Five years of meteorological and hydrological data from a typical New England watershed where winter snow cover is significant were used to drive and validate two off-line land surface schemes suitable for use in the Goddard Institute for Space Studies GCM: a baseline scheme that does not model the physics of a snowpack and therefore neglects the insulating properties of snow cover; and a modified scheme in which a three-layer snowpack is modeled. Comparing baseline model results with validation data reveals several model deficiencies. Surface radiation temperatures could not adequately be modeled and the ground froze to unreasonable depths. Furthermore, because of ground cooling resulting from large surface heat fluxes to the atmosphere from the uninsulated surface, deeper model layers did not unfreeze until midsummer. As such, the normal hydrologic processes of runoff, ground water infiltration, and movement, etc., are compromised for a good part of the year. With the inclusion of a simple thre... Abstract Five years of meteorological and hydrological data from a typical New England watershed where winter snow cover is significant were used to drive and validate two off-line land surface schemes suitable for use in the Goddard Institute for Space Studies GCM: a baseline scheme that does not model the physics of a snowpack and therefore neglects the insulating properties of snow cover; and a modified scheme in which a three-layer snowpack is modeled. Comparing baseline model results with validation data reveals several model deficiencies. Surface radiation temperatures could not adequately be modeled and the ground froze to unreasonable depths. Furthermore, because of ground cooling resulting from large surface heat fluxes to the atmosphere from the uninsulated surface, deeper model layers did not unfreeze until midsummer. As such, the normal hydrologic processes of runoff, ground water infiltration, and movement, etc., are compromised for a good part of the year. With the inclusion of a simple thre...