364 - Seasonal Snowpack Forecasting Using Fokker-Planck Equation Method with Non-Gaussian Information Metrics for Incorporating the Modes of Low-frequency Climate Variability

The effects of the El Niño–Southern Oscillation (ENSO) climate patterns on seasonal water availability have been recognized while it has not been explicitly incorporated into operational seasonal forecasting. It is because the conventional Gaussian regression approach cannot well characterize the complexities of the earth circulation system under influence of non-linear climatic variation. However, recently, a non-Gaussian approach was found to improve water resource predictions, including asymmetricity and kurtosis effects in addition to mean and variance metrics. This methodology has been successfully applied to inland European nations for seasonal stream flow predictions. The mathematical techniques used to determine saddle point joint distributions between multiple random variables are readily available. This presentation focuses on the stochastic snow model based on Fokker-Planck Equation framework, which can describe basin-scale snowpack evolution as probability density function (PDF). This adaptive stochastic-physically-based method can effectively incorporate the climatic variations. The proposed seasonal forecasting system uses the historical snowpack databases based on the SNODAS (the National Operational Hydrologic Remote Sensing Center, 2015) as well as the SNOTELs. This presentation will discuss the demonstrative case study based on the readily available mathematical technique to incorporate the earth system variables as a predictor of the atmospheric forcing for the snowpack evolution.