The increasing contribution of potential evapotranspiration to severe droughts in the Yellow River basin

Under global warming, droughts have become more severe and more frequent. In this context, it is widely accepted that both water supply (e.g., precipitation, PRE) and demand (e.g., potential evapotranspiration, PET) should be considered in drought assessments due to increasing evapotranspiration. China’s Yellow River basin is located mainly in arid and semi-arid regions that suffer severe droughts. Analyzing the temporal and spatial patterns of droughts in this area and exploring the relative effects of supply and demand on drought occurrence and variations are thus necessary to adapt to increased droughts. Here, we compared three kinds of drought indices, including the standardized precipitation evapotranspiration index (SPEI), the standardized precipitation index (SPI), and the standardized evapotranspiration deficit Index (SEDI), and used SPEI to identify droughts in the Yellow River basin during the six decades from 1956 to 2016. We identified droughts and described drought dynamics based on drought duration, intensity, frequency, and spatial extent. We then evaluated the relative contributions of PRE and PET to droughts of varying severity and long-term trend, as well as the relationships between ocean signals and droughts. Our results show that most of the basin has undergone progressively drier conditions since the 1950 s. This increased drought occurrence has manifested as longer, more frequent droughts affecting a larger spatial extent. PRE was the dominant driver (64%) of drought occurrence in an area of ∼ 86% of the basin, whereas PET was the main driver (55%) in the remaining 14% of the basin. However, changes in PET contributed more than PRE to drought variation, i.e., long-term SPEI trends, in about ∼ 60% of the area. And effects of PET increased during severer droughts due to the intensive interaction between soil moisture and atmospheric water demand in highly arid conditions. Physically, droughts were closely related to several ocean signals, such as the Multivariate ENSO Index (MEI), the North Atlantic Oscillation (NAO), the Atlantic Multidecadal Oscillation index (AMO), and the Arctic Oscillation index (AO), at different time scales. Understandings of drought dynamics, their physical mechanisms, and the effects of PRE and PET on droughts of varying severity are essential for water resource management, especially in arid regions.