Explaining Great Lakes water level variability through interpretable ensemble machine learning

Rahim Barzegar

doi:10.1016/j.scitotenv.2025.181302

Explaining Great Lakes water level variability through interpretable ensemble machine learning

2026-01-03

0

PRE

AI

DOI：10.1016/j.scitotenv.2025.181302

Original

Original request for help

Save

Abstract

En 中文

• Boosting-based ML models significantly improve Great Lakes water-level prediction. • SHAP–VARS framework identifies dominant drivers and reveals their lagged sensitivities. • Runoff and inflow govern Superior and Michigan, while inflow overwhelmingly drives Erie. • Temperature and evaporation exert strong long-lag effects, especially in Lake Ontario.

Explaining Great Lakes water level variability through interpretable ensemble machine learning

Abstract

Journal

Organization