Better downscaling results for the right reasons - A process based evaluation of the ICAR model

The evaluation of models in general is a non-trivial task. Even a well established model may yield correct results for the wrong reasons, i.e. by a different chain of processes than found in observations. While guidelines and strategies exist to maximize the chances that results match measurements for the right reasons, these are mostly applicable to full-physics models, such as numerical weather prediction models. The Intermediate Complexity Atmospheric Research (ICAR) model is a comparatively novel atmospheric model employed to downscale atmospheric fields. ICAR uses linear mountain wave theory to represent the wind field and advects atmospheric quantities, such as temperature and moisture in this wind field. Additionally a microphysics scheme is applied to represent the formation of clouds and precipitation.

We conducted an in-depth process-based evaluation of ICAR, employing idealized simulations to increase the understanding of the model and develop recommendations to improve its results. We contrast the ICAR simulations to Weather Research and Forecasting (WRF) model simulations and asses the impact of our recommendations with a case study for the South Island of New Zealand.

Our results suggest two key aspects relevant for ICAR to obtain the correct results for the right reasons. Firstly, the representation of the wind field within the domain improves when the dry and the moist Brunt-Väisälä frequencies are calculated in accordance to linear mountain wave theory from the unperturbed base state rather than from the time-dependent perturbed atmosphere. Secondly, the results show that there is a lowest possible model top elevation that should not be undercut to avoid influences of the model top on cloud and precipitation processes within the domain. We analysed the causes for the differences between the idealized ICAR and WRF simulations and attribute them to the non-linearities in the WRF wind field and additional simplifications in the governing equations of ICAR. With our recommended ICAR setup applied to the real case study we find an upwind spatial shift of the precipitation maximum in comparison to the results obtained with the original ICAR setup. Additionally our results show that when model skill is evaluated from statistical metrics based on comparisons to surface observations only, such analysis may not reflect the skill of the model in capturing atmospheric processes such as gravity waves and cloud formation.

Overall our findings have consequences for the interpretation of past results obtained with ICAR and suggest improvements to ICAR in future studies.