A downscaling framework with WRF-UCM and LES/RANS models for urban microclimate simulation strategy: Validation through both measurement and mechanism model

Abstract

Microscale numerical simulation models are widely applied to explore potential factors and adaptive strategies for localized high temperatures in urban surface or near-surface environments. However, few studies address the limited availability of meteorological input data and the use of multiple meteorological outputs to investigate the mechanisms between factors as a theoretical verification for simulation. This study used the WRF-UCM model outputs in Tianjin, China, as the basic background meteorological field for microclimate simulation and compared the improvement in simulation accuracy of LES-based scheme (PALM-4U) and RANS-based software (ENVI-met) in predicting pedestrian-level air temperature and relative humidity during the downscaling simulation. Subsequently, attribution analysis of land surface temperature imbalance is performed using the two-resistance model (TRM) based on surface and atmospheric simulation outputs which also aids in verifying the applicability of the one-way downscaling simulation framework. It is found that the WRF-UCM-RANS framework exhibits superior overall performance, reducing the error in 2-m height relative humidity by approximately 50 % at the same location compared to mesoscale results. The attribution results indicate that localized high temperature on impervious surfaces within urban neighborhood are primarily driven by surface resistance ($r_s$) during the daytime heating process and ground heat storage ($G$) during nighttime cooling. However, surface resistance ($r_s$) remains the dominant driving factor influencing land surface temperature throughout both daytime and nighttime. The framework reduces the challenge of obtaining initial meteorological data and provides technical support for expanding microclimate research to multi-site simulations and future scenario predictions in complex urban environment.