Instant download Spatiotemporal changes of drought characteristics and their dynamic drivers in cana by Education Libraries

https://ebookmass.com/product/spatiotemporal-

More products digital (pdf, epub, mobi) instant download maybe you interests ...

Elsevier Weekblad - Week 26 - 2022 Gebruiker

https://ebookmass.com/product/elsevier-weekbladweek-26-2022-gebruiker/

Jock Seeks Geek: The Holidates Series Book #26 Jill Brashear

https://ebookmass.com/product/jock-seeks-geek-the-holidatesseries-book-26-jill-brashear/

The New York Review of Books – N. 09, May 26 2022

Various Authors

https://ebookmass.com/product/the-new-york-review-ofbooks-n-09-may-26-2022-various-authors/

Synthesis, Modelling and Characterization of 2D Materials and their Heterostructures (Micro and Nano Technologies) 1st Edition Eui-Hyeok Yang (Editor)

https://ebookmass.com/product/synthesis-modelling-andcharacterization-of-2d-materials-and-their-heterostructuresmicro-and-nano-technologies-1st-edition-eui-hyeok-yang-editor/

E-Paper Displays Bo Ru Yang

https://ebookmass.com/product/e-paper-displays-bo-ru-yang/

Multimaterial 3D Printing Technology Jiquan Yang

https://ebookmass.com/product/multimaterial-3d-printingtechnology-jiquan-yang/

Advanced Polyimide Materials Shi-Yong Yang

https://ebookmass.com/product/advanced-polyimide-materials-shiyong-yang/

Mathematical Physics with Differential Equations Yisong. Yang

https://ebookmass.com/product/mathematical-physics-withdifferential-equations-yisong-yang/

Pavement Analysis and Design (2nd Edition) Yang H. Huang

https://ebookmass.com/product/pavement-analysis-and-design-2ndedition-yang-h-huang/

AtmosphericResearch

journalhomepage: www.elsevier.com/locate/atmosres

Spatiotemporalchangesofdroughtcharacteristicsandtheirdynamicdrivers inCanada

YangYanga,ThianYewGana,⁎,XuezhiTana,b

a DepartmentofCivilandEnvironmentalEngineering,UniversityofAlberta,Edmonton,AlbertaT6G1H9,Canada

b DepartmentofWaterResourcesandEnvironment,SunYat-senUniversity,Guangzhou510275,PRChina

ARTICLEINFO

Keywords:

Droughtcharacteristics

StandardizedPrecipitationEvapotranspiration Index

Dynamiclinearmodel

Large-scaleclimatedrivers Canada

ABSTRACT

Therehasbeenagrowingconcernregardingimpactsofglobalwarmingondroughts,whichcanhavedevastatingeffectsontheenvironment,society,andeconomyofnationsworldwide.Droughtcharacteristicsinterms ofduration,frequency,area,andseverityareinvestigatedusingtheStandardizedPrecipitation EvapotranspirationIndex(SPEI)atseasonal(3-month)andannual(12-month)timescalesforCanadaover 1950–2016derivedfromtheClimaticResearchUnit(CRU)TS4.03griddeddata.Usingk-meansclustering, Canadaisdividedintofoursub-regions,eachwithdistinctdroughtcharacteristics.Next,theinfluenceofclimate driverssuchastheElNiño-SouthernOscillation(ENSO)andPacificDecadalOscillation(PDO)onregional droughtvariabilitywereexaminedusingaBayesianDynamicLinear(BDL)model.

Theresultsshowthatbetween1950and2016(1)therehasbeenaprevalentdryingtrendinsouthwestern Canadaduringwinter;(2)changesinmaximumdroughtdurationshaveoccurredwithdipolarpatterns,i.e., northernCanadahasexperiencedalongerdroughtdurationthansouthernCanada;(3)droughtfrequency,area, andseverityhavepredominantlyshownstatisticallysigniﬁcantdecreasingtrends,indicatingthatdroughtsin Canadahavegenerallybecomelesssevere;and(4)therelationshipsbetweenclimateanomaliesanddrought variabilityhavechangedovertime.DroughtsaregenerallymorenegativelycorrelatedtoENSOandPDOafter 1970s,butmorepositivelycorrelatedtotheAtlanticMulti-decadalOscillation(AMO)andArcticOscillation (AO)afterthe1980s.Theresultsprovideabetterunderstandingofthecharacteristicsofmeteorological droughtsinCanada,essentialforimprovingtheriskmanagementandmitigationstrategiesontheimpactof droughts.

1.Introduction

Inrecentdecades,therehavebeengrowingconcernsregarding droughtsgivenincreasedincidencesinthefrequencyandintensityof droughtshavebeenobservedworldwide,givingrisetobillionsofdollarsineconomiclosses.Moreover,theseverityofdroughtwilllikely increaseinmanyregionsbytheendofthe21stcentury(Dai,2013).For example,theseveredroughtacrossnorthernChinain1997resultedin 226daysofzero flowinthelowerreachoftheYellowRiver(Cong etal.,2009),andanotherdroughtin2000damaged>40million hectaresofcrops(Yuetal.,2014).Fromatrendanalysis, Joshietal. (2016) foundthatdroughtoccurrencesincreasedsignificantlyin northeastandcentralIndiaoverthesecondhalfofthe20thcentury.In Mexico, Escalante-SandovalandNuñez-Garcia(2017) projectedthat climatechangeimpactunderRCP4.5and8.5climatescenarioswould significantlyincreasethedurationandintensityofdroughts.

⁎ Correspondingauthor.

E-mailaddress: tgan@ualberta.ca (T.Y.Gan).

https://doi.org/10.1016/j.atmosres.2019.104695

Meanwhile,ananalysisofthespatialandtemporalcharacteristicsof droughtsinthecontinentalUnitedStates(US)alsorevealthatdrought duration,severity,andintensityhaveincreasedinthewesternand easternUSoverthepastcentury(Geetal.,2016).Onthewhole,the globalpercentageofdryareashasincreasedby1.74%perdecadefrom 1950to2008(Dai,2011a).

Droughtisacreepingrecurrentnaturalhazardbutunfortunately, thereisnoeasyidentifiableonsetsandterminations.Itisanextreme climaticphenomenonthatcanlastforweeks,months,orevenyears, andthenumberofpeopleaffectedbydroughtandthespatialextentof droughtaretypicallylargerthanthatofothernaturalhazardssuchas floodsandhurricanes.Asmentionedinthe fifthassessmentreport (AR5)oftheIntergovernmentalPanelonClimateChange(IPCC), droughtisdefinedas “aperiodofabnormallydryweatherlongenough tocauseaserioushydrologicalimbalance”,butitisarelativetermthat requiresthespecificationoftheappropriateprecipitation-related

Received13March2019;Receivedinrevisedform6August2019;Accepted29September2019

Availableonline15October2019

activity(IPCC,2014).

WilhiteandGlantz(1985) classifieddroughtsintofourcategories: a)meteorologicaldrought,b)hydrologicdrought,c)agricultural drought,andd)socio-economicdrought.Thisstudywillfocusonmeteorologicaldrought,whichisdefinedasadeficiencyofprecipitation acrossaregionduringacertaintimeperiod,asitisaprecursorof hydrological,agricultural,andsocio-economicdroughts.Toquantifya drought,droughtindiceshavebeenwidelyappliedbecausetheycan provideacomprehensivedelineationofdrought,makingdrought characteristicssuchasduration,severity,andintensitymeasurable (Jiangetal.,2014a,2014b).Moreimportantly,theseindicesallowfor directcomparisonbetweenregionsofdifferentclimaticconditions.The commonlyusedmeteorologicaldroughtindicesincludethePalmer DroughtSeverityIndex(PDSI)(Palmer,1965),theCropMoistureIndex (CMI)(Palmer,1968),theStandardizedPrecipitationIndex(SPI) (Mckeeetal.,1993),theReconnaissanceDroughtIndex(RDI)(Tsakiris etal.,2007),theStandardizedPrecipitationEvapotranspirationIndex (SPEI)(Vicente-Serranoetal.,2010),andtheWaterSurplusVariability Index(WSVI)(GocicandTrajkovic,2015).

Specifically,inCanada,droughtshavebeenidentifiedasoneofthe mostdamagingnaturaldisastersthathaveenormousimpactsonagriculture,industry,municipalservices,andhumanhealth,amongother sectors.Forexample,the2001–2002droughtintheCanadianPrairies (CP)sweptalmosttheentiresouthernpartofthecountry,anditwas oneofthetoptenworstdroughtsobservedovertheinstrumental period.InAlberta,cropproducerslost$413millionand$1.33billionin 2001and2002,respectively,whiletheestimatedreductionincrop productioninSaskatchewanaccountedforlossesof$925millionand $1.49billionin2001and2002,respectively(Wheatonetal.,2008). Furthermore,thedroughtthatoccurredinthespringandsummerof 2015isnoteworthyintermsofitsseverity,extent,andimpacts.Large areasinsouthernBritishColumbiawereassignedthehighestdrought rating,andtheAlbertagovernmentdesignatedtheprovinceasan AgriculturalDisasterArea(Szetoetal.,2016).Moreover,theextreme dryandwarmconditionsledtooneofthemostactiveandpersistent wildfireseasonsforwesternCanada,andsomeriversexperiencedtheir lowesthistoric flowlevelsin100years(CMOS,2016).

Theseaforementionedoverwhelmingdroughthazardshavebeenof majorconcerntoboththeCanadiangovernmentandthegeneralpublic inrecentdecades.Asaresult,manystudieshaveevaluatedthedryness orwetnessvariationsoverCanadausingdiﬀerentdroughtindices.For example, QuiringandPapakryiakou(2003) comparedfouragricultural droughtindices,whoseresultsrevealedthattheZ-index(derivedfrom thePalmermodel)wasbetterformeasuringagriculturaldroughtand predictingcropyieldduringgrowingseasonsfrom1920to1999. GobenaandGan(2013) whoassessedthetrendinsummermoisture availabilityinwesternCanadausingthesc_PDSIdetectedasigniﬁcant increasingdrynessandwetnessintheCPandwestofthecontinental divide,respectively,from1950to2003. Bonsaletal.(2012),afterassessingthevariabilityofsummerdroughtdurationandintensityinthe CPusingPDSIandSPI,advocatedthenecessitytoconsiderthepotential impactofclimatewarminginstudiesrelatedtofuturedroughtsofCP. VariousstudieshavebeenconductedregardingdroughtsintheCP,but so faronlylimitedresearchhasbeenconductedtoexaminethedrought characteristicsacrossCanada.Eventhough Asongetal.(2018) investigatedthecoherencebetweenCanadiandroughtandlarge-scale climateoscillationsusingwaveletanalysis,thedynamicandtimevaryingrelationshipsbetweenlarge-scaleclimatesignalsandCanadian droughtcharactershaveyettobeexamined.

Therefore,theobjectivesofthisstudyareto(1)assessthespatial andtemporalvariabilityofdroughtduration,frequency,area,andseverityinCanadafrom1950to2016;(2)identifysubregionscharacterizedbydistinctdroughtbehaviors;and(3)investigatethedynamiceﬀectsoflarge-scaleclimatedriversondroughtsinCanada.The paperisorganizedasfollows.Thedataandmethodsaredescribedin Section2,while Sections3and4 providetheresultsanddiscussion,

respectively.Summaryandconclusionsarepresentedin Section5.

2.Dataandmethods

2.1.Data

Inthisstudy,weusedtheSPEIfordroughtidentiﬁcationandevaluation.Themonthlyprecipitationandpotentialevapotranspiration datafrom1950to2016withaspatialresolutionof0.5°×0.5°were obtainedfromtheCRUTS4.03(http://www.cru.uea.ac.uk/data/ ).The CRUTSdatasethasbeenadoptedbyIPCCasithasundergonestrict timeuniformitytestsindatareconstruction,hasahighspatialresolution,andalongtimeseries(Wangetal.,2014).Potentialevapotranspiration(PET)istakenfromtheCRUTSdataset,inwhichPETwas estimatedusingtheFAO-56Penman-Monteithmethod(Harrisetal., 2014)consideredtobeanaccurateandrobustPETequation.

TheCRUdatasethasbeenwidelyappliedindroughtanalysis.For example, GuenangandMkankamKamga(2014) foundthattheSPI calculatedusingtheCRUexhibitssimilarresultswiththosecalculated fromstationdatainCameroonfrom1951to2005.BasedontheCRU dataset, Guoetal.(2018) examineddroughtsinCentralAsiaover 1966–2015anddetectedadryingtendencyafter2003. Huetal.(2018) foundthattheCRUdatasetperformedbetterincapturingdrought eventsincentralAsiathantheGlobalPrecipitationClimatologyCentre (GPCC)andWillmottandMatsuura(WM)precipitationdatasets.In addition, GobenaandGan(2013) usedgriddedvaporpressure,cloud cover,andwindspeeddatafromtheCRUdatasettocalculatePETin westernCanada.TheCRUdatasethasalsobeenusedtovalidatemodel simulatedprecipitationanddroughtcharacteristicsovertheCP (PaiMazumderetal.,2013). SwainandHayhoe(2015) usedtheCRU datatocomputehistoricalSPIindextoassesshowspringandsummer droughtwillchangeoverNorthAmerica.Bycomparingofeightgridded datasets(CANGRD,CRU-TS3.1,CRUTEM4.1,GISTEMP,GPCC,GPCP, HadCRUT3,andUDEL)witheightreanalysesdatasets(20CR,CFSR, ERA-40,ERA-Interim,JRA25,MERRA,NARR,andNCEP2)forthe CanadianArctic, Rapaić etal.(2015) foundthatCRUTS3.1agreedmost closelywithCANGRDinallseasons.

Tostudytheinfluenceofvariousclimateanomaliesondrought conditions,weinvestigatedtheinfluenceofclimateanomalieson droughtsofCanada,particularlyhowcertainlarge-scaleclimateoscillationshavecontributedtotheprecipitationandtemperaturevariabilityofCanada(Coulibaly,2006; Jiangetal.,2014a,2014b; Yang etal.,2019).Forexample,monthlyindicesofENSO,PDO(https:// www.esrl.noaa.gov/psd/gcos_wgsp/Timeseries/ ),Pacific-NorthAmericanOscillation(PNA),AtlanticMulti-decadalOscillation(AMO), Arctic Oscillation(AO)(https://www.esrl.noaa.gov/psd/data/ climateindices/list/),andsolaractivityrepresentedbysunspots (http://sidc.oma.be/index.php3 )wereselected.

2.2.Therunstheory

Diversedroughtconditionsarecharacterizedintermsoffrequency, duration,severity,andintensity.Toprovideananalyticalsolutionto droughtevents, Yevjevich(1967) proposedtherunstheoryforidentifyingdroughtparameters.Arunisdeﬁnedasaportionofthetime seriesofavariable,inwhichallvaluesareeitherlessorgreaterthanthe selectedtruncatedvalue;accordingly,itisreferredtoaseitheranegativeorapositiverun.Droughtfrequency(DF)isthenumberof droughteventsoccurredintheperiodconsidered;droughtduration (DD),expressedinmonths,isthedurationinwhichadroughtparameteriscontinuouslylessthanthecriticallevel;anddroughtseverity (DS)isthecumulativedeﬁcitofadroughtparameterbelowthecritical level.Detailsaboutvariablestodescribedroughtsaregivenin Mishra andSingh(2010)

2.3.Thestandardizedprecipitationevapotranspirationindex(SPEI)

PDSIisamilestoneinthedevelopmentofdroughtindicesasitintegratespriorprecipitation,moisturesupply,runoff,andevaporation demandatthesurfacelevel.Nevertheless,itsuffersfromhavinga fixed temporalscaleandanautoregressivecharacteristic,wherebyindex valueshavealong-termmemoryofpreviousconditions(VicenteSerranoetal.,2010),evenwiththeimprovedsc_PDSI.TheSPI,transformedfromaprecipitationtimeseriesusingaprobabilisticmethod, canbeestimatedatdifferenttimescaleswhichallowdroughteventsto becomparedacrosstimeandspace.However,SPIiscriticizedforonly usingprecipitationdatawithoutincludingotherimportantclimate variablessuchastemperature,waterbalance,andwindspeedthat couldalsocontributetodroughts.

