FROSTBYTE: Forecasting River Outlooks from Snow Timeseries: Building Yearly Targeted Ensembles

Release corresponding to data-driven forecasting workflow described in Arnal et al. (2024). This release will be superseded if peer review indicates further changes are needed.

FROSTBYTE is a reproducible data-driven workflow for probabilistic seasonal streamflow forecasting, based on streamflow and snow water equivalent station observations. The workflow leverages snow water equivalent (SWE) measurements as predictors and streamflow observations as predictands, drawn from reliable datasets like CanSWE, NRCS, SNOTEL, HYDAT, and USGS. Gap filling for SWE datasets is done using quantile mapping from nearby stations and Principal Component Analysis is used to identify independent predictor components. These components are employed in a regression model to generate ensemble hindcasts of seasonal streamflow volumes. This workflow was applied by Arnal et al. (2024) to 75 river basins with a nival (i.e., snowmelt-driven) regime and with minimal regulation across Canada and the USA, for generating hindcasts from 1979 to 2021. The study presents a user-oriented hindcast evaluation, offering valuable insights for snow monitoring experts, forecasters, workflow developers, and decision-makers.

Arnal, L., Clark, M. P., Pietroniro, A., Vionnet, V., Casson, D. R., Whitfield, P. H., Fortin, V., Wood, A. W., Knoben, W. J. M., Newton, B. W., and Walford, C.: FROSTBYTE: A reproducible data-driven workflow for probabilistic seasonal streamflow forecasting in snow-fed river basins across North America, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2023-3040, 2024.