Comparison of Numerical Characteristics in the Regional Model for Prediction across Scales (MPAS) and Weather Research and Forecasting (WRF) Models Using Low-Level Wind Analysis during Heavy Snowfall Episodes over Complex Terrain

Abstract

We present a comprehensive evaluation of the regional Model for Prediction Across Scales (rMPAS) in simulating heavy snowfall events during the International Collaborative Experiment for Pyeongchang Olympics and Paralympics field campaign in the Korean Peninsula. We compared rMPAS simulations with those of the Weather Research and Forecasting (WRF) Model and observational data for three significant snowfall events, emphasizing on their performances in capturing precipitation patterns and synoptic environments. Both the rMPAS and WRF Models effectively replicated the complex meteorological conditions of heavy snowfall events, although they overestimated precipitation. We included sensitivity experiments to compare the dynamical cores in rMPAS and WRF, which use the same physics packages. Kinetic energy spectral analysis showed remarkable similarities between rMPAS and WRF despite differences in their numerical approaches, especially at smaller spatial scales. Analysis of horizontal and vertical wind speeds revealed a consistent overestimation in both models, particularly pronounced at the surface and in the lower troposphere and gradually diminished with altitude. This trend was prominent over complex terrains, where both models simulate stronger vertical velocities compared to those of the observational data. The rMPAS model demonstrated reliable performance in winter weather prediction, comparable to that of the WRF Model, especially in the eastern coastal region of the Korean Peninsula. Further investigation is necessary to address the overestimation of low-level wind speeds for enhancing numerical weather prediction in complex topographical regions.

Publication
Monthly Weather Review, 153(3)
Uju Shin
Uju Shin
Academic research professor
Sang-Hun Park
Sang-Hun Park
Professor