ToaddresstheshortcomingofSPI,SPEIisdevisedtocombinethe multi-temporalnatureofSPIandthesensitivityofPDSIinprovidinga morecomprehensiveapproachtoinvestigatetheeffectofglobal warmingondroughtconditions.Inthisstudy,SPEIisbasedonmonthly climaticwaterbalance,thedifferencebetweenprecipitationandpotentialevapotranspiration.(Vicente-Serranoetal.,2010).SinceSPEI couldindicatewaterdeficitscausednotonlybyprecipitationshortages butalsobyexcessiveevapotranspiration,itisarobustindexfordrought monitoringandanalysisinthecontextofclimatechangeatglobaland regionalscales.DetailsaboutSPEIhavebeenextensivelydescribed (Vicente-Serranoetal.,2010).Onaglobalscale, Wangetal.(2014) studiedthechangingcharacteristicsofseveredroughtsbasedonthe SPEIandrevealedanoverallincreasingtrendintheglobalextentof droughtareasfrom1902to2008.

A3-monthSPEIindexforagivenmonthisbasedonthecurrent monthandprecedingtwomonths,e.g.,the3-monthSPEIinFebruary representstheDecember–January-FebruaryprecipitationandPET. Thus,the3-monthSPEIvaluesforFebruary,May,August,and Novemberwereusedintheseasonalanalysisforwinter,spring, summer,andautumn,respectively,whereasthe12-monthSPEIvalues forDecemberwereusedintheannualanalysis. Table1 showsthe classiﬁcationofdroughtbasedonSPEIvalues.

2.4.Trenddetection

Therank-basednon-parametricMann-Kendall(MK)trendtest (Kendall,1948; Mann,1945),recommendedbytheWorldMeteorologicalOrganization(WMO),wasappliedtodetecttrendsinautocorrelatedSPEIdata.Astheexistenceofautocorrelationinthedatacan affecttheprobabilityofdetectingtrends, HamedandRao(1998) recommendedsubtractingatrendestimatorfromtheoriginaltimeseries toevaluatetheautocorrelation.ThemodifiedMK(MMK)testwasthen appliedtotheSPEItimeseriesandtrendswith p-values<.05were consideredstatisticallysignificant.Trendmagnitudeswereevaluated usingthenon-parametricSen'sslopeestimator(Sen,1968)andabrupt changepointsweredetectedusingthePettitttest(Pettitt,1979).

2.5.Droughtregionalization

Duetothepossibleimpactsofcomplicatedtopographicalproperties anddiverseclimatetypeswithinCanada,thestudyareawasdivided intodiﬀerenthomogeneoussub-regions.Inthisstudy,k-means

Table1

ClassiﬁcationofSPEIdroughtcategory. Droughtcategory Indexvalue

clusteringwasconductedonseasonalSPEItopartitionCanadainto mutuallyexclusivesub-regions,andtheoptimalnumberofclusterswas determinedusingtheRpackage “NbClust” (Charrad etal.,2014).The packageprovides30validityindices(e.g.GapStatistic,Silhouette index,etc.)todeterminethenumberofclustersinadatasetbasedon diﬀerentdistancemeasuresandaggregationmethods.Foreachindex, NbClustproposestheoptimalnumberofclustersfromwhichwecompareallindicestochoosethebestnumberofclustersforthestudy. Typically,k-meansclusteringcanproducedistinctclusterswiththe aimof(1)minimizingvariabilitywithinclustersand(2)maximizing variabilityamongclusters. GongandRichman(1995) statedthatnonhierarchicalmethods,suchasthek-meansalgorithm,outperformed hierarchicalmethods(Ward'smethodandaveragelinkagemethod) whendealingwithprecipitationdata.Giventhelargenumberofgrids (over6000)inthedataset,k-meansclusteringprovidesamoredecent partitionresult.Toinvestigatethespatialandtemporalvariabilityof droughtsinPortugal, Santosetal.(2010) appliedk-meansclusteringto theSPIseriesandfoundthreeclusters,while Zhangetal.(2001) used thek-meansclusteringtoassessthespatialandtemporalcharacteristics ofheavyprecipitationeventsoverCanada.Inordertoconductapooled frequencyanalysisofdroughtsintheCP, SadriandBurn(2014) also appliedthek-meansclusteringtodividetheCPintosub-regions.More recently, Liuetal.(2018) partitionedsouthwestChinabyk-means clusteringandidentiﬁed3zonesbasedontheSilhouetteindextostudy thelong-termchangeofPETfrom1961to2013.

2.6.Bayesiandynamiclinearmodel

Toassessthetime-varyinginfluencesoflarge-scaleclimatedrivers ondrought,theBayesianDynamicLinear(BDL)modelfromtheR package “dlm” wasused(Petris,2010).Unlikethetraditionallinear regressionwhichhasstaticregressioncoefficients,BDLisabletomodel thedynamicandtime-varyingrelationshipsbetweenclimatedrivers anddrought,andthereforeithashigheraccuracythanthetraditional linearregression.BDLhasbeenwidelyappliedtoidentifythetimevaryingcharacteristicsoftimeseriesinhydrologyandclimateresearch. Forexample, Ciupaketal.(2015) usedBDLtomodelannualhydrographsand1-,2-,and3-dayleadtimestreamflowforecastinginPoland. Gaoetal.(2017) appliedBDLtoassessthedynamicinfluenceofENSO, IOD,PDO,NAO,andAMOonextremeregionalprecipitationinChina from1960to2014.However,onlylimitedstudieshaveemployedthe BDLmodelindroughtanalysis.

TheBDLmodelcanbedescribedasfollows(Gaoetal.,2017; Petris etal.,2013):

tt βtβtβt 1,,, 1 ,,, where yt istheresponsevariable(adroughtindex), xt isthecovariate(a climatepattern),and αt and βt arethedynamicinterceptandslope coeﬃcients,respectively,attime t.

3.Results

3.1.Homogenousregions

GivenCanadahasmanyclimaticregimesandwesternCanadais subjectedtosignificantorographiceffectsoftheCanadianRockies,it hasexperiencedawiderangeofspatialvariabilityinprecipitationand temperature.Resultsshowthatgridsassignedtothesamehomogenous regionsaregenerallylocatedtogether,whichlikelymeansthatSPEIis spatiallycoherent.From Fig.1,foursubregionswereidentifiedinCanada,namely,subregion1(S1)inthesouthandsoutheast,sub-region2 (S2)inArcticCanada,sub-region3(S3)inthewestcoast,andsubregion4(S4)innorth-central. Fig.2 showsthemeanannualSPEIseries

Fig.1. Studyareaandspatialdistributionoffoursubregions.

1950196019701980199020002010 SPEI−0.50.00.51.0(a)

1950196019701980199020002010 SPEI−1.5−0.50.51.5(c)

1950196019701980199020002010 SPEI−2−1012(b)

1950196019701980199020002010

Fig.2. AnnualSPEIseriesforfoursubregions:S1(a),S2(b),S3(c),andS4(d).TheredlineisthelineartrendandSisthetrendusingSen'sslope(signiﬁcantvalues areinbold).(Forinterpretationofthereferencestocolourinthis ﬁgurelegend,thereaderisreferredtothewebversionofthisarticle.)

forthefoursub-regionsofCanada.Amongthem,bothS1andS4showa signiﬁcantincreasingtrendin1950–2016,whichmeansthatdroughts inthesouthandnorth-centralCanadahasbecomelesssevereinrecent decades.

Usingk-meansclustering, Zhangetal.(2001) investigatedheavy precipitationeventsoverCanada(excludingthehighArctic)during 1900–98andidentifiedfourclustersofheavyrainfallforspringand twoclustersofheavyrainfallandsnowfallinotherthreeseasons.Accordingtotheirclusters,southernCanadaconsistsofsouthernBritish Columbia,southernCP,andsoutheasternCanada,whichgenerally agreeswithourClusterS1.Fromexaminingtheseasonalvariabilityof temperatureandprecipitationofCanada, Whitfieldetal.(2002) found thatthevariationsinclimategenerallyoccuratlargerthanthelocal scale,e.g.,mosttemperatureandprecipitationclustersspanacross morethanoneecozonesdespitetheirdifferencesinclimaticconditions. Thiscouldbeattributedtochangesinlarge-scaleatmosphericcirculationsaffectingregionsmuchlargerthansingleclimatologicalorecologicalzones.

3.2.Seasonaldroughttrends

Tofurtherstudythetemporalevolutionofdrought,theMMKtest wasemployedtoestimatetheSPEItrendineachseason. Fig.3 presents thespatialdistributionoftheMMKtrendstatisticoftheseasonalSPEI.

Ingeneral,droughtsshowremarkablespatialvariationsindifferent seasons;however,positivetrendsdominateovermajorpartsofCanada whilemostnegativetrendsinspring,summer,andautumnarenot statisticallysignificant,whichisconsistentwiththeresultsin Fig.2.In spring,scatteredsignificantincreasingtrendsarepredominantlyfound inS1andS4.Insummer,S3exhibitsmoresignificantincreasingtrends comparedwiththatin Fig.3(a).Furthermore,southernAlbertaand Saskatchewan(S1)haveexhibitedmorenegativetrendsinsummer comparedwiththatinspring.Moreover,themajorityofdecreasing trendsarelocatedinthesoutheast,althoughtheyarenotsignificant.In autumn,thespatialdistributionissimilartothatforspring,exceptthat therearemorenegativetrendsinAlberta,eastS2,andeastS3,whereas forwinter(Fig.3d),aprevalentdecreasingtrendappearslargelyin southwestCanada,whichmeansthatamongthefourseasons,winter hasanapparentdryingtendency.

From1950to1998, Zhangetal.(2000) foundthatprecipitationin Canadahasincreasedby5%to35%,butwithsigniﬁcantnegative trendsfoundinthesouthwestduringwinter(MekisandVincent,2011). Inaddition,warminginthesouthandwest(Bonsaletal.,2017)and coolinginthenortheastisevidentinwinterandspring.Together,these trendscouldhavecontributedtothetendencyofdryingdetectedin southwestCanadainwinterandwetterspringdetectedinnorthern Canada.Usingthesingularvaluedecomposition(SVD), Shabbarand Skinner(2004) reportedthatthe ﬁrstSVDpatternofthesummerPDSI

Fig.3. SpatialdistributionsoftheMMKtrendstatisticoftheseasonalSPEIat95%significancelevel.TheabbreviationsareP:positivetrend,SP:significantpositive trend,N:negativetrend,andSN:significantnegativetrend.Theyellowcurvesdenotetheclusterboundaries.(Forinterpretationofthereferencestocolourinthis figurelegend,thereaderisreferredtothewebversionofthisarticle.)

1952196219721982199220022012

Fig.4. TemporalvariationofquinquennialdroughtfrequencyforS1(a),S2(b),S3(c),andS4(d).1952denotes1950–1954and1962denotes1960–1964,etc.The redlineisthelineartrendandSisthetrendusingSen'sslope(signiﬁcantvaluesareinbold).(Forinterpretationofthereferencestocolourinthis ﬁgurelegend,the readerisreferredtothewebversionofthisarticle.)

inCanadaexhibitedbothinterannualanddecadalvariabilitywitha generalwettingtrendfrom1940to2002. Meynetal.(2010) analyzed trendsinwildfireandsummerdroughtinBritishColumbiafrom1920 to2000andrevealedthatthesummersc_PDSIexhibitedapositive trend,whichsignificantlycontributedtothedecreaseinwildfirefrequencyandareaburned.Similarly, GobenaandGan(2013) reported thatsummermoistureavailabilityinsouthernBritishColumbiahas shownasignificantincreasingtrendfrom1950to2003,whichcorrespondswiththeincreasingtrendin Fig.3(b).Inadditiontotheprevious research,ourresultsfurtheremphasizethatthedroughtriskinwinter hasworsened.

3.3.Droughtcharacteristics

3.3.1.Changesindroughtfrequency(DF)

Changesinthedroughtfrequency(SPEI<-1)overCanadawere investigatedfor5-year(lustrum)periods(1950–1954,1955–1959,etc.) in1950–2016,asshownin Fig.4.Overall,thereweremoredrought eventsduringthe firsttwolustrums(1950–1959)thanthelasttwo lustrumsperiods(1996–2015)forallregions.Additionally,S1andS4 experiencedsignificantdecreasingtrendsof 0.15and 0.32per decade,respectively. Spinonietal.(2014) reportedthatNorthAmerica hasexperiencedasignificantdecreaseindroughtfrequencyover 1951–2010,whichtheyestimatedusingSPI,andisverifiedinthisstudy usingSPEI(Fig.4).

Amongallsub-regions,droughtsinS1arelessfrequent(4.7/lustrum),comparedwithS2(6.0/lustrum),S3(5.1/lustrum)andS4(6.2/ lustrum).Ontheotherhand,aspointedoutby Bonsaletal.(2011),the lowfrequencyhasgenerallyledtoaloweradaptivecapability,thus makingtheregionmorevulnerabletotheimpactsofdrought.Forexample,theCPandinteriorvalleysofBritishColumbiaaresusceptibleto droughtprimarilyduetotheirlocationintheleewardsideofmajor mountainranges,withlowprecipitationofhighspatialandtemporal variability.AlthoughthedroughtfrequencyinnorthernCanadais higher(S2andS4),itislessofaconcernlargelybecauseofitslow

populationdensityandalackofagriculturalactivity.Eventhough signiﬁcantwarminghasbeenreportedoverCanada,ithasnotresulted inanincreaseinthedroughtfrequency,whichtendstoexhibitdecadal variabilityconsistentwiththatofprecipitation(Bonsaletal.,2011).

3.3.2.Changesindroughtduration

3.3.2.1.Longestdroughtduration(LDD).Inthisstudy,thelongest droughtduration(LDD)isdeﬁnedasthelongestconsecutivemonths withSPEI<-1over1950–2016.FromthespatialdistributionofLDD shownin Fig.5(a),itisevidentthattheLDDinmostpartsofCanada is<12months.Inaddition,itseemsthattheLDDinCanadahasa dipolarpattern:northernCanadatendstohavealongerLDD (12–24 months),whereastheCPandsoutheastregionhaveashorter LDD(3–6months).Thedecadalspatialdistributionoftheoccurring timeofLDDin Fig.5(b),illustratesthatS2andS4(northeastCanada) havebeenstrickenbytheirLDDsprimarilyinthe1950sand1960s. However,theCPandeastS1havebeenaﬀectedinmorerecentdecades (afterthe1990s),suggestingthatsouthandeastCanadahave experiencedmorelong-lastingdroughtsinthepastfewdecades.

