Does the performance of different physical parameterizations in the WRF regional climate model depend on the synoptic weather pattern in the large-scale driving fields?

It has been indicated by multiple studies that the performance of a particular physical parameterization scheme in a numerical model is determined by several factors and displays high spatial and temporal variability. One factor that could potentially lead to distinctive behavior for the different schemes is the large-scale weather situation. Here we aim to answer the following questions: does the performance of the WRF regional climate model with different physical parameterizations depend on the synoptic weather situation? Can simulation results be improved with the utilization of different physics schemes depending on the synoptic weather situation?

To answer these questions, we produce 5-year (2006–2010) regional climate simulations with the WRF model. First, a 50-km simulation is performed using the ERA5 reanalysis as input data. Then, this coarse-resolution run is further downscaled to a set of 10 km simulations with daily re-initialization utilizing different planetary boundary layer (PBL), surface layer, and deep convection schemes. A simulation with the cumulus convection turned off is also included in the analysis.

The weather pattern in the 50-km run for each day is determined based on model-derived relative vorticity at the 500 and 950 hPa isobaric levels (i.e., cyclonic or anticyclonic flow in the upper and lower levels of the atmosphere). Then, we evaluate the performance of the 10-km simulations separately for periods with a specific synoptic pattern prevailing in the driving fields, using the E-OBS dataset as reference. This way, it can be deduced if a physical parameterization is superior to the other for a particular large-scale weather situation. Finally, based on the results of this sensitivity study, we produce further 10-km simulations in which we initialize the 10-km model with different physics schemes each day based on the dominant synoptic pattern in the outer domain.