| First, I reexamine the basis for a widely cited critique that climate models overstate the strength of the water vapor feedback; a key amplification of the projected warming due to increasing concentrations of carbon dioxide. This critique states that the correlation between interannual variations of tropical average temperature and humidity is stronger and less height dependent than observations suggest. This work raises several caveats to this finding. First, if this model-data discrepancy exists it is independent of moist convective parameterization. Second, the discrepancy is greatly reduced by sampling the model in a manner more consistent with the observations. In the end, it is suggested that the original critique is inconclusive at best and cannot therefore be interpreted to mean that climate models greatly overestimate the water vapor feedback.; Next, I examine the climate model for the narrow filamentary plumes of intense water vapor transport sometimes called tropospheric rivers. Such plumes have never been documented in a climate model before despite clear signs that they are key agents of water vapor transport. Indeed, similar to observations these plumes carry some 80% of all upward and poleward water vapor transport in the model. This strongly suggests that small scale plumes conduct much of the eddy moistening behind the strong positive water vapor feedback in this climate model.; Finally, I address the humidity variations associated with baroclinic cyclones. This is done with a combination of cyclone tracking and composite analysis. While this model makes generally realistic cyclones it also lacks enough cyclones and especially enough strong ones. Moreover, model cyclones need to move faster, extend further and more connectively into the upper troposphere, and sport more vigorous circulations. Indeed, specific differences exist in how this climate model links cyclone propagation to the three-dimensional wind field, and thus to the movement of water vapor, when compared to observations and weather models. Such differences account for much of this model's extra-tropical dryness. The role of model resolution and formulation is addressed as is the issue of how the models cyclone deficiencies might affect the water vapor feedback of the model. |
