A new method of reducing dynamic downscaling precipitation bias in drylands: Wet-day Quantile Delta Mapping (WQDM)

Quantile Delta Mapping (QDM), as the most popular bias correction method after dynamical downscaling, has been widely applied across various climate regions. However, it is characterized by a notable shortcoming on the generation of an excessive number of random numbers in drylands compared to other regions. To address this issue, we developed an innovative Wet-day Quantile Delta Mapping (WQDM) method. This approach overcomes the limitations of QDM by mapping the proportion of wet-day precipitation, thereby achieving a more effective reduction of precipitation bias from dynamical downscaling. In terms of wet-day frequency bias, WQDM achieves a maximum reduction of over 50% compared to QDM. For extreme precipitation, the bias is reduced by 52.22% with QDM and further lowered to 61.58% with WQDM. These results demonstrate the superior bias correction performance of WQDM, particularly in low-frequency precipitation regions. This study provides new methodological support for processing dryland precipitation data and lays a robust foundation for a systematic understanding of its characteristics, evolution patterns, and influencing mechanisms.