Evaluation of city-scale PALM model simulations and intra-urban thermal variability in Vienna, Austria using operational and crowdsourced data

Abstract

The PALM model system for urban applications is used to simulate the spatio-temporal thermal variability under heat wave conditions in Vienna, Austria. Model simulations covering the city area and its surroundings with a spatial resolution of 20 m are performed for four consecutive clear-sky hot days in August 2022. The results for hourly air temperature are evaluated with data from conventional weather stations of professional monitoring networks and quality-controlled data from private weather stations of the company NETATMO. The observations show high intra-urban variability during daytime and distinct spatial patterns at night with higher air temperatures in densely built city centre and lower temperatures in surrounding low-density urban region with prevailing green areas. The model shows lower variability than the observations, but similar large-scale spatial patterns. Direct comparison with observational data indicates a good model performance with a high coefficient of determination (R2 > 0.85), low bias (0.2 °C – 0.4 °C) and a root mean squared error (RMSE) of 1.8 °C. Evaluation of intra-urban thermal variability based on classification per land use and land cover type shows a statistically significant air temperature difference between the built-up areas and high vegetation surfaces during the night of about 1 °C in the observation and about 2 °C simulated by the model.