution of seagrasses as a response of the wave energy disposed at the seabed. The methodology has been applied using forty-four years of wave data from 1958 to 2001 in order to obtain the mean wave climate at deep water in front of an open sandy beach in the Balearic Islands, Western Mediterranean where the seagrass Posidonia oceanica forms an extensive meadow. Mean wave conditions were propagated to the shore using a 2D parabolic model over the detailed bathymetry. The resulting hydrodynamics has been correlated with bottom type and the distribution of P. oceanica. Results showed a predicted near-bottom orbital velocity determining the P. oceanica upper depth limit between 38 - 42 cm sâ&#x2C6;&#x2019;1 . This Chapter shows the importance of the interdisciplinary effort in ecological modelling and in particular the need of hydrodynamical studies to elucidate the distribution of seagrasses at shallow depths. Besides, the use of predictive models would permit the evaluation of the effects of coastal activities (construction of ports, artificial reefs, beach nourishments, dredging) on the benthic ecosystems.

2.2 2.2.1

Material and methods Study Area and Regional Settings

The analysis was carried out in Cala Millor, located on the northeast coast of Mallorca Island (Fig. 2.1a,b). The beach is in an open bay with an area of ca.

14 km2 . Near the coast to 8 m depth, there is a regular slope indented with sand bars near the shore (Fig. 2.1c), these bars migrate from offshore to onshore between periods of gentle wave conditions. At depths from 6 m to 35 m, the seabed is covered by a meadow of P. oceanica. This area was chosen for this study because of the availability of data from previous studies. The tidal regime is microtidal, with a spring range of less than 0.25 m (G´omez-Pujol et al. 2007).

The bay is located on the east coast of the island of Mallorca and it is therefore exposed to incoming wind and waves from NE to ESE directions. According to the criteria of Wright and Short (1983), Cala Millor is an intermediate barred sandy beach formed by biogenic sediments with median grain values ranging between 0.28 and 0.38 mm at the beach front (L. G´omez-Pujol et al. 2007).

2.2 Material and methods

15

Figure 2.1: (a) Location of Mallorca in the Mediterranean Sea. (b) Location of the study area of Cala Millor in Majorca. The asterisk (* indicates HIPOCAS node 1433. (c) Bathymetry of Cala Millor with isobaths (in meters).

2.2.2

HIPOCAS database (1958 - 2001) and deep water wave characterization

Wave data used are part of the Hindcast of Dynamic Processes of the Ocean and Coastal Areas of Europe (HIPOCAS) project. This database consists of a high resolution, spatial and temporal, long-term hindcast dataset (Soares et al.

2002, Ratsimandresy et al. 2008). The HIPOCAS data were collected hourly for the period 1958 to 2001, providing 44 years of wave data with a 0.125 spatial resolution. This is the most complete wave data base currently available for the Mediterranean Sea. These hindcast models have become powerful tools, not only for engineering or predictive purposes, but also for long-term climate studies (CaË&#x153; nellas et al. 2007). These data were produced by the Spanish Harbor Authority by dynamical downscaling from the National Center for Environ-

16

Wave energy and the upper depth limit of P. oceanica

Figure 2.2: Directional wave histogram for HIPOCAS node 1433 10 km from Cala Millor

mental Prediction (NCEP) and the National Center for Atmospheric Research (NCAR) global reanalysis using the regional atmospheric model REMO. Data from HIPOCAS node 1433 (see Fig. 2.1) located 10 km offshore at 50 m depth, which is the closest HIPOCAS node to Cala Millor was used. The long-term distribution of significant wave height and wave direction at this node (Fig. 2.2) shows that the most energetic waves usually come from N-NNE. These wave directions are also the most frequent during the 44-year dataset.

Data contained in the HIPOCAS node consist of a set of sea states (one per hour) defined by their significant wave height, spectral peak period, and direction. An estimation of the long-term distribution of the mean significant wave height (Hm ) and its standard deviation was carried out making use of the lognormal probability distribution (Castillo et al. 2005). The long-term probability distribution identifies the most probable sea state for the 44-year period. Before estimation of this wave regime, data were preselected taking into account their incoming directions. Only sea states directed towards the beach were included in the analysis. Mean wave climate for HIPOCAS node 1433 provides an Hm value of 1.53 m Âą 0.96 m ( Âą 1 SD), a peak period of 7.3 s, and a direction of 11.258. This mean (most probable) wave climate was propagated to the shore using a parabolic model.

2.2 Material and methods

2.2.3

17

Shallow-water wave conditions

As water waves propagate from the region where they are generated to the coast, both wave amplitude and wavelength are modified. The surf zone is a highly dynamic area where energy from waves is partially dissipated through turbulence in the boundary layer and transformed into short and long waves, mean sea level variations, and currents (Dean and Dalrymple 2002). In the present work, waves were propagated using a gentle slope parabolic model (OLUCAMC), which includes refraction-diffraction effects as well as energy losses due to wave breaking (Kirby and Dalrymple 1983, GIOC 2003). Detailed bathymetry obtained with echo-sounding was used to generate the numerical mesh. The model solves continuity and momentum equations assuming a smooth bottom (e.g., variations of the bottom negligible within a wave length) and converting the hyperbolic system to a parabolic system (e.g., with wave propagation in one direction).

Two grids were generated: the external grid (122 x 81 nodes), which covers the deeper area with 75 m resolution between nodes, and the internal grid (140 x 311 nodes), which covers the shallow area with 15 m resolution. The model output provides the wave field (significant wave height and direction of the mean flux energy) in the whole grid.

Maximum near-bottom orbital amplitudes ab were calculated following the linear wave theory (see Appendix A): ab =

Hs 2sinh(2πh/L)

(2.1)

where h is the water depth and L is the wave length calculated iteratively as: L=

Tp2 2πh tanh( ) 2π L

(2.2)

where Tp is the peak period and g is the acceleration of gravity. The maximum near-bottom orbital velocity ub ) is: ub = 2πab /Tp

(2.3)

18

2.2.4

Wave energy and the upper depth limit of P. oceanica

Bottom typology and bathymetry

Remote sensing of the seabed from air or space is commonly used for mapping seagrass habitats over a wide range of spatial scales (McKenzie et al. 2001).

Satellite spectral images from Ikonos are suitable for the detection and mapping of the upper depth limit of seagrass distribution in shallow clear waters (Fornes et al. 2006). Aerial color photographs have been used in some studies to describe temporal changes in the distribution of seagrasses (Hine et al.

1987), to evaluate the effect of wave exposure (Frederiksen et al. 2004), and the influence of anthropogenic activities (Leriche et al. 2006). Aerial photographs for mapping the area covered by P. oceanica, dead P. oceanica rhizome, rocks, and sand up to 11 m depth were used. Aerial color photographs were taken in August of 2002 with a resolution of 0.4 m. Polygons were drawn around the different areas with Arc/GIS software (Arc/Info and Arc/Map v9.0, ESRI) and classified by bottom types. In those areas where bottom recognition was not possible from the aerial photographs, field surveys were carried out to identify the typology of the substratum (some areas tend to accumulate seagrass leaves which can lead to false interpretation of aerial images). Image classification was validated with bathymetric filtered echo-soundings and field observations.

During 2005, an acoustic survey was carried out to determine the bathymetry of the inner mesh and to test the classification of the seagrass coverage from the aerial photographs. Acoustic mapping of Posidonia oceanica was performed with a shipmounted Biosonics DE-4000 echo sounder (BioSonics, Inc., Seattle, USA) equipped with a 200 KHz transducer. The draught of the boat allowed sampling up to depths of approximately 0.5 m. Inshore-offshore echo-sounding transects were sampled perpendicularly to the bathymetric gradient, with a separation of 50 m between transects. Acoustic pulse rate was set to 25 sâ&#x2C6;&#x2019;1 and the sampling speed was set to 3 knots, which allowed for a horizontal resolution of 1 m (Orfila et al. 2005). Bottom typology was estimated as the most probable after echogram examination using the first to second bottom echo ratio technique (Orlowski 1984, Chivers et al. 1990). The resulting echo sounding points were filtered, averaged (1 output equals 20 pings) and clustered into three groups (P. oceanica meadows, sandy, and hard bottoms) taking into consideration calibrations performed for previously classified bottoms. Hard bottoms include rocks, P. oceanica rhizome mats, and zones with poor seagrass coverage (Fig. 2.3).

Afterwards, this map was verified by direct observation at random points, also distinguishing those bottom types that the algorithm was not able to identify

2.3 Results

19

(i.e., dead rhizome and rocks). The map of bottom typology in Fig. 2.3 represents the final classification (aerial photograph verified with echo-sounding and direct SCUBA observations).

Bathymetric data were interpolated using the kriging technique to create 1-m scale depth contours that were overlaid with bottom typology. Similarly, maximum near-bottom orbital velocity data were interpolated to create 5 cm sâ&#x2C6;&#x2019;1 scale ub contours that were overlaid with bottom typology. Percent coverage of rocks, sand, dead rhizomes of P. oceanica, and P. oceanica were calculated for each depth and near-bottom orbital velocity interval. A total of 400 points were randomly selected throughout the study area, and the corresponding bottom type and near-bottom orbital velocity interval were used to calculate the average near-bottom orbital velocity for each bottom type. Additionally, 400 points along the upper limit of the P. oceanica meadow were randomly selected and the corresponding near-bottom orbital velocities were used to estimate an average along that edge. Differences in near-bottom orbital velocity between each bottom type and the upper limit of P. oceanica were evaluated using Kruskal-Wallis non-parametric analysis of variance (due to heterogeneity of data variance).

2.3

Results

The study site has a total area of 121.44 ha wherein sand and Posidonia oceanica meadow are the most abundant substrata. Shallow bottoms in Cala Millor (depths between 0 and 6 m) are mostly sandy, with some patches of rock, particularly in the center and south parts of the beach (Table 2.1). The upper depth limit of P. oceanica is located between 5 and 6 m and the meadow continues down to 30 - 35 m depth (acoustic survey data not shown). The limit between seagrass and sand is irregular and has several areas of hard substratum between 4 and 7 m that correspond to dead rhizomes of P. oceanica partly covered by sand and algae, which can be indicative of meadow regression (Fig. 2.3). P. oceanica reaches the highest percentage cover at depths greater than 8 m, and no stands are found in depths shallower than 4 m. Most of the rocky bottom is in the shallowest water, between 0 and 2 m. Two large sand fingers cross the seagrass meadow at the center of the study area; these may have resulted from sand transport from the exposed beach after storm events.

20

Wave energy and the upper depth limit of P. oceanica

Figure 2.3: Bottom typology off Cala Millor (Mallorca, western Mediterranean Sea).

2.3 Results

21

Mean wave conditions propagated over the beach resulted in wave heights between 0.2 and 0.4 m, with significant wave breaking in the shallow sandy area (between 0.5 and 1 m) (Fig. 2.4a) and near-bottom orbital velocities up to 110 cm s−1 . The highest velocities occurred in the shallow sandy part of the beach (between 1 and 3 m) (Fig. 2.4b).

Depth (m) 1−2 2−3 3−4 4−5 5−6 6−7 7−8 8−9 9 − 10 10 − 11 Total area (m2 )

P. oceanica 0 0 0 0.46 11.55 42.34 73.87 83.82 94.40 97.98 565, 349.42

Sand 96.63 97.73 97.31 83.36 51.03 25.27 21.10 14.64 5.59 2.02 528, 241.5

Rock 3.24 1.75 0 0 0 0 0 0 0 0 25,192.7

Dead P. oceanica 0.13 0.52 2.69 16.17 37.42 32.39 5.03 1.54 0 0 95,632.92

Table 2.1: Percent cover of bottomtype at different depths. The total area of coverage of each bottom type is presented in the last row.

Sandy areas are associated with high values of near-bottom orbital velocities, while P. oceanica is associated with lower velocities (Fig. 2.5). P. oceanica is not present in areas with velocities higher than 38 - 42 cm s−1 . This velocity interval might be considered a first estimate of the threshold near-bottom orbital velocity that allows the P. oceanica to occur in Cala Millor.

Variance of near-bottom orbital velocity was higher in sand, rock, and dead Posidonia oceanica than in the P. oceanica meadow (Fig. 2.6). Kruskal-Wallis non-parametric analysis of variance detected significant differences in the average near-bottom orbital velocity between bottom types [H (4, n = 800) = 310.34, p < 0.001], and post-hoc multiple comparisons of mean ranks showed that velocities were lower in P. oceanica stands and at the P. oceanica upper limit than in rock, sand, and dead P. oceanica (Fig. 2.6).

22

Wave energy and the upper depth limit of P. oceanica

Figure 2.4: (a) Distribution of wave heights (m) at the beach derived from mean wave conditions (11.25o , Hm = 1.53 m, Tp = 7.30 s). (b) Distribution of nearbottom orbital velocities (ub , m sâ&#x2C6;&#x2019;1 ) at the beach and cover of the P. oceanica meadow (gray) and dead rhizomes (dark gray).

2.4

Discussion

In this Chapter, a methodology to estimate the wave energy that determines the upper depth limit of Posidonia oceanica is presented. Data show that an increase in wave energy is related to a decrease of P. oceanica cover, and that above a threshold level of wave energy seagrass is not present. It is important to emphasize that the evidence provided in this study is correlative and applies to this study site only. Other locations, with different sediment characteristics and wave climates might provide different threshold values. Additional sources of disturbance (both natural and anthropogenic) will also introduce variability

2.4 Discussion

23

Figure 2.5: Percent coverage of the different bottom types in each of the nearbottom orbital velocity (ub ) intervals established in Cala Millor.

Figure 2.6: Mean near-bottom orbital velocity above each bottom type in Cala Millor. Different capital letters indicate significant differences between bottom types (post-hoc multiple pairwise comparison of mean ranks, p < 0.05). Error bars show 95 % confidence intervals. n-values indicate the number of points selected randomly in each bottom type. Differences in n-values between bottom types are driven by the differences in percentage covers of each bottom type in the study area.

24

Wave energy and the upper depth limit of P. oceanica

in the threshold estimates. The near-bottom orbital velocities are computed from numerical model predictions and real velocities within the meadow could be lower due to wave attenuation by the seagrass meadow and wave-current interactions. However, the predicted velocities where P. oceanica is not present provide an estimate of the threshold velocities that would impede the persistence of this seagrass species. The usefulness of this methodology is that it provides quantitative estimates of wave energy that sets the upper depth limit of Posidonia oceanica and, therefore, these can be compared to those obtained in other locations.

An estimate of the threshold value of near-bottom orbital velocity that allows the long-term persistence of Posidonia oceanica (38 - 42 cm s−1 ) is estimated.

It has been suggested that Zostera marina L. can tolerate unidirectional current velocities up to 120-150 cm s−1 and that the meadows formed by this species become spatially fragmented at tidal current velocities of 53 cm s−1 (Fonseca et al. 1983). Shallow mixed meadows of Z. marina and Halodule wrightii Asch. seem to remain spatially fragmented at tidal current speeds above 25 cm s−1 (Fonseca and Bell 1998). Experimental transplantations of Z. marina along depth gradients in intertidal zones indicate that this species cannot persist at sites where the maximum bottom orbital velocity during the tidal cycle reaches 53 - 63 cm s−1 . Furthermore, this species can survive when exposed to waves for less than 60 % of the time and maximum orbital velocity is less than 40 cm s−1 (van Katwijk and Harmus 2000). Hence, this estimate of the threshold value of near-bottom orbital velocity that allows the persistence of P. oceanica is within the range of values of current velocity proposed for other seagrass species.

Frederiksen et al. (2004) used aerial photographs to follow changes in the distribution of Zostera marina from 1954 to 1999 and showed that seagrass landscapes can change extensively over long periods of time, especially in the more wave-exposed areas. Comparison of aerial photographs taken in 1956 and 2004 indicates that the upper depth limit of Posidonia oceanica in Cala Millor has regressed in the south part of the beach (IMEDEA 2005), which is a sector of the beach well exposed to the most energetic waves (those from the N-NNE). However, that other processes may determine the upper depth limit of P. oceanica are not rule out. This seagrass species is able to grow in certain locations, usually sheltered, almost to sea level (Ribera et al. 1997), which suggests that neither photo-inhibition nor temperature fluctuations associated with

2.4 Discussion

25

shallow depths are important in setting the upper depth limit. Except for the desiccation effects associated with exposure to air at low tides (negligible in the Mediterranean Sea), the consensus is that the upper depth limit of seagrasses is set by the energy of currents and waves (Koch et al. 2006). Wave breaking might have a role in establishing the upper limit, but the position at which waves break will change depending on wave height, wave period, and presence of submerged sandbars. In addition, the effect of waves of certain energy may depend on the amount of time the plants are exposed to different levels of wave energy. Average wave conditions have been used to estimate the average field of near-bottom orbital velocities, but other wave statistics could be used, such as extreme events (De Falco et al. 2008), and elaborate a prediction of those velocities that allow the long-term persistence of P. oceanica. 44 years of data are used, as this would be an adequate period of time over which to average the effects of waves on the slow-growing P. oceanica (Marb`a and Duarte 1998).

Sediment density and grain size characteristics have not been considered in this study. These granulometric characteristics may play an important role in determining the location of seagrasses, because wave energy will move and resuspend the sediment, which might bury or erode a seagrass meadow. Differences in sediment characteristics between locations will then contribute to the variability of the estimate of the threshold wave energy that sets the upper depth limit of Posidonia oceanica.

I conclude that the methodology presented here is a useful tool to estimate wave energy on the bottom and to identify the level that sets the upper depth limit of Posidonia oceanica meadows. This approach provides testable estimates of the threshold level of wave energy that allows this species to persist and indicates a research program to validate them. This Chapter also highlights the importance of interdisciplinary and multidisciplinary research in ecological modeling and, in particular, the need for hydrodynamical studies to elucidate the distribution of seagrasses at shallow depths.

26

Wave energy and the upper depth limit of P. oceanica

Chapter 3

Posidonia oceanica and Cymodocea nodosa seedling tolerance to wave exposure Part of this Chapter has been submitted for publication as an article on Limnology and Oceanography1

3.1

Summary

In this Chapter the role of hydrodynamics in the establishment of seagrass seedlings for two Mediterranean seagrass species, Posidonia oceanica and Cymodocea nodosa is studied, by combining flume and field experiments. Flume measurements under both unidirectional and oscillatory flow showed that P. oceanica seedlings experienced higher drag forces than C. nodosa, which could be related to the larger total leaf area. Drag coefficients were between 0.01 and 0.1, for Reynolds numbers of 103 and 105 . As a result, P. oceanica seedlings required 40 - 50 % of root length anchored to the sediment before being dislodged while C. nodosa required â&#x2030;&#x2C6; 20 %. To validate the flume results, seedling survival in sandy beds was evaluated for two depths (12 and 18 m) at two field locations. To calculate near-bottom orbital velocities at the planting sites, deep 1 Infantes E, Orfila A, Bouma TJ, Simarro G, Terrados J. Posidonia oceanica and Cymodocea nodosa seedling tolerance to wave exposure. (Submitted to: Limnology and Oceanography).

27

28

Seedling tolerance to wave exposure

water waves were propagated to shallow water using a numerical model. Results showed that P. oceanica seedlings experienced high losses after the first autumn storms when near-bottom orbital velocities exceeded 18 cm s−1 . The loss of C. nodosa seedlings was much lower and some seedlings survived velocities as high as 39 cm s−1 . Thus, flume and field results are consistent in explaining relative higher losses of P. oceanica seedlings than for C. nodosa.

3.2

Methods

Posidonia oceanica and Cymodocea nodosa seedlings used in the flume and field studies were obtained from fruits collected at sea. P. oceanica fruits were collected from drift material on several beaches of Mallorca and Ibiza Islands (Western Mediterranean Sea) during April-May 2009. Fruits were manually opened and seeds were placed in natural salinity seawater. Seed germination took place within a week or two after collection. C. nodosa seeds were collected during February-April 2009 in a shallow (3 m) meadow in Mallorca Island. Seeds were kept in aquariums until May 2009, the month when C. nodosa seeds germinate at sea (Buia and Mazzella 1991). Germination of C. nodosa was induced by reducing the salinity to 10 (Caye and Meinesz 1986). Both P. oceanica and C. nodosa germinated seeds were kept in aquariums with natural seawater during early seedling development. Aquariums were in a temperature-controlled room at 20◦ C with a 14 h photoperiod and a photon irradiance of 60 µmol m−2 s−1 .

Seedlings were haphazardly distributed into two groups, one for flume experiments and another for field experiments.

3.2.1

Drag forces and drag coefficient

Drag measurements were carried out to obtain the effective drag forces acting on individual seedlings and to calculate the drag coefficients of both species under unidirectional and oscillatory flows. Experiments were carried out in the NIOO racetrack-shaped channel at Yerseke, Netherlands. The channel is 17.55 m long, 0.6 m wide and 0.45 m deep (Hendricks et al. 2006). In all experiments, water depth was 0.32 m, water temperature was 20 ± 0.5◦ C and salinity 36. The flume allows running both unidirectional and oscillatory flows. Currents in the flume were generated by a conveyor belt. Waves were generated by a vertical wave maker driven by a piston. At the end of the test section waves were dampened by

3.2 Methods

29

a porous gentle slope. Flow was characterised by measuring vertical profiles of the velocity at 5 cm intervals along the water column using an Acoustic Doppler Velocimeter (ADV Vectrino, Nortek) at the same cross section where seedlings were tested. The ADV sampling rate was 25 Hz and measurements were taken for 5 minutes, sampling volume of 7 mm and a nominal velocity range of 1 m s−1 .

Drag was measured on individual P. oceanica and C. nodosa seedlings fixed to a force transducer inside the flume. Seedlings were exposed to 7 different wave conditions that include wave heights (H) ranging from 1 to 7 cm and periods (T ) from 1.5 to 4 seconds. These waves gave near-bottom orbital velocities between 4 and 16 cm s−1 . Seedlings were also exposed to 6 different current velocities ranging between 5 and 36 cm s−1 . Hereafter, u represents the near-bottom orbital velocity (for waves) or the mean velocity (currents).

Each wave and current flow conditions was repeated for 5 seedlings. Seedlings were attached to a 2 cm long metal screw by the stem using a small cable tie removing the roots and seed. The metal screw was then fixed into the drag sensor. The drag force of the metal screw with cable ties alone is a constant, and was subtracted from the drag force measured on the seedlings. No breaking of leaves or plant fragments was observed during the drag experiments. Drag forces and wave heights were measured at 20 Hz during 3 min. Drag forces were measured using a drag transducer developed by WL-Delft Hydraulics (Bouma et al. 2005) and calibrated as Stewart (2004). Waves were measured with a pressure sensor (Druck, PT1830). Drag measurements were performed on P. oceanica in October 2009 and February 2010, but drag in C. nodosa was only measured in October 2009 since seedlings did not survive until February 2010.

The force acting over individual plants measured by the transducer can be expressed as, 1 (3.1) FD = ρCD av u|u| 2 where ρ is the density of water, av is the total leaf area of the plant and u the flow velocity (either unidirectional or oscillatory). The drag coefficient, CD , which depends on the Reynolds number, can be readily obtained from the experiments using Eq. A.9. Defining the characteristic length as (Martone and Denny 2008): √ l∗ = av (3.2) the corresponding Reynolds number is defined as, Re =

ul∗ ν

(3.3)

30

Seedling tolerance to wave exposure

Figure 3.1: Sketch of the flume experimental set-up showing the critical erosion depth and minimum rooting length (Ld ) of seedlings. (a) Seedling in the flume, (b) seedling before dislodging from the sediment after the discs addition. Critical erosion depth is equivalent to the total height of discs added when the seedling is dislodged. Not drawn to scale being ν the molecular kinematic viscosity of seawater. Most aquatic vegetation is flexible and becomes streamlined with increasing flow velocities and it is very difficult to measure accurately the frontal area under waves and currents (Sand-Jensen 2003, 2005). For this reason, the total surface area was used as the frontal area as it can be accurately measured.

3.2.2

Critical erosion depth and minimum rooting length

Once the drag force was measured for the seedlings the next step was to analyze the capacity of both species to withstand dislodgement by sediment disturbances via root anchoring. Individual P. oceanica and C. nodosa seedlings were planted in cylindrical pots (height x diameter: 12 x 12 cm) over sand with a diameter (d50 ) of 592 ± 23µm. Pots were placed in the flume under two sets of experiments: i) periodic waves of H = 2.5 cm and T = 2.6 s (u = 5 cm

3.2 Methods

31

sâ&#x2C6;&#x2019;1 ) and ii) periodic waves of H = 5.1 cm and T = 1.9 s (u = 10 cm sâ&#x2C6;&#x2019;1 ) during 15 minutes (Fig. 3.1a). Seedlings were repeatedly exposed to 15 minutes of wave action, while in between wave treatments, sediment erosion was mimicked.

This was repeated until the critical erosion depth at which seedlings dislodged was found. Sediment erosion was simulated by progressively adding discs of 3 mm thickness underneath the pots removing carefully the pushed-up top layer of sediment. This process was repeated until the waves dislodged the seedling from the sediment. A total of 18 plants were used for P. oceanica and 9 for C. nodosa. For each seedling, root lengths and total leaf area were measured.

The critical erosion depth was measured as the thickness of disks added before the plant became dislodged. The minimum rooting length (Ld ) or length of the root inside the sediment before the plant become dislodged was obtained by the difference of the maximum root length and the critical erosion depth.

The length of the longest root was used for all calculations since in the end, regardless of the number of roots, seedlings remained anchored to the substrata with only one root, even when the rest were dislodged.

3.2.3

Field study

Seedlings were planted at sea to assess their survival and to validate the principles observed in the flume. The field study was performed from August 2009 to February 2010 in Mallorca Island, Spain, Western Mediterranean Sea (Fig. 3.2a,b). In this area, tides are almost negligible, e.g., less than 25 cm (Orfila et al. 2005). P. oceanica and C. nodosa seedlings were planted at two locations in the Natural Reserve of Cap de Enderrocat (Fig. 3.2c). The first location, Cap Enderrocat is a sandy area between the upper limit of P. oceanica and the coastline exposed to Southwest and Southeast waves (triangles in Fig. 3.2c).

The second location, Cala Blava is located in a large P. oceanica meadow with sand gaps (circles in 3.2c), which is only exposed to Southwest waves. The upper depth limit of P. oceanica in Cap Enderrocat is at 19 m while that in Cala Blava is at 12 m. The upper depth limit of C. nodosa in Cap Enderrocat and Cala Blava is at 12 m. Within each location two depths were selected at 18 m and 12 m depth for the short-term experiment.

Six plots (three for each species) separated by 3 m intervals were established at the four sites. Twelve seedlings were planted on each plot directly on the sediment without any artificial supporting aid, in order to mimic natural condi-

32

Seedling tolerance to wave exposure

Figure 3.2: (a) Location of Mallorca Island in the Mediterranean Sea. (b) Location of study area and WANA node. (c) Location of the experimental sites, Cap Enderrocat (triangles) and Cala Blava (circles). Bathymetric contours in meters.

Figure 3.3: Experimental plots with a) Posidonia oceanica and b) Cymodocea nodosa seedlings at the beginning of the field experiment.

3.2 Methods

33

tions where the roots are the only attachment structure to the substrata (Fig. 3.3a,b). A total of 144 seedlings of each species were planted after measuring the number of roots, root lengths and root diameter in the laboratory. Additionally, the number of leaves of each seedling and the length and width of each leaf were measured once planted in the field. The number of seedlings in each plot was counted in September, October, and November 2009 and in February 2010. Whenever seedlings were no longer observed in the plots, it was determined that the loss was not caused by burial.

Survival rate was related to wave conditions as well as to averaged nearbottom orbital velocities between sampling dates at the four experimental sites.

Significant wave height (Hs ), peak period (Tp ), and direction (Θ) were acquired at deep waters approximately 15 km from the study sites (Fig. 3.2b). These wave conditions were propagated to the study sites using a numerical model based on the mild-slope parabolic approximation (Kirby and Dalrymple 1983, Alvarez-Ellacuria 2010, Infantes et al. 2011). Wave conditions in the study area ranged between a Hs of 0.5 - 4 m, Tp of 2 - 12 s and Θ of 130 - 270 degrees. Near-bottom orbital velocities at the four locations where P. oceanica and C. nodosa seedlings were transplanted were computed from the model output between sampling periods using linear wave theory (Dean and Dalrymple 1991).

3.2.4

Statistical analysis

Differences in drag forces under currents and waves of both species were tested using a Student t-test. Differences among drag coefficients were evaluated using species (P. oceanica, C. nodosa) and flow (unidirectional, oscillatory) as fixed factors and Reynolds number as the covariate in the analysis of covariance (ANCOVA) model CD = species + flow + Reynolds + (species x flow) + (species + Reynolds) + (flow x Reynolds) + (species x flow x Reynolds) (Quinn and Keough 2002). Critical erosion depths and minimum rooting lengths in the sediment were evaluated using univariate test of significance ANOVA. “Wave velocity” (orbital velocities of 5 cm s−1 and 10 cm s−1 ), “Species” (P. oceanica and C. nodosa) were the between-subjects factors (fixed). Differences of seedling survivorship in the field experiment were also evaluated using univariate test of significance ANOVA. The test was performed only in October 2009 because we observed differences between sites on that date. After October 2009 no plants were left in most of the plots (see Results). “Sites” (Cap Enderrocat and Cala

34

Seedling tolerance to wave exposure

Figure 3.4: Drag forces acting on seedlings in (a) unidirectional flow and (b) oscillatory flow. (mean, SE, n = 5).

Blava), “Depth” (18 m and 12 m), “Species” (P. oceanica and C. nodosa) were the factors (fixed). A Cochran’s C-test was used to test for heterogeneity of variances. All data were normally distributed.

3.3

Results

Posidonia oceanica and Cymodocea nodosa have flexible leaves which easily bend with the currents or waves. Drag forces acting on individual seedlings increased with current and wave velocities (see Fig. 3.4a,b). Drag forces also increased with foliar surface area for both seagrass species (Fig. 3.5). Drag forces were higher (t-test: t = 7.08, df = 14, p < 0.01) for P. oceanica (0.006 ± 4 x 10−4 N, mean ± SE) than for C. nodosa (0.003 ± 3 x 10−4 N) at u = 16 cm s−1 in unidirectional flow (see Fig. 3.5a). Similarly, drag forces were also higher (t-test: t = 3.2, df = 14, p < 0.01) for P. oceanica (0.011 ± 1 x 10−3 N) than for C. nodosa (0.005 ± 6 x 10−4 N) at u = 16 cm s−1 in oscillatory flow (see Fig. 3.5b).

ANCOVA shows that drag coefficient (CD ) depends on Reynolds number

3.3 Results

35

Figure 3.5: Drag forces acting on individual seedlings of different surface area in (a) unidirectional flow and (b) oscillatory flow. Flow velocity of 16 cm s−1 .

(Re), (F = 298.7, p < 0.001) and that it is also influenced by the type of flow (CD currents > CD waves, F = 4.8, p < 0.05), but not by the species (F = 1.4, p = 0.24) (Fig. 3.6. A linear fitting of the drag coefficient for both species under oscillatory motion is, log10 CD = −0.6653 ∗ log10 Re + 1.1886, (R2 = 0.77, r = −0.87)

(3.4)

and similarly for unidirectional flow log10 CD = −0.7269 ∗ log10 Re + 1.6253, (R2 = 0.92, r = −0.95)

(3.5)

The experiments in the flume have shown that there are significant differences in the minimum rooting length that each species tolerates (Table 3.1). P. oceanica seedlings require 40 - 50 % (2 ± 0.2 cm, mean ± SE) of root length anchored into the sediment while C. nodosa requires only 20 % (2.8 ± 0.2 cm, mean ± SE) of root length anchored into the sediment (Fig. 3.7). P. oceanica seedlings tolerate less sediment erosion (2.7 ± 0.3 cm, mean ± SE) to become dislodged than C. nodosa (5.6 ± 0.4 cm, mean ± SE) under waves.

The capacity of a plant to remain anchored depends on the force that the roots can manage (which mainly depends on their length Ld ), and on the force

36

Seedling tolerance to wave exposure

Figure 3.6: Drag coefficient versus Reynolds number for Posidonia oceanica and Cymodocea nodosa under undirectional and oscillatory flow. Experimental data under uniform flow for P. oceanica seedlings (triangles) and for C. nodosa (squares). For oscillatory flow P. oceanica seedlings (circles), and C. nodosa (crosses). Solid lines are the linear fitting for the different flow conditions.

that the exposed leaves receive from the ambient (which is highly dependent on the total leaf area av ). The above suggests that an important dimensionless √ coefficient in the problem is the ratio Ld / av . Since other variables could play √ a role in the problem, the ratio Ld / av will not be constant in general, and it could depend on other dimensionless groups representing the plant species, flow velocity, type of flow (wave or current), etc. √ From the experimental results, only the influence of the species on Ld / av has been clearly identified as a measure of the dislodgment risk, which can be termed as the dislodgement safety factor. Results presneted in this Chapter indicate that under periodic flows, seedlings of P. oceanica will remain anchored as long as P. oceanica ≈ 0.35, i.e., if their root lengths are on average 0.35 times the square root of the leaves area. In the case of C. nodosa the condition for √ the seedlings to remain anchored is Ld / av ≈ 1.6.

3.3 Results

37

Minimum rooting Length

df 1 1 1 34

MS 6828 63 196 129

F 52.9 ∗∗∗ 0.5 ns 1.5 ns

Table 3.1: Two-factor way ANOVA testing differences between species and orbital velocities and minimum rooting length. Significant differences are expressed in bold as, ∗∗∗ p < 0.001, and ns = not significant. Cochran’s C-test no significant. df = degrees of freedom and MS = mean square.

The foliar surface of P. oceanica seedlings at the beginning of the field experiment was 4 times larger than that of C. nodosa seedlings (Table 3.2). As seedlings develop the difference in foliar surface increases. In contrast, C. nodosa maximum root lengths are higher than P. oceanica. Root diameters are higher for P. oceanica than for C. nodosa. Survival was related to wave conditions as well as to averaged near-bottom orbital velocities between sampling dates (Fig. 3.8; Table 3.3). No loss of seedlings occurred from August to September 2009. During this period no storms affected the study area and the computed near-bottom orbital velocities were below 5 cm s−1 .

Total foliar surface (cm2 ) Number of leaves Number of roots Total root length (cm) Max. root length (cm) Root diameter (cm)

P. oceanica 12.9 ± 1.6 6.1 ± 0.6 5.2 ± 0.2 15.8 ± 0.2 5.5 ± 0.2 1.78 ± 0.1

C. nodosa 3.2 ± 0.4 2.3 ± 0.2 4.8 ± 0.2 26.1 ± 0.4 6.8 ± 0.4 0.51 ± 0.03

Table 3.2: Morphological characteristics of seedlings at the beginning of the field experiment (mean ± SE, n = 144).

38

Seedling tolerance to wave exposure

Figure 3.7: (a) Critical erosion depth and (b) Minimum rooting length of Posidonia oceanica and Cymodocea nodosa exposed to two orbital velocities (u = 5 and 10 cm s−1 ).

Between September and October 2009, a storm with Hs of 1.5 m and Tp of 7 s was measured at deep waters. During this period, the first losses of P. oceanica seedlings were observed. Computed u for this period in Cap Enderrocat was 8.6 cm s−1 at 18 m depth and 28 cm s−1 at 12 m depth while in Cala Blava was 7.3 cm s−1 at 18 m depth and 14.7 cm s−1 at 12 m depth (Table 3.3). The percentage of P. oceanica seedlings surviving in Cala Blava was significantly higher than that in Cap Enderrocat (3.8a; Table 3.4). Moreover, the percentage was significantly higher in the deeper (18 m) than in the shallow (12 m) sites.

The loss for C. nodosa was also significant but no differences in the percentage of survivorship among sites and depths was detected.

3.4 Discussion

Aug-Sep Sep-Oct Oct-Nov Nov-Feb

Cap Enderrocat 12 m <5 18.0 ± 1.1 34.4 ± 2.2 46.0 ± 2.3

39

18 m <5 8.6 ± 0.3 18.2 ± 1.2 25.5 ± 1.2

Cala Blava 12 m <5 14.7 ± 0.9 29.4 ± 1.9 39.1 ± 1.5

18 m <5 7.3 ± 0.2 17.0 ± 1.1 23.4 ± 1.1

Table 3.3: Computed near-bottom orbital velocities (cm s−1 ) at the experimental locations during the sampling periods (mean ± SD).

From October to November 2009, two storms from the southwest with Hs of 4 and 2.5 m and Tp of 10 and 8 s respectively were recorded at deep waters.

After these storms, all P. oceanica seedlings disappeared in both Cap Enderrocat and Cala Blava sites while C. nodosa seedlings persisted in Cala Blava at both depths (Fig. 3.8b). Computed u at the experimental sites varied between 17 to 34 cm s−1 (Table 3.3).

From November 2009 to February 2010 several storms were recorded with maximum Hs of 5 m and Tp of 11 s at deep waters (Fig. 3.8c). Remaining seedlings of C. nodosa persisted during this period. Computed u in this site varied between 23 and 39 cm s−1 during that period of time.

3.4

Discussion

This Chapter offers a mechanistic insight into how the combination of drag and erosion may cause seedling dislodgement from the substrata. Results showed that drag forces were higher in P. oceanica seedlings than in C. nodosa under both current and waves, and that P. oceanica seedlings are earlier dislodged than C. nodosa seedlings. Thus flume results predict survival of C. nodosa seedlings to exceed that of P. oceanica seedlings. Field experiments confirmed this, by showing that seedlings were dislodged during storms, and that C. nodosa seedlings survived longer than P. oceanica.

Measured drag force per unit total leaf surface area is on average (3 ± 1.2 N/m2 mean ± SE) for both the P. oceanica and C. nodosa seedlings at u =

40

Seedling tolerance to wave exposure

Location Depth Species Location x Depth Location x species Depth x species Location x depth x species Error

df 1 1 1 1 1 1 1 16

MS 9401 350 3151 651 2109 2433 2109 234

F ∗∗∗

40.1 1.5 ns 13.5 ∗∗ 2.8 ns 9.1 ∗∗ 10.3 ∗∗ 9.2 ∗∗

Table 3.4: Results of three-factor way ANOVA of seedling survival percentage in October 2009. Significant differences are expressed in bold as ∗∗ p < 0.01 , ∗∗∗ p < 0.001, and (ns) not significant. Cochran’s C-test no significant. df = degrees of freedom and MS = mean square

16 cm s−1 , which was in the same order of magnitude as earlier measurements for seagrass model plants with flexible shoots (5.5 - 7.8 N/m2 ) at 37 cm s−1 (Bouma et al. 2005). For seedling survival, it is however the absolute drag force rather than the drag per unit surface area that matters. Present flume experiments show that drag coefficient, CD , depends on Reynolds number but not on the species and that this relationship is different for current and waves (Fig. 3.6). The computed drag coefficients for the individual seedlings of P. oceanica and C. nodosa are between 0.01 and 0.1 for Reynolds numbers between 103 and 105 , which is in accordance with those provided for flexible macrophytes (Sand-Jensen 2003, Martone and Denny 2008). It should be noted that drag coefficients for meadows can depend on many factors such as the height and density of the plants, the distance travelled by waves, etc. and that drag coefficient for meadows thus may differ from the values presented for single plants.

Sediment mobilisation is related to hydrodynamics (wave and currents) which can bury or dislodge seagrasses (Madsen et al. 2001, Teeter et al. 2001, Cabaco et al. 2008). Previous studies have shown that 2 cm of sediment erosion caused 75 % of mortality in C. nodosa seedlings in stagnant water (Marb`a and Duarte 1994), but to my knowledge there are no data about the effect of sediment erosion on P. oceanica seedlings. In this Chapter is shown that P. oceanica seedlings under waves tolerate less sediment erosion (2 - 3 cm) to become dis-

3.4 Discussion

41

Figure 3.8: Seedling survival on the experimental plots in Mallorca, August 2009 to February 2010. (a) Posidonia oceanica, (b) Cymodocea nodosa, and (c) wave heights in deep water (WANA node) shown in grey line and propagated wave heights in Cala Blava at 12 m depth shown in black line. Gaps in propagated Hs corresponds to wave directions other than Southwest to Southeast no affecting the study area. lodged than C. nodosa (5 - 6 cm). P. oceanica seedlings need a root length of 0.35 times the square root of the total leaf area to remain anchored while in C. nodosa this condition is increased up to 1.6 times the square root of leaf area. However, seedlings of P. oceanica in the flume have a foliar area that is on average 20 times higher than those of C. nodosa and therefore the seedling capacity to remain anchored is much more limited for P. oceanica than C. nodosa.

This suggests that P. oceanica seedlings are adapted to grow at ambients less energetic than C. nodosa. This may also be inferred from the proportionally higher root biomass and lower leaf biomass that C. nodosa allocates compared to P. oceanica (Guidetti et al. 2002). Anderson et al. (2006) described the dislodgement of coenocytic green algae from soft sediments by waves measuring

42

Seedling tolerance to wave exposure

the mean force to dislodge the algae (4.9 - 12.7 N) which increased as the leaf surface area increased.

The seedling survival in the field also showed that P. oceanica and C. nodosa seedlings behave differently under the same hydrodynamic conditions. C. nodosa seedlings survived longer than P. oceanica in Cala Blava during the field experiment. P. oceanica seedlings survived at a higher proportion in the deep than in the shallow sites at both locations during October 2009. In November 2009, all P. oceanica seedlings disappeared in the plots while C. nodosa seedlings persisted at both depths in Cala Blava. Two storms took place before the complete loss of P. oceanica seedlings in November 2009. It can not be determined which of these two events caused seedling loss but it can conclude that these storms had a deleterious effect on the survival of P. oceanica seedlings.

The presence of a large P. oceanica meadow from the shallow areas to 30 - 35 m depths demonstrates that the seagrass is not light limited at the study site (Fig. 3.2c). Herbivory is not considered to be a cause of seedling loss because bite scars on the leaves were not observed. Different environmental conditions between sites associated with hydrodynamics such as sediment resuspension and turbidity could have an effect on the seedling survival, although, these are indirect effects to wave exposure. Some studies have shown high mortality of seedlings in the field after storms. A mortality of 75 % to 100 % natural and transplanted P. oceanica seedlings on pebbles occurred during the first winter after planting, associated to storms (Balestri et al. 1998). Individual seedlings of four different Posidonia spp. planted in a blow out area did not succeed after 1 year (Kirkman 1998).

Water depth had an effect on seedling survival, as wave induced orbital velocities attenuate with depth, near-bottom orbital velocities will be higher at 12 m than at 18 m. In October 2009, a significantly lower survival was observed for P. oceanica at 12 m than at 18 m (Fig. 3.8a; Table 3.4). Piazzi et al. (1999) observed lower survival rates of P. oceanica seedlings at 2 m depth than at 10 m depth. A correlative study in a P. oceanica meadow (Infantes et al. 2009) suggested that at depths between 0 to 5 m near-bottom orbital velocities were too high for P. oceanica to be present while at depths deeper than 7 m velocities were lower and a dense meadow was present.

3.4 Discussion

43

Infantes et al. (2009) estimated a threshold in near-bottom orbital velocities of 38 - 42 cm s−1 for a P. oceanica meadow to be present in a shallow bay.

In this study, P. oceanica seedlings had a 100 % survivorship at near-bottom orbital velocities below 5 cm s−1 , but for velocities between 7 - 18 cm s−1 the percentage of seedlings surviving decreased and all seedlings disappeared at velocities above 18 cm s−1 . Some C. nodosa seedlings survived at velocities of 39 cm s−1 . Thus, these results suggest that seedlings tolerate lower orbital velocities than mature plants. This is not surprising, because mature plants develop a network of roots and rhizomes that penetrate deeper into the substrata than seedlings. Other studies showed that transplanted Zostera marina survived at orbital velocities of 40 cm s−1 but not at 60 cm s−1 (van Katwijk and Hermus 2000). For meadows, Cabaco et al. (2010) estimated that the decline of C. nodosa occurs at velocities over 60 cm s−1 .

Flow velocity is an important factor determining the size of the marine organisms that inhabit benthic ecosystems (Denny et al. 1985). Present results suggest that the higher drag of large species may make small species relatively more successful in the colonisation of exposed sites. This agrees with the convention (based on size-related differences of growth rates) that large seagrass species generally occupy low energy or sheltered habitats while small seagrass species are able to colonize more energetic habitats because the capacity of small species to recover from disturbance through growth is higher that that of large species (Duarte 1991, Idestam-Almquist and Kautsky 1995, Laugier et al. 1999). Results presented in this Chapter that seedlings of C. nodosa are able to persist under higher wave energies than those of P. oceanica, are also in agreement with the old but untested assumption that C. nodosa is a pioneer species that stabilises the substrata and facilitates the establishment of P. oceanica (Den Hartog 1970).

The findings presented in this Chapter provide the basis for understanding seedling survivorship of two Mediterranean seagrass species in sandy substratum under different hydrodynamic conditions. Morphological characteristics of each species such as foliar surface and root length are key factors in colonisation processes. Moreover, drag and sediment erosion play an important role in seedling survival, a crucial stage of seagrass life. Habitat degradation and fragmentation caused by anthropogenic effects around the world’s coastal zones raises the need to manage and restore seagrass beds (Short and Wyllie-Echeverria 1996, Orth

44

Seedling tolerance to wave exposure

et al. 2006). Seedling plantings could be more effective in seagrass restoration if more quantitative knowledge on the hydrodynamic effects that limit the survival of seagrass seedlings is gathered.

Chapter 4

Assessment of substratum effect on the distribution of two invasive Caulerpa (Chlorophyta) species Part of this Chapter has been publisehd as an article on Estuarine, Coastal and Shelf Science1

4.1

Summary

Two-year monitoring of the invasive marine Chlorophyta Caulerpa taxifolia and Caulerpa racemosa var. cylindracea shows the great influence of substratum on their spatial distribution. The cover of C. taxifolia and C. racemosa was measured in shallow (< 8 m) areas indicating that these species are more abundant in rocks with photophilic algae and in the dead matte of the seagrass Posidonia oceanica than in sand or inside the P. oceanica meadow. A short-term experiment comparing the persistence of C. taxifolia and C. racemosa planted either in a model of dead matte of P. oceanica or in sand shows that the persistence of these species was higher in the dead matte model than in sand. Correlative 1 Infantes E, Terrados J, Orfila A (2011) Assessment of substratum effect on the distribution of two invasive Caulerpa (Chlorophyta) species. Estuar. Coast. and Shelf Sci. 91: 434-441.

45

46

Substratum effect on invasive Caulerpa species

evidence suggests that C. taxifolia and C. racemosa tolerate near-bottom orbital velocities below 15 cm s−1 and that C. taxifolia cover declines at velocities above that value. These results contribute to understand the process of invasion of these Caulerpa species predicting which substrata would be more susceptible to be invaded and to the adoption of appropriate management strategies.

4.2 4.2.1

Methods Study Sites

The study was carried out in four shallow (depth less than 8 m) locations on the South coast of Mallorca Island, Western Mediterranean sea (Fig. 4.1a). Cala D’Or was chosen for this study because it is the only location in the Balearic Islands where Caulerpa taxifolia is present. Caulerpa racemosa is mainly found in the South of Mallorca, and the locations of Sant Elm, Cala Estancia and Portals Vells where chosen (Fig. 4.1b). Sand patches, rocky reefs covered by an erect stratum of photophilous macroalgae and meadows of the seagrass Posidonia oceanica are the main components of the undersea landscape at the study sites.

