Development of a Nonstationary Standardized Precipitation Evapotranspiration Index Using Climate Indices and NDVI as Covariates: A Case Study in the Yellow River Basin, China

With the intensification of climate change and human activities, meteorological and hydrological elements have exhibited nonstationary characteristics. In this changing environment, it is essential to develop a drought index that reflects the nonstationary characteristics of various meteorological and hydrological elements, such as precipitation and evapotranspiration, to assess and monitor drought events more effectively. However, previous studies have predominantly focused on the nonstationary impacts of temporal or climatic indices on meteorological and hydrological elements, neglecting the nonstationary effects resulting from the feedback of vegetation cover changes on these factors. In this context, this study employs the generalized additive models for location, scale, and shape with the climate indices (IOD, NAO, and PDO) and normalized difference vegetation index (NDVI) as covariates to construct a nonstationary standardized precipitation evapotranspiration index (NSPEI). The developed NSPEI is employed to analyze the spatiotemporal characteristics of drought evolution in the Yellow River Basin (YRB) from 1982 to 2022, and the driving factors of NSPEI are quantitatively attributed. The results indicate that (1) NSPEI adapts to the nonstationary impacts induced by the changing environment and more accurately identifies drought characteristics in the YRB than SPEI. (2) There are significant spatiotemporal variations in the drought characteristics of the YRB. The most severe droughts occurred in the upper reaches of the YRB, with the most severe period being from 2000 to 2009. (3) The most significant contributing factor to NSPEI in the YRB was relative humidity, with an average contribution rate of 0.357, followed by NDVI, with an average contribution rate of −0.221. Therefore, considering NSPEI, with climate indices and NDVI as covariates, is more advantageous in revealing the spatiotemporal variations of drought in the YRB, providing valuable information for drought management decisions in a changing environment.