Sensitivity of Fine-Resolution Urban Heat Island Simulations to Soil Moisture Parameterization

Urban areas experience the impact of natural disasters, such as heatwaves and flash floods, disparately in different neighbourhoods across a city. The demand for precise urban hydrometeorological and hydroclimatological modelling to examine this disparity, and the interacting challenges posed by climate change and urbanisation, has thus surged. The Weather Research and Forecasting (WRF) model has served such operational and research purposes for decades. Recent advancements in WRF, including enhanced numerical schemes and sophisticated urban atmospheric-hydrological parameterizations, have empowered the simulation of urban geophysical processes at high resolution (~1 km), but even this resolution misses significant urban microclimate variability. This study applies the large-eddy simulations (LES) mode within WRF, coupled with single-layer urban canopy models (SLUCM), to enable even finer-scale modelling (150 m) of the Urban Heat Island (UHI) effect in the Baltimore metropolitan area. We run nine scenarios to evaluate various methods of initializing soil moisture and various spinup lead times, and to assess the impact of WRF's Mosaic approach in depicting subgrid-scale processes. We evaluate the scenarios by comparing the WRF simulated land surface temperature (LST) against Landsat LST and the WRF simulated hourly 2-m air temperatures (AT) with observations from eight weather stations across the domain. Results underscore the paramount influence of the lead spinup time on the spatiotemporal distribution of simulated soil moisture, consequently shaping WRF's efficacy in predicting the UHI. Furthermore, interpolating soil moisture-related parameters from the parent for child domain initialization yields a notable reduction in mean and root-mean-squared errors. This improvement was particularly evident in simulations with the longest spinup time, affirming the importance of carefully designing the initialization of soil moisture for improved urban temperature predictions.