AN IMPROVED SNOW HYDROLOGY FOR GCMS .1. SNOW COVER FRACTION, ALBEDO, GRAIN-SIZE, AND AGE

A new, physically-based snow hydrology has been implemented into the N CAR CCM1. The snow albedo is based on snow depth, solar zenith angle, snow cover pollutants, cloudiness, and a new parameter, the snow grain size. Snow grain size in turn depends on temperature and snow age. An improved expression is used for fractional snow cover which relates i t to surface roughness and to snow depth. Each component of the new sn ow hydrology was implemented separately and then combined to make a ne w control run integrated for ten seasonal cycles. With the new snow hy drology, springtime snow melt occurs more rapidly, leading to a more r easonable late spring and summer distribution of snow cover. Little im pact is seen on winter snow cover, since the new hydrology affects sno w melt directly, but snowfall only indirectly, if at all. The influenc e of the variable grain size appears more important when snow packs ar e relatively deep while variable fractional snow cover becomes increas ingly important as the snow pack thins. Variable surface roughness aff ects the snow cover fraction directly, but shows little effect on the seasonal cycle of the snow line. As an applicaion of the new snow hydr ology, we have rerun simulations involving Antarctic and Northern Hemi sphere glaciation; these simulations were previously made with CCM1 an d the old snow hydrology. Relatively little difference is seen for Ant arctica, but a profound difference occurs for the Northern Hemisphere. In particular, ice sheets computed using net snow accumulations from the GCM are more numerous and larger in extent with the new snow hydro logy. The new snow hydrology leads to a better simulation of the seaso nal cycle of snow cover, however, our primary goal in implementing it into the GCM is to improve the predictive capabilities of the model. S ince the snow hydrology is based on fundamental physical processes, an d has well-defined parameters, it should enable model simulations of c limatic change in which we have increased confidence.