Simulating Moisture Transport Over the Tibetan Plateau in Summer of 2015 Across Scales With a Global Variable‐Resolution Model (MPAS‐A)

Abstract Accurately simulating moisture transport in summer over the TP is uncertain for current numerical models with one important factor being horizontal resolution. In this study, in order to investigate the moisture transport across scales, three experiments are conducted for summer of 2015 using a global variable‐resolution model (MPAS‐A), including one with globally quasi‐uniform resolution of 60 km (U60km) and two with regional refinements over the TP at resolutions of 16 km (V16km) and 4 km (V4km). The wet bias of summer rainfall within the TP increase from U60km to V16km but is significantly improved in V4km. One important source of rainfall bias is the moisture transport across scales. The differences in moisture transport among three simulations are significantly influenced by the changes in wind fields through the Himalayas and eastern TP in two layers, 700–600 and 600–400 hPa, which is largely modulated by their difference in large‐scale circulations particularly monsoon depression. At convection‐parameterized scale (from 60 to 16 km), the scale‐aware Grell‐Freitas convection scheme produces more rainfall and latent heat due to its large sensitivity to the integrating timestep. This sensitivity along with further resolved dynamical processes, collectively strengthen the monsoon depression to the south of TP and make it shift northward in conjunction with the mid‐latitude westerlies. With resolution increasing to convection‐permitting scale (from 16 to 4 km), the resolved moist convection releases significantly less latent heat and then reproduces a weaker monsoon depression. This causes a discrepancy that exceeds the resolution‐related difference at convection‐parameterized scale.