The Role of Mean State Bias in a Climate Model on Atmospheric Blocking Frequency

Atmospheric blocking often causes extreme weather events in the mid-latitudes, yet climate models cannot faithfully reproduce the observed climatology of blocking. The cause of this issue remains unclear – two possibilities are the incorrect model physics and/or the mean state bias in climate models. When assessing climate models, the two factors are interconnected. To better determine the root cause, we vary aspects of the model physics while maintaining a consistent model mean state. One particular focus is the moistphysics. We use two models: one a standard CAM with conventional moistphysics parameterization schemes and the other a similar CAM with embedded cloud-resolving models instead of parameterization (SPCAM). We do not find any noticeable difference in the blocking frequencies between the two, suggesting that any issues that do exist with the model moist physics have only a weak direct influence on blocking frequencies. We also find that both CAM and SPCAM substantially underestimate the frequency of blocking events when compared with the reanalysis products. This motivates us to test the hypothesis that the bias in the model mean state is responsible for most of the bias in the blocking frequency. We reduce the mean state bias in the CAM by correcting the mean state toward a reanalysis product (MERRA2), and then quantify the change in the blocking frequency. Our results will shed light on the important role of the basic climate state in determining blocking frequencies, and will help improve blocking climatology in climate models.