Publication Date : 2004-12-27
Author : Sheffield, J.Goteti, G.Wen, F.Wood, E. F.
Disaster Management Theme :
Disaster Type : Drought
Document Type : Research Paper
Languange : en
Link : http://onlinelibrary.wiley.com/doi/10.1029/2004JD005182/full
Droughts have severe economic, environmental and social impacts. Timely determination of the current level of drought may aid the decision making process in reducing the impacts from drought. In this study, high-resolution, land surface hydrology simulations using the Variable Infiltration Capacity (VIC) model are used to derive a hydrologically based drought index. Soil moisture data from a retrospective simulation from 1950 to 1999 over the continental United States are used to develop probability distributions of monthly average soil moisture, and the relative position of soil moisture fields within the historic distribution provides a measure of drought in relation to the long-term behavior. The index is able to identify the major drought events during the latter part of the twentieth century and shows good agreement with the time series of U.S. drought from two Palmer Drought Severity Index (PDSI) data sets. On average, 30% of the United States experienced dry conditions (<10% soil moisture quantile) during 1950–1999, peaking at over 70% coverage at the height of the 1950s drought. Many dry events exhibit long-term persistence, especially in the West, which is important in terms of the cumulative impacts. The physical basis of the model allows the index to take into account a number of processes, which contribute to the development of drought, such as snow accumulation and melt that other indices ignore or treat unsatisfactorily. Furthermore, the high spatial and temporal resolution of the simulations ensure that the drought index is able to allow for the effects of short-term changes in meteorology as well as longer-term climate variations, and resolve the high spatial variability in soil moisture and drought occurrence. The potential for implementing the analysis in an operational mode exists by using data from the near real-time simulations within the North American Land Data Assimilation System (NLDAS).