Diagnosis and monitoring of drought using regional climate model over Bangladesh

Abstract

Drought is a major natural hazard that causes the earth to parch and a considerable hydrologic or water imbalance resulting water shortages, wells to dry, depletion of ground water and soil moisture, stream flow reduction, crops to wither leading to crop failure and scarcity in fodder for livestock. This study analyzed drought diagnosis and monitoring over Bangladesh. Emphasis is given to analyze meteorological drought through calculation of Standardized Precipitation Index (SPI). Other popular drought indices are not utilized due to the lack of observed data. Besides observed dataset, Regional Climate Model output is utilized in calculation of SPI at different station locations over Bangladesh. The historical records of drought event extracted from Bangladesh Statistical Bureau (BBS) and International Disaster Database EM-DAT are used to confirm the drought which detected by the model data SPI calculation for particular drought month/year. The main goal of this research is to identify the utility of climate model outputs in projecting drought phenomenon in this region. The SPI is calculated for month 1, month 2... month 6, month 9, month 12, month l5, month 18, month 2l and month 24 using observed and model datasets. In calculation of SPI, 27 station locations data over Bangladesh during 1961 to 1990 is utilized. It is found that point information at a particular station location is not enough to detect drought condition using model data with reference to the observed values. Sometimes observed data also show drought in one location but opposite situation (surplus of rainfall) in other nearby location due to the huge variation of rainfall in place-to-place. Regional average information provides better understanding in detection of drought compared to point information because drought is a regional phenomena. It is not possible to detect same month/year as a drought situation by both datasets used in this study suggests further study with more datasets and new technique. Overall, model outputs employed SPI during 196l-1990 can detect moderate and severe drought compared to the extreme cases over Bangladesh. Historical record of drought confirms drought detection by the model data. This infers that statistics of drought may be helpful in preparedness from impacts of drought in agricultural and water resources sectors even though particular drought month/event may not be projected very accurately.