Characterization of palmer drought index as a precursor for drought mitigation

Coping with droughts involves two phases. In the first phase drought susceptibility of a region should be assessed for developing proper additional sources of supply which will be exploited during the course of a drought. The second phase focuses on the issuance of drought warnings and exercising mitigation measures during a drought . These kinds of information are extremely valuable to decision making authorities.

In this dissertation three broad schemes i) time series modeling, ii) Markov chain analysis, and iii) dynamical systems approach are put forward for computing the drought parameters necessary for understanding the scope of the drought. These parameters include drought occurrence probabilities, duration of various drought severity classes which describe a region's drought susceptibility, and first times of arrival for non drought classes which signify times of relief for a drought-affected region. These schemes also predict drought based on given current conditions.

In the time series analysis two classes of models; the fixed parameter and the time varying models are formulated. To overcome the bimodal behavior of the Pallner Drought Severity Index (PDSI), primarily due to the backtracking scheme to reset the temporary index values as the PDSI values, the models are fitted to the Z index in addition to the PDSI for the forecasting of the PDSI.