Abstract:IT is widely known that seasonal snow cover in mountainous regions is a reliable indicator of water supply during the following spring season in snow-FED watersheds

The University of Wyoming, doing business as Sponsored Programs, was awarded a $603,274 Cooperative Agreement from the Bureau of Reclamation under the Snow Water Supply Forecasting Program (CFDA 15.078). The project aims to improve seasonal and sub-seasonal water supply predictions in Wyoming basins by incorporating the effects of climate oscillations, such as the Pacific North American Pattern and El Niño-Southern Oscillation, into stochastic snow models. The university will employ advanced non-Gaussian information metrics and stochastic snow modeling techniques to quantify the asymmetric and non-linear response of snowpack to these large-scale climate patterns. The findings from this 3-year project, which runs from January 2024 through September 2026, are expected to enhance water resource management and forecasting in the region.

This proposed study aims to improve seasonal and sub-seasonal water supply predictions by incorporating effects associated with the modes of low-frequency variability earth oscillations, such as the pacific north american pattern (PNA) and the EL niosouthern oscillation (enso). Cutting edge non-gaussian information metrics and stochastic snow models will be implemented for seasonal water resource forecasting for wyoming basins. This study will employ pires and perdigo 2007 non-gaussian information metrics (PP07), which can effectively project global oscillation effects on inland climatic and hydrologic conditions. In this analysis specifically, snow state variables (snow water equivalent, SWE), observables (e.g. snotel and modis data), and climatic indicators (e.g. PNA and enso) can be regarded as random variables, and joint probability density functions can be estimated as saddlepoint distributions based on edgeworth expansion. This analytical approximation describes a PDF with an edgeworth series over a hermite polynomial basis that includes higher-order statistical cumulants as coefficients, such as skewness, kurtosis, and high-order covariates. The expectation of conditional probability of SWE can also be interpreted as non-linear correlation. This novel information metric is proven effective for quantifying the asymmetric and non-gaussian response of monthly precipitation in most of europe to the north atlantic oscillation (NAO) index. For the proposed sub-seasonal water resource prediction, a stochastic snow model is implemented with initial conditions prepared by the PP07 metric from satellite driven data as well as ground-based snow survey data. Such a stochastic modeling approach is effective at integrating the sources of heterogeneities and uncertainties into probability distributions using a numerical model framework for all nineteen watersheds in wyoming. This snow model will be driven by the regional atmospheric force or snow mass exchanges informed by an information metric that uses the modes of low-frequency variability. Accordingly, this project aims to add non-gaussian effects to the basin-wide seasonal and sub-seasonal water supply forecasts in wyoming basins.