Evaluation of snowfall estimates from the Grand Junction, Colorado WSR-88D during winter 2012-2013

Estimation of radar-observed snowfall in snow water equivalent (SWE) is perhaps the last frontier for radar quantitative precipitation estimation (QPE). Extreme challenges exist due to the geomorphic and microphysical properties of snow and limited radar sampling during the cool season. Microphysical variations cause uncertainty via the radar equation that includes ice density and particle size distribution. Additionally, radar coverage is especially poor in the western U.S. where three WSR-88Ds are sited near 3000 m MSL.

The Grand Mesa in western Colorado is an important region for water resource management from snow pack. Grand Mesa drainage supplies the City of Grand Junction and surrounding areas in addition to major agriculture and mining interests. This paper documents the performance of the Grand Junction KGJX WSR-88D, located on the Grand Mesa, during the winter of 2012-2013.

Three QPE algorithms are compared and validated: the WSR-88D new dual polarization (DP) algorithm, the WSR-88D Legacy (PPS) algorithm and the National Severe Storm Laboratory's “Q2” algorithm.

Ground validation includes a 2-D video disdrometer, a heated tipping bucket gauge, and three SNOTELs. The Stage 4 operational precipitation analysis product issued by the Colorado Basin River Forecast Center, and products from the Snow Data Assimilation System (SNODAS) issued by the National Hydrologic Remote Sensing Center are used as additional fields for comparison.

Both point and spatial validation are performed above 2500 m MSL providing a computation of acre-feet of water. Results indicate that radar SWE derived from the traditional reflectivity (Zh) provides the best estimates. Stage 4 is somewhat limited in spatial extent due to total reliance on gauge values although point SWE values are consistent with radar. SNODAS daily SWE changes appear to overestimate compared to the other products. Differences are amplified when AF of water is considered. Vertical correction methods within Q2 mitigate blockage on low radar elevation angles seen in the DP and Legacy products, thus improving spatial coverage and providing the optimal QPE overall.