DOWNSTREAM CHANNEL CHANGES AFTER A SMALL DAM REMOVAL: USING AERIAL PHOTOS AND MEASUREMENT ERROR FOR CONTEXT; CALAPOOIA RIVER, OREGON

Dam removal is often implemented without adequate baseline monitoring to distinguish background variability from any channel changes due to the removal. Thus, this study evaluated aerial photos as substitutes for multiple-year pre-removal field data to assess the downstream channel changes associated with a small dam removal. The Brownsville Dam, a 2.1 m tall concrete dam on the Calapooia River, Oregon, was removed in 2007. We surveyed cross sections during the summers prior to (2007) and following (2008) removal, and we analyzed aerial photos from 1994 to 2008. The field surveys and aerial photos were used to map the low flow channel and bars over 1.3 km downstream from the dam and in an upstream control reach. Three types of measurement error (position, identification, and wetted boundary datum error) were assessed for the field surveys and aerial photos in order to make the two measurement methods more comparable. The locations and magnitudes of changes between affected and unaffected areas revealed by these comparisons indicate that channel changes below the dam were a result of removal. For example, the first year after the dam removal, we observed the maximum changes in bar area (390 ± 125 m²) and wetted width (17 ± 3 m) immediately below the dam. In contrast, the maximum changes observed after dam removal in the upstream control were smaller (less than error for bar area and 9 ± 5 m for wetted width), while those observed prior to dam removal with aerial photos were predominantly far downstream and larger (32 ± 11 m for wetted width, and at least 880 ± 320 m² for bar area). From our analysis, we found aerial photos to be an acceptable but coarse substitute for multi-year pre-removal field data which could be improved by choosing and processing aerial photos to minimize errors. The errors which most impacted our analysis were photo specific but included wetted boundary datum error for photos taken at different times of year, position error for photos not processed expressly for this study, and identification error for small scale photos not scanned from film.