3.3.2.2.Totaldroughtduration(TDD) Fig.6(a-d)depictsthetemporal variationsoftheannualTDDforthefoursub-regions.ExceptforS3 (Fig.6c),allsub-regions(S1,S2,andS4)haveexhibitedsigniﬁcant decreasingtrendsat 0.20month/decade, 0.42month/decade,and 0.69month/decade,respectively.Moreover,S1,S2,andS4all experiencedshort-durationdroughtscircathe1980s;incontrast,S3 encountereddroughtsofshortdurationsinthe2000s.Achangepoint analysis(Table2)showsthatCanada'swestcoastunderwentanabrupt changeinthe2000s,whereastheremainderofthecountrywas subjectedtoanearlierabruptchangeinTDDin1980s.

3.3.2.3.Severeandextremedroughtduration(SEDD) Fig.6(e-f)depicts thetemporalevolutionoftheannualSEDD(SPEI< 1.5)forfour sub-regions.UnliketheTDD,theannualSEDDsexhibitasigniﬁcant decreasingtendencyforallsub-regionsat 0.11month/decade,

Fig.5. Spatialdistributionof(a)thelongestdroughtdurationand(b)thelongestdroughtdurationoccurringtime.

0.32month/decade, 0.14month/decade,and 0.48month/ decade,respectively.Inaddition,theannualSEDDshaveexperienced changepointssimilartotheannualTDD(Table2).Consistentwith Fig.6(d),S4hasthelargestdecreasingSEDDtrendamongallsubregions.Inaddition,giventhedecreasingtrendsestimatedusingSen's slopeforS2( 0.42month/decadeand 0.32month/decade),S4 ( 0.69month/decadeand 0.48month/decade),andS3( 0.19/ decadeand 0.14/decade)in Fig.6,itisevidentthatsevereand extremeeventshavecontributedsigniﬁcantlytothegeneraldeclineof TDDinnorthernCanada.

Amongallsub-regions,themeanTDDsinS1arelesspersistent (2.1month/year),comparedwithS2(3.2month/year),S3(2.2month/ year)andS4(3.1month/year).However,theCPsuﬀeredoneofthe mostsevereandprolongeddroughtsinCanadianhistoryfrom1999to 2005,whichwaswellcapturedbytheSPEIin Fig.5(b).Furthermore, theregionthatexperiencedtheLDDintheCPwaspredominantlylocatedinAlbertaandSaskatchewan(S1),whichisconsistentwith Greeneetal.(2011).From Fig.5(a),thedroughtsinsouthernOntario andQuebecaretypicallyshortinduration,andtheytendtooccurin morerecentdecades.

3.3.3.Changesindroughtarea

3.3.3.1.Totaldroughtarea(TDA).Theannualtemporalvariabilityof theTDAduring1950–2016overdiﬀerentsub-regionsispresentedin Fig.7(a-d).TheTDAisthepercentageofgridswithSPEI<-1overthe totalgrids.ThemostsigniﬁcantdecreasingtrendisfoundinS4,witha rateof 6%/decade(Fig.7d).Overthestudyperiod,TDAsaremore widespreadinS2andS4(24%and23%,respectively),whereasTDAin S1isthelowest(15%).Onanannualtimescale,widespreaddroughts haveoccurredinthe1950sforallsub-regions.Additionally,striking droughtsoccurredin1961and1962forS1(>40%),in1965,1972, and1990forS2(>70%),in1983,1995,and2004forS3(>50%), andin1964and1965forS4(>70%).Asshownin Fig.7(a),there seemstobeanincreasingtrendinTDAafter2005forS1,however,the magnitudeismuchsmallerthanthatduringtheperiod1950–1970.In otherwords,spatiallylessextensivedroughtshadoccurredinCanada after2000s.AbruptchangepointsofTDAalsooccurredinthe1970s (Table2),whicharesimilartothatofdroughtfrequency.

3.3.3.2.Severeandextremedroughtarea(SEDA). Fig.7(e-f)exhibitsthe timeseriesoftheannualSEDA(SPEI<-1.5)forthefoursub-regions. Similarto Fig.7aandd,themostsigniﬁcantdecreasingtrendsforSEDA werefoundinS1( 1%/decade)andS4(3%/decade).Moreover, SEDAsarewidespreadinS2andS4(14%and13%,respectively)and thatforS1isthelowest(6%).Again,similartoTDA,abruptchange

pointsappearedinthe1970s(Table2). Fromdecreasingtrends estimatedforTDAandSEDAofS2( 0.03/decadeand 0.009/ decade)andS3( 0.02/decadeand 0.007/decade)shownin Fig.7, itseemsthattherehasbeenamarginaldeclineindroughtareasinS2 andS3aﬀectedbymoremoderatedroughteventsoccurringinrecent decades.

AnoveralldeclineinthedroughtareasinCanadaisalsodetectedby VanDerSchrieretal.(2013).Asdemonstratedby Szetoetal.(2016), theextremedroughteventinBritishColumbiaandAlberta(westernS1) in2015wasexceptionalbecauseofitsseverityandextent,whichwas wellcapturedbytheridgeduringthe2010sin Fig.7(a)and(e). Bonsal etal.(2012) alsonotedthatdroughtconditionsinthetwentiethcentury wererelativelymildwhencomparedtothepre-instrumentalperiodon theCP,reinforcingour ﬁndingsofageneraldeclineinthedrought areasofS1(Table3).

3.3.4.Changesindroughtseverity

3.3.4.1.Totaldroughtseverity(TDS).TDSisdefinedastheabsolute valueofSPEI(< 1)multipliedwiththeduration.Asshownin Fig.8(a-d),theannualTDSinallsub-regionshasdeclinedat 0.33/ decade, 0.81/decade, 0.35/decade,and 1.25/decade, respectively(Table3),whichmeansthattheseverityofdrought eventshavegenerallymitigatedfrom1950to2016,especiallyforS4. Allsub-regionssuffereddroughtsofhigh-severityinthe1950sand 1995–2005wasadifficultperiodforS3.ForS1,1999–2005isalso markedbyhighseverity,whichhasbeenreportedby Hryciwetal. (2013).

3.3.4.2.Severeandextremedroughtseverity(SEDS). Fig.8(e-f)presents theevolutionoftheannualSEDS(SPEI<-1.5)forfoursub-regions. Coincidentally,allsub-regionshavethesamechangepointsasthatof TDS(Table3).Similarly,allsub-regionsexperiencedanegativetrendin SEDSespeciallyforS4( 0.96/decade),anddroughtsofhigh-severity inthe1950s.From Fig.8(e),eventhoughtrendsofSEDSbecame positiveafter2005forS1,themagnitudewassmallcomparedtothat forthenegativetrendin1950–1970,indicatingthatthe1950swasa decadeimpactedbysevereandprolongeddroughts.Givenconsistent negativetrendsestimatedbytheSen'sslopeforS1( 0.33/decadeand 0.21/decade),S2( 0.81/decade and 0.69/decade),S3( 0.35/ decadeand 0.27/decade),andS4( 1.25/decadeand 0.96/ decade)in Fig.8,thegeneraldeclineinsevereandextremedroughts havecontributedsigniﬁcantlytothegeneraldeclineofTDSacross Canadaover1950–2016.

Theaboveresultsagreewiththatofpastdroughtstudies.Forexample,usingsc_PDSIandintermediatefutureclimatescenarios

1950196019701980199020002010

Fig.6. Temporalvariationofannualtotaldroughtduration(TDD)(a,b,c,d)andsevere-extremedroughtduration(SEDD)(e,f,g,h)forS1,S2,S3,andS4.Thered lineisthelineartrendandSisthetrendperdecadeusingSen'sslope(signiﬁcantvaluesareinbold).(Forinterpretationofthereferencestocolourinthis ﬁgure legend,thereaderisreferredtothewebversionofthisarticle.)

simulatedby14globalclimatemodels(GCMs)ofCMIP5, Dai(2013) projectedalargeincreaseinwetnessoverhighlatitudesinNorth America.Moreover,underclimatewarming,summerandfallprecipitationisprojectedtoincreaseathighnorthernlatitudes(Swainand Hayhoe,2015),whichwouldleadtowetterconditionsinCanada. Choi andKim(2018) assessedhowwarmingcouldaﬀectspringandearly summerdroughtsinNorthAmericaandfoundthatregionswith droughtreliefarepredominantlylocatedinCanadaandAlaska,partly duetomorespringsnowmelt.Basedonprojectionsfrom21GCMsof

CMIP5, SwainandHayhoe(2015) projectedapronouncedincreasein wetconditionsacrossCanadainspringofthe2020sto2080s,andthe magnitudeofprojectedchangesinwetnessmayscalewithglobal temperature.EventhoughregionssuchasYukon(S3)isprojectedto becomewetterwithhigherspringSPI,giventhedistributionofSPIis highlyskewed,itcouldalsoexperienceprolongeddryconditions.

Table2

Changepointsofannualdroughtduration,droughtarea,anddroughtseverity. SubregionDroughtdurationDroughtarea Droughtseverity

Total drought Severe and extreme drought

S1 197619761974197219761976

S2 199319941972197219941994

S3 20042004 19831985 20042004

S4 198319831978197819831983

Whole 197819781972197219781978

Changepointsthatarestatisticallysigniﬁcantareinbold.

3.4.Compositeanalysis

Fig.9 depictsthecompositemeanandanomalyofprecipitationand surfacetemperatureacrossCanadabetween1950and1979and 1987–2016.From Fig.9(e),Canadahasexperiencedincreasingprecipitationexceptforsomeregionsinthesouth,resultinginanoverall lesssevereandlessfrequentoccurrencesofdroughtsinCanada.The annualrainfallinCanadahasincreasedbyabout12.5%whiletheannualsnowfallbyabout4%from1950to2009;however,theincreasein snowfallwasnotconsistentacrossthecountry.Forexample,western provinces(BritishColumbia,Alberta,andSaskatchewan)exhibited signiﬁcantdecreasingtrends(Fig.9e)(MekisandVincent,2011).In addition,highersurfacetemperatureswereobservedacrossCanada, especiallyinthenorth(Fig.9f)(Vincentetal.,2018).Therefore,these ﬁndingssuggestthatCanadahasgraduallybecomewetterandwarmer sincethe1950s.

BasedonGCMoutputsofCMIP5, Cooketal.(2014) assessedthe relativecontributionfromchangesinprecipitationversusevapotranspirationtothemagnitudeandextentofdryinginducedbyglobal warming.Theyshowthatoverallmoreprecipitationmaycausethe entireNorthernHemispheretobecomewetter,especiallyinhighlatitudeswhereprecipitationincreaseisprojectedtobethehighest,but changesinmid-latitudeswouldbenearneutralormarginallywetter.In contrast,anincreaseinPETcouldresultindryingacrossalllatitudes.In viewofprojectedchangesinPETandprecipitation,thenetresultisa robustwettingoccurringinNH'shighlatitudes(Dai,2013). Naumann etal.(2018) investigatedglobaldroughtconditionsunderdiﬀerent projectedglobalwarminglevels.Theyrevealedthatthedroughtmagnitudewillhalvewitha1.5°CwarmingprimarilyintheRussianFederation,southernAlaska,andCanada.Further,droughtsareprojected tobeshorterinlengthformostlandareasnorthof55°Nlatitude,which furtherreinforcesour ﬁndings.Therefore,theincreaseinprecipitation isexpectedtosubstantiallyoutweightheimpactofincreasedPET(resultingfromincreasedtemperature)tofuturedroughtsinCanada.

AccordingtotheClausius–Clapeyronrelationship,theatmospheric waterholdingcapacitywillincreasebyabout7%/Kriseintemperature,andsowarmingwillgiverisetoincreasedevaporation,atmosphericmoisture,moiststaticenergyandthereforestormsareexpected tobemoreintensive(O'GormanandSchneider,2009).WiththeClausius–Clapeyronrelationshipasthebasis,the7%/oKriseinwater holdingcapacityisgenerallytrue,butdiﬀerentratesofincreasein extremerainfallwithrespecttoincreaseintemperaturehavebeen detected(Tanetal.,2018b),whichisexpectedsinceprecipitationis highlyvariablespatially.Contrastingdriversofprecipitationchangeat theregionalandglobalscales,globallythereishighconﬁdencethat globalmeanprecipitationincreases~2%/Kofglobalmeanwarming (HeldandSoden,2006).Thereforelikelyattributedtotheregional impactofglobalwarming,theoccurrencesofsevereandextreme droughtsacrossthefoursub-regionsofCanadahasgenerallydeclined over1950–2016.

3.5.Dynamicinﬂuenceofclimatedrivers

Fig.10 presentstheestimateddynamicregressionslopecoefficients ofsixclimatedriversonfourregionalSPEIseries. Fig.10(a)illustrates thatthecorrelationbetweenNINO3andSPEIatS1experiencedaphase changefrompositivebefore1970tonegativeuntil2015.However, NINO3wasnegativelycorrelatedwithSPEIbefore1985(Fig.10s)in S4,andtherewasatroughcirca1970andaridgeduringthe1990s.For S3,theslopeexhibitedadecadalvariabilityandgraduallychanged frompositivetonegative(Fig.10m).Additionally,theslopecoefficients ofNINO3tendtoapproach0after2005,implyingthattheeffectsofthe ENSOweakenedforallsub-regionsofCanadainrecentdecades.

Astheworld'smostsignificantinter-annualclimatepattern,ENSO playsasubstantialroleinthevariabilityofclimateworldwide.In Canada,ElNiñoistypicallyassociatedwithwarmeranddrierwinters, whereasLaNiñahastheoppositeeffect.Forexample,theENSOexertsa stronginfluenceonthewinterprecipitationofsouthwestCanadaandEl Niño(LaNiña)mayleadtoa14%decrease(20%increase)inthemean winterprecipitation(Ganetal.,2007).ElNiñoeventsaretypically accompaniedbyasummermoisturedeficitinmostofwesternCanada, whereasLaNiñaeventsproduceanabundanceofsummermoisturein extremewesternCanadaandinthesoutheasternpartoftheCP (ShabbarandSkinner,2004).OuranalysisfurtherrevealsthetimevaryinginfluencesofENSOondroughtsinCanada.Giventhelarge negativeslopecoefficientofS1in2001–2002(Fig.10a),thestrong influenceofElNiñoinS1leadingtoasignificantdeclineinSPEIis expectedin2001–2002,whichcorrespondstothesevere2001–2002 droughtintheCP.Furthermore,theextremedroughtinwesternCanadaisalsoamplifiedbythestrongElNiñoepisodeof2015–16(Szeto etal.,2016).

ForPDO,theslopecoefficientofS1changedfrompositivetonegative inthe1970s(Fig.10b). Fig.10(t)depictsthatPDOwasnegativelycorrelatedwithSPEIbutitsimpactdecreasedfrom1950to2016 inS4.TwotroughsarefoundinbothS1andS3circa2000respectively (Fig.10n),suggestingthatPDO'seffectbecamestrongeratthattime. ForS2,thenegativecorrelationwasrelativelystableafter2000 (Fig.10h). Bonsaletal.(1993) reportedthatpositivePDOtendstogive risetoextendeddryspellsduringthegrowingseasonthroughoutthe CP.Similarly, Ganetal.(2007) foundthatapositive(negative)phaseof thePDOisassociatedwithan8%decrease(9%increase)inthemean winterprecipitationinsouthwesternCanada.Ourresultsrevealthat PDOistypicallynegativelycorrelatedwiththeSPEI,whichfurther confirmsthatapositivePDOcontributestodrierconditions.

