Improving climate model representations of snow hydrology

In this study previous work that led to an improved representation of snow albedo in global climate models is expanded to include regional simulations of snow cover over three distinctly different regions of the United States. We also discuss broadening our work in snow cover parametrization to include representation of the following processes: (i) the 'dirtiness' of snow (due to dust and other particulate loading); (ii) partitioning of energy between melt, evaporation of meltwater, and refreezing within the snow pack; (iii) heat added to a snowpack by rain; and (iv) the vertical temperature profile within a snowpack. In addition, a new parameter, called 'WFLUX', is introduced as a useful glaciological parameter that is based on the net energy available for snowmelt. The snow hydrology formulation of Marshall and Oglesby is implemented into two widely used climate models: a global GCM, the NCAR CCM3/LSM, and a regional climate model, the RegCM2 version of MM4. A suite of new simulations has been made with RegCM2 for three regions of the United States for the 1992-1993 snow season. The regions of interest include the western US/Rocky Mountain region, the region surrounding the Great Lakes and the northeastern United States. Calibration and validation of the new snow hydrology is accomplished by comparison to available ground and satellite based datasets of snowcover, including both mean conditions and year-to-year variability. The goal is not just to improve the simulation of the present-day seasonal cycle (which is reasonably well simulated by many current climate models) but also to improve the predictive capability of models when used to address questions of past climate (especially those involving glaciation) and possible future climatic change. (C) 1999 Elsevier Science Ltd. All rights reserved.