Metrics to identify meaningful downscaling skill in WRF simulations of intense rainfall events

Dynamical downscaling attempts to provide regional detail to climate change projections that subsequently can be used as input to climate change impact models. However, unlike forecasts by numerical weather prediction models, downscaled projections cannot be tested for skill because the future of interest is decades away. Nevertheless, models can be tested in terms of how well they simulate current weather or climate, thus giving an indication of skill in representing the process of interest. Here, six configurations using different combinations of three microphysics and two planetary boundary layer schemes are assessed on their skill to simulate desired characteristics in daily rainfall fields from three two week simulations in southeast Australia; 'desired' meaning desirable in relation to the intended application. Of different metrics and analysis assessed, a metric based on variography analysis, summarising characteristics about spatial variability and dissimilarity, is shown to provide the most informative guidance relative to the desirable characteristics.