4.2.2

Presence of Caulerpa over different substrata

The cover of Caulerpa taxifolia in Cala D’Or and that of Caulerpa racemosa in Cala Estancia and Portals Vells was measured during two consecutive summers and winters (years 2007 and 2008) which represent the periods of maximum and minimum vegetative development of these species in the Western Mediterranean (Piazzi and Cinelli 1999). Four to eight 20 m long transects were laid haphazardly at each location and the percentage of substratum covered by Caulerpa was estimated at 1 m intervals along the transects by placing two 20 cm x 20 cm quadrats divided in 25 cells of 4 cm2 . The number of cells corresponding to each type of substratum in each quadrat was counted as well as the number of cells of each substratum type colonized by Caulerpa. Substrata were classified either as sand, rocks (always covered by photophilic algae), Posidonia oceanica meadow or dead matte of P. oceanica. The matte of P. oceanica is formed by the accumulation of sediment, rhizomes and/or roots of this seagrass species over time, building a terraced structure that retains the sediment even after seagrass death and that may last hundreds of years (Mateo et al. 1997).

4.2 Methods

47

Figure 4.1: Location of the study areas. (a) Mallorca Island in the Mediterranean Sea. (b) Location of the four study areas and deep water wave data (WANA nodes). (c) Bathymetry of Cala D’Or with location of WANA node. (d) Bathymetry of Sant Elm. Experimental plots, WANA nodes and ADCP location in Cala D’Or.

4.2.3

Short-term experiment evaluating substratum effect on Caulerpa persistence

The experiment evaluated the persistence of Caulerpa fragments planted in sand plots (Fig. 4.2a) and in plots of a model of Posidonia oceanica dead matte (Fig. 4.2b). The experiment was performed in two locations: Cala D’Or for Caulerpa taxifolia and Sant Elm for Caulerpa racemosa (Fig.4.1b). These two locations were selected because C. racemosa is not present in Cala D’Or, which is the only place where C. taxifolia is present in the Balearic Islands, and it was not consider appropriate to contribute to the dispersion of these invasive species by transplanting them to other locations. Besides, the C. racemosa experiment was not perform in Cala Estancia or Portals Vells because these sites are highly visited by swimmers and snorkelers and the risk of disturbance to experimental plots was high. A sand patch large enough to hold all the experimental plots at a depth of 7 - 8 m was selected at each location. These sand patches were surrounded by P. oceanica meadows.

48

Substratum effect on invasive Caulerpa species

Figure 4.2: Photographs of the experimental set up for Caulerpa taxifolia in (a) natural sandy bottom and (b) the model of dead matte of the seagrass Posidonia oceanica. (c) Illustration of Caulerpa fragments fixed to the plots using pickets.

Six 50 cm x 50 cm plots were established in each location separated every 5 m. Plots were delimited at the corners with galvanized iron bars. The model of dead matte of Posidonia oceanica was constructed by attaching fragments of dead P. oceanica rhizomes collected from fresh beach-cast plants after a storm to 50 cm x 50 cm squares of plastic gardening mesh, using plastic cable ties (Fig. 4.2b). A density of around 500 vertical rhizomes per m2 was used as this is a typical P. oceanica shoot density found at 8 m depth in clear Mediterranean waters (Marb` a et al 2002, Gobert et al. 2003, Terrados and Medina-Pons 2008).

A model of dead matte of P. oceanica was used instead of natural occurring dead matte to standardize the methodology having the same substratum types in both locations. The dead matte models were attached to the bars delimiting the randomly assigned plots (n = 3) with cable ties and using four pickets to anchor it firmly to the sediment (Fig. 4.2b). The sand plots (n = 3) were made of the naturally occurring sand and placement of the galvanized iron bars delimiting plot corners was the only manipulation performed.

Fragments of Caulerpa including at least one stolon apex and 3-4 fronds each were manually collected from the same area and depth where the experimental

4.2 Methods

49

plots were located and held in aerated seawater at sea surface until planting.

Sixteen Caulerpa fragments were fixed to each plot within 2h after collection by threading them in groups of four into four cable-tie loops anchored in each plot using galvanized iron pickets (see Fig. 4.2c). The experiment started in July of 2007 as previous studies on biomass seasonality of Caulerpa taxifolia in shallow waters (6 - 10 m) have shown that maximum biomass occurs in summer and autumn (Thibaut et al. 2004). Monitoring of Caulerpa cover in each plot was carried out approximately every 30 days for C. taxifolia and 15 days for Caulerpa racemosa as the latter species has a higher growth rate (Piazzi et al.

2001) using underwater digital photography.

Photographic sampling techniques have been previously used to measure C. racemosa cover (Piazzi et al. 2003). In this study, underwater photographs were taken to evaluate the persistence of Caulerpa in the plots by analyzing the area covered by the algae over time. This method allowed taking many samples in the field as diving time is limited. Each plot was divided on 4 sub-plots of 25 x 25 cm to improve the delineation of the algae on the photographic images. As algae can swing with wave orbital movements that could affect the area facing up on the images, three images of each sub-plotwere taken on calm days. A scale was placed oneach sub-plot next to the algae when taking the underwater photographs. Open software (ImageJ, NIH) was used to measure the area of Caulerpa on each photograph. To determine the absolute area on the images a relationship between image pixels and real dimensions ratio was established on each image using the scale. In order to test the method, photographs were taken in the laboratory and in the sea before the experiment began. A set of figures of known areas was used to test the output of theimage processing analysis, providing an error around 6 %.

4.2.4

Wave climate and modelling at the short-term experimental sites

The propagation of the deep water mean wave climate between sampling dates was performed using a numerical model which allows the calculation of the wave induced near-bottom orbital velocity at the study sites. Significant wave height (Hs ), peak period (Tp ) and direction were acquired from July 2007 to March 2008 at deep water which corresponds to the duration of the short-term experiment (see 4.1b, c, for the location of these points). The deep water data have

50

Substratum effect on invasive Caulerpa species

been obtained from the WANA system, a reanalysis with real data of a third generation spectral WAM model and was provided by the Spanish harbour authority operational system. These deep water wave conditions were propagated to the study sites using a numerical model based on the mild slope parabolic approximation (Kirby and Dalrymple, 1983) and near-bottom orbital velocities (ub ) at the experimental locations were obtained from the model outputs using linear wave theory (Infantes et al. 2009). These models are widely used in engineering and scientific approaches in order to obtain the accurate wave field in coastal areas from open sea conditions since refraction and diffraction are well captured (Alvarez-Ellacuria et al. 2010).

4.2.5

Field wave measurements and velocity profiles

To test the results of deep water wave propagations, field measurements were performed on the 8th of November of 2007 in Cala D’Or measuring wave conditions (Hs and Tp ) at deep and shallow water as well as near-bottom velocities in shallow water on both substrata. An Acoustic Doppler Current Profiler (ADCP) RDI, Sentinel 600 kHz, was deployed at approximately 400 m from the experimental site at a depth of 26 m. The ADCP was set to record 20 min burst intervals every 60 min with a 2 Hz frequency. It was mounted on a tripod at 0.5 m above the bed. Hs obtained by propagating the deep water conditions to shallow waters, are within aw 15 % error to those measured in situ at deep and shallow waters, showing the accuracy of modelled wave conditions. Table 4.1 shows the data provided by the numerical model and those measured by the ADCP at deep water (26 m) and with the ADV at shallow water (7.5 m) on the 8th of November of 2007.

Hs

Deep water (26 m) Shallow water (7.5 m)

Model outputs 0.53 ± 0.02 m 0.26 ± 0.02 m

Field measurements 0.47 ± 0.03 m 0.21 ± 0.02 m

Table 4.1: Wave heights from model outputs and field measurements in Cala D’Or (means ± 95 % CL).

Additionally, vertical profiles of velocity were measured on the P. oceanica

4.2 Methods

51

dead matte model and on sand using Acoustic Doppler Velocimeters (ADV), Vector, Nortek. Simultaneous measurements on both substrata were performed using two ADVs at seven positions (2.5, 5, 10, 20, 40, 60 and 100 cm above the bottom). The instruments were set to measure during 25 min with a frequency of 32 Hz, a nominal velocity range of 0.1 m s−1 and a sampling volume of 11.8 mm. The ADVs were mounted in a down-looking position on stainless steel bottom fixed structures. Stability of the equipmentwas verified with the compass, tilt and roll sensors. Raw data were processed by first filtering for low beam correlation (< 70 %). Velocity data were then filtered with a low pass filter to remove high frequency Doppler noise (Lane et al. 1998). Horizontal velocity was calculated as the root mean square (rms). From the velocity profiles, friction coefficients Cf over sand and dead matte substrata were computed following Simarro et al. (2008).

Cf =

τb ρu2c

(4.1)

where τb is the bottom shear stress, ρ the density of seawater and uc a characteristic velocity outside the boundary layer. The bed shear stress can be found from, du )Z=δ (4.2) dz where νT is the eddy viscosity and du/dz the vertical variation of the velocity evaluated at the outside of the boundary layer δ.

R Temperature HOBO StowAway Tidbit loggers were moored next to the plots in Cala D’Or and Sant Elm during the experiment. Temperature was recorded every 2 h and averages of water temperature between sampling dates were calculated. The numbers of daylight hours were obtained from the US Naval Observatory (http://aa.usno.navy.mil/) for the study areas and averages of daylight hours between sampling dates were also calculated. τb = ρ(νT

4.2.6

Statistical Analysis

A Chi-squared test (χ2 ) was used to evaluate if the colonization of benthic communities by Caulerpa species depends on the type of substratum, that is, if some substrata are more favourable/unfavourable than others for the colonization of these species. The null hypothesis of the test assumes substratumindependent, random colonization and allows the calculation of expected cover

52

Substratum effect on invasive Caulerpa species

of Caulerpa species in the different substrata that is compared with the cover observed. Changes in cover of Caulerpa species through time in the short-term experiments were evaluated using univariate repeated-measures ANOVA. “Substratum” (sand versus model of dead matte of P. oceanica) was the betweensubjects factors (fixed) and “Time” was the within-subjects factor (random) for both Caulerpa experiments. A significant Time x Substratum interaction was to be expected if the type of substratum has an effect on the persistence of Caulerpa in the plots. A Cochran’s C-test was used to test for heterogeneity of variances.

4.3

Results

Caulerpa taxifolia was more abundant in the dead matte of Posidonia oceanica than in the rest of substratum types available in Cala D’Or (Fig. 4.3a). C. racemosa was also more abundant in the dead matte of P. oceanica but it additionally colonized other substrata such as rocks with photophilic algae and sand (Fig. 4.3b). The percent cover of C. taxifolia over dead matte showed small changes between summer and winter (Fig. 4.3a). In contrast, the percent cover of C. racemosa was highly seasonal since large differences between summer and winter were observed (Fig. 4.3b).

The relative frequencies of the different types of substratum in the total surface of sea bottom surveyed and of the presence of Caulerpa on each substratum are shown in Fig. 4.4 for each sampling event. Different distribution of frequencies indicates that Caulerpa species do not distribute randomly in the substrata (Table 4.2). Results from the χ2 test show that both Caulerpa species are found in the dead matte of P. oceanica and in rock on a higher proportion than expected in relation to the amount of substratum type available at the study sites. In contrast, Caulerpa species are found in a lower proportion than expected in sand and inside the P. oceanica meadow. Experimental χ2 were largely above critical values (Table 4.2).

The significant “Time x Substratum” interaction in both species (Table 4.3) indicates that substratum type affected the persistence of the two Caulerpa species. The temporal evolution of Caulerpa cover in the plots (Figs. 4.5a and 4.6a) shows that the persistence of Caulerpa was higher in the model of dead

4.3 Results

53

Figure 4.3: Percent cover of (a) Caulerpa taxifolia and (b) Caulerpa racemosa in the different types of substratum during two consecutive years.

matte of P. oceanica than in sand. C. taxifolia cover on dead matte increased until the 28th of September while C. taxifolia cover on sand did not change during the same period of time (Fig. 4.5a). After the 28th of September, C. taxifolia cover decreased with time in both substrata and by the end of winter no C. taxifolia was present in the plots. The cover of C. racemosa in the model of dead matte increased during the first five weeks of the experiment while on sand was roughly maintained (Fig. 4.6a). After August 30th, the C. racemosa experiment was disturbed by a storm that buried the plots with a 30 - 70 cm thick layer of drifting macroalgae and P. oceanica leaves causing the death of all C. racemosa fragments in the plots.

Seawater temperature in Cala Dâ&#x20AC;&#x2122;Or (Caulerpa taxifolia site) changed from 26 C in July to 14o C in March, while the number of daylight hours was reduced from 14.5 h to 9 h (Fig. 4.5b). Conditions in Sant Elm (Caulerpa racemosa site) during the experiment were rather constant with a seawater temperature o

54

Substratum effect on invasive Caulerpa species

Figure 4.4: Relative frequency of total substratum surveyed and of the presence of Caulerpa taxifolia and Caulerpa racemosa on each substratum. Low presence of rocks was recorded at the C. taxifolia site. Rocky bottom with photophilic algae is indicated as Rocks.

4.3 Results

55

Rock C. taxifolia

Winter 07 Summer 07 Winter 08 Summer 08

− − − −

C. racemosa

Winter 07 Summer 07 Winter 08 Summer 08

3− 80 M 1M 146 M

P. oceanica 64 20 23 77

L L L L

3− 879 L 34 L 796 L

Sand 36 79 65 32

L L L L

7M 142 L 16 L 2L

Dead matte P. oceanica 240 M 336 M 121 M 69 M 7 1153 91 629

M M M M

χ2 340∗∗∗ 465∗∗∗ 208∗∗∗ 178∗∗∗ 20∗∗∗ 2272∗∗∗ 142∗∗∗ 1573∗∗∗

Table 4.2: Chi-square (χ2 ) test of observed Caulerpa cover on different substrata. Caulerpa taxifolia has a critical χ2 p=0.001∗∗∗ , 2df = 13.81 and Caulerpa racemosa has a critical χ2 p=0.001∗∗∗ , 3df = 16.26. C. taxifolia was not found on rock substratum in enough quantities for the test. The type of contribution was classified as more than expected (M) or less than expected (L).

of 26o C and the number of daylight hours between 14 and 13 h (Fig. 6b).

Caulerpa taxifolia cover was maintained or increased with estimated nearbottom orbital velocities below 15 cm s−1 and declined above that value (Fig. 4.5b). Caulerpa racemosa was also able to maintain or increase its cover at the same velocity values (Fig. 4.6b). The vertical profiles of velocity indicate that the width of the boundary layer is around 10 cm (Fig. 4.7) showing also a significantly higher reduction of the horizontal rms velocity in the model of dead matte of P. oceanica (1.5 ± 0.2 cm s−1 , mean ± 95 % CL) than in sand (2.2 ± 0.2 cm s−1 , mean ± 95 % CL), (t-test: t = 115, df = 103598, p < 0.01).

The friction coefficient computed for the dead matte model was higher (1.04 x 10−1 ) than for sand (2.33 x 10−2 ). Increased friction by the dead matte model would reduce the magnitude of the velocities compared to sand. Wave data from the ADCP showed that wave conditions did not change during the ADV profile measurements, with Hs of 0.5 ± 0.1 m, Tp of 9.5 ± 0.5 s and South East direction.

56

Substratum effect on invasive Caulerpa species

df

MS

F

p

C. taxifolia Intercept Substratum Error Time Time x Substratum Residual

1 1 4 3 3 12

469997 17685 1600 5580 7086 1884

293 11

0.000∗∗∗ 0.029 ns

2.9 3.7

0.075 ns 0.041∗

C. taxifolia Intercept Substratum Error Time Time x Substratum Residual

1 1 4 3 3 12

41205 8321 1128 1790 1929 390

36 7

0.003∗∗∗ 0.053 ns

4.6 4.9

0.023∗ 0.018∗

Table 4.3: Results of repeated measures of ANOVA performed to evaluate if the cover of Caulerpa taxifolia and Caulerpa racemosa was different between a model of dead matte of Posidonia oceanica and a sandy substratum. ∗ p < 0.05, ∗∗∗ p < 0.001, ns = not significant. Cochran tests were not significant.

4.4

Discussion

The results of this Chapter indicate that substratum plays an important role in the distribution of these invasive Caulerpa species. Field measurements show that the presence of Caulerpa taxifolia and Caulerpa racemosa on the substrata is not at random. Dead matte of the seagrass Posidonia oceanica and rock covered with photophilic algae are more favourable than sand to the presence of Caulerpa species. Experimental results show that substratum has a significant effect on the persistence of both Caulerpa species and confirm that is higher in the model of dead matte of P. oceanica than in sand. Hence, the results of both analyses are consistent to each other and suggest that the type of substratum influences the abundance of these invasive Caulerpa species.

4.4 Discussion

57

Figure 4.5: (a) Caulerpa taxifolia cover (normalised means Âą SE, n = 3). (b) Water temperature at the experimental site, hours of light per day and nearbottom orbital velocities, ub (means Âą SD). Cover area on each sampling date (A) in cm2 normalised by initial cover area (A0). Dotted square indicates the period used for the statistical analysis ANOVA, Table 4.3.

Understanding the factors that regulate the establishment and spread of introduced Caulerpa species is important to predict future pathways of invasion as well as the susceptible areas to be invaded (Carlton and Geller 1993, Bax et al. 2003, Cebrian and Ballesteros 2009). Previous studies mainly based on direct observation indicated that favourable substrata for the establishment of Caulerpa spp. are dead matte of Posidonia oceanica (Piazzi and Cinelli 1999, Piazzi et al. 2001, Ruitton et al. 2005) and rocky bottoms covered by turf algae (Ceccherelli et al. 2002, Piazzi et al. 2003, Bulleri and Benedetti-Cecchi 2008). A propagation model of Caulerpa taxifolia assumed that the settlement probabilities of this species were high for dead matte, harbour mud, and rocks with photophilic algae and low for P. oceanica and unstable sand (Hill et al.

1998) but no explanation of how these probabilities were set is given. This Chapter provides the first quantitative evaluation of previous descriptions of dead matte of P. oceanica as being a favourable substratum for the colonization of invasive Caulerpa species. It also provides quantitative evidence that rocks covered with photophilic algae are also a favourable substratum for Caulerpa racemosa colonization. Sand and P. oceanica meadows are less favourable than

58

Substratum effect on invasive Caulerpa species

Figure 4.6: (a) Caulerpa racemosa cover (normalised means Âą SE, n 3). (b) Water temperature at the experimental site, hours of light per day and nearbottom orbital velocities, ub (means Âą SD). Cover area on each sampling date (A) in cm2 normalised by initial cover area (A0).

the dead matte of P. oceanica and rocks covered with photophilic algae for the colonization of C. taxifolia and C. racemosa.

The positive association between consolidated substrata (rocks and dead matte) and the presence of invasive Caulerpa species suggests that near-bottom orbital velocity may be an important determinant of the spatial distribution of these species at shallow depths. The capacity of Caulerpa to acquire nutrients from the sediment through the rhizoids (Williams 1984, Chisholm et al. 1996) might be invoked to explain the preferential distribution of invasive Caulerpa on dead matte of P. oceanica but not over rocky substrata where the amount of sediment is much lower than in dead matte. The structural complexity of the substratum facilitates the retention of Caulerpa taxifolia fragments (Davis et al.

2009) and could be a factor to explain why rocky substrata covered by an erect layer of photophilous macroalgae and the P. oceanica dead matte are favourable substrata for the establishment of Caulerpa. The calculated friction coefficients for the dead matte model and sand indicate that the surface of the dead matte is structurally more complex than sand and that Caulerpa fragments established on it will experience lower flow velocities than in sand.

4.4 Discussion

59

Figure 4.7: Vertical profiles of velocity (means Âą 95 % CL). For convenience only the 40 cm above the bottom is shown.

Caulerpa taxifolia and Caulerpa racemosa were rarely observed inside the P. oceanica meadow in this study and the Ď&#x2021;2 test shows that P. oceanica is not a favourable substratum for the establishment of the introduced Caulerpa species even though the matte of the P. oceanica meadow is a favourable substratum. Previous studies have also described that dense P. oceanica meadows are not invaded by C. taxifolia and C. racemosa in the short-term (Ceccherelli and Cinelli 1999, Ceccherelli et al. 2000). Dense P. oceanica meadows appear to resist the invasion of C. taxifolia for more than eight years (Jaubert et al.

1999). Results presented in this Chapter suggest that degraded, low-density P. oceanica meadows would be suitable locations for the settlement and growth of invasive Caulerpa species if shading by the leaf canopy was the main constraint of Caulerpa invasion (Ceccherelli and Cinelli 1999, Ceccherelli et al. 2000), and thus protecting existing meadows would likely reduce the spread of these species in the Mediterranean Sea.

60

Substratum effect on invasive Caulerpa species

There has been little effort to correlate wave and current energy with the spatial distribution and growth of invasive algae. The effects of wave exposure, morphology and drag forces of the invasive Codium fragile were studied under field and flume conditions by D’Amours and Scheibling (2007). These authors suggest that the C. fragile became dislodged at flow velocities exceeding 50 cm s−1 . Moreover, Scheibling and Melady (2008) found that a sheltered location was a more suitable habitat than an exposed location for this species. The effect of wave exposure on C. racemosa its not clear since it can be found in both exposed and sheltered locations (Thibaut et al. 2004, Klein and Verlaque 2008).

To date, there are no available data on near-bottom orbital velocities that allow the establishment and development of invasive Caulerpa species. Results in this Chapter show that C. taxifolia and C. racemosa do only occur at near-bottom orbital velocities below 15 cm s−1 and that C. taxifolia cover could decline at velocities above that value. This value, however, should not be considered as a velocity threshold for the persistence of C. taxifolia because the evidence provided by in this Chapter is correlative only and other factors such as seawater temperature and light availability declined as velocities increased in the experimental site (see Figs. 4.5b and 4.6b).

C. racemosa was able to colonize sand in summer but disappeared in winter, a season with higher wave energy. One of the potential problems for aquatic macrophytes growing on unconsolidated substrata is sediment movement and deposition that causes uprooting and burial (Fonseca and Kenworthy 1987, Williams 1988, Cabaco et al. 2008). Previous works described that C. racemosa was not colonizing fine sand with large ripple marks (Ruitton et al. 2005).

Sand with large ripples indicates that orbital velocities are or have been high enough to mobilize the sediment suggesting that only those macrophytes with high anchoring capacity could survive in that location.

Results presented in this Chapter provide novel quantitative evidence highlighting the importance of substratum for the distribution of invasive Caulerpa species at shallow depths. It is shown that the two introduced species of Caulerpa in the Mediterranean Sea are more abundant in dead matte of P. oceanica or rock than inside the P. oceanica meadow or sand. These results also suggest the quantitative hydrodynamic conditions that shape colonization as their cover started to decline at near-bottom orbital velocities over 15 cm s−1 . Further studies should be done to elucidate the validity of this velocity

4.4 Discussion threshold of the colonization of introduced Caulerpa.

61

62

Substratum effect on invasive Caulerpa species

Chapter 5

Effect of a seagrass (Posidonia oceanica) meadow on wave propagation Part of this Chapter has been submitted as an article on Marine Ecology Progress Series1

5.1

Summary

Wave damping induced by the seagrass Posidonia oceanica is evaluated using field data from bottom mounted acoustic doppler velocimeters (ADVs). Using the data from one storm event, the bottom roughness is calculated for the meadow as ks = 0.42 m. This roughness length is then used to predict the wave friction factor fw , the drag coefficient on the plant, CD , and ultimately the wave damping for two other storms. Values for CD ranged from 0.1 to 0.4 during both storms. The expected wave decay coefficient for seagrass shoot densities of 200 - 800 shoots mâ&#x2C6;&#x2019;2 is also predicted. The utility of the bottom roughness for calculating the friction factor and the drag coefficient of a seagrass meadow for conditions in which the meadow height is small compared to the water depth is 1 Infantes E, Orfila A, Terrados J, Luhar M, Simarro G, Nepf H. Effect of a seagrass (P. oceanica) meadow on wave propagation (Submitted to: Marine Ecology Progress Series)

63

64

Effect of a seagrass meadow on wave propagation

demonstrated.

5.2

Theory

The bottom boundary layer is a region adjacent to the bottom where the velocity field drops from the value in the core of the fluid to zero at the bed. When seagrasses are present near the bottom, the boundary layer is modified by the seagrass canopy, which influences the mean velocity, turbulence and mass transport (Nepf & Vivoni 2000, Ghisalberti & Nepf 2002, Luhar et al. 2010). A good understanding of the physics of the boundary layer is important because the damping of the waves is mainly due to the friction in the boundary layer where the seagrass is present. At this point, I want to remark that P. oceanica leaves occupy only a small fraction of the water column and under these conditions the drag associated with the meadow into the bed friction formulation are merged.

The main question to be solved in the analysis of the boundary layer under oscillatory flow is the relationship between the near bed velocity, ub , and the bed shear stress transmitted to the combined seagrass and bottom, τb . The dimensionless parameter relating both quantities is the wave friction factor defined as fw ≡ 2|τb |/ρu2b , where ρ is the fluid density. The friction factor fw depends on the Reynolds number and the relative roughness, defined respectively as Re ≡

u2b , νω

and

ε≡

ks ω . ub

Above, ν is the kinematic viscosity of the water, ω is the wave angular frequency (ω = 2π/T ) with T being wave period and ks is a length characterizing the bottom roughness, or Nikuradse length. In this approach, ks includes both the characteristics of the bottom and of the meadow. If the boundary layer is smooth (ks ub /ν < 5), then fw becomes a function of the Reynolds number only.

If the boundary layer is rough (ks ub /ν > 70), the friction factor depends mainly on the relative roughness ε (Schlichting 1979).

The friction factor proposed by Nielsen (1992) as a modification of the semiempirical formula of Swart (1974) for sandy bottoms is, ) ( 0.2 ks ω − 6.3 .

(5.1) fw = exp 5.5 ub

5.2 Theory

65

Accounting for signs, the definition of fw implies Ď&#x201E;b = Ď fw ub |ub | /2. This is known to be an oversimplification of the problem for oscillatory motions. For instance, it is known that the shear under monochromatic waves is not in phase with velocity. This has led to modifications of the friction factor to introduce the phase lag (Nielsen 1992), and to redefine the friction factor as (Jonsson 1966) 2Ď&#x201E;b,max fw â&#x2030;Ą 2 , (5.2) Ď ub,max where the subindex refers to the maximum value of the variable within a wave period. In fact, the expression Eq. (5.1) is actually valid for fw defined as in Eq. (5.2). More serious problems regarding the use of the friction factor appear for saw-tooth shaped waves or for wave-current interactions. In general, for oscillatory flows the use of the friction factor fw is recommended as a first order approximation of time averaged values.

Assuming that linear wave theory is valid and assuming straight and parallel bathymetric contours, the conservation of wave energy may be written as â&#x2C6;&#x201A;(Ecg ) = â&#x2C6;&#x2019; D , â&#x2C6;&#x201A;x

(5.3)

with E â&#x2030;Ą Ď gHs2 /8 being g the acceleration of gravity, Hs the significant wave height, D the energy dissipation and cg the group velocity given by

Ď&#x2030; 2kh cg = 1+ , (5.4) 2k sinh 2kh with k being the wave number (k = 2Ď&#x20AC;/L), where L is the wave lenght and h the local water depth.

Assuming a meadow composed by a set of N plants per unit of horizontal area where each plant has a leaf length given by lv and neglecting the swaying by blades, the energy dissipation D due to seagrass can be written (Dalrymple et al. 1984) as, Z â&#x2C6;&#x2019;h+lv 1 Ď CD a0v N |u|u2 dz.

(5.5) D â&#x2030;Ą 2 â&#x2C6;&#x2019;h where a0v is the plant area per unit height, u is the orbital velocity and CD is the drag coefficient for the forces exerted by the fluid on the plant. In Eq. (5.5)

66

Effect of a seagrass meadow on wave propagation

the integral is evaluated from the bed (z = â&#x2C6;&#x2019;h) to the vegetation height lv and the overbar indicates the time average over a wave period.

Previous theoretical, numerical and observational works on wave propagation over vegetated fields have dealt with the question of obtaining the dissipation term in order to integrate Eq. (5.3) (Kobayashi et al. 1993, M´endez & Losada 2004, Bradley & Houser 2009). Some authors have suggested that the morphology of the P. oceanica leaf can be simulated by a vertical cylinder (M´endez & Losada 2004). By doing so, an analytical expression for the dissipation term can be obtained under the hypothesis of linear wave theory (Dalrymple et al.

1984) as,

3 2Ď gkHs sinh3 klv + 3 sinh klv 0 D = CD av N .

(5.6) 9Ď&#x20AC;k 2Ď&#x2030; cosh3 kh Since one could alternatively approximate D as Ď&#x201E;b ub , assuming that the velocity distribution is nearly uniform within the plant height lv , it follows that, fw CD = 0 .

(5.7) a v lv N However, the main problem in the determination of the rate of energy dissipation arises from the ignorance of the value of the drag coefficient CD in Eq.(5.6).

Here, it is proposed to estimate CD from fw .

If wave propagation is studied over constant depth, an analytical solution can be found for the wave damping as, 1 Hs = Hs,0 1 + Kd Hs,0 x

(5.8)

where Kd is a decay coefficient given by, Kd =

4 sinh3 klv + 3 sinh klv CD a0v N k 9Ď&#x20AC; (sinh 2kh + 2kh) sinh kh

(5.9)

In this work the wave damping induced by a Posidonia oceanica meadow is evaluated. Data from bottom mounted acoustic doppler velocimeters for one of the three storms measured are used to obtain the roughness length ks for the meadow. The friction factor for two other storms are then calculated using Eq. (5.1) and the value of CD related with fw through the meadow parameters.

5.3 Material and Methods

67

This approach assumes that lv /h 1, so that the roughness length formulation accounts for the effects of both the bottom and the meadow. Finally, the wave damping is evaluated along a transect by numerical integration of Eq.(5.3).

5.3

Material and Methods

Measurements were carried out in Cala Millor, located on the northeast coast of Mallorca Island, Mediterranean Sea (Fig. 5.1, top). Cala Millor is an intermediate barred sandy beach formed by biogenic sediments with median grain values ranging between 0.28 and 0.38 mm at the beach front (G´omez-Pujol et al.

2007). The beach is in an open bay with an area of around 14 km2 . At depths from 6 m to 35 m, the seabed is covered by a meadow of Posidonia oceanica (Infantes et al. 2009). The tidal regime is microtidal, with a spring range of less than 0.25 m (Orfila et al. 2005). The bay is located on the East coast of the island, and it is therefore exposed to incoming wind and waves from NE to ESE directions.

From July 7th to July 23rd 2009, four self-contained Acoustic Doppler Velocimeters, ADVs (Nortek, Vector) with pressure sensors were deployed in a transect perpendicular to the coast at depths of 6.5, 10, 12.5 and 16.5 m and total length of 942 m (see Fig. 5.1, bottom). ADVs were mounted over galvanized iron structures taking care that pressure sensors were placed at 80 - 100 cm above the bottom, just above the seagrass canopy (Fig. 5.2). Stability of the equipment was verified with the compass, tilt and roll sensors (Infantes et al. 2011). Velocity data were collected in bursts of 15 minutes every two hours at 80 cm above the bottom at a sampling rate of 4 Hz, sampling volume of 14.9 mm and a nominal velocity range of 1 m sâ&#x2C6;&#x2019;1 .

Wave data was processed using Nortek software (QuickWave v.2.04) that uses the PUV method to estimate from pressure data the wave height (Hs ), peak period (Tp ) and the combination of the horizontal components and pressure to calculate wave direction (Nortek AS, 2002). Wave data was filtered to remove waves not approaching perpendicular to the beach. To exclude wave energy lost by white capping, wave records when mean wind velocities higher than 10 m sâ&#x2C6;&#x2019;1 were not used (Puertos del Estado wind data).

68

Effect of a seagrass meadow on wave propagation

Figure 5.1: (a) Location of Mallorca Island in the Mediterranean Sea. (b) Location of the study area. (c) Location of the transect and deployment sites in Cala Millor. (d) Bathymetric profile and distance between the deployment sites.

The area is under cyclogenetic activity throughout the year (CaË&#x153; nellas et al.

2007). Indeed, during the instrument deployment, three storms affected the area with significant wave heights (Hs ) larger than 1 m at the deeper location (Fig. 5.3, top). The first storm began on the 13th July with a duration of 2 days and with a maximum Hs of 1.85 m. The second storm began on the 18th July with a duration of 16 hours reached Hs of 1.58 m. The third storm began on the 21st July with a duration of 24 hours reached Hs of 1.1 m. The roughness length ks , was calculated independently from the third storm and then used to compute fw and the corresponding CD for the first and second storms.

5.3 Material and Methods

69

Figure 5.2: Acoustic doppler velocimeter (ADV) deployed in the Posidonia oceanica seagrass meadow.

5.3.1

Determination of Nikuradse roughness length

As stated, roughness length ks is a critical parameter for the characterization of the friction factor in rough turbulent flows. ks is assumed to be a property of the meadow morphology (number of shoots, shoot height and leaf width mainly), which in turn is assumed to be constant in time. This is a fair assumption since measurements were done in summer during a short period of time (16 days). P. oceanica leaf number, leaf lengths and leaf widths were measured in ten vertical shoots collected at the locations 1, 2, 3 and 4 (Fig. 5.1, top). The mean shoot length was set as lv = 0.8 ± 0.1 m (mean ± SE) and mean leaf surface area as 211 ± 23 cm2 (mean ± SE). This is the averaged maximum leaf length of individual shoots measured randomly at field. Shoot density was also measured at the same locations and the mean shoot density as 615 ± 34 shoots m−2 (mean ± SE).

Measured field velocities and wave heights along the transect during the

70

Effect of a seagrass meadow on wave propagation

Hs (m)

1.5 1 0.5

0 07

08

09

10

11

12

13

14

15 16 July

17

18

19

20

21

22

23

08

09

10

11

12

13

14

15 16 July

17

18

19

20

21

22

23

0.25

ub (m s!1)

0.2 0.15 0.1 0.05 0 07

Figure 5.3: Top panel: Significant wave height at 16.5 m depth (solid), at 12.5 m depth (doted), at 10 m depth (dash-doted) and at 6.5 m depth (dashed). Grey areas indicate the three storms. Bottom panel: near bottom orbital velocities at the same locations (depths and lines are the same as the top panel).

third storm (July 21st) are used to obtain ks in the following procedure. A set of CD ranging from 0.01 to 2 are provided to compute the wave damping by numerical integration of Eqs. (5.3)-(5.6). Departing from the wave height measured at location 1, the corresponding Hs,i at locations i = 2, 3, 4 (x = 337 m, x = 664 m and x = 942 m respectively) is computed obtaining the error between modeled and measured wave as, v u 4

2 uX computed measured , â&#x2C6;&#x2019; Hs,i =t

Hs,i

i=2

(5.10)

5.4 Results

71

The value of CD that minimizes the error is selected to compute fw and its corresponding ks through Eq. (5.1). ks is computed for a total of 20 hours (10 records) and the average for all records is calculated. The drag coefficient is related with the friction factor through the measured plant characteristics a0v and lv and N , Eq. (5.7). Therefore and for the sake of simplicity hereinafter the value of CD will be used.

5.3.2

Determination of drag coefficient

Once ks is determined for the Posidonia oceanica meadow, fw (ks /ab ) is computed for the other two storms. A total of 24 hours have been analyzed for the July 13th storm (12 records) and 18 hours for the July 18th storm (9 records).

For these events, near-bottom orbital velocities ranged between 12 and 25 cm sâ&#x2C6;&#x2019;1 (Fig. 5.3, bottom) with wave periods between 4.5 and 5.5 s during both storms. To obtain the friction factor from Eq. (5.1) it is assumed purely oscillatory motion. This is a fair assumption from the experimental data since mean velocities are one order of magnitude lower than the oscillatory component.

In order to study the energy dissipation within the P. oceanica meadow, the energy conservation Eqs. (5.3)-(5.6) are solved. Bathymetry is adjusted by fitting a third order polynomial to the depths at the four moorings (Fig. 5.1, bottom).

5.4

Results

During the three storms the significant wave heights Hs decreases between the depths 16.5 m to 6.5 m (Fig. 5.3, top). In contrast, but as expected by linear wave theory, near-bottom orbital velocities ub increase at shallower depths (Fig. 5.3, bottom). The roughness length obtained during July 21st is ks = 0.42Âą0.05, which is the mean and the standard deviation of the values estimated for stations 2, 3 and 4.

On July 13th at 4am significant wave heights of Hs = 0.86 m were measured at mooring 1 (16.5 m depth). Waves heights above 1 m were recorded for 24 hours with maximum Hs = 1.85 m. Measured significant wave height normalized by incident wave height (Hs,0 ) at the four moorings are displayed for this period in Fig. (5.4). Numerical integration of of Eq. (5.3) are plotted

72

Effect of a seagrass meadow on wave propagation

as black lines, and the uncertainty in the predictions, based on a 15 % error in the measurement of the initial wave heigh at period are displayed as dashed lines. For each set of data the aprori estimate of CD , based on ks and Eq. (5.1), is presented for completeness. Values of CD ranged from 0.12 to 0.43. As seen, fairly good agreement is obtained between the measured and predicted Hs . The second storm lasted 18 hours starting on July 18th at 12am. Fig. (5.5) present this event at a 2 hours interval. Similarly, the measured data (dots) are well represented by the predictions (lines) although some discrepancies appear at the shallow moorings. Hs /Hs,0 between the deepest (16.5 m) and shallowest (6.5 m) measurement locations is reduced by 50 % Âą 4 % (mean Âą SD) during the first storm and by 39 % Âą 9 % during the second storm.

As P. oceanica shoot density varies between locations (Marb`a et al. 1996), the effect of this variation on wave damping is assessed by computing the wave decay (Hs /Hs,0 ) over the bathymetry depicted in Fig. (5.1, bottom), but for different shoot densities. The wave decay for an incident wave of Hs,0 = 1.56 m and Tp = 4.6 s (CD = 0.41) for shoot densities ranging between 200 to 800 shoots mâ&#x2C6;&#x2019;2 is shown in Fig. (5.6). After propagation over 1000 m, Hs /Hs,0 is reduced 40 % in a meadow of 200 shoots mâ&#x2C6;&#x2019;2 . In a high density meadow of 800 shoots mâ&#x2C6;&#x2019;2 , the reduction is > 55 % over the same distance.

Wave decay over constant depth can be readily obtained from Eq. (5.8).

The coefficient Kd in Eq. (5.9) depends on plant parameters (e.g., a0v , lv , N and ks ), hydrodynamical conditions (e.g., CD , Ď&#x2030;) and water depth. For a fixed Ď&#x2030; = 1.3 sâ&#x2C6;&#x2019;1 representing a period, T â&#x2030;&#x2C6; 4.8 s, which is typical condition for the Mediterranean Sea and Hs,0 = 1.1 m (CD = 0.42) the value of Kd (N, h) is displayed as a function of the shoot density and the length of the blade in Fig. (5.7) for water depths of h = 5 m (top), h = 10 m (middle) and h = 15 m (bottom). The set of shoot densities were varied from sparse meadows of N = 200 plants/mâ&#x2C6;&#x2019;2 to very dense meadows of N = 1000 plants/mâ&#x2C6;&#x2019;2 . Blade length was varied from the length of seedlings, around 0.1 m, to blades of 1.4 m. The large values of Kd obtained at 5 m water depths have to be taken with caution. Two conditions are violated for this wave condition at this depth; first, the nonlinearity for the condition presented is = 0.1 which is too large to satisfy the hypothesis of linear wave theory, and second, the large values of lv /h suggest that the meadow cannot be characterised simply by the roughness length concept to obtain the drag coefficient. Values of Kd at 10 m depth are

73

0 0

100

Hs,0 = 1.08 m! " = 1.31 s-1!

CD = 0.56 !

100

0.5 0 0

Hs,0 = 1.11 m! " = 1.19 s-1!

CD = 0.42 !

1 0.5 0 0

100

Hs,0 = 1.46 m! " = 1.19 s-1!

CD = 0.31 !

0 0

Hs,0 = 1.85 m! " = 1.23 s-1!

CD = 0.21 !

1 0.5

Hs,0 = 1.53 m! " = 1.13 s-1!

CD = 0.17 !

0 0

100

1 0.5

Hs,0 = 1.50 m! " = 1.13 s-1!

CD = 0.16 !

0 0

100 kx

0.5 0 0

0.5 0 0

Hs,0 = 1.59 m! " = 1.18 s-1!

CD = 0.17 !

100 kx

200

Hs,0 = 1.56 m! " = 1.38 s-1!

CD = 0.31 !

100

200

kx

1 0.5 0 0

100

1

100

1

kx Hs/Hs,0

Hs/Hs,0

kx

Hs,0 = 1.1 m! " = 1.31 s-1!

CD = 0.55 !

kx

kx Hs/Hs,0

Hs/Hs,0

0.5

0.5 0 0

200

100

kx

1

1

kx Hs/Hs,0

Hs/Hs,0

kx

1

Hs/Hs,0

0.5

Hs/Hs,0

0 0

1

Hs/Hs,0

0.5

Hs,0 = 0.86 m! " = 1.27 s-1!

CD = 0.59 !

Hs,0 = 1.79 m! " = 1.13 s-1!

CD = 0.16!

100 kx

Hs/Hs,0

1

Hs/Hs,0

Hs/Hs,0

5.4 Results

1 0.5 0 0

Hs,0 = 1.31 m! " = 1.11 s-1!

CD = 0.16 !

100 kx

Figure 5.4: Normalized significant wave height (by incident) along nondimensional distance (by wave length) for the first storm. The first panel corresponds to conditions on July 13th at 4am and sequence are elapsed 2 hours. The solid black line displays the computed normalized significant wave height including the dissipation due to P . oceanica meadow. The grey line corresponds to the Hs assuming no dissipation by the seagrass. Dashed lines displays the predictions for wave decay for a 15 % error in the measurement of initial wave height and wave period.

74

Effect of a seagrass meadow on wave propagation

100

0 0

200

kx

Hs,0 = 1.22 m! " = 1.27 s-1!

CD = 0.20 !

100

0.5 0 0

200

kx

0 0

200

Hs,0 = 0.89 m! " = 1.45 s-1!

CD = 0.36 !

100 kx

200

0.5 0 0

Hs,0 = 1.58 m! " = 1.49 s-1!

CD = 0.19 !

100

200

kx

1 Hs,0 = 1.25 m! " = 1.51 s-1!

CD = 0.28 !

100

0.5 0 0

200

Hs,0 = 1.25 m! " = 1.44 s-1!

CD = 0.31 !

100

200

kx

1 Hs/Hs,0

Hs/Hs,0

0 0

0.5

kx

1 0.5

100

1 Hs/Hs,0

Hs/Hs,0

0 0

Hs,0 = 1.69 m! " = 1.30 s-1!

CD = 0.20 !

kx

1 0.5

Hs/Hs,0

0.5

Hs/Hs,0

0 0

Hs,0 = 1.31 m! " = 1.46 s-1!

CD = 0.45 !

1

1 Hs/Hs,0

0.5

1 Hs/Hs,0

Hs/Hs,0

1

Hs,0 = 0.91 m! " = 1.39 s-1!

CD = 0.44 !

100 kx

200

0.5 0 0

Hs,0 = 0.67 m! " = 1.31 s-1!

CD = 0.42 !

100

200

kx

Figure 5.5: Normalized significant wave height (by incident) along nondimensional distance (by wavelength) for the second storm. The first panel correspond to conditions on July 18th at 12am and sequence are elapsed 2 hours. The solid black line displays the computed normalized significant wave height including the dissipation due to P . oceanica meadow. The grey line corresponds to the Hs assuming no dissipation by the seagrass. Dashed lines displays the predictions for wave decay for a 15 % error in the measurement of initial wave height and wave period.

5.5 Discussion

75

1 0.95

0.99

0.9 0.8

0.7

0.6

Figure 5.6: Wave damping H/H0 along the transect for different conditions of shoot density. Initial wave height is Hs,0 = 1.56 m, ω = 1.38 s−1 and CD = 0.41.

between 5 · 10−4 m−2 for short blades in low density meadow to 2 · 10−3 m−2 for large blades in high density meadows which gives wave reduction of 20 % to 60 % in 500 m propagation respectively. Values of Kd at 15 m depth are roughly one order of magnitude smaller.

5.5

Discussion

This study suggests that for waves of height Hs,0 = 1.56 m and period T = 4.5 s, the decay coefficient ranges from Kd = 5 x 10−4 to Kd = 3 x 10−3 (Fig. 5.7) for a meadow of density N = 200 - 1000 plants/m2 occupying lv /h = 0.04 - 0.14 in water depth h = 10 m. In dimensionless terms, Kd Hs,0 L = 0.024 - 0.14, which can be interpreted as a 2.4 % to 14 % decay in wave height per wavelength (see Fig. 5.7). Bradley and Houser (2009) measured an exponential decay coefficient that ranged between 0.004 m−1 to 0.02 m−1 for waves of peak period 1.5 s in water depth ∼ 1.0 m, which equates to approximately a 1.3 % - 6.6 % reduction in wave height per wavelength (L ≈ 3.3 m). Further, Bradley and Houser (2009) suggest a bottom roughness length of ks = 0.16 m, which is broadly consistent with the value, ks = 0.42 m, obtained in this study. The lower roughness length obtained by Bradley and Houser (2009) may be explained by the fact that T. testudinum (lv ≈ 0.3 m) is shorter than P. oceanica (lv ≈ 0.8 m). Our results are also broadly consistent with Fonseca

76

Effect of a seagrass meadow on wave propagation

(

/+,-ĂŻ .

,&&& +&&

%"#$

%&!"#'

*&&

!"#$

)&& (&& !"#$% &'(

&')

&'*

&'+

,

,'(

)*+,-.+ /+,-ĂŻ .

,&&&

'"#$

%"#$

+&&

%&!"#$

*&& )&&

!"#0

(&& !",&$% &'(

'&!"#$

&')

&'*

&'+

,

,'(

)*+,-.+ (

/+,-ĂŻ .

,&&& +&& *&& )&&

HĂŻ

HĂŻ

%&'"#0

HĂŻ

%&0"#0 %"#0

HĂŻ

(&& !",#$% &'(

&')

&'*

&'+

,

,'(

) +,-.+ *

Figure 5.7: Decay coefficient (Kd ) versus seagrass shoot density (N) and leaf length (lv ) for an incident wave Hs,0 = 1.56 m, Ď&#x2030; = 1.38 sâ&#x2C6;&#x2019;1 and CD = 0.41 propagating over constant depth h = 5 m (top), h = 10 m (middle) and h = 15 m (bottom). Units in mâ&#x2C6;&#x2019;2 .

5.5 Discussion

77

and Cahalan (1992), who evaluated wave attenuation for four seagrass species (Zostera marina, Halodule wrightii, Syringodium filiforme and Thalassia testudinum) in a laboratory study. They found that, on average, wave energy was reduced by ∼ 40 % per meter when the length of seagrasses was equal to the water depth. The wavelengths tested by Fonseca and Cahalan were 0.38 m and 0.67 m, leading to a ∼ 15 % to ∼ 27 % reduction in wave energy per wavelength.

