A new approach to net solar radiation in a spatially distributed snow energy balance model to improve snowmelt timing

Authors: MEYER, JOACHIM - UNIVERSITY OF UTAH; Hedrick, Andrew; SKILES, MCKENZIE - UNIVERSITY OF UTAH

Submitted to: Journal of Hydrology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/27/2024
Publication Date: 6/13/2024
Citation: Meyer, J., Hedrick, A., Skiles, M. 2024. A new approach to net solar radiation in a spatially distributed snow energy balance model to improve snowmelt timing. Journal of Hydrology. 638. Article 131490. https://doi.org/10.1016/j.jhydrol.2024.131490.
DOI: https://doi.org/10.1016/j.jhydrol.2024.131490

Interpretive Summary: Seasonal mountain snowpacks are undergoing unprecedented declines in extent and amount, which is wreaking havoc on hydrologic predictions for downstream water supplies. A primary controlling factor on spring snowmelt is the amount of the sun's radiation that a snowpack absorbs, which is modulated by the reflectivity of the snow surface. This reflectivity, also known as snow albedo, can exhibit a great degree of spatial variability over short distances and is susceptible to dust and tree litter deposition causing increased absorption of the sun's radiation and thus faster melt. Physics-based snowmelt models often treat albedo simplistically due to its lack of ground validation and its interdependence with complex atmospheric processes. This study seeks to use satellite-derived snow albedo coupled with incoming solar radiation from a Numerical Weather Prediction (NWP) model in place of the simplified treatment of albedo in a physically based energy balance snow modeling framework. Model estimates of timing of complete melt out were improved to within one to six days of observations at mountain weather stations. Also, the prediction of areas without snow were improved by over 10%, resulting in higher confidence in model results.

Technical Abstract: Mountain headwaters have a demonstrated decline in the extent and amount of snow, putting this previously consistent natural reservoir at risk. The timing and magnitude of available freshwater from snowmelt are primarily driven by the amount of absorbed (net) solar radiation controlled by the snow albedo. However, solar radiation and snow albedo are not commonly measured at instrumentation sites in the mountains, yet have a high degree of spatial variability. With the sparsity of observations, physically based snow models typically use simplified modeling or time-decay functions, leading to errors in snowmelt rate and snow depletion timing. One option to advance the net solar representation is to use the incoming solar radiation output from the High-Resolution Rapid Refresh (HRRR; National Weather Service) numerical weather prediction model and a snow albedo remote sensing product based on the Moderate-Resolution Imaging Spectroradiometer (MODIS). The observed snow albedo product from MODIS combines clean snow albedo from Snow Covered Area and Grain Size (MODSCAG) and reduction in visible albedo from Dust Radiative Forcing in Snow (MODDRFS) and is spatially and temporally complete. This study showcases the use of hourly updated net solar radiation by combining HRRR and the snow albedo product in a spatially distributed snow energy balance model in the East River Watershed, Colorado, USA. The results improved simulated snow depletion days within one to six days compared to point observations and increased spatial agreement of areas with no snow from 86.6% to 96.8% during melt season. These achievements will increase the model combination potential for faster adaptation into local water forecaster operations to improve runoff predictions.