Seasonal Drought Prediction in India

Abstract:

Drought is among the most costly natural disasters in India. Seasonal prediction of drought can assist planners to manage agriculture and water resources. Such information can be valuable for a country like India where 60% of agriculture is rain-fed. Here we evaluate precipitation and temperature forecast from the NCEP’s CFSV2 for seasonal drought prediction in India. We demonstrate the utility of the seasonal prediction of precipitation and temperature for drought forecast at 1-2 months lead time at a high spatial resolution. Precipitation from CFSv2 showed moderate correlations with observed up to two months lead. For one month lead, we found a significant correlation between CFSv2 and observed precipitation during winter season. Air temperature from the CFSv2 showed a good correlation with observed temperature during the winter. We forced the Variable Infiltration Capacity (VIC) model with the CFSv2 forecast of precipitation and air temperature to generate forecast of hydrologic variables such as soil moisture and total runoff. We find that errors of the prediction reduce for the two month lead time in the majority of the study domain except the northern India. Skills of Initial Hydrologic Conditions combined with moderate skills of forcings based on the CFSv2 showed ability of drought prediction in India. The developed system was able to successfully predict observed top layer soil moisture and observed drought based on satellite remote sensing in India.