Stream Temperature Modeling: A Modeling Comparison for Resource Managers and Climate Change Analysis
Richard N. Palmer
In the Northeast U.S. increasing stream temperatures due to climate change pose a serious threat to cool and cold water fish communities, as well as aquatic ecosystems as a whole. In this study, three stream temperature models were implemented for two different case-study basins in the Northeast Climate Science Center region. Two coupled hydrology-stream temperature (physical) models were used: VIC-RBM and SWAT-Ficklin et al. (2012). The third model implemented was a nonlinear regression (statistical) model developed by Mohseni et al. (1998). Metrics were developed to assess these models regarding their prediction skill, data input requirements, spatial and temporal resolutions, and “user-friendliness.” This comprehensive assessment will be employed by aquatic resource managers in need of projected stream temperatures for management decisions in the face of climate change. Additionally, these models were used to predict stream temperatures under a range of future air temperature and precipitation scenarios for the study basins. These basins were the Westfield Basin (1,338 km2) in western Massachusetts and the Milwaukee Basin (2,220 km2) in Wisconsin. The climate change analysis was performed using a range of potential precipitation changes and air temperature increases (similar to a climate stress test). Precipitation scenarios ranged from 90% of observed to 130% of observed (in increments of 10%) and daily air temperature increases ranged from 0° C to 7° C (in increments of 1° C); the combinations of 5 precipitation scenarios and 8 air temperature scenarios yielded 40 different climate scenarios that were evaluated by each model. The impacts of climate change on these temperature and precipitation ranges was determined for the two watersheds and during specific seasons of the year.