Seasonal changes in the seagrass meadow (shoot density, leaf number and leaf lengths) increases or reduces the degree of wave damping. This work suggests that for a given wave conditions (Hs,0 = 1.56 m, ω = 1.38 s−1 and CD = 0.41) Hs /Hs,0 is reduced by > 55 % in a dense meadow (800 shoots m−2 ) and a 40 % reduction in a low density meadow (200 shoots m−2 ) over the same distance (Fig. 5.6). Newell & Koch (2004) observed that wave height reduction was higher during June when the seagrass canopy of Ruppia maritima was dense (> 1000 shoots m−2 ) and occupied most of the water column than during October. P. oceanica shoot densities do not vary along the year, but leave number and length is higher in summer. Instead, P. oceanica shoot densities vary between locations and depth (Marb`a et al. 1996, Pergent-Martini et al.

1994, Terrados & Medina Pons). Wave damping would be higher in shallow locations with higher shoot densities. Anthropogenic disturbances affecting the shoot density of the meadow could later revert in a change in the upper limit by the increase of wave energy reaching the shallow areas of the beach (Infantes et al. 2009).

The energy dissipation term in Eq. (5.6), obtained by Dalrymple et al.

(1984), considers vertical extent of plants as rigid cylinders. As already noted by these authors, a proper characterization of CD is crucial in order to cover the ignorance on the plant movement. Since the value of the drag coefficients is estimated directly from wave height measurements, plant motion is accounted for to a certain extent. However it would be desirable to perform specific experiments to characterize the importance of the shoot swaying on the determination of CD .

The approach followed in the present study is based on the hypothesis that hydrodynamics is mainly induced by waves. For the experiments presented in this Chapter, this is a reasonable assumption for the two storms analysed since near-bottom orbital velocities measured by the ADVs are one order of magnitude larger than mean currents. In coastal environments, where the flow is the

78

Effect of a seagrass meadow on wave propagation

result of the interaction of waves and currents, the boundary layer interacts directly with the bottom roughness and the turbulence generated near the bottom by the orbital motion increasing the apparent roughness. In those situations, one has to solve explicitly the boundary layer by using a specific model (Grant & Madsen 1979, Simarro et al. 2008).

0.9

0.8

0.7

0.6

CD

0.5

0.4

0.3

0.2

0.1

0 30

40

50

60

70

80

90

100

KC

Figure 5.8: Drag coefficient versus Keulegan-Carpenter number for experimental data -crosses-, for S´ anchez-G´ onzalez et al. (2011) -black solid-, Kobayashi et al.

(1993) -black dashed- and for M´endez and Losada (2004) -grey-.

M´endez & Losada (2004) pointed out that a suitable relation is obtained between the drag coefficient and the Keulegan-Carpenter (KC) number defined as K = uc Tp /lv where uc is a characteristic velocity acting over the plant. The term ab /ks used in this Chapter is equivalent to the KC number used by those authors. The data presented in this Chapter and several recent studies show similar trends in CD with KC (Fig. 5.8). S´ anchez-G´onzalez et al. (2011) studied wave attenuation due to seagrass meadows in a scaled flume experiment using artificial models of P. oceanica and conducted that CD is better related with KC than with Reynolds number. Specifically, these authors found that

5.5 Discussion

79

CD = 22.9/K 1.09 for 15 ≤ K ≤ 425. Fig. (5.8) shows the drag coefficient predicted for the two storms considered in this study versus the KC number (dots). As seen, good agreement is obtained between the present approach and the expression given by S´ anchez-G´ onzalez et al. (2011), shown as a black line in Fig (A.3). Kobayashi et al. (1993) measured CD = 35.5/K 1.12 (black dashed line in Fig. 5.8). M´endez and Losada (2004) measured CD = 6.7/K 0.73 (grey line in Fig. 5.8).

Two aspects should be considered in detail in future research. First, the link between the roughness length and meadow properties has to be explored for different meadow conditions. Second, the limitation of this approach for increasing non-dimensional blade length (lv /h) has to be assessed. Understanding the interaction between waves and bottom canopies such as P. oceanica seagrass meadow is crucial for assessing the importance of these communities in coastal protection as well as to determine the final wave parameters which will drive sediments. This work shows that a P. oceanica seagrass meadow reduces the wave height that is reaching the beach during two storm events. Parameters such as ks and CD are necessary in order to run more precise wave propagation models over a seagrass meadows. Moreover, for meadows that occupy a small fraction of the water depth, it may be appropriate to use relations for bare beds, specifically Eq. (5.1), to characterize the drag imparted by the canopy.

80

Effect of a seagrass meadow on wave propagation

Chapter 6

General discussion and conclusions This Ph.D Thesis shows that hydrodynamics and bottom type play and important role on the spatial distribution of a seagrass meadow of Posidonia oceanica, the invasive macroalgae Caulerpa taxifolia and Caulerpa racemosa and seagrass seedlings of P. oceanica and Cymodocea nodosa. The coastal areas studied in this Thesis are exposed during winter to storm events with wave heights of 3 to 5 m. During these storms near-bottom orbital velocities are elevated and can shape the distribution of macrophytes. Testable estimates of the threshold level of wave energy that allows macrophyte species to persist are provided for two Mediterranean seagrass species and two invasive macroalgae. This Chapter will summarise the various experiments undertaken and link them together.

Previous studies have related the minimum depth of occurrence of macrophytes to wave action. Chambers (1987) calculated the wave conditions from fetch data in lakes obtaining a relationship between minimum depth of macrophyte colonization and wave mixing depth. Vacchi et al. (2010) related the distance of the upper limit of P. oceanica meadow from the shoreline and the morphodynamic domain of the beach. Taking into account sedimentological parameters, coastal dynamics was correlated with the position of the P. oceanica meadow upper limit and shoot density of its shallow portions. The methodology presented in Chapter 2 allows to estimate the near-bottom orbital velocities that set the upper depth limit of a Posidonia oceanica meadow. Mean wave conditions are propagated to the shore using a 2D parabolic model over the detailed 81

82

General discussion and conclusions

bathymetry. A correlation between hydrodynamics, bottom type and the spatial distribution of P. oceanica shows that the percent cover of seagrass over sand is reduced with increasing orbital velocities up to 38 - 42 cm sâ&#x2C6;&#x2019;1 . Above these velocities, there is no presence of P. oceanica meadow and therefore they set the upper limit of the bathymetric distribution of this species in this location and for the associated wave conditions. These results shows that there is a relation between the wave orbital velocities and the upper limit of seagrass.

This has been previously hypothesized (Koch et al. 2006) but it is the first time that near-bottom orbital velocities are quantified, setting a threshold value for the presence of this seagrass species in unconsolidated substratum. P. oceanica growing in rocks or dead matte likely supports higher velocities than those when it grows in sand.

Water motion plays an important role in the initial life stages of seagrass (Koch et al. 2010, Rivers et al. 2011) because seedlings are vulnerable to burial or dislodgement. The basis for understanding seedling survivorship of two Mediterranean seagrass species in sandy substratum under different hydrodynamic conditions is provided in this Thesis. Morphological characteristics of each species such as foliar surface and root length are key factors in colonisation processes. P. oceanica has larger leaves and shorter roots compared to C. nodosa. Drag forces are higher in P. oceanica under both unidirectional and oscillatory flow, due to the larger leaf surface area of this species compared to C. nodosa. In contrast, C. nodosa tolerates larger sediment erosion before becoming dislodged from the sediment than P. oceanica due to its longer roots.

The drag coefficients of individual seedlings obtained in a biological flume under waves ranged between 0.01 and 0.1. Drag measurements and seedling dislodgement results obtained in the flume agree with those obtained in the field where C. nodosa survived longer than P. oceanica. These results are important to understand the spatial distribution of these species in sandy substratum.

The strong decline of P. oceanica and C. nodosa has generated a major interest in restoring both Mediterranean species (Molenaar and Meinesz 1995, Balestri et al 1998). Understanding the habitat requirements of seedlings is necessary before a large-scale seagrass restoration plan is carried out. Results obtained in this Thesis for seagrass seedlings could be applied to restoration purposes. Macrophytes morphology and rooting capacity could determine their success under physical disturbances. This Thesis indicates that seagrass seedlings

83 were dislodged during storms. The disappearance of P. oceanica suggests that water dynamics was too severe at these depths for this species. In contrast, C. nodosa was better adapted to stand higher orbital velocities. Restoration success in sand will be improved if planting of P. oceanica is carried out in a relatively wave sheltered location with low orbital velocities. P. oceanica seedlings started to decline at near-bottom orbital velocities between 7 and 18 cm s−1 and completely disapeared at velocities between 17 and 34 cm s−1 . C. nodosa seedlings started to decline at velocities between 17 and 34 cm s−1 and some seedlings survived velocities between 23 and 39 cm s−1 . Anchoring the seedlings to the substratum could improve the rooting capacity and their survivorship in unconsolidated substratum under different hydrodynamic conditions, although this has not been assessed in this Thesis. Planting in consolidated substrata such as dead rhizomes of P. oceanica could also increase the seedling rate of success during restoration (Balestri et al. 1998, Infantes et al 2011, Dominguez et al. 2011, Rivers et al. 2011).

Wave exposure determines the morphology of the organisms. Many studies have documented the general trend of decreasing plant size with increasing wave exposure in intertidal systems (Gaylord et al. 1994, Blanchette 1997, Martone and Denny 2008). Hydrodynamic forces are the mechanism responsible for these observed patterns of size with respect to exposure. As seen in Chapter 3, drag forces are related to leaf surface area. The small size of C. nodosa will be an advantage over the large size of P. oceanica when colonising wave exposed locations. Fast-growing, colonizing species are often small (e.g. C. nodosa), whereas slow-growing or climax species are often large (e.g. P. oceanica). Results obtained in this Thesis agrees with the general assumption that small seagrass species occupy frequently wave disturbed habitats, while large seagrass species require more stable environments (Marb`a et al. 1996), which probably become even more stable when supporting a continuous seagrass cover through current reduction, wave attenuation and sediment trapping.

Assuming that physical dislodgement is the primary mechanism for macrophytes failure, an appropriate area to grow and persist would be one that enhances anchoring ability or reduces physical disturbance. Novel quantitative evidence highlighting the importance of substratum for the distribution of invasive Caulerpa species at shallow depths is provided in this Thesis. Observations carried out during two years showed that both Caulerpa species are more abun-

84

General discussion and conclusions

dant in dead matte of P. oceanica or rock than inside the P. oceanica meadow or sand. A short time experiment where fragments of both Caulerpa species where fixed to sand and dead matte also showed that cover of Caulerpa was higher in dead matte than on sand. C. taxifolia and C. racemosa cover started to decline at velocities above 15 cm s−1 . The friction coefficient computed for the dead matte model was higher (1.04 x 10−1 ) than for sand (2.33 x 10−2 ).

The calculated friction coefficients for the dead matte model and sand indicate that the surface of the dead matte is structurally more complex than sand and that Caulerpa fragments established on it will experience lower flow velocities than in sand. These results are important because indicate which substrates will be more suitable to be colonised by the invasive Caulerpa species.

The importance of hydrodynamics on the spatial distribution of macrophytes ecosystems is clearly shown in Chapters 2 to 4. On the other hand, the presence of macrophytes reduce current velocity and damping wave heights (Nepf and Vivoni 2000, Ghisalberti and Nepf 2002, Fonseca and Cahalan 1992). There is a general agreement that macrophytes can attenuate waves, but little quantitative evidence is available. Field data on wave damping in macrophytes is scarce as most studies have been performed using numerical models (Kobayashi et al. 1993, M´endez et al. 1999, M´endez and Losada 2004, Chen et al. 2007) and flume data (Fonseca and Cahalan 1992, Kobayashi et al. 1993, M´endez and Losada 2004, Augustin et al. 2009, S´ anchez-Gonz´alez et al. 2011). This Thesis shows that a P. oceanica seagrass meadow attenuates wave height during two storm events. The bottom roughness is calculated for the meadow as ks = 0.42 m using flow velocities above the seagrass. Wave heights are reduced between 39 and 50 % by a meadow over a distance of 942 m. Computed drag coefficients CD , ranged from 0.1 to 0.4 during two storms. These values agree with the CD values obtained in a flume using artificial plant models of P. oceanica (S´ anchez-G´ onzalez et al. 2011). The predicted wave decay coefficient for seagrass shoot densities of 200 - 800 shoots m−2 ranged between 40 % to 55 % of wave reduction respectively. This Thesis indicates that parameters such as bottom roughness ks , wave friction coefficient fw and drag coefficient CD are necessary in order to run more precise wave propagation models over a seagrass meadows.

In a case scenario of global warming and sea-level rise, the shallow areas will become deeper and orbital velocities reduced. A change in depth due to sea-

85 level rise could modify positively the upper limit as new unoccupied areas would become potentially available for colonization, but findings by Wicks (2005) at the Chincoteague Bay suggest that Zostera marina seagrass distribution might be negatively affected by sea level rise as seagrasses would be unable to migrate shoreward due to unsuitable sediments. Another consequence of global warming is an increase in the number and intensity of extreme storm events. These events could have serious implications on the spatial distribution of macrophytes and the upper limit of seagrass. As seen in Chapter 2, the upper depth limit of seagrass is determined by hydrodynamics. An increase in wave energy reaching the shallow areas of a meadow could revert on a deeper position of the upper depth limit.

Hydrodynamics effects operate on several scales from molecules (Âľm) to leaves and shoots (cm), seagrass canopies (m) and seagrass meadows (km), therefore having complex implications for entire macrophyte systems (see review Koch et al. 2006). During this Thesis, hydrodynamic effects have been studied at the scales of individual shoots, seagrass canopies and seagrass meadows. To understand these processes a link between different fields must be build.

Quantitative studies on interactions between waves and seagrasses are fairly recent (last â&#x2C6;ź 10 years). New, smaller and autonomous wave-current sensors (e.g.

Doppler technology) can be easily deployed in shallow coastal areas, providing high quality wave data. The continuous development of new computers with higher computational power can nowadays run complex numerical models that computers made 15 years ago were not able to run.

This Thesis highlights the importance of interdisciplinary and multidisciplinary research in ecological modeling and, in particular, the need for hydrodynamical studies to elucidate the distribution of macrophytes at shallow depths.

A relation between fluid dynamics, bottom type and benthic macrophytes is shown. Predicted and measured quantitative results are provided such as nearbottom orbital velocities and drag coefficients that could be tested and compared to other benthic marine macrophytes species on other locations. This Thesis demonstrates that water motion is an important parameter in seagrass ecology and requires serious considerations in seagrass research, conservation and restoration programmes.

86

General discussion and conclusions

Chapter 7

Recomendations for future work This Ph.D. Thesis shows that the upper depth limit of a seagrass meadow of Posidonia oceanica is related to hydrodynamics. Future work will be devoted to study the near-bottom velocities that determine the upper limit of Posidonia oceanica meadows in other locations in the Mediterranean Sea. The relation between hydrodynamics and the upper limit has been applied in a unconsolidated subtrata such as sand. This methodology could be also applied to P. oceanica colonizing consolidated substrates such as rocks. It can also be applied to other benthics species i.e., seagrass, macroalgae, sponges, bivalves.

There are few numerical models studying flow interactions with sediment and macrophytes. The complexities involved in the calculations involve an understanding of physical, biological and geological parameters (Teeter et al. 2001).

One of these limitations in the numerical calculations is the information on the friction coefficient due to the bottom topography and drag caused by the vegetation. In order to improve the accuracy of ecological models, the friction coefficient and bottom roughness must be first determined. Bottom roughness and drag coefficients values obtained for P. oceanica meadows can be included in wave propagation models. Additionally, wave induced movement of segrass canopy and its effects on wave damping remains unknown. Several numerical models have been developed to account for leaf swaying (M´endez et al. 1999, Bradley and Houser 2009) but more effort needs to be made in this topic.

87

88

Recomendations for future work

Most studies have dealt with interactions of either current or waves on macrophytes. In nature, both types of flow are usually present in a higher or lower degree. To date, little is known about the interaction of macrophytes with the combination of current and waves. This type of studies will be first carried out in hydraulic flumes where the current flow, wave conditions and meadow parameters can be easily adjusted.

Macrophytes distribution is influenced by both sediment inputs and hydrodynamics (Frederiksen et al. 2004). Studies have been carried out to determine interactions between hydrodynamics and sediment movements such as bars, rip´ ples, sand waves, beach profiles, etc (Alvarez-Ellacuria et al. 2010, G´omez-Pujol et al. 2011). Some studies have carried out on interactions between vegetation and hydrodynamics, but little work has been devoted to study interactions between hydrodynamics, sediment dynamics and macrophytes.

Appendix A

Background review on fluid dynamics Fluid dynamics is the study of the movement of fluid. Among other things it addresses velocity, acceleration, and forces exerted by or upon fluids in motion.

In this Appendix, a review is made on the basic concepts of fluid dynamics that affect the life of macrophytes. Linear wave theory, wave propagation and benthic boundary layers are summarised.

A.1

Waves

Water flow in coastal marine ecosystems can be divided into unidirectional and oscillatory flows. Unidirectional flows (also described as currents) are those in which water particles tend to move in the same direction over time. These flows in marine ecosystems are mainly generated by tides. Oscillatory flows (also described as waves) are those in which water particles move in two directions at a periodic interval. These flows are generated by waves. Depending on the dominance of wave periodicity, at marine ecosystem can be characterized as tide-dominated or wave-dominated. Because this Thesis is mainly focused on wave-dominated flows, for simplicity, the currents and tides characteristics will not be reviewed in this Appendix.

Wind generated waves are formed by the frictional stress produced by two fluid layers of different speed which creates a transfer of energy from the wind 89

90

Background review on fluid dynamics

to the water surface. Wind generated waves are surface waves that occur on the ocean, sea, lake, rivers, and canals or even on small puddles and ponds. Wave size depends on the wind speed blowing over a distance of sea surface (known as fetch) for a lenght of time.

Wave data can be obtained by direct measurements of the sea conditions using oceanographic instruments or estimated using numerical models. Several oceanographic instruments are used to measure waves such as pressure sensors, accelerometers, acoustic dopplers, radars or altimeters mounted on satellites.

Instruments can be deployed anchored at the sea bottom or fixed at meteorological buoys or at the shore. Wave records are generally very limited in space and time due to the cost of the instruments and measuring platforms. Also, waves are usually recorded on a specific location over a period of time. For these reasons, numerical models are used to estimate the wave conditions in locations where wave measurements are not available. In order to obtain a wave field over larger portions of the sea surface area where wave measurements are not available, numerical models are used. Numerical models are widely used, for example, to estimate wave fields over the Mediterranean Sea (Ca˜ nellas et al.

2007), to study a specific coastal area with an ecological, economical or turistic ´ ´ interest (Alvarez-Ellacuria et al. 2009, Alvarez-Ellacuria et al. 2010).

Waves are characterized by a wave height H (from trough to crest) or wave amplitude a (half of the wave height), wavelength L, (from crest to crest), wave period T , (time interval between arrival of consecutive crests at a stationary point) and wave propagation direction (Fig. 2). Waves in a given area typically have a range of heights. For weather reporting and for scientific analysis of wind wave statistics, their characteristic height over a period of time is usually expressed significant wave height, Hs , which represents an average height of the highest one-third of the waves in a given time period. The angular frequency is ω = 2π/T , and the wave number is k = 2π/L. Airy wave theory or linear wave theory gives a linear system description of the wave propagation of gravity waves on the surface of a homogeneous fluid layer. The theory assumes that the fluid layer has a uniform mean depth, and that the fluid flow is inviscid, incompressible and irrotational. Linear wave theory uses a potential flow approach to describe the motion of gravity waves on a fluid surface. In nature, a combination of waves and currents are present, but

!"#$%&"'()'*+%

,%

# ! = $" # !

!

"#$!

A.1 Waves

91

in watts per unit of wave crest length.

! !

#

$

"

! ! !

!

! ! ! %&'()*!#+!!,-.*/01&-!2&0')0/!34!0!5.306&7'!809*!05!&1!0::)30-.*5!5.3)*+!!;0)&0<6*!70/*5!0)*!

2*4&7*2!&7!1.*!1*=1+!!>5!1.*!2*:1.!!!<*-3/*5!5.06638*)!7*0)5.3)*?!1.*!5:**2!34!1.*!809*!

Figure A.1: Diagram of a shoaling wave approaching to shore. Varible names 2*-)*05*5+!!@.*!809*!:*)&32!-07731!-.07'*?!53!&7!3)2*)!43)!1.*!5:**2!13!2*-)*05*!1.*!

809*6*7'1.!/(51!2*-)*05*+!!@.*!809*!.*&'.1!/(51!0653!&7-)*05*!13!-375*)9*!809*!:38*)?! are defined in the text.

(76*55!2&55&:01&37!&5!5&'7&4&-071+!!@.*!809*!51**:*75!05!0!)*5(61!34!<31.!-.07'*5?!072! *9*71(066A!<)*0B5+!

Note thatwe !" is not necessarily the same as the wave phase speed c shown for simplicity have deal with them independently. in Figure 4. In general, the group velocity is given by:

Assuming linear wave theory, theª dispersion relationship in deep and shallow º "π % « » water describes the field of propagating waves as # & !

$" =

«# + »! $« § "π · » %&'( ¨ %¸ ω = «¬gk tanh© (kh), & ¹ »¼ 2

"C$!

(A.1)

wherewhere g is the of gravity andtranscendental h is the water depth. Substituting ω L isacceleration given by the solution of the function: and k, Eq. (A.1) can also be written as: $

!

$π % · § $π · 2$π " § ¨ ¸ T= g ¨ )*'(2πh ¸ ! ' & & )¹ tanh( ©L =¹ ©

2π

L

"D$!

(A.2)

Waves approachingwaves to the coast different and For shallow-water with #$% travel < 0.04,through where # is the waterwater depth,depths the phase speed group velocitydeep, are theshallow same and depend only on water are classified intoand three regimes; and intermediate depth waves. depth: The speed of the wave propagating, C, can be expressed as !

g = "% ! " = $tanh(kh) C 2 $= k

"E$!

(A.3)

In deep water, kh is large and tanh(kh) = 1, therefore L = T 2 g/2π or p C = g/k. For a water depth larger than half the wavelength, h > L/2, the orbits are circular and the phase speed of the waves is hardly influenced by ! depth (this is the case for most wind waves on the sea and ocean surface). √ In shallow water, kh is small and tanh(kh) = kh, therefore C = gh. For a water depth smaller than the wavelength divided by 20, h < L/20, the orbits become elliptical or flatter due to the influence of the bottom. The phase speed

92

Background review on fluid dynamics

sinh kh

cosh kh kh = kh

tanh kh

kh 1/20

Shallow water waves

h/L

Intermediate waves

1/2

Deep water waves

Figure A.2: Relative depth and asymptotes to hyperbolic functions.

of the waves is only dependent on water depth, and no longer a function of periodic function or wavelength. All other cases, L/20 < h < L/2, are termed intermediate depth waves, where both water depth and period (or wavelength) have a significant influence on the solution of Airy wave theory. In this case waves combine characteristics of deep and shalow water waves and speed is described as Eq. (A.3).

Assuming linear wave theory, deep and shallow waves differ in characteristics other than wave velocity. In deep-water waves, the individual particles describe circles whose radius decreases with depth. In the surface the radius is the same

A.2 Wave propagation

93

as the wave amplitude, and the particle velocity is the circumference of the circle divided by the wave period. In shallow-water waves the particles describe ellipses. The radius along the minor axis is equal to the wave amplitude at the surface and decreases linearly with depth. At the bottom the minor axis is zero and the motion is horizontal. The radius of the major axis is a function of water depth, wavelength and amplitude. In this Thesis, near-bottom orbital velocity, ub , is given by ub = 2Ď&#x20AC;ab /T (A.4) where ab is the wave orbital amplitude calculated as ab =

Hs , 2sinh(2Ď&#x20AC;h/L)

(A.5)

where L is calculated iteratively from Eq A.2.

Orbital velocities can be also measured from instruments deployed in the seafloor. Records from acoustic doppler velocimeters (ADV) provide large datasets to calculate velocity and turbulence. Field records are usually filtered to remove spikes and low beam correlations. Then velocity data can be filtered with a low pass filter to remove high frequency Doppler noise (Lane et al. 1998). Horizontal velocity is calculated as the root mean square (rms). v u N u1 X urms = t x2 + yi2 (A.6) N i=1 i where x and y are the two horizontal velocities.

Wave energy possessed by a wave is in two forms, kinetic and potential.

Kinetic energy is the energy inherent in the orbital motion of the water particles while potential energy possessed by the particles when they are displaced from their mean position. The total energy E per unit area of a wave is, E=

A.2

1 Ď gHs2 .

8

(A.7)

Wave propagation

As water waves propagate from the region where they were generated to the coast, both wave amplitude and wavelength are modified due to refraction (e.g.

94

Background review on fluid dynamics

changes in the bathymetry or interactions with wind induced currents), diffraction (e.g. intense variations of the bottom), the loss of energy due to the shoaling, damping and finally wave breaking. The surf zone is a highly dynamic area where energy from waves is partially dissipated through turbulence in the boundary layer and transformed in short and long waves, mean sea level variations and currents. Therefore, the energy dissipated in the surf zone is used for sediment transport providing a highly variable morphological environment (Dean and Dalrymple 2002).

Early efforts to study wave transformation from deep to shallow water were based on the geometrical ray theory. As a modification to the linear wave theory, the mild slope approach which assumes that the water depth changes slowly in a typical wave length appears as an improvement to the former since it includes wave diffraction. In this approach, the vertical structure of the velocity is the same as that for a progressive wave over a constant depth being the governing partial differential equations of the elliptic type. Moreover, these equations can be easily converted into parabolic type by assuming that the wave amplitude is primarily a function of water depth due to the shoaling. This approach, known as the parabolic approach, can be seen as a modification of the ray theory where wave energy can be diffused along the rays as wave propagate (Liu and Losada 2002).

In this Thesis, deep-water wave conditions were propagated to the study sites using a numerical model (Oluca) based on the mild slope parabolic approximation (Kirby and Dalrymple 1983) and was developed by the University of Cantabria, Spain. The Oluca model is a variation of the REF/DIF model (Kirby and Dalrymple 1983). These models are widely used in engineering and scientific approaches in order to obtain the accurate wave field in coastal areas from open sea conditions since refraction and diffraction are well captured ´ (Alvarez-Ellacuria et al. 2010).

A.3

Boundary Layers

The boundary layer is that layer of fluid in the immediate vicinity of a bounding surface where effects of viscosity of the fluid are considered in detail. The boundary layer effect occurs at the field region in which all changes occur in the

A.3 Boundary Layers

95

flow pattern. The boundary layer distorts surrounding non-viscous flow. It is a phenomenon of viscous forces. This effect is related to the Reynolds number.

The thickness of the boundary layer, δ, is normally defined as the distance from the solid body at which the flow velocity is 99 % of the free stream velocity, that is, the velocity that is calculated at the surface of the body in an inviscid flow solution.

When benthic marine macrophytes are present, the boundary layer is modified by the canopy which influences the mean velocity, turbulence and mass transport (Nepf and Vivoni 2000, Ghisalberty and Nepf 2002, Luhar et al.

2010). Macrophytes reduce the flow velocity near the bottom changing the logarithmic velocity profile. The goal of studying benthic boundary layers in ecological studies is to understand the water movement around organisms that live on or near the substratum.

Figure A.3: Vertical profile of velocity. Top, weak currents (black line) and fast currents (dashed line). Bottom, velocity reduction due to benthic macrophytes.

96

Background review on fluid dynamics

Water flow can either be smooth and regular as the fluid flows in layers (laminar flow) or rough and irregular (turbulent flow). This depends on the velocity and the length scale (temporal and spatial scales respectively) and is defined by the Reynolds number. The Reynolds number Re is a dimensionless number that gives a measure of the ratio of inertial forces ρν 2 /l∗ to viscous forces µν/l∗2 and consequently quantifies the relative importance of these two types of forces for given flow conditions and is expressed as Re =

ul∗ , ν

(A.8)

where u is the flow velocity, l∗ is the characteristic length and ν is the kinematic water viscosity.

Macrophytes are a source of drag, they reduce near-bottom water flow, and dissipate current and wave energy. Drag forces act in a direction opposite to the oncoming flow velocity. Drag on macrophytes increases with water velocity and foliar surface area. In flexible organisms, foliar surface area will change with increasing flow velocity, and this change is encapsulated in the drag coefficient.

Drag force is expressed as: FD =

1 ρaf CD u2 2

(A.9)

where ρ is density of the water, af is the frontal area, CD is the drag coefficient and u is the water velocity. The drag coefficient CD is a dimensionless parameter used to quantify the drag or resistance of an object in a fluid environment. It is used in the drag equation, where a lower drag coefficient indicates the object will have less hydrodynamic drag. The drag coefficient is always associated with a particular surface area.

Bibliography Ackerman JD and Okubo A (1986) Reduced Mixing in a Marine Macrophyte Canopy. Funct. Ecol. 7(3): 305-309. ´ Alvarez-Ellacuria A, Orfila A, Olabarrieta M, G´omez-Pujol L, Medina R, Tintor´e J (2009) An Alert System for Beach Hazard Management in the Balearic Islands. Coast. Manage. 37: 569-584 ´ Alvarez-Ellacuria A, Orfila A, Olabarrieta M, Medina R, Vizoso G, Tintor´e J (2010) A nearshore wave and current operational forecasting system. J. Coast.

Res. 26: 503-509.

Anderson K, Close L, DeWreede RE, Lynch BJ, Ormond C, Walker M (2006) Biomechanical properties and holdfast morphology of coenocytic algae (Halimedales, Chlorophyta) in Bocas del Toro, Panama. J. Exp. Mar. Biol. Ecol.

328: 155-167.

Augustin LN, Irish JL, Lynett P (2009) Laboratory and numerical studies of wave damping by emergent and near-emergent wetland vegetation. Coast.

Eng. 56: 332-340.

Balestri E, Piazzi L, Cinelli F (1998) Survival and growth of transplanted and natural seedlings of Posidonia oceanica (L.) Delile in a damaged coastal area.

J. Exp. Mar. Biol. Ecol. 228: 209-225.

Bax N, Williamson A, Aguero M, Gonzalez E, Geeves W (2003) Marine invasive alien species: a threat to global biodiversity. Mar. Policy 27: 313-323.

Bell SS, Robbins BD, Jensen SL (1999) Gap dynamics in a seagrass landscape.

Ecosystems 2(6): 493-504.

Blanchette CA (1997) Size and survival of intertidal plants in response to wave action: a case study with Fucus gardneri. Ecology 78: 1563-1578.

97

98

BIBLIOGRAPHY

Bold HC, Wynne MJ (1978) Introduction to the Algae. Structure and reproduction. Prentice-Hall, Inc., Englewood Cliffs.

Boudouresque CF and Verlaque M (2002) Biological pollution in the Mediterranean Sea: invasive versus introduced macrophytes. Mar. Pollut. Bull. 44: 32-38.

Boudouresque CF, Bernard G, Bonhomme P, Charbonnel E, Diviacco G, Meinesz A, Pergent G, Pergent-Martini C, Ruitton S, Tunesi L (2006) Pr´eservation et conservation des herbiers ` a Posidonia oceanica. Lâ&#x20AC;&#x2122; Accord RAMOGE, 202 pp.

Boudouresque CF, Bernard G, Pergent G, Shili A, Verlaque M (2009) Regression of Mediterranean seagrasses caused by natural processes and anthropogenic disturbances and stress: A critical review. Bot. Mar. 52: 395-418.

Bouma TJ, de Vries MB, Low E, Peralta G, Tanczos C, van de Koppel J, Herman PMJ (2005) Trade-offs related to ecosystem engineering: A case study on stiffness of emerging macrophytes. Ecology 86: 2187-2199.

Bouma TJ, de Vries MB, Herman PMJ. (2010) Comparing ecosystem engineering efficiency of two plant species with contrasting growth strategies. Ecology 91(9): 2696-2704 Bradley K and Houser C. (2009) Relative velocity of seagrass blades: Implications for wave attenuation in low-energy environments. J. Geophys. Res. 114: F01004 Buia MC and Mazzella L (1991) Reproductive phenology of the Mediterranean seagrasses Posidonia oceanica (L.) Delile, Cymodocea nodosa (Ucria) Aschers, and Zostera noltii Hornem. Aquat. Bot. 40: 343-362.

Bulleri F, Benedetti-Cecchi L (2008) Facilitation of the introduced green alga Caulerpa racemosa by resident algal turfs: experimental evaluation of underlying mechanisms. Mar. Ecol. Prog. Ser. 364: 77-86.

Cabaco S, Santos R, Duarte CM (2008) The impact of sediment burial and erosion on seagrasses: A review. Estuar. Coast. and Shelf Sci. 79: 354-366.

Cabaco S, Ferreira O, Santos R (2010) Population dynamics of the seagrass Cymodocea nodosa in Ria Formosa lagoon following inlet artificial relocation.

Estuar. Coast. and Shelf Sci. 87: 510-516.

BIBLIOGRAPHY

99

Carlton JT, Geller JB (1993) Ecological roulette - The global transport of nonindigenous marine organisms. Science 261: 78-82.

Ca˜ nellas B, Orfila A, M´endez FJ, Men´endez M, Tintor´e J (2007) Application of POT model to estimate the extreme significant wave height levels around the Balearic Sea (Western Mediterranean). J. Coast. Res. 50: 329-333.

Castillo MC, Baquerizo A, Losada MA (2005) Temporal and spatial statistical variability of the wave height in the surf zone. 29th International Conference on Coastal Engineering, 2004. Lisbon, Portugal. Vols 1-4: 997-1008.

Caye G and Meinesz A (1986) Experimental study of seed germination in the seagrass Cymodocea nodosa. Aquat. Bot. 26: 79-87.

Cebrian E and Ballesteros E (2009) Temporal and spatial variability in shallowand deep-water populations of the invasive Caulerpa racemosa var. cylindracea in the Western Mediterranean. Estuar. Coast. and Shelf Sci. 83: 469474.

Ceccherelli G and Cinelli F (1999) Effects of Posidonia oceanica canopy on Caulerpa taxifolia size in a north-western Mediterranean bay. J. Exp. Mar. Biol. Ecol. 240: 19-36.

Ceccherelli G, Piazzi L, Balata D (2002) Spread of introduced Caulerpa species in macroalgal habitats. J. Exp. Mar. Biol. Ecol. 280: 1-11.

Chambers PA (1987) Nearshore occurrence of submersed aquatic macrophytes in relation to wave action. Can. J. Fish. Aquat. Sci. 44: 1666-1669.

Chen SN, Sanford LP, Koch EW, Shi F, North EW (2007) A nearshore model to investigate the effects of seagrass bed geometry on wave attenuation and suspended sediment transport. Estuar. and Coasts 30: 296-310 Chisholm JRM and Moulin P (2003) Stimulation of nitrogen fixation in refractory organic sediments by Caulerpa taxifolia (Chlorophyta). Limnol. Oceanogr.

48: 787-794.

Chisholm JRM, Dauga C, Ageron E, Grimont PAD, Jaubert JM (1996) Roots in mixotrophic algae. Nature 381: 382.

Chivers R, Emerson N, Burns D (1990) New acoustic processing for underway surveying. The Hydrographic Journal 56: 8-17.

100

BIBLIOGRAPHY

Clarke SM and Kirkman H (1989) Seagrass dynamics. In: Larkum AWD, McComb AJ, Shepherd SA (Eds.) Biology of Seagrasses: a treatise on the biology of seagrasses with special reference to the Australasian region. Elsevier, North Holland.

Costanza R, d’Arge R, de Groot R, Farber S, Grasso M, Hannon B, Limburg K, Naeem S, O’neill RV, Paruelo J, Raskin RG, Sutton P, van den Belt M (1997) The value of the world’s ecosystem services and natural capital. Nature 387: 253-260.

Cornelisen CD and Thomas FIM (2004) Ammonium and nitrate uptake by leaves of the seagrass Thalassia testudinum: impact of hydrodynamic regime and epiphyte cover on uptake rates. J. Mar. Systems 29: 177-194.

Cornelisen CD and Thomas FIM (2006) Water flow enhances ammonium and nitrate uptake in a seagrass community. Mar. Ecol. Prog. Ser. 312: 1-13.

Dalrymple RA, Kirby JT, Hwang PA. (1984) Wave diffraction due to areas of energy-dissipation. J. Waterw. Port C-ASCE 110: 67-79.

D’Amours O and Scheibling RE (2007) Effect of wave exposure on morphology, attachment strength and survival of the invasive green alga Codium fragile ssp. tomentosoides. J. Exp. Mar. Biol. Ecol. 351: 129-142.

Davis AR, Ferguson AM, Wright JT (2009) Structural complexity facilitates accumulation and retention of fragments of the invasive alga, Caulerpa taxifolia.

J. Exp. Mar. Biol. Ecol. 371: 163-169.

Dean RG. and Dalrymple RA (1991) Water wave mechanics for engineers and scientist. Advanced series on ocean engineering, volume 2. World Scientific.

Dean RG. and Dalrymple RA (2002) Coastal processes with engineering applications. Cambridge University Press, Cambridge. pp. 475.

De Falco G, Baroli M, Cucco A, Simeone S (2008) Intrabasinal conditions promoting the development of a biogenic carbonate sedimentary facies associated with the seagrass Posidonia oceanica. Cont. Shelf Res. 28(6): 797-812.

Den Hartog C (1970) The seagrasses of the world. North Holland Publishibg Company, Amsterdam.

Den Hartog C (1971) The dynamic aspect of the ecology of sea-grass communities. Conference paper. Thalassia Jugoslavica 7(1): 101-112.

BIBLIOGRAPHY

101

Dennison WC and RS Alberte RS (1985) Role of daily light period in the depth distribution of Zostera marina (Eelgrass). Mar. Ecol. Prog. Ser. 25(1): 51-61.

Dennison WC (1987) Effects of light on seagrass photosynthesis, growth and depth distribution. Aquat. Bot. 27(1): 15-26.

Denny MW, Daniel TL, Koehl MAR (1985) Mechanical limits to size in waveswept organisms. Ecol. Monogr. 55: 69-102.

Dom´ınguez M, Infantes E, Terrados J (2010) Seed maturity of the Mediterranean seagrass Cymodocea nodosa. Vie et Milieu 60: 1-6.

Dom´ınguez M, Celdr´ an-Sabater D, Mu˜ noz-Vera A, Infantes E, Martinez-Ba˜ nos P, Mar´ın A, Terrados J (2011) Experimental evaluation of the restoration capacity of a fish-farm impacted area with Posidonia oceanica (L.) Delile seedlings. Rest. Ecol. 10.1111/j.1526-100X.2010.00762.x Duarte CM (1991) Seagrass depth limits. Aquat. Bot. 40(4): 363-377.

Duarte CM, Terrados J, Agawin NSR, Fortes MD, Bach SS, Kenworthy WJ (1997) Response of a mixed Philippine seagrass meadow to experimental burial. Mar. Ecol. Prog. Ser. 147: 285-294.

Folkard AM (2005) Hydrodynamics of model Posidonia oceanica patches in shallow water. Limnol. Oceanogr. 50(5): 1592-1600.

Fonseca MS, Fisher JS, Zieman JC, Thayer GW (1982) Influence of the seagrass, Zostera marina L, on current flow. Estuar. Coast. Shelf Sci. 15(4): 351-364.

Fonseca MS, Zieman JC, Thayer GW, Fisher JS (1983) The role of current velocity in structuring eelgrass (Zostera marina L.) meadows. Estuar. Coast.

Shelf Sci. 17: 367-380.

Fonseca MS and Fisher JS (1986) A comparison of canopy friction and sediment movement between 4 species of seagrass with reference to their ecology and restoration. Mar. Ecol. Prog. Ser. 29(1): 15-22.

Fonseca MS and Kenworthy WJ (1987) Effects of current on photosynthesis and distribution of seagrasses. Aquat. Bot. 27: 59-78.

Fonseca MS and Cahalan JA (1992) A preliminary evaluation of wave attenuation by 4 species of seagrass. Estuar. Coast. and Shelf Sci. 35: 565-576.

102

BIBLIOGRAPHY

Fonseca MS (1996) The role of seagrasses in nearshore sedimentary processes: a review. In: Nordstrom KF, Roman CT (eds) Estuarine Shores: Evolution, Environments and Human Alterations. John Wiley & Sons LTD, p 261-281 Fonseca MS and Bell SS (1998) Influence of physical settings on seagrass landscapes near Beauford, North Carolina, USA. Mar. Ecol. Prog. Ser. 171: 109121.

Fonseca MS, Whitfield PE, Kelly NM, Bell SS (2002) Modeling seagrass landscape pattern and associated ecological attributes. Ecol. Appl. 12(1): 218237. ´ Fornes A, Basterretxea G, Orfila A, Jordi A, Alvarez A, Tintor´e J (2006) Mapping Posidonia oceanica from IKONOS. Journal of Photogrammetry and Remote Sensing 60(5): 315-322.

Frederiksen M, Krause-Jensen D, Holmer M, Laursen JS (2004) Spatial and temporal variation in eelgrass (Zostera marina) landscapes: influence of physical setting. Aquat. Bot. 78(2): 147-165.

Gacia E and CM Duarte (2001) Sediment retention by a Mediterranean Posidonia oceanica meadow: the balance between deposition and resuspension. Estuar.

Coast. Shelf Sci. 52: 505-514.

Gambi MC, Nowell ARM, Jumars PA (1990) Flume observations on flow dynamics in Zostera marina (Eelgrass) beds. Mar. Ecol. Prog. Ser. 61(1-2): 159-169.

Ghisalberti M and Nepf HM (2002) Mixing layers and coherent structures in vegetated aquatic flows. J. Geophys. Res. 107(C2), doi:10.1029/2001JC000871.

GIOC (2003) Spectral wave propagation model (Oluca-SP). State Coastal Office - Spanish Environmental Ministry and University of Cantabria. pp. 170. (in Spanish).

Glasby TM, Gibson PT, Kay S (2005) Tolerance of the invasive marine alga Caulerpa taxifolia to burial by sediment. Aquat. Bot. 82: 71-81.

Gobert S, Kyramarios M, Lepoint G, Pergent-Martini C, Bouquegneau JM (2003) Variations at different spatial scales of Posidonia oceanica (L.) Delile beds; effects on the physico-chemical parameters of the sediment. Oceanol.

Acta 26: 199-207.

BIBLIOGRAPHY

103

´ G´ omez-Pujol L, Orfila A, Ca˜ nellas B, Alvarez-Ellacuria A, M´endez FJ, Medina R, Tintor´e J (2007) Morphodynamic classification of sandy beaches in low energetic marine environment. Mar. Geol. 242(4): 235-246. ´ G´ omez-Pujol L, Orfila A, Alvarez-Ellacuria A, Tintor´e J (2011) Controls on sediment dynamics and medium-term morphological change in a barred microtidal beach (Cala Millor, Mallorca, Western Mediterranean). Geomorphology. doi: 10.1016/j.geomorph.2011.04.026 Granata TC, Serra T, Colomer J, Casamitjana X, Duarte CM, Gacia E, Petersen JK (2001) Flow and particle distributions in a nearshore seagrass meadow before and after a storm. Mar. Ecol. Prog. Ser. 218: 95-106 Grant WD and Madsen OS (1979) Combined wave and current interaction with a rough bottom. J. Geo. Res. 84:C4 Green EP and Short FT (2003) World atlas of seagrasses. University of California Press.

Greve T and Krause-Jensen D (2005) Predictive modelling of eelgrass (Zostera marina) depth limits. Mar. Biol. 146: 849-858.

Guidetti P, Lorenti M, Buia MC, Mazzella L (2002) Temporal dynamics and biomass partitioning in three Adriatic seagrass species: Posidonia oceanica, Cymodocea nodosa, Zostera marina. Mar. Ecol-P. S. N. I. 23: 51-67.

Hemminga MA and Nieuwenhuize J (1990) Seagrass wrack-induced dune formation on a tropical coast (Banc-Darguin, Mauritania). Estuar. Coast. and Shelf Sci. 31: 499-502.

Hendriks IE, van Duren LA, Herman PMJ (2006) Turbulence levels in a flume compared to the field: Implications for larval settlement studies. J. Sea Res. 55: 15-29.

Hill D, Coquillard P, Vaugelas J, Meinesz A (1998) An algorithmic model for invasive species: Application to Caulerpa taxifolia (Vahl) C. Agardh development in the North-Western Mediterranean Sea. Ecol. Model. 109: 251-265.

Hine AC, Evans MW, Davis RA, Belknap DF (1987) Depositional response to seagrass mortality along a low-energy, barrier-island coast: west-central Florida. J. Sedim. Petrol. 57: 431-439.

104

BIBLIOGRAPHY

Hovel KA, Fonseca CR, Myer DL, Kenworthy WJ, Whitfield PE (2002) Effects of seagrass landscape structure, structural complexity and hydrodynamic regime on macrofaunal densities in North Carolina seagrass beds. Mar. Ecol. Prog.

Ser. 243: 11-24.

Idestam-Almquist J and Kautsky L (1995) Plastic responses morphology of Potamogeton pectinatus L. to sediment and above-sediment conditions at two sites in the northern Baltic proper. Aquat. Bot. 52: 205-216. ˜ IMEDEA (2005) Variabilidad y DinA¡mica Sedimentaria De Las Playas de Cala Millor y Cala Sant Vicent. Volumen 1. Cala Millor. Govern de les Illes Balears. (in Spanish) ´ Infantes E, Terrados J, Orfila A, Ca˜ellas A, Alvarez-Ellacuria (2009) Wave energy and the upper depth limit distribution of Posidonia oceanica. Bot. Mar. 52: 419-427.

Infantes E, Terrados J, Orfila A (2011) Assessment of substratum effect on the distribution of two invasive Caulerpa (Chlorophyta) species. Estuar. Coast. and Shelf Sci. 91: 434-441 Jaubert JM, Chisholm JRM, Ducrot D, Ripley HT, Roy L, Passeron-Seitre G (1999) No deleterious alterations in Posidonia beds in the bay of menton (France) eight years after Caulerpa taxifolia colonization. J. Phycol. 35: 11131119.

Jonsson IG (1967) Wave boundary layers and friction factors. Proc. 10th Int. Conf. Coastal Engineering, ASCE, New York, 127.

Keddy PA (1984) Quantifying a within-lake gradient of wave energy in Gillfillan Lake, Nova-Scotia. Can. J. Bot. 62(2): 301-309.

Kendrick GA and Burt JS (1997) Seasonal changes in epiphyte macro-algae assemblages between offshore exposed and inshore protected Posidonia sinuosa seagrass meadows, Western Australia. Bot. Mar. 40: 77-85.

Kenworthy WJ and Fonseca MS (1996) Light requirements of seagrasses Halodule wrightii and Syringodium filiforme derived from the relationship between diffuse light attenuation and maximum depth distribution. Estuaries 19(3): 740-750.

BIBLIOGRAPHY

105

Kirby JT and Dalrymple RA (1983) A parabolic equation for the combined refraction-diffraction of Stokes waves by mildly-varying topography. J. Fluid Mech. 136: 453-466.

Kirkman H and Kuo J (1990) Pattern and process in southern Western Australian seagrasses. Aquat. Bot. 37: 367-382.

Kirkman H (1998) Pilot experiments on planting seedlings and small seagrass propagules in western Australia. Mar. Pollut. Bull. 37: 460-467 Knutson PL, Bronchu RA, Seelig WN (1982) Wave damping in Spartina alterniflora marshes. Wetlands 2: 87-104.

