A Novel Framework for Short-Term Urban Meteorological Simulations

Urban-scale downscaling faces significant challenges due to model biases and errors propagated from initial and boundary conditions provided by reanalysis datasets or global climate models. To address these issues, we introduce a novel global-to-local modeling framework that employs one-way coupling between the Model Prediction Across Scales (MPAS) and Weather Research and Forecasting (WRF) models. This framework is applied to a short-term (10 days) summertime meteorological simulation for the Phoenix Metropolitan Area (PMA).

 

The MPAS-to-WRF approach maintains physics and static field representation consistency across models and reduces spatiotemporal interpolation errors from the driving data. Simulations conducted with the MPAS-to-WRF framework are compared to a traditional approach using the ERA-5 reanalysis product to drive high-resolution (2 km grid spacing) WRF simulations for a 10-day summertime period (June 1-10, 2020). Results demonstrate improved accuracy in near-surface air temperature, moisture, and wind predictions using the MPAS-to-WRF framework, for the first 4-5 days, but a tendency for degradation appears thereafter.

This study introduces a robust method for short-term urban downscaling and highlights the potential for future research to incorporate data assimilation for enhanced long-term simulation accuracy. Future work - ongoing - will incorporate data assimilation methods to improve predictability for monthly to seasonal scale summertime simulations, and potentially, beyond.