Hydrologic Model Parameter and Process Sensitivity to Stable Isotope Data

Abstract:

Accurate prediction of streamflow in cold regions, for ungauged or recently gauged locations, or under changing climatic conditions, requires hydrologic models capable of representing the processes generating streamflow. Often, however, data for individual processes and storages are even less common than meteorologic or hydrometric data in mid-to-high latitude regions, and is more costly and less feasible to obtain. Stable isotope tracers have proven to be useful for process-based hydrological modelling, in calibration and model verification. The isoWATFLOOD hydrologic model, which has embedded deuterium and oxygen-18 simulations, has been applied to the Athabasca basin in northern Canada to evaluate the differences between simulation types in parameter and simulated process sensitivity.

The simulation response due to variation in parameter values will be analyzed using variogram analysis with the VARS-TOOL software. The sensitivity analysis will consider multiple model responses, including both the flow and isotope simulation responses, and the response in different sub-basins of the Athabasca. The impact of performance metric choice will also be explored. It is hypothesized that the embedded isotope tracer simulation will be sensitive to parameters and hydrologic processes that the streamflow model is insensitive to.