Recent work has demonstrated that hydrology models calibrated under historical climate may provide biased estimates of future runoff under a changed future climate, and that bias appears to vary with model. One reason is that model parameterization for a given model type has been shown to be dependent on climate. Another reason is structural, as other work has shown that the sensitivity of simulated runoff to climate change varies among different hydrology model types. The findings challenge the scoping process for planners tasked with assessing climate change vulnerabilities on water and environmental resources, selecting a hydrology model to serve the assessment, demonstrating that the chosen model is preferable among available options, and demonstrating that the model will simulate reliable estimates of runoff under climate change conditions.

This study explores whether model preferences may exist for such applications, focusing on four surface water hydrologic models that have been widely used to assess climate change impacts on hydrology in the western United States. The goal is to apply the models consistently to several basin settings, calibrating and verifying across historical yearsets that have contrasting climate conditions. The case study models include (ordered from more to less conceptual or from less to more physically-based): Thornthwaite Monthly Water Balance (TMWB), gridded SacSMA/Snow17 – also known as the Research Distributed Hydrology Model (RDHM), Precipitation Runoff Modeling System (PRMS), and Variable Infiltration Capacity model (VIC). The three case study basins are tributary to the Colorado River, feature runoff seasonality influenced by snowmelt, and represent basin settings varying from cooler alpine conditions (East at Almont, CO; Animas at Durango, CO) to warmer and more arid conditions (Salt at Chrysotile, AZ). For calibration and verification, three sets of seven discontinuous years were identified during 1987-2007 and relatively classified for each basin as warmer/drier, middle, or cooler/wetter. Models were then calibrated for both warmer/drier and cooler/wetter climates, and respectively verified during the yearsets having contrasting climate.

Presentation will summarize model development, challenges encountered when trying to consistently apply models, and modeling results revealing relative success in simulating monthly to annual runoff under climate change.

Collaborators include: U.S. Geological Survey (Lauren Hay, Greg McCabe), NWS Colorado Basin River Forecast Center (Kevin Werner), and Bureau of Reclamation (Tom Pruitt, Nancy Parker)