CR 74:263-281 (2018)  -  DOI: https://doi.org/10.3354/cr01500

ABSTRACT: Changes in regional aridity have complicated spatio-temporal features during the past several decades and have important implications for hydrologic cycles under global climate change. However, nonlinear changes in aridity and its spatial variability in China remain poorly understood. Here, we analyzed the spatial-temporal changes in hydroclimatic variables, including reference evapotranspiration (ET_o), precipitation (P), aridity index (AI) and their relationships, mainly based on the ensemble empirical mode decomposition method. A climate dataset from 573 meteorological stations in China was used for the period 1961-2015. The results show that the nonlinearity of changes in aridity was complex, with distinct spatial differences. Approximately 76.09% of stations showed reversed trends in AI during 1961-2015, among which AI initially decreased and then increased, mainly in Inner Mongolia and the areas south of the middle and lower reaches of the Yangtze River. However, their reversal times were not synchronous. The drying to wetting transition mainly occurred in northern areas, including the southern part of Northeast China, the eastern Loess Plateau and the southern North China Plain. Although the majority of stations experienced reversals from downward to upward trends of ET_o (53.92%) mainly around the 1990s, increasing precipitation exerted a larger effect on AI. For most eco-geographical regions in China, the nonlinear variations in ET_o and P jointly led to the decline in AI over the past 55 yr, and the P trend contributed more than the ET_o trend to the change in AI, except in subtropical regions, where ET_o variations played a key role.

KEY WORDS:  Aridity · Precipitation · Reference evapotranspiration · Nonlinear · Climate change · Ensemble empirical mode decomposition · EEMD