An Analysis of Non-stationary Drought Conditions in Parana State Based on Climate Change Scenarios.

Climate change affects the hydrological cycle and has a significant influence on water resources, which can lead to environmental and socioeconomic damage caused by droughts. This phenomenon has become more acute in recent years. Among the indices used to assess the extent of droughts, the Reconnaissance Drought Index (RDI) is being adopted throughout the world. The RDI requires a stationarity assumption in its statistical calculations; however, most of the hydro-meteorological time series show non-stationarity. Hence, this paper seeks to assess meteorological droughts by means of RDI by adopting both stationary and non-stationary approaches. Non-stationarity was assessed by applying the Mann–Kendall test for trend detection and the Pettitt test for change point detection. The Thornthwaite method was employed to estimate the potential evapotranspiration required for RDI application. The bias in the future series were corrected employing the Linear Scaling (LS) technique for precipitation series and the Empirical Quantile Mapping (EQM) technique for the temperature series. Historical and future simulated monthly temperature and precipitation data were obtained from 34 meteorological stations located within Parana State, Brazil. The results showed that the maximum drought magnitude, duration, and intensity have increased during the 21^st Century. When compared with historical patterns up to 2100, the results showed that drought magnitude is expected to increase by 105%, while drought duration is expected to increase by 59% in meteorological drought projections. [ABSTRACT FROM AUTHOR]