Jean Stein

Snowmelt in a boreal forest site: an integrated model of meltwater quality (SNOQUALI)

An integrated model (SNOQUAL1) for the simulation of meltwater quality has been developed for a boreal forest site (Lake Laflamme, Quebec). The model consists of a physical module and a chemical module; it has been validated for H+, NO 3 - and SO 4 2- concentrations in discharge from the snowcover during the spring melt of 1984. Good agreement of simulated ion concentrations with those measured in the field were obtained for the strong acid anions NO 3 - and SO 4 2- . Greater deviations of simulated concentrations from observed concentrations were recorded in the case of H+. These are attributed to ion exchange reactions between the ion and the organic debris which is deposited continously onto the snowcover from the forest canopy during the winter season.

Simulation of snowmelt runoff pathways on the Lac Laflamme watershed

Abstract The contributions of subsurface and surface runoff to streamflow during snowmelt in a balsam fir forest located on a laurentidian upland watershed (Lac Laflamme) of Quebec (47°N, 71°W) were determined with a hydrological model (HYFOR). Lysimeter and calculated snowmelt runoff at the base of the snowpack were used as an input to the model during the snowmelt periods of 1985, 1986, and 1987. The calculated values were obtained from a temperature index snow cover model. The initial simulations showed a poor ability to predict runoff. The soil porosity and hydraulic conductivity were modified to account for the ground frost effect. Simulations were substantially improved by reducing soil porosities by 31% and hydraulic conductivities by 53% of their original values measured under the frost free conditions. Surface flow volume, which lumps rapid through-flow and overland flow, computed with the calibrated model varied between 32 and 47% of the total volume for the three snowmelt periods. Because of the significant modifications of soil parameters necessary to obtain reasonable model performance, it is concluded that better field observations on soil hydrologic properties are needed to improve snowmelt runoff simulations.

Évolution du couvert de neige et des profils hydrique et thermique du sol dans un bassin de la sapinière laurentidienne

Abstract Physical characteristics of the snowpack (depth, water-equivalent, density temperature) and the thermal and hydrological profiles of the soil have been observed in the balsam fir ecosystem of the Lac Laflamine basin, during three accumulation and snowmelt seasons. A network of 85 sampling points of snow and soil water covering the basin was installed in 1985. Temperatures within the snow cover and at different levels in the soil were measured in a station installed on a southerly exposed slope of the basin. As on other forested sites, the high spatial variability of snow water-equivalent (coefficient of variation of 20% during accumulation and up to 56% during snowmelt) is attributed to snow interception by the crowns and to the redistribution on the ground of densified snow from the branches. The statistical analysis shows a higher accumulation of 1–7 cm water equivalent and a faster snowmelt in the upper part of the slopes and on the divide compared to the downslope stations. The detailed study of snow profiles indicates temperature gradients of 0.1–0.2°C cm −1 and strong density gradients in the snow cover. Because of the thermal properties of snow, the soil temperature remains above −2°C and the freezing depth lies between 40 and 60 cm in spite of air temperatures frequently reaching −30°C in January. The soil hydrological regime would indicate a porous frost during winter.

Évaluation d'un modèle de fonte nivale en forêt boréale

Abstract SNOW-17, a temperature-index snowmelt model developed by Anderson (1973) was tested on an experimental plot in a boreal fir forest, 80 km north of Quebec city. Simulations were produced with a one hour time step for the melt seasons of 1981 to 1985. The calibration was done with 1982 and 1984 data. The outputs considered were: snowpack water equivalent, cumulated outflow, daily outflow, hourly outflow and some characteristics of daily hydrographs. The results are verified with field data and compared with those of a degree-day model previously calibrated on the same site. The calibration modified very little the a-priori attributed values of the model parameters. Some parameters appeared to be sensitive to alteration of the forest cover which occurred during the test period. In general SNOW-17 gave an accurate and precise estimation of the considered variables. Some weak points are identified. The results from SNOW-17 stayed consistently better than those provided by the degree-day model. SNOW-17 seems to be an efficient snowmelt model especially adequate for the boreal forest environment of eastern Canada.

Snowmelt Runoff in the Lake Laflamme Experimental Watershed, Quebec: Methodology and Preliminary Results

The application of two conceptual models (VSAS2 and HYFOR) to simulate the routing of water during the snowmelt period, has been carried out at the lake Laflamme forested watershed, near Quebec City, Canada. These models are based on the variable source area concept. They are coupled with a snow cover mass and energy balance model which computes snowmelt; the computed snowmelt is then used as an input in the water routing models. The methodology to simulate the snowmelt and the water flow in the basin requires several measurements which are described in this paper. The first results of snowmelt simulation obtained with the energy balance model are also presented.

Prediction of Snow Density and Temperature Changes within Layers of the Snowpack using a Point Energy and Mass Balance Model

The point energy and mass balance model of a snow cover developed by Anderson in 1976 has been used to simulate the temperatures and densities within the snowpack on two sites. The first site is located in a subarctic setting, near Fairbanks Alaska. The second one is located in a mid latitude temperate climate, north of Quebec City, Canada. For Alaska, simulations are presented and compared with observed data for the period between November 1981 and May 1982 while for the lac Laflamme site the results are presented for April and May 1985. In both cases the initialization is done at the beginning of the run. The comparisons show that the model simulates both variables quite well even in the subarctic setting.