A new multivariate composite drought index considering the lag time and the cumulative effects of drought

Frequent and intense droughts pose significant threats to ecosystem health and human society under global change, making timely and rapid detection of such events crucial. Drought index is an essential tool for drought monitoring and risk assessment. Univariate drought indices cannot effectively characterize comprehensive drought characteristics and rarely account for the time lag between different types of droughts as well as their cumulative effects. Considering this, we developed a multivariate composite drought index (MCDI) using the Gringorten empirical formula based on four drought indices representing meteorological drought (Standardized Precipitation Actual Evapotranspiration Index, SPAEI), agricultural drought (Standardized Soil Moisture Index SSI), and hydrological drought (Standardized Runoff Index (SRI), and Water Storage Deficit Index (WSDI)). To verify the effectiveness of MCDI, we first calculated the Pearson correlation coefficients (PCC) between scPDSI (Self-calibrating Palmer Drought Severity Index) and MCDI in China. The results showed that the percentage of pixels with PCC greater than 0.5 was 70.02 % (p < 0.05). Then we analyzed the spatial/temporal drought trends in China and different hydroclimatic zones. Drought indices performed differently, and MCDI generally exhibited a "dry areas become wetter, wet areas become drier" pattern in China based on the trends shown by regional means. However, in terms of spatial distribution, the wet and dry trends in China are highly spatially heterogeneous. In addition, we selected two typical drought events that occurred in Arid/Semi-Arid (Inner Mongolia) and Humid/Semi-Humid (Yunnan Province) zones, respectively, to assess the ability of the MCDI to characterize drought. Compared with other drought indices and drought indicators (Soil Moisture and Solar-induced Chlorophyll Fluorescence), MCDI characterized the drought event most consistently with official records and responded faster to drought. Overall, the MCDI has good potential for drought monitoring and assessment.