MPhil Thesis Presentation - Developing and Evaluating the New MPAS-Urban Modeling System

Global warming is increasing the exposure of densely populated cities to heatwaves, intense rainfall, and other hazardous weather, yet these risks remain difficult to represent in current global climate models (GCMs). Because most GCMs operate at coarse resolutions and rely on simplified urban representations, they often fail to capture the heterogeneous surface fluxes, urban morphology, and fine-scale land--atmosphere interactions that strongly influence urban climate extremes. To address this gap, this study develops an MPAS--Urban modeling system by integrating a single-layer urban canopy model (SLUCM) and a bulk urban parameterization into the Noah-MP land surface model within the Model for Prediction Across Scales (MPAS) framework. Using a variable-resolution mesh from 30~km to 500~m, the system enables seamless refinement over urban regions without conventional nested downscaling. High-resolution Local Climate Zone (LCZ) data are incorporated specifically for MPAS and applied consistently in both the bulk and SLUCM configurations to represent urban morphology and land-use heterogeneity.

The system is evaluated through a retrospective simulation of a 10-day heatwave over Hong Kong, a dense subtropical coastal city. MPAS-Urban successfully reproduces the onset and peak of the extreme event and captures the diurnal cycles of near-surface air temperature, humidity, and wind speed with generally small biases. SLUCM performs better for near-surface temperature and wind speed, whereas the bulk scheme shows slightly better skill for humidity. These results demonstrate that MPAS-Urban provides a flexible multiscale framework for simulating urban climate extremes and for improving urban climate assessment in rapidly urbanizing regions.