Nonstationary Modeling of Meteorological Droughts: Application to a Region in India

Abstract

In the continually climate change scenario, it is of great concern to revisit, rethink, and improve the existing computational aspects of drought indexes. In general, the commonly used drought indexes do not encompass environmental changes. Hence, the primary objective of this study was to incorporate large-scale climatic oscillations in developing the nonstationary Standardized Precipitation Index (SPIN) and Reconnaissance Drought Index (RDIN). The nonstationary modeling of the drought index is performed using the generalized additive model for location, scale and shape (GAMLSS) over 103 grid points covering Maharashtra state in India. The computed nonstationary drought indexes are then used to examine the spatiotemporal variability of drought over the province. The outcomes from the analysis indicate that nonstationary modeling outperforms a stationary approach over the study area for different drought scales. In addition, notable differences are observed while comparing the different drought properties using stationary and nonstationary drought indexes. A comparative analysis of the historical drought that occurred in 2013 reveals that nonstationarity in the meteorological data sets cannot be ignored for developing a sustainable mitigation and adaptation strategy against drought hazard. The proposed nonstationarity-based drought indexes for Maharashtra can be a feasible alternative of drought assessment in a changing climate scenario.