Snow Never Falls on Satellite Radiometers: A Compelling Alternative to Ground Observations

Snowmelt is an important source of surface water for ecosystems, river flow, drinking water, and production of hydroelectric power. Thus accurate modeling of snow accumulation and melt is needed to improve our understanding of the impact of climate change on mountain snowpack and for use in water resource forecasting and management decisions. One of the largest potential sources of uncertainty in modeling mountain snow is the net radiative flux. This is because while net irradiance makes up the majority of the surface energy balance, it is one of the most difficult forcings to measure at remote mountain locations.

Here we investigate the use of irradiances derived from satellite measurements in the place of surface observations. NASA’s Clouds and the Earth’s Radiant Energy System (CERES) SYN satellite product provides longwave and shortwave irradiances at the ground on three-hourly temporal and one degree spatial resolution.Although the low resolution of these data is a drawback, their availability over the entire globe for the full period of March 2000 through December 2010 (and beyond, as processing continues) makes them an attractive option for use in modeling. We first assessed the accuracy of the SYN downwelling solar and longwave fluxes by comparison to measurements at NOAA’s Surface Radiation Network (SURFRAD) reference stations and at remote mountain stations. The performance of several snow models of varying complexity when using SYN irradiances as forcing data was then evaluated. Simulated snow water equivalent and runoff from cases using SYN data fell in the range of those from simulations forced with irradiances from higher quality surface observations or more highly-regarded empirical methods. We therefore judge the SYN irradiances to be suitable for use in snowmelt modeling and preferable to in situ measurements of questionable quality.