Comparison of 300 m and 100 m uSINGV clear-day simulations for Singapore

Increasing urbanization and its implication for human health, outdoor thermal comfort, air quality or energy consumption necessitate a need for high-resolution urban modelling. In this study, we evaluate uSINGV, a coupled urban-atmosphere research model used by the Singapore Meteorological Service, for four clear-day over Singapore using two different spatial resolutions of 300 and 100 m, respectively. The model is modified to incorporate urban morphology and land use/land cover datasets which are based on European Space Agency climate change initiative data (ESA CCI) at 300 and local datasets at 100 m spatial resolution. The evaluation is carried out for near-surface variables such as temperature, specific humidity, wind speed, and turbulent surface fluxes using Kling-Gupta efficiency (KGE'), root mean square error (RMSE) and mean absolute error (MAE) as model evaluation metrics. Results suggest that temperature and specific humidity are similar for 300 m and 100 m spatial resolution. On the other hand, for 100 m (300 m), the 10 m wind speed has a KGE’ of 0.45 (0.15), RMSE of 0.69 (1.42) m/s, and MAE of 0.55 (1.26) m/s, hence showing improvements from 300 to 100 m spatial resolution. In addition, sensible heat flux for 100 m resolution simulations is closer to observations, while latent heat flux is overestimated. Overall, uSINGV is able to produce reliable simulations at 100 m spatial resolution, thereby showing promise for improved understanding of detailed urban climate processes.