Estimation of Potato Yield in and Around Munshigonj Using Remote Sensing and GIS techniques S. K. Bala1, Md. Ali2 and A.K.M.S. Islam3
Abstract Monitoring of the crop growth and forecasting its yield well before harvest is very important for an agricultural country like Bangladesh for crop and food management. This enables planners and decision makers to take decision on import or on export in due time. It also helps the government to plan for redistribution of food during times of disaster. So, early yield prediction together with monitoring of crop development and its growth are generally important. Conventional techniques for yield estimation in many countries are based on ground-based field visits, data collection for crop and reports. These are often subjective, costly, time consuming and bear large errors for incomplete ground observations. Satellite and remote sensing images are capable of not only identifying crop classes but also predicting crop yield. In recent years, studies using remote sensing data done at field level reported high correlation between Normalized Difference Vegetation Index (NDVI) and yield. Assessment and monitoring of vegetation parameters like NDVI and crop vigor and green biomass may be ascertained through use of remote sensing. Therefore, NDVI can be used to estimate yield before harvesting.
In this light, a study is being carried out to estimate crop yield of potato at field level by applying remotely sensed Normalized Differential Vegetation Index (NDVI). The study has been conducted on Munshigonj area which is the main potato producing district of Bangladesh. The main objectives of the study were estimation of the spatial and temporal pattern of the crop (potato) productivity in and around Munshigonj district and establishment of a relationship between NDVI of potato crop and field level crop yield of potato. Satellite images used for the study include NOAA AVHRR (resolution 1.1 km) and TERRA MODIS (resolution 250m). NOAA AVHRR images are received from Space Research and Remote Sensing Organization, Bangladesh (SPARRSO), while TERRA MODIS images are downloaded from internet. Groundtruthing of farmerâ€™s field has been made by Global Position System (GPS) survey. Satellite image has been processed and analyzed by using ERDAS IMAGINE software. Land use data for potato and yield during December 2005 to April 2006 are collected from fields. The satellite
image is processed in a GIS environment using ArcGIS software and an NDVI map is generated using band 3 (NIR) and band 2 (Red). The relationship between NDVI and yield data for potato at field level is established through regression analysis.
The correlation coefficients are
calculated between yields and NDVIs. Accordingly, the predictive yield equation of potato as a function of NDVI is generated from regression analysis. The functional relations of weather data with potato NDVI and yield are also investigated. The yield of potato for the year 2006 for different administrative units of Munsigonj district is compared with that of Bangladesh Bureau of Statistics (BBS). The results showed that there is significant correlation between remotely sensed NDVI and field level potato yield. It was found that NDVIs are also capable to explain the variability of potato yield at field level. Hence, remotely sensed NDVI data may be an effective tool for early prediction of potato yield as well as production deficit, which are very important for planners to prevent disaster like famine in a country.
Key words: AVHRR, crop yield, ERDAS IMAGINE, GPS, NDVI, TERRA MODIS
Associate Prof., Institute of Water and Flood Management (IWFM), Bangladesh University of Engineering and Technology (BUET), Dhaka-1000, Bangladesh, e-mail: firstname.lastname@example.org
M. Sc. Student & Research Assistant in the Research Project, â€œEstimation of Potato Yield in and Around of Munshigonj Using RS and GISâ€? under IWEM, BUET, Dhaka-1000, Bangladesh, e-mail: email@example.com
Assistant Prof., IWFM, BUET, Dhaka-1000, Bangladesh, e-mail: firstname.lastname@example.org