Development and analysis of a Bayesian water balance model for large lake systems
share
Water balance models are often employed to improve understanding of drivers of change in regional hydrologic cycles. Most of these models, however, are physically-based, and few employ state-of-the-art statistical methods to reconcile measurement uncertainty and bias. Here, we introduce a framework for developing, analyzing, and selecting among alternative formulations of a statistical water balance model for large lake systems that addresses this research gap. We demonstrate our new analytical framework using a model customized for Lakes Superior and Michigan-Huron, the two largest lakes on Earth by surface area. The selected model (from among 26 alternatives) closed the water balance across both lakes and had a computation time of roughly 95 minutes - an order of magnitude less than prototype versions of the same model. We expect our new framework will be used to improve computational efficiency and skill of water balance models for other lakes around the world.
READ FULL TEXT
