Evaluation of urban surface parameterizations in WRF model using energy fluxes measurements in Portugal

Abstract

The performance of WRF model was investigated for the simulation of urban microclimate, with particular focus on energy fluxes, using different urban surface parameterizations. The model performance was evaluated using measurements carried out between August and December 2014 in Portugal. Several simulations were performed over two different areas, Porto urban area and Aveiro suburban area, for the entire measurement period. Distinct simulations were performed using different urban parametrizations: (i) the Noah Land Surface Model (LSM); (ii) a single-layer urban canopy model (UCM); and (iii) a modelling system composed by WRF and SUEWS models (WRF-SUEWS).The results showed that both UCM and SUEWS are able to simulate the energy partitioning over the low and high intensity residential. At the low intensity residential area, the majority of energy is partitioned to latent heat flux, accounting on average for 47% and 49% of the daytime available energy, for UCM and SUEWS, respectively. At the high intensity residential area, the greatest share of energy goes to sensible heat flux (42% [UCM] and 50% [SUEWS]), followed by the storage heat flux (33% [UCM] and 43% [SUEWS]). For both areas, a completely different energy partitioning was obtained when the LSM was used. The analysis performed showed that the UCM are able to provide a more accurate turbulent energy partitioning (sensible and latent heat), which contribute to enhance the urban microclimate simulation results; the systematic model biases in the LSM simulation was reduced by 1–2 °C in air temperature and by 0.5–1 m s−1 in wind speeds at near surface layer, on average, depending on the urban density. The overall results suggest that an appropriate representation of urban physical processes are crucial to improve numerical tools suited for the modelling of the urban atmospheric boundary layer.