Kobayashi N, Raichle AW, Asano T (1993) Wave Attenuation by Vegetation. J. Waterw. Port C-ASCE. 119: 30-48 Koch EW and Beer S (1996) Tides, light and the distribution of Zostera marina in Long Island Sound, USA. Aquat. Bot. 53(1-2): 97-107.

Koch EW and Gust G (1999) Water flow in tide- and wave-dominated beds of the seagrass Thalassia testudinum. Mar. Ecol. Prog. Ser. 184: 63-72.

Koch EW (2001) Beyond light: Physical, geological, and geochemical parameters as possible submersed aquatic vegetation habitat requirements. Estuaries 24: 1-17.

Koch EW, Ackerman JD, Verduin J, van Keulen M (2006) Fluid Dynamics in Seagrass Ecology - from Molecules to Ecosystems. In: (Larkum, A.W.D., R.J. Orth, C.M. Duarte, eds). Seagrasses: Biology, Ecology and Conservation.

Springer, Netherlands. pp. 193-224.

Koch EW, Barbier EB, Silliman BR, Reed DJ, Perillo GME, Hacker SD, Granek EF, Primavera JH, Muthiga N, Polasky S, Halpern BS, Kennedy CJ, Kappel CV, Wolanski E (2009) Non-linearity in ecosystem services: temporal and spatial variability in coastal protection. Front. Ecol. Environ. 7: 29-37.

Krause-Jensen D, Pedersen MF, Jensen C (2003) Regulation of eelgrass (Zostera marina) cover along depth gradients in Danish Coastal Waters. Estuaries 26: 866-877.

Lacy JR, Sherwood CR, Wilson DJ, Chisholm TA, Gelfenbaum GR (2005) Estimating hydrodynamic roughness in a wave-dominated environment with a high-resolution acoustic Doppler profiler. J. Geophys. Res. 110, C06014, doi:10.1029/2003JC001814

106

BIBLIOGRAPHY

Lane SN, Biron PM, Bradbrook KF, Butler JB, Chandler JH, Crowell MD, McLelland SJ, Richards KS, Roy AG (1998) Three-dimensional measurement of river channel flow processes using acoustic Doppler velocimetry. Earth Surf.

Processes 23: 1247-1267.

Laugier T, Rigollet V, de Casabianca ML (1999) Seasonal dynamics in mixed eelgrass beds, Zostera marina L. and Z. noltii Hornem., in a Mediterranean coastal lagoon (Thau lagoon, France). Aquat. Bot. 63: 51-69.

Leriche A, Pasqualini V, Boudouresque CF, Bernard G, Bonhomme P, Clabaut P, Denis J (2006) Spatial, temporal and structural variations of a Posidonia oceanica seagrass meadow facing human activities. Aquat. Bot. 84(4): 287293.

Lowe RJ, Koseff JR, Monismith SG, Falter JL (2005) Oscillatory flow through submerged canopies: 1. Velocity structure. J. Geophys. Res. 110, C10016, doi:10.1029/2004JC002788.

Luque AA, Templado J (Eds.) (2004) Praderas y bosques marinos de Andaluc´ıa.

Consejer´ıa de Medio Ambiente, Junta de Andaluc´ıa, Sevilla, 336 pp.

Luhar M, Rominger J, Nepf H (2008) Interaction between flow, transport and vegetation spatial structure. Env. Fluid Mech. 8: 423-439.

Luhar M, Coutu S, Infantes E, Fox S, Nepf H (2010) Wave-induced velocities inside a model seagrass bed. J. Geophys. Res. 115, C12005, doi:10.1029/2010JC006345.

Madsen JD, Chambers PA, James WF, Koch EW, Westlake DF (2001) The interaction between water movement, sediment dynamics and submersed macrophytes. Hydrobiologia 444: 71-84.

Marb` a N and Duarte CM (1994) Growth response of the seagrass Cymodocea nodosa to experimental burial and erosion. Mar. Ecol. Prog. Ser. 107: 307311.

Marb` a N, Duarte CM, Cebrian J, Gallegos ME, Olesen B, Sand-Jensen K (1996) Growth and population dynamics of Posidonia oceanica on the Spanish Mediterranean coast: Elucidating seagrass decline. Mar. Ecol. Prog. Ser. 137: 203-213 Marb` a N and Duarte CM (1998) Rhizome elongation and seagrass clonal growth.

Mar. Ecol. Prog. Ser. 174: 269-280.

BIBLIOGRAPHY

107

Marb` a N, Duarte CM, Holmer M, Martinez R, Basterretxea G, Orfila A, Jordi A, Tintor´e J (2002) Effectiveness of protection of seagrass (Posidonia oceanica) populations in Cabrera National Park (Spain). Environment. Conser. 29: 509518.

Martone PT and Denny MW (2008) To break a coralline: mechanical constraints on the size and survival of a wave-swept seaweed. J. Exp. Biol. 211: 34333441.

Mateo MA, Romero J, Perez M, Littler MM, Littler DS (1997) Dynamics of millenary organic deposits resulting from the growth of the Mediterranean seagrass Posidonia oceanica. Estuar. Coast. Shelf Sci. 44: 103-110.

McKenzie LJ, Finkbeine MA, Kirkman H (2001) Methods for mapping seagrass distribution. In: (Short, F.T. and Coles, R.G. eds). Global Seagrass Research Methods. Elsevier, Amsterdam. pp. 101-121.

Meinesz A and Hesse B (1991) Introduction of the tropical alga Caulerpa taxifolia and its invasion of the Northwestern Mediterranean. Oceanol. Acta 14: 415426.

M´endez FJ, Losada IJ, Losada MA (1999) Hydrodynamics induced by wind waves in a vegetation field. J. Geophys. Res. 104: 18383-18396 M´endez FJ, Losada IJ (2004) An empirical model to estimate the propagation of random breaking and nonbreaking waves over vegetation fields. Coast. Eng. 51: 103-118 Molenaar H and Meinesz A (1995) Vegetative reproduction in Posidonia oceanica: survival and development of transplanted cuttings according to different spacings, arrangements and substrates. Bot. Mar. 38: 313-322 Nepf HM and Vivoni ER (2000) Flow structure in depth-limited, vegetated flow.

J. Geophys. Res. 105(C12) 28: 547-557, doi:10.1029/2000JC900145.

Nielsen P (1992) Coastal bottom boundary layer and sediment transport. World Scientific, Singapore Nielsen P and Callaghan DP (2003) Shear stress and sediment transport calculations for sheet flow under waves. Coast. Eng. 47(3): 347-354 Nortek AS (2002) Wave measurements using the PUV method, TN-19. Doc. No. N4000-720

108

BIBLIOGRAPHY

Orfila A, Jordi A, Basterretxea G, Vizoso G, Marb`a N, Duarte CM, Werner FE, Tintor´e J (2005) Residence time and Posidonia oceanica in Cabrera Archipielago National Park, Spain. Cont. Shelf. Res. 25: 1339-1352.

Orth RJ, Carruthers TJB, Dennison WC, Duarte CM, Fourqurean JW, Heck KL, Hughes AR, Kendrick GA, Kenworthy WJ, Olyarnik S, Short FT, Waycott M, Williams SL (2006) A global crisis for seagrass ecosystems. Bioscience 56: 987-996.

Orlowski A (1984) Application of multiple echo energy measurements for evaluation of sea bottom type. Oceanologia 19: 61-78.

Paling EI, van Keulen M, Wheeler KD, Phillips J, Dyhrberg R (2003) Influence of spacing on mechanically transplanted seagrass survival in a high wave energy regime. Rest. Ecol. 11: 56-61.

Pasqualini V, Pergent-Martini C, Clabaut P, Pergent G (1998) Mapping of Posidonia oceanica using aerial photographs and side scan sonar: Application off the Island of Corsica (France). Estuar. Coast. Shelf Sci. 47(3): 359-367.

Pergent-Martini C, Rico-Raimondino V, Pergent G (1994) Primary production of Posidonia oceanica in the Mediterranean Basin. Mar Biol 120: 9-15 Piazzi L, Cinelli F (1999) Developpement et dynamique saisonniere d’un peuplement mediterraneen de l’algue tropicale Caulerpa racemosa (Forsskal) J. Agardh. Crytogamie Algol. 20: 295-300. (in French, with English Abstract) Piazzi L, Acunto S, Cinelli F (1999) In situ survival and development of Posidonia oceanica (L.) Delile seedlings. Aquat. Bot. 63: 103-112.

Piazzi L, Ceccherelli G, Cinelli F (2001) Threat to macroalgal diversity: effects of the introduced green alga Caulerpa racemosa in the Mediterranean. Mar. Ecol. Progr. Ser. 210: 149-159.

Piazzi L, Ceccherelli G, Balata D, Cinelli F (2003) Early Patterns of Caulerpa racemosa recovery in the Mediterranean Sea: the influence of algal turfs. J. Mar. Biol. Assoc. UK. 83: 27-29.

Piazzi L, Balata D, Ceccherelli G, Cinelli F (2005) Interactive effect of sedimentation and Caulerpa racemosa var. cylindracea invasion on macroalgal assemblages in the Mediterranean Sea. Estuar. Coast. Shelf Sci. 64: 467-474.

BIBLIOGRAPHY

109

Preen AR, Lee Long WJ, Coles RG (1995) Flood and cyclone related loss, and partial recovery, of more than 1000 km2 of seagrass in Hervey Bay, Queensland, Australia. Aquat. Bot. 52: 3-17.

Procaccini G, Buia MC, Gambi MC, Perez M, Pergent G, Pergent-Martini C, Romero J (2003) The seagrasses of the Western Mediterranean. In: (E.P. Green, F.T. Short, eds), World Atlas of Seagrasses, UNEP World Conservation Monitoring Centre. University of California Press, Berkeley, pp. 48-58.

Quinn GP and Keough MJ (2002) Experimental design and data analysis for biologists. Cambridge University Press, pp 357 ´ Ratsimandresy AW, Sotillo MG, Carretero Albiach JC, Alvarez Fanjul E, Hajji H (2008) A 44-year high-resolution ocean and atmospheric hindcast for the Mediterranean Basin Developer within the HIPOCAS Project. Coast. Eng. 55: 827-842.

Riedel HP (1972) Direct measurement of bed shear stress under waves. Ph.D. thesis, Queen’s University, Ont., Canada Ribera G, Coloreu M, Rodr´ıguez-Prieto C, Ballesteros E (1997) Phytobenthic assemblages of Addaia Bay (Menorca, Western Mediterranean): Composition and distribution. Bot. Mar. 40: 523-532.

Riis T and Hawes I (2003) Effect of wave exposure on vegetation abundance, richness and depth distribution of shallow water plants in a New Zealand lake. Freshwater Biol. 48: 75-87.

Rivers DO, Kendrick GA, Walker DI (2011) Microsites play an important role for seedling survival in the seagrass Amphibolis antarctica. J. Exp. Mar. Biol.

Ecol. 40: 29-35 Robertson AI and Mann KH (1984) Disturbance by ice and life-history adaptations of the seagrass Zostera marina. Mar. Biol. 80(2): 131-141.

Ruitton S, Javel F, Culioli JM, Meinesz A, Pergent G, Verlaque M (2005) First assessment of the Caulerpa racemosa (Caulerpales, Chlorophyta) invasion along the French Mediterranean coast. Marine Pollution Bulletin 50: 10611068.

Ru´ız GM, Fofonoff PW, Carlton JT, Wonham MJ, Hines AH (2000) Invasion of coastal marine communities in North America: Apparent patterns, processes, and biases. Annu. Rev. Ecol. Syst. 31: 481-531.

110

BIBLIOGRAPHY

S´ anchez-G´ onzalez JF, S´ anchez-Rojas V, Memos CD (2011) Wave attenuation due to Posidonia oceanica meadows. J. Hydr. Res. In Press. doi:10.1080/00221686.2011.552464 Scheibling RE and Melady RA (2008) Effect of water movement and substratum type on vegetative recruitment of the invasive green alga Codium fragile ssp. tomentosoides. Bot. Mar. 51: 341-349.

Shepherd SA and Womersley HBS (1981) The algal and seagrass ecology of Waterloo Bay, South-Australia. Aquat. Bot. 11(4): 305-371.

Short FT and Short CA (1984) The seagrass filter: purification of estuarine and coastal waters. In: VC Kennedy (ed.), The Estuary as a Filter, Academic Press.

Short FT and Wyllie-Echeverria S (1996) Natural and human-induced disturbance of seagrasses. Environ. Conserv. 23: 17-27.

Short FT, Davis RC, Kopp BS, Short CA, Burdick DM (2002) Site-selection model for optimal transplantation of eelgrass Zostera marina in the northeastern US. Mar. Ecol. Prog. Ser. 227: 253-267.

Simarro G, Orfila A, Liu PLF (2008) Bed-shear stress in turbulent wave-current boundary layers. J. Hydr. Eng. Asce 134: 225-230. ´ Soares CG, Carretero JC, Weisse R, Alvarez E (2002) A 40 years hindcast of wind, sea level and waves in European waters. Proceedings of OMAE 2002: 21st International Conference on Offshore Mechanics and Artic Engineering.

Oslo, Norway.

Stevens CL and Hurd CL (1997) Boundary-layers around aquatic macrophytes.

Hydrobiologia 346: 199-128.

Stewart HL (2004) Hydrodynamic consequences of maintaining an upright posture by different magnitudes of stiffness and buoyancy in the tropical alga Turbinaria ornata. J. Marine Syst. 49: 157-167.

Swart DH (1974) Offshore sediment transport and equilibrium beach profiles.

Delf Hydr. Lab. Publ. No 131 Taylor WR (1960) Marine algae of the Eastern tropical and subtropical coasts of the Americas. University of Michigan Press, Ann Arbor.

BIBLIOGRAPHY Teeter AM, Parchure emphasis lagoons).

111

Johnson BH, Berger C, Stelling G, Scheffner NW, Garcia MH, TM (2001) Hydrodynamic and sediment transport modeling with on shallow-water, vegetated areas (lakes, reservoirs, estuaries and Hydrobiologia 444: 1-24.

Terrados J (1997) Is light involved in the vertical growth response of seagrasses when buried by sand?. Mar. Ecol. Prog. Ser. 152: 295-299.

Terrados J and Duarte CM (2000) Experimental evidence of reduced particle resuspension within a seagrass (Posidonia oceanica) meadow. J. Exp. Mar. Biol. Ecol. 243(1): 45-53.

Terrados J and Medina-Pons FJM (2008) Epiphyte load on the seagrass Posidonia oceanica (L.) Delile does not indicate anthropogenic nutrient loading in Cabrera Archipelago National Park (Balearic Islands, western Mediterranean). Scientia Marina 72: 503-510.

Thibaut T, Meinesz A, Coquillard P (2004) Biomass seasonality of Caulerpa taxifolia in the Mediterranean Sea. Aquat. Bot. 80: 291-297.

Thomas FIM, Cornelisen CD, Zande JM (2000) Effects of water velocity and canopy morphology on ammonium uptake by seagrass communities. Ecology 81: 2704-2713 Thomas FIM and Cornelisen CD (2003) Ammonium uptake by seagrass communities: effects of oscillatory versus unidirectional flow. Mar. Ecol. Prog. Ser. 247: 51-57 Vacchi M, Montefalcone M, Bianchi CN, Morri C, Ferrari M (2010) The influence of coastal dynamics on the upper limit of the Posidonia oceanica meadow.

Mar. Ecol. 31: 546-554. van Katwijk MM and Hermus DCR (2000a) Effects of water dynamics on Zostera marina: Transplantation experiments in the intertidal Dutch Wadden Sea. Mar. Ecol. Prog. Ser. 208: 107-118. van Katwijk MM, Hermus DCR, de Jong DJ, Asmus RM, de Jonge VN (2000b) Habitat suitability of the Wadden Sea for restoration of Zostera marina beds.

Helgoland Mar. Res. 54(2-3): 117-128. van Katwijk MM, Bos AR, de Jonge VN, Hanssen LSAM, Hermus DCR, de Jong DJ (2009) Guidelines for seagrass restoration: Importance of habitat selection

112

BIBLIOGRAPHY

and donor population, spreading of risks, and ecosystem engineering effects.

Mar. Pollut. Bull. 58: 179-188.

Verduin JJ and Backhaus JO (2000) Dynamics of plant-flow interactions for the seagrass Amphibolis antarctica: Field observations and model simulations.

Estuar. Coast. and Shelf Sci. 50: 185-204 Ward LG, Kemp WM, Boynton WR (1984) The influence of waves and seagrass communities on suspended particulates in an estuarine embayment.

Mar. Geol. 59: 85-103 Wicks EC, Koch EW, Oâ&#x20AC;&#x2122;Neil JM, Elliston K (2009) Effects of sediment organic content and hydrodynamic conditions on the growth and distribution of Zostera marina. Mar. Ecol. Prog. Ser. 378: 71-80.

Williams SL (1984) Uptake of sediment ammonium and translocation in a marine green macroalga Caulerpa cupressoides. Limnol. Oceanogr. 29: 374-379.

Williams SL (1988) Disturbance and recovery of a deep-water Caribbean seagrass bed. Mar. Ecol. Prog. Ser. 42: 63-71.

Worcester SE (1995) Effects of eelgrass beds on advection and turbulent mixing in low current and low shoot density environments. Mar. Ecol. Prog. Ser. 126(1-3): 223-232.

Wright LD and AD Short AD (1983) Morphodynamics of beaches and surf zones in Australia. In: (P.D. Koma, eds). Handbook of Coastal Processes and Erosion. CRC Press, Boca Raton, FL. pp. 35-64.

Other publications ´ • E. Infantes, J. Terrados, A. Orfila, B. Ca˜ nellas, A. Alvarez-Ellacuaria (2009). Wave energy and the upper depth limit distribution of Posidonia oceanica. Botanica Marina 52: 419-427 • E. Infantes, J. Terrados, A. Orfila (2011). Assessment of substratum effect on the distribution of two invasive Caulerpa (Chlorophyta) species.

Estuarine, Coastal and Shelf Science 91: 434-441 • Luhar M, Coutu S, Infantes E, Fox S, Nepf H (2010) Wave-induced velocities inside a model seagrass bed. J. Geophys. Res. 115, C12005.

• Dom´ınguez M, Infantes E, Terrados J (2010) Seed maturity of the Mediterranean seagrass Cymodocea nodosa. Vie et Milieu 60: 1-6.

• Dom´ınguez M, Celdr´ an-Sabater D, Mu˜ noz-Vera A, Infantes E, MartinezBa˜ nos P, Mar´ın A, Terrados J (2010) Experimental evaluation of the restoration capacity of a fish-farm impacted area with Posidonia oceanica (L.) Delile seedlings. (In Press: Restoration Ecology, doi:10.1111/j.1526100X.2010.00762.x).

• Luhar M, Infantes E, Nepf H. Seagrass blade motion under waves and its impact on wave decay. (In preparation).

• March D, Al´ os J, Cabanellas M, Infantes E, Palmer M. Probabilistic mapping of Posidonia oceanica cover: a Bayesian geostatistical model from drop camera system data. (In preparation).

• Klaassen PC, Bouma TJ, Balke T, Infantes E, Van Dalen J, Temmerman S, Herman PMJ. Quantifying the critical sediment disturbance depth for seedling survival of coastal vegetations: a method. (In preparation).

113

114

Other publications

• Terrados J, Dom´ınguez M, Mar´ın A, Celdr´an D, Mu˜ noz-Vera A, Infantes E, Mart´ınez-Ba˜ nos P. Survivorship and leaf development of Posidonia oceanica seedlings transplanted in a fish-farm impacted bay. (In preparation).

Article in press - uncorrected proof Botanica Marina 52 (2009): 419–427 2009 by Walter de Gruyter • Berlin • New York. DOI 10.1515/BOT.2009.050

Wave energy and the upper depth limit distribution of Posidonia oceanica

Eduardo Infantes*, Jorge Terrados, Alejandro Orfila, Bartomeu Can˜ellas and Amaya A´lvarez-Ellacuria Instituto Mediterra´neo de Estudios Avanzados, IMEDEA (CSIC-UIB), Miquel Marque´s 21, 07190, Esporles, Spain, e-mail: eduardo.infantes@uib.es * Corresponding author

Abstract It is widely accepted that light availability sets the lower limit of seagrass bathymetric distribution, while the upper limit depends on the level of disturbance by currents and waves. The establishment of light requirements for seagrass growth has been a major focus of research in marine ecology, and different quantitative models provide predictions for seagrass lower depth limits. In contrast, the influence of energy levels on the establishment, growth, and maintenance of seagrasses has received less attention, and to date there are no quantitative models predicting the evolution of seagrasses as a function of hydrodynamics at a large scale level. Hence, it is not possible to predict either the upper depth limit of the distribution of seagrasses or the effects that different energy regimes will have on these limits. The aim of this work is to provide a comprehensible methodology for obtaining quantitative knowledge and predictive capacity for estimating the upper depth limit of seagrasses as a response to wave energy dissipated on the seafloor. The methodology has been applied using wave data from 1958 to 2001 in order to obtain the mean wave climate in deep water seaward from an open sandy beach in the Balearic Islands, western Mediterranean Sea where the seagrass Posidonia oceanica forms an extensive meadow. Mean wave conditions were propagated to the shore using a two-dimensional parabolic model over the detailed bathymetry. The resulting hydrodynamics were correlated with bottom type and the distribution of P. oceanica.

Results showed a predicted near-bottom orbital velocity of between 38 and 42 cm s-1 as a determinant of the upper depth limit of P. oceanica. This work shows the importance of interdisciplinary effort in ecological modeling and, in particular, the need for hydrodynamical studies to elucidate the distribution of seagrasses in shallow depths. Moreover, the use of predictive models would permit evaluation of the effects of coastal activities (construction of ports, artificial reefs, beach nutrientinput, dredging) on benthic ecosystems.

Keywords: near-bottom orbital velocity; parabolic model; Posidonia oceanica upper depth limit; seagrass distribution; wave energy.

Introduction Seagrasses are marine flowering plants that cover large areas in shallow coastal waters; the meadows they form are among the most biologically diverse and productive components of coastal systems. Posidonia oceanica (L.) Delile, an endemic seagrass species of the Mediterranean Sea, grows between the depths of 0.5 and 45 m (Procaccini et al. 2003) and covers an estimated surface area between 2.5 and 5 million ha, approximately 1–2% of the 0–50 m depth zone (Pasqualini et al. 1998).

The study of light requirements for seagrass growth has been a major focus of research in marine ecology, and different quantitative models provide predictions of the seagrass lower depth limits (Dennison and Alberte 1985, Dennison 1987, Duarte 1991, Kenworthy and Fonseca 1996, Koch and Beer 1996, Greve and Krause-Jensen 2005). Seagrasses can thrive up to depths where the irradiance at the top of the leaf canopy is above 11% of surface irradiance (Duarte 1991), or where the number of hours with values of irradiance above photosynthetic saturation is )6–8 h (Dennison and Alberte 1985). The upper depth limit of distribution of seagrasses has been related to their tolerance of desiccation at low tide (Koch and Beer 1996) and to ice scour (Robertson and Mann 1984). In seas with low tidal range and no ice formation, such as the Mediterranean Sea, the upper depth limit of seagrasses is determined mainly by their tolerance of wave energy (Koch et al. 2006). However, quantitative estimates of the wave energy that sets the upper depth limit of seagrasses are still scarce to date.

Studies in flumes have evaluated the influence of seagrass on unidirectional water flow (Fonseca et al. 1982, Gambi et al. 1990, Folkard 2005), sediment stabilization (Fonseca and Fisher 1986), and wave attenuation (Fonseca and Cahalan 1992). Field measurements and experiments have also been performed to study water flow in seagrass meadows (Fonseca 1986, Worcester 1995, Koch and Gust 1999) and evaluate their effect on particle resuspension (Terrados and Duarte 2000, Gacia and Duarte 2001). However, few studies have analyzed the effects of local hydrodynamics on seagrasses.

Storms can resuspend sediments and uproot seagrasses, whereas sediment deposition can bury them and cause mortality (Williams 1988, Preen et al. 1995, Duarte et al. 1997, Bell et al. 1999). Conceptual models have been proposed to explain differences in shape, bottom relief, and cover of seagrass meadows (Fonseca et al. 1983, Fonseca and Kenworthy 1987) or the depth distribution of intertidal seagrass (van Katwijk et al. 2000) as a function of wave energy and/or current velocity. Local hydrodynamics (indirectly estimated from the weight loss of clod cards) have been related to depth and seagrass presence and have been used to identify the habitat requirements of South Australian seagrass species

2009/27

Article in press - uncorrected proof 420 E. Infantes et al.: Wave energy and upper depth limit distribution of Posidonia oceanica

(Shepherd and Womersley 1981). Keddy (1984) developed a relative wave exposure index (REI) in order to quantify the degree of wave exposure by using wind speed, direction, and fetch measurements. REI values have been correlated to different attributes of seagrass meadows, such as the content of silt-clay and organic matter of the sediment, seagrass cover, shoot density (Fonseca and Bell 1998, Fonseca et al. 2002, KrauseJensen et al. 2003, Frederiksen et al. 2004), and biomass (Hovel et al. 2002). Shallow seagrass populations tend to be more spatially fragmented in wave-exposed environments than in wave-sheltered environments (Fonseca and Bell 1998, Frederiksen et al. 2004), and temporal changes in seagrass cover are also greater at the most wave-exposed sites (Frederiksen et al. 2004). Plaster clod cards and REI provide only semi-quantitative idiosyncratic estimates of current and/or wave energy and, therefore, comparison between studies is difficult. Additionally, REI estimates do not consider the influence of depth on wave damping (but see van Katwijk and Hermus 2000). Direct quantitative measurements of the energy of waves and currents impinging on seagrasses are necessary to elucidate their effects on them, to identify their habitat requirements, to predict their spatial distribution and their response to both natural (i.e., storms, hurricanes, etc.) and anthropogenic (i.e., dredging and beach eutrophication, coastal development, etc.) disturbances.

The main goal of this study was to develop a methodology to estimate the wave energy impinging on seagrass meadows (Figure 1) and to obtain a quantitative and testable relationship between wave energy and the upper depth limit of the Mediterranean seagrass Posidonia oceanica. Wave energy estimation was obtained by analyzing 44 years of wave data. We consider long-term historical wave data, rather than present-day wave measurements, are more appropriate to link wave energy to the presence of P. oceanica, given the low rates of vegetative growth and space occupancy in this seagrass species (Marba` and Duarte 1998). Aerial color photographs and bathymetric data were also used for an accurate mapping of the meadow and to delimit the upper

Figure 1 Schematic diagram of the approach presented.

HIPOCAS, hindcast of dynamic processes of the ocean and coastal areas of Europe.

depth boundary of P. oceanica. We provide a testable estimate of the near-bottom current velocity that could set a threshold for the presence of P. oceanica and, therefore, determine the upper depth limit of this seagrass.

Materials and methods Study areas and regional settings This work was carried out in Cala Millor, located on the northeast coast of Majorca Island (Figure 2A,B). The beach is in an open bay with an area of ca. 14 km2. Near the coast to 8 m depth, there is a regular slope indented with sand bars near the shore (Figure 2C); these bars migrate from offshore to onshore between periods of gentle wave conditions. At depths from 6 m to 35 m, the seabed is covered by a meadow of Posidonia oceanica.

This area was chosen for this study because of the availability of data from previous studies. The tidal regime is microtidal, with a spring range of less than 0.25 m (Go´mez-Pujol et al. 2007). The bay is located on the east coast of the island of Majorca and it is therefore exposed to incoming wind and waves from NE to ESE directions.

According to the criteria of Wright and Short (1983), Cala Millor is an intermediate barred sandy beach formed by biogenic sediments with median grain values ranging between 0.28 and 0.38 mm at the beach front (L. Go´mezPujol and A. Orfila, unpublished data).

HIPOCAS database (1958–2001) and deep water wave characterization Wave data used are part of the Hindcast of Dynamic Processes of the Ocean and Coastal Areas of Europe (HIPOCAS) project. This database consists of a high resolution, spatial and temporal, long-term hindcast dataset (Soares et al. 2002, Ratsimandresy et al. 2008).

The HIPOCAS data were collected hourly for the period 1958 to 2001, providing 44 years of wave data with a 0.1258 spatial resolution; this is the most complete wave data base currently available for the Mediterranean Sea. These hindcast models have become powerful tools, not only for engineering or predictive purposes, but also for long-term climate studies (Can˜ellas et al. 2007). These data were produced by the Spanish Harbor Authority by dynamical downscaling from the National Center for Environmental Prediction (NCEP) and the National Center for Atmospheric Research (NCAR) global reanalysis using the regional atmospheric model REMO. We used the data from HIPOCAS node 1433 (see Figure 2B) located 10 km offshore at 50 m depth, which is the closest HIPOCAS node to Cala Millor. The long-term distribution of significant wave height and wave direction at this node (Figure 3) shows that the most energetic waves usually come from N–NNE. These wave directions are also the most frequent during the 44-year dataset.

Data contained in the HIPOCAS node consist of a set of sea states (one per hour) defined by their significant wave height, spectral peak period, and direction. An estimation of the long-term distribution of the mean significant wave height (Hm) and its standard deviation was

Article in press - uncorrected proof E. Infantes et al.: Wave energy and upper depth limit distribution of Posidonia oceanica 421

Figure 2 (A) Location of Majorca in the Mediterranean Sea. (B) Location of the study area of Cala Millor in Majorca. The asterisk (*) indicates HIPOCAS node 1433. (C) Bathymetry of Cala Millor with isobaths (in meters).

carried out making use of the lognormal probability distribution (Castillo et al. 2005). The long-term probability distribution identifies the most probable sea state for the 44-year period. Before estimation of this wave regime, data were pre-selected taking into account their incoming directions. Only sea states directed towards the beach were included in the analysis. Mean wave climate for Hipocas node 1433 provides an Hm value of 1.53"0.96 ("1 SD) m, a peak period of 7.3 s, and a direction of 11.258. This mean (most probable) wave cli-

Figure 3 Directional wave histogram for HIPOCAS node 1433 (50 km from Cala Millor).

mate was propagated to the shore using a parabolic model.

Shallow-water wave conditions As water waves propagate from the region where they are generated to the coast, both wave amplitude and wavelength are modified. The surf zone is a highly dynamic area where energy from waves is partially dissipated through turbulence in the boundary layer and transformed into short and long waves, mean sea level variations, and currents (Dean and Dalrymple 2002). In the present work, waves were propagated using a gentle slope parabolic model (OLUCA-MC), which includes refraction-diffraction effects as well as energy losses due to wave breaking (Kirby and Dalrymple 1983, GIOC 2003). Detailed bathymetry obtained with echo-sounding was used to generate the numerical mesh. The model solves continuity and momentum equations assuming a smooth bottom (e.g., variations of the bottom negligible within a wave length) and converting the hyperbolic system to a parabolic system (e.g., with wave propagation in one direction).

Two grids were generated: the external grid (122=81 nodes), which covers the deeper area with 75 m resolution between nodes, and the internal grid (140=311 nodes), which covers the shallow area with 15 m resolution. The model output provides the wave field (significant wave height and direction of the mean flux energy) in the whole grid.

Article in press - uncorrected proof 422 E. Infantes et al.: Wave energy and upper depth limit distribution of Posidonia oceanica

Maximum near-bottom orbital amplitudes (Ab) were calculated following the linear wave theory: Abs

Hs 2sinhŽ2pDyl.

(1)

where D is the water depth and l is the wave length calculated iteratively as:

ls

B 2pD E T 2pg F tanhC D l G 2p

(2)

where Tp is the peak period and g is the acceleration of gravity. The maximum near-bottom orbital velocity (Ub) is: Ubs2pAbyTp

(3)

Bottom typology and bathymetry Remote sensing of the seabed from air or space is commonly used for mapping seagrass habitats over a wide range of spatial scales (McKenzie et al. 2001). Satellite spectral images from Ikonos are suitable for the detection and mapping of the upper depth limit of seagrass distribution in shallow clear waters (Fornes et al. 2006). Aerial color photographs have been used in some studies to describe temporal changes in the distribution of seagrasses (Hine et al. 1987), to evaluate the effect of wave exposure (Frederiksen et al. 2004), and the influence of anthropogenic activities (Leriche et al. 2006). We used aerial photographs for mapping the area covered by Posidonia oceanica, dead P. oceanica rhizome, rocks, and sand up to 11 m depth. Aerial color photographs were taken in August of 2002 with a resolution of 0.4 m. Polygons were drawn around the different areas with Arc/GIS software (Arc/Info and Arc/Map v9.0, ESRI) and classified by bottom types. In those areas where bottom recognition was not possible from the aerial photographs, field surveys were carried out to identify the typology of the substratum (some areas tend to accumulate seagrass leaves which can lead to false interpretation of aerial images). Image classification was validated with bathymetric filtered echo-soundings and field observations.

During 2005, an acoustic survey was carried out to determine the bathymetry of the inner mesh and to test the classification of the seagrass coverage from the aerial photographs. Acoustic mapping of Posidonia oceanica was performed with a ship-mounted Biosonics DE-4000 echo sounder (BioSonics, Inc., Seattle, USA) equipped with a 200 KHz transducer. The draught of the boat allowed sampling up to depths of approximately 0.5 m.

Inshore-offshore echo-sounding transects were sampled perpendicularly to the bathymetric gradient, with a separation of 50 m between transects. Acoustic pulse rate was set to 25 s-1 and the sampling speed was set to 3 knots, which allowed for a horizontal resolution of 1 m (Orfila et al. 2005). Bottom typology was estimated as the most probable after echogram examination using the first to second bottom echo ratio technique (Orlowski 1984, Chivers et al. 1990). The resulting echo sounding points were filtered, averaged (1 output equals 20 pings) and

clustered into three groups (P. oceanica meadows, sandy, and hard bottoms) taking into consideration calibrations performed for previously classified bottoms.

Hard bottoms include rocks, P. oceanica rhizome mats, and zones with poor seagrass coverage (Figure 4). Afterwards, this map was verified by direct observation at random points, also distinguishing those bottom types that the algorithm was not able to identify (i.e., dead rhizome and rocks). The map of bottom typology in Figure 4 represents the final classification (aerial photograph verified with echo-sounding and direct SCUBA observations).

Bathymetric data were interpolated using the kriging technique to create 1-m scale depth contours that were overlaid with bottom typology. Similarly, maximum nearbottom orbital velocity data were interpolated to create 5-cm s-1 scale Ub contours that were overlaid with bottom typology. Percent coverage of rocks, sand, dead rhizomes of Posidonia oceanica, and P. oceanica were calculated for each depth and near-bottom orbital velocity interval. A total of 400 points were randomly selected throughout the study area, and the corresponding bottom type and near-bottom orbital velocity interval were used to calculate the average near-bottom orbital velocity for each bottom type. Additionally, 400 points along the upper limit of the P. oceanica meadow were randomly selected and the corresponding near-bottom orbital velocities were used to estimate an average along that edge. Differences in near-bottom orbital velocity between each bottom type and the upper limit of P. oceanica were evaluated using Kruskal-Wallis non-parametric analysis of variance (due to heterogeneity of data variance).

Results The study site has a total area of 121.44 ha wherein sand and Posidonia oceanica meadow are the most abundant substrata. Shallow bottoms in Cala Millor (depths between 0 and 6 m) are mostly sandy, with some patches of rock, particularly in the center and south parts of the beach (Table 1). The upper depth limit of P. oceanica is located between 5 and 6 m and the meadow continues down to 30â&#x20AC;&#x201C;35 m depth (acoustic survey data not shown). The limit between seagrass and sand is irregular and has several areas of hard substratum between 4 and 7 m that correspond to dead rhizomes of P. oceanica partly covered by sand and algae, which can be indicative of meadow regression (Figure 4). P. oceanica reaches the highest percentage cover at depths greater than 8 m, and no stands are found in depths shallower than 4 m.

Most of the rocky bottom is in the shallowest water, between 0 and 2 m. Two large sand fingers cross the seagrass meadow at the center of the study area; these may have resulted from sand transport from the exposed beach after storm events.

Mean wave conditions propagated over the beach resulted in wave heights between 0.2 and 0.4 m, with significant wave breaking in the shallow sandy area (between 0.5 and 1 m) (Figure 5A) and near-bottom orbital velocities up to 110 cm s-1. The highest velocities occurred in the shallow sandy part of the beach (between 1 and 3 m) (Figure 5B).

Article in press - uncorrected proof E. Infantes et al.: Wave energy and upper depth limit distribution of Posidonia oceanica 423

not present in areas with velocities higher than 38–42 cm s-1. This velocity interval might be considered a first estimate of the threshold near-bottom orbital velocity that allows the P. oceanica to occur in Cala Millor.

Variance of near-bottom orbital velocity was higher in sand, rock, and dead Posidonia oceanica than in the P. oceanica meadow (Figure 7). Kruskal-Wallis non-parametric analysis of variance detected significant differences in the average near-bottom orbital velocity between bottom types wH (4, ns800)s310.34, p-0.001x, and post hoc multiple comparisons of mean ranks showed that velocities were lower in P. oceanica stands and at the P. oceanica upper limit than in rock, sand, and dead P. oceanica (Figure 7).

Discussion

Figure 4 Bottom typology off Cala Millor (Majorca, western Mediterranean Sea). P., Posidonia.

Sandy areas are associated with high values of nearbottom orbital velocities, while Posidonia oceanica is associated with lower velocities (Figure 6). P. oceanica is

In this study, we present a methodology to estimate the wave energy that determines the upper depth limit of Posidonia oceanica. Our data show that an increase in wave energy is related to a decrease of P. oceanica cover, and that above a threshold level of wave energy seagrass is not present. It is important to emphasize that the evidence provided in this study is correlative and applies to our study site only. Other locations, with different sediment characteristics and wave climates might provide different threshold values. Additional sources of disturbance (both natural and anthropogenic) will also introduce variability in the threshold estimates. The near-bottom orbital velocities are computed from numerical model predictions and real velocities within the meadow could be lower due to wave attenuation by the seagrass meadow and wave-current interactions. However, the predicted velocities where P. oceanica is not present provide an estimate of the threshold velocities that would impede the persistence of this seagrass species.

The usefulness of this methodology is that it provides quantitative estimates of wave energy that sets the upper depth limit of Posidonia oceanica and, therefore, these can be compared to those obtained in other locations.

We obtained an estimate of the threshold value of nearbottom orbital velocity that allows the long-term persistence of Posidonia oceanica (38–42 cm s-1). It has been suggested that Zostera marina L. can tolerate unidirec-

Table 1 Percent cover of bottom type at different depths.

Depth (m)

Posidonia oceanica

Sand

Rock

Dead Posidonia oceanica

1–2 2–3 3–4 4–5 5–6 6–7 7–8 8–9 9–10 10–11 Total area (m2)

0 0 0 0.46 11.55 42.34 73.87 83.82 94.40 97.98 565,349.42

96.63 97.73 97.31 83.36 51.03 25.27 21.10 14.64 5.59 2.02 528,241.5

3.24 1.75 0 0 0 0 0 0 0 0 25,192.7

0.13 0.52 2.69 16.17 37.42 32.39 5.03 1.54 0 0 95,632.92

Total area of coverage of each bottom type is presented in the last row.

Article in press - uncorrected proof 424 E. Infantes et al.: Wave energy and upper depth limit distribution of Posidonia oceanica

Figure 5 (A) Distribution of wave heights (m) at the beach derived from mean wave conditions (11.258, Hms1.53 m, Tps7.30 s).

(B) Distribution of near-bottom orbital velocities (Ub, m s-1) at the beach and cover of the Posidonia oceanica meadow (gray) and dead rhizomes (dark gray).

Figure 6 Percent coverage of the different bottom types in each of the near-bottom orbital velocity (Ub) intervals established in Cala Millor (Majorca, Western Mediterranean Sea). P., Posidonia.

tional current velocities up to 120â&#x20AC;&#x201C;150 cm s-1 and that the meadows formed by this species become spatially fragmented at tidal current velocities of 53 cm s-1 (Fonseca et al. 1983). Shallow mixed meadows of Z. marina and Halodule wrightii Asch. seem to remain spatially fragmented at tidal current speeds above 25 cm s-1 (Fonseca and Bell 1998). Experimental transplantations of Z. marina along depth gradients in intertidal zones indicate that this species cannot persist at sites where the maximum

bottom orbital velocity during the tidal cycle reaches 53â&#x20AC;&#x201C;63 cm s-1. Furthermore, this species can survive when exposed to waves for less than 60% of the time and maximum orbital velocity is less than 40 cm s-1 (van Katwijk and Harmus 2000). Hence, our estimate of the threshold value of near-bottom orbital velocity that allows the persistence of P. oceanica is within the range of values of current velocity proposed for other seagrass species.

Article in press - uncorrected proof E. Infantes et al.: Wave energy and upper depth limit distribution of Posidonia oceanica 425

Figure 7 Mean near-bottom orbital velocity above each bottom type in Cala Millor (Majorca, Western Mediterranean Sea).

Different capital letters indicate significant differences between bottom types (post hoc multiple pairwise comparison of mean ranks, p-0.05). Error bars show 95% confidence intervals. N values indicate the number of points selected randomly in each bottom type.

Differences in N values between bottom types are driven by the differences in percentage covers of each bottom type in the study area. P., Posidonia.

Frederiksen et al. (2004) used aerial photographs to follow changes in the distribution of Zostera marina from 1954 to 1999 and showed that seagrass landscapes can change extensively over long periods of time, especially in the more wave-exposed areas. Comparison of aerial photographs taken in 1956 and 2004 indicates that the upper depth limit of Posidonia oceanica in Cala Millor has regressed in the south part of the beach (IMEDEA 2005), which is a sector of the beach well exposed to the most energetic waves (those from the N–NNE). We do not rule out, however, that other processes may determine the upper depth limit of P. oceanica. This seagrass species is able to grow in certain locations, usually sheltered, almost to sea level (Ribera et al. 1997), which suggests that neither photo-inhibition nor temperature fluctuations associated with shallow depths are important in setting the upper depth limit. Except for the desiccation effects associated with exposure to air at low tides (negligible in the Mediterranean Sea), the consensus is that the upper depth limit of seagrasses is set by the energy of currents and waves (Koch et al. 2006).

Wave breaking might have a role in establishing the upper limit, but the position at which waves break will change depending on wave height, wave period, and presence of submerged sandbars. In addition, the effect of waves of certain energy may depend on the amount of time the plants are exposed to different levels of wave energy. Average wave conditions have been used to estimate the average field of near-bottom orbital velocities, but other wave statistics could be used, such as extreme events (De Falco et al. 2008), and elaborate a prediction of those velocities that allow the long-term persistence of P. oceanica. We used 44 years of data, as this would be an adequate period of time over which to average the effects of waves on the slow-growing P. oceanica (Marba` and Duarte 1998).

We note that sediment density and grain size characteristics have not been considered in this study. These

granulometric characteristics may play an important role in determining the location of seagrasses, because wave energy will move and resuspend the sediment, which might bury or erode a seagrass meadow. Differences in sediment characteristics between locations will then contribute to the variability of the estimate of the threshold wave energy that sets the upper depth limit of Posidonia oceanica.

We conclude that the methodology presented here is a useful tool to estimate wave energy on the bottom and to identify the level that sets the upper depth limit of Posidonia oceanica meadows. Our approach provides testable estimates of the threshold level of wave energy that allows this species to persist and indicates a research program to validate them. This study also highlights the importance of interdisciplinary and multidisciplinary research in ecological modeling and, in particular, the need for hydrodynamical studies to elucidate the distribution of seagrasses at shallow depths.

Acknowledgements E. Infantes would like to thank the Spanish Ministerio de Educacio´n y Ciencia (MEC) FPI scholarship program (BES-200612850) for financial support. Financial support from grants CTM2005-01434, CTM2006-12072/MAR, and the Integrated Coastal Zone Management (UGIZC) project are acknowledged.

The authors would like to thank Puertos del Estado for the HIPOCAS dataset and Marta Fuster for assistance with the data.

References Bell, S.S., B.D. Robbins and S.L. Jensen. 1999. Gap dynamics in a seagrass landscape. Ecosystems 2: 493–504.

Can˜ellas, B., A. Orfila, F.J. Me´ndez, M. Mene´ndez and J. Tintore.

2007. Application of POT model to estimate the extreme significant wave height levels around the Balearic Sea (Western Mediterranean). J. Coast. Res. 50: 329–333.

Article in press - uncorrected proof 426 E. Infantes et al.: Wave energy and upper depth limit distribution of Posidonia oceanica

Castillo, M.C., A. Baquerizo and M.A. Losada. 2005. Temporal and spatial statistical variability of the wave height in the surf zone. 29th International Conference on Coastal Engineering, Lisbon, Portugal, 2004: 997–1008.

Chivers, R., N. Emerson and D. Burns. 1990. New acoustic processing for underway surveying. Hydrogr. J. 56: 8–17.

Dean, R.G. and R.A. Dalrymple. 2002. Coastal processes with engineering applications. Cambridge University Press, Cambridge. pp. 475.

De Falco, G., M. Baroli, A. Cucco and S. Simeone. 2008. Intrabasinal conditions promoting the development of a biogenic carbonate sedimentary facies associated with the seagrass Posidonia oceanica. Continent. Shelf Res. 28: 797–812.

Dennison, W.C. 1987. Effects of light on seagrass photosynthesis, growth and depth distribution. Aquat. Bot. 27: 15–26.

Dennison, W.C. and R.S. Alberte. 1985. Role of daily light period in the depth distribution of Zostera marina (eelgrass). Mar. Ecol. Prog. Ser. 25: 51–61.

Duarte, C.M. 1991. Seagrass depth limits. Aquat. Bot. 40: 363–377.

Duarte, C.M., J. Terrados, N.S.R. Agawin, M.D. Fortes, S.S. Bach and W.J. Kenworthy. 1997. Response of a mixed Philippine seagrass meadow to experimental burial. Mar. Ecol.

Prog. Ser. 147: 285–294.

Folkard, A.M. 2005. Hydrodynamics of model Posidonia oceanica patches in shallow water. Limnol. Oceanogr. 50: 1592–1600.

Fonseca, M.S. and J.S. Fisher. 1986. A comparison of canopy friction and sediment movement between 4 species of seagrass with reference to their ecology and restoration. Mar. Ecol. Prog. Ser. 29: 15–22.

Fonseca, M.S. and W.J. Kenworthy. 1987. Effects of current on photosynthesis and distribution of seagrasses. Aquat. Bot. 27: 59–78.

Fonseca, M.S. and J.A. Cahalan. 1992. A preliminary evaluation of wave attenuation by 4 species of seagrass. Est. Coast.

Shelf Sci. 35: 565–576.

Fonseca, M.S. and S.S. Bell. 1998. Influence of physical settings on seagrass landscapes near Beaufort, North Carolina, USA. Mar. Ecol. Prog. Ser. 171: 109–121.

Fonseca, M.S., J.S. Fisher, J.C. Zieman and G.W. Thayer. 1982.

Influence of the seagrass, Zostera marina L., on current flow.

Est. Coast. Shelf Sci. 15: 351–364.

Fonseca, M.S., J.C. Zieman, G.W. Thayer and J.S. Fisher. 1983.

The role of current velocity in structuring eelgrass (Zostera marina L.) meadows. Est. Coast. Shelf Sci. 17: 367–380.

Fonseca, M.S., P.E. Whitfield, N.M. Kelly and S.S. Bell. 2002.

