Multi-fidelity Temperature Modeling to Better Understand Sensitivities and Uncertainties in a Managed Reservoir-River System
Abstract
Reservoir and in-stream temperature simulation models for the Upper Sacramento River provide a means to explore options for meeting summer demands for irrigation delivery, cold water habitat for endangered Winter Run Chinook salmon, and other environmental flows. Uncertainties in these models arise from model structure, parameters, hydrologic inputs, and meteorological forcing. The longer simulation time of the high-fidelity models typically used for seasonal analysis for the upper Sacramento River system (CE-QUAL-W2 and RAFT) limits the extent to which these uncertainties can be explored. Trading process fidelity for faster simulation, a simplified temperature modeling approach can provide an opportunity to quickly identify important sensitivities and uncertainties. We present an analysis of reservoir temperature sensitivity and uncertainty that uses results from an ensemble of simplified model runs to inform more targeted runs of the higher-fidelity CE-QUAL-W2 reservoir temperature model. The simpler model uses a coarser discretization and a simplified representation of internal reservoir thermal processes while 1) maintaining a transparent mapping between meteorological forcing and simulated water temperatures and 2) allowing a reasonable fit to historic temperature data. We analyze sensitivity of reservoir temperature to meteorological and hydrologic inputs obtained from each modeling approach separately and investigate the benefits of the combined multi-fidelity approach in providing insight relevant to cold water management in the Upper Sacramento River.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2020
- Bibcode:
- 2020AGUFMH150...12G
- Keywords:
-
- 1831 Groundwater quality;
- HYDROLOGY;
- 1869 Stochastic hydrology;
- HYDROLOGY;
- 1871 Surface water quality;
- HYDROLOGY;
- 1873 Uncertainty assessment;
- HYDROLOGY