Large-scale regional model biases in the extratropical North Atlantic storm track and impacts on downstream precipitation

Global climate models have circulation biases that the community aims to reduce, for instance through high-resolution dynamical downscaling. We used the Weather Research and Forecasting model (WRF) to downscale both ERA-Interim and a bias-corrected version of the Norwegian climate model NorESM1-M on a high-resolution grid. By varying the domain size, we investigated the influence of the driving data and highly resolved topography on the North Atlantic storm track and the precipitation in its exit region. In our largest domains, we found large-scale circulation and storm track biases similar to those seen in global models and with spatial patterns independent of the driving data. The biases in the smaller domains were more dependent on the quality of the driving data. Nevertheless, the biases had little effect on the simulated precipitation in Norway. Although the added value of downscaling was clear with respect to the global climate models, all the downscaled simulations showed similar precipitation frequencies and intensities. We posit that, because the precipitation is so strongly governed by the local topographic forcing, a correct storm track is less critical for the precipitation distribution.