PNAexertedapredominantandpersistentnegativeinfluenceonthe SPEIinS1over1950–2106(Fig.10c),butamorepositiveinfluencein S4(Fig.10u).ForS2andS3,theeffectsofthePNAhavebeentypically weak:thecorrelationforS2experiencedaphasechangecirca1980 (Fig.10i);andtheimpactofthePNAforS3wasminimalafterthe 1970s(Fig.10o),implyingareductioninthestrengthandapossible phasechangeinthefuture.ThepositivePNAandassociatedgeopotentialheightanomaliesoverwesternCanadagiverisetolarge-scale subsidenceoverthearea,leadingtowarmeranddrierconditions (Bonsaletal.,2001).GivenslopecoefficientsinS3arepredominantly negative(Fig.10o),whilePNAwasmostlyinpositivephasessincethe mid-1970s,SPEIisexpectedtodecrease,resultingindrierclimatein westernCanada.Theresultisconsistentwiththatof Ganetal.(2007) whoshowthatastrongpositive(negative)PNAwouldleadtoa12% decrease(9%increase)inmeanprecipitationinsouthwesternCanada. DistincteffectsofAMOaredisplayedinS1andS3.ForS1,theslope coefficientsdecreasedfrompositivetonegativecirca1980(Fig.10d), butanoppositerelationshipisobservedforS3in1985(Fig.10p).The influenceofAMOinS2weakenedovertheyears,althoughitremains positive.However,theslopeforS3changedcirca1985(Fig.10p)from negativetopositive.From Fig.10(d),theAMOwarmphaseinthelate 1990scontributedtoanegativeslopethatresultedindryerconditions inS1,whilethecoolphasefrom1960to1970ledtoapositiveslope

SEDA (S3)

0.00.10.20.30.4(a)

1950196019701980199020002010

0.00.20.40.6(c)

1950196019701980199020002010

0.000.100.200.30(e)

1950196019701980199020002010

0.00.20.40.6(g)

1950196019701980199020002010

0.00.20.40.60.8(h)

1950196019701980199020002010

Fig.7. Temporalevolutionofannualtotaldroughtarea(TDA)(a,b,c,d)andsevere-extremedroughtarea(SEDA)(e,f,g,h)forS1,S2,S3,andS4.Theredlineisthe lineartrendandSisthetrendperdecadeusingSen'sslope(signiﬁcantvaluesareinbold).(Forinterpretationofthereferencestocolourinthis ﬁgurelegend,the readerisreferredtothewebversionofthisarticle.)

andwetterconditionsinS3,whichagreeswiththe ﬁndingsof Shabbar andSkinner(2004).Inaddition,theAMOwarmphasewithanegative slopecoeﬃcientfrom2015to2016(Fig.10d)couldstimulatedrier conditionsinS1,whichmayhavefurtheraggravatedthe2015–16extremedroughtinwesternCanada.

TheeffectsofAOchangedfromnegativetopositiveforbothS2and S4in1995and1975,respectively.AtroughappearedinS1in1970 (Fig.10e),whichcouldmeanastrongereffectofAOoverS1atthat time,butafter1985thateffectdecreasedtoaminimumuntil2005.

Afterthattheslopecoefficientstendtodecrease,implyingareinforcing effectofAOoverS1.ForS3,thecorrelation fluctuatesover1950–2016, withanegativeslopedetectedafter2010(Fig.10q).TheAOisa dominantclimatedriveraffectingthewintertemperatureineastern Canada,suchthatwinterstendtobecolderwhenAOispositiveand viceversa(Bonsaletal.,2001).Asshownin Fig.10(k),aphasechange occurredcirca1995,showingapositiveinfluenceofAOoverS2since then.

TheeﬀectsofsunspotsonSPEIappeartobegenerallystableand

Table3

TheSen'sslopeofdroughtduration,droughtarea,anddroughtseverity.

SubregionDroughtdurationDroughtarea Droughtseverity

Total

Trendsthatarestatisticallysigniﬁcantareinboldandtheunitisperdecade.

predominantlynegativeafter1990forS2,S3,andS4;whileforS1,the influencechangedfrompositivetonegativeinthe1980s,withanenhancednegativeslopeafterthe1990s(Fig.10f).Noteworthyisthatthe slopecoefficientstendtoincreaseafterthe2000sforS1,S3,andS4, indicatingthattheeffectsofsunspotshavestrengthenedinrecent decades.Thesolarcycles,whicharedrivenbythesun'smagneticturbulenceatan11-yearcycle,havethepotentialtoinfluenceclimate systemsonEarth(Meehletal.,2002).Forexample, Fangetal.(2018) showedthatsunspotactivitiesarecloselyassociatedwithdryconditionsinChina.InSouthernCanada, Fuetal.(2012) examinedthe combinedinfluenceofsolaractivityandElNiñoepisodesonstreamflow,and Prokophetal.(2012) alsofoundthatyearswithmajor floods weremostlikelytooccurduringyearswithalowsunspotnumber.In contrast,the1999–2005Canadiandroughtoccurredduringaperiod withahighsunspotnumber.SunspotsareprimarilynegativelyassociatedwiththeSPEIafterthe1980sinS1(Fig.10f),whichfurther confirmsthatincreasingsunspotscouldleadtoadecreaseinSPEI,indicatingthatdryeventsaremorelikelytooccurduringperiodswith highsunspotnumbers.

4.Discussion

Globallyoceanscoverover70%oftheEarth'ssurfaceandtheyserve asenormousreservoirsofwater,energy,carbon,andothersubstances. Thereforeoceansthatinteractdirectlywiththeatmosphereplayapivotalroleintheglobalclimatesystem(BushandLemmen,2019). Variousstudieslinkingdroughtvariabilitytomajorcirculationsofthe NorthernHemispherehaveidentifiedadynamicrelationshipoverthe past300years;however,becausethemechanismslinkingregionalclimatevariabilitywithocean-atmosphericcirculationpatternsarenot firmlyestablished,itremainsdifficulttoprovideapreciseexplanation ofthecauseoftheatmosphericcirculationshiftanditsdynamicinfluencesondroughts,posingexcessivechallengesforpredictingthe onset,duration,andseverityofdroughtsinCanada(Bonsaletal.,2017; Hanesiaketal.,2011; Tanetal.,2018a).Theaboveresultsrevealthe haphazardanddynamicnatureofclimateanomalies'influenceoverthe drynessofCanadadividedintofoursub-regionsrepresentedbytheSPEI in1950–2016.Thisisexpectedbecauseclimateanomalieshavebeen changingbetweenpositiveandnegativephasesandatvaryingstrength andfrequencies,rangingfrominter-annualtointer-decadalscales.

Fig.10 presentsmultiplephasechangesforthedynamicregression slopecoefficientsinthe1970s,whichcouldbeattributedtotheextraordinaryclimaticshiftofthe1970scharacterizedbyasignificant shiftfromcoolertowarmertropicalPacificseasurfacetemperatures (SSTs)(Meehletal.,2009).Climatechangeandanthropogenicimpacts maycausenonstationaritiesinhydrologicalextremes(TanandGan, 2015),andglobal-scaleabruptchangesinatmosphericcirculationsand climatehavebeendetected(Jacques-Coper andGarreaud,2015),e.g., the2ndSVDpatternofthesummerPDSIinCanadadisplayedadecreasingtrendfrom1940tothemid-1970s,butthereafterincreased abruptly(ShabbarandSkinner,2004).Theatmosphere-oceanclimate

systemovertheNorthPacificOceanabruptlychangeditsnormalstate, wherebySSTscooledinthecentralPacificandwarmedoff thecoastof westernNorthAmerica(Milleretal.,1994). Meehletal.(2009) found thattheobserved1970s'climateshiftmayhavebeentheresultof changesinexternalforcingsuperimposedonaninherent fluctuationof thePacificclimatesystem,whichinclude:(a)areductioninthe northwardoceanicheat fluxassociatedwiththeNorthAtlanticthermohalinecirculation,and(b)arapidincreaseinanthropogenicaerosol emissions,particularlyoverEuropeandNorthAmerica(Bainesand Folland,2007).

DroughtsinsouthernCanadaareassociatedwithpositive500-hPa geopotentialheightanomaliescenteredovertheGulfofAlaskaand BaffinBay(Girardinetal.,2004b). Bonsaletal.(2001) suggestedthat ElNiñoepisodeswithapositivePDOareassociatedwithstrongpositive wintertemperatureanomaliesovermostofCanada,whicharetheresultofadeeperthannormalAleutianlow,anamplificationandeastwarddisplacementofthewesternCanadianridge,andnegative500hPaheightanomaliesoverthesoutheasternUS. Girardinetal.(2004a) foundthatdroughtovertheAbitibiPlainsecoregion(easternCanada) displayedashiftin1850:droughtwascorrelatedwithPDObefore 1850,butafterwhichitwasmorecorrelatedwiththeNAO,suggesting thediminishingeffectsofPacificforcing.Moreover,theshiftthatoccurredcirca1850reflectsanorthwarddisplacementofthepolarjet streaminducedbyawarmerseasurfacetemperaturealongtheNorth Pacificcoast,whichinhibitstheoutflowofcoldanddryArcticairover mostofCanadaandallowstheincursionofairmassesfromtheAtlantic subtropicalregions.Usingredundancyanalysis, Girardinetal.(2004a) alsodescribedthechangingrelationshipbetweenatmosphericcirculationindicesanddroughtsinCanada.Theydemonstratedthatthe first principalcomponent(PC1)wasnegativelycorrelatedwiththePDOand positivelywiththeNAOandSouthernOscillation(SO)from1706to 1998.However,forPC2,itsrelationshipwiththePDOchangedfrom positivetonegativeduring1880–1979,andtheeffectoftheSOweakenedduring1850–1949.

Dai(2011b) showsthatmeteorologicaldroughtsoftenresultfrom persistentanomalouslarge-scaleatmosphericcirculationpatternsinducedbyanomaloustropicalSSTorotherremoteclimaticconditions, andthatiswhymanystatisticalmethodshaveemployedlarge-scale atmosphericcirculationsasnaturalprecursorstopredictdroughts. However,the1999–2005CPdroughtwasrelatedtoanorthwardextensionofapersistentdroughtinAmerica,insteadofdroughtsnormally attributedtodistinctmeridional flowsovertheNorthPacific andNorth America(Bonsaietal.,2005).Furthermore,alackofconsistentpositive PNAandPDOpatternsduringtherecentmostseveredroughtin 2001–02differsfrompastdroughtsthattendtobeassociatedwith large-scaleteleconnections(Hanesiaketal.,2011).Therefore,froma large-scaleteleconnectionperspective,this1999–2005droughtisdifferentfrompastdroughtsoftheCP,whichagainconfirmsour findings onthenon-steadyanddynamicrelationshipsbetweenlarge-scaleclimateoscillationsanddroughtsinCanada.

Similarly, Rajagopalanetal.(2000) whoexaminedtheteleconnectionofpastsummerU.S.droughtstoENSO,recommendedthe applicationoftheBDLmodeltobettercapturenon-stationaritiesinthe relationshipbetweendroughtsandENSOovertheTwentiethcentury. Meanwhile,fromCMIP5climatemodels'simulations, Coatsetal. (2013) investigatedtheteleconnectionbetweentropicalPaciﬁcSSTs and200mbgeopotentialheightinNorthAmerica.Theyconcludedthat thenon-stationarityofthisteleconnectionwasassociatedwithchanges intropicalPaciﬁcSSTs,whichagainhighlightstheroleofoceandynamicsinthenon-stationarynatureofteleconnectionsobservedin NorthAmerica.

Exceptforlarge-scaleclimateoscillations, Greeneetal.(2011) who studiedcloudcharacteristicsduringthe1999–2005droughtintheCP, foundthatmonthswithbelow-averageprecipitationtendtohavenegativecloudamountanomalies,andtheoccurrenceofthickand mediumcloudsdecreasedwithdroughtseverity.Ourresultsfurther

1950196019701980199020002010 02468(a)

1950196019701980199020002010

SEDS (S3) SEDS (S4)

1950196019701980199020002010

Fig.8. Temporalvariabilityofannualtotaldroughtseverity(TDS)(a,b,c,d)andsevere-extremedroughtseverity(SEDS)(e,f,g,h)forS1,S2,S3,andS4.Thered lineisthelineartrendandSisthetrendperdecadeusingSen'sslope(signiﬁcantvaluesareinbold).(Forinterpretationofthereferencestocolourinthis ﬁgure legend,thereaderisreferredtothewebversionofthisarticle.)

demonstratethedynamicinfluencesoflarge-scaleclimateoscillations ondroughtsinCanadaandrevealconsiderablechangesintherelationship.ItseemsthattheeffectsoftheENSOacrossCanadahave generallyweakenedinthepastfewdecades,especiallyafterthe2000s. ForS1,alllarge-scaleclimatedriversexertedmorenegativeinfluences ondroughtvariabilityafterthe1980s,whileENSO,AMO,andsunspots underwentaphasechangein1970,1980,and1990,respectively.For S2,exceptforPDO,otherclimateindicesaremorepositivelycorrelated withSPEIafterthe1990s.ForS3,PDO,PNA,andsunspotsare

negativelycorrelatedwithdroughtvariability.ForS4,morepositive relationshipscanbefoundwithPNAandAOafterthe1970s,andthe eﬀectofPDOweakenedover1950–2016.

5.Summaryandconclusions

Givenunderstandingthespatiotemporalcharacteristicsofdrought iscrucialfordroughtriskmitigation,thisarticleaimstoprovidea comprehensiveassessmentofdroughtconditionsacrossCanada.The

Fig.9. Compositemeanof(a,c)precipitation,(b,d)temperature,and(e,f)theiranomaliesduring1950–1979and1987–2016.Regionswithdotsareat95% signiﬁcancelevel.

Slope

1950196019701980199020002010

Slope

1950196019701980199020002010

−0.6 −0.2 0.2 0.4 SPEI~NINO3 Slope

1950196019701980199020002010

−2−1 SPEI~AMO0123 Slope

1950196019701980199020002010

−0.6−0.2 0.20.6 SPEI~NINO3 Slope

1950196019701980199020002010

SPEI~AMO Slope

−2−101 2

1950196019701980199020002010

−0.6−0.20.2 0.6 SPEI~NINO3 Slope

1950196019701980199020002010

0.05 0.15 SPEI~PDO Slope

1950196019701980199020002010

SPEI~AO−0.15−0.050.05 Slope

Slope

SPEI~PNA−0.050.000.05 Slope

1950196019701980199020002010

1950196019701980199020002010 −0.4 SPEI~Sunspot−0.20.00.20.4 Slope

1950196019701980199020002010

SPEI~PNA−0.15−0.050.050.15 Slope

1950196019701980199020002010

0.2 SPEI~ AO Slope

1950196019701980199020002010

SPEI~Sunspot−0.50.00.5 Slope

1950196019701980199020002010

SPEI~PDO−0.3−0.10.00.1 Slope

1950196019701980199020002010

0.10 SPEI~ AO Slope

1950196019701980199020002010

Slope

1950196019701980199020002010

A O Slope

SPEI~PNA Slope

−0.100.00 0.05 0.10

1950196019701980199020002010

SPEI~Sunspot−0.40.00.20.4 Slope

1950196019701980199020002010

1950196019701980199020002010 −0.15−0.050.05 0.15 SPEI~PNA Slope

1950196019701980199020002010 SPEI~Sunspot−0.6−0.20.2 Slope

1950196019701980199020002010

Fig.10. VariationsintherelationshipbetweenregionalSPEIandlarge-scaleclimateoscillationsforS1(a-f),S2(g-l),S3(m-r),andS4(s-x).Theblacksolidline denotestheestimatedtime-varyingslopes,alongwiththe25thand75thpercentilecredibleintervallines(reddottedlines)fromtheBayesiandynamiclinearmodel. (Forinterpretationofthereferencestocolourinthis ﬁgurelegend,thereaderisreferredtothewebversionofthisarticle.)

