Time for 3D: UAV-based lidar for modelling splash erosion under vegetation

<p>Rain throughfall under vegetation is determined by characteristics of the vertical structure and the associated plant traits. It goes both ways: A protective layer of ground covering vegetation or leaf litter can decrease throughfall kinetic energy (TKE), whereas the formation of large drips in the canopy layers has been found to increase TKE. Abstracting the three-dimensional vegetation structure into usable quantitative metrics is challenging, and therefore these processes have not yet been sufficiently integrated into spatial erosion models. The vegetation splash factor (VSF) was designed to close this gap (Senn et al. 2020, DOI: 10.1002/esp.4820). The VSF quantifies the influence of vegetation on TKE and can be calculated from aerial lidar point clouds. In the first step, we derive the vegetation cover in a voxel space, which then allows modelling the proportional contribution of drips per layer to reach the ground. Hence, the approach is strictly based on the 3D structure rather than conventional forestry parameters, e.g. crown diameter or leaf sizes. Here, we present the result of the first application of the VSF in a small scale field study using splash cup measurements to validate and refine the concept.</p><p>We implemented the experiment in a mixed-broadleaf forest near Bretten, Germany with a beech and an oak-dominated plot to cover a variety of vertical forest structure configurations and a diverse composition of species. Each plot comprised two transects of ten splash cups to measure sand loss - as a proxy for TKE - during six individual rainfall events. In addition, we used micro-scale runoff plots to determine the effect of soil covering layers such as leaf litter or biological soil crusts in comparison to bare soil. The VSF was calculated in R with a voxel resolution of 0.5 x 0.5 x 0.25 m using a UAV lidar dataset. </p><p>Initial results from the splash cup measurements showed that young oak induced about 70 % higher TKE than adult beech trees. Among the individual cup positions, the lowest energy values were measured without canopy influence as freefall kinetic energy (FKE), TKE at positions with an intermediate young growth and shrub layer showed medium values. In near-trunk and mid-positions without intermediate layers, we measured TKE values more than twice as high as FKE. This resulted in significant sediment removal beneath the tree layer when the ground covering vegetation layer was removed, which is in accordance with studies from other ecosystems. Grouped according to these conventional vegetation structural criteria, we found that the calculated VSF values clustered around similar values and correlated with sand loss from splash cups. From these initial results, we assume general suitability of the VSF to reflect the influence of vegetation structure on TKE. Further, more detailed analysis will now be done to adjust and calibrate the VSF model to produce more indicative results. The preliminary findings presented here will be further expanded to be presented at vEGU21.</p>

Deep Learning Applied to Vegetation Identification and Removal Using Multidimensional Aerial Data

When performing structural inspection, the generation of three-dimensional (3D) point clouds is a common resource. Those are usually generated from photogrammetry or through laser scan techniques. However, a significant drawback for complete inspection is the presence of covering vegetation, hiding possible structural problems, and making difficult the acquisition of proper object surfaces in order to provide a reliable diagnostic. Therefore, this research’s main contribution is developing an effective vegetation removal methodology through the use of a deep learning structure that is capable of identifying and extracting covering vegetation in 3D point clouds. The proposed approach uses pre and post-processing filtering stages that take advantage of colored point clouds, if they are available, or operate independently. The results showed high classification accuracy and good effectiveness when compared with similar methods in the literature. After this step, if color is available, then a color filter is applied, enhancing the results obtained. Besides, the results are analyzed in light of real Structure From Motion (SFM) reconstruction data, which further