| Résumé | The purpose of this work was to improve the Canadian hydrokinetic (HK) energy database produced in 2014 using newly available data and methods. The two areas of the database that were identified to have the greatest opportunity for improvement were river width estimates and completion of the database in the Arctic. The previous database was not able to provide flow, depth, width, velocity, or power estimates in the Arctic area of Canada; thus, with newly available hydrologically and hydraulically driven flow data, this update could be achieved. Additionally, with the greatly increased availability of medium-resolution, openly accessible satellite imagery (i.e., Sentinel-2 Multispectral Instrument or S2MSI, Sentinel-1 Synthetic Aperture Radar or S1SAR, and the Landsat series), satellite imagery surface water detection methods have improved. Detecting waterbodies from space is now achievable at a 20m or 10m resolution; thus, river width can be extracted directly from imagery. The imagery used to generate river width estimates was S2MSI and the imagery was classified using the spectral index AWEI_NSH with a uniform threshold of -0.3. Imagery-derived widths were extracted in New Brunswick, Nova Scotia, and southern Gaspe (i.e., the Maritimes watershed, Watershed 1, or WS1). Additionally, river widths throughout Canada were extracted from Canada’s National Hydro Network (NHN) database.
Based on comparison to highly resolved satellite images, both NHN-derived and imagery-derived widths were able to substantially improve measurements of river width from the previous 2014NRC database. NHN-derived widths may be more accurate for small rivers and imagery-derived widths are likely to be more accurate for larger rivers. Nonetheless, the imagery-derived widths are expected to improve more the river width estimates, as it was observed that the NHN polygons can under- or overestimate the river width where the NHN polygons are non-representative of current conditions. The relationship between imagery-derived width and percent flow exceedance was weak in major watershed 1 (i.e., the Maritimes watershed), so the power estimates were derived with the mean imagery-derived width and the maximum and minimum widths were used to calculate the lower and upper limits of hydrokinetic power, respectively.
The updated HK power database allows the user to locate specific cross sections of rivers that have a high potential for feasible HK power extraction, rather than the reach-averaged approach used in the 2014NRC database. It is, however, less continuous than the 2014NRC dataset as estimates of HK power in lakes and reservoirs are not included. British Columbia contained the most rivers with continuous high HK power potential where the river flow was constrained within canyons or high slope landforms. In other major watersheds, such as St. Lawrence, Northern Quebec, Labrador, and Albany, high HK power cross sections were identified and generally corresponded to a location of white water. Braided rivers and rivers located at low slope landforms were generally predicted to have low HK power, unless a width constriction was present. The flow estimates from the NRC2014 database were less representative in underfit streams when compared to historical gauge data, and the NHN polygons sometimes poorly represented waterbodies in the prairies and the Yukon where rivers had meandered from their original position, causing errors in the HK power estimates.
Planned future work to upgrade the updated HK energy database are to link the river flow estimates in the Arctic to the database points representing Arctic rivers, derive river depth estimates for Arctic rivers based on the regionalization relationships used in the previous 2014NRC database, optimize the imagery-derived widths and apply the imagery-based width estimation in the ten other major watersheds in Canada, and perform further validation on the final HK power estimates in the database by comparing the estimates with power values calculated from field data and hydrodynamic modelling. |
|---|
