Narasimhan, Balaji (2005-11). Development of indices for agricultural drought monitoring using a spatially distributed hydrologic model. Doctoral Dissertation.
Thesis
Farming communities in the United States and around the world lose billions of dollars every year due to drought. Drought Indices such as the Palmer Drought Severity Index (PDSI) and Standardized Precipitation Index (SPI) are widely used by the government agencies to assess and respond to drought. These drought indices are currently monitored at a large spatial resolution (several thousand km2). Further, these drought indices are primarily based on precipitation deficits and are thus good indicators for monitoring large scale meteorological drought. However, agricultural drought depends on soil moisture and evapotranspiration deficits. Hence, two drought indices, the Evapotranspiration Deficit Index (ETDI) and Soil Moisture Deficit Index (SMDI), were developed in this study based on evapotranspiration and soil moisture deficits, respectively. A Geographical Information System (GIS) based approach was used to simulate the hydrology using soil and land use properties at a much finer spatial resolution (16km2) than the existing drought indices. The Soil and Water Assessment Tool (SWAT) was used to simulate the long-term hydrology of six watersheds located in various climatic zones of Texas. The simulated soil water was well-correlated with the Normalized Difference Vegetation Index NDVI (r ~ 0.6) for agriculture and pasture land use types, indicating that the model performed well in simulating the soil water. Using historical weather data from 1901-2002, long-term weekly normal soil moisture and evapotranspiration were estimated. This long-term weekly normal soil moisture and evapotranspiration data was used to calculate ETDI and SMDI at a spatial resolution of 4km ?? 4km. Analysis of the data showed that ETDI and SMDI compared well with wheat and sorghum yields (r > 0.75) suggesting that they are good indicators of agricultural drought. Rainfall is a highly variable input both spatially and temporally. Hence, the use of NEXRAD rainfall data was studied for simulating soil moisture and drought. Analysis of the data showed that raingages often miss small rainfall events that introduce considerable spatial variability among soil moisture simulated using raingage and NEXRAD rainfall data, especially during drought conditions. The study showed that the use of NEXRAD data could improve drought monitoring at a much better spatial resolution.