Modeling seagrass landscape pattern and associated ecological attributes. Ecol. Appl. 12: 218–237.

Fornes, A., G. Basterretxea, A. Orfila, A. Jordi, A. A´lvarez and J. Tintore. 2006. Mapping Posidonia oceanica from IKONOS. J. Photogramm. Remote Sens. 60: 315–322.

Frederiksen, M., D. Krause-Jensen, M. Holmer and J.S. Laursen.

2004. Spatial and temporal variation in eelgrass (Zostera marina) landscapes: influence of physical setting. Aquat. Bot. 78: 147–165.

Gacia, E. and C.M. Duarte. 2001. Sediment retention by a Mediterranean Posidonia oceanica meadow: the balance between deposition and resuspension. Est. Coast. Shelf Sci. 52: 505–514.

Gambi, M.C., A.R.M. Nowell and P.A. Jumars. 1990. Flume observations on flow dynamics in Zostera marina (eelgrass) beds. Mar. Ecol. Prog. Ser. 61: 159–169.

GIOC. 2003. Spectral wave propagation model (Oluca-SP). State Coastal Office, Spanish Environmental Ministry and University of Cantabria, Santander. pp. 170 (in Spanish).

Go´mez-Pujol, L., A. Orfila, B. Canellas, A. Alvarez-Ellacuria, F.J. Mendez, R. Medina, et al. 2007. Morphodynamic classification of sandy beaches in low energetic marine environment.

Mar. Geol. 242: 235–246.

Greve, T. and D. Krause-Jensen. 2005. Predictive modelling of eelgrass (Zostera marina) depth limits. Mar. Biol. 146: 849–858.

Hine, A.C., M.W. Evans, R.A. Davis and D.F. Belknap. 1987. Depositional response to seagrass mortality along a low-energy, barrier-island coast: west-central Florida. J. Sediment. Petrol.

57: 431–439.

Hovel, K.A., C.R. Fonseca, D.L. Myer, W.J. Kenworthy and P.E. Whitfield. 2002. Effects of seagrass landscape structure, structural complexity and hydrodynamic regime on macrofaunal densities in North Carolina seagrass beds. Mar. Ecol.

Prog. Ser. 243: 11–24.

IMEDEA. 2005. Variabilidad y dina´mica sedimentaria de las playas de Cala Millor y Cala Sant Vicent. Volumen 1. Govern de les Illes Balears, Cala Millor. pp. 103.

Keddy, P.A. 1984. Quantifying a within-lake gradient of wave energy in Gillfillan Lake, Nova-Scotia. Can. J. Bot. 62: 301– 309.

Kenworthy, W.J. and M.S. Fonseca. 1996. Light requirements of seagrasses Halodule wrightii and Syringodium filiforme derived from the relationship between diffuse light attenuation and maximum depth distribution. Estuaries 19: 740–750.

Kirby, J.T. and R.A. Dalrymple. 1983. The propagation of weakly nonlinear waves in the presence of varying depth and currents. In: Proceedings of the XXth Congress I.A.H.R., Moscow 7, Moscow, September, 1983: 198–202.

Koch, E.W. and S. Beer. 1996. Tides, light and the distribution of Zostera marina in Long Island Sound, USA. Aquat. Bot. 53: 97–107.

Koch, E.W. and G. Gust. 1999. Water flow in tide- and wavedominated beds of the seagrass Thalassia testudinum. Mar. Ecol. Prog. Ser. 184: 63–72.

Koch, E.W., J.D. Ackerman, J. Verduin and M. van Keulen. 2006.

Fluid dynamics in seagrass ecology – from molecules to ecosystems. In: (A.W.D. Larkum, R.J. Orth, C.M. Duarte, eds).

Seagrasses: biology, ecology and conservation. Springer, Dordrecht. pp. 193–224.

Krause-Jensen, D., M.F. Pedersen and C. Jensen. 2003. Regulation of eelgrass (Zostera marina) cover along depth gradients in Danish coastal waters. Estuaries 26: 866–877.

Leriche, A., V. Pasqualini, C.F. Boudouresque, G. Bernard, P. Bonhomme, P. Clabaut, et al. 2006. Spatial, temporal and structural variations of a Posidonia oceanica seagrass meadow facing human activities. Aquat. Bot. 84: 287–293.

Marba`, N. and C.M. Duarte. 1998. Rhizome elongation and seagrass clonal growth. Mar. Ecol. Prog. Ser. 174: 269–280.

McKenzie, L.J., M.A. Finkbeine and H. Kirkman. 2001. Methods for mapping seagrass distribution. In: (F.T. Short and R.G. Coles, eds). Global seagrass research methods. Elsevier, Amsterdam. pp. 101–121.

Orfila, A., A. Jordi, G. Basterretxea, G. Vizoso, N. Marba, C.M. Duarte, et al. 2005. Residence time and Posidonia oceanica in Cabrera Archipelago National Park, Spain. Continent. Shelf Res. 25: 1339–1352.

Orlowski, A. 1984. Application of multiple echo energy measurements for evaluation of sea bottom type. Oceanologia 19: 61–78.

Pasqualini, V., C. Pergent-Martini, P. Clabaut and G. Pergent.

1998. Mapping of Posidonia oceanica using aerial photographs and side scan sonar: application off the Island of Corsica (France). Est. Coast. Shelf Sci. 47: 359–367.

Preen, A.R., W.J. Lee Long and R.G. Coles. 1995. Flood and cyclone related loss, and partial recovery, of more than 1000 km2 of seagrass in Hervey Bay, Queensland, Australia. Aquat.

Bot. 52: 3–17.

Procaccini, G., M.C. Buia, M.C. Gambi, M. Perez, G. Pergent, C. Pergent-Martini and J. Romero. 2003. The seagrasses of the Western Mediterranean. In: (E.P. Green, F.T. Short, eds).

World atlas of seagrasses, UNEP World Conservation Monitoring Centre. University of California Press, Berkeley, CA. pp. 48–58.

Article in press - uncorrected proof E. Infantes et al.: Wave energy and upper depth limit distribution of Posidonia oceanica 427

Ratsimandresy, A.W., M.G. Sotillo, J.C. Carretero Albiach, E. A´lvarez Fanjul, H. Hajji. 2008. A 44-year high-resolution ocean and atmospheric hindcast for the Mediterranean Basin Developer within the HIPOCAS Project. Coast. Eng. 55: 825–826.

Ribera, G., M. Coloreu, C. Rodrı´guez-Prieto and E. Ballesteros.

1997. Phytobenthic assemblages of Addaia Bay (Menorca, Western Mediterranean): composition and distribution. Bot. Mar. 40: 523–532.

Robertson, A.I. and K.H. Mann. 1984. Disturbance by ice and life-history adaptations of the seagrass Zostera marina. Mar. Biol. 80: 131–141.

Shepherd, S.A. and H.B.S. Womersley. 1981. The algal and seagrass ecology of Waterloo Bay, South-Australia. Aquat. Bot. 11: 305–371.

Soares, C.G., J.C. Carretero, R. Weisse and E. Alvarez. 2002. A 40 years hindcast of wind, sea level and waves in European waters. In: Proceedings of OMAE 2002: 21st International Conference on Offshore Mechanics and Artic Engineering, Oslo, Norway 2000: 1–7.

Terrados, J. and C.M. Duarte. 2000. Experimental evidence of

reduced particle resuspension within a seagrass (Posidonia oceanica) meadow. J. Exp. Mar. Biol. Ecol. 243: 45–53. van Katwijk, M.M. and D.C.R. Hermus. 2000. Effects of water dynamics on Zostera marina: transplantation experiments in the intertidal Dutch Wadden Sea. Mar. Ecol. Prog. Ser. 208: 107–118. van Katwijk, M.M., D.C.R. Hermus, D.J. de Jong, R.M. Asmus and V.N. de Jonge. 2000. Habitat suitability of the Wadden Sea for restoration of Zostera marina beds. Helgol. Mar. Res. 54: 117–128.

Williams, S.L. 1988. Disturbance and recovery of a deep-water Caribbean seagrass bed. Mar. Ecol. Prog. Ser. 42: 63–71.

Worcester, S.E. 1995. Effects of eelgrass beds on advection and turbulent mixing in low current and low shoot density environments. Mar. Ecol. Prog. Ser. 126: 223–232.

Wright, L.D. and A.D. Short. 1983. Morphodynamics of beaches and surf zones in Australia. In: (P.D. Koma, ed). Handbook of coastal processes and erosion. CRC Press, Boca Raton, FL. pp. 35–64.

Received 26 February, 2008; accepted 23 September, 2008; online first 13 February, 2009

Estuarine, Coastal and Shelf Science 91 (2011) 434e441

Contents lists available at ScienceDirect

Estuarine, Coastal and Shelf Science journal homepage: www.elsevier.com/locate/ecss

Assessment of substratum effect on the distribution of two invasive Caulerpa (Chlorophyta) species Eduardo Infantes*, Jorge Terrados, Alejandro Orfila Instituto Mediterráneo de Estudios Avanzados, IMEDEA (CSIC-UIB), Miquel Marqués 21, 07190 Esporles, Spain

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 February 2010 Accepted 28 November 2010 Available online 3 December 2010

Two-year monitoring of the invasive marine Chlorophyta Caulerpa taxifolia and Caulerpa racemosa var. cylindracea shows the great influence of substratum on their spatial distribution. The cover of C. taxifolia and C. racemosa was measured in shallow (<8 m) areas indicating that these species are more abundant in rocks with photophilic algae and in the dead matte of the seagrass Posidonia oceanica than in sand or inside the P. oceanica meadow. A short-term experiment comparing the persistence of C. taxifolia and C. racemosa planted either in a model of dead matte of P. oceanica or in sand shows that the persistence of these species was higher in the dead matte model than in sand. Correlative evidence suggests that C. taxifolia and C. racemosa tolerate near-bottom orbital velocities below 15 cm s 1 and that C. taxifolia cover declines at velocities above that value. These results contribute to understand the process of invasion of these Caulerpa species predicting which substrata would be more susceptible to be invaded and to the adoption of appropriate management strategies. Ó 2010 Elsevier Ltd. All rights reserved.

Keywords: Caulerpa racemosa var. cylindracea Caulerpa taxifolia Posidonia oceanica substratum invasive macroalgae

1. Introduction Light and temperature are identified as the main determinants for the spatial distribution of marine macrophytes but substratum type and hydrodynamic conditions are also key factors limiting the distribution of marine vegetation (Shepherd and Womersley, 1981; Fonseca and Kenworthy, 1987; Riis and Hawes, 2003; Frederiksen et al., 2004; Koch et al., 2006). Consolidated substrata (i.e., rocks) offer a stable, non-mobile surface where macroalgae and some seagrass species attach effectively. Unconsolidated substrata (i.e., sand, mud) are unstable, mobile and only seagrasses and macroalgae with root-like structures can colonize them. Waves and currents can mobilize sediment particles producing sediment erosion and accretion that may affect seagrasses negatively through uprooting and burial (Fonseca and Kenworthy,1987; Williams,1988; Terrados,1997; Cabaço et al., 2008). Thresholds of maximum wave energy tolerated by some seagrasses have been determined (Fonseca and Bell, 1998; van Katwijk and Hermus, 2000; Koch, 2001; Infantes et al., 2009) as well as the relationship between wave exposure and temporal variability of seagrass coverage in shallow sands (Fonseca et al.,1983; Fonseca and Bell, 1998; Frederiksen et al., 2004). By contrast, information about the wave energy levels that macroalgae tolerate is limited (D’Amours and Scheibling, 2007; Scheibling and Melady, 2008), especially in unconsolidated substrata. * Corresponding author.

E-mail address: eduardo.infantes@uib.es (E. Infantes).

0272-7714/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.ecss.2010.11.005

Most marine macroalgae grow over consolidated substrata such as rocks or over other macrophytes but a few Chlorophyta species of the order Caulerpales are also able to grow on unconsolidated substrata (Taylor, 1960). The thallus of Caulerpa species is composed of a creeping portion, the stolon, that it is attached to the substrata by root-like structures, the rhizoids, and an erect portion, the fronds that have different shapes depending on the species (Taylor, 1960; Bold and Wynne, 1978). The rhizoids of Caulerpa are able to bind sediment particles (Chisholm and Moulin, 2003) and anchor the plant in sandy sediments or other unconsolidated substrata, and to attach to rocks and other macrophytes (Taylor, 1960; Meinesz and Hesse, 1991; Klein and Verlaque, 2008). Hence Caulerpa species may be present both in wave exposed, rocky bottoms and sheltered, sandy-muddy sediments (Thibaut et al., 2004).

Caulerpa taxifolia (Vahl) C. Agardh and Caulerpa racemosa var. cylindracea (Sonder) Verlaque, Heisman et Boudouresque (hereafter C. racemosa) have been introduced in the Mediterranean Sea within the last 25 years showing an invasive behaviour (Meinesz and Hesse, 1991; Boudouresque and Verlaque, 2002; Klein and Verlaque, 2008).

Understanding the factors that influence the establishment and spread of these introduced Caulerpa species is essential to identify benthic communities susceptible of being invaded and to have scientific tools to predict future invasions. Initial observations of C. taxifolia invasion suggested that rocks and dead matte of Posidonia oceanica were appropriate for the establishment of this species but sandy or muddy sediments in sheltered conditions were also colonized by this species (Hill et al., 1998). A short-term experiment

E. Infantes et al. / Estuarine, Coastal and Shelf Science 91 (2011) 434e441

435

showed that algal turfs were more favourable for the establishment of C. taxifolia and C. racemosa than other macroalgal communities on rocky bottoms (Ceccherelli et al., 2002). Both Caulerpa species seem to tolerate a certain level of sediment deposition and burial (Glasby et al., 2005; Piazzi et al., 2005) a feature that likely facilitates their development in sandy sediments.

In this study we evaluate the effect of substratum on the persistence of two invasive Caulerpa species. First, we quantify the presence of Caulerpa taxifolia and Caulerpa racemosa in different types of substrata during two consecutive summers and winters testing if their presence was independent of the type of substratum.

Then, we perform a short-term experiment where fragments of C. taxifolia and C. racemosa are established in a model of dead matte of Posidonia oceanica and in sand and their development followed from summer to winter. The cover of both Caulerpa species is correlated to near-bottom orbital velocities and friction coefficients over sand and the model dead matte are computed.

2. Material and methods 2.1. Study sites The study was carried out in four shallow (depth less than 8 m) locations on the South coast of Mallorca Island, Western Mediterranean sea (Fig. 1a). Cala D’Or was chosen for this study because it is the only location in the Balearic Islands where Caulerpa taxifolia is present. Caulerpa racemosa is mainly found in the South of Mallorca, and the locations of Sant Elm, Cala Estancia and Portals Vells where chosen (Fig. 1b). Sand patches, rocky reefs covered by an erect stratum of photophilous macroalgae and meadows of the seagrass Posidonia oceanica are the main components of the undersea landscape at the study sites.

2.2. Presence of Caulerpa over different substrata The cover of Caulerpa taxifolia in Cala D’Or and that of Caulerpa racemosa in Cala Estancia and Portals Vells was measured during two consecutive summers and winters (years 2007 and 2008) which represent the periods of maximum and minimum vegetative development of these species in the Western Mediterranean (Piazzi and Cinelli, 1999). Four to eight 20-m long transects were laid haphazardly at each location and the percentage of substratum covered by Caulerpa was estimated at 1-m intervals along the

Fig. 2. Photographs of the experimental set up for Caulerpa taxifolia in a) natural sandy bottom and b) the model of dead matte of the seagrass Posidonia oceanica. c) Illustration of Caulerpa fragments fixed to the plots using pickets.

transects by placing two 20 cm 20 cm quadrats divided in 25 cells of 4 cm2. The number of cells corresponding to each type of substratum in each quadrat was counted as well as the number of cells of each substratum type colonized by Caulerpa. Substrata were classified either as sand, rocks (always covered by photophilic algae), Posidonia oceanica meadow or dead matte of P. oceanica. The matte of P. oceanica is formed by the accumulation of sediment, rhizomes and/or roots of this seagrass species over time, building a terraced structure that retains the sediment even after seagrass death and that may last hundreds of years (Mateo et al., 1997).

2.3. Short-term experiment evaluating substratum effect on Caulerpa persistence The experiment evaluated the persistence of Caulerpa fragments planted in sand plots (Fig. 2a) and in plots of a model of Posidonia oceanica dead matte (Fig. 2b). The experiment was performed in two

Fig. 1. Location of the study areas. (a) Mallorca Island in the Mediterranean Sea. (b) Location of the four study areas and deep water wave data (WANA nodes). (c) Bathymetry of Cala D’Or with location of WANA node. (d) Bathymetry of Sant Elm. C Experimental plots, , WANA nodes and B ADCP location in Cala D’Or.

436

E. Infantes et al. / Estuarine, Coastal and Shelf Science 91 (2011) 434e441

locations: Cala D’Or for Caulerpa taxifolia and Sant Elm for Caulerpa racemosa (Fig.1b). We did so because C. racemosa is not present in Cala D’Or, which is the only place where C. taxifolia is present in the Balearic Islands, and we did not consider appropriate to contribute to the dispersion of these invasive species by transplanting them to other locations. Besides, we did not perform the C. racemosa experiment in Cala Estancia or Portals Vells because these sites are highly visited by swimmers and snorkelers and the risk of disturbance to experimental plots was high. A sand patch large enough to hold all the experimental plots at a depth of 7e8 m was selected at each location. These sand patches were surrounded by P. oceanica meadows.

Six 50 cm 50 cm plots were established in each location separated every 5 m. Plots were delimited at the corners with galvanized iron bars. The model of dead matte of Posidonia oceanica was constructed by attaching fragments of dead P. oceanica rhizomes collected from fresh beach-cast plants after a storm to 50 cm 50 cm squares of plastic gardening mesh, using plastic cable ties (Fig. 2b). A density of around 500 vertical rhizomes per m2 was used as this is a typical P. oceanica shoot density found at 8 m depth in clear Mediterranean waters (Marbà et al., 2002; Gobert et al., 2003; Terrados and MedinaPons, 2008). A model of dead matte of P. oceanica was used instead of natural occurring dead matte to standardize the methodology having the same substratum types in both locations. The dead matte models were attached to the bars delimiting the randomly assigned plots (n ¼ 3) with cable ties and using four pickets to anchor it firmly to the sediment (Fig. 2b). The sand plots (n ¼ 3) were made of the naturally occurring sand and placement of the galvanized iron bars delimiting plot corners was the only manipulation performed.

Fragments of Caulerpa including at least one stolon apex and 3e4 fronds each were manually collected from the same area and depth where the experimental plots were located and held in aerated seawater at sea surface until planting. Sixteen Caulerpa fragments were fixed to each plot within 2 h after collection by threading them in groups of four into four cable-tie loops anchored in each plot using galvanized iron pickets (see Fig. 2c). The experiment started in July of 2007 as previous studies on biomass seasonality of Caulerpa taxifolia in shallow waters (6e10 m) have shown that maximum biomass occurs in summer and autumn (Thibaut et al., 2004). Monitoring of Caulerpa cover in each plot was carried out approximately every 30 days for C. taxifolia and 15 days for Caulerpa racemosa as the latter species has a higher growth rate (Piazzi et al., 2001) using underwater digital photography. Photographic sampling techniques have been previously used to measure C. racemosa cover (Piazzi et al., 2003). In this study, underwater photographs were taken to evaluate the persistence of Caulerpa in the plots by analyzing the area covered by the algae over time. This method allowed taking many samples in the field as diving time is limited. Each plot was divided on 4 sub-plots of 25 25 cm to improve the delineation of the algae on the photographic images. As algae can swing with wave orbital movements that could affect the area facing up on the images, three images of each sub-plot were taken on calm days. A scale was placed on each sub-plot next to the algae when taking the underwater photographs. Open software (ImageJ, NIH) was used to measure the area of Caulerpa on each photograph. To determine the absolute area on the images a relationship between image pixels and real dimensions ratio was established on each image using the scale. In order to test the method, photographs were taken in the laboratory and in the sea before the experiment began. A set of figures of known areas was used to test the output of the image processing analysis, providing an error around 6%.

2.4. Wave climate and modelling at the short-term experimental sites The propagation of the deep water mean wave climate between sampling dates was performed using a numerical model which

allows the calculation of the wave induced near-bottom orbital velocity at the study sites. Significant wave height (Hs), peak period (Tp) and direction (q) were acquired from July 2007 to March 2008 at deep water which corresponds to the duration of the short-term experiment (see Fig. 1bec for the location of these points). The deep water data have been obtained from the WANA system, a reanalysis with real data of a third generation spectral WAM model and was provided by the Spanish harbour authority operational system.

These deep water wave conditions were propagated to the study sites using a numerical model based on the mild slope parabolic approximation (Kirby and Dalrymple, 1983) and near-bottom orbital velocities (Ub) at the experimental locations were obtained from the model outputs using linear wave theory (Infantes et al., 2009). These models are widely used in engineering and scientific approaches in order to obtain the accurate wave field in coastal areas from open sea conditions since refraction and diffraction are well captured (Alvarez-Ellacuria et al., 2010).

2.5. Field wave measurements and velocity profiles To test the results of deep water wave propagations, field measurements were performed on the 8th of November of 2007 in Cala D’Or measuring wave conditions (Hs and Tp) at deep and shallow water as well as near-bottom velocities in shallow water on both substrata. An Acoustic Doppler Current Profiler (ADCP) RDI, Sentinel 600 kHz, was deployed at approximately 400 m from the experimental site at a depth of 26 m. The ADCP was set to record 20 min burst intervals every 60 min with a 2 Hz frequency. It was mounted on a tripod at 0.5 m above the bed. Hs obtained by propagating the deep water conditions to shallow waters, are within a w15% error to those measured in situ at deep and shallow waters, showing the accuracy of modelled wave conditions. Table 1 shows the data provided by the numerical model and those measured by the ADCP at deep water (26 m) and with the ADV at shallow water (7.5 m) on the 8th of November of 2007.

Additionally, vertical profiles of velocity were measured on the Posidonia oceanica dead matte model and on sand using Acoustic Doppler Velocimeters (ADV), Vector, Nortek. Simultaneous measurements on both substrata were performed using two ADVs at seven positions (2.5, 5, 10, 20, 40, 60 and 100 cm above the bottom).

The instruments were set to measure during 25 min with a frequency of 32 Hz, a nominal velocity range of 0.1 m/s and a sampling volume of 11.8 mm. The ADVs were mounted in a downlooking position on stainless steel bottom fixed structures. Stability of the equipment was verified with the compass, tilt and roll sensors.

Raw data were processed by first filtering for low beam correlation (<70%). Velocity data were then filtered with a low pass filter to remove high frequency Doppler noise (Lane et al., 1998). Horizontal velocity was calculated as the root mean square (rms). From the velocity profiles, friction coefficients Cf over sand and dead matte substrata were computed following Simarro et al. (2008).

Cf ¼

sb ruc 2

(1)

where sb is the bottom shear stress, r the density of seawater and uc a characteristic velocity outside the boundary layer. The bed shear stress can be found from, Table 1 Wave heights from model outputs and field measurements in Cala D’Or (means 95% CL).

Hs

Deep water (26 m) Shallow water (7.5 m)

Model outputs

Field measurements

0.53 0.02 m 0.26 0.02 m

0.47 þ 0.03 m 0.21 0.02 m

E. Infantes et al. / Estuarine, Coastal and Shelf Science 91 (2011) 434e441

du sb ¼ r nT dz z¼d

(2)

where nT is the eddy viscosity and du/dz the vertical variation of the velocity evaluated at the outside of the boundary layer d.

2.6. Water temperature and light at the short-term experimental sites Temperature HOBO StowAway TidbitÒ loggers were moored next to the plots in Cala D’Or and Sant Elm during the experiment.

Temperature was recorded every 2 h and averages of water temperature between sampling dates were calculated. The numbers of daylight hours were obtained from the US Naval Observatory (http://aa.usno.navy.mil/) for the study areas and averages of daylight hours between sampling dates were also calculated.

2.7. Statistical analysis A Chi-squared test (c2) was used to evaluate if the colonization of benthic communities by Caulerpa species depends on the type of substratum, that is, if some substrata are more favourable/unfavourable than others for the colonization of these species. The null hypothesis of the test assumes substratum-independent, random colonization and allows the calculation of expected cover of Caulerpa species in the different substrata that is compared with the cover observed. Changes in cover of Caulerpa species through time in the short-term experiments were evaluated using univariate repeated-measures ANOVA. “Substratum” (sand versus model of dead matte of Posidonia oceanica) was the between-subjects factors (fixed) and “Time” was the within-subjects factor (random) for both Caulerpa experiments. A significant Time Substratum interaction was to be expected if the type of substratum has an effect on the persistence of Caulerpa in the plots. A Cochran’s C-test was used to test for heterogeneity of variances.

3. Results Caulerpa taxifolia was more abundant in the dead matte of Posidonia oceanica than in the rest of substratum types available in Cala D’Or (Fig. 3a). Caulerpa racemosa was also more abundant in

437

the dead matte of P. oceanica but it additionally colonized other substrata such as rocks with photophilic algae and sand (Fig. 3b).

The percent cover of C. taxifolia over dead matte showed small changes between summer and winter (Fig. 3a). In contrast, the percent cover of C. racemosa was highly seasonal since large differences between summer and winter were observed (Fig. 3b).

The relative frequencies of the different types of substratum in the total surface of sea bottom surveyed and of the presence of Caulerpa on each substratum are shown in Fig. 4 for each sampling event. Different distribution of frequencies indicates that Caulerpa species do not distribute randomly in the substrata (Table 2).

Results from the c2 test show that both Caulerpa species are found in the dead matte of Posidonia oceanica and in rock on a higher proportion than expected in relation to the amount of substratum type available at the study sites. In contrast, Caulerpa species are found in a lower proportion than expected in sand and inside the P. oceanica meadow. Experimental c2 were largely above critical values (Table 2).

The significant “Time Substratum” interaction in both species (Table 3) indicates that substratum type affected the persistence of the two Caulerpa species. The temporal evolution of Caulerpa cover in the plots (Figs. 5a and 6a) shows that the persistence of Caulerpa was higher in the model of dead matte of Posidonia oceanica than in sand. Caulerpa taxifolia cover on dead matte increased until the 28th of September while C. taxifolia cover on sand did not change during the same period of time (Fig. 5a). After the 28th of September, C. taxifolia cover decreased with time in both substrata and by the end of winter no C. taxifolia was present in the plots. The cover of Caulerpa racemosa in the model of dead matte increased during the first five weeks of the experiment while on sand was roughly maintained (Fig. 6a). After August 30th, the C. racemosa experiment was disturbed by a storm that buried the plots with a 30e70 cm thick layer of drifting macroalgae and P. oceanica leaves causing the death of all C. racemosa fragments in the plots.

Seawater temperature in Cala D’Or (Caulerpa taxifolia site) changed from 26 C in July to 14 C in March, while the number of daylight hours was reduced from 14.5 h to 9 h (Fig. 5b). Conditions in Sant Elm (Caulerpa racemosa site) during the experiment were rather constant with a seawater temperature of 26 C and the number of daylight hours between 14 and 13 h (Fig. 6b).

Caulerpa taxifolia cover was maintained or increased with estimated near-bottom orbital velocities below 15 cm s 1 and declined above that value (Fig. 5b). Caulerpa racemosa was also able to maintain or increase its cover at the same velocity values (Fig. 6b).

The vertical profiles of velocity indicate that the width of the boundary layer is around 10 cm (Fig. 7) showing also a significantly higher reduction of the horizontal rms velocity in the model of dead matte of Posidonia oceanica (1.5 0.2 cm s 1, mean 95% CL) than in sand (2.2 0.2 cm s 1, mean 95% CL), (t-test: t ¼ 115, df ¼ 103598, p < 0.01).

The friction coefficient computed for the dead matte model was higher (1.04 E 01) than for sand (2.33 E 02). Increased friction by the dead matte model would reduce the magnitude of the velocities compared to sand. Wave data from the ADCP showed that wave conditions did not change during the ADV profile measurements, with Hs of 0.5 0.1 m, Tp of 9.5 0.5 s and South East direction.

4. Discussion

Fig. 3. Percent cover of (a) Caulerpa taxifolia and (b) Caulerpa racemosa in the different types of substratum during two consecutive years.

The results of this study indicate that substratum plays an important role in the distribution of these invasive Caulerpa species. Field measurements show that the presence of Caulerpa taxifolia and Caulerpa racemosa on the substrata is not at random.

438

E. Infantes et al. / Estuarine, Coastal and Shelf Science 91 (2011) 434e441

Fig. 4. Relative frequency of total substratum surveyed and of the presence of Caulerpa taxifolia and Caulerpa racemosa on each substratum. Low presence of rocks was recorded at the C. taxifolia site. Rocky bottom with photophilic algae is indicated as Rocks.

Dead matte of the seagrass Posidonia oceanica and rock covered with photophilic algae are more favourable than sand to the presence of Caulerpa species. Experimental results show that substratum has a significant effect on the persistence of both

Caulerpa species and confirm that is higher in the model of dead matte of P. oceanica than in sand. Hence, the results of both analyses are consistent to each other and suggest that the type of substratum influences the abundance of these invasive Caulerpa species.

Table 2 Chi-square (c2) test of observed Caulerpa cover on different substrata. Caulerpa taxifolia has a critical c2 p¼0.001***, 2df ¼ 13.81 and Caulerpa racemosa has a critical c2 p¼0.001***, 3df ¼ 16.26. C. taxifolia was not found on rock substratum in enough quantities for the test. The type of contribution was classified as more than expected (M) or less than expected (L).

Rock C. taxifolia

Winter 07 Summer 07 Winter 08 Summer 08

e e e e

C. racemosa

Winter 07 Summer 07 Winter 08 Summer 08

380 M 1M 146 M

P. oceanica 64 20 23 77

L L L L

3897 L 34 L 796 L

Sand 36 79 65 32

Dead matte P. oceanica

c2 Experimental

L L L L

240 366 121 69

M M M M

340*** 465*** 209*** 178***

7M 142 L 16 L 2L

7 1153 91 629

M M M M

20*** 2272*** 142*** 1573***

E. Infantes et al. / Estuarine, Coastal and Shelf Science 91 (2011) 434e441 Table 3 Results of repeated measures of ANOVA performed to evaluate if the cover of Caulerpa taxifolia and Caulerpa racemosa was different between a model of dead matte of Posidonia oceanica and a sandy substratum (Subst.). *p < 0.05, ***p < 0.001, ns ¼ not significant. Cochran tests were not significant. df

MS

F

p

Caulerpa taxifolia Intercept Subst.

Error Time Time Subst.

Residual

1 1 4 3 3 12

469997.4 17685.9 1600.7 5580.2 7086.1 1884.1

293.6 11.0

0.0000*** 0.0292 ns

2.9 3.7

0.0749 ns 0.0410*

Caulerpa racemosa Intercept Subst.

Error Time Time Subst.

Residual

1 1 4 3 3 12

41205.3 8321.6 1128.9 1790.3 1929.1 390.2

36.5 7.4

0.0037*** 0.0532 ns

4.6 4.9

0.0231* 0.0184*

Understanding the factors that regulate the establishment and spread of introduced Caulerpa species is important to predict future pathways of invasion as well as the susceptible areas to be invaded (Carlton and Geller, 1993; Bax et al., 2003; Cebrian and Ballesteros, 2009). Previous studies mainly based on direct observation indicated that favourable substrata for the establishment of Caulerpa spp. are dead matte of Posidonia oceanica (Piazzi and Cinelli, 1999; Piazzi et al., 2001; Ruitton et al., 2005) and rocky bottoms covered by turf algae (Ceccherelli et al., 2002; Piazzi et al., 2003; Bulleri and Benedetti-Cecchi, 2008). A propagation model of Caulerpa taxifolia assumed that the settlement probabilities of this species were high for dead matte, harbour mud, and rocks with photophilic algae and low for P. oceanica and unstable sand (Hill et al., 1998) but no explanation of how these probabilities were set is given. Our study provides the first quantitative evaluation of previous descriptions of dead matte of P. oceanica as being a favourable substratum for the colonization of invasive Caulerpa species. It also provides quantitative evidence that rocks covered with photophilic algae are also a favourable substratum for Caulerpa racemosa colonization. Sand and P. oceanica meadows are less favourable than the dead matte of P. oceanica and rocks covered with photophilic algae for the colonization of C. taxifolia and C. racemosa.

439

The positive association between consolidated substrata (rocks and dead matte) and the presence of invasive Caulerpa species suggests that near-bottom orbital velocity may be an important determinant of the spatial distribution of these species at shallow depths. The capacity of Caulerpa to acquire nutrients from the sediment through the rhizoids (Williams, 1984; Chisholm et al., 1996) might be invoked to explain the preferential distribution of invasive Caulerpa on dead matte of Posidonia oceanica but not over rocky substrata where the amount of sediment is much lower than in dead matte. The structural complexity of the substratum facilitates the retention of Caulerpa taxifolia fragments (Davis et al., 2009) and could be a factor to explain why rocky substrata covered by an erect layer of photophilous macroalgae and the P. oceanica dead matte are favourable substrata for the establishment of Caulerpa. The calculated friction coefficients for the dead matte model and sand indicate that the surface of the dead matte is structurally more complex than sand and that Caulerpa fragments established on it will experience lower flow velocities than in sand.

Caulerpa taxifolia and Caulerpa racemosa were rarely observed inside the Posidonia oceanica meadow in this study and the c2 test shows that P. oceanica is not a favourable substratum for the establishment of the introduced Caulerpa species even though the matte of the P. oceanica meadow is a favourable substratum.

Previous studies have also described that dense P. oceanica meadows are not invaded by C. taxifolia and C. racemosa in the short-term (Ceccherelli and Cinelli, 1999; Ceccherelli et al., 2000).

Dense P. oceanica meadows appear to resist the invasion of C. taxifolia for more than eight years (Jaubert et al., 1999). Our results suggest that degraded, low-density P. oceanica meadows would be suitable locations for the settlement and growth of invasive Caulerpa species if shading by the leaf canopy was the main constraint of Caulerpa invasion (Ceccherelli and Cinelli, 1999; Ceccherelli et al., 2000), and thus protecting existing meadows would likely reduce the spread of these species in the Mediterranean Sea. There has been little effort to correlate wave and current energy with the spatial distribution and growth of invasive algae. The effects of wave exposure, morphology and drag forces of the invasive Codium fragile were studied under field and flume conditions by D’Amours and Scheibling (2007). These authors suggest that the C. fragile became dislodged at flow velocities exceeding 50 cm s 1. Moreover, Scheibling and Melady (2008) found that a sheltered location was a more suitable habitat than an exposed location for this species. The effect of wave exposure on Caulerpa

Fig. 5. (a) Caulerpa taxifolia cover (normalised means SE, n ¼ 3). (b) Water temperature at the experimental site, hours of light per day and near-bottom orbital velocities (Ub) (means SD). Cover area on each sampling date (A) in cm2 normalised by initial cover area (A0). Dotted square indicates the period used for the statistical analysis ANOVA, Table 3.

440

E. Infantes et al. / Estuarine, Coastal and Shelf Science 91 (2011) 434e441

Fig. 6. (a) Caulerpa racemosa cover (normalised means SE, n ¼ 3). (b) Water temperature at the experimental site, hours of light per day and near-bottom orbital velocities (Ub) (means SD). Cover area on each sampling date (A) in cm2 normalised by initial cover area (A0).

racemosa its not clear since it can be found in both exposed and sheltered locations (Thibaut et al., 2004; Klein and Verlaque, 2008).

To date, there are no available data on near-bottom orbital velocities that allow the establishment and development of invasive Caulerpa species. Our results show that Caulerpa taxifolia and C. racemosa do only occur at near-bottom orbital velocities below 15 cm s 1 and that C. taxifolia cover could decline at velocities above that value. This value, however, should not be considered as

a velocity threshold for the persistence of C. taxifolia because the evidence provided by our study is correlative only and other factors such as seawater temperature and light availability declined as velocities increased in the experimental site (see Figs.

5b and 6b).

We observed that Caulerpa racemosa was able to colonize sand in summer but disappeared in winter, a season with higher wave energy. One of the potential problems for aquatic macrophytes growing on unconsolidated substrata is sediment movement and deposition that causes uprooting and burial (Fonseca and Kenworthy, 1987; Williams, 1988; Cabaço et al., 2008). Previous works described that C. racemosa was not colonizing fine sand with large ripple marks (Ruitton et al., 2005). Sand with large ripples indicates that orbital velocities are or have been high enough to mobilize the sediment suggesting that only those macrophytes with high anchoring capacity could survive in that location.

Our results have provided novel quantitative evidence highlighting the importance of substratum for the distribution of invasive Caulerpa species at shallow depths. We have shown that the two introduced species of Caulerpa in the Mediterranean Sea are more abundant in dead matte of Posidonia oceanica or rock than inside the P. oceanica meadow or sand. Our results also suggest the quantitative hydrodynamic conditions that shape colonization as their cover started to decline at near-bottom orbital velocities over 15 cm s 1. Further studies should be done to elucidate the validity of this velocity threshold of the colonization of introduced Caulerpa.

Acknowledgments

Fig. 7. Vertical profiles of velocity (means 95% CL). For convenience only the 40 cm above the bottom is shown.

E. Infantes would like to thank financial support from the Spanish Ministerio de Educación y Ciencia (MEC) FPI scholarship program (BES-2006-12850). Research funds were provided by MEC grants CTM2005-01434/MAR, CTM2006-12072/MAR and by the Govern of the Balearic Islands, project Integrated Coastal Zone Management (UGIZC). The authors thank Puertos del Estado for WANA wave climate data and the Marina of Cala D’Or (Santanyi) which kindly made available its harbour facilities for executing this study. We thank Drs. G. Zarruk and G. Simarro for their advice in the velocimeter data. Thanks to B. Casas, F. Medina-Pons, I. Castejón and T. Box for field assistance.

E. Infantes et al. / Estuarine, Coastal and Shelf Science 91 (2011) 434e441

References Alvarez-Ellacuria, A., Orfila, A., Olabarrieta, M., Medina, R., Vizoso, G., Tintore, J., 2010. A nearshore wave and current operational forecasting system. Journal of Coastal Research 26, 503e509.

Bax, N., Williamson, A., Aguero, M., Gonzalez, E., Geeves, W., 2003. Marine invasive alien species: a threat to global biodiversity. Marine Policy 27, 313e323.

Bold, H.C., Wynne, M.J., 1978. Introduction to the Algae. Structure and Reproduction.

Prentice-Hall, Inc., Englewood Cliffs.

Boudouresque, C.F., Verlaque, M., 2002. Biological pollution in the Mediterranean Sea: invasive versus introduced macrophytes. Marine Pollution Bulletin 44, 32e38.

Bulleri, F., Benedetti-Cecchi, L., 2008. Facilitation of the introduced green alga Caulerpa racemosa by resident algal turfs: experimental evaluation of underlying mechanisms. Marine Ecology Progress Series 364, 77e86.

Cabaço, S., Santos, R., Duarte, C.M., 2008. The impact of sediment burial and erosion on seagrasses: a review. Estuarine, Coastal and Shelf Science 79, 354e366.

Carlton, J.T., Geller, J.B., 1993. Ecological roulette - the global transport of nonindigenous marine organisms. Science 261, 78e82.

Cebrian, E., Ballesteros, E., 2009. Temporal and spatial variability in shallow- and deep-water populations of the invasive Caulerpa racemosa var. cylindracea in the Western Mediterranean. Estuarine, Coastal and Shelf Science 83, 469e474.

Ceccherelli, G., Cinelli, F., 1999. Effects of Posidonia oceanica canopy on Caulerpa taxifolia size in a north-western Mediterranean bay. Journal of Experimental Marine Biology and Ecology 240, 19e36.

Ceccherelli, G., Piazzi, L., Cinelli, F., 2000. Response of the non-indigenous Caulerpa racemosa (Forssakal) J. Agardh to the native seagrass Posidonia oceanica (L.) Delile: effect of density of shoots and orientation of edges of meadows. Journal of Experimental Marine Biology and Ecology 243, 227e240.

Ceccherelli, G., Piazzi, L., Balata, D., 2002. Spread of introduced Caulerpa species in macroalgal habitats. Journal of Experimental Marine Biology and Ecology 280,1e11.

Chisholm, J.R.M., Moulin, P., 2003. Stimulation of nitrogen fixation in refractory organic sediments by Caulerpa taxifolia (Chlorophyta). Limnology and Oceanography 48, 787e794.

Chisholm, J.R.M., Dauga, C., Ageron, E., Grimont, P.A.D., Jaubert, J.M., 1996. Roots in mixotrophic algae. Nature 381, 382.

D’Amours, O., Scheibling, R.E., 2007. Effect of wave exposure on morphology, attachment strength and survival of the invasive green alga Codium fragile ssp. tomentosoides. Journal of Experimental Marine Biology and Ecology 351, 129e142.

Davis, A.R., Ferguson, A.M., Wright, J.T., 2009. Structural complexity facilitates accumulation and retention of fragments of the invasive alga, Caulerpa taxifolia.

Journal of Experimental Marine Biology and Ecology 371, 163e169.

Fonseca, M.S., Bell, S.S., 1998. Influence of physical settings on seagrass landscapes near Beauford, North Carolina. USA. Marine Ecology Progress Series 171, 109e121.

Fonseca, M.S., Kenworthy, W.J., 1987. Effects of current on photosynthesis and distribution of seagrasses. Aquatic Botany 27, 59e78.

Fonseca, M.S., Zieman, J.C., Thayer, G.W., Fisher, J.S., 1983. The role of current velocity in structuring eelgrass (Zostera marina L) meadows. Estuarine, Coastal and Shelf Science 17, 367e380.

Frederiksen, M., Krause-Jensen, D., Holmer, M., Laursen, J.S., 2004. Spatial and temporal variation in eelgrass (Zostera marina) landscapes: influence of physical setting. Aquatic Botany 78, 147e165.

Glasby, T.M., Gibson, P.T., Kay, S., 2005. Tolerance of the invasive marine alga Caulerpa taxifolia to burial by sediment. Aquatic Botany 82, 71e81.

Gobert, S., Kyramarios, M., Lepoint, G., Pergent-Martini, C., Bouquegneau, J.M., 2003.

Variations at different spatial scales of Posidonia oceanica (L.) Delile beds; effects on the physico-chemical parameters of the sediment. Oceanologica Acta 26, 199e207.

Hill, D., Coquillard, P., Vaugelas, J., Meinesz, A., 1998. An algorithmic model for invasive species: application to Caulerpa taxifolia (Vahl) C. Agardh development in the North-Western Mediterranean Sea. Ecological Modelling 109, 251e265.

Infantes, E., Terrados, J., Orfila, A., Cañellas, B., Alvarez-Ellacuria, A., 2009. Wave energy and the upper depth limit distribution of Posidonia oceanica. Botanica Marina 52, 419e427.

Jaubert, J.M., Chisholm, J.R.M., Ducrot, D., Ripley, H.T., Roy, L., Passeron-Seitre, G., 1999.

No deleterious alterations in Posidonia beds in the bay of menton (France) eight years after Caulerpa taxifolia colonization. Journal of Phycology 35, 1113e1119.

Kirby, J.T., Dalrymple, R.A., 1983. A parabolic equation for the combined refraction diffraction of stokes waves by mildly varying topography. Journal of Fluid Mechanics 136, 453e466.

441

Klein, J., Verlaque, M., 2008. The Caulerpa racemosa invasion: a critical review.

Marine Pollution Bulletin 56, 205e225.

Koch, E.M., 2001. Beyond light: physical, geological, and geochemical parameters as possible submersed aquatic vegetation habitat requirements. Estuaries 24, 1e17.

Koch, E.W., Ackerman, J.D., Verduin, J., van Keulen, M., 2006. Fluid dynamics in seagrass ecology - from molecules to ecosystems. In: Larkum, A.W.D., Orth, R.J., Duarte, C.M. (Eds.), Seagrasses: Biology, Ecology and Conservation. Springer, Netherlands, pp. 193e224.

Lane, S.N., Biron, P.M., Bradbrook, K.F., Butler, J.B., Chandler, J.H., Crowell, M.D., McLelland, S.J., Richards, K.S., Roy, A.G., 1998. Three-dimensional measurement of river channel flow processes using acoustic Doppler velocimetry. Earth Surface Processes and Landforms 23, 1247e1267.

Marbà, N., Duarte, C.M., Holmer, M., Martinez, R., Basterretxea, G., Orfila, A., Jordi, A., Tintore, J., 2002. Effectiveness of protection of seagrass (Posidonia oceanica) populations in Cabrera National Park (Spain). Environmental Conservation 29, 509e518.

Mateo, M.A., Romero, J., Perez, M., Littler, M.M., Littler, D.S., 1997. Dynamics of millenary organic deposits resulting from the growth of the Mediterranean seagrass Posidonia oceanica. Estuarine, Coastal and Shelf Science 44, 103e110.

Meinesz, A., Hesse, B., 1991. Introduction of the tropical alga Caulerpa taxifolia and its invasion of the Northwestern Mediterranean. Oceanologica Acta 14, 415e426.

Piazzi, L., Cinelli, F., 1999. Developpement et dynamique saisonniere d’un peuplement mediterraneen de l’algue tropicale Caulerpa racemosa (Forsskal) J. Agardh.

Cryptogamie Algol 20, 295e300 (in French, with English Abstract).

Piazzi, L., Ceccherelli, G., Cinelli, F., 2001. Threat to macroalgal diversity: effects of the introduced green alga Caulerpa racemosa in the Mediterranean. Marine Ecology Progress Series 210, 149e159.

Piazzi, L., Ceccherelli, G., Balata, D., Cinelli, F., 2003. Early patterns of Caulerpa racemosa recovery in the Mediterranean Sea: the influence of algal turfs. Journal of the Marine Biological Association of the United Kingdom 83, 27e29.

Piazzi, L., Balata, D., Ceccherelli, G., Cinelli, F., 2005. Interactive effect of sedimentation and Caulerpa racemosa var. cylindracea invasion on macroalgal assemblages in the Mediterranean Sea. Estuarine, Coastal and Shelf Science 64, 467e474.

Riis, T., Hawes, I., 2003. Effect of wave exposure on vegetation abundance, richness and depth distribution of shallow water plants in a New Zealand lake. Freshwater Biology 48, 75e87.

Ruitton, S., Javel, F., Culioli, J.M., Meinesz, A., Pergent, G., Verlaque, M., 2005. First assessment of the Caulerpa racemosa (Caulerpales, Chlorophyta) invasion along the French Mediterranean coast. Marine Pollution Bulletin 50, 1061e1068.

Scheibling, R.E., Melady, R.A., 2008. Effect of water movement and substratum type on vegetative recruitment of the invasive green alga Codium fragile ssp tomentosoides. Botanica Marina 51, 341e349.

Shepherd, S.A., Womersley, H.B.S., 1981. The algal and seagrass ecology of Waterloo bay, South-Australia. Aquatic Botany 11, 305e371.

Simarro, G., Orfila, A., Liu, P.L.F., 2008. Bed-shear stress in turbulent wave-current boundary layers. Journal of Hydraulic Engineering-Asce 134, 225e230.

Taylor, W.R., 1960. Marine Algae of the Eastern Tropical and Subtropical Coasts of the Americas. University of Michigan Press, Ann Arbor.

Terrados, J., 1997. Is light involved in the vertical growth response of seagrasses when buried by sand? Marine Ecology Progress Series 152, 295e299.