CRUdatasetwasusedtocalculateSPEIforfoursub-regionsofCanada for1950–2016.Spatialandtemporalvariationsofdroughtduration, frequency,percentarea,andseveritywereinvestigated.Thedynamic inﬂuencesoflarge-scaleclimatedriversonthedroughtvariabilityof CanadaintermsofSPEIwerealsoidentiﬁed.Theprimaryconclusions canbesummarizedasfollows:

(1)The1950swasoneoftherecentdecadesmostinﬂuencedbysevere andprolongeddroughtevents.Seasonally,wettingtrendsdominate overmajorpartsofCanadainspring,summer,andautumn;however,aprevalentwinterdryingtrendhasbeendetectedinsouthwestCanada.

(2)Forthedroughtfrequency(DF),S1andS4haveexperiencedsigniﬁcantdecliningtrends.AlthoughDFishigherinnorthern Canada,itislessofaconcernlargelybecauseofitslowpopulation densityandalackofagriculturalactivity.Asigniﬁcantincreasein temperaturehasbeenreportedoverCanada,butthishasnotresultedinanincreaseinthedroughtfrequency.

(3)Thelongestdroughtduration(LDD)inCanadaexhibitsadipolar pattern:thenorthernpartofCanadahasalongerdroughtduration, whereasthesouthernparthasashorterduration.IntermsofTDD, themostsignificantdecreasingtrendswerefoundinS4.Spatially, dryareaswerewidespreadinS2andS4.Temporally,widespread droughtsoccurredacrossCanadainthe1950sandtherewasan increasingtrendafter2005forS1.Forthesevere&extreme droughtduration(SEDD),allsub-regionsexhibitedasignificant decreasingtrend,whichcontributedsignificantlytothegeneral declineofthetotaldroughtduration(TDD)innorthernCanada.

(4)Theannualtotaldroughtseverity(TDS)andsevereandextreme droughtseverity(SEDS)havealsoexhibitednegativetrends,especiallyforS4,implyingthatdroughtshavegenerallydeclinedfrom 1950to2016.Allsub-regionssuﬀeredextremelyseveredroughtsin the1950s.Thegeneraldeclineinsevereandextremedroughts acrossCanadahasresultedinthegeneraldeclineofannualTDS acrossCanada.

(5)OurresultsfurtherdemonstratethedynamicinfluenceoflargescaleclimatedriversondroughtsinCanadabutwithconsiderable changesintherelationshipover1950–2016,whichispartlyassociatedwiththeclimateshiftdetectedinthe1970s.Theeffectsof ENSOweakenedacrossCanadainthepastfewdecades,especially afterthe2000s;andENSO,AMO,andsunspotsexperiencedaphase changein1970,1980,and1990,respectively.ForS1,climate patternsgenerallyexertedmorenegativethanpositiveinfluenceon thedroughtvariabilityexpressedintermsofdynamicregression slopecoefficientsafterthe1980s.ForS2,exceptforPDO,other climateindicesweremorepositivelycorrelatedwiththeSPEIafter the1990s.ForS3,PDO,PNA,andsunspotswerenegativelycorrelatedwiththedroughtvariability,whenbothENSOandAMO experiencedaphasechangecirca1985.ForS4,morepositiverelationshipswereidentifiedforPNAandAOafterthe1970s,andthe effectofPDOhasweakenedduring1950–2016.Overall,likelyattributedtotheregionalimpactofglobalwarming,theoccurrences ofsevereandextremedroughtsacrossthefoursub-regionsof Canadahasgenerallydeclinedoverthelastsevendecadessince 1950.

Acknowledgments

ThisprojectwaspartlyfundedbytheNaturalScienceand EngineeringResearchCouncil(NSERC)ofCanada.The ﬁrstauthorwas alsopartlyfundedbytheChineseScholarshipCouncilofChinaandby theUniversityofAlberta.Wesincerelythankthethreereviewersfor theirconstructivecommentsandsuggestionswhichhavesigniﬁcantly improvedourmanuscript.TheClimaticResearchUnitVersion4.03 dataset,CRUTS4.03,wasdownloadedfrom http://www.cru.uea.ac. uk/data/

Appendix A.Supplementarydata

Supplementarydatatothisarticlecanbefoundonlineat https:// doi.org/10.1016/j.atmosres.2019.104695

References

Asong,Z.E.,Wheater,H.S.,Bonsal,B.,Razavi,S.,Kurkute,S.,2018.Historicaldrought patternsoverCanadaandtheirteleconnectionswithlarge-scaleclimatesignals. Hydrol.EarthSyst.Sci.22,3105–3124. https://doi.org/10.5194/hess-22-3105-2018

Baines,P.G.,Folland,C.K.,2007.Evidenceforarapidglobalclimateshiftacrossthelate 1960s.J.Clim.20,2721–2744. https://doi.org/10.1175/JCLI4177.1

Bonsai,B.R.,Wheaton,E.E.,Bonsal,B.R.,2005.Atmosphericcirculationcomparisons betweenthe2001and2002andthe1961and1988CanadianPrairiedroughts. Atmosphere-Ocean43,163–172. https://doi.org/10.3137/ao.430204

Bonsal,B.R.,Chakravarti,A.K.,Lawford,R.G.,1993.TeleconnectionsbetweenNorth PaciﬁcSSTanomaliesandgrowingseasonextendeddryspellsonthecanadian prairies.Int.J.Climatol.13,865–878. https://doi.org/10.1002/joc.3370130805

Bonsal,B.R.,Shabbar,A.,Higuchi,K.,2001.Impactsoflowfrequencyvariabilitymodes onCanadianwintertemperature.Int.J.Climatol.21,95–108. https://doi.org/10. 1002/joc.590.

Bonsal,B.R.,Wheaton,E.E.,Chipanshi,A.C.,Lin,C.,Sauchyn,D.J.,Wen,L.,2011. DroughtresearchinCanada:areview.Atmosphere-Ocean49,303–319. https://doi. org/10.1080/07055900.2011.555103

Bonsal,B.R.,Aider,R.,Gachon,P.,Lapp,S.,2012.AnassessmentofCanadianprairie drought:past,present,andfuture.Clim.Dyn.41,501–516. https://doi.org/10.1007/ s00382-012-1422-0

Bonsal,B.R.,Cuell,C.,Wheaton,E.,Sauchyn,D.J.,Barrow,E.,2017.Anassessmentof historicalandprojectedfuturehydro-climaticvariabilityandextremesoversouthern watershedsintheCanadianPrairies.Int.J.Climatol.37,3934–3948. https://doi.org/ 10.1002/joc.4967

Bush,E.,Lemmen,D.S.,2019.Canada’sChangingClimateReport.Canadian MeteorologicalandOceanographicSociety(CMOS),Ottawa,ON CanadianMeteorologicalandOceanographicSociety,2016.Canada'sTopTenWeather Storiesfor2015[WWWDocument].URL. http://www.cmos.ca/site/top_ten?a= 2015

Charrad,M.,Ghazzali,N.,Boiteau,V.,Niknafs,A.,2014.NbClust:AnRpackagefor determiningtherelevantnumberofclustersinadataset.J.Stat.Softw.61,1–36. https://doi.org/10.18637/jss.v061.i06

Choi,W.,Kim,K.,2018.Physicalmechanismofspringandearlysummerdroughtover NorthAmericaassociatedwiththeborealwarming.Sci.Rep.8,1–8. https://doi.org/ 10.1038/s41598-018-25932-5

Ciupak,M.,Ozga-Zielinski,B.,Adamowski,J.,Quilty,J.,Khalil,B.,2015.Theapplication ofDynamicLinearBayesianModelsinhydrologicalforecasting:VaryingCoeﬃcient RegressionandDiscountWeightedRegression.J.Hydrol.530,762–784. https://doi. org/10.1016/j.jhydrol.2015.10.023 . Coats,S.,Smerdon,J.E.,Cook,B.I.,Seager,R.,2013.Stationarityofthetropicalpaciﬁc teleconnectiontoNorthAmericainCMIP5/PMIP3modelsimulations.Geophys.Res. Lett.40,4927–4932. https://doi.org/10.1002/grl.50938

Cong,Z.,Yang,D.,Gao,B.,Yang,H.,Hu,H.,2009.Hydrologicaltrendanalysisinthe YellowRiverbasinusingadistributedhydrologicalmodel.WaterResour.Res.45. https://doi.org/10.1029/2008WR006852

Cook, B.I.,Smerdon,J.E.,Seager,R.,Coats,S.,2014.Globalwarmingand21stcentury drying.Clim.Dyn.43,2607–2627. https://doi.org/10.1007/s00382-014-2075-y Coulibaly,P.,2006.SpatialandtemporalvariabilityofCanadianseasonalprecipitation (1900-2000).Adv.WaterResour.29,1846–1865. https://doi.org/10.1016/j. advwatres.2005.12.013

Dai,A.,2011a.CharacteristicsandtrendsinvariousformsofthePalmerDroughtSeverity Indexduring1900-2008.J.Geophys.Res.Atmos.116. https://doi.org/10.1029/ 2010JD015541

Dai,A.,2011b.Droughtunderglobalwarming:Areview.WileyInterdiscip.Rev.Clim. Chang.2,45–65. https://doi.org/10.1002/wcc.81

Dai,A.G.,2013.Increasingdroughtunderglobalwarminginobservationsandmodels. Nat.Clim.Chang.3,52–58. https://doi.org/10.1038/nclimate1633

Escalante-Sandoval,C.,Nuñez-Garcia,P.,2017.Meteorologicaldroughtfeaturesin northernandnorthwesternpartsofMexicounderdiﬀerentclimatechangescenarios. J.AridLand9,65–75. https://doi.org/10.1007/s40333-016-0022-y

Fang,W.,Huang,S.,Huang,G.,Huang,Q.,Wang,H.,Wang,L.,Zhang,Y.,Li,P.,Ma,L., 2018.Copulas-basedriskanalysisforinter-seasonalcombinationsofwetanddry conditionsunderachangingclimate.Int.J.Climatol. https://doi.org/10.1002/joc. 5929

Fu,C.,James,A.L.,Wachowiak,M.P.,2012.Analyzingthecombinedinﬂuenceofsolar activityandElNiñoonstreamﬂowacrosssouthernCanada.WaterResour.Res.48. https://doi.org/10.1029/2011WR011507

Gan,T.Y.,Gobena,A.K.,Wang,Q.,2007.PrecipitationofsouthwesternCanada:Wavelet, scaling,multifractalanalysis,andteleconnectiontoclimateanomalies.J.Geophys. Res.Atmos.112. https://doi.org/10.1029/2006JD007157

Gao,T.,Wang,H.J.,Zhou,T.,2017.Changesofextremeprecipitationandnonlinear inﬂuenceofclimatevariablesovermonsoonregioninChina.Atmos.Res.197, 379–389. https://doi.org/10.1016/j.atmosres.2017.07.017

Ge,Y.,Apurv,T.,Cai,X.,2016.Spatialandtemporalpatternsofdroughtinthe ContinentalU.S.duringthepastcentury.Geophys.Res.Lett. https://doi.org/10. 1002/2016GL069660

Girardin,M.P.,Tardif,J.,Flannigan,M.D.,Bergeron,Y.,2004a.Multicentury

Another random document with no related content on Scribd:

carelessly without speaking to the horse, seeing that he stands over, or otherwise responds to his call, is himself to blame if he gets kicked. The attendant who does things to a dangerous or questionable horse for mere bravado cannot blame the owner if he gets himself injured. If a person teases a horse so as to tempt him to retaliate, not only is he responsible for his own consequent injuries, but largely also for the habits of the horse and for such injuries as others may subsequently sustain from him.

A dog or a bull shown in a public place, and which breaks loose and injures spectators or others, manifestly renders his master responsible for all such damage.

Treatment of aggressive vice. In mild dispositions in which the vice is roused by temporary suffering, it may often be cured by removal of the cause of such suffering. Indeed, without the healing of sores under the collar or saddle the vice cannot be arrested. Considerate and gentle treatment, too, will go far to restore confidence and to gradually do away with the aggressive disposition.

In wicked stallions castration will usually restore to a good measure of docility. The exceptional cases appear to be those that are hereditarily and constitutionally vicious, or in which the habit has been thoroughly developed and firmly fixed by long practice.

Mares, too, which become vicious and dangerous at each recurrence of œstrum, can usually be completely cured by the removal of the ovaries especially if this is done early in the disease.

The inveterate cases may usually be subdued and rendered controllable for a time by one of the methods of subjugation employed by the professional tamers, but unless they are thereafter kept in good hands they are liable to relapse into the old habit. Among the more effective methods are the Rarey mode of throwing which may be repeated again and again until the animal is thoroughly impressed with a sense of the domination of man and the futility of resistance; the resort of tying the head and tail closely together and letting the animal weary and daze himself by turning in a circle, first to the one side and then to the other; the application of the Comanche bridle made of a small rope, one loop of which is passed through the mouth and back of the ears and drawn tightly, then another loop is made to encircle the lower jaw, and the chin is drawn in against the trachea by passing the free end of the rope

round the upper part of the neck and again through the loop encircling the lower jaw and drawing it tight; or a similar small rope is passed a number of times through the mouth and back of the ears and drawn tightly so as to compress the medulla and stupify the animal. This is supposed to be rendered more effective by passing one turn each between the upper lip and gums and between the lower lip and the gums.

CATALEPSY.

Definition. Tetanic and paralytic forms. Balance of flexors and extensors Cataleptoid. No constant lesion. Hysterical. Hypnotic. Subjects: horse, ox, wolf, cat, chicken, Guineapig, snake, frog, crayfish. Causes: strong mental impression, indigestion, etc., in susceptible system. Lesion: inconstant, muscular degeneration, etc. Symptoms: wax-like retention of position given, voluntary movement in abeyance, mental functions impaired, secretions altered. Duration and frequency variable. Treatment: shock; cold; ammonia, pepper, snuff, electricity, amyle nitrite, nitro-glycerine, apomorphine, bromides, purgatives, bitters, iron, zinc, silver, open air exercise.

Definition. This is a functional nervous disorder, characterized by paroxysms of impaired or perverted consciousness, diminished sensibility, and above all a condition of muscular rigidity, by means of which the whole body, or it may be but one or more limbs retain any position in which they may be placed.

Laycock describes two forms in man—the catochus or tetanic form, and the paralytic form. Mills would restrict the name catalepsy to cases in which the muscular tone is such that the affected part may be bent or moulded like wax or a leaden pipe, and will not vary from this when left alone. Other forms in which this waxen flexibility (flexibilitas cerea) is absent or imperfect he would designate as cataleptoid.

The disease is not associated with any constant cerebral lesion, though it may supervene in the course of other nervous disorders, and therefore may own an exciting cause in existing lesions of the brain. The immediate cause must however be held to be functional, and this is in keeping with its most common form in man (hysterical), and with the hypnotic form which is observed both in man and animals. This latter may be looked on as a form of induced or hypnotic sleep, in which the retention of the position given to a

limb or part is the most prominent symptom. In all cases there is an impaired condition of the sensory functions of the cerebral convolutions, and an insusceptibility of the motor centres to the control of the will, or the reflex stimulus.

