Extratropical cyclones derive much of their energy from the release of latent heat. Within cyclones, precipitation occurs largely on the meteorological mesoscale. This suggests that cyclone development cannot be fully simulated at climate-model grid scales of 100 km or more. At the same time, projections of global warming indicate that specific humidity will increase with global warming at a pace close to that dictated by the Clausius-Clapeyron relation. Taken together, these facts facts lead to two hypotheses:

1) Simulated cyclone development is sensitive to model grid scale, and enhanced development at improved resolution can be attributed to improved simulation of diabatic processes.

2) The sensitivity of simulated storm tracks to global warming will increase with model resolution (and the sensitivity of the storm tracks to resolution will increase with warming).

These hypotheses have been examined in regional simulations of the North American/North Atlantic storm track, varying resolution and carried out under current and future climatic conditions. Results are consistent with expectations (Wiillison et al., 2013 & 2015). Complexity emerges, however, when simulations are carried out in a global domain, thus relaxing constraints imposed by lateral boundaries. In a global model, it is found that the sensitivity of the storm track to global warming does increase with enhanced resolution in the North Pacific, but not in the North Atlantic. in the latter region, the loss of energy from synoptic eddies to variability in the large-scale flow increases in future simulations.

These result imply that current climate model resolutions are inadequate for projecting future changes extratropical cyclones and the storm tracks and that results from regional modeling (a.k.a. dynamical downscaling) must be treated with extreme caution wherever upscale transfers of energy are important.