Evaluation of ICON-LAM and COSMO high-resolution simulations with urban parameterisation for Warsaw, Poland

The rapid development of urban areas and the impacts of global climate change make cities increasingly vulnerable to environmental challenges, including extreme weather and climate events. While operational numerical weather prediction (NWP) models are crucial in supporting emergency management systems, they often lack the necessary detail and do not adequately account for the complex physical interactions between buildings, artificial surfaces, and meteorological processes. Enhancing these models could significantly improve the accuracy of weather predictions in cities, providing better guidance for preparing and responding to weather-related emergencies. This improvement is essential for safeguarding urban populations and infrastructure against the adverse effects of climate change and urbanisation. The ICON-LAM and COSMO models have a bulk urban canopy parameterisation, TERRAURB, which is additionally implemented with spatially variable urban canopy fields in the COSMO model based on the Local Climate Zones (LCZ) approach. By considering the varied characteristics of urban surfaces, such as their materials, structures, and layouts, we can significantly enhance the precision of urban meteorology models. This comprehensive approach allows a deeper understanding of how cities influence local weather patterns and climate conditions, ultimately leading to more accurate predictions and better urban planning strategies. This work presents the results of evaluating NWP hindcasts at hectometric scales for Warsaw agglomeration. For the test simulation, a period of heat wave and strong convection in the city area was selected, covering the end of June and the beginning of July 2022. The National Hydrological and Meteorological Service's measurement and observation data were used. During this period, on 26-28 June, the maximum temperature in urban areas exceeded 30C, and on 30 June and 1 July, the highest monthly temperatures of 34C and 36C, respectively, were observed. On 1 July, the front passed through with heavy precipitation the daily total of 11 mm was recorded, and there was a rapid cooling. The comparison study results show a significant enhancement in the accuracy of surface weather predictions. This is due to implementing urban parametrisation with detailed spatial land use characteristics. The findings provide strong evidence for the effectiveness of this approach in weather forecasting.