Hering has recorded the disease in the horse, Landel in the ox, and Leisering in the prairie wolf. The hypnotic form has been shown in cats, chickens, and Guinea pigs. The serpent charming of the Indian dervishes and similar effects on frogs and crayfish have been attributed to hypnotic catalepsy.

Causes. Strong mental emotions and diseases which profoundly affect the nervous system have been adduced as causes (fear, excitement, chills). Indigestible food has even been charged with causing it. There is undoubtedly, to begin with, a specially susceptible nervous system, and hence it is liable to prove hereditary, and in man to appear as a form of hysteria, or to alternate in the same family with epilepsy, chorea, alcoholism, opium addiction and other neurosis.

Hypnotism as a cause is claimed by various writers. Azam says that in the fairs in the South of France, jugglers hypnotize cocks by placing the bill on a board, on which they trace a black line passing between the two feet of the bird. Cadeac adds that Father Kircher, in the 17th century, employed a similar method to put fowls to sleep. Alix put cats to sleep by securing them firmly, and then looking steadily into their eyes. The condition attained varies according to the degree of the sleep, the will being dominated first, and later, consciousness of external objects is lost. Hypnotism, however, appears to be difficult and uncertain in the lower animals, in keeping with the limited development of intelligence and will, as compared with the human being. Cadeac states that the very old and the very young are completely refractory to hypnotizing influences.

Lesions. No constant pathological changes are found, though different nervous lesions may serve to rouse the disease in a predisposed subject. Fröhner found in the affected muscles granular swelling, fatty degeneration, hæmorrhages, and waxy (amyloid) degeneration of the cardiac muscles, corresponding to what has been found in tetanus; also hæmorrhages on the stomach and intestines.

Symptoms. The leading objective symptom is the tonic condition of the muscles by which a perfect balance is established and

maintained between the flexors and extensors so that the affected part maintains the same position which it had when the attack began, or any other position which may be given to it during the progress of the paroxysm. The position is only changed when the muscles involved have become completely exhausted. During the attack the affected muscles are swollen and firm, so that their outline may often be traced through the skin, later as the attack subsides they become soft and flaccid. Voluntary movement of the affected muscles is impossible until after the paroxysm. The attack usually comes on suddenly and in this respect resembles epilepsy; at other times there are premonitory symptoms of nervous anxiety, excitement or irritability. There is usually considerable impairment of consciousness, intelligence, common sensation, and even of the special senses. In a cataleptic dog Fröhner noted mental and motor troubles, considerable anæsthesia, and loss of sight, smell, and hearing. The eyes are fixed, the pupils either contracted or dilated, and the urine passed may be albuminous or even icteric.

Course, Duration. Like other functional nervous disorders this is extremely uncertain in its progress. There may be but one attack or a succession; they may last from a few minutes, to 7 days (Fröhner), or even several weeks (Hertwig); they may end in recovery or less frequently they may prove fatal usually by inanition.

Treatment. During a seizure a sudden shock will sometimes cut short the attack, douching with cold water, an inhalation of ammonia, of capsicum or of snuff, or the application of electricity in an interrupted current through the spine and affected muscles. Ether anæsthesia will not always relax (Sinkler). Inhalation of a few drops of nitrite of amyle has proved effective in man, as has also the injection subcutem of three drops of a 1 per cent solution of nitroglycerine, apomorphine hypodermically is usually effective (Sinkler). Bromide of potassium has also been advised, and in case of coldness of the surface, a warm bath.

When there is overloaded stomach and gastric indigestion an emetic is indicated, and in constipation a purgative (for speedy action chloride of barium or physostigma subcutem).

In the intervals between attacks tonics and general hygiene should be invoked to build up the weakened nervous system. Quinine, and

salts of iron, zinc or silver with a nourishing diet and out door exercise are especially indicated.

INSOLATION. HEAT EXHAUSTION. SUNSTROKE. THERMIC FEVER.

Definition: two forms. Heat exhaustion. Causes: prolonged heat, and moisture, overexertion. Impaired vaso-motor centre. Failing heart. Carbon dioxide poisoning. Symptoms: weak, fluttering pulse, perspiration, muscles flaccid, prostration, no hyperthermia. Treatment: stimulant, digitalis, digitalin, subcutem, nitro-glycerine, warm baths. Thermic fever. Hyperthermia excessive. Causes: insolation, prolonged heat and impure air, furnace heat, moist and dry heat, electric tension, overwork, muscular exhaustion, coagulation of myosin, constant heat on one part (head), excess of carbon dioxide, stiffening of bodies when killed in hot weather, debility, weakness, fatigue, chest constriction, tight girths or collars, short bearing reins, plethora, obesity, open cars and yards, fever, privation of water, heavy fleece. Lesions: right heart and systemic veins full, blood black fluid or diffluent, left ventricle empty, congested meninges, effusions in or on brain, or hæmorrhages. Symptoms: horse: dull, stupid, stubs toes, sways quarters, droops head, hangs on bit, props on feet, breathes rapidly, pants, stertor, dilated nostrils, gasping, fixed eyes, dilated pupils, tumultuous heartbeats, gorged veins, epistaxis, perspiration, convulsions: ox: parallel symptoms: sheep: open mouth, stertor, fixed eyes, pupils dilated, panting, swaying, fall, convulsions: dog: dull, prostrate, pants, congested veins and mucosæ, weakness, spasms, syncope, speedy rigor mortis. Overheating. Diagnosis: early excessive hyperthermia, venous congestion, shallow panting breathing, violent heart action, loss of sensory and motor functions, convulsions. Prevention: avoid violent, prolonged heat, and exertion, especially in case of fat animals or those new to hot climate, keep emunctories acting, shade head, water on head and to drink, protect fat cattle, shear sheep, water. Treatment: shade and laxatives; if severe, cold water from hose, ice bags to poll, rub legs, acetanilid subcutem, stimulant enemata, later mineral tonics, iron or zinc.

Definition. A morbid condition produced by the exposure to extreme heat, and marked by profound disorder of the vaso-motor and heat centres.

The single term of sun-stroke or heat-stroke has been replaced by two,—heat exhaustion and sun-stroke, indicating two distinct conditions, brought about by exposure to heat and manifested by different states of the body and distinctive symptoms.

Heat Exhaustion.

This appears as an exaggerated form of the general sense of relaxation, weakness and languor which follows on prolonged violent exertion in a hot atmosphere. There is more or less impairment of the vaso-motor nerve centre in the medulla, relaxation of the capillary system, and flagging of the heart’s action, which loses its customary stimulus, by reason of the defective supply of blood returned by the veins. This may become so extreme that the patient dies by syncope. In other cases the paresis is mainly shown in the vaso-motor system, and its centres in the medulla, the blood is delayed in the distended capillaries and veins, it becomes overcharged with carbon dioxide, the heart’s action is accelerated and feeble, the pulse rapid, weak and fluttering, perspiration breaks out on the skin, and the temperature is normal or subnormal. The muscular weakness, the flaccid condition of the facial muscles, and general depression suggest a state of collapse. This condition is not necessarily due to exposure to the intensity of the sun’s rays, but may come on in animals subjected for a length of time to artificial heat, and especially if the air is impure, and if the subject has to undergo severe physical exertion.

Treatment. In slight cases of this kind a stimulant is usually desirable and ammonium carbonate in bolus or solution will usually serve a good purpose. In its absence alcohol or spirits of nitrous ether may be given. Digitalis is of great value in sustaining the flagging action of the heart and has the advantage that as digitalin it can be given hypodermically when it is impossible to give ammonia, alcohol or ether by the mouth. For the same reason nitro-glycerine may be resorted to, or even atropia as a vaso-motor stimulant. Active friction of the body and limbs will aid circulation and indirectly stimulate the heart, and in case of subnormal temperature it may be supplemented by a warm bath in the smaller animals, kept up until the normal temperature in the rectum has been restored.

Thermic Fever. Sun-stroke.

This is readily distinguished from heat exhaustion by the predominance of the hyperthermia. While in heat exhaustion the temperature is usually subnormal, in sun-stroke it is excessive, (108°–113° F.).

Causes. The immediate cause of sun-stroke is exposure to undue heat, but this need not be the heat of the sun’s rays direct. A large proportion of cases in the human subject are attacked during the night, and again at sea where an attack in a passenger is practically unknown, it is terribly common among stokers working in a close atmosphere of 100° to 150° F.

The attendant conditions have much influence in determining an attack, thus it is generally held that heat with excess of moisture is the most injurious, yet in Cincinnati, statistics showed a greater number of cases in man when the air was dry. The suppression of perspiration and the arrest of cooling by evaporation in the latter case would tend to a rapid increase of the body temperature, and the condition would be aggravated by the electric tension usually present with the dry air. With the hot, moist air perspiration might continue, but evaporation would be hindered, and there would be arrest of the cooling process and an extreme relaxation of the system.

Again, if is usually found that seizures take place during or after hard muscular exertion in a hot period, and much importance is attached to the attendant exhaustion, the excess of muscular waste, and the alteration of the myosin, which latter coagulates at a lower temperature in the overworked animal. But on the other hand, experiment shows that the animal confined to absolute inactivity in the hot sunshine or in a high temperature (at 90°), dies in a few hours, whereas another animal left at liberty in the same temperature does not suffer materially. The explanation appears to be that the dog, kept absolutely still, has the continuous action of the heat on the same parts and on the same blood, for the capillaries dilate, and the blood is delayed, overheated, and surcharged with carbon dioxide, and the result is either syncope from heart failure, or

asphyxia from excessive carbonization of the blood. Back of these and concurring with them is the paralysis of the vaso-motor and heat generating nerve centres, from the high temperature or the condition of the blood.

The excessive carbonization of the blood deserves another word. The prolonged contact of the blood and air in the lungs is essential to the free interchange of oxygen and carbon dioxide. Vierordt showed that with sixty respirations per minute the expired air became charged with but 2.4 per cent. of this gas, whereas with fourteen respirations it contained 4.34 per cent. Therefore, with violent muscular work (which charges the blood with carbon dioxide) and rapid breathing (which fails to secure its elimination), the overdriven animal soon perishes from asphyxia. Under a high temperature of the external air, this condition is aggravated since the rarefied air contains just so much the less oxygen, the absorption of which is the measure of the exhalation of carbon dioxide.

Dr. H. C. Wood, who has experimented largely on the subject in animals, finds the cause of heart failure in the coagulation of the myosin, which takes place under ordinary circumstances at 115° F., but at a much lower temperature when a muscle has been in great activity immediately before death. As the temperature of thermic fever frequently reaches 113°, or even higher, he easily accounts for the sudden syncope occurring during active work in a high temperature. As an example of such sudden rigor, he adduces the sudden stiffening of the bodies of some soldiers killed in battle during hot weather.

Wood further shows that all the symptoms of thermic fever can be produced in the rabbit by concentrating the temperature on its head, which seems to imply a direct action on the brain and in particular on the heat producing and vaso-motor centres. This becomes the more reasonable that the temperature attained does not impair the vitality of the blood but, leaves the leucocytes possessed of their amœboid motion. He found, moreover, that if the heat were withdrawn before it has produced permanent injury to the nervous system, blood or other tissues, the convulsions and unconsciousness are immediately relieved and the animal recovers.

Other conditions may be adduced as predisposing or concurrent causes of thermic fever. Whatever impairs the animal vigor has this

effect. Fatigue, as already noticed, is a potent factor, in man a drinking habit; in all animals a long persistence of the heat during the night as well as the day; impure air in badly ventilated buildings; and mechanical restriction on the freedom of breathing. In military barracks with the daily temperature at 118° F. and the night temperature 105, the mortality became extreme, and in close city car stables the proportion of sun-strokes is enhanced. In all such cases, the air becomes necessarily more and more impure continually. The atmosphere has the same heat as the animal body, so that no upward current from the latter can be established, to create a diffusion. The carbon dioxide and other emanations from the lungs, the exhalations from the skin, dung and urine, accumulate in the air immediately surrounding the animal and respiration becomes increasingly imperfect and difficult. This condition is further aggravated by the accumulation of the animal heat in the body. The blood circulating in the skin can no longer be cooled, to return with refrigerating effect on the interior of the body, the cooling that would come from the evaporation of sweat is obviated by the suppression of that secretion, as well as by the saturation of the zone of air immediately surrounding the body, and thus the tendency is to a steady increase of the body temperature until the limit of viability has been passed.

The mechanical restriction of respiration should not be overlooked. In European soldiers landed in India and marched in the tight woolen clothing and close stocks a high mortality has been induced and in horses with tight girths or collars and short bearing reins, and oxen working in collars a similar result is observed. Any condition of fever is a potent predisposing factor.

Horses or cattle that are put to violent or continued exertion when too fat or out of condition are especially subject to sun-stroke. Fat cattle driven to market under a hot sun, or shipped by rail, crowded in a car and delayed on a siding under a hot sun, with no circulation of air, often have insolation in its most violent form. The same may be seen in the hot stockyard, with a still atmosphere and the fat animals subjected to the full blaze of a July sun. The chafed feet caused by travel, and the muscular weariness caused by standing in the moving car are material additions to the danger.

Similarly horses suffer on the race track when subjected to protracted and severe work in hot weather, or again dragging loads

in a heated street under a vertical sun, or on a side hill with the sun’s rays striking perpendicularly to its surface.

A change in latitude has a decided effect, the Northern horse suffering much more frequently than the one which is native to the Southern States and which has inherited the habit of heat endurance.

Finally faults in feeding and above all watering are appreciable factors. The privation of water in particular is to be dreaded. Tracy in his experience with American soldiers in Arizona, found that the command could usually be guarded against sun-stroke when a supply of water was kept on hand. It should be used guardedly, but nothing would act better in obviating an attack. On the other hand, when the canteens were empty, under the hot sun the seizures increased disastrously.

Sheep are especially liable to suffer from heat by reason of their dense fleece, which hinders the evaporation of perspiration, and the cooling effect of air on the skin. When the temperature rises, respiration is accelerated and panting, the lungs seeking to supplement the work of the skin. When traveling in a heavy fleece, or in the hot sunshine in July or August sun-stroke is not uncommon among them.

Lesions. Among the lesions may be named, vacuity of the left ventricle and fullness of the right ventricle and veins with fluid blood or a diffluent clot; congestion of the pia or dura mater, effusion into the ventricles, hæmorrhages into the subserous tissues, and degeneration of the muscles.

Symptoms. Horse. When premonitory symptoms are observed the animal fails to respond to whip or voice, lessens his pace, stubs with his fore feet and sways with the hind, depresses his head and hangs heavily on the bit.

Too often these are omitted or overlooked, and the horse suddenly stops, props himself on his four limbs, drops and extends the head, breathes with great rapidity, panting and even stertor, dilates the nostrils widely, retracts the angle of the mouth and even gapes, has the eyes fixed, the pupils dilated and the beats of the heart tumultuous. The superficial veins are distended, the visible mucosæ congested with dark blood, and blood may escape from the nose. Perspiration usually sets in.

The animal may fall and die in a few minutes in convulsions, or, if stopped sufficiently early and suitably treated, he may in a measure recover in 15 to 20 minutes.

Symptoms. Ox. The premonitory symptoms are like those in the horse: dullness, rapid, panting breathing, the mouth is opened and the pendent tongue is covered with frothy saliva, a frothy mucus escapes from the nose, the eyes are congested and fixed, the pupils dilated, the nostrils and flanks work laboriously, the heart palpitates, the animal sways or staggers and falls. Death follows in convulsions, or it may be delayed, the animal struggling ineffectually to rise, or having periods of comparative quiet. The rectal temperature is very high, 107° to 114° F. If able to stand, there is usually blindness and heedlessness of surrounding objects.

