Submitted to: Journal of Hydrometeorology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/26/2003
Publication Date: 2/1/2004
Citation: Dery, S.T., Crow, W.T., Stieglitz, M., Wood, E.F. 2004. Modeling snowcover heterogeneity over complex terrain for regional and global climate models. Journal of Hydrometeorology. 5:33-48. Interpretive Summary: A number of recent observational studies suggest the presence of ongoing climate change in the Artic environment. Understanding the causes and consequences of this change requires accurate climatic modeling of this region. An important component of such modeling is the hydrologic representation of snow accumulation during the winter and snow melt during the late spring. This paper investigates the suitability of various physical representations of snow cover and snowmelt within the land surface component of a climate model. Particular emphasis in placed on determining the impact of topographic heterogeneity and subsequent wind redistribution of the snow pack on model forecasting of the spring snow melt. The paper provides advice for climate modelers concerning appropriate representations of snow accumulation and melt processes at high latitudes. The research should eventually contribute to an enhanced understanding of climate change within the region.
Technical Abstract: We investigate some of the physical representations of snowcover and the subsequent impact on a range of processes, including snowmelt, land/atmosphere interactions, and the surface energy and water budgets, that requires special consideration in the Artic for regional and global climate models. The effects of small-scale variations in snowcover are identified as a critical process that controls to a large extent the hydrology of the Upper Kuparuk watershed, a 142 km2 river basin on the Alaskan North Slope. Incorporation of small-scale snowcover heterogeneity in two TOPMODEL-based land surface models leads to improved simulations of the water and energy budgets during the spring transition period, including the timing and amount of water discharge, of the Upper Kuparuk River. The two numerical models employ different techniques to incorporate these additional processes. In a quasi-dynamical application of TOPLATS formulations, one model uses a distributed approach to simulate snowpack conditions at a 100-m resolution across the entire Upper Kuparuk watershed. It captures successfully variations in the snowpack that arise from local differences in elevation and exposure to sunlight. By contrast, the other land surface model is quasi-statistical in nature and uses a simple parameterization to resolve the effects of wind-blown snow on the hydrology of the Upper Kuparuk Basin. It is demonstrated that by using the distributed snowmelt simulations, in addition to remote sensing data, constraints on the new model parameters required by the simple representation my be imposed for future applications in regional and climate model simulations.