Modelling the frequency distributions of peak streamflows on the Canadian Prairies

Kevin Shook and John W Pomeroy, University of Saskatchewan, Centre for Hydrology, Saskatoon, SK, Canada

Contact First Author: Kevin Shook; kevin.shook@usask.ca

Abstract ID#: 35026

 

English Abstract:
The peak flows of streams in the Canadian Prairies are difficult to fit to existing frequency distributions for the design of hydraulic infrastructure. The problem is exacerbated by the very low density of stream gauging in most of the region, and by the short record lengths of most of the gauges. Furthermore, nonstationarity in the forcing climate and in the landscape (i.e. in vegetation, and drainage) prevents the use of historical data for estimating future design flows. Prior research has shown that the most commonly used method for estimating design flows for ungauged Prairie streams is hydrologically and statistically invalid, leading to poor designs for infrastructure in the region.

It is demonstrated that physically-based hydrological modelling using the Cold Regions Hydrological Modelling platform (CRHM) is able to reproduce the streamflows, and therefore their frequency distributions, of Prairie streams over their periods of record. Using downscaled climate data, the frequency distributions of Prairie streamflows could be simulated for any future interval. From the generated frequency distributions of peak flows, design flows could be extracted for any desired return period. The logistical requirements for producing return-period flows over the Prairie region are discussed.