Symptoms. Sheep. The open mouth, protruding tongue, frothy saliva, reddened fixed eyes, rapid breathing, beating flanks, stertor, and unsteady gait are characteristic when taken along with the manifest causes. Swaying movements followed by a sudden fall and death in convulsions form the usual termination of the disease.

Symptoms. Dog. These have been mainly produced experimentally and consisted in hyperthermia, dullness, prostration, accelerated breathing and heart action, congested veins, and mucosæ, muscular weakness, convulsions, and syncope or asphyxia. After death the muscles became speedily rigid, and the blood accummulated in the venous system, was fluid or only loosely coagulated. In these animals, if the experiment were stopped in time the animal could be restored to health.

Slighter cases may occur in the different animals, more particularly from overdriving in hot weather, and in such cases the overheated animal recovers, but there is liable to remain a special sensitiveness to excessive heat and a tendency to be dull, sluggish and short winded, to hang the head in hot weather, and to seek shelter from the direct rays of the sun.

Diagnosis is largely based on the suddenness of the attack, on the occurrence of high temperature before the seizure, not after as it is liable to be, if at all, in apoplexy, on the dark congestion of the mucosæ, and of the venous system, on the rapidity and shallowness of the respirations, on the tumultuous action of the heart, and on the general loss of sensory and especially of motor function, in

circumstances calculated to induce sun-stroke. Localized paralysis or spasm would suggest the formation of a cerebral effusion or clot.

Prevention. This will depend on the class of animal and its conditions of life and work. In horses care should be taken to regulate the work by the heat of the season and condition of the animal. When the temperature ranges from 80° to 100° F. the work should be lessened and every attention should be given to maintain the healthy functions (bowels, kidneys, skin) in good working condition. If the horse is young, fat, or out of condition from idleness or accumulation of fat he must have the greater consideration. So it is with a horse recently come from a colder latitude, and with a heavy draught horse that may be called on to do rapid work. Some protection is secured by wearing a sunshade or a wet sponge over the poll, and much may be expected from an occasional rest in the shade, a swallow of cool water and sponging of the head.

Very heavy fat cattle should not be driven far nor shipped on the hottest days, and the packed car should not be left in the full sunshine in a still atmosphere. Yards with sheds under which they can retreat must be secured if possible.

The heavily fleeced sheep must have equal care and the pastures for fat sheep and cattle should have available shade in form of trees, walls or sheds. Access to water is an important condition.

Treatment. In slight cases (overheated) a few days of rest, under an awning rather than in a close stable, with a restricted and laxative diet.

In severe thermic fever the first consideration is to lower temperature. If available turn a hose on the head, neck and entire body for five or ten minutes, or until the rectal temperature approaches the normal. In the absence of such a water supply, dash cold water from a well on the body but especially the head and neck, and if available tie a bag of ice around the poll. Active friction to the legs and body is often of great advantage. A large dose of antipyrin or acetanilid may be given hypodermically. On the other hand stimulants, and especially carbonate of ammonia, or sweet spirits of nitre may be given as an enema. This may be repeated in an hour in case the pulse fails to acquire force and tone.

Should the temperature rise again later it may often be kept in check by cold sponging and scraping followed by rubbing till dry.

In case of continued elevation of temperature, with heat of the head, and perversion of sensory or motor functions, meningitis may be suspected and appropriate treatment adopted.

For the prostration and weakness that is liable to follow thermic fever, mineral tonics such as the salts of iron or zinc may be resorted to.

EPILEPSY. FALLING SICKNESS.

Definition. Frequency. Susceptibility: dogs, pigs, cattle, horses, parrots, sparrows. Divisions: slight and severe: Jacksonian (partial): symptomatic; idiopathic. Lesions: inconstant: of brain, cranium, cerebral circulation, myelon, poisons in blood, dentition, cortical and ganglionic lesions, cerebral asymmetry, stenosis of vertebral canal. Medullar asymmetry, traumas of cranium, anæmia, bleeding, carotid ligation, spinal reflexes, irritation of skin, creatinin, cinchonoidin, lead, ergot, nitro-pentan, nitro-benzol, ptomaines, toxins, parasites, nerve lesions, local hyperæsthesia (withers of horse, recurrent ophthalmia), indigestion, constipation, sciatic neuritis. Causes: nervous predisposition, heredity (man, cat, dog, ox), sexual excitement, fear, sudden strong visual impression, uric acid in blood, meat diet. Symptoms: horse, sudden seizure, bracing feet and limbs, swaying, fall, convulsive rigidity, jaws working or clenched, eyes rolling, salivation, stertor, dyspnœa, sensation absent. Duration. Symptoms of localized epilepsy. Cattle, bellow, stertor, rolling eyes, jerking, rigidity, fall. Sheep. Swine premonitory malaise, jerking, champing jaws, fall, trembling, rigidity, involuntary discharges. Dog trembles, cries, falls, rigidity, clonic contractions, stertor, sequelæ. Diagnosis: sudden attack, unconsciousness, spasms, quick recovery, no spasms in syncope, vertigo has no spasms, thrombosis has symptoms developed by exercise. Jurisprudence: animal returnable after twenty-eight days (Wurtenberg, etc.,) thirty days (France). Treatment: of susceptible brain, and peripheral irritant. Correct all irritation or disease, or expel parasites. Nerve sedations: bromides, opium, valerian, belladonna, hyoscine, duboisine. Tonics: zinc, arsenic, silver, baths, electricity. Borax. Vegetable diet. Castration. Avoidance of excitement. Surgical operations. Trephining. Excision of cortex. Outdoor life. During a fit: amyle nitrite, chloroform, ether, chloral, warm bath, cold or warmth to head, quiet secluded place.

Epilepsy is the name given to a class of cases characterized by a sudden and transient loss of consciousness with a convulsive seizure, partial or general. It appears to be due to a sudden explosive discharge of convulsive nervous energy, which may be generated by a great number of causes of morbid irritation—pathological, traumatic, or toxic. As a rule the epileptic seizure is but the symptomatic

expression of a complex derangement which may be extremely varied as to its nature and origin.

Frequency in different animals. The affection is far less frequent in the domestic animals than in man, doubtless because of the absence of the special susceptibility which attends on the more highly specialized brain, the disturbing conditions of civilization, and the attendant vices.

Among domestic animals, dogs are the most frequent victims in keeping with their relatively large cerebral development, their emotional and impressionable nature and the unnatural and artificial conditions in which as house pets they are often kept. Their animal food and the consequent uric acid diathesis is a probable cause, as it is in man. In ten years of the dog clinic at Alfort they made an average of 3 per cent. of all cases. Next to the dog the pig kept in confinement is the most frequent victim, while cattle and horses come last. At the Alfort clinic epileptic horses were not more than 1 per 1000 patients. It is not at all infrequent in birds, especially canaries and parrots. Reynal has seen it in sparrows.

Divisions. The disease has long been divided into petit mal and grand mal (haut mal). The petit mal (slight attack) is usually a transient seizure affecting a group of muscles only and associated with only a momentary or very transient loss of consciousness. The loss of consciousness is uncertain as to many cases. Under partial epilepsies must be included the hemi-epilepsy, or Jacksonian epilepsy, which is confined to one side of the body.

The grand mal (severe attack) is one in which the loss of consciousness is complete, and the convulsions are general in the muscles of animal life.

Another division is into symptomatic and idiopathic cases, and if this distinction could always be made it would be of immense value in the matters of prognosis and treatment as the removal of the morbid state of which epilepsy is the symptom will usually restore the patient to health. Thus the removal of worms from the alimentary canal, of indigestible matters from the stomach, of a depressed bone or tumor from the surface of the brain may in different cases be the essential condition of a successful treatment.

Morbid Anatomy and Pathology. The literature of epilepsy is very rich and extensive and yet no constant lesions of the nervous system can be fixed on as the local cause of the disease. A review of the whole literature leads rather to the conclusion that irritations coming from lesions of the most varied kind, acting on a specially susceptible brain will rouse the cerebral centres to an epileptic explosion. Thus epilepsy has been found to be associated with lesions of the following kinds:

1st. Brain lesions of almost every kind, including malformations.

2nd. Lesions of the walls of the cranium.

3d. Disorders of the cerebral circulation.

4th. Lesions of the spinal cord.

5th. Morbid states of the circulating blood (excess of urea, uric acid, creatinin, lead poisoning).

6th. Reflected irritation, as from dentition, worms, sexual excesses, injuries to certain nerves, notably the sciatic, or to particular parts of the skin.

1st. Brain lesions. Those which affect the medulla and the cortical convolutions around the fissure of Rolando would be expected to be implicated because these centres preside over the principal motor actions of the body and limbs. Yet though these parts are found to be affected with various morbid lesions in a certain number of cases of epilepsy, such lesions are exceptional, rather than the rule. In 20 cases of epilepsy in man, 15 showed no lesion whatever of the brain. Blocq and Marinesco, pupils of Charcot, recently made a critical examination of the medulla and Rolandic cortex in nine cases that died during the fit. All showed granular bodies (degenerated myelin or blood pigment) in the perivascular sheaths but they found these in disseminated sclerosis and even in healthy brains as well. The neuroglia cells of the first cortical layer contained black granules. Otherwise four cases had no change, while five showed sclerosis of the cortex. The medulla was sound in all cases excepting one which showed punctiform hæmorrhages. Visible lesions may be present in other parts of the brain; Wenzel long ago claimed constant lesion of the pituitary body. Beside the cerebral cortex, lesions have been found in the bulb, the hypoglossal nucleus, the olivary body, the hippocampi, the thalamus, the corpus striatum,

the quadrigemini, the cerebellum, etc. Hughlings-Jackson who made an extended investigation of the subject concludes that any part of the gray matter of the encephalon may become over-excitable and give rise to a convulsive attack. Not only may the lesion be in any part of the brain, but it may be of any kind: meningitis, cerebritis, softening, tubercle, tumor, hydatid, embolism, or dropsy. Marie Bra found an extreme asymmetry of the cerebral lobes in epileptics. Kussmaul and others found stenosis of the vertebral canal and asymmetry of the two lateral halves of the medulla.

2nd. Cranial lesions. These consist largely in blows or falls upon the head, with osteitis, periostitis, fractures with depressions, fibrous neoplasia implicating or not the meninges and pressing on the brain, hæmorrhages from minute arteries, etc. The diagnosis of such lesions will often open a way to a successful treatment. Baker found most of the severe cases from head injuries.

3d. Disorders of the cerebral circulation. Burrows, Kussmaul and Turner showed that in animals, loss of consciousness and epileptiform convulsions followed on cerebral anæmia caused by profuse bleeding or by compression of the carotids. The same has been observed in surgical cases after ligation of one common carotid. Hermann caused convulsions in a rabbit by ligating both anterior and posterior venæ cavæ.

4th. Lesions of the Spinal Cord. Brown-Sequard determined epileptiform convulsions by transverse section of one-half of the spinal cord, or of its superior, lateral or inferior columns. The later development of the doctrine of interrupted spinal inhibition, suggests that, many of the seizures in question are but exaggerated spinal reflexes, which are no longer restrained by cerebral inhibition. That all are not of this spurious kind may be fairly inferred from his further demonstration that bruising of the great sciatic in animals tended to produce epilepsy. In such cases., the irritation of certain areas by pinching the skin, served to produce a seizure. Not only so, but the animals in which such artificial epilepsy had been induced tended to transmit the infirmity to their progeny. The prevailing view of epilepsy however, would consider such lesions as sources of peripheral irritation by which the brain is affected sympathetically, while the real explosion is the result of the sudden discharge of the

pent up excitement caused in the encephalic centres by the irritation at such distant points.

5th. Morbid States of the Circulating Blood. Certain poisons, when brought in contact with encephalic nerve centres produce epileptic seizures. Gallerani and Lussana applied creatinin directly to the cerebral cortex and quickly induced epileptiform convulsions and choreiform movements. Injected subcutaneously it failed to produce the same effect. Cinchonoidin acted on the basal ganglia of the brain producing convulsions but no choreiform movement. Poisoning with lead, ergot, nitro-pentan, nitro-benzol and a number of other poisons brings about intermittent convulsive seizures. The same may be inferred of ptomaines and toxins, in the convulsions that appear in the advanced stages of infectious diseases (canine distemper, hog cholera, etc.).

6th. Reflex Irritation. Perhaps no peripheral irritation more frequently causes epilepsy, than parasites. In young dogs worms in the intestines (tænia cœnurus, tænia tenuicollis, tænia serrata, tænia echinococcus, and ascarides) have been especially incriminated. Also linguatula tænioides in the nasal sinuses. In young pigs the echinorrhynchus gigas, ascarides and trichocephalus. In horses ascarides have been principally blamed.

Wounds implicating nerves, and tumors pressing on nerves, have served as sources of nervous excitement which accumulates in the cerebral ganglia and bursts forth as an epileptic explosion. Bourgelat mentions the case of a horse which fell in a fit the moment he was touched on his tender withers, also a case in which a seizure coincided with an attack of recurrent ophthalmia. Gerlach saw a horse which had an epileptic fit the instant he was touched on his sensitive withers. In kittens and puppies the irritation attendant on dentition is a common cause of attacks. In nervous dogs and pigs indigestion or constipation may serve as the occasion of an explosion. In the experimental cases of Brown-Sequard, not only did the injury to the sciatic nerve develop in the brain a latent tendency to epilepsy, but the subsequent pinching of the skin in certain areas (epileptigenous zones) promptly brought about a seizure.

Causes. Most of the causes of epilepsy have been given above under the head of pathology and morbid anatomy. The nervous predisposition may, like any other peculiarity or function, become

hereditary. In the human race nothing is more certain than the tendency to some form of nervous disorder (insanity, dementia, alcoholism, morphinism, epilepsy, chorea, etc.) in a special family line. Reynal records the case of an epileptic cat (belonging to an employe of the Alfort veterinary school) the progeny of which for three generations, became affected with epilepsy and mostly died before they were a year old. Also four epileptic dogs (3 males and 1 female) which produced a number of epileptic puppies. LaNotte records the cases of two bulls affected with epilepsy, in the progeny of which numerous cases of epilepsy appeared; the cows being attacked after the first calving, and the oxen soon after they were first put to work. Breeding stallions are particularly liable to attacks, the high feeding, lack of muscular work in the open air, and above all the oft repeated nervous excitement attendant on copulation being directly exciting causes. The heredity of the artificial epilepsy induced by Brown-Sequard in Guinea pigs, serves to strengthen the doctrine of heredity in ordinary forms.

Among emotional causes fear easily heads the list. Bernard states that a horse became epileptic in connection with the terror caused by the giving way of a wooden bridge over which he was passing. Bourgelat and Reynal adduce instances, in cavalry horses when first put under fire. Reynal records the case of another which had his first attack when facing a moving locomotive, and which never again could see an engine in motion without suffering another attack. La Notte mentions the case of a horse attacked when frightened by a sky rocket; Romer, the case of a horse scared by the sudden display of a white sheet in front of him, and Friedberger and Fröhner relate cases of attacks caused by intense rays of light, as in racing toward the declining sun, or the dazzling reflection from the surface of water. Liedesdorf saw it in a dog scared by a locomotive.

