Computer Modeling Snow Relationships in the Big Eau Pleine Watershed, Wisconsin

Abstract

To reduce problems in computer modeling the Big Eau Pleine Reservoir's winter oxygen conditions, it was considered desirable to have an index of solar radiation penetration through snow. To accomplish this, eleven measurements of solar radiation penetration through snow and ice were performed during the 1976 and 1977 winter using a Whitney submarine photometer and Thornthwaite net radiometer. A series of equations were obtained from literature and used to calculate both the reflectivity (albedo) and extinction coefficient of snow based on snow density. A comparison of these calculations with the field data showed a strong correlation for solar radiation penetration through snow (r = .961) and a slightly lower correlation for albedo (r = .928). Since a significant relation (alpha .05) was obtained in applying these equations, it was decided to combine them with a snow model so as to obtain a predictive tool of light penetration through snow. Two snow models, capable of calculating snow density, were obtained and calibrated for the winter periods of 1974 and 1975. Model calibration data was obtained from the Wisconsin Valley Improvement Company and consisted of snow data collected biweekly from a snow course within the Big Eau Pleine watershed. The two snow models selected were the LEAF (Leaf and Brink, 1973) modified according to Soloman (et al, 1976) arid the NPS (Donigian and Crawford, 1976). To maintain objectivity a computer controlled optimization technique (Monroe, 1971) was employed For a stability test the models were run over an additional two year period {1976 and 1977). A comparison of results from the simulations and snow course data showed the NPS model was superior (r = .945, s = .316 in., n = 36) to the LEAF model in simulating conditions within the watershed. Based on this superiority the NPS model was recalibrated with snow data collected on the reservoir during 1977. Comparing the calibration simulation with the field data showed good correlations for water equivalent (r = .978, s = .109 in., n =· 19), snow depth (r = .964, s = 1.112 in., n = 19) and snow density data (r = .957, s = 3.16%, n = 19). Once calibrated the NPS model was used to simulate the previous three winters. Output from the simulations was then utilized in conjunction with the solar equations to calculate solar radiation penetration through snow. Although limited data was available and ignoring the minimal effects of an ice cover, initial research shows a reasonable correlation (r = .881, n = 10) between the solar radiation calculations and average surface chlorophyll a concentrations within the reservoir during the 1976 snow accumulation season.