Variability of Cloud Cover and Its Relation to Spring Snowmelt in the Sierra Nevada

Estimation of surface radiation is necessary for understanding mountain snowmelt processes and remains a fundamental challenge due to the high variability of cloud cover. We use 17 years of NASA/NOAA GOES West albedo imagery at 4-km resolution and half-hourly sampling to characterize spatial and temporal variations of incident short wave radiation over the western U.S. with emphasis on the Sierra Nevada in California. The derived surface radiation is then related to snowmelt and streamflow variations during early springs through summers. Albedo data over areas with land surfaces below 800-m above sea level are masked to avoid influence of coastal marine and low stratus clouds, and effects of snow cover on albedo are removed using a 'clear-sky' filtering operation. High correlation between the albedo and in situ radiation from selected surface station observations indicates that the remotely sensed albedo is a valid proxy of the surface radiation. A rotated empirical orthogonal function (REOF) analysis of albedo data reveals that the first 2 REOFs account for 35% of the total anomalous variance over the western U.S., with the leading pattern (20% of the total variance) being centered over the Sierra Nevada. This variability is largest in the early snowmelt season and significantly smaller in the summer. Correlations between cloud albedo and Merced River discharge anomalies composited for the early snowmelt season (Mar-May) reveals that the radiation anomaly accounts for 9% of discharge variation, with a maximum variation of 14% in April. A composite of albedo anomalies mapped for early and late snowmelt years in the Sierra Nevada confirms that early (late) years are characterized by anomalously low (high) cloud albedo during the early spring period preceding the snowmelt pulse. Results thus far suggest that improving understanding of cloud cover variability in the high elevation region will be an indispensable contribution to hydrologic applications.