Urban areas are characterized by a clearly different meteorology than rural areas. With the refinement of the resolution of numerical weather prediction models, the consideration of the surface energy balance of urban areas require a specific representation to account for radiative trapping, urban vegetation, anthropogenic heat flux, heat storage in the urban fabric. Many urban canopy models of different complexity have been developed in the last decades. Grimmond et al (2010, 2011) compared a myriad of the models on the performance of the surface energy balance, and identified a ranking in the essentials of processes to be represented and parameter space. In that particular study the models were driven by observations taken above the canopy, which is a state of the art method to evaluate land surface models. However, in the real world, these urban canopy schemes are operating in conjunction with boundary-layer schemes that are responsible for transport of heat, moisture and momentum from the surface through the lower atmosphere, as well as with the free atmosphere due to entrainment. This coupling leads to feedbacks and dependencies on the schemes that have so far not been quantified. Here we propose a modelling experiment in which we further evaluate the modelling infrastructure for the urban boundary layer coupled to the urban land surface.
This modelling exercise specifically aims to:
Evaluate single-column models coupled to the urban surface for the urban environment against field observations at the surface as well and in the PBL.
Identify key strengths and weaknesses in these model approaches.
Identify feedbacks and their strengths between urban canopy schemes and boundary-layer scheme.
Provide a benchmark case study for later use in the community.
In this sense the proposed work build upon earlier experiments in the GABLS (Holtslag et al, 2013) and DICE (Best and Lock, 2016) communities.
Since this will be the first model comparison study for urban areas, we propose to start relatively simple and search for a clear-sky period with relatively low winds (geowind < 5 m/s) for a period of 48-72 h. In general vertical information of the structure of the atmosphere is scarce from observations, though for London (Bohnenstengel et al 2013) a wide suite of observations is at hand. Therefore we propose to select a case study for London.
Since the setup of such a case study is not a trivial thing to do, we propose to perform the intercomparison in two phases. Phase 0 covers the time available at the eWUDAPT workshop in which the participants thoroughly prepare the intercomparison and analyse pro and cons of the setups. In phase 1 the single column model intercomparison is being released to the whole research community, and results will be discussed at later workshops (e.g. during ICUC-2018). Moreover, each phase will cover multiple stages.
Stage 1: In this stage only the urban land surface schemes will be evaluated, analogous to PILPS-urban. Urban morphological parameters will be provided by WG1 in the eWUDAPT workshop and formulated in terms of local climate zones.
Stage 2: In this stage the same urban morphological parameters will be provided as in stage 1, but now to the single column model will be run. In this way one can identify the model behaviour of the land surface scheme in connection to the PBL scheme.
Stage 3: In this stage modellers are asked to apply their default model settings for the urban scheme. This allows for model evaluation against real world observations. For phase 1 identical steps will be undertaken, but for the whole community.