Categorical Drought Monitoring and Prediction in the United States Based on NLDAS-2

Abstract:

Drought is a pervasive natural hazard and is a billion-dollar disaster in the United States, which is comparable to hurricanes and tropical storms with greater economic impacts than extratropical storms, wildfires, blizzards, and ice storms combined. Drought early warning is of critical importance for drought preparedness planning and mitigation efforts to reduce potential impacts of drought, for which drought monitoring and prediction are the essential components. The U.S. Drought Monitor (USDM) has been widely used to track droughts and their impacts. USDM is a composite product that blends quantitative drought indicators and qualitative drought information from multiple sources and classifies drought conditions into different drought categories. Due to the wide application of USDM products, drought monitoring and prediction in the categorical form would be of great importance to aid decision makers to take appropriate measures for drought managements. Based on drought indices from North American Land Data Assimilation System Phase 2 (NLDAS-2), this study proposes a statistical method for the categorical drought monitoring and prediction in the United States. The probabilities of drought conditions falling into different USDM drought categories can be estimated from the proposed method. The method is found to satisfactorily reconstruct historical USDM drought categories and predict future USDM drought categories, and has considerable potential to aid early drought warning in the United States.