Musselman et al., 2010

Talk/Poster

Simulating plot-scale variability of snowpack states in conifer forests using hemispherical photography and a process based one-dimensional snow model.

Musselman, K.N., Molotch, N.P., Margulis, S.A., Lehning, M., Kirchner, P.B., Bales, R.C. (2010)
Fall meeting, American Geophysical Union, December 2010. Abstract C33E-0590.  

Abstract

Energy balance and snowmelt were studied using distributed measurements and model simulations spanning a 2290 - 2660 meter elevation band of a dense red-fir forest in the Wolverton Basin of Sequoia National Park. The site is part of the Southern Sierra Nevada Critical Zone Observatory. This research was aimed at developing a better process understanding of how forest cover affects the magnitude and variability of sub-canopy snow accumulation and melt, as well as the distribution of soil moisture and catchment water yield. The representation of processes controlling sub-canopy snow dynamics remains a source of significant uncertainty in land-surface and hydrologic models. To improve understanding of the feedbacks between forest canopy structure and the spatio-temporal evolution of snow properties and soil moisture, we apply a one-dimensional snow, vegetation and soil model (SNOWPACK) with a modified treatment of sub-canopy solar radiation. With the use of hemispherical photography, the model separately estimates direct and diffuse shortwave canopy transmissivity. When forced at high temporal resolution with above-canopy meteorological data, the model adequately captures the variable nature of sunflecks, which explains much of the heterogeneity in sub-canopy incident shortwave radiation. Three-year comparisons of model results with the in-situ measurements indicate improved estimation of depth and SWE using the detailed canopy information versus the same model but with canopy transmissivity parameterized as a function of leaf area index. The modified model improved estimates of melt out date (MAE < 5 days) compared to the parameterized canopy model (MAE > 11 days). The results of this work may be used to inform the development of improved forest cover parameterizations necessary for regional climate models, macro-scale hydrologic models, and ecosystem models. In this regard, our future efforts aim to improve predictions of the hydrologic impacts of changes in climate and vegetation distribution.

Citation

Musselman, K.N., Molotch, N.P., Margulis, S.A., Lehning, M., Kirchner, P.B., Bales, R.C. (2010): Simulating plot-scale variability of snowpack states in conifer forests using hemispherical photography and a process based one-dimensional snow model. Fall meeting, American Geophysical Union, December 2010. Abstract C33E-0590..