Under the influence of climate change and high intensity human activities

it is difficult to accurately describe the evolution of non-stationary drought series under the traditional stationary hypothesis. This study uses the complete ensemble empirical mode decomposition with adaptive noise and the run theory to extract the reconstructed drought characteristics of the standardized precipitation evapotranspiration index (SPEI)

based on monthly meteorological data from the Yellow River basin of 1960-2020. And the Pettitt method and rotating empirical orthogonal functions are combined to examine its non-stationary meteorological drought characteristics. Results indicate that its meteorological drought manifested a significant increasing trend on multiple time scales

and significant changes in SPEI occurred around 1996 on all time scales. More pronounced seasonal variations in time and more realistic distributions in space are found in the reconstructed drought characteristics when taking into account the non-stationarity. This study is of great significance in clarifying the spatiotemporal evolution patterns of meteorological drought under non-stationary conditions

and helps drought prevention and mitigation for the Yellow River basin.