An empirical approach to separate camera-based elevation change measurements due to sediment yield from other soil erosion masking processes

Over the past years studies (e.g., Hnsel et al. 2016 or Yang et al. 2021) have shown the feasibility of camera-based soil erosion assessment. This cost-efficient and non-invasive photogrammetric approach is a valuable tool to meassure soil surface changes (Balaguer-Puig et al., 2018). A challenging aspect nevertheless represents the masking of the sediment yield by surface lowering processes such as soil consolidation and compaction (Ehrhardt et al. 2022). Based on the camera elevation changes and measured field observations, we developed an approach to estimate these masking effects in the beginning of rainfall events and approximate a correction function. We conducted ten rainfall simulations at plots with 3 m length and 1 m width on agricultural slopes. The runoff and sediment concentration were measured at the plots outlet, while a time-lapse camera system surrounding the plot took images every few seconds. We furthermore collected data on soil bulk density, soil moisture, grain size distribution, total organic carbon, slope steepness, surface cover and surface roughness. To describe the changes of the soil surface at the beginning of the rainfall events, dominated by the masking effects, S-shaped curves were fitted via non-linear regression for each rainfall experiment. We then used the variables of those functions as well as the field observations as input values for an adjustment to estimate masking effects at the beginning of rainfall simulations as functions of soil and plot characteristics. The best results were achieved using four observations: grain size distribution, slope, bulk density and total carbon content. Our approach shows the potential to disentangle soil surface changes due to erosion and non-erosion processes at the onset of rainfall events. While the model worked well for most of the rainfall simulations, predictions were challenging for those events with strongly varying field observations. Especially difficult were those simulations conducted on freshly tilled soils. They showed high elevation changes at the beginning of the event that had great potential for soil consolidation and thus the mixed signals regarding the different processes were not separable by our approach. Nevertheless our study showed potential to increase the informative value of camera-based soil erosion measurements on agricultural fields. References Balaguer-Puig, M.; Marqus-Mateu, .; Lerma, J.L.; Ibez-Asensio, S. Quantifying small-magnitude soil erosion: Geomorphic change detection at plot scale. Land Degrad Dev 2018, 29, 825-834, doi:10.1002/ldr.2826. Ehrhardt, A.; Deumlich, D.; Gerke, H.H. Soil Surface Micro-Topography by Structure-from-Motion Photogrammetry for Monitoring Density and Erosion Dynamics. Front. Environ. Sci. 2022, 9, doi:10.3389/fenvs.2021.737702. Hnsel, P.; Schindewolf, M.; Eltner, A.; Kaiser, A.; Schmidt, J. Feasibility of High-Resolution Soil Erosion Measurements by Means of Rainfall Simulations and SfM Photogrammetry. Hydrol 2016, 3, 38, doi:10.3390/hydrology3040038. Yang, Y.; Shi, Y.; Liang, X.; Huang, T.; Fu, S.; Liu, B. Evaluation of structure from motion (SfM) photogrammetry on the measurement of rill and interrill erosion in a typical loess. Geomorphology 2021, 385, 107734, doi:10.1016/j.geomorph.2021.107734.