Evaluating the use of structure-from-motion photogrammetry to monitor eroding banks of the black vermillion river in northeast Kansas

• Evaluating use of Structure-from-Motion (SfM) to monitor riverbanks with steep, high banks. • Comparing point clouds using a horizontal projection increased change detected. • Comparing models, matching cells were higher from leaf-off to leaf-off conditions. • SfM accurately estimated erosion from eroding banks but overestimated deposition. • Bank erosion rates were as high as 1.58 3 m per m 2 of bank face. Remote sensing techniques, such as terrestrial laser scanner (TSL), airborne light detection and ranging (LiDAR), and Structure-from-Motion Photogrammetry (SfM), can be used to survey riverbanks having three-dimensional complexity. This research focuses on SfM, which has comparable accuracies to LiDAR and TLS but with lower costs and shorter data collection times. SfM was used to monitor selected banks of the Black Vermillion River (BVR) in Northeast Kansas typified by seasonally varying amounts of herbaceous bank vegetation and high, vertical banks. Data was collected six times between December 2020 and July 2022. Newly established vegetation was removed from the SfM point clouds leaving voids, and volume change was calculated with empty and interpolated voids. Comparison of each study bank’s leaf-off to leaf-off-condition point cloud pair in both the typical vertical and horizontal directions, showed an 18 % increase in net erosion, 13 % increase in model overlap, yet 4.9-mm increase in error, when horizontal was specified. All banks (822 m) experienced an estimated −1,729 m 3 of net erosion during the study; however, one study bank experienced net deposition of 10.74 m 3 . Study banks with the highest amounts of net erosion (−409 m 3 , and –233 m 3 ) experienced mass-wasting erosional processes. The data suggest SfM is better at monitoring actively eroding banks where vegetation did not establish, as interpolated voids led to the overestimation of deposition. Understanding rates of sediment production will inform the management of Tuttle Creek Reservoir, which is downstream of the BVR and has lost approximately 50 % of its capacity from sedimentation.