Improving the accuracy of covariance estimates used in spatial modeling and estimation of snow water equivalent

With the increasing demand for water in the United States, it is essential that water resources be accurately monitored to insure that the avajlable water supply is used optimally. Hence, the National Weather Service (NWS) maintains a set of conceptual, continuous, hydrologic simulation models to generate extended streamflow predictions, water supply outlooks, and flood forecasts for river systems throughout the U.S. The accuracy of these forecasts is very dependent on the accuracy of snow water equivalent estimates simulated by the models. Hence, periodically throughout the snow season, the simulated snow water equivalent estimates are updated using snow water equivalent estimates obtained from snow data collected in river basins around the country. The estimates are obtained using a geostatistical model and snow course, SNOTEL, and airborne snow data. In this research, we exa.Inine ways to improve the accuracy of spatial covariance estimates that are necessary to obtain snow water equivalent estimates used to update the snow simulation models. We illustrate our methodology using snow data collected in the North Fork Clearwater River basin in Idaho.