Impact of New Regional Land Cover Data on WRF Simulations of Heat Waves in Shanghai

Accurate land cover data are essential for improving simulations of near-surface meteorological conditions in numerical weather prediction models. In this study, a newly developed 30-m resolution regional land cover dataset (ModelLand30) is applied to the Weather Research and Forecasting (WRF) model. Compared with the default land cover dataset (MODIS) in the WRF model, the ModelLand30 dataset provides a more nuanced depiction of land cover characteristics, particularly in urban areas. To evaluate the impact of ModelLand30 on the simulation of near-surface meteorological variables, simulations are conducted for high-impact heat waves in Shanghai using the MODIS dataset and the ModelLand30 dataset. Based on the ModelLand30 dataset, an additional simulation using the mosaic approach is conducted to further examine the influence of sub-grid surface heterogeneity. The results show that compared with the MODIS dataset, the ModelLand30 dataset reduces the overestimation of the surface sensible heat flux and successfully reproduces the diurnal cycle of the surface latent heat flux at the Fengxian station, because it provides more accurate description on land cover types. The ModelLand30 dataset also reduces the underestimation of the 2-m temperature during nighttime but overestimates the 2-m temperature during daytime in Shanghai urban areas. Due to consideration of the effect of non-urban land types within sub-grid cells, the mosaic approach further improves the simulation performance of the surface latent heat flux. It also lessens the daytime temperature overestimation in the Shanghai urban area. This study highlights the application of the ModelLand30 dataset in the WRF model and the importance of the representation of sub-grid surface heterogeneity for heatwave prediction.