Agricultural and meteorological drought assessment using remote sensing and GIS in North - West region of Bangladesh

Abstract

Drought is the most complex but least understood of all natural hazards in Bangladesh. Timely information about the onset of drought, extent, intensity, duration and impacts can limit drought related losses of life, human suffering and decrease damage to economy and environment. In this study an attempt has been made to apply RS and GIS techniques in the field of drought detection. An effort has been made to define drought risk areas facing agricultural and meteorological drought by using temporal images from MODIS terra surface reflectance (resolution 250m) based Normalized Difference Vegetation Index (NDVI) (2000-2008) and meteorological information based Standardized Precipitation Index (SPI). Correlation and regression analyses were performed with NDVI, SPI and Food grain yield anomaly. Analysis shows that NDVI anomaly and SPI are significantly (R²>0.36) correlated in 13 districts of the 16 districts of N-W region of Bangladesh. The correlation is highly significant (R²>0.60) in 5 districts. Correlation between NDVI anomaly and crop yield anomaly is significant (R²>0.26) in 8 districts and highly significant (R²>0.67) in 3 districts. SPI values were interpolated to determine the spatial pattern of meteorological drought. Threshold value for different types of drought frequency was defined from McKee et. al.’s (1993) drought classification table, which was used to derive meteorological drought risk. Food grain yield trend was plotted and an equivalent NDVI threshold was identified to get the agricultural drought risk. Results of correlation and regression analysis with SPI and crop yield shows that SPI can be used as an indicator of regional crop production. Since each of the factors NDVI, SPI and detrended food grain yield anomaly has positive linear correlation with each other these factors can be effectively used for monitoring and assessing the food grain production, and thereby appropriate agricultural practices can be adopted to minimize drought effects. From the interpolated SPI and NDVI anomaly maps, the highest levels of frequency in agricultural drought was 5, 3 and 2 in three classes; slight, moderate and severe drought, respectively and found in three districts, Dinajpur, Panchagarh and Thakurgaon. In case of meteorological drought the highest level of drought frequency was found in four districts, Dinajpur, Panchagarh, Thakurgaon and Nilphamari. Meteorological and agricultural drought risk maps were prepared by integrating the three classes of drought. It is found that 15.47 % area did not face agricultural drought risk and 16.86 % area did not face meteorological drought risk during the year 2000 to 2008. Approximately 36% and 29 % area faced slight drought in terms of agricultural and meteorological drought risk respectively. Approximately 33% and 33% area faced moderate agricultural and meteorological drought risk respectively. Approximately 16% and 21% area are within severe agricultural and meteorological drought risk. Finally, a resultant risk map was obtained by integrating agriculture and meteorological drought risk maps which indicate the areas facing a combined drought. The combined risk map shows that approximately 17% area has no risk, 23 % area face slight risk, 30 % area face moderate risk and 31 % area face severe to very severe risk. It was evident from the study that central, northern and southwestern districts of N-W region of Bangladesh are more prone to agricultural or meteorological drought. The research results can be used in taking mitigation measures to minimize the loss in agricultural production in drought prone areas. The results also provide information on prevalence, severity level and persistence of drought conditions, which will be helpful to the resource managers in optimally allocating scarce resources.