A strong impression like that caused by transition from bright light into darkness, by seeing shadows of trees crossing the road, or violent suffering caused by severe forms of constraint have been named as causes.

Speaking in “Brain,” of epilepsy in man, Alexander Haig attributes the fits to the fluctuations of uric acid in the blood. Headache (migraine) he finds to be very closely allied to epilepsy and convulsions and to be a result in a susceptible system of a liberal

flesh diet. By a vegetable and fruit diet he reduces the ingestion and formation of uric acid, so that the largest quantity which a patient is likely to get into his blood, shall never or only very rarely, affect the blood pressure and increase the intracranial circulation to a dangerous extent. In predisposed subjects, all flesh food, soup, and meat extracts must be avoided, while even tea, coffee, cocoa and other vegetable articles containing xanthin compounds are to be regarded as producing uric acid, and to be denied, or employed only as the merest flavoring.

This position is greatly strengthened by the fact that epilepsy is so much more frequent in the carnivora (dog, cat, bird) than in the herbivora. It also suggests very strongly a light vegetable diet for both prophylactic and curative purposes in our domestic animals. In the same line the frequent and liberal drinking of warm water, the use of diuretics and the flushing of the large intestine are indicated.

For other causes see under pathology.

Symptoms in the Horse. It has been claimed that premonitory symptoms, such as dullness, lack of energy and quick, nervous or startled movements herald an attack, but in animals as in man, the disease usually attacks suddenly without any antecedent indication. If at work the horse stops suddenly, or if in the stable he ceases eating, seems frightened, stands for an instant immovable, braces his feet, sways, trembles, and falls heavily to the ground. Or he may remain for an instant supported on his rigid limbs, the jaws moving or firmly closed, the eyes rolling, and the facial muscles drawn or twitching. When down there are convulsive movements of the limbs, so that the animal may kick out violently, and tense contractions or twitchings may occur in the muscles of the croup, chest and abdomen. There is usually an increase of the salivary secretion with frothy accumulation about the angles of the mouth. The respiration is stertorous, dyspnœic, and interrupted, the nostrils widely dilated, the nasal mucosa of a dark brownish red, and the superficial veins distended. The pulse is weak, slow, irregular, intermittent and sometimes imperceptible. Sensation seems to be in abeyance. No attention is paid to loud sounds, nor to pinching, pricking, or even cauterizing the skin. Perspirations may break out on the flank or over the whole surface of the body.

The duration of the attack may be from one to four minutes, or exceptionally ten or fifteen, after which the muscles relax, the twitching ceases, the horse raises his head, extends his fore limbs and finally rises.

After rising some are dull and stupid for an hour or so, and may continue to perspire, some move the limbs, jaws or head automatically, turn in a circle, or seek seclusion and darkness, while some take at once to eating and seem as if nothing had happened.

In partial or localized epilepsy the spasms are confined to a limited group of muscles like those of the jaws, neck, or fore limbs. These may alternately contract and relax, or they may remain rigid for a minute or less, the mouth being held open or firmly closed with grinding of the teeth, the eyes rolled backward and upward, or affected with strabismus, the face drawn and distorted, the head turned to one side or downward, or the limbs fixed and immovable.

At the conclusion of an attack it is not uncommon to see a discharge of urine or fæces, or in stallions, of semen.

The horse often contracts a fear of the place where the attack occurred, and this contributes, with the re-appearance of the former object of dread (car, locomotive, rifles, cannon, etc.) to precipitate a new attack if he is compelled to go to such a place.

Symptoms in Cattle. In cattle the animal is attacked without premonition, bellows, breathes hard and with effort, has dilatation of the nostrils, and squinting or rolling upward and backward of the eyes and falls to the ground rigid and trembling. There may be violent succussions of the limbs, head or neck, movements of the jaws, grinding of the teeth, and the appearance of frothy saliva and elements of food about the lips. The beats of the heart are violent, the pulse slow and small, and sometimes intermittent. Involuntary micturition, defecation, or discharge of semen may occur. In slight cases one or more of these symptoms may be absent, and the victim may not even fall to the ground but support himself against a wall or other object.

The duration of the attack may be from one to five minutes, rarely more, and there is often a slow and progressive subsidence of the spasms. When recovered the animal may get up and go to eating or rumination as in health.

Symptoms in Sheep. In sheep the attack is sudden. The animal ceases eating or stops in its walking, and after turning or other involuntary movement falls to the ground, head extended, mouth open, eyes rolling or squinting, and with rigidity or twitching of the muscles of the neck or limbs. There is the same loss of sensation, frothing from the mouth, and grinding of the teeth as in the larger animals. The attack may last 40 to 50 seconds.

Symptoms in Swine. In pigs a state of discomfort and restlessness often marks the approach of an attack, referable probably to the digestive disturbance or to parasites which furnish the occasion of the disease. Uneasy, wandering movements, jerkings of head or limbs, rolling of the eyes, and champing of the jaws may first appear. Then the animal falls, extending its limbs and head, with open mouth, retracted lips, and a free flow of saliva. Trembling and jerking of the head, neck and limbs, hurried, short, difficult breathing, and complete loss of sensation may be noted. Discharges of urine, semen, and prostatic fluid are not uncommon. The attack usually lasts 2 or 3 minutes, and exceptionally 10 to 15. In the shorter seizures, frequent repetition is not uncommon, Delafond having observed 5 or 6 attacks in the course of an hour.

Symptoms in Dogs. The attack is sudden and unheralded by prodromata. The animal stops, trembles, cries plaintively and falls; he may manage to rise or to do so in part but instantly falls anew. The limbs stiffen, tremble or twitch, the head is extended or flexed, or jerked, violently striking the ground, the mouth open, with abundant saliva, or firmly closed though the tongue may be between the teeth. The trunk may be firm and rigid or alternately twisted in one direction or the other. The eyes roll or squint, and the breathing is stertorous and difficult. Insensibility is complete. Toward the end of the attack there may be a discharge of urine, fæces or semen, the stools often containing worms. The body is often wet with perspiration during or after an attack.

The attack usually lasts for two or three minutes, then the convulsions gradually lessen in intensity and finally cease, the dog raises his head, opens his eyes, and gazes inquiringly around. Then he gets on his feet shakes himself and may at once resume his customary habits. In other cases the restoration is less sudden. The dog remains for 30 to 60 minutes dull and stupid, or seems to have

little power of control over its muscles and staggers as if intoxicated, or as if the muscles were benumbed. It may drop on its knees and then fall with the head on the ground and repeat this several times. In other cases the dog wanders around, or trots off and may snap at any one interfering with him, so that the case is often mistaken for one of rabies. Finally the animal may remain prostrate and fall into a deep sleep marked by stertorous breathing.

Diagnosis. The diagnosis of epilepsy is usually easy. The suddenness of the attack, the loss of consciousness, the muscular spasms, the complete temporary recovery and the tendency to recur, form a toute ensemble, which is pathognomonic. The danger of confounding this with other nervous disorders is on the whole greatest in the slight cases in which the symptoms are less typical.

From Syncope it is easily distinguished by the spasms which are not present in syncope.

From eclampsia it is not so easy to diagnose, but the line between eclampsia and epilepsy has not been accurately drawn, and some have even shown a disposition to drop eclampsia from medical nomenclature. Eclampsia may be defined as general convulsions dependent on some eccentric irritation, and which do not recur after such irritation has been removed. This would remove from the category of epileptic attacks the cases of convulsions in which the attacks were due to intestinal or nasal parasites, dentition irritation, tumors pressing on nerves, canine distemper and other infectious diseases. So far the distinction might be made by the diagnosis of the particular disease on which the convulsions depend. There remains however a class of cases in which the centric nervous disorder on which the epileptic seizure depends is present, and also the peripheral source of irritation (worms, etc.). In such a case the presence of the worms or other eccentric source of irritation, even if added to the fact that this was the immediate exciting cause of the epileptic explosion, could not do away with the fact that the essential conditions of epilepsy are permanently present in the nervous centres. The difficulty therefore of making an accurate differential diagnosis, resides largely in the impossibility of drawing a definite line of pathological separation between eclampsia and epilepsy.

From Vertigo epilepsy is distinguished by the absence in the former of marked spasmodic contractions. It is only in the milder

forms of epilepsy those in which the spasmodic action is so slight as to be overlooked, that this disease can be confounded with vertigo.

From Thrombosis or embolism of the iliac or femoral arteries epilepsy is easily distinguished by the absence of exercise as the essential cause in the development of the latter. In thrombosis on the other hand, the loss of control over the hind limbs is developed at will by active motion (walking, trotting). In thrombosis too the absence of pulsation at the fetlocks or at any point below the seat of obstruction is conclusive.

Question of Soundness and Jurisprudence.

Manifestly a horse or bull subject to attacks of epilepsy is not sound. It is moreover a disease, the symptoms of which are only shown for a very short period at one time, after a long interval of apparently perfect health. It is, therefore, a disease against which a purchaser cannot be expected to protect himself and he should have the right to annul the sale and return the animal in case the infirmity should appear within a reasonable period after purchase. This is provided for in the laws of different countries of Europe, thus in Wurtenberg, Baden and Hesse, a purchased animal may be returned within 28 days; in France within 30 days, and in Bavaria within 40 days. The greatest difficulty arises from the frequent impossibility of obtaining expert testimony on a seizure which is likely to occur at any moment, without premonition, and in which the testimony of a non-expert may easily be misleading. It seems as if complaint having been made within the specified legal time, an extension of guarantee should be given by the court, the animal to be meanwhile kept under the supervision of a veterinarian.

Another question has arisen as to the position of an animal suffering from reflex epilepsy. If the attacks are caused by intestinal worms or nasal acarina which are easily removed, it is quite evident that this cannot be considered as a permanent unsoundness, and one for which a contract of sale can be justly annulled. But on the other hand, while the eccentric source of irritation which is easily curable may have been the active agent in developing the seizures, it may be none the less true that the central infirmity which determines the abnormal susceptibility, to excessive generation and epileptic explosion of nervous force, may also be present and the animal cannot be considered as sound until a sufficient length of time has

elapsed after the removal of the peripheral irritation and no new seizure has taken place.

Treatment of Symptomatic Epilepsy. In cases due to an eccentric irritant the first step must be the removal of such irritant. In case of intestinal worms the various vermicides and tæniacides must be resorted to. (See Intestinal Parasites). For the linguatula tænioides the injection of benzine or tobacco water into the nose, or into the sinuses, with or without trephining may be resorted to. In diseased teeth extraction or filling may be demanded. In dentition irritation, lancing of the gums. In all other cases in which a peripheral nervous irritation can be traced every available means should be taken to remove it.

Treatment of Central Epilepsy. Bearing in mind that peripheral irritation is a frequent exciting cause of a seizure, too much care cannot be given to the conservation of the general health and especially to make the diet wholesome in quantity, quality and time of feeding and watering, and to guard against constipation and indigestion. In dogs a too stimulating meat diet is to be avoided.

The medicinal agents employed have been mainly such as are sedative, or tonic to the nervous system. Valerian was long extolled as a valuable remedy (Gohier, Delafond, Delwart), and this has been improved upon more lately by substituting valerianate of zinc. Belladonna and its alkaloid atropia have been strongly advocated (Tisserant, Bernard, Williams, Friedberger) and it has the recommendation that it causes vaso-motor contraction and tends to lessen cerebral congestion. Hyoscine or duboisine may be used as a substitute. Cyanide of iron has been lauded by Jourdier and Tabourin, as far superior to valerian. Of late years the nerve tonics, zinc compounds (oxide, sulphate, chloride) and silver salts (nitrate) and arsenic have been used, often with excellent results. Borax strongly recommended for man (1 to 1¾ drachm daily) by Pastena is worthy of a trial for dogs. It is given largely diluted in syrup to avoid gastric irritation.

Of all agents employed up to the present the bromides still claim a foremost place. They should be given in a large dose, on an empty stomach and at such a time as to occupy the system at the hour when the seizure is expected to recur. Thus for morning attacks the dose may be given at night, while for night attacks it may be given in the

afternoon. Müller uses sodium bromide in the dog as least liable to disturb the stomach, while Peterson, for man, advises the potassium salt for the same reason. For man, McLane Hamilton advocates a combination of the sodium and ammonium salts, Eulenberg adds the potassium compound, while Berkley uses strontium bromide, and Bourneville camphor monobromide.

Given at night in full dose (30 grs. for dog) the bromides tend to secure a quiet sleep, with brain rest and recuperation. If beneficial they should be repeated daily until a cure or other sign of bromism appears. This may be somewhat checked by arsenic or chloral hydrate.

Wesley Mills finds potassium iodide useful in some dogs when bromides fail. Bromohydrate is advocated by Müller. Flechsig and others have had excellent results in man from the opium bromide treatment. Full doses of opium are given three times a day for six weeks, when they are replaced by full doses of bromides four times a day.

Improvement should be shown in the shortening of the convulsions and the lengthening of the intervals between them. Should the bromides fail in this, resort may be had to other treatment.

Toulouse, Clark and others find that privation of salt, in man, allows the bromine salt to replace the chlorine one in the tissues, and the hydrobromic acid the hydrochloric in the gastric juice, and in this way the bromine can be introduced safely in larger amount into the tissues and is longer retained, though given in half the doses.

A most important element in the treatment is a vegetable diet with or without milk, to obviate excessive production of uric acid. Anything which will disagree and produce gastric or intestinal fermentations with toxins must be carefully guarded against and these will differ in different individuals.

Stallions and other excitable males, and females may often be cured by castration. Patients should be very carefully guarded against all sources of excitement, reports of guns, sight of locomotives or automobiles, waving flags, instrumental music, sudden exposure to sunshine or other bright light, reflection from water, snow, or ice, the contrast of dark shadows, as of trees,

alternating with bright light, etc. Dogs, becoming excited at a show, may have a convulsion if not removed, and much more so in presence of another dog in a fit.

A surgical operation often places the disease in abeyance for many months, but, unless in the case of the removal of a diseased organ which has acted as a factor, this is not permanent. Hence in man transient benefit has been secured from operations on the eyes, the brain, the testicles, the ovaries, etc. In local (Jacksonian) epilepsy, which can be traced to a definite cortical area in the brain, the trephining of the skull and the excision of the cortex at that point, has given temporary relief, with a local palsy, but too often the irritation from the resulting cicatrix has in time aroused the disorder anew. Even independently of the removal of the cortex, the trephining has been successfully resorted to, by savage as well as civilized peoples, securing a temporary relief. Though not in practice in veterinary medicine it seems as if this were even more applicable than in man. It would be fully justified if it preserved for a year or more an animal in usefulness which must otherwise be destroyed, even if the disease should return at the end of this time.

Plunge or douche baths (60° to 70° F.) and rubbing dry will often tone up the nervous system, and a course of bitters, or iron, or both, may prove valuable. An outdoor life and moderate muscular exercise are important.

During a convulsion the animal should be freed from all harness, halters, girths, etc., that would impair respiration, the jaws may be kept apart with a cloth to prevent biting the tongue, and the animal held with head and neck in natural position. To arrest the spasms the best agent is amyle nitrite inhaled from a handkerchief. It may be replaced by a mixture in equal parts of chloroform and ether. Or rectal injections may be given of chloral. Nitro-glycerine will sometimes cut short an attack or prevent it. Small animals may have the body immersed in a warm bath, and cold applied to the head. Congested buccal and conjunctival mucosæ would indicate cold to the head, while pallor would suggest warm fomentations.

When the fit is over the animal should be kept in a quiet, dark place until the excitement or stupor has completely passed.