Terrados, J., Medina-Pons, F.J.M., 2008. Epiphyte load on the seagrass Posidonia oceanica (L.) Delile does not indicate anthropogenic nutrient loading in Cabrera Archipelago National Park (Balearic Islands, western Mediterranean). Scientia Marina 72, 503e510.

Thibaut, T., Meinesz, A., Coquillard, P., 2004. Biomass seasonality of Caulerpa taxifolia in the Mediterranean sea. Aquatic Botany 80, 291e297.

US Naval Observatory., 2008. Astronomical Applications Department (accessed 25.09.2008). http://aa.usno.navy.mil/. van Katwijk, M.M., Hermus, D.C.R., 2000. Effects of water dynamics on Zostera marina: transplantation experiments in the intertidal Dutch Wadden Sea. Marine Ecology Progress Series 208, 107e118.

Williams, S.L., 1984. Uptake of sediment ammonium and translocation in a marine green macroalga Caulerpa cupressoides. Limnology and Oceanography 29, 374e379.

Williams, S.L., 1988. Disturbance and recovery of a deep-water Caribbean seagrass bed. Marine Ecology Progress Series 42, 63e71.

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 115, C12005, doi:10.1029/2010JC006345, 2010

Wave‐induced velocities inside a model seagrass bed Mitul Luhar,1 Sylvain Coutu,2 Eduardo Infantes,3 Samantha Fox,1 and Heidi Nepf1 Received 15 April 2010; revised 8 July 2010; accepted 13 August 2010; published 2 December 2010.

[1] Laboratory measurements reveal the flow structure within and above a model seagrass meadow (dynamically similar to Zostera marina) forced by progressive waves.

Despite being driven by purely oscillatory flow, a mean current in the direction of wave propagation is generated within the meadow. This mean current is forced by a nonzero wave stress, similar to the streaming observed in wave boundary layers. The measured mean current is roughly four times that predicted by laminar boundary layer theory, with magnitudes as high as 38% of the near‐bed orbital velocity. A simple theoretical model is developed to predict the magnitude of this mean current based on the energy dissipated within the meadow. Unlike unidirectional flow, which can be significantly damped within a meadow, the in‐canopy orbital velocity is not significantly damped. Consistent with previous studies, the reduction of in‐canopy velocity is a function of the ratio of orbital excursion and blade spacing.

Citation: Luhar, M., S. Coutu, E. Infantes, S. Fox, and H. Nepf (2010), Wave‐induced velocities inside a model seagrass bed, J. Geophys. Res., 115, C12005, doi:10.1029/2010JC006345.

1. Introduction [2] Seagrasses, which occupy 10% of shallow coastal areas [Green and Short, 2003], are essential primary producers, forming the foundation for many food webs. Seagrass beds also damp waves, stabilize the seabed, shelter economically important fish and shellfish, and enhance local water quality by filtering nutrients from the water. On the basis of nutrient cycling services alone, the global economic value of seagrass beds was estimated to be 3.8 trillion dollars per year by Costanza et al. [1997]. Furthermore, numerous studies note higher infaunal density within seagrass beds [e.g., Santos and Simon, 1974; Stoner, 1980; Irlandi and Peterson, 1991]. Peterson et al. [1984] found that clams in seagrass beds had higher growth rates and also that the density of the bivalve, Mercenaria mercaneria, was five time greater in seagrass meadows than on adjacent sand beds. By damping near‐bed water velocities, seagrasses reduce local resuspension and promote the retention of sediment [e.g., Fonseca and Cahalan, 1992; Gacia et al., 1999; Granata et al., 2001], thereby stabilizing the seabed. Reduced resuspension improves water clarity, leading to greater light penetration and increased productivity [Ward et al., 1984].

Seagrasses are also a source of drag. Hence, waves propagating over seagrass beds lose energy [Fonseca and Cahalan, 1992; Chen et al., 2007; Bradley and Houser, 2009].

1 Department of Civil and Environmental Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts, USA. 2 Institut d’Ingénieurie de l’Environnement, Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland.

3 Instituto Mediterráneo de Estudios Avanzados, IMEDEA (CSIC‐UIB), Esporles, Spain.

Copyright 2010 by the American Geophysical Union.

0148‐0227/10/2010JC006345

[3] Some of the ecosystem services mentioned above (stabilizing the seabed, wave decay, shelter for fish and shellfish) arise because seagrasses are able to alter the local flow conditions. The nutrient cycling capability of seagrasses is limited both by the rate of water renewal within the bed and the rate at which the seagrasses are able to extract nutrients from the surrounding water, which, under some conditions, is limited by the diffusive boundary layer on the blades. Clearly, the hydrodynamic condition plays a major role in determining both the health of seagrass beds and their ecologic contribution. Previous studies have successfully described the flow structure for submerged vegetation subjected to unidirectional flow (currents) using rigid and flexible vegetation models [e.g., Finnigan, 2000; Nepf and Vivoni, 2000; Ghisalberti and Nepf, 2002, 2004, 2006].

A summary of unidirectional flow over submerged canopies can be found in the study by Luhar et al. [2008].

[4] For the case of oscillatory flow (waves), previous work has focused primarily on quantifying wave decay [e.g., Fonseca and Cahalan, 1992; Kobayashi et al., 1993; Méndez et al., 1999; Mendez and Losada, 2004; Bradley and Houser, 2009]. Other studies involving oscillatory flow include those by Thomas and Cornelisen [2003], who showed that nutrient uptake in seagrass beds was higher for wave conditions, and Koch and Gust [1999], who suggested that the periodic motion of seagrass blades could lead to enhanced mass transfer between the meadow and the overlying water column. Despite these insights, researchers have only recently begun to study the detailed hydrodynamics of submerged vegetation subjected to wave‐driven oscillatory flow.

Lowe et al. [2005a] studied the flow structure within a model canopy comprising rigid vertical cylinders and developed an analytical model to predict the magnitude of in‐canopy velocity in the presence of waves. They used a simple friction coefficient to characterize the shear stress at the top of the

C12005

1 of 15

C12005

LUHAR ET AL.: WAVE‐INDUCED FLOW INSIDE A SEAGRASS BED

C12005

Figure 1. Schematic of experimental setup. The bold dashed line indicates measurement locations for the vertical profile. Not to scale. canopy and drag and inertia coefficients to parameterize the hydrodynamic impact of the canopy elements.

[5] The present laboratory study investigates the flow structure within a model seagrass bed subject to propagating waves. The model is constructed with flexible blades that are dynamically similar to real seagrass blades [see Ghisalberti and Nepf, 2002]. Our experiments reveal that a unidirectional current is generated within the model seagrass bed when it is forced by purely oscillatory, wave‐driven flow. This is, in some ways, analogous to viscous [Longuet‐ Higgins, 1953] and turbulent [e.g., Longuet‐Higgins, 1958; Johns, 1970; Trowbridge and Madsen, 1984; Davies and Villaret, 1998, 1999; Marin, 2004] streaming observed in wave boundary layers. The induced current could speed up the rate of water renewal within a meadow, enhancing the nutrient cycling capabilities of the seagrass. A directional bias in the dispersal of seeds and pollen could affect seagrass meadow structure. The induced current also has the potential to influence the net transport of sediment. Finally, the hydrodynamic drag exerted by the model vegetation leads to a reduction of in‐canopy orbital velocities. We observed that the ratio of in‐canopy to over‐canopy velocity is significantly higher when the flow is oscillatory (tested here) compared to the unidirectional case tested by Ghisalberti and Nepf [2006]. This is in agreement with Lowe et al. [2005a]. As noted by Lowe et al. [2005b], larger in‐canopy velocities could explain the higher nutrient uptake observed under oscillatory flows [Thomas and Cornelisen, 2003].

and wave‐induced dynamic pressure ( pw), ! cosh kz cosðkx !t Þ: pw ¼ Uw ¼ ga k cosh kh

ð2Þ

Uw ¼ a!

cosh kz cosðkx !tÞ; sinh kh

ð1aÞ

In the equations above, r is the fluid density, g is the gravitational acceleration, a is the wave amplitude, w is the wave radian frequency, k is the wave number, h is the water depth, x and z are the horizontal and vertical coordinates (z = 0 at the bed), and t is time. The water depth h refers to the distance from the bed to the mean water level (Figure 1).

The frequency, wave number, and water depth are related by the dispersion relation, w2 = (kg)tanh(kh). Throughout this paper, the subscript w refers to purely oscillatory flows (i.e., time average of zero). When we refer specifically to unidirectional flows (currents), the subscript c is used. Turbulent, fluctuating velocities are represented by lowercase letters with prime symbols (u′, w′).

[7] In addition to neglecting the nonlinear terms in the Navier‐Stokes equations, linear wave theory assumes perfectly inviscid, irrotational motion. Under these assumptions, the horizontal and vertical velocities are exactly 90° out of phase with each other, as evidenced by equations (1a) and (1b). However, this solution does not satisfy the no‐slip boundary condition at the bed. While the inviscid assumption is valid for most of the water column, viscosity is important in the bottom boundary layer, which for laminar flows is of thickness O[(n/w)1/2]. Here, n is the kinematic viscosity of water. The horizontal oscillatory velocity decays from the inviscid value (1a) at the outer edge of the boundary layer to zero at the bed because of viscosity. This modification to the inviscid solution causes a phase shift in the oscillatory velocities. The horizontal and vertical velocities are no longer exactly 90° out of phase, creating a steady, nonzero wave stress, hUwWwi ≠ 0 (h i denotes a time average). This wave stress is analogous to turbulent Reynolds stress. It represents a time‐invariant momentum transfer out of the oscillatory flow and generates a mean current in the boundary layer. For laminar flows, the magnitude of this current, Uc, at the outer edge of the boundary layer is [Longuet‐Higgins, 1953]

Ww ¼ a!

sinh kz sinðkx !tÞ; sinh kh

ð1bÞ

3 a!

Uc ¼ ðkaÞ : 4 sinh2 kh

2. Theory 2.1. Linear Wave Theory and Boundary Layer Streaming [6] In the absence of a canopy, linear wave theory [e.g., Mei et al., 2005] leads to the following solutions for the horizontal (Uw) and vertical (Ww) oscillatory velocity fields for waves propagating over a flat bed,

2 of 15

ð3Þ

C12005

LUHAR ET AL.: WAVE‐INDUCED FLOW INSIDE A SEAGRASS BED

The forces exerted by our model seagrass canopy also lead to a phase shift between the oscillatory velocities, resulting in a nonzero wave stress. Below, we present an overview of the forces exerted by the canopy on the wave flow [based on Lowe et al., 2005a] followed by a new analysis estimating the mean flow generated within the canopy.

2.2. Canopy Forces [8] Lowe et al. [2005a] described the water motion within a rigid canopy (a model coral reef) relative to the undisturbed flow above the canopy. Here, we consider the application of their model to a flexible model canopy (a model seagrass meadow). The geometry of the canopy is described by two dimensionless parameters, the frontal area per bed area, lf = av hv, and the planar area per bed area, lp.

Here, av is the frontal area per unit volume, and hv is the vegetation height. Because of the forces exerted by the ^ m, is vegetation, the velocity scale within the meadow, U reduced relative to that above the meadow, U∞. The velocity ^ m, represents a vertical average scale inside the canopy, U over the canopy height (denoted by the over‐hat symbol). ^ m /U∞, depends on the relative [9] The velocity ratio, a = U importance of the shear stress at the top of the meadow (ru2*hv), ^ m∣U ^ m/ the drag force exerted by the meadow ((1/2)rCD av∣U (1 − lp)), and the inertial forces including added mass ((Cmlp / ^ m /∂t), with Cm the inertial force coefficient. These (1 − lp))∂U three forces are characterized by the following length scales, respectively, the shear length scale, LS ¼ hv

U1 u hv *

!2 ¼

2hv ; Cf

ð4Þ

where Cf = 2(u*hv /U∞)2 is the meadow friction factor, the drag length scale, LD ¼

2hv 1 p ; CD f

ð5Þ

and the oscillation length scale, which is simply the wave orbital excursion, A∞ = U∞ /w above the meadow.

[10] Conceptually, the drag and shear length scales describe the scale at which the effects of drag and shear begin to influence fluid motion. With these forces, the governing equation becomes [Lowe et al., 2005a]

^ m U1 ^ m jU ^m ^m @ U Cm p @ U jU1 jU1 jU :

¼ LS LD 1 p @t @t

ð6Þ

^ m = Re{bA∞weiwt} By introducing the complex variables U and U∞ = Re{A∞ weiwt}, and considering only the first Fourier harmonic, we suggest simplifying equation (6), after some straightforward algebra, to ið 1Þ ¼

8 A1 8 A1 Cm p :

j j i 1 p 3 LS 3 LD

ð7Þ

To obtain this result, we assume that A∞ is real and positive while b may be complex. The ratio of in‐canopy velocity to the velocity above the canopy is simply a = ∣b∣.

[11] From equation (7), we see that if the wave excursion is smaller than the drag and shear length scales (A∞ LS

C12005

and LD), the wave motion is unaffected by the drag and shear stress, and the flow is dominated by inertia. At this limit, the velocity ratio is given by the following, with subscript i used to emphasize inertia‐dominated conditions [Lowe et al., 2005a], i ¼

1 p : 1 þ ðCm 1Þ p

ð8Þ

At the other limit of flow behavior, when the wave excursion, A∞, is much longer than LS and LD, the flow resembles a current. At this limit, the inertial forces drop out, as the acceleration term is negligibly small. Flow within the meadow is determined by the balance of shear and drag forces. Using subscript c to denote the current‐only limit, as in the study by Lowe et al. [2005a], we have the following velocity ratio: c ¼

rffiffiffiffiffiffi LD : LS

ð9Þ

For the intermediate case, where the effects of both drag and inertia are important, equation (7) must be solved iteratively to yield a = ∣b∣. Lowe et al. [2005a] solved equation (6) numerically by providing an initial condition and marching forward in time until a quasi‐steady state is achieved.

Alternatively, we propose the use of the Fourier decomposition shown in equation (7), which yields identical results for the inertia and current‐only limits and can be more easily solved for the general case.

[12] Equation (6) assumes that the drag‐generating elements are rigid, which is not the case with our canopy modeled on flexible seagrass. However, incorporating the impact of wave‐induced blade movement in a predictive model is extremely difficult and requires the development of a coupled fluid‐structure interaction model. On the basis of observations of blade motion, we argue in a companion study (M. Luhar et al., Seagrass blade motion under waves and its impact on wave decay, submitted to Marine and Ecology Progress Series, 2010) that a rigid vegetation model is appropriate if an effective rigid canopy height is used. The effective height is defined as the vertical extent of the meadow over which the blades do not move significantly relative to the water. For the wave conditions considered by Luhar et al. (submitted manuscript, 2010), the effective height was typically less than half the blade length.

However, in unidirectional flows, there is some relative motion between the blades and the water along the entire blade length [e.g., Ghisalberti and Nepf, 2006]. Furthermore, in a recent field study measuring wave decay over seagrass beds, Bradley and Houser [2009] observed relative motion along the whole blade in low‐energy, broadband wave conditions. For generality, therefore, we will assume that the effective height is equal to the blade length.

2.3. Wave‐Induced Current [13] We propose that the forces generated within the model meadow lead to a nonzero mean (time‐invariant) wave stress at the top of the canopy, which drives a mean current through the meadow. The magnitude of this wave stress and the mean flow generated by it can be estimated based on the arguments shown by Fredsøe and Deigaard

3 of 15

C12005

C12005

LUHAR ET AL.: WAVE‐INDUCED FLOW INSIDE A SEAGRASS BED

[1992] for wave boundary layers. Within the meadow, the horizontal wave velocity, Uw,m, deviates from the linear wave theory solution, Uw (equation (1a)), because of the forces exerted by the vegetation. From continuity, this deviation leads to the generation of a vertical velocity, Ww,m, in addition to that predicted by linear wave theory. This additional velocity at the top of the meadow (z = hv) can be expressed as Ww; m ð z ¼ hv Þ ¼

@ @x

Zhv 0

k Uw; m Uw dz ¼ !

Zhv 0

@ Uw; m Uw dz; @t

phase with the flow acceleration, leading to a zero time average when multiplied by the velocity. Furthermore, for typical values of Cf and CDav hv, the energetic contribution of the work done by the shear stress, which is of O[rCf Uw3], is negligible compared to the total energy dissipation, which is of O[rCDav hvUm3 ]. Specifically, for the wave conditions tested here, the magnitude of the in‐canopy velocity, Um, is comparable to the outer flow velocity, Uw (Tables 1 and 2), and Cf is an O[0.01] constant while the parameter CDav hv is O[1]. Under these assumptions, the time‐averaged wave stress at the top of the canopy is

ð10Þ

where the relation between the spatial and temporal derivatives for propagating waves, ∂/∂x = −(w/k)∂/∂t, is used.

Note that the integral on the right‐hand side of equation (10) resembles the vertically averaged momentum balance for the meadow shown in equation (6) when multiplied by (1/hv).

Above the meadow, the horizontal oscillatory velocity is described by linear wave theory (Uw; equation (1a)), but the vertical velocity field includes both the velocity predicted by linear theory (Ww; equation (1b)) and the additional vertical velocity, Ww,m, shown in equation (10). Unlike the linear wave solution, Ww,m is not perfectly out of phase with the horizontal oscillatory velocity, leading to a nonzero time‐ averaged wave stress, hUwWw,mi, acting at the top of the meadow.

[14] To estimate the magnitude of this wave stress, we consider the energy balance for the meadow. Wave energy is transferred from the outer flow into the meadow via the work done by the wave‐induced pressure at the top of the meadow, −h pw(Ww + Ww,m)i, and the work done by the shear stress at the interface, ht wUwi. The energy transfer is balanced by dissipation within the meadow, hEDi:

h pw Ww þ Ww;m i þ h w Uw i ¼ hED i:

ð11Þ

Note that hEDi includes dissipation attributed to the forces exerted by the vegetation, dissipation caused by bed stress, and shear‐induced viscous dissipation. Above the meadow, we assume the horizontal oscillatory velocity and pressure fields are specified by the linear wave solution; hence, pw = r(w/k)Uw as shown in equation (2). Then, equation (11) may be rearranged to yield the time‐averaged wave stress at the top of the canopy: hUw Ww;m i ¼

k ðh w Uw i hED iÞ: !

ð12Þ

Assuming that energy dissipation is dominated by the drag force exerted by the vegetation, fD, i.e., excluding bed friction and viscous dissipation [see also Mendez and Losada, 2004; Lowe et al., 2007; Bradley and Houser, 2009; Luhar et al., submitted manuscript, 2010], *Zhv hED i ¼

hUw Ww;m i ¼

fD Um dz ;

ð13Þ

where Um is the velocity inside the meadow. We ignore the contribution of the inertia force since this tends to be in

*Zhv

+ fD Um dz :

ð14Þ

0

Integrating the momentum equation over the height of the meadow and time averaging leads to the following physically intuitive mean momentum balance [e.g., Fredsøe and Deigaard, 1992]: hUw Ww;m i þ h w i h b i ¼ hFD i;

ð15Þ

where, ht wi and ht bi are the mean shear stresses at the top of the canopy and at the bed, and *Zhv hFD i ¼

+ fD dz ;

ð16Þ

0

is the time‐averaged drag force integrated over the height of the canopy. For simplicity, the ∂/∂x convective acceleration term, caused by slow wave decay in the x‐direction, and the mean pressure gradient have been assumed negligible.

Assuming that the mean shear stresses are negligible compared to the vegetation drag, equations (14) to (16) can be combined to yield k !

*Zhv

+ fD Um dz

0

*Zhv ¼

+ fD dz :

ð17Þ

0

Recognizing that the velocity inside the canopy, Um, consists of both an oscillatory (Uw, m) and a mean flow (Uc, m) component, the drag force using a standard quadratic law is fD = (1/2)rCD av∣Uw, m + Uc, m∣(Uw, m + Uc, m). However, on the basis of experimental results [Sarpkaya and Isaacson, 1981] and numerical simulations [Zhou and Graham, 2000], the use of a two‐term formulation for the drag force is more appropriate in combined wave‐current flows. Following Zhou and Graham [2000], we decompose the drag force into its steady and time‐varying components with separate drag coefficients, CDc and CDw respectively, for each term, 1 1 fD ¼ fDc þ fDw ¼ CDc av Uc;m 2 þ CDw av jUw;m jUw;m : 2 2

+

0

k !

ð18Þ

Both drag coefficients, C Dw and C Dc [e.g., Zhou and Graham, 2000] depend on the Reynolds number (Re = Uw d/n, where d is a typical length scale for the drag‐ generating elements), the Keulegan‐Carpenter number (KC = UwT/d, where T is the wave period), and the ratio of

4 of 15

C12005

C12005

LUHAR ET AL.: WAVE‐INDUCED FLOW INSIDE A SEAGRASS BED

Table 1. Wave and Vegetation Parameters for Each Experimenta

−2

z (Uc = 0)c (cm)

Uc,m (21)d (cm s−1)

Us (22)d (cm s−1)

UR,m (23)d (cm s−1)

ns (cm )

H (cm)

T (s)

a (cm)

A∞ /S

a (z = 1 cm)

a (7)

D1 D2 D3 D4e D5 D6 D6f

0.03 0.06 0.09 0.12 0.15 0.18 0.18

39 39 39 39 39 39 39

1.4 1.4 1.4 1.4 1.4 1.4 1.4

3.2 3.1 3.1 3.1 3.0 2.9 3.1

0.5 0.7 0.8 1.0 1.1 1.1 1.2

0.95 0.92 0.92 0.94 0.94 0.92 0.79

0.97 0.94 0.91 0.87 0.84 0.82 0.81

0.97 0.94 0.92 0.89 0.87 0.84 0.84

0.5 0.9 1.0 1.8 1.9 1.6 1.9

11.2 12.7 13.1 10.9 13.3 14.1 11.1

2.1 1.9 1.9 1.9 1.9 1.8 2.0

0.8 0.7 0.7 0.7 0.7 0.6 0.7

0.5 0.3 0.3 0.2 0.2 0.2 0.2

H1 H2 H3 H4e

0.12 0.12 0.12 0.12

16 24 32 39

1.4 1.4 1.4 1.4

0.9 1.7 2.4 3.1

0.6 0.8 0.9 1.0

0.95 0.94 0.93 0.94

0.89 0.88 0.88 0.87

0.89 0.89 0.89 0.89

0.3 0.8 1.4 1.8

5.2 9.0 9.2 10.9

1.0 1.5 1.7 1.9

0.2 0.4 0.5 0.7

0.5 0.3 0.3 0.2

T1 T2 T3e T4

0.12 0.12 0.12 0.12

39 39 39 39

0.9 1.1 1.4 2.0

2.8 3.3 3.1 3.2

0.4 0.7 1.0 1.6

0.93 0.92 0.94 0.89

0.89 0.88 0.87 0.85

0.89 0.89 0.89 0.89

0.4 1.1 1.8 2.2

1.9 9.1 10.9 13.4

0.7 1.6 1.9 2.5

0.7 0.9 0.7 0.7

0.2 0.3 0.2 0.3

A1 A1f A2 A3e A4 A5 A5f

0.12 0.12 0.12 0.12 0.12 0.12 0.12 (0.003)

39 39 39 39 39 39 39 (0.5)

1.4 1.4 1.4 1.4 1.4 1.4 1.4 (0.05)

0.8 0.8 1.7 3.1 4.3 5.2 5.3 (0.2)

0.2 0.3 0.5 1.0 1.4 1.6 1.7 (0.1)

0.94 0.92 0.95 0.94 0.94 0.91 0.80 (0.03)

0.89 0.89 0.89 0.87 0.86 0.85 0.85 (0.01)

0.89 0.89 0.89 0.89 0.89 0.89 0.89 (0.01)

0.1 0.1 0.4 1.8 3.3 4.2 4.3 (0.3)

1.7 4.9 10.2 10.9 12.0 11.9 12.2 (0.5)

0.3 0.3 0.8 1.9 3.2 4.2 4.3 (0.1)

0.0 0.0 0.2 0.7 1.4 2.0 2.1 (0.1)

0.0 0.0 0.1 0.2 0.5 0.6 0.6 (0.05)

c

ai (8)

Mean (Uc)c (cm s−1)

Run

b

d

d

a

The values shown in the last rows represent typical experimental uncertainty.

The wave amplitude was calculated by fitting the linear theory solution (1a) to measured oscillatory velocities at the highest four measurement locations (z ≥ 21 cm). Exceptions are runs H3, where the top three measurements were used, and runs H1 and H2, for which the top two measurements were used. c Indicates measurements from experiment. d Indicates equations used to arrive at the predicted values. e Identical runs; listed in multiple locations for clarity. f Repeats with wooden dowels left in place in the clearing. b

mean to oscillatory velocity. However, the two coefficients are typically of comparable magnitude.

[15] Substituting (18) into (17) and time averaging under the assumption that the parameters CDc, CDw, and av are constants leads to k !

Zhv 0

Zhv 1 4 1 av CDc Uc;m 3 þ CDw uw;m 3 dz ¼ CDc av Uc;m 2 dz; 2 3 2 0

ð19Þ

where uw, m is the magnitude of the in‐canopy oscillatory flow, Uw, m = uw, mcos(wt). The mean current is a second‐ order phenomenon, generated because of nonlinear interaction between the oscillatory velocities. As a result, Uc, m uw, m. Then, assuming CDw and CDc are of comparable magnitude, the energy dissipation within the meadow is dominated by the oscillatory drag force and k !

Zhv 0

Zhv 1 4 1 av CDw uw;m 3 dz ¼ CDc av Uc;m 2 dz: 2 3 2

ð20Þ

0

Given that the model seagrass blades are vertically uniform (see section 3) and that the parameter cosh khv, which is the ratio of the horizontal oscillatory velocity at the top of the canopy to that at the bed (equation (1a)), is smaller than 1.2 for all the cases tested here, we ignore any vertical variation in av, CDw, CDc, and uw, m. For simplicity, we also assume Uc, m to be constant over the height of the meadow

and solve equation (20) to obtain an estimate for the mean‐ current generated within the meadow, Uc;m ¼

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 4 CDw k uw;m 3 : 3 CDc !

ð21Þ

Equation (21) indicates that the magnitude of the mean current is controlled primarily by wave parameters (k, w, and uw, m) and does not depend on the canopy parameters (hv or av). However, below we discuss how the conditions under which (21) applies is dependent on the ratio of blade spacing and wave excursion (i.e., it will have some dependence on av). In addition to the wave conditions, an important Table 2. Observed and Predicted Velocity Ratio for Unidirectional Flow Over Seagrass Modela Run

hv (cm)

hCDavi (cm−1)

u*hv /Uhv

Cf b

^ m /U∞ U

ac (9)c

F1 F2 F3 F4 F5 F6

21.5 21.3 20.0 18.6 17.0 15.5

0.064 0.060 0.047 0.045 0.040 0.034

0.20 0.17 0.17 0.18 0.15 0.14

0.08 0.06 0.06 0.06 0.05 0.04

0.22 0.23 0.22 0.23 0.26 0.28

0.25 0.21 0.24 0.27 0.26 0.27

(0.5)

(0.003)

(0.02)

(0.01)

(0.01)

(0.01)

a

Data are from Ghisalberti and Nepf [2006]. Run numbers follow convention used by the above authors. The last row in the table indicates typical uncertainty. b Estimated based on Cf = 0.5 (u*hv /Uc,hv)2. c Indicates equation used.

5 of 15

C12005

LUHAR ET AL.: WAVE‐INDUCED FLOW INSIDE A SEAGRASS BED

quantity governing the magnitude of the mean current is the ratio of drag coefficients, CDw /CDc. Zhou and Graham [2000] performed numerical simulations estimating the force acting on a single circular cylinder in combined wave‐current flows. Simulation results for Uc /Uw = 0.25 showed the drag coefficient ratio to decrease from CDw /CDc ≈ 1.8 for KC = 0.2 (Re = 40) to CDw /CDc ≈ 0.5 for KC = 26 (Re = 5200), with Re and KC based on cylinder diameter. For the experimental conditions considered here, the Keulegan‐ Carpenter number, based on blade width and near‐bed orbital velocity, ranges from KC ≈ 14 (Re ≈ 90) to KC ≈ 94 (Re ≈ 590). If the stem diameter is used instead of blade width (see section 3), which might be more appropriate near the base of the model plants used for the experiments, the ranges are KC ≈ 5.8–39 (Re ≈ 220–1400). Given the overlap in range between the experiment conditions considered here, and the numerical simulations performed by Zhou and Graham [2000], it is reasonable to assume that CDw /CDc is an O[1] parameter.

2.4. Return Current [16] The preceding analysis considers wave‐induced oscillatory and mean flow at a fixed point (i.e., an Eulerian perspective). However, it is well known that even for purely oscillatory wave motion, individual water parcels (i.e., a Lagrangian perspective) tend to drift in the direction of wave propagation. This phenomenon is called Stokes’ drift [see Fredsøe and Deigaard, 1992 for a discussion]. Because the flume is a closed system, mass transport in the direction of wave propagation, attributed to both the wave‐induced mean current described above and the Stokes’ drift, sets up a surface slope. The pressure gradient caused by this surface setup drives a return current, leading to zero depth‐averaged net transport. This return current may modify the measured meadow drift, relative to the theoretical prediction derived above. To estimate the magnitude of this return current, we assume that the flow field attributed to the pressure gradient can simply be superimposed onto the existing wave‐induced oscillatory and mean velocity fields. Most of the return current will be diverted above the meadow because of canopy drag.

We assume that the ratio of the return current within the meadow (UR, m) to the return current above the meadow (UR) is given by the parameter ac, shown in equation (9). The depth‐integrated Stokes’ drift (per unit width) is Qs ¼ Us h ¼

1 a2 !

2 tanhðkhÞ

ð22Þ

[e.g., Fredsøe and Deigaard, 1992], where Us is the depth‐ averaged mass transport velocity attributed to Stokes’ drift.

We can now write the following mass balance, Qs þ Uc; m hv ¼ UR ðh hv Þ þ UR; m hv ;

ð23Þ

from which we can estimate the return current within the meadow, UR,m = acUR. Experimental results reported later (Table 1) suggest that, for most cases, the impact of the return current within the meadow is negligible.

3. Methods [17] The experiments were performed in a 24 m long, 38 cm wide, and 60 cm deep flume equipped with a paddle‐

C12005

type wave maker. The vertical paddle was actuated using a hydraulic piston driven by a Syscomp WGM‐101 arbitrary waveform generator. The waveform generator was programmed to produce surface waves of the desired amplitude and frequency based on the closed‐form solution for paddle motion described by Madsen [1971]. A plywood beach of slope 1:5 and covered with rubberized coconut fiber limited reflections to less than 10% of the incident wave. The model canopy was 5 m long. The canopy comprised model plants placed in four predrilled baseboards 1.25 m long. Two additional baseboards were placed both upstream and downstream of the model vegetation to ensure a uniform bed roughness across the test section. Each model plant consisted of six polyethylene (density rb = 920 kg m−3; elastic modulus E = 3 × 108 Pa) blades of length lb = 13 cm, width wb = 3 mm, and thickness tb = 0.1 mm attached to a 2 cm long wooden dowel of 0.64 cm diameter using rubber bands. With the rubber bands in place, the maximum diameter of the dowels was distributed with a mean of 0.92 cm and a standard deviation of 0.03 cm. Where necessary, a mean stem diameter, d = 0.78 cm, is used. When inserted into the baseboard, the stem (dowel) protruded 1 cm above the bed.

[18] Velocity measurements were made with a 3‐D Acoustic Doppler Velocimeter (ADV; Nortek Vectrino).

Synchronous measurements of the wave height were made at the same x‐location using a wave gauge of 0.2 mm accuracy. The analog output from the wave gauge was amplified and logged to a computer using an analog‐digital converter (NI‐USB6210, National Instruments). Both the ADV and wave gauge were mounted on a trolley moving on precision rails. Vertical profiles of velocity were measured at two longitudinal locations, midway through the canopy and upstream of the canopy. The model bed was shifted longitudinally along the flume to ensure that the measurement location midway through the canopy corresponded to an antinode of the partially standing waves created by reflections from the downstream end of the flume. The other measurement location was chosen to be an antinode at least half a wavelength upstream of the canopy. This eliminates the lower‐order spatially periodic variation in wave and velocity amplitude associated with the 10% reflection.

Velocities were measured at 1 cm vertical intervals. At each location, velocities and surface displacement were measured for 6 min at 25 Hz. The height of the lowest measurement location varied between 0.1 and 0.9 cm above the bed (z = 0).

[19] A schematic of the setup is shown in Figure 1. Wave period (T = 0.9–2.0 s) and amplitude (a = 0.8–5.3 cm), water depth (h = 16–39 cm), and vegetation density (ns = 300– 1800 stems m−2, or nb = 1800–10,800 blades m−2) were varied systematically. These parameter ranges were chosen based on typical field values for the dimensionless parameters a/h, kh, hv /h, and av hv (see Luhar et al. [submitted manuscript, 2010] for details). The conditions for each experimental run are shown in Table 1. To measure velocities close to the bed within the meadow, all vegetation was removed from a circular area approximately 10 cm in diameter, which was the minimum cleared area necessary to prevent blades from entering the measurement control volume. To test whether the clearing had an appreciable impact on the velocity structure near the bed, three runs were repeated with the wooden dowels (with rubber bands but no

6 of 15

C12005

LUHAR ET AL.: WAVE‐INDUCED FLOW INSIDE A SEAGRASS BED

C12005

Figure 2. Qualitative overview of the flow pattern at the meadow scale. The decay in wave height (fine black line) along the meadow results in a proportional decrease in the oscillatory velocity fields. The black ellipses with arrows indicate the wave orbitals. Vertical profiles of the mean current (heavy gray lines) are shown at an upstream, downstream, and in‐meadow position. At each position, the vertical dashed lines indicate the axis position for the profile. The local circulation pattern, shown by the large gray arrows, results from the difference in the velocity profile within and outside the meadow. The direction of wave propagation is from left to right. Not to scale. blades attached) replaced in the cleared area. These runs are marked with a superscript “f” in Table 1. The dynamic influence of this cleared area on both unsteady and steady velocity components is discussed below.

[20] The velocity measurements were decomposed into mean (Uc, Wc), root‐mean‐square (RMS) oscillatory (Uw, RMS, Ww, RMS), and turbulent (u′, w′) components using a phase‐averaging technique. The velocity readings were binned into different phases based on the upward zero‐ crossings (8 = 0 rad) of the synchronous wave elevation measurements. Wave elevation is defined as the instantaneous surface displacement minus the mean water level. The wave crest and wave trough correspond to 8 ≈ p/2 rad and 8 ≈ 3p/2 rad, respectively. The velocity measurements for each phase bin where then ensemble averaged for the entire record (180–396 waves, depending on frequency) to yield the phase‐averaged velocity values (u(8), w(8)). The mean and RMS velocity components were then calculated by performing the following operations (only the u‐component is shown for brevity): Uc ¼

1 2

Z

2

0

uð8Þd8;

ð24Þ

and Uw;RMS ¼

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Z 1 2 ðuð8Þ Uc Þ2 d8: 2 0

ð25Þ

Similarly, the turbulent Reynolds stress, u′w′ (8), was calculated by subtracting the phase‐averaged velocities from the instantaneous velocities, multiplying the vertical and horizontal components, and ensemble‐averaging over all data within that phase bin. The time‐averaged turbulent Reynolds stress (as before h i denotes a time average.) was then calculated as hu 0 w 0 i ¼

1 2

Z

0

2

u 0 w 0 ð8Þd8:

ð26Þ

4. Results [21] A qualitative overview of the observations at the scale of the entire bed is presented in Figure 2. Upstream of

the model seagrass bed, we observe very little wave decay (less than 1.5% decrease in wave height per wavelength), which is due to viscous dissipation at flume bed and walls [e.g., Hunt, 1952]. The RMS oscillatory velocities match predicted values based on linear wave theory. A small mean flow is generated close to the bed; the magnitude of this mean current is in reasonable agreement with the Longuet‐ Higgins [1953] solution for induced drift in laminar wave boundary layers.

[22] As an example, the velocity measurements shown in Figures 3a–3c, made upstream of the meadow for run A5, support this qualitative description of the velocity structure.

The RMS oscillatory velocities are predicted to within 5% by linear wave theory (Figure 3a), and the mean velocity is maximum at the measurement position closest to the bed (z = 0.4 cm, Uc = 2.4 cm/s) (Figure 3b). The magnitude of this mean current is consistent with the laminar boundary layer solution shown in equation (3), which predicts that the induced drift will be Uc = 1.9 cm/s outside the wave boundary layer. For laminar flows, the boundary layer thickness is O[(n/w)1/2] ∼ 0.05 cm. Over a smooth bottom, the boundary layer transitions from laminar to turbulent for a wave Reynolds number, Rew = U∞ A∞ /n > 5 × 104 [e.g., Fredsøe and Deigaard, 1992]. For the wave conditions tested here, Rew ≤ 10,000. Hence, we expect the bottom boundary layer to remain laminar upstream of the canopy.

[23] Figure 3c shows that the turbulent Reynolds stress is essentially zero within uncertainty throughout the water column upstream of the canopy, as expected for linear waves. Note that the Reynolds stress measurements at heights z = 8.4 and 9.4 cm (Figure 3c) are not reliable because these locations correspond roughly to the “weak spots” of the ADV (Nortek Forum data are available at http://www.nortek‐as.com/en/knowledge‐center/forum/ velocimeters/30180961; last accessed on 29 March 2010).

At this height, acoustic reflections from the bed interfere with the signal from the measurement volume, resulting in occasional spikes in both the horizontal and vertical components of velocity. We observe that the spikes tend to be more frequent during set phases of the wave cycle, resulting in a coherent bias of the hu′w′i estimate. To summarize, for the measurement location upstream of the meadow, the linear wave solution coupled with laminar boundary layer theory describes the velocity structure well.

7 of 15

C12005

LUHAR ET AL.: WAVE‐INDUCED FLOW INSIDE A SEAGRASS BED

C12005

Figure 3. Vertical profiles of RMS wave velocity, mean velocity, and Reynolds stress for run A5.

Figures 3a–3c correspond to the measurement location upstream of the meadow. Figures 3d–3f show profiles for the measurement location within the meadow. Results for the case in which a 5 cm radius circle was completely cleared of vegetation are plotted as white squares. Black squares represent the case in which wooden dowels were left in this clearing and only blades were removed. Solid lines in Figures 3a and 3d represent RMS velocity profiles predicted by linear wave theory (equation (1a)).

The horizontal dashed lines in Figures 3d–3f show the estimated maximum and minimum canopy heights over a wave cycle. The bottom line shows the canopy height under a wave crest, when the blades tend to lie streamwise, while the top line shows the canopy height under a wave trough, when the model blades are more upright.

[24] In contrast to the observations upstream of the meadow, wave decay is significant over the model seagrass bed (as much as 13% per wavelength for a water depth of 39 cm and 27% per wavelength for a water depth of 16 cm [see Luhar et al., submitted manuscript, 2010]). Furthermore, a mean current in the direction of wave propagation is generated within the model meadow, as shown schematically in Figure 2. This mean current is stronger and extends over a larger vertical distance than the boundary layer drift observed upstream of the meadow. Qualitative observations using a passive tracer (food coloring) indicate that the mean current is established within ∼50 cm of the start of the meadow and persists for a similar distance downstream of the meadow, beyond which the velocity structure resembles the observations made upstream of the meadow. The mean current induced within the meadow results in the local circulation pattern, indicated by large gray arrows in Figure 2.

[25] The schematic of in‐meadow velocity structure shown in Figure 2 is supported by the measurements shown in Figure 3. Vertical profiles of the RMS orbital velocity, the mean current, and the turbulent Reynolds stresses for run A5 are shown in Figures 3d–3f, respectively. The RMS oscillatory velocity is reduced relative to predictions based on linear theory below z ≈ 4 cm. For the 10 cm clearing completely devoid of model vegetation, the RMS orbital velocity is reduced to 91% of the predicted linear wave velocities at the lowest measurement location (z = 0.6 cm) (Figure 3d, white squares). However, with the stems left in the clearing, the RMS orbital velocity is reduced to 73% of the value predicted by linear wave theory at z = 0.3 cm (Figure 3d, black squares). The presence of wooden dowels in the clearing leads to an additional reduction (91% to 73%) in the RMS orbital velocity for z ≤ 1 cm, suggesting that our

8 of 15

C12005

LUHAR ET AL.: WAVE‐INDUCED FLOW INSIDE A SEAGRASS BED

C12005

Figure 4. Measured mean currents plotted against theoretical predictions. (a) Comparison of the maximum mean currents, Max(Uc), with the current induced in laminar boundary layers. White circles represent upstream measurements, gray squares indicate in‐canopy measurements for the complete clearing, and black squares represent repeat in‐canopy measurements with the model stems left in place. The dashed line represents the theoretical value shown in equation (3). (b) Comparison of the canopy‐averaged measured current, Mean(Uc), with the theoretical prediction (solid line) shown in equation (21) assuming uw, m = aw/sinh(kh), and CDw /CDc = 1. Gray and black squares are as described for Figure 4a. Note the different x‐axis scales in Figures 4a and 4b. measurements within the meadow underestimate the reduction of RMS velocity because of the clearing.

[26] Importantly, note that the presence or absence of model stems within the clearing does not affect the mean current significantly, as shown in Figure 3e. The maximum measured mean current is Uc = 7.3 cm/s for the complete clearing and Uc = 7.6 cm/s with the stems left in place; these values agree within experimental uncertainty. The mean current recorded at the lowest measurement location is close to that predicted for a laminar boundary layer. The magnitude of this mean current increases away from the bed and is greatest at approximately the elevation for which the RMS velocity begins to deviate from linear wave theory (z ≈ 4 cm in Figure 3d). Because the flume is a closed system, a return current develops above the meadow (z > 13 cm in Figure 3e). Vertical profiles of the turbulent Reynolds stress are physically consistent with the profiles of mean velocity (Figure 3f). The turbulent stress is opposite in sign to ∂Uc / ∂z, and it crosses zero at the same height as ∂Uc/∂z ≈ 0 (Figures 3e and 3f).

[27] Figure 4a compares the maximum mean current measured upstream of (white markers), and within (gray and black markers), the canopy with the predicted mean velocity for laminar boundary layers, given in equation (3). Consistent with Figure 3b, the maximum measured currents upstream of the canopy agree reasonably well with predicted values for boundary layers. However, the currents generated within the meadow can be 3–4 times larger than the laminar boundary layer prediction. The simple theory developed earlier (equation (21); Figure 4b, solid line) gives a better prediction of the measured in‐canopy currents. Note that equation (3) predicts the maximum current outside the boundary layer, whereas equation (21) predicts the vertically averaged mean flow in the seagrass meadow. To reflect this, the maximum measured mean current, Max(Uc), is plotted

in Figure 4a, while the canopy‐averaged mean current, Mean(Uc), is plotted in Figure 4b. Mean(Uc) is defined as the vertical average of the measured mean flow profile below the zero crossing for Uc (e.g., z ≤ 13 cm in Figure 3e).

[28] To arrive at a prediction for in‐canopy currents using equation (21), the following assumptions are made: the in‐ canopy oscillatory velocity is equal to the near‐bed velocity predicted by linear wave theory, uw,m = aw/sinh(kh), and the ratio of drag coefficients is CDw /CDc = 1. Under these assumptions, equation (21) simplifies to Uc,m = [(4/3p)(ka) a2w2/sinh3(kh)]1/2. The use of the near‐bottom oscillatory velocity in equation (21) is justified because the increase in horizontal oscillatory velocities over the height of the canopy is modest. As mentioned earlier, the ratio of the oscillatory velocity at the top of the canopy to the near‐bed velocity based on linear theory (1a) is smaller than 1.2 for all the cases tested here. Furthermore, vegetation resistance only leads to a limited reduction of in‐canopy oscillatory velocities as discussed below. The drag coefficient ratio CDw /CDc = 1 is chosen based on the range suggested by Zhou and Graham [2000], CDw /CDc ≈ 0.5–1.8.

[29] For cases D1–D3 (Figure 4b), the observed mean current is significantly lower than the values predicted by equation (21). These cases correspond to the lowest stem densities (ns in Table 1) tested here. Deviation at the lowest stem densities is not surprising, as the drift must transition back to the boundary layer drift below some threshold density. Figure 4b also suggests that equation (21) overpredicts the mean current for the cases with smaller wave orbital excursions (i.e., small a/sinh(kh)). It is tempting to attribute this overprediction to a variation in the drag coefficient ratio, CDw /CDc, based on KC and Re. However, the simulations performed by Zhou and Graham [2000] (see section 2) showed that the drag coefficient ratio, CDw /CDc, is highest for low KC and Re, suggesting that the mean current

9 of 15

C12005

LUHAR ET AL.: WAVE‐INDUCED FLOW INSIDE A SEAGRASS BED

C12005

Figure 5. (a) Canopy‐averaged mean current normalized by the theoretical prediction, Uc,m = [(4/3p) (CDw /CDc)(ka)a2w2/sinh3(kh)]1/2, plotted against the ratio of orbital excursion to stem center‐center spacing, A∞ /S. CDw /CDc is assumed to be 1. (b) Vertical elevation for the zero crossing in the mean current, z(Uc = 0), normalized by the blade length, lb, plotted against A∞ /S. In both panels, gray squares indicate in‐canopy measurements for the runs where a clearing was made for ADV access, and black squares represent repeat in‐canopy measurements with the model stems left in place. should be under‐predicted by CDw /CDc = 1 for waves with smaller orbital velocities and excursions. Clearly, variations in drag based on KC and Re do not explain the observations.

[30] We suggest that the ratio of orbital excursion, A∞, to , dictates the transition stem center‐center spacing, S = n−1/2 s between boundary layer drift and canopy‐induced current.

This is confirmed by Figure 5a, which shows the observed canopy‐averaged mean currents normalized by the predicted values plotted against the ratio A∞/S. The observed velocity matches the predictions very well for A∞ /S ≥ 1, whereas equation (21) overpredicts Mean(Uc) for A∞ /S < 1. The vertical extent over which the mean flow is positive within the canopy, z(Uc = 0), is also a function of A∞/S (Table 1 and Figure 5b). The height z(Uc = 0) is roughly equal to the blade length for A∞ /S ≥ 1 but is smaller than the blade length for A∞ /S < 1, consistent with a transition to boundary layer streaming for A∞ /S < 1. Finally, if we consider only the cases for which A∞ /S ≥ 1, the measured profiles collapse to a similar form when normalized by the predicted velocity scale (Figure 6), further confirming the theoretical model.

Physically, the large orbital excursions ensure that all the water parcels moving back and forth encounter the model vegetation for A∞ /S > 1. Hence, the bulk representation of seagrass canopy drag used here is accurate. In contrast, for A∞ /S < 1, only the water parcels moving back and forth in the vicinity of the model plants interact with vegetation, and the hydrodynamic impact of the canopy on the wave‐induced orbital velocities is diminished. In effect, a bulk representation of canopy drag is strictly valid only for A∞ /S ≥ 1. However, if we retain the distributed drag model for simplicity, the wave canopy drag coefficient is reduced for A∞ /S < 1 but not the current drag coefficient, resulting in a lower drag coefficient ratio, CDw /CDc. [31] Finally, we consider the possible impact of the expected return current. In a closed system, a return flow must balance the mass transport in the direction of wave propagation attributed to the wave‐induced current (Uc,mhv) and Stokes’ drift (Qs). We use equation (23) to estimate the

magnitude of the return current within the meadow, UR,m (Table 1). For the cases that satisfy the assumptions of equation (21), A∞ /S ≥ 1, the return flow within the meadow is small compared to the measured current. Specifically, UR,m is, at most, 15% of the measured mean current (Table 1).

