Development of a non-stationary Standardized Precipitation Index and its application to a South Australian climate

In a changing climate, while hydroclimatic variables such as precipitation may show non-stationary behaviour, a traditional Standardized Precipitation Index (SPI) is not capable of accurately predicting extreme meteorological droughts. In this study, we have developed a non-stationary Standardized Precipitation Index (NSPI) within the Generalized Additive Model in Location, Scale and Shape (GAMLSS) modelling framework. This incorporates various climate indices such as SOI, Nino3.4, PDO, SAM and DMI as external covariates to capture the non-stationary and nonlinear characteristics of precipitation and thereby droughts. This idea has been applied at 46 high quality rainfall stations in the state of South Australia. The results indicate that a non-stationary model that considers climate indices can reproduce the rainfall variability better than a stationary model thereby NSPI is better than a traditional stationary SPI (SSPI) at capturing drought characteristics. Bivariate frequency analysis shows that the recurrence interval of drought events exceeding any severity and duration of interest is significantly different for NSPI compared to SSPI. This study demonstrates the need to use a non-stationary drought index in a changing climate to accurately represent the drought characteristics. (C) 2018 Elsevier B.V. All rights reserved.