Prairie Hydrological Model Study Final Report

This report describes the development of the Prairie Hydrological Model (PHM), a model that is suitable for hydrological process simulations in the prairie pothole region of Western Canada. The model considers all major prairie hydrological cycle, wetland storage, and runoff generation mechanisms and is capable of addressing the influences of changing land use, wetland drainage and climate variability. The purpose of this report is to describe the model, examine the performance of the model, and to demonstrate the model as a predictive tool for prairie hydrology. This purpose is achieved by using the model to analyze the impacts of wetland drainage and restoration as well as changes in surrounding upland land use on downstream hydrology. This focus on wetland drainage impacts required the development and testing of a new volume-area-depth (v-a-h) method for estimating wetland volume in the prairie pothole region. The method was incorporated into the PHM and improved the model’s ability to estimate wetland volume. The Cold Regions Hydrological Model platform (CRHM) is a computational toolbox developed by the University of Saskatchewan to set up and run physically based, flexible, object oriented hydrological models. CRHM was used to create the PHM for Smith Creek Research Basin (~400 km2 ), Saskatchewan. Two types of PHM runs were performed to estimate the basin hydrology. The non-LiDAR (Light Detection and Ranging) runs used a photogrammetric based DEM (digital elevation model) to estimate drainage area and hydrograph calibration to determine maximum depressional storage. The LiDAR runs used a fine-scale LiDAR derived DEM to determine drainage area and maximum depressional storage; use of LiDAR information meant that calibration was not required to set any parameter value. In both cases all non-topographic parameters were determined from basin observations, remote sensing and field surveys. Both LiDAR and non-LiDAR model predictions of winter snow accumulation were very similar and compared quite well with the distributed snow survey results. The simulations were able to effectively capture the natural sequence of snow redistribution and relocate snow from ‘source’ areas (e.g. fallow and stubble fields) to ‘sink’ or ‘drift’ areas (e.g. tall vegetated wetland area and deeply incised channels). This is a vital process in controlling the water balance of prairie basins as most water in wetlands and prairie river channels is the result of redistribution of snow by wind and subsequent snowmelt runoff. Soil moisture status is an important factor in determining the spring surface runoff and in controlling agricultural productivity. Unfrozen soil moisture content at a point during melt was adequately simulated from both modelling approaches. Both modelling approaches were capable of matching the spring streamflow hydrographs with good accuracy; the non-LiDAR approach performed slightly better than the LiDAR approach because the streamflow hydrograph was calibrated, whereas no calibration was involved in the LiDAR simulation. However, the LiDAR approach to simulation shows promise for application to ungauged basins or to changing basins and demonstrates that prairie hydrology can be simulated based on our current understanding of physical principles and good basin data that provides “real” parameters. The approach uses a ii LiDAR DEM, SPOT 5 satellite images and involved automated basin parameters delineation techniques and a new wetland depth-area-volume calculation. The new wetland depth-area-volume calculation used a LiDAR-derived DEM to estimate maximum depressional storage, a substantial improvement over estimates generated from simpler area-volume methods. This was likely due to the inclusion of information on depression morphology when calculating volume. Further, the process to retrieve the coefficients from a LiDAR DEM was automated and wetland storage was estimated at a broad spatial scale. A GIS model was created that can automatically extract the elevation and area data necessary for use in the new depth-area-volume method. Using the Prairie Hydrological Model, PHM, a series of scenarios on changing land use and wetland and drainage conditions was created from 2007-08 meteorological data. The scenario simulations were used to calculate cumulative spring basin discharge, total winter snow accumulation, blowing snow transport and sublimation, cumulative infiltration, and spring surface depression storage status. From these simulations, spring streamflow volumes decreased by 2% with complete conversion to agriculture and by 79% with complete restoration of wetlands; conversely it increased by 41% with complete conversion to forest cover and by 117% with complete wetland drainage. The greatest sensitivity was to further drainage of wetlands which substantially increased streamflow. Additional sensitivity analysis of scenarios on basin streamflow using historical (29-year periods: 1965-82 and 1993-2005) meteorology and initial conditions and current land use was carried out. Results showed that the effects of land use change and wetland drainage alteration on cumulative basin spring discharge volume and peak daily spring discharge were highly variable from year to year and depended on the flow condition. For both forest conversion and agricultural conversion and wetland drainage scenarios increased the long-term average peak discharge from current conditions, whereas wetland restoration reduced it. Forest conversion, agricultural conversion and wetland drainage scenarios increased the long-term average spring discharge volume by 1%, 19%, and 36% respectively; whilst the wetland restoration scenario reduced volumes by 45%. Several recommendations were made regarding the modelling challenges faced by this study and value of local meteorological data collection and using a LiDAR generated DEM for Prairie hydrological modelling purposes. It is recommended that similar studies be conducted in other geographic areas of the prairies where climate, soils, wetland configuration and drainage may produce differing results.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.