This comparison suggests for A∞ /S ≥ 1 that the wave‐ induced drift within the meadow measured in this study is representative of the magnitudes that will occur in the field (i.e., in the absence of the flume‐associated return

Figure 6. Vertical profiles of measured mean velocity, Uc, normalized by magnitude of the wave‐driven current, Uc, m, predicted by equation (21) for all the runs with A∞ /S ≥ 1.

The vertical coordinate, z, is normalized by blade length, lb.

Runs are as denoted in the legend.

10 of 15

C12005

LUHAR ET AL.: WAVE‐INDUCED FLOW INSIDE A SEAGRASS BED

flow) and that equation (21) needs no adjustment for application in the field.

[32] Next, we consider the reduction in oscillatory velocity within the canopy, which is characterized by the ratio of observed to predicted (from linear theory) horizontal RMS velocity. The velocity reduction is estimated for all cases at z = 1 cm. When measurements are not available at z = 1 cm, we interpolate linearly between the two lowest velocity measurements. The resulting velocity ratio, a (z = 1 cm) is listed in Table 1. Table 1 also lists velocity reductions predicted by equation (8) for the inertia‐ dominated limit and by the general solution shown in equation (7). The elevation z = 1 cm was chosen as the basis for comparison for two reasons. First, velocity reductions were greatest near the bed (see Figure 3d), making the relative uncertainty smaller. Second, the forces exerted by the vegetation for z > 1 cm (recall the z ≤ 1 cm corresponds to the stem region) depend on blade posture and the relative motion between the water and the flexible blades. Predictive quantitative models for blade posture and motion are outside the scope of this study. Since the elevation z = 1 cm corresponds to the stem region, the velocity reduction for the inertia‐only limit (equation (8)) is calculated using the planar area parameter for the stems, lp = nspd2/4. The inertia coefficient is assumed to be Cm = 2, as is the case for cylinders. To estimate the velocity reduction using the general solution (equation (7)), we assume the frontal area parameter to be lf = ns dhs + nbwblb, where hs = 1 cm is the height of the stem and lb is the blade length. Furthermore, we use a drag coefficient, CD = 1, based on the typical values for a cylinder at Re ≥ O[100] and a shear coefficient, Cf = 0.05, based on velocity and Reynolds stress profiles measured by Ghisalberti and Nepf [2006] for a similar model seagrass meadow in unidirectional flow (see Table 2 and discussion below).

[33] As Table 1 shows, for all the wave conditions tested here, there is very little difference between velocity reductions predicted by the general solution, ∣a∣ compared to the inertia‐dominated limit, ai. This is in agreement with Lowe et al. [2005a], who note that the general solution diverges substantially from the inertia‐dominated limit only when the wave excursion to spacing ratio A∞ /S is greater than unity; this ratio is smaller than 2 for all the cases tested here.

Consistent with the observation made earlier, the velocity ratio is higher than predicted for most of the cases in which the model vegetation was removed to allow ADV access.

The wave‐induced flow adjusts locally to the clearing.

Hence, the removal of the model vegetation results in higher velocities locally. The exceptions are cases D1–D3, where the observed velocity ratios agree with the predictions within experimental uncertainty. Given the low vegetation densities for these cases, the clearing is not sufficiently distinct from the rest of the sparse meadow. For the cases in which the model stems were left in place in the clearing, agreement between the observed and predicted velocity ratios improves. Given that the smallest velocity ratio we observe is 80% (i.e., a reduction of 20%), the experiments suggest that the reduction in oscillatory velocities within seagrass meadows is limited for wave‐dominated conditions, consistent with the assumptions made in predicting the wave‐induced mean current.

C12005

[34] In contrast, velocities are significantly reduced in similar model seagrass meadows for unidirectional flows [Ghisalberti and Nepf, 2006], as shown in Table 2.

Ghisalberti and Nepf [2006] measured unidirectional velocity profiles over a similar model seagrass meadow of density 230 stems m−2 (1380 blades m−2). The shear stress coefficient is estimated using the relation Cf = 2(u*hv /Uhv)2 shown earlier. Here, u*hv = [− hu′w′i ]1/2 is the friction velocity, and Uhv is the unidirectional flow velocity at the top of the meadow, z = hv. Consistent with observations of natural seagrass [e.g., Grizzle et al., 1996], the meadow height decreased with increasing flow speed (Table 2). The compression of blades with increasing flow speed makes the interface with the overflow hydraulically smoother, reducing the friction coefficient of this interface (Cf ), a trend that was also noted by Fonseca and Fisher [1986]. The blade density considered by Ghisalberti and Nepf [2006] is at the lower limit of the conditions used here for the wave ^ m /U∞, is 28% experiments (Table 1), yet the velocity ratio, U or less (i.e., a reduction of 72% or more). With denser meadows, the reduction will be greater. For a typical dense canopy used here (i.e., 1200 stems m−2, av hv ≈ 2.9), equation (9) predicts a velocity ratio of ac = 0.13, a velocity reduction of 87%. The implications of these vastly different in‐canopy velocities under wave‐ and current‐dominated conditions are discussed in section 5.

5. Discussion [35] Perhaps the most interesting aspect of this study is the mean current induced within the model seagrass canopy. A significant body of analytical, numerical, and experimental work regarding wave‐induced mean currents within laminar and turbulent boundary layers over smooth, rippled, and rough beds already exists (see Davies and Villaret [1998, 1999] and Marin [2004] for relatively recent reviews).

However, to our knowledge, this is the first instance of a similar current being observed within submerged canopies.

For field applications, our results suggest that, in addition to wind and tidal forcing, mean currents within submerged canopies can also be induced by wave forcing.

[36] The wave‐induced current was established within 50 cm of the upstream edge of the canopy. We propose the following scaling for this development length, Lc. When the mean current is fully developed, the conservation of momentum reduces to a balance between the wave stress and the meadow drag (equation (15)). The balance hUwWw, mi /hv = 2 = 2 hUwWw, mi /CD c av hv . Over (1/2)CDc avUc,2 m leads to Uc,m the development length scale, drag is unimportant, and the wave stress is balanced by the convective acceleration term, Uc,2m /Lc ∼ hUwWw, mi /hv . Equating the above expressions leads to the development length Lc ∼ 2/CD c av , which is consistent with the discussion by Luhar et al. [2008]. Given that CDc is an O[1] parameter, Lc ∼ 2/av. For the vegetation densities tested here, av = 0.054–0.32 cm−1, leading to Lc ∼ 6.2–37 cm. Note that this analysis assumes that the wave stress hUwWwi is set up instantaneously at the upstream edge of the canopy. It is likely that the wave stress develops over a distance on the order of the wave excursion (A∞ = 0.7–4.8 cm for this study). Hence, in order for the mass drift to be generated within a meadow, the meadow must be longer than both the wave excursion and Lc ∼ 2/av. In the field, av ∼ 0.01–

11 of 15

C12005

C12005

LUHAR ET AL.: WAVE‐INDUCED FLOW INSIDE A SEAGRASS BED

Table 3. Predicted Velocity Reductions for Current‐ and Wave‐Dominated Conditions in the Field for a Range of Seagrass Speciesa Species

nb (m−2)

lb (cm)

wb (cm)

tb (mm)

l fb

l pb

ai (8)c

ac (9)c

Emaxd

References

P. oceanica

Maxe Mean Min

5600 3500 1400

81 50 32

0.9

0.8 0.5 0.3

40.8 15.8 4.0

0.041 0.014 0.003

0.68 0.78 0.90

0.03 0.06 0.11

4.4 3.7 2.8

Pergent‐Martini et al., 1994; Marbà et al., 1996; Fourqurean et al., 2007

Cymodocea nodosa

Max Mean Min

6000 3000 1500

30 17 10

0.3

1.0 0.5 0.2

5.4 1.5 0.5

0.018 0.005 0.001

0.95 0.97 0.99

0.10 0.18 0.33

3.1 2.3 1.7

Cancemi et al., 2002; Guidetti et al., 2002

Halodule wrightii

Max Mean Min

30000 18000 6000

20

0.2

0.4

12.0

3.7

Creed, 1997, 1999; Fonseca and Bell, 1998

1.2

0.89 0.93 0.98

0.06

10

0.024 0.014 0.005

0.20

2.2

Ruppia maritima

Max Mean Min

5700 4800 3600

100 70 40

0.2

0.8 0.6 0.2

8.6 5.0 2.2

0.007 0.004 0.001

0.99 0.99 0.99

0.08 0.10 0.15

3.6 3.2 2.6

Koch et al., 2006

T. testudinum

Max Mean Min

3020 2400 2000

35

1

0.5 0.4 0.4

10.6

0.77 0.80 0.83

3.3

Lee and Dunton, 2000; Terrados et al., 2008

2.0

0.015 0.010 0.007

0.07

10

0.16

2.3

Zostera marina

Max Mean Min

3850 2500 1350

80 60 32

0.8

0.3 0.3 0.3

23.1 11.3 3.2

0.010 0.006 0.003

0.82 0.88 0.93

0.05 0.07 0.12

4.2 3.6 2.7

Fonseca and Bell, 1998; Laugier et al., 1999; Guidetti et al., 2002

Zostera nolti

Max Mean Min

30000 21000 12000

20 21 8

0.2

0.5 0.4 0.2

9.0 6.6 1.4

0.021 0.011 0.003

0.94 0.95 0.97

0.07 0.09 0.19

3.5 3.3 2.3

Laugier et al., 1999; Cabaço et al., 2009

Enhalus acoroides

Mean

137

30

1.5

0.5

0.6

0.001

0.97

0.28

1.8

Vermaat et al., 1995; Green and Short, 2003

Halodule uninervis

Mean

7772

10

0.3

0.2

2.3

0.005

0.93

0.15

2.5

Vermaat et al., 1995; Green and Short, 2003

Thalassia hemprichii

Mean

2481

20

0.8

0.3

4.0

0.005

0.86

0.11

2.8

Vermaat et al., 1995; Green and Short, 2003

a

An estimate for the maximum mass transfer enhancement factor, Emax = (ai /ac)1/2, is also provided.

Vegetation parameters, l f = nbwblb and l p = nbwbtb.

Indicates equation used. d Emax = (ai /ac)0.5, as suggested by Lowe et al. [2005b]. e Max, maximum; Min, minimum. b c

0.3 cm−1, leading to Lc ∼ 6–200 cm (see Table 3 and also Luhar et al. [2008]), while the near‐bed wave excursion can be of O[10–500 cm], suggesting that the wave‐induced current is likely to exist for seagrass meadows longer than a few meters.

[37] At this point, it must be noted that while the laminar boundary layer theory developed by Longuet‐Higgins [1953] predicts a mean current in the direction of wave propagation, subsequent experimental and analytical studies [e.g., Davies and Villaret, 1998; Marin, 2004] show that both the magnitude and direction of the mean current change as the flow transitions from laminar to turbulent. The change in magnitude and direction of the mean current is attributed to the fact that wave asymmetry introduces asymmetries in the turbulence generated near the bed, or the vortices being shed from individual roughness elements. The drag characteristics of the vegetation depend on the periodic shedding and advection of vortices around the plants. Furthermore, qualitative observations of our model meadow indicate (see also Luhar et al., submitted manuscript, 2010) that the induced drift introduces an asymmetry in blade posture, whereby the blades lie streamwise in the direction of wave propagation under the wave crest and remain more upright

under the wave trough. The resulting increase in frontal area can lead to greater drag under the wave trough, when the horizontal oscillatory velocity is negative, thereby reinforcing the mean current. Asymmetries in turbulent processes and blade posture can be accounted for by using time‐ varying drag coefficients in equation (18) (cf. the time‐ varying eddy viscosity used by Trowbridge and Madsen [1984]); however, significant additional experimental and analytical work is required before this can be quantified with any certainty. [ 38] The simulations performed by Zhou and Graham [2000] were for a cylinder in isolation. For an array of model plants, processes such as wake interaction between neighboring plants could influence the drag coefficients. A combined wave‐current flow would also move the wakes back and forth and advect them downstream. To our knowledge, no previous studies have considered this interaction. As a result, we assume that the results obtained by Zhou and Graham [2000] for a single cylinder apply to the array of model plants used in this study.

[39] The generation of a mean current within submerged seagrass meadows has important implications for the health of meadows and for the ecologic services provided by the

12 of 15

C12005

LUHAR ET AL.: WAVE‐INDUCED FLOW INSIDE A SEAGRASS BED

seagrasses. As mentioned above, the mean current can lead to a bias in blade posture over a wave cycle. Blade posture can control light uptake and, hence, productivity in seagrass meadows [Zimmerman, 2003]. The predictive model developed by Zimmerman [2003] shows that the fraction of downwelling irradiance absorbed by submerged seagrass meadows increases as the bending angle of the seagrass blades increases. Increased absorption, caused by the increase in horizontal projected leaf area, leads to higher photosynthesis rates until a threshold bending angle of ∼20°.

Above this threshold, photosynthesis rates decrease because of self‐shading; a larger fraction of the incoming light is absorbed in the upper part of the canopy where photosynthesis is no longer limited by light availability.

[40] Notably, the wave‐induced current also has the potential to transport sediment and organic matter in the direction of wave propagation. Oscillatory wave velocities can generate turbulence close to the bed and suspend sediment but can only move the suspended sediment back and forth. In contrast, the wave‐induced current revealed in this study can advect the material away. Advection could be especially important for fine sediment and organic matter, where the majority of transport is in the form of suspended load. The mean current can also introduce a directional bias in the dispersal of spores, thereby dictating the direction of meadow expansion. Furthermore, the mean currents induced within the meadow may play a role in mediating the economically important nutrient cycling services provided by seagrasses. Nutrient cycling slows down if the rate at which seagrasses extract nutrients from the water is faster than the rate at which the water, and hence nutrients, are replenished within the meadow as a whole. In oscillatory flows, one mechanism of water renewal for seagrass meadows is turbulent exchange with the overlying water column. By systematically flushing the meadow (Figure 2), a wave‐induced mean current may provide a second mechanism of water renewal.

[41] The model developed by Lowe et al. [2005a] can be used to estimate the velocity reductions expected in field seagrass meadows. Table 3 shows the anticipated velocity ratios for the inertia‐ and current‐dominated limits, ai and ac, for a range of real seagrass species. These ratios are estimated using equations (8) and (9) based on the ranges of seagrass blade density (nb), length (lb), width (wb), and thickness (tb) listed in Table 3. The frontal area parameter is lf = nbwblb, and the planar area parameter is lf = nbwbtb. The drag and inertia coefficients are assumed to be CD = 1 and Cf = 0.05 as before (Tables 1 and 2), whereas the inertia coefficient is simply the ratio of blade width to blade thickness, Cm = wb /tb (see discussion in the study by Vogel [1994]). As Table 3 indicates, only a few species are likely to experience a significant reduction in in‐canopy velocity at the inertia‐dominated limit (e.g., Posidonia oceanica, ai = 0.68–0.90); however, at the current‐only limit, in‐canopy velocities are likely to be reduced by 67% or more for all the species listed (e.g., P. oceanica, ac = 0.03–0.11). Recall that the inertia‐dominated limit is applicable when the wave excursion is much smaller than the drag and shear length scales. Because near‐bed wave excursions in the field are likely to range from O[10–500 cm], while the drag length scale, LD = 2hv(1 − lp)/CDlf is O[5–10 cm] for all the

C12005

species listed in Table 3, the inertia‐dominated limit may not be very relevant for field conditions. For field conditions, therefore, the wave velocity reduction is likely to lie somewhere between the predictions for the inertia‐dominated and current‐only limits. However, the general conclusion that oscillatory velocities are attenuated less within submerged canopies compared to unidirectional currents remains valid.

[42] The higher magnitude of in‐canopy velocities for oscillatory flows relative to comparable unidirectional flows leads to an enhancement of mass transfer at the surface of individual canopy elements [e.g., Lowe et al., 2005b]. Via experiments measuring the rate of dissolution of gypsum blocks, Lowe et al. [2005b] confirmed the oft‐cited dependence of the convective mass transfer velocity (uCT) on the flow speed within the canopy, u CT ∼ U 1/2 m . Lowe et al.

[2005b] introduced the “enhancement factor,” which is the ratio of mass transfer in oscillatory flow relative to that in unidirectional flow. The maximum possible value for this mass transfer enhancement factor is found in flows at the inertial limit, i.e., Emax = (ai /ac)1/2. This ratio is listed in Table 3 for typical field conditions. Estimates of Emax show that convective mass transfer in and out of submerged seagrass meadows may be as much as ∼1.7 to ∼4.4 times larger for oscillatory flows compared to unidirectional currents of the same magnitude. This is in good agreement with the observations made by Thomas and Cornelisen [2003], which show ammonium uptake in a meadow of Thalassia testudinum to be ∼1.5 times greater in oscillatory flow relative to unidirectional flow. The maximum enhancement predicted for this species of seagrass ranges from ∼2.3 to ∼3.3 (Table 3).

[43] The weaker damping of in‐canopy velocity observed for oscillatory flows compared to mean currents may lead to different horizontal spatial structure within a meadow. In the presence of currents, a meadow can greatly reduce the near‐ bed velocity, and hence bed stress (e.g., as shown by the typical range of velocity ratios shown in Tables 2 and 3).

This can create a feedback that maintains a fragmented meadow structure, as described by Luhar et al. [2008]. An isolated patch of seagrass reduces the bed stress within the patch, and the diversion of flow away from the patch enhances the bed stress on the adjacent bare bed. Similarly, flow is enhanced locally within channels cutting through the meadow, inhibiting regrowth and thereby stabilizing the channels [Temmerman et al., 2007]. The scenario is different in wave‐dominated conditions, because the meadow does not significantly reduce near‐bed wave velocity (and associated bed stress), relative to adjacent bare bed (e.g., as seen in Tables 1 and 3). When a local area of meadow is lost, the bed stress in the bare patch does not increase appreciably, and the vegetation can grow back.

[44] On the basis of this difference in wave‐ and current‐ dominated conditions, we anticipate that regions dominated by currents will have more fragmented meadows, because any channels and cuts in the meadow will be maintained by the local adjustment in near‐bed flow and bed stress. In contrast, regions dominated by waves will have more uniform vegetation distributions, because under waves, there is little local flow adjustment to the meadow. Some support for the above hypothesis can be found in the field literature.

Fonseca et al. [1983] observed that as the hydrodynamic

13 of 15

C12005

LUHAR ET AL.: WAVE‐INDUCED FLOW INSIDE A SEAGRASS BED

conditions became more current dominated, the meadows became more fragmented. Similarly, Fonseca and Bell [1998] measured the percentage of meadow cover across 50 m × 50 m plots and found higher correlations in linear regression between percentage cover and current (r 2 = 0.60) than between percentage cover and wave exposure (r 2 = 0.45). Using a multiple regression, they found that percentage cover was predominately explained by current (r 2 = 0.54) with only a minor contribution from wave exposure (r 2 = 0.07).

6. Conclusions [45] Velocity profiles measured within and above a model seagrass meadow show that a mean current is generated within the meadow under wave forcing. Similar to boundary layer streaming, this mean current is forced by a nonzero wave stress. A simple model, developed in section 2, is able to predict the magnitude of this mean current. By introducing a bias in blade posture, the induced mean current may affect light uptake [Zimmerman, 2003] and hence photosynthesis rates. Furthermore, the mean current can play an important role in the net transport of suspended sediment and organic matter. Finally, by continuously renewing the water within the meadow, the induced current may also mediate the ecologically and economically important nutrient cycling services [Costanza et al., 1997] provided by seagrass meadows.

[46] In agreement with Lowe et al. [2005a], the velocity reduction within the meadows is lower for oscillatory flows compared to unidirectional flows. The higher in‐ canopy velocities associated with wave‐dominated conditions have been observed to enhance nutrient and oxygen transfer between the seagrasses and the water [Thomas and Cornelisen, 2003]. Finally, the limited reduction of in‐canopy oscillatory velocities suggests that in wave‐ dominated regions, the bed stress is not sufficiently distinct in any cuts or channels compared to areas of healthy meadow. Hence, seagrasses may be able to recolonize areas of lost meadow, leading to more uniform meadow structure. This is in contrast to tidal‐ or current‐dominated regions where any cuts or channels tend to be stable because of the local increase in flow and hence bed stress [Temmerman et al., 2007].

[47] Acknowledgments. This study received support from the U.S. National Science Foundation under grant OCE 0751358. Any conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation. Sylvain Coutu was supported by the Boston‐Strasbourg Sister City Association during his stay at MIT, and he thanks the cities of Strasbourg and Boston for making this collaboration possible. Eduardo Infantes thanks the Spanish Ministerio de Ciencia e Innovación (MICINN) FPI scholarship program (BES‐2006‐12850) for financial support. We also thank Ole Madsen, Amala Mahadevan, and Shreyas Mandre for helpful discussions.

References Bradley, K., and C. Houser (2009), Relative velocity of seagrass blades: Implications for wave attenuation in low‐energy environments, J. Geophys. Res., 114, F01004, doi:10.1029/2007JF000951.

Cabaço, S., R. Machas, and R. Santos (2009), Individual and population plasticity of the seagrass Zostera noltii along a vertical intertidal gradient, Estuarine Coastal Shelf Sci., 82(2), 301–308, doi:10.1016/j.ecss.2009.

01.020.

C12005

Cancemi, G., M. C. Buia, and L. Mazzella (2002), Structure and growth dynamics of Cymodocea nodosa meadows, Sci. Mar., 66, 365–373.

Chen, S. N., L. P. Sanford, E. W. Koch, F. Shi, and E. W. North (2007), A nearshore model to investigate the effects of seagrass bed geometry on wave attenuation and suspended sediment transport, Estuaries Coasts, 30(2), 296–310.

Costanza, R., et al. (1997), The value of the world’s ecosystem services and natural capital, Nature, 387, 253–260, doi:10.1038/387253a0.

Creed, J. C. (1997), Morphological variation in the seagrass Halodule wrightii near its southern distributional limit, Aquat. Bot., 59(1–2), 163–172, doi:10.1016/S0304-3770(97)00059-4.

Creed, J. C. (1999), Distribution, seasonal abundance and shoot size of the seagrass Halodule wrightii near its southern limit at Rio de Janeiro state, Brazil, Aquat. Bot., 65(1–4), 47–58, doi:10.1016/S0304-3770(99)00030-3.

Davies, A. G., and C. Villaret (1998), Wave‐induced currents above rippled beds and their effects on sediment transport, in Physics of Estuaries and Coastal Seas, edited by J. Dronkers and M. Scheffers, pp. 187–199, Balkema, Brookfield, VT. Davies, A. G., and C. Villaret (1999), Eulerian drift induced by progressive waves above rippled and very rough beds, J. Geophys. Res., 104(C1), 1465–1488, doi:10.1029/1998JC900016.

Finnigan, J. (2000), Turbulence in plant canopies, Annu. Rev. Fluid Mech., 32(1), 519–571, doi:10.1146/annurev.fluid.32.1.519.

Fonseca, M. S., and S. S. Bell (1998), Influence of physical settings on seagrass landscapes near Beauford, North Carolina, USA, Mar. Ecol. Prog.

Ser., 171, 109–121, doi:10.3354/meps171109.

Fonseca, M. S., and J. A. Cahalan (1992), A preliminary evaluation of wave attenuation by four species of seagrass, Estuarine Coastal Shelf Sci., 35, 565–576, doi:10.1016/S0272-7714(05)80039-3.

Fonseca, M. S., and J. S. Fisher (1986), A comparison of canopy friction and sediment movement between four species of seagrass with reference to their ecology and restoration, Mar. Ecol. Prog. Ser., 29, 15–22, doi:10.3354/meps029015.

Fonseca, M. S., J. C. Zieman, G. W. Thayer, and J. S. Fisher (1983), The role of current velocity in structuring eelgrass (Zostera marina L.) meadows, Estuarine Coastal Shelf Sci., 17, 367–380, doi:10.1016/0272-7714(83)90123-3.

Fourqurean, J. W., N. Marba, C. M. Duarte, E. Diaz‐Almela, and S. Ruiz‐ Halpern (2007), Spatial and temporal variation in the elemental and stable isotopic content of the seagrasses Posidonia oceanica and Cymodocea nodosa from the Illes Balears, Spain, Mar. Biol. Berlin, 151, 219–232, doi:10.1007/s00227-006-0473-3.

Fredsøe, J., and R. Deigaard (1992), Mechanics of Coastal Sediment Transport, World Sci., Singapore.

Gacia, E., T. C. Granata, and C. M. Duarte (1999), An approach to measurement of particle flux and sediment retention within seagrass (Posidonia oceanica) meadows, Aquat. Bot., 65(1–4), 255–268, doi:10.1016/ S0304-3770(99)00044-3.

Ghisalberti, M., and H. M. Nepf (2002), Mixing layers and coherent structures in vegetated aquatic flows, J. Geophys. Res., 107(C2), 3011, doi:10.1029/2001JC000871.

Ghisalberti, M., and H. M. Nepf (2004), The limited growth of vegetated shear layers, Water Resour. Res., 40(7), W07502, doi:10.1029/ 2003WR002776.

Ghisalberti, M., and H. Nepf (2006), The structure of the shear layer in flows over rigid and flexible canopies, Environ. Fluid Mech., 6, 277–301, doi:10.1007/s10652-006-0002-4.

Granata, T. C., T. Serra, J. Colomer, X. Casamitjana, C. M. Duarte, and E. Gacia (2001), Flow and particle distributions in a nearshore seagrass meadow before and after a storm, Mar. Ecol. Prog. Ser., 218, 95–106, doi:10.3354/meps218095.

Green, E. P., and F. T. Short (2003), World Atlas of Seagrasses, Univ. of Calif. Press, Berkeley.

Grizzle, R. E., F. T. Short, C. R. Newell, H. Hoven, and L. Kindblom (1996), Hydrodynamically induced synchronous waving of seagrasses: Monami and its possible effects on larval mussel settlement, J. Exp. Mar. Biol. Ecol., 206, 165–177, doi:10.1016/S0022-0981(96)02616-0.

Guidetti, P., M. Lorenti, M. C. Buia, and L. Mazzella (2002), Temporal dynamics and biomass partitioning in three Adriatic seagrass species: Posidonia oceanica, Cymodocea nodosa, Zostera marina, Mar. Ecol.

Berlin, 23, 51–67, doi:10.1046/j.1439-0485.2002.02722.x.

Hunt, J. N. (1952), Amortissement par viscosite de la houle sur un fond incline dans un canal de largeur finie, Houille Blanche, 7, 836.

Irlandi, E. A., and C. H. Peterson (1991), Modification of animal habitat by large plants: Mechanisms by which seagrasses influence clam growth, Oecologia, 87, 307–318, doi:10.1007/BF00634584.

Johns, B. (1970), On the mass transport induced by oscillatory flow in a turbulent boundary layer, J. Fluid Mech., 43, 177–185, doi:10.1017/ S0022112070002306.

14 of 15

C12005

LUHAR ET AL.: WAVE‐INDUCED FLOW INSIDE A SEAGRASS BED

Kobayashi, N., A. W. Raichle, and T. Asano (1993), Wave attenuation by vegetation, J. Waterw. Port Coastal Ocean Eng., 119, 30–48, doi:10.1061/(ASCE)0733-950X(1993)119:1(30).

Koch, E. W., and G. Gust (1999), Water flow in tide‐ and wave‐dominated beds of the seagrass Thalassia testudinum, Mar. Ecol. Prog. Ser., 184, 63–72, doi:10.3354/meps184063.

Koch, E. W., L. P. Sandford, S. N. Chen, D. J. Shafer, and J. M. Smith (2006), Waves in seagrass systems: Review and technical recommendations, U.S. Army Corps of Engineers Tech. Rep. ERDC TR‐06‐15.p., Eng. Res. and Dev. Cent., Alexandria, Va. Laugier, T., V. Rigollet, and M. L. de Casabianca (1999), Seasonal dynamics in mixed eelgrass beds, Zostera marina L. and Zostera noltii Hornem., in a Mediterranean coastal lagoon (Thau lagoon, France), Aquat. Bot., 63, 51–69, doi:10.1016/S0304-3770(98)00105-3.

Lee, K. S., and K. H. Dunton (2000), Effects of nitrogen enrichment on biomass allocation, growth, and leaf morphology of the seagrass Thalassia testudinum, Mar. Ecol. Prog. Ser., 196, 39–48, doi:10.3354/meps196039.

Longuet‐Higgins, M. S. (1953), Mass transport in water waves, Philos.

Trans. R. Soc. A, 245(903), 535–581, doi:10.1098/rsta.1953.0006.

Longuet‐Higgins, M. S. (1958), The mechanics of the boundary‐layer near the bottom in a progressive wave, Proceedings of the 6th International Conference on Coastal Engineeering, pp. 184–193, Am. Soc. of Civil Eng., New York.

Lowe, R. J., J. R. Koseff, and S. G. Monismith (2005a), Oscillatory flow through submerged canopies: 1. Velocity structure, J. Geophys. Res., 110, C10016, doi:10.1029/2004JC002788.

Lowe, R. J., J. R. Koseff, S. G. Monismith, and J. L. Falter (2005b), Oscillatory flow through submerged canopies: 2. Canopy mass transfer, J. Geophys. Res., 110, C10017, doi:10.1029/2004JC002789.

Lowe, R. J., J. L. Falter, J. R. Koseff, S. G. Monismith, and M. J. Atkinson (2007), Spectral wave flow attenuation within submerged canopies: Implications for wave energy dissipation, J. Geophys. Res., 112, C05018, doi:10.1029/2006JC003605.

Luhar, M., J. Rominger, and H. Nepf (2008), Interaction between flow, transport and vegetation spatial structure, Environ. Fluid Mech., 8, 423–439, doi:10.1007/s10652-008-9080-9.

Madsen, O. (1971), On the generation of long waves, J. Geophys. Res., 76(36), 8672–8683, doi:10.1029/JC076i036p08672.

Marbà, N., C. M. Duarte, J. Cebrian, M. E. Gallegos, B. Olesen, and K. Sand‐Jensen (1996), Growth and population dynamics of Posidonia oceanica on the Spanish Mediterranean coast: Elucidating seagrass decline, Mar. Ecol. Prog. Ser., 137(1–3), 203–213, doi:10.3354/ meps137203.

Marin, F. (2004), Eddy viscosity and Eulerian drift over rippled beds in waves, Coastal Eng., 50, 139–159, doi:10.1016/j.coastaleng.2003.

10.001.

Mei, C. C., M. Stiassnie, and D. K.‐P. Yue (2005), Theory and Applications of Ocean Surface Waves, World Sci., Singapore.

Mendez, F. J., and I. J. Losada (2004), An empirical model to estimate the propagation of random breaking and non‐breaking waves over vegetation fields, Coastal Eng., 51, 103–118, doi:10.1016/j.coastaleng.2003.11.003.

Méndez, F. J., I. J. Losada, and M. A. Losada (1999), Hydrodynamics induced by wind waves in a vegetation field, J. Geophys. Res., 104(C8), 18,383–18,396, doi:10.1029/1999JC900119.

C12005

Nepf, H. M., and E. R. Vivoni (2000), Flow structure in depth‐limited, vegetated flow, J. Geophys. Res., 105(C12), 28,547–28,557, doi:10.1029/ 2000JC900145.

Pergent‐Martini, C., V. Rico‐Raimondino, and G. Pergent (1994), Primary production of Posidonia oceanica in the Mediterranean Basin, Mar. Biol.

Berlin, 120, 9–15.

Peterson, C. H., H. C. Summerson, and P. B. Duncan (1984), The influence of seagrass cover on population structure and individual growth rate of a suspension‐feeding bivalve, Mercenaria mercenaria, J. Mar. Res., 42(1), 123–138, doi:10.1357/002224084788506194.

Santos, S. L., and J. L. Simon (1974), Distribution and abundance of the polychaetous annelids in a south Florida estuary, Bull. Mar. Sci., 24, 669–689.

Sarpkaya, T., and M. Isaacson (1981), Mechanics of Wave Forces on Offshore Structures, Van Nostrand Reinhold, New York.

Stoner, A. W. (1980), The role of seagrass biomass in the organization of benthic macrofaunal assemblages, Bull. Mar. Sci., 30, 537–551.

Temmerman, S., T. J. Bouma, J. Van de Koppel, D. Van der Wal, M. B. De Vries, and P. M. J. Herman (2007), Vegetation causes channel erosion in a tidal landscape, Geology, 35, 631–634, doi:10.1130/G23502A.1.

Terrados, J., P. Ramirez‐Garcia, O. Hernandez‐Martinez, K. Pedraza, and A. Quiroz (2008), State of Thalassia testudinum Banks ex Konig meadows in the Veracruz Reef system, Veracruz, Mexico, Aquat. Bot., 88(1), 17–26, doi:10.1016/j.aquabot.2007.08.003.

Thomas, F. I. M., and C. D. Cornelisen (2003), Ammonium uptake by seagrass communities: Effects of oscillatory versus unidirectional flow, Mar. Ecol. Prog. Ser., 247, 51–57, doi:10.3354/meps247051.

Trowbridge, J. H., and O. S. Madsen (1984), Turbulent wave boundary layers, 2, Second‐order theory and mass transport, J. Geophys. Res., 89(C5), 7999–8007, doi:10.1029/JC089iC05p07999.

Vermaat, J. E., N. S. R. Agawin, C. M. Duarte, M. D. Fortes, N. Marbà, and J. S. Uri (1995), Meadow maintenance, growth and productivity of a mixed Philippine seagrass bed, Mar. Ecol. Prog. Ser., 124(1–3), 215–225, doi:10.3354/meps124215.

Vogel, S. (1994), Life in Moving Fluids, Princeton Univ. Press, Princeton, N. J. Ward, L. G., W. Michael Kemp, and W. R. Boynton (1984), The influence of waves and seagrass communities on suspended particulates in an estuarine embayment, Mar. Geol., 59(1–4), 85–103, doi:10.1016/0025-3227 (84)90089-6.

Zhou, C. Y., and J. M. R. Graham (2000), A numerical study of cylinders in waves and currents, J. Fluids Structures, 14, 403–428, doi:10.1006/ jfls.1999.0276.

Zimmerman, R. C. (2003), A bio‐optical model of irradiance distribution and photosynthesis in seagrass canopies, Limnol. Oceanogr., 48, 568–585, doi:10.4319/lo.2003.48.1_part_2.0568.

S. Coutu, Institut d’Ingénieurie de l’Environnement, Ecole Polytechnique Fédérale de Lausanne, GR A1 445, Station 2, CH‐1015 Lausanne, Switzerland.

S. Fox, M. Luhar, and H. Nepf, Department of Civil and Environmental Engineering, Massachusetts Institute of Technology, 15 Vassar St., 48‐216, Cambridge, MA 02139, USA. (mluhar@mit.edu) E. Infantes, Instituto Mediterráneo de Estudios Avanzados, IMEDEA (CSIC‐UIB), Miquel Marqués 21, 07190 Esporles, Spain.

15 of 15

Vie et milieu - life and environment, 2010, 60 (4): 307-312

Seed maturity of the Mediterranean seagrass Cymodocea nodosa M. Domínguez *, E. Infantes, J. Terrados Instituto Mediterráneo de Estudios Avanzados IMEDEA (CSIC-UIB), C/ Miquel Marqués 21, 07190 Esporles, Mallorca, Spain * corresponding author: mdominguez@imedea.uib-csic.es

CYMODOCEA NODOSA SEEDS TIME OF COLLECTION INDUCTION OF GERMINATION SEEDLINGS VEGETATIVE DEVELOPMENT

ABSTRACT. – The germination of seeds in aquaria and posterior planting in the sea is an approach that may be applied in seagrass restoration projects. Hypo-salinity is known to induce germination of Cymodocea nodosa but the appropriate period for seed collection is useful information that has not been considered until now. The aim of this study was to evaluate the maturity of C. nodosa seeds by comparing the effectiveness of the hypo-salinity induction of germination in seeds collected in August, September and February and detect possible differences in the development of the seedlings. None of the seeds collected in August germinated after hyposalinity induction and died during the following months of storage in aquaria. The September and February seeds showed high germination percentage after hypo-salinity induction (> 90 %) and most of them develop into seedlings that showed similar vegetative development. Germination of seeds under full salinity was much slower and scarce (less than 15 % after six months). It was feasible to produce well-developed seedlings of C. nodosa with 3-4 leaves and 2 roots within five weeks from seeds collected from September to February.

Introduction Seagrass meadows play important roles in coastal processes because they enhance biological productivity and biodiversity, stabilize the sediment and reduce coastal erosion, and trap suspended particles and dissolved nutrients in the water column (Hemminga & Duarte 2000).

These ecosystems are highly sensitive to water quality degradation and seagrass meadows are one of the most threatened coastal ecosystems (Orth et al. 2006, Waycott et al. 2009). The renewed interest in seagrass restoration during the last decade (Christensen et al. 2004, Orth et al. 2006), increased the need for seagrass research, from the production of new knowledge about seagrass biology and ecology relevant to restoration (Kenworthy et al.

2006) to the adaptation of existing restoration techniques to local species and conditions (Calumpong & Fonseca 2001).

Cymodocea nodosa (Ucria) Ascherson is the second most important seagrass concerning the surface area covered in the Mediterranean after Posidonia oceanica (L.) Delile (Procaccini et al. 2003). Like in most seagrasses, clonal growth is the species main mechanism of proliferation, but in restoration projects the use of seeds or seedlings is considered to be more advantageous as it reduces the damages to donor meadows, facilitates the logistics involved in the production and handling of planting units, and assures a certain level of genetic diversity of the planted populations (Fonseca et al. 1998, Procaccini & Piazzi 2001, Christensen et al. 2004).

Cymodocea nodosa produces flowers and fruits yearly, and seed banks are common (Caye & Meinesz 1985, Ter-

rados 1993, Buia & Mazzella 1991, Cancemi et al. 2002).

The experiments performed by Caye & Meinesz (1986) showed that it is possible to induce the germination of C. nodosa seeds in the laboratory using a hypo-saline treatment (germination might occur in 2-6 days), but the state of maturity of the fruits and, consequently, the appropriate period for their collection is relevant information that has not been considered until now. The aim of this study was to evaluate the maturity of C. nodosa fruits by comparing the capacity of fruits collected in different months to develop into seedlings.

The hypo-saline induction of C. nodosa seeds may allow obtaining high germination percentages in short time and a sufficiently high number of seedlings for restoration. Here we investigated the feasibility of producing seedlings of Cymodocea nodosa from seeds collected at different times along the year considering that seed maturity may change with time. C. nodosa flowers from April to June, the fruits develop from July to October (Caye & Meinesz 1985, Terrados 1993, Buia & Mazzella 1991, Cancemi et al. 2002), and the seeds start to germinate in April or May of the following year (Pirc et al. 1986, Buia & Mazzella 1991). Seeds were collected in August 2008, when fruits are attached to the shoot, in September 2008, when fruits are still attached to the shoot but fruit pericarp starts to degrade, and in February 2009, when all the fruits had lost the pericarp and had become loose seeds in the sediment. A hypo-saline induction (Caye & Meinesz 1986) of seed germination was used to assess if the seeds were viable and sufficient mature to produce seedlings.

The early vegetative development of the seedlings produced was compared.

308

M. DOMÍNGUEZ, E. INFANTES, J. TERRADOS

Methods The fruits/seeds of Cymodocea nodosa were collected by SCUBA diving between depths of 2-3 m inside a patchy meadow located in the Bay of Pollença, North coast of Mallorca, Balearic Islands, Spain (N 39º 51’ 41.97”; W 3º 6’ 47.46”). The sampling was carried out in August 2008, in September 2008 and in February 2009, resulting in the collection of 79, 330 and 233 fruits/ seeds, respectively. We consider that all the fruits/seeds collected were produced during the summer of 2008, as no loose seeds were found in the sediment in August 2008 and September 2008.

The fruits/seeds were transported to the laboratory in aerated seawater and ambient temperature within 3 hours of collection.

They were maintained until the beginning of the experiment in February 2009, inside a temperature-controlled room in 60 L aquaria over a layer of sediment (fine sand) of the collection site, at full salinity (36-37) seawater, 23 ºC, and a light:dark photoperiod of 14:10 h.

Eighteen 20 L aquaria were set inside the temperature-controlled room and each was equipped with a filter, an air pump and a 9W white-light fluorescent lamp (Philips MASTER PLS9W/840/222P) that provided a photosynthetically active photon flux density of 17.5 ± 4.2 µmol m-2 s-1. Half of the aquaria was filled with full salinity (36-37) seawater and the other half was filled with seawater mixed with distilled water to achieve a salinity of 10. This hypo-saline treatment induces germination in less than a week (Caye & Meinesz 1986). Salinity was controlled regularly with an YSI Model 30 handheld salinity probe. No sediment was placed in the aquaria to assure the constancy of the salinity treatment of the fruits/seeds. The fruits/ seeds from each month of collection were distributed in 6 of the 20 L aquaria, 3 with full salinity seawater and 3 with a salinity of 10. The number of fruits/seeds per aquaria was 13 for the August 2008 collection, 55 for the September 2008 collection, and 39 for the February 2009 collection. The aquaria were kept at a temperature 20-21 ºC, which allows a seed germination rate above 90 % within two weeks and a good development of the seedlings (Caye & Meinesz 1986).

Seeds were considered germinated when the dorsal ridge of the endocarp opened and began to show the cotyledon (Fig. 1A).

The hypo-saline treatment was maintained for 15 days. Caye & Meinesz (1986) showed that germinated seeds should be returned to full salinity Mediterranean seawater for proper seedling development. For this reason, the seeds that did not germinate were removed from the aquaria to separate containers to maintain the hypo-saline treatment. The salinity in the aquaria was gradually increased to 36-37 of salinity within twelve days.

The seeds kept in the separate containers at salinity of 10, continue to germinate and after 15 additional days most of them had germinated. Similarly to the aquaria, the salinity in the containers was gradually increased to reach full salinity. Five weeks after the beginning of the experiment the number of leaves and roots were counted, and the length and width of each leaf and the length of each root were measured for all the seedlings obtained (Fig. 1B). The leaf surface area of each seedling and total root length (sum of all roots per seedling) was estimated.

As the germination of seeds in the full salinity aquaria was very slow and scarce, it was possible to identify each seedling.

Each germinated seed was placed individually in glass Petri dishes inside the same aquaria and labelled. Five weeks after germination, the leaves and the roots of each seedling were counted and measured as in the hypo-saline treatment.

Data analysis: the t-test for independent samples was used to compare the cumulative percentages of germination of seeds collected in September 2008 and February 2009 in the two salinity treatments. The t-test for independent samples was also used to evaluate the difference for the number of leaves, the leaf surface area, the number of roots and the total root length of the seedlings between September 2008 and February 2009. Prior to the t-test analysis, the homogeneity of variances was tested

by Levene’s test. Because the seedling measurements were done at different intervals of time (the “full salinity” seedlings were measured exactly after 5 weeks of germination) while the “hypo-saline” seedlings were measured after 5 weeks of the initiation of the experiment, the early vegetative development

Fig. 1. – Photographs of Cymodocea nodosa seeds after hypo-salinity induction. A: seed showing the dorsal ridge opened, and the elongating cotyledon.

B: seedling five weeks after the germination.

Vie Milieu, 2010, 60 (4)

MATURITY OF CYMODOCEA NODOSA SEEDS

309

each aquarium had germinated (Fig. 2). The February 2009 seeds germinated faster than the September 2008 seeds showing about 70 % higher percentage of germination by the 3rd day of experiment (t-test: t = -5.11, P = 0.007). The germination rate of the February 2009 seeds slowed down after that and by the 15th day after the initiation of germination, the percentage of germinated seeds was 16 % lower (t-test: t = 3.24, P = 0.032) than that of the September 2008 seeds. No differences of the cumulative percentage of germinated seeds were detected in any other measuring dates (Table I). The maximum percentage of seed germination reached 27 days after the start of the experiment was of 94 ± 1.6 % for the seeds collected in September 2008 and of 95 ± 2.6 % for those collected in February 2009.

Almost all the seeds that germinated by hypo-saline induction developed into a seedling: 99 ± 1.3 % for the September 2008 seeds and 100 % for the February 2009 seeds. The number of leaves, the leaf surface area, the number of roots and the total root length of the September 2008 and February 2009 seedlings were similar after 5 weeks of germination (Table II).

Seed germination and seedling development in full salinity conditions Similar to the hypo-saline conditions the Cymodocea nodosa fruits collected in August 2008 and maintained at salinity 37 did not germinate and died. The seeds colFig. 2. – Cumulative percentage of germination of Cymodocea nodosa fruits/ lected in September 2008 started to germiseeds collected in August 2008, September 2008 and February 2009 in the Bay nate in April 2009 (0.61 %) and those colof Pollença (Mallorca, Balearic Islands) after hypo-saline induction of germilected in February 2009 started to germinate nation (Salinity 10) or kept in full salinity (Salinity 37). The error bars indicate ± 1 standard error of the mean (n = 3). Notice the considerable different scale of in May 2009 (0.85 %). During the followY and X-axes. ing months the seeds germinated slowly and by the end of the experiment, in September achieved by the “hypo-saline” and “full salinity” seedlings was 2009, the cumulative percentage of seed germination was not tested. of 13 ± 2.8 % for the September 2008 seeds and of 6 ± 0.8 % for the February 2009 seeds (Fig. 2). The percentage of germinated seeds tended to be higher in the SepResults tember 2008 seeds than in the February 2009 seeds but the difference was never significant (Table I).

Seed germination in hypo-salinity conditions and The percentage of germinated seeds that developed seedling development into a seedling was of 61 ± 9.6 % for seeds collected in September 2008 and of 100 % for seeds collected in The Cymodocea nodosa seeds collected in August February 2009 group. The number of leaves, the leaf sur2008 did not germinate and eventually died. On the conface area, the number of roots and the total root length of trary, the seeds collected in September 2008 and Februthe seedlings of the September 2008 and February 2009 ary 2009 started to germinate 3 days after the reduction groups were similar five weeks after germination (Table of salinity, and by the 6th day about half of the seeds in II).

Vie Milieu, 2010, 60 (4)

310

M. DOMÍNGUEZ, E. INFANTES, J. TERRADOS

Table I. – Statistical differences of cumulative percentages recorded in seed germination between dates of collection (September 2008 and February 2009). Significant differences of the t-test results are in bold.

