Comparing Remote Sensing Techniques in Detecting Salmonid Habitat, Salmon River, Oregon

Abstract:

Many restoration projects in the Pacific Northwest are implemented to improve habitat quality, quantity, and complexity for fish. Although numerous engineered log structures have been constructed in the hopes of achieving these goals, relatively few projects have been rigorously monitored to determine their success. This research seeks to compare the utility and application between photogrammetric and spectral depth approaches in detecting fish habitat in order to determine which method is more accurate and affordable for monitoring channel bathymetry. While each of these techniques has been individually studied, previous research has not directly compared and quantified their differences. Channel bathymetry data were collected by combining pre- and post-restoration digital photographs of the Salmon River in Northeast Clackamas County, Oregon, using structure-from-motion (SfM). The resulting 3D point cloud will be used to estimate water depths using photogrammetry and spectral depth. The photogrammetric method applies a refraction correction to the extracted water depth from the SfM topography to derive water depth. A regression between the surveyed water depth values and digital number values of surface pixels will derive depth. The resulting water depths from these two methods will be compared to the surveyed water depths for their accuracy and precision, particularly in critical salmonid habitats. The quantification of these differences will be an important contribution to river restoration science as it will allow for more accurate measurement and monitoring of changes in fish habitat. In the future, these data will be used in an eco-hydraulic River2D model to simulate changes in salmonid habitat availability after restoration.