Evaluation of surface processes in the PALM model system 6.0 for a real urban environment: a case study in Dejvice, Prague

In recent years, the the Large-eddy simulation (LES) model PALM has been rapidly developed its capability to simulate physical processes within urban environments. For example, this includes energy-balance solvers for building and land surfaces, a radiative transfer model to account for multiple reflections and shading, a plant-canopy model to consider the effects of plants on flow (thermo-)dynamics, and a chemistry transport model, as well as nesting capabilities that enable “hot-spot” analysis, to name a few.

This contribution provides an evaluation of modeled meteorological as well as ground and wall-surface quantities against dedicated in-situ measurements taken in an urban environment in Dejvice, Prague. Measurements included monitoring of surface temperature and wall heat fluxes. Simulations were performed for multiple days during several summer and winter episodes, characterized by different atmospheric conditions. To consider time-evolving synoptic conditions, boundary conditions were obtained from mesoscale WRF simulations.

For the simulated episodes, the resulting temperature and wind speed within street canyons show a realistic representation of the observed state, except that the LES did not adequately capture night-time cooling near the surface in some scenarios. At most of the evaluation points, the simulated surface temperature reproduces the observed surface temperature reasonably well, for both, absolute and daily amplitude values. However, especially for the winter episodes and for modern buildings with multi-layer wall structure, the heat transfer through the walls is not well captured in some cases, leading to discrepancies between the modeled and observed wall-surface temperature. Moreover, we also show that the model performance with respect to the observations strongly depends on the accuracy of the input data. To name a few, this includes e.g. the prescribed initial soil moisture, the given leaf-area densities to account for correct shading, or if a facade is insulated or not. Additionally, we will point out current model limitations, particularly implications accompanied by the step-like topography on the Cartesian grid, or wide glass facades that are not fully represented in terms of radiative processes.

With our findings we are able to evaluate the representation of physical processes in PALM, while also pointing out specific shortcomings.