Simulating Mesoscale Convective Systems Downstream of the Tibetan Plateau at Kilometer-Scale: Insights from the First Ensemble Run of ‘WY2020’ in CORDEX-FPS-CPTP

Mesoscale convective systems (MCSs), as major contributors to extreme precipitation events, have garnered significant attention in the context of global warming. In 2020, vast areas of East Asia—downstream of the Tibetan Plateau (TP) and home to over 30% of the global population—experienced an anomalously wet rainy season, marked by intense MCS-induced precipitation that resulted in severe socio-economic impacts and extensive losses.

This study leverages the first ensemble run of kilometer-scale (~4 km) WRF simulations for the Water Year 2020 (WY2020) under the CORDEX-FPS-CPTP framework. The performance of these simulations in representing MCS precipitation characteristics has been assessed using the GPM-IMERG precipitation product and the CMA Multi-Source Merged Precipitation Analysis (CMPA).

The results demonstrate that while all ensemble members can generally capture the spatial distribution of MCS precipitation downstream of the TP, notable differences arise among the simulations with single-physics perturbations. Specifically, simulations using the Morrison and WSM5 microphysics schemes exhibit strong agreement with observations. In contrast, simulations employing the SBU_YLin or WDM6 microphysics schemes significantly underestimate MCS precipitation in the region. Regarding planetary boundary layer (PBL) scheme sensitivity, simulations utilizing the YSU and Shin-Hong schemes outperform those employing the MYNN3 scheme.

Despite these variations, a common bias emerges across all seven kilometer-scale WRF simulations: they collectively underestimate the rainfall area of MCSs by 30.9% to 43.0%, while simultaneously overestimating precipitation intensity of MCSs by 59.4% to 64.1%. These results suggest a consistent tendency for K-scale WRF simulated MCS precipitation to exhibit smaller spatial extents yet greater magnitudes compared to observations.

To explore potential improvements, we expanded the model domain from [15.0°N–50.0°N; 65.0°E–125.0°E] to [5.0°N–55.0°N; 45.0°E–160.0°E; almost doubled] and conducted additional WY2020 simulations. Preliminary results indicate that the expanded domain not only enhances the model’s ability to capture heavy MCS rainfall centers during Mei-yu season, particularly over the Western North Pacific Ocean (south to Japan), but also better reproduces MCS precipitation features than other K-scale WRF runs. Remarkably, the expanded domain simulation even outperforms the GPM data in representing MCS precipitation over the middle and lower reaches of the Yangtze River basin in eastern China, compared to CMPA.

The K-scale WY2020 ensemble run represents a valuable resource for advancing our understanding of the K-scale model uncertainties on MCS and the hydrological cycle over the TP and its downstream regions.