Digital Surface Models (DSMs) of the Whale's Tail Marsh region, South San Francisco Bay, CA

Horizontal accuracy was estimated by comparing ground control point (GCP) positions measured with PPK GPS measurements to their SfM-estimated positions. For each data set (images from a single acquisition date), five or more GCP positions were used as GCP check points. These positions were marked in the imagery, but they were left disabled (unchecked) during image alignment and sparse point cloud error reduction. The residual errors of the SfM-estimated positions of the check points relative to their GPS-measured position were used to calculate the root-mean-square error (RMSE), mean-absolute error (MAE), and mean error (bias). Following the Federal Geographic Data Committee National Standard for Spatial Data Accuracy (1998) guidelines, we use the RMSE to calculate the horizontal positional accuracy at the 95 percent confidence level. To account for additional uncertainty in the GPS-measured positions of the GCPs (0.020 meters) we include this as an additional error term through summation in quadrature to arrive at an estimate of total positional horizontal accuracy. The error statistics and horizontal accuracy estimates for each of the DSMs are below:
WhaleTailMarsh DSM 2021-05-30: Horizontal RMSE (m): 0.034 Horizontal MAE (m): 0.028 Mean error, north (m): 0.015 Mean error, east (m): 0.003 Horizontal accuracy at the 95 percent confidence level (m): 0.058 Total horizontal accuracy estimate at the 95 percent confidence level, including GPS uncertainty (m): 0.062
WhaleTailMarsh DSM 2021-09-09: Horizontal RMSE (m): 0.039 Horizontal MAE (m): 0.035 Mean error, north (m): 0.013 Mean error, east (m): 0.012 Horizontal accuracy at the 95 percent confidence level (m): 0.067 Total horizontal accuracy estimate at the 95 percent confidence level, including GPS uncertainty (m): 0.070
WhaleTailMarsh DSM 2021-11-14: Horizontal RMSE (m): 0.046 Horizontal MAE (m): 0.043 Mean error, north (m): 0.010 Mean error, east (m): 0.008 Horizontal accuracy at the 95 percent confidence level (m): 0.079 Total horizontal accuracy estimate at the 95 percent confidence level, including GPS uncertainty (m): 0.081
WhaleTailMarsh DSM 2022-02-10: Horizontal RMSE (m): 0.036 Horizontal MAE (m): 0.032 Mean error, north (m): 0.025 Mean error, east (m): 0.010 Horizontal accuracy at the 95 percent confidence level (m): 0.062 Total horizontal accuracy estimate at the 95 percent confidence level, including GPS uncertainty (m): 0.065
WhaleTailMarsh DSM 2022-05-19: Horizontal RMSE (m): 0.039 Horizontal MAE (m): 0.036 Mean error, north (m): 0.029 Mean error, east (m): 0.007 Horizontal accuracy at the 95 percent confidence level (m): 0.067 Total horizontal accuracy estimate at the 95 percent confidence level, including GPS uncertainty (m): 0.070
It should be noted that these accuracy estimates are for areas of bare ground or low vegetation in the locations where GCPs were placed. Although the GPS control on the image positions (from the post-processed camera-synchronized GPS) can reduce the need for on-the-ground control points, the overall accuracy of the final DSM may vary as a function of the distance to a measured ground control point. Areas where no ground control points were placed (Whale's Tail marsh north of Alameda Creek and any areas east of the unbreached north-south levy on the eastern edge of the marsh, as well as some areas along the marsh edge may have accuracies lower than these estimates. Additional sources of error such as poor image-to-image point matching due to vegetation or uniform substrate texture (such as mud) resulting in poor surface reconstruction may also cause additional localized errors in some portions of the DSM which exceed these estimates.