Can localized vegetation parameters improve microclimate simulation accuracy? A case study in the south subtropical hot-humid area

Abstract

Vegetation parameters significantly impact the accuracy of microclimate simulations, yet their values vary across different climate zones. This study investigates whether using localized vegetation parameters can enhance simulation accuracy. Four scenarios were established with varying vegetation parameters and meteorological boundary conditions. Localized vegetation parameters were obtained from tree measurements in Guangzhou, a representative city in the southern subtropical hot-humid area. On-site microclimate data were used to validate the simulations, with evaluation metrics including R², d, RMSE, and MAE. The results show that (1) Simulations using localized vegetation parameters under full forcing meteorological boundary condition closely matched observed values (R² = 0.89, d = 0.98, NRMSE = 0.17, and NMAE = 0.40); (2) ENVI-met can simulates microclimatic effects for different vegetation types, but air temperature simulations on the lawn being more accurate than under tree shade；(3) 63 % of frequently used trees in Guangzhou exhibit LAD values between 0.3 and 0.6 m²/m³. Overall, the microclimate simulations are more accurate when the localized vegetation parameters are used.