Objectivization of an expert assessment framework for drought monitoring

Global warming leads to an increased occurrence and severity of regional drought events. However, recognizing drought risks across scales in a timely and accurate manner is still challenging due to the complexity and multifaceted nature of drought evolution. Here, we present a newly developed Comprehensive Drought Monitoring Model (CDMM) that objectivizes an expert assessment framework for drought monitoring using the Random Forest algorithm. The CDMM integrates multiple drought monitoring indices and the expert subjective experience derived from the U.S. Drought Monitor (USDM). This allows us to effectively capture different drought categories across regions. Our model validation analysis shows that the CDMM accurately reproduces the USDM drought categories and their spatial distributions. By applying CDMM to China, it is found that CDMM effectively captures drought dynamics across different climatic regions and timescales. Results show that CDMM identifies a higher frequency of short-duration drought events in the eastern monsoon region of China, while the northwest and southwest regions experience prolonged drought events. These findings shed light on the potential of CDMM for drought monitoring at long-term reginal scales and demonstrate its transferability for tracking drought evolution across scales.