VanDeWeghe et al. (2026) Advancing seasonal water supply forecasting for lakes and reservoirs using copulas

Identification

• Journal: Journal of Hydrology
• Year: 2026
• Date: 2026-07-13
• Authors: Alexander VanDeWeghe, Andrew D. Gronewold
• DOI: 10.1016/j.jhydrol.2026.136009

Research Groups

• Department of Civil and Environmental Engineering, University of Michigan
• School for Environment and Sustainability, University of Michigan

Short Summary

This paper introduces a novel copula-based seasonal-to-annual (S2A) water supply forecasting system for large lakes and reservoirs. The new approach uses vine copulas to capture the correlation structure among hydrological variables and outperforms current operational forecasts.

Objective

• Develop a statistical approach using high-dimensional distributions described by copulas to generate water supply ensembles that preserve the correlation structure among hydrological variables.
• Apply this approach to seasonal-to-annual forecasting on the Laurentian Great Lakes.

Study Configuration

• Spatial Scale: Continental-basin scale, focusing on the Laurentian Great Lakes system.
• Temporal Scale: Seasonal-to-annual (S2A) horizon, with a 12-month forecast period.

Methodology and Data

• Models used: Vine copulas, specifically regular (R-) vines, to capture high-dimensional dependencies among water balance components.
• Data sources: Historical monthly estimates of water balance component values from the Large Lake Statistical Water Balance Model, which uses a Bayesian framework to assimilate multiple data products.

Main Results

• The copula-based model reproduces the serial correlation in historical Great Lakes water balance components, especially for components with high autocorrelation.
• The model outperforms current operational forecasts, extending the forecast horizon from 6 to 12 months.

Contributions

• This study advances seasonal-to-annual forecasting for large lakes and reservoirs by introducing a novel copula-based approach that captures high-dimensional dependencies among hydrological variables.
• The results demonstrate the potential of vine copulas in water resources management planning and forecasting frameworks.

Funding

• This research was supported by grants from the National Science Foundation (NSF) and the U.S. Environmental Protection Agency (EPA).

Citation

@article{VanDeWeghe2026Advancing,
  author = {VanDeWeghe, Alexander and Gronewold, Andrew D.},
  title = {Advancing seasonal water supply forecasting for lakes and reservoirs using copulas},
  journal = {Journal of Hydrology},
  year = {2026},
  doi = {10.1016/j.jhydrol.2026.136009},
  url = {https://doi.org/10.1016/j.jhydrol.2026.136009}
}