Development and Applications of a Regional Climate Model Based on MM5 (Invited Presentation)

In 1985 I became aware of the Penn State-NCAR mesoscale model (MM4 and 5) developed by Rick Anthes and his students and colleagues. MM5 was made available freely to the international community and was well known through its documentation and publications. Beginning with MM4, I and my students at Seoul National University developed several versions suitable for studying the weather and climate of eastern Asia. We have used these versions to study such important phenomena as heavy rainfall events and tropical cyclones that affect the Korean peninsula. MM5 and WRF have served as an operational model for Korea Meteorological Administration (KMA) in regional NWP through research projects between KMA and NCAR. Meanwhile, Rick Anthes and his student Bill Kuo have made great contributions to the SNU NWP research lab until my retirement.

My presentation describes the regional climate model (SNURCM) based on MM5 and a recent application of tropical cyclone simulations within the Coordinated Regional Climate Downscaling Experiment (CORDEX)-East Asia, WCRP/WMO. SNURCM and four other regional climate models with horizontal resolution of 50 km were driven by the ERA-Interim reanalysis data over the western North Pacific for 20 years (1989–2008). Individual models showed significant biases of simulated tropical cyclone with intensity underestimated due to relatively low horizontal resolutions (Figure 1). Nonetheless, they reasonably captured observed climatological spatial distribution and interannual variability of tropical cyclone activity. For simulated interannual variability, most models showed approximately one half of observed mean accumulated cyclone energy and high correlation coefficients above 0.6. In general, Multi-RCM ensemble mean based on model performance outperformed the individual models with smaller biases and higher correlations on the spatial and temporal variation of tropical cyclone activity. Multi-RCM ensembles driven by multi-GCM projections can provide useful regional information on climate change projections of tropical cyclone activity.