Online and Offline Evaluation of the Single-Layer Urban Canopy Model in Summertime Conditions over London

At present the majority of the world’s population is living in urban areas. This creates a demand for accurate forecasts on human thermal comfort, air quality, energy demand and weather in the urban environment. To meet this demand, mesoscale numerical weather prediction (NWP) models need to forecast the urban microclimate with a reasonable accuracy. However, the complexity and heterogeneity of urban areas cannot be explicitly resolved at resolutions still common to NWP models. Thus, NWP models utilize so called urban parameterisation scheme, which represents processes (radiation trapping, anthropogenic forcing, etc.) occurring in the urban environment. These schemes are essential for weather forecasting in urban areas. However, they often require a large number of prescribed parameters, ranging from urban fraction and building height to thermal conductivities and thickness of walls, roofs etc. Some of these parameters are highly uncertain and they can reduce model performance if not prescribed accurately. Recent studies have investigated the sensitivities of urban canopy model’s performance to these parameters, but mainly in an offline mode (no coupling with the overlying urban boundary layer). However, such an approach limits the understanding of how urban canopy models interact with the overlying atmosphere and fails to identify feedback mechanisms between the land surface and the atmosphere. In this study, we investigate the model sensitivity of the single-layer urban canopy model (SLUCM) (Kusaka, 2004) included in the Weather and Forecasting model (WRF) to surface parameters in two different set-ups. For the evaluation and comparison of both set-ups we use a 2-day period (54 hours) during summer time in London, with clear skies and low wind speeds. In the first set-up, we test the single-layer urban canopy model in an offline configuration to understand how the model reacts to fixed atmospheric forcing. Then, we evaluate the urban canopy scheme coupled to the overlying atmosphere in the 1D configuration of WRF, with prescribed large scale forcing. An analysis of model response to changes in surface parameters reveals that the urban canopy model reacts differently, when coupled to the atmosphere compared to the offline analysis. For instance, the change in daily net radiation, due to changes in the roof albedo, is up to 11% of the mean value for daily net radiation in the coupled experiment, while in the offline one the change is up to 13%. This indicates that the coupling with the overlying atmosphere should be considered, when investigating the performance of urban canopy models, due to the presence of feedback mechanisms between the land surface and the atmosphere.