Salinity 10

Salinity 37

Times

t

df

P

Day 3 Day 6 Day 9 Day 12 Day 15 Day 21 Day 24 Day 27 April May June July August September

-5.11 -1.14 1.45 2.19 3.24 1.35 0.37 -0.31 1.00 0.93 0.66 0.92 1.83 2.32

4 4 4 4 4 4 4 4 4 4 4 4 4 4

0.01 0.32 0.22 0.09 0.03 0.25 0.73 0.77 0.37 0.41 0.54 0.41 0.14 0.08

Discussion Our results showed that the Cymodocea nodosa fruits collected in August 2008 were not mature enough to germinate and produce seedlings. However, in September, one month later, the fruits/seeds were already mature even if most of them were still attached to the shoot. Most of the seeds collected in September 2008 germinated and produced a seedling when the hypo-saline treatment was applied. The seeds collected in February 2009 showed a percentage of germination and an early seedling development similar to that of the seeds collected in September 2008 indicating that they were still mature and no loss of viability occurred after five months. Hence, the collection of mature fruits/seeds of C. nodosa in the meadow studied may be performed from September to February. Our results also show that specific care is not needed to maintain C. nodosa seeds viable in aquaria. They can be easily

maintained in good condition in full salinity seawater and ambient temperature for several months. Consequently it is possible to accumulate a stock of seeds collected during the period from September to February prior to any restoration labours. After germination, the seedlings can also be easily maintained in aquaria with full salinity seawater, at a temperature of 20-21 ºC and a photoperiod of 14:10 h light:dark for at least 5 weeks before planting. Therefore, it is also possible to avoid the harsh growing conditions associated with winter time and produce seedlings in aquaria with at least 3-4 leaves and 1-2 roots ready to be transplanted to the field when the environmental conditions for plant growth improved (spring or summer).

Our results corroborate previous studies that showed that the germination of Cymodocea nodosa seeds is to a large extent inhibited by the osmotic pressure of full salinity seawater and that a reduction of salinity may induce seed germination within a few days (Caye & Meinesz 1986, Caye et al. 1992). Although these studies concluded that seed germination will occur only if the external osmotic pressure is reduced, we showed that germination of seeds also occurred at full salinity seawater. Differences between locations in the percentage of seeds that germinate at full salinity have been detected previously (Buia & Mazzella 1991) and interpreted as indicators of genetic variability. Partial inhibition of seed germination by the external osmotic pressure provided by full salinity seawater provides a mechanism for the formation of a seed bank, which may be an advantage for C. nodosa populations to overcome years of poor reproductive output and to initiate population recovery after disturbances.

The induction of germination of Cymodocea nodosa seeds by a hypo-saline treatment may be used for restoration purposes as it allows maximizing the number of seedlings produced relative to the number of mature seeds collected and obtain the seedlings at a faster rate. Seedlings produced after the hypo-saline induction of germination showed a slightly lower vegetative development compared to that of the seedlings produced at constant, full salinity. However, the combined results of the seed

Table II. – Vegetative development of Cymodocea nodosa seedlings originated from seeds collected in September 2008 and February 2009 after 5 weeks (mean ± standard error, n = 3), and t-test results for the differences between dates of collection.

Variable

September 08

February 09

Mean

± SE

Mean

± SE

t

P

Salinity 10

Number of leaves Leaf area (cm2) Number of roots Total root length (cm)

3.2 1.2 1.6 1.1

0.1 0.2 0.2 0.2

3.1 1.2 1.6 1.6

0.1 0.03 0.1 0.4

0.62 0.58 -0.25 -1.28

0.57 0.59 0.81 0.27

Salinity 37

Number of leaves Leaf area (cm2) Number of roots Total root length (cm)

4.5 2.0 1.8 1.8

0.05 0.3 0.4 0.9

4.2 2.0 2.2 3.8

0.2 0.4 0.2 0.4

1.21 0.66 0.38 -0.44

0.29 0.54 0.72 0.68

Vie Milieu, 2010, 60 (4)

MATURITY OF CYMODOCEA NODOSA SEEDS

germination and the seedling development make it a good option and interesting alternative for C. nodosa restoration and research.

The use of early life stages in restoration is currently an important focus of interest in many parts of the world where seagrass populations have been damaged (Kirkman 1999, Reed et al. 1998, Balestri et al. 1998, Balestri & Bertini 2003, Holbrook et al. 2002, Bull et al. 2004, Bos & van Katwijk 2007) and, recently, studies with in vitro germinated seedlings (Zarranz et al. 2010) regarding to optimizing restoration of Cymodocea nodosa have been carried out. Each seedling is a physiologically independent small size unit that may be easily manipulated in field and laboratory experiments, contrasting to large size, clonal adult plants. For restoration purposes the use of laboratory-produced seedlings avoids damage of donor meadows and the small size of the seedlings facilitates their manipulation and planting. In addition, they guarantee the maintenance of a high level of genetic diversity in the restored area (Ruggiero et al. 2005).

In summary, the fruits/seeds of Cymodocea nodosa in the meadow studied in the Bay of Pollença (Mallorca, Balearic Islands) reach maturity in September even if they are still attached to the parental shoot and can be easily maintained in aquaria at full salinity, a temperature of 20-21ºC, and a photoperiod of 14 h:10 h light:dark for several months. Seed germination by hypo-saline induction (salinity of 10 during 2-4 weeks) reached germination percentages higher than 90 % and the seedlings produced developed 3-4 leaves and 1-2 roots within five weeks when growing in full salinity (37) seawater.

Acknowledgements.- This work was supported by the Ministerio de Medio Ambiente of Spain (grant number 116/ SGTB/2007/1.3). We thank FJ Medina-Pons for his SCUBA diving help in seed collections. We also thank reviewers’ comments that greatly improved the manuscript.

References Balestri E, Bertini S 2003. Growth and development of Posidonia oceanica seedlings treated with plant growth regulators: possible implications for meadow restoration. Aquat Bot 76: 291-297.

Balestri E, Piazzi L, Cinelli F 1998. Survival and growth of transplanted and natural seedlings of Posidonia oceanica (L.) Delile in a damaged coastal area. J Exp Mar Biol Ecol 228: 209-225.

Bos AR, van Katwijk MM 2007. Planting density, hydrodynamic exposure and mussel beds affect survival of transplanted intertidal eelgrass. Mar Ecol Prog Ser 336: 121-129.

Buia MC, Mazzella L 1991. Reproductive phenology of the Mediterranean seagrasses Posidonia oceanica (L.) Delile, Cymodocea nodosa (Ucria) Aschers., and Zostera noltii Hornem. Aquat Bot 40(4): 343-362.

Bull JS, Reed DC, Holbrook SJ 2004. An experimental evaluation of different methods of restoring Phyllospadix torreyi (Surfgrass). Restor Ecol 12: 70-79.

311

Calumpong HP, Fonseca MS 2001. Seagrass transplantation and other seagrass restoration methods. In Short FT & Coles RG eds, Global Seagrass Research Methods. Elsevier, Amsterdam: 425-443.

Cancemi G, Buia MC, Mazzella L 2002. Structure and growth dynamics of Cymodocea nodosa meadows. Sci Mar 66(4): 365-373.

Caye G, Meinesz A 1985. Observations on the vegetative development, flowering and seeding of Cymodocea nodosa (Ucria) Ascherson on the Mediterranean coasts of France.

Aquat Bot 22: 277-289.

Caye G, Meinesz A 1986. Experimental study of seed germination in the seagrass Cymodocea nodosa. Aquat Bot 26: 79-87.

Caye G, Bulard C, Meinesz A, Loquès F 1992. Dominant role of seawater osmotic pressure on germination in Cymodocea nodosa. Aquat Bot 42: 187-193.

Christensen PB, Díaz Almela E, Diekmann O 2004. Can transplanting accelerate the recovery of seagrasses? In Borum J, Duarte CM, Krause-Jensen D & Greve TM, eds, European seagrasses: an introduction to monitoring and management.

The M&MS project, Amsterdam: 77-82.

Fonseca MS, Kenworthy WJ, Thayer GW 1998. Guidelines of the conservation and restoration of seagrasses in the United States and adjacent waters. NOAA Coastal Ocean Program Decision Analysis Series No. 12. NOAA Coastal Ocean Office, Silver Spring, MD. Hemminga M, Duarte CM 2000. Seagrass Ecology. Univ Cambridge Press.

Holbrook SJ, Reed DC, Bull JS 2002. Survival experiments with outplanted seedlings of surfgrass (Phyllospadix torreyi) to enhance establishment on artificial structures. ICES J Mar Sci 59: S350-S355.

Kenworthy WJ, Wyllie-Echevarria S, Coles RG, Pergent G, Pergent-Martini C 2006. Seagrass conservation biology: An interdisciplinary science for protection of the seagrass biome.

In Larkum AWD, Orth RJ & Duarte CM eds, Seagrasses: Biology, Ecology and Conservation. Springer, Dordrecht: 595-623.

Kirkman H 1999. Pilot experiments on planting seedlings and small seagrass propagules in Western Australia. Mar Poll Bull 37: 460-467.

Orth RJ, Carruthers TJB, Dennison WC, Duarte CM, Four­ qurean JW, Heck Jr KL, Hughes AR, Kendrick GA, Kenworthy WJ, Olyarnik S, Short FT, Waycott M, Williams SL 2006.

A global crisis for seagrass ecosystems. BioScience 56: 987996.

Pirc H, Buia MC, Mazzella L 1986. Germination and seed development of Cymodocea nodosa (Ucria) Ascherson under laboratory conditions and “in situ”. Aquat Bot 26: 181-188.

Procaccini G, Piazzi L 2001. Genetic polymorphism and transplantation success in the Mediterranean seagrass Posidonia oceanica. Restor Ecol 9(3): 332-338.

Procaccini G, Buia MC, Gambi MC, Pérez M, Pergent G, Pergent-Martini C, Romero 2003. The seagrasses of the Western Mediterranean. In Green EP & Short FT eds, World Atlas of Seagrasses. UNEP World Conservation Monitoring Centre, Univ California Press, Berkeley: 48-58.

Reed DC, Holbrook SJ, Solomon E, Anghera M 1998. Studies on germination and root development in the surfgrass Phyllospadix torreyi: implications for habitat restoration. Aquat Bot 62: 71-80.

Vie Milieu, 2010, 60 (4)

312

M. DOMÍNGUEZ, E. INFANTES, J. TERRADOS

Ruggiero MV, Reusch TBH, Procaccini G 2005. Local genetic structure in a clonal dioecious angiosperm. Mol Ecol 14: 957-967.

Terrados J 1993. Sexual reproduction and seed banks of Cymodocea nodosa (Ucria) Ascherson meadows on the southeast Mediterranean coast of Spain. Aquat Bot 46: 293-299.

Waycott M, Duarte CM, Carruthers TJB, Orth RJ, Dennison WC, Olyarnik S, Calladine A, Fourqurean JW, Heck Jr KL, Hughes AR, Kendrick GA, Kenworthy WJ, Short FT, Williams SL 2009. Accelerating loss of seagrasses across the globe threatens coastal ecosystems. Proc Natl Acad Sci USA 106: 12377-12381.

Zarranz ME, González-Henríquez N, García-Jiménez P, Robaina R 2010. Restoration of Cymodocea nodosa (Ucria) Ascherson seagrass meadows through seed propagation: seed storage and influences of plant hormones and mineral nutrients on seedling growth in vitro. Bot Mar 53: 439-448.

Vie Milieu, 2010, 60 (4)

Received March 29, 2010 Accepted November 19, 2010 Associate Editor: G Tita

RESEARCH ARTICLE

Experimental Evaluation of the Restoration Capacity of a Fish-Farm Impacted Area with Posidonia oceanica (L.) Delile Seedlings Marta Domínguez,1,2 David Celdr´an,3 Ana Mu˜noz-Vera,4 Eduardo Infantes,1 Pedro Martinez-Ba˜nos,4 Arnaldo Marín,3 and Jorge Terrados1 Abstract Marine aquaculture is an activity that has induced severe local losses of seagrass meadows along the coastal areas.

The purpose of this study was to evaluate the capacity of an area degraded by fish-farm activities to support Posidonia oceanica seedlings. In the study site, a bay in the southeast coast of Spain where part of a meadow disappeared by fishfarm activities, seedlings inside mesh-pots were planted in three areas. Two plots were established in each area, one in P. oceanica dead matte and another inside a P. oceanica meadow. To evaluate if sediment conditions were adequate for the life of the seedlings, half of them were planted in direct contact with the sediment and the other half were planted above the surface of the sediment in each plot.

Monitoring during 1 year showed that there were large

Introduction Seagrasses are key components of shallow coastal ecosystems for their contribution to biological productivity and the maintenance of biodiversity, the control of water quality, and the protection of the shoreline (Hemminga & Duarte 2000). Seagrass populations have decreased due to the impact caused by human activities such as coastal development, pollution, and trawling (Boudouresque et al. 2009). The increasing demand for marine products by human population has promoted the growth of marine aquaculture, which is the fastest growing animal food-producing sector (FAO 2009). Coastal fish aquaculture has contributed to the decline of seagrass populations (Holmer et al. 2003a). For this reason, there is a growing concern about the environmental impact of marine aquaculture 1 IMEDEA (CSIC-UIB) Instituto Mediterr´aneo de Estudios Avanzados, C/Miquel Marqu´es 21, 07190 Esporles, Mallorca, Spain 2 Address correspondence to M. Domínguez, email mdominguez@imedea. uib-csic.es 3 Departamento de Ecología e Hidrología, Facultad de Biología, Universidad de Murcia, Campus Universitario de Espinardo, 30100 Murcia, Spain 4 Contesma & Comprotec SLP (C&C Medio Ambiente), C/Antonio Oliver 17, 30204 Cartagena, Spain

© 2011 Society for Ecological Restoration International doi: 10.1111/j.1526-100X.2010.00762.x

Restoration Ecology

differences in seedling survival between the dead matte and the P. oceanica meadow. While seedlings planted in dead matte had a high survivorship after 1 year (75%), seedlings planted in P. oceanica progressively died (survivorship of 20% after 1 year). The average leaf length of the seedlings surviving in the two substrata was not different, but the leaf area per seedling was lower in the seedlings growing inside the P. oceanica meadow during most part of the year. Seedling survivorship and vegetative development were not affected by the level of planting and suggest that the sediment conditions are adequate for the life of P. oceanica seedlings.

Key words: fish-farm, planting level, Posidonia oceanica,

seedlings, substratum development.

type,

survival,

vegetative

(Naylor et al. 2000) and as a result severe environmental impacts have been documented regarding net cage production (Hall et al. 1990, 1992; Holmer & Kristensen 1992; Karakassis et al. 2000; Holmer et al. 2002, 2003a).

The loading of organic matter coming from fish feces and uneaten fish food is considered the main driver of habitat degradation (Pergent-Martini et al. 2006; Holmer et al.

2008). Fish-farm-derived organic matter inputs to the sediment increase sulfate reduction rates (Holmer & Frederiksen 2007) and promote the invasion of the rhizomes and roots by sulfide which is associated with reduced growth and increased mortality of seagrasses (Calleja et al. 2007; Frederiksen et al.

2007, 2008). The nutrients released by the mineralization of organic matter may also cause local nutrient enrichments and the increase of the epiphytes (P´erez et al. 2008) and, likely, of their palatability to herbivores. Shading by epiphytic overgrowth has been considered another process involved in seagrass losses in fish-farm impacted areas (Pergent-Martini et al. 2006).

Posidonia oceanica (L.) Delile forms extensive meadows in the Mediterranean Sea but its surface has declined by 5–20% during the last century. Coastal development, pollution, and trawling are considered the main causes of this loss (Boudouresque et al. 2009). Coastal fish aquaculture

1

Survival of Posidonia oceanica Seedlings

´ Figure 1. Hornillo Bay, Aguilas, Spain. Present distribution of the Posidonia oceanica meadow is represented by light gray. The three study areas are represented by dark gray rectangles. Spatial arrangement of the seedlings planted in each area with indication of planting level.

is a fast growing activity in the Mediterranean that also threats P. oceanica (Holmer et al. 2003b). Indeed, reports of P. oceanica losses associated to fish farming are numerous (Delgado et al. 1997; Pergent et al. 1999; Ruiz et al. 2001; Cancemi et al. 2003; Pergent-Martini et al. 2006; Díaz-Almela et al. 2008) and may be observed within months of the initiation of fish-farm activities (Marb`a et al. 2006; P´erez et al. 2007).

The changes in sediment biogeochemistry may persist for some years even if organic inputs are stopped and sustain further declines of P. oceanica (Delgado et al. 1999; Holmer et al. 2003a). The recovery of P. oceanica is also limited by its slow growth (Marb`a & Duarte 1998; Kendrick et al. 2005), low flowering (Díaz-Almela et al. 2006), and high rates of fruit abortion and predation (Balestri & Cinelli 2003).

Various studies have revealed the great importance that careful habitat selection has for seagrass transplantation (van Katwijk 2009) and pilot studies have been carried out to develop effective restoration methods (Kirkman 1998). The aim of this study was to evaluate the capacity of one area impacted by fish-farming activities to support the life of seagrasses under the current environmental conditions. To that end, we planted in the impacted area P. oceanica seedlings cultured in the laboratory from fruits collected in nearby beaches. The seedlings were planted inside pots anchored to substratum and their vegetative development was followed for a year.

2

Methods Study Site

´ The study was performed in Hornillo Bay, Aguilas, Spain (Fig. 1), a small bay that once supported a continuous Posidonia oceanica meadow between the depths of 3–5 and 25 m (Calvín et al. 1989). Fish farming (Seriola dumerili (Risso) L., Sparus auratus (L.), and Dicentrarchus labrax (L.)) started in 1989 on the west side of the bay and by 1998, 11 ha (28% of total surface) of the P. oceanica meadow had been completely lost while another 10 ha (25% of total surface) were degraded (Ruiz et al. 2001). Fish farming caused a 10% reduction in light availability and increases in dissolved inorganic phosphorus (from 0.1 to 0.2–0.6 μM) and ammonium (from 3–4 to 30–40 μM), and also of the organic matter content of the sediment (from 0.5–0.9% DW to 1.8–2.4% DW). The fishfarm was progressively dismantled in three phases between 2000 and 2003. Two years after the stopping of fish farming, the concentration of organic matter and acid-volatile reduced sulfur compounds was higher in the impacted area than in a non-impacted, control site (Ramos et al. 2003; Sanz-L´azaro and Marín 2006). Where P. oceanica survived, the size and leaf growth rate of the shoots were up to 75% lower than those of shoots in non-impacted control sites while the herbivore pressure inferred from grazing marks on the leaves was 20% higher at least (Ruiz et al. 2001). An in situ herbivore exclusion experiment indicates that grazing by sea urchins that were

Restoration Ecology

Survival of Posidonia oceanica Seedlings

(a)

(b)

Figure 2. (a) Photograph of mesh-pot with seedling inside, totally buried in sediment of a dead matte plot. (b) Photograph of mesh-pot with seedling inside planted above the surface of the sediment in a dead matte plot.

5â&#x20AC;&#x201C;59 times more abundant than previously reported for other non-impacted P. oceanica meadows was the major contributor to the loss of P. oceanica (Ruiz et al. 2009). Sea urchins have disappeared since then and no P. oceanica recruitment has been observed in the impacted area where the substratum is made of a dead matte, the characteristic structure that the rhizomes and roots of P. oceanica form.

Experimental Design

Posidonia oceanica seedlings were used to evaluate the capacity of the fish-farming impacted area to be restored under current environmental conditions. Seedlings were chosen for several reasons: (1) they are physiologically independent planting units of small size and this facilitates handling; (2) given the distance to the nearest meadows and the slow growth of the rhizomes, the recruitment of P. oceanica into the impacted area should proceed through seed/seedling recruitment; (3) the use of seedlings makes the estimation of plant mortality straightforward; (4) avoids the damage caused to donor meadows if fragments of adult plants were used; and also (5) the interference that the disturbance caused to the fragments after extraction might exert on the response of the planting units. The seedlings were cultured in the laboratory from fruits collected in nearby beaches that were maintained in

Restoration Ecology

aquaria where they germinated and produced five to six leaves and one to three roots within 2â&#x20AC;&#x201C;3 months. To assure the persistence of the seedlings at the experimental site and minimize damage, to roots specially, during handling and planting each seedling was grown inside a plastic mesh-pot filled with fiberglass wool (Fig. 2). Each pot was anchored to the substratum by a 60-cm corrugated iron bar and a 30-cm galvanized iron spike fixed in opposite sides.

Two factors were evaluated: substratum type (dead matte versus P. oceanica) and planting level (above vs. below sediment surface). We wanted to compare the responses of seedlings when planted in a substratum where fish farming caused the complete loss of P. oceanica and no recruitment has occurred since then to those planted in a substratum where P. oceanica persisted and currently supports a well-developed, 100% cover meadow. As sediment conditions, particularly its state of reduction, may affect seagrass development (Calleja et al. 2007; Frederiksen et al. 2007, 2008), we also wanted to compare the performance of seedlings planted in direct contact with the sediment to that of seedlings planted above the surface of the sediment and consequently in a more oxic substratum. We planted the seedlings following a two-way factorial design with three replicates of each combination of factor levels. To that end, we selected three experimental areas (I, II, and III) along the current edge between the dead matte

3

Survival of Posidonia oceanica Seedlings

and the P. oceanica meadow (37◦ 24.6 N, 1◦ 33.4 W), and we established a “dead matte” plot and a “P. oceanica” plot at the two sides of the edge in each of the three areas (Fig. 1).

Distance between the consecutive areas was 25 m and water depth changed from 12.1 to 8.3 m. Each plot was made up of four parallel lines of six seedlings: the distance between consecutive seedlings in a line was 2 m and the distance between lines was also 2 m. Thus a total of 24 seedlings were planted in each plot and the planting level (either below or above the sediment surface) was assigned alternatively to the seedlings along the lines, that is, there were 12 seedlings per planting level. Planting of the seedlings was performed on 28–29 July 2008.

The seedlings were haphazardly assigned to each combination of factor levels and the vegetative development of the seedlings at the time of planting was evaluated in the seedlings assigned to area I by counting the number of leaves of each seedling and the length and width of all the leaves. These measurements provide an estimation of the initial vegetative development of the seedlings in the experiment.

The cover of P. oceanica in each of the three “P. oceanica” plots was 90–100%. The density of shoots was estimated from counts of the number of shoots present in ten 30 × 30 cm quadrats placed haphazardly in each plot.

Six samples of the sediment were collected in each plot to characterize the sediment types as well as to estimate the content of organic matter. Sediments were sampled using clear acrylic tubes (length, 50 cm; internal diameter, 3.6 cm) that were inserted 40 cm into the sediment by hand. The upper 10-cm sediment layer was retained for analysis. Three core samples were used to determine the size distribution of sediment particles and were analyzed by first removing all organic matter by washing with hydrogen peroxide and deaggregating with sodium hexametaphosphate. Samples were oven-dried at 60◦ C for 48 hours and at 100◦ C for 24 hours.

Samples were then sieved (mesh sizes 2, 1, 0.5, 0.25, 0.125, and 0.063 mm). Sediment type and D50 were calculated using the grain size analysis program GRADISTAT v4.0 (Blott & Pye 2001). The other three core samples were used to determine the organic matter content as percentage of sediment dry weight and it was assessed by loss on ignition (450◦ C for 6 hours) on a sediment subsample.

The vertical attenuation coefficient of light in the water column (Kd : Kirk 1983) and the shading provided by the P. oceanica leaf canopy were estimated by measuring PAR photon flux density immediately below the sea surface and above the canopy, and inside the canopy just above the sediment using a LiCor LI-193SA spherical quantum sensor.

Photon flux density was measured at noon under clear skies and at three haphazardly chosen places within the meadow.

Shoot density counts, photon flux density measurements, and sediment sampling and analysis were performed at the initiation of the experiment (July 2008).

The plots were monitored every 3 months (October 2008, January 2009, May 2009, and July 2009) since the initiation of the experiment to assess seedling survival and vegetative development. Seedling survival at each time was expressed

4

as the percentage of seedlings surviving relative to the initial number of seedlings. The number of leaves of each seedling was counted and the length and width of each leaf of each seedling were measured to quantify the vegetative development achieved by the seedlings at each time.

Statistical Analysis

One-way ANOVA was used to evaluate the significance of the differences of P. oceanica shoot density between areas while differences in sediment properties (organic matter content, D50 ) between areas and substratum type (dead matte, P. oceanica) were evaluated using two-way ANOVA. If ANOVA was significant, multiple comparisons were performed using Tukey HSD test. Data transformation was applied if Cochran’s test indicated that variances were not homogeneous.

One-way ANOVA was also used to evaluate if the seedlings haphazardly assigned to each combination of factor levels had similar vegetative development. To that end, we compared the number of leaves, and the average leaf length and leaf area per seedling, of the seedlings (n = 12) assigned to each combination of factor levels in area I. Two-way factorial ANOVA was used to evaluate the significance of the effect of substratum type, planting level and the interaction of both factors on seedling survival and vegetative development at each time. Cochran’s test indicated that most data variances were homogeneous. The variances of leaf length of the seedlings measured in July 2009 were heterogeneous but data transformation (logarithm and square root) did not homogenize them. Significance level was set to 0.05.

Results Environmental Characterization

The density of Posidonia oceanica shoots varied between 124 and 187 shoots m−2 (Table 1) and was not different between the areas. The content of organic matter of the sediment varied between 2.75 and 3.95% in P. oceanica and between 2.2 and 3.48% of sediment dry weight in the dead matte (Table 1).

There were significant differences in the content of organic matter between dead matte and P. oceanica and the content of organic matter in the area III was significantly lower than that in areas I and II. The average grain size of the sediment was similar in areas I and II (predominance of medium-fine sand), whereas sediment of area III was coarser. Regarding the grain size, statistical analysis (Tukey HSD test) shows that dead matte plot in area III was significantly different from the other plots. The average grain size of the sediment was also different between substratum types (P. oceanica vs. dead matte) being smaller in P. oceanica.

PAR photon flux density at the top of the P. oceanica leaf canopy varied between 490 and 540 μmol m−2 s−1 . The light reaching the canopy was 30.3 ± 0.9% of that at the sea surface and Kd was 0.109 ± 0.003 m−1 . PAR photon flux density at the surface of the sediment inside the P. oceanica meadow

Restoration Ecology

Survival of Posidonia oceanica Seedlings

Table 1. Depth, shoot density, organic matter content of the sediment and size distribution of sediment particles in each of the experimental plots.

Variables Measured (July 2008)

Depth (m) Shoot density (number of shoots m−2 ) Organic matter in sediment (% DW)

D50 (μm)

Area I

Area II

Area III

P. oceanica

Dead matte

P. oceanica

Dead matte

P. oceanica

Dead matte

11.4 124 ± 14

12.1 —

9.6 187 ± 16

11 —

8.3 151 ± 32

9.5 —

3.9 ± 0.2

3.1 ± 0.2

3.5 ± 0.4

3.5 ± 0.0

2.75 ± 0.2

2.2 ± 0.1

Results of ANOVA

F : 1.963, p: 0.1599

Area: F : 13.309, p: 0.0009 Substrate: F : 5.407, p: 0.0384 Area × substrate: F : 1.821, p: 0.2038 253.9 ± 38.9 254.6 ± 16.8 242.2 ± 39.9 224.4 ± 7.2 369.8 ± 10.1 576.2 ± 26.4 Area: F : 49.866, p: 0.0000 Substrate: F : 8.450, p: 0.0132 Area × Substrate: F : 10.966, p: 0.0002

Mean ± 1 SE. Result of one-way (depth) and two-way ANOVA (organic matter in sediment and D50 ). Bold type indicates statistical significance at p < 0.05.

varied from 14 to 60 μmol m−2 s−1 . The shading provided by the leaf canopy was high, for only 5.8 ± 2.4% of the light at the top of the canopy reached the sediment surface inside the meadow.

Seedling Performance

During the first few months of the experiment, from July 2008 to October 2008, the seedlings experienced low mortality (Fig. 3). An effect of the substratum type on the survival of the seedlings became evident by May 2009 and confirmed by two-way factorial ANOVA. The percentage of seedlings surviving in the dead matte was higher than that in P. oceanica both in May 2009 (F = 26.036, p = 0.0009) and in July 2009 (F = 68.762, p < 0.0001) (Fig. 3). The percentage of seedlings surviving in dead matte was 75 ± 4.3% in May 2009 and 75 ± 4.8% in July 2009 but only 37.5 ± 6.7% and 22.2 ± 4.1%, respectively, in P. oceanica. Seedling survival was not affected by the planting level.

No differences of vegetative development of the seedlings assigned to each combination of factor levels in area I were detected (one-way ANOVA). The average number of leaves per seedling in this area at the initiation of the experiment was 8.4 ± 0.3 leaves, the average leaf length was 4.3 ± 0.1 cm and the average leaf area per seedling was 17.9 ± 0.8 cm2 .

The number of leaves per seedling decreased after 3 months in all combinations of factor levels and this reduction was maintained during the rest of the study (Fig. 4a). The number of leaves per seedling in dead matte after October 2008 was from 6.3 ± 0.22 to 5.6 ± 0.16, always higher than in P. oceanica (from 4.3 ± 0.18 to 3.4 ± 0.39) (October 2008: F = 34.377, p = 0.0004; January 2009: F = 24.340, p = 0.0011; May 2009: F = 50.629, p = 0.0001; July 2009: F = 53.545, p = 0.0001). The number of leaves per seedling was not affected by the level of planting of the seedlings.

Restoration Ecology

Figure 3. Survival percentage of Posidonia oceanica seedlings for each combination of factor levels. PA, Posidonia oceanica —above sediment surface; PB, Posidonia oceanica —below sediment surface; MA, Dead matte—above sediment surface; MB, Dead matte—below sediment surface.

The average length of the leaves increased after plantings in all combinations of factor levels (Fig. 4b) reaching a maximum of 13.0 ± 0.5 cm in May 2009. No effect of substratum type and planting level was detected. In July 2009, the average leaf length tended to decrease in the seedlings planted in dead matte but the difference with those planted in P. oceanica was not significant (F = 4.244, p = 0.0733).

The leaf area per seedling increased progressively in all combinations of factor levels since the initiation of the

5

Survival of Posidonia oceanica Seedlings

(a)

experiment, reaching the maximum surface in May 2009 (Fig. 4c). The seedlings planted in dead matte had a higher leaf area than those planted in P. oceanica in October 2008 (F = 13.266, p = 0.0066) and May 2009 (F = 6.294, p = 0.0364) (Fig. 4). The reduction of leaf area of the seedlings planted in dead matte from May 2009 to July 2009 turned the effect of substratum to not significant in July 2009. The level of planting had no effect on the leaf area of the seedlings at any of the measurement dates.

Discussion

(b)

(c)

Figure 4. Vegetative development of Posidonia oceanica seedlings for each combination of factor levels. (a) Number of leaves per seedling, (b) average leaf length per seedling, (c) average leaf area per seedling.

Initial value (M) of the seedlings measured in July 2008 are represented by the horizontal dashed line. The significance of factors is indicated by asterisks: ∗ p < 0.05; ∗∗ p < 0.01; ∗∗∗ p < 0.001. PA, Posidonia oceanica —above sediment surface; PB, Posidonia oceanica —below sediment surface; MA, Dead matte—above sediment surface; MB, Dead matte—below sediment surface.

6

Our results show that seedlings of Posidonia oceanica produced in the laboratory from beach-cast fruits and planted in a fish-farm impacted area were able to survive for 1 year.

This suggests that the current environmental conditions in the impacted area make the restoration of P. oceanica feasible using this type of planting unit.

We found that dead matte in this area is the most adequate substratum for the development of P. oceanica seedlings, at least during their first year of life. After an initial mortality during the first 3 months, the percentage of seedlings surviving in dead matte remained stable for the rest of the experiment yielding a 75% survival after 1 year. On the contrary, the seedlings planted inside the P. oceanica meadow died progressively during the experiment so that the percentage of seedlings surviving at the end of experiment was only about 20%. The shading imposed by the leaf canopy is a possible reason to explain the low survivorship of the seedlings planted inside the P. oceanica meadow. Our measurements of the percentage of light reaching the top of the leaf canopy that is absorbed by it (around 95%) indicate that there is not much light available for a small seedling growing inside the leaf canopy. PAR photon flux density inside the meadow was lower than 60 μm−2 s−1 and this light level is known to limit the photosynthetic rate of this species (Alcoverro et al. 1998).

In the studied meadow, only 5% of the light reaching the top of the leaf canopy is available at the lower level which means that less than 1% of surface irradiance reaches the bottom. The shading we measured might be considered as maximal because it is in summer when the P. oceanica leaf canopy reaches the maximum development (Alcoverro et al. 1995). Shading will likely be less in other seasons what might explain why the mortality of the planted seedlings in P. oceanica has not been completed.

The use of early life stages in restoration is currently an important focus of interest in many parts of the world where seagrass populations have been damaged (Kirkman 1998; Reed et al. 1998; Holbrook et al. 2002; Balestri & Bertini 2003; Bull et al. 2004; Bos & van Katwijk 2007). One of the main problems of using early life stages for transplanting is the low capacity to anchor in the substratum during the first months, and consequently the high loss of seedlings (Cooper 1979; Balestri et al. 1998; Balestri & Bertini 2003). Methods have been developed for facilitating the settlement of seagrass propagules, either enhancing early root growth (Reed et al.

Restoration Ecology

Survival of Posidonia oceanica Seedlings

1998) or using artificial structures (Kirkman 1998; Harwell & Orth 1999; Holbrook et al. 2002; Bull et al. 2004). In Australia, seedlings of Posidonia and sprigs of Amphibolis were used in a pilot experiment to discover the ability of some propagules to grow after being planted (Kirkman 1998).

This study showed that after about 1 year, plantings of single seedlings of Posidonia were not successful in either of the artificial media used (Jiffy pots and Growool blocks) because the rhizomes do not spread quickly. The planting system we used facilitated the establishment of the seedlings and assured their persistence during 1 year. It seems feasible, therefore, to consider the restoration of the meadow through the planting of seedlings inside the mesh-pots. Furthermore, this structure permits the direct contact of the roots with the sediment, allowing to evaluate if sediment conditions are adequate for seedling development.

Our results show that sediment conditions in the fish-farm impacted area do not seem detrimental for the development of P. oceanica seedlings because (1) seedling survival after 1 year was similar to that in other locations (Balestri et al.

1998; Piazzi et al. 1999) and (2) there was no effect of the planting level and, therefore, of the degree of contact of the seedlings with the sediment on seedling survival.

The content of organic matter of the sediment increased as a result of fish farming from less than 1% of sediment dry weight to 1.8–2.4% after 5 years of farming activities (from 1989 to 1994, Ruiz et al. 2001). The organic matter content of the sediment at the initiation of our study varied from 2.2 to 3.9% in July 2008 which indicates that the organic matter content of the sediment 5 years after the cease of farming activities has not been reduced to values considered as reference, indicative of non-impacted conditions in the area (<1–2% of sediment dry weight (Ruiz et al. 2001; Ramos et al. 2003)).

In spite of this, the survival and vegetative development of the seedlings did not seem to be affected negatively by these contents of organic matter. This is not surprising as the negative effects of high levels of organic matter in the sediment on seagrass performance have been related more to the accumulation of reduced sulfur in the sediment as a consequence of increased sulfate reduction rates than to high organic matter levels per se (Holmer & Frederiksen 2007).

The few planted seedlings surviving inside the P. oceanica meadow tended to have longer leaf lengths in July 2009 than those planted in dead matte. The observation of bite marks on the leaves of the seedlings growing in the dead matte but not in those living inside the P. oceanica meadow suggests that this reduction of leaf length in summer might be due to fish herbivory. The seedlings living in dead matte might be more easily located by herbivore fish than those inside the leaf canopy and probably support more intense pressure from herbivores. Although sea urchin herbivory was the main driver of P. oceanica loss in the fish-farm impacted area in Hornillo Bay (Ruiz et al. 2009) our results indicate that herbivore pressure is not an important process in the early life of P. oceanica seedlings but this conclusion requires further evaluation.

Restoration Ecology

Implications for Practice • Seedlings of seagrass may be used successfully in

experiments to evaluate if sediment and environmental conditions in disturbed areas are appropriate for meadow restoration.

• Plastic mesh-pots filled with fiberglass wool and firmly anchored to the substratum provide a reliable planting structure, allowing the persistence of seedlings for at least a year.

• Dead matte is a more adequate substratum for the development of planted seedlings than inside of a welldeveloped P. oceanica meadow.

Acknowledgments Funds for this study were provided by the research grant 116/SGTB/2007/1.3 of the Ministerio de Medio Ambiente, Rural y Marino and the Ministerio de Ciencia e Innovaci´on Espa˜nol, Programa Nacional de Formaci´on de Profesorado Universitario of Spain. We are very grateful to Club N´autico ´ de Aguilas for granting the use of its facilities to execute this study and we thank P. Navarrete and J. Martínez for their help in the monitoring of the seedlings.

LITERATURE CITED Alcoverro, T., C. M. Duarte, and J. Romero. 1995. Annual growth dynamics of Posidonia oceanica: contribution of large-scale versus local factors to seasonality. Marine Ecology Progress Series 120:203–210.

Alcoverro, T., M. Manzanera, and J. Romero. 1998. Seasonal and agedependent variability of Posidonia oceanica (L.) Delile photosynthetic parameters. Journal of Experimental Marine Biology and Ecology 230:1–13.

Balestri, E., and S. Bertini. 2003. Growth and development of Posidonia oceanica seedlings treated with plant growth regulators: possible implications for meadow restoration. Aquatic Botany 76:291–297.

Balestri, E., and F. Cinelli. 2003. Sexual reproductive success in Posidonia oceanica. Aquatic Botany 75:21–32.

Balestri, E., L. Piazzi, and F. Cinelli. 1998. Survival and growth of transplanted and natural seedlings of Posidonia oceanica (L.) Delile in a damaged coastal area. Journal of Experimental Marine Biology and Ecology 228:209–225.

Bos, A. R., and M. M. van Katwijk. 2007. Planting density, hydrodynamic exposure and mussel beds affect survival of transplanted intertidal eelgrass. Marine Ecology Progress Series 336:121–129.

Boudouresque, C. F., G. Bernard, G. Pergent, A. Shili, and M. Verlaque.

2009. Regression of Mediterranean seagrasses caused by natural processes and anthropogenic disturbances and stress: a critical review.

Botanica Marina 52:391–418.

Bull, J. S., D. C. Reed, and S. J. Holbrook. 2004. An experimental evaluation of different methods of restoring Phyllospadix torreyi (Surfgrass).

Restoration Ecology 12:70–79.

Calleja, M., N. Marb`a, and C. M. Duarte. 2007. The relationship between seagrass (Posidonia oceanica) decline and porewater sulfide pools in carbonate sediments. Estuarine, Coastal and Shelf Science 73:583–588.

Calvín, J. C., R. Ballester, I. Franco Navarro, A. M. Martínez Ingl´es, A. Marín, and A. Belando. 1989. Caracterizaci´on, valoraci´on ecol´ogica y determinaci´on de a´ reas a proteger en el litoral sumergido de la Regi´on de Murcia.

Agencia para el Medio Ambiente y la Naturaleza, Murcia.

7

Survival of Posidonia oceanica Seedlings

Cancemi, G., G. De Falco, and G. Pergent. 2003. Effects of organic matter input from a fish farming facility on a Posidonia oceanica meadow.

Estuarine, Coastal and Shelf Science 56:961–968.

Cooper, G. 1979. Jardinier de la mer. In: Cooper, G., Association-Fundation Marin-pˆecheur pour la Reconquˆete des Milieux Naturales D´etruits, Cahier 3: 67 pp.

Delgado, O., A. Grau, S. Pou, F. Riera, C. Massuti, M. Zabala, and E. Ballesteros. 1997. Seagrass regression caused by fish cultures in Fornells Bay (Menorca, Western Mediterranean). Oceanologica Acta 20:557–563.

Delgado, O., J. M. Ruiz, M. P´erez, J. Romero, and E. Ballesteros. 1999.

Effects of fish farming on seagrass (Posidonia oceanica) in a Mediterranean Bay: seagrass decline after organic loading cessation. Oceanologica Acta 22:109–117. ´ Díaz-Almela, E., N. Marb`a, E. Alvarez, E. Balestri, J. M. Ruiz, and C. M. Duarte. 2006. Patterns of seagrass (Posidonia oceanica) flowering in the Western Mediterranean. Marine Biology 148:723–742. ´ Díaz-Almela, E., N. Marb`a, E. Alvarez, R. Santiago, M. Holmer, A. Grau, et al. 2008. Benthic input rates predict seagrass (Posidonia oceanica) fish farm-induced decline. Marine Pollution Bulletin 56:1332–1342.

FAO (Food and Agriculture Organization of the United Nations). 2009. The state of world fisheries and aquaculture 2008. ISBN 978-92-5-106029-2.

Frederiksen, M. S., M. Holmer, E. Díaz-Almela, N. Marb`a, and C. M. Duarte.

2007. Sulfide invasion in the seagrass Posidonia oceanica at Mediterranean fish farms: assessment using stable sulfur isotopes. Marine Ecology Progress Series 345:93–104.

Frederiksen, M. S., M. Holmer, M. P´erez, O. Invers, J. M. Ruiz, and B. B. Knudsen. 2008. Effect of increased sediment sulphide concentrations on the composition of stable sulphur isotopes (δ 34 S) and sulphur accumulation in the seagrasses Zostera marina and Posidonia oceanica.

Journal of Experimental Marine Biology and Ecology 358:98–109.

Hall, P. O. J., L. G. Anderson, O. Holby, S. Kollberg, and M. Samuelsson.

1990. Chemical fluxes and mass balances in a marine fish cage farm.

I. Carbon. Marine Ecology Progress Series 61:61–73.

Hall, P. O. J., O. Holby, S. Kollberg, and M. Samuelsson. 1992. Chemical fluxes and mass balances in a marine fish cage farm. IV. Nitrogen. Marine Ecology Progress Series 89:81–91.

Harwell, M. H., and R. J. Orth. 1999. Eelgrass (Zostera marina L.) seed protection for field experiments and implications for large-scale restoration.

Aquatic Botany 64:51–61.

Hemminga, M. A., and C. M. Duarte. 2000. Seagrass ecology. Pages 298.

Cambridge University Press, Cambridge, U.K. Holbrook, S. J., D. C. Reed, and J. S. Bull. 2002. Survival experiments with outplanted seedlings of surfgrass (Phyllospadix torreyi ) to enhance establishment on artificial structures. ICES Journal of Marine Science 59:S350–S355.

Holmer, M., M. Argyrou, T. Dalsgaard, R. Danovaro, E. Diaz-Almela, C. M. Duarte, et al. 2008. Effects of fish farm waste on Posidonia oceanica meadows: synthesis and provision of monitoring and management tools.

Marine Pollution Bulletin 56:1618–1629.

Holmer, M., C. M. Duarte, A. Heilskov, B. Olesen, and J. Terrados. 2003a.

Biogeochemical conditions in sediments enriched by organic matter from net-pen fish farms in the Bolinao area, Philippines. Marine Pollution Bulletin 46:1470–1479.

Holmer, M., C. M. Duarte, and N. Marb`a. 2003b. Sulfur cycling and seagrass (Posidonia oceanica) status in carbonate sediments. Biogeochemistry 66:223–239.

Holmer, M., and M. S. Frederiksen. 2007. Stimulation of sulphate reduction rates in Mediterranean fish farm sediments inhabited by the seagrass Posidonia oceanica. Biogeochemistry 85:169–184.

Holmer, M., and E. Kristensen. 1992. Impact of marine fish cage farming on sediment metabolism and sulfate reduction of underlying sediments.

Marine Ecology Progress Series 80:191–201.

8

Holmer, M., N. Marb´a, J. Terrados, C. M. Duarte, and M. D. Fortes. 2002.

Impacts of milkfish (Chanos chanos) aquaculture on carbon and nutrient fluxes in the Bolinao area, Philippines. Marine Pollution Bulletin 44:685–696.

Karakassis, I., M. Tsapakis, E. Hatziyanni, K.-N. Papadopoulou, and W. Plaiti.

2000. Impact of cage farming of fish on the seabed in three Mediterranean coastal areas. ICES Journal of Marine Science 57:1462–1471.

Kendrick, G. A., C. M. Duarte, and N. Marb`a. 2005. Clonality in seagrasses, emergent properties and seagrass landscapes. Marine Ecology Progress Series 290:291–296.

Kirk, J. T. O. 1983. Light and photosynthesis in aquatic ecosystems. Cambridge University Press, Cambridge, U.K. Kirkman, H. 1998. Pilot experiments on planting seedlings and small seagrass propagules in Western Australia. Marine Pollution Bulletin 37:460–467.

Marb`a, N., and C. M. Duarte. 1998. Rhizome elongation and seagrass clonal growth. Marine Ecology Progress Series 174:269–280. ´ Marb`a, N., R. Santiago, E. Díaz-Almela, E. Alvarez, and C. M. Duarte. 2006.

Seagrass (Posidonia oceanica) vertical growth as an early indicator of fish-farm-derived stress. Estuarine, Coastal and Shelf Science 67:475–483.

Naylor, R. L., R. J. Goldburg, J. H. Primavera, N. Kautsky, M. C. M. Beveridge, J. Clay, C. Folke, J. Lubchenco, H. Mooney, and M. Troell.

2000. Effect of aquaculture on world fish supplies. Nature 405: 1017–1024.

P´erez, M., T. García, O. Invers, and J. M. Ruiz. 2008. Physiological responses of the seagrass Posidonia oceanica as indicators of fish farm impact.

Marine Pollution Bulletin 56:869–879.

P´erez, M., O. Invers, J. M. Ruiz, M. S. Frederiksen, and M. Holmer. 2007.

Physiological responses of the seagrass Posidonia oceanica to elevated organic matter content in sediments: an experimental assessment. Journal of Experimental Marine Biology and Ecology 344:149–160.

Pergent, G., S. Mendez, C. Pergent-Martini, and V. Pasqualini. 1999. Preliminary data on the impact of fish farming facilities on Posidonia oceanica meadows in the Mediterranean. Oceanologica Acta 22:95–107.

Pergent-Martini, C., C.-F. Boudouresque, V. Pasqualini, and G. Pergent. 2006.

Impact of fish farming facilities on Posidonia oceanica meadows: a review. Marine Ecology 27:310–319.

Piazzi, L., S. Acunto, and F. Cinelli. 1999. In situ survival and development of Posidonia oceanica (L.) Delile seedlings. Aquatic Botany 63:103–112.

Ramos, M., A. Marín, R. V. Barber´a, L. M. Guirao, A. C´esar, and Lloret J. 2003. Evoluci´on de las comunidades bent´onicas de la bahía de El Hornillo ´ (Aguilas, Murcia) (sureste de Espa˜na) finalizado el cultivo en jaulas flotantes de dorada Sparus auratus L., 1758 y lubina Dicentrarchus labrax L., 1758. Boletín Instituto Espa˜nol de Oceanografía 19:379–389.

Reed, D. C., S. J. Holbrook, E. Solomon, and M. Anghera. 1998. Studies on germination and root development in the surfgrass Phyllospadix torreyi : implications for habitat restoration. Aquatic Botany 62:71–80.

Ruiz, J. M., M. P´erez, and J. Romero. 2001. Effects of fish farm loadings on seagrass (Posidonia oceanica) distribution, growth and photosynthesis.

Marine Pollution Bulletin 42:749–760.

Ruiz, J. M., M. P´erez, J. Romero, and F. Tomas. 2009. The importance of herbivory in the decline of a seagrass (Posidonia oceanica) meadow near a fish farm: an experimental approach. Botanica Marina 52:449–458.

Sanz-L´azaro, C., and A. Marín. 2006. Benthic recovery during open sea fish farming abatement in Western Mediterranean, Spain. Marine Environmental Research 62:374–387. van Katwijk, M. M., A. R. Bos, V. N. de Jonge, L. S. A. M. Hanssen, D. C. R. Hermus, and D. J. de Jong. 2009. Guidelines for seagrass restoration: importance of habitat selection and donor population, spreading of risks, and ecosystem engineering effects. Marine Pollution Bulletin 58:179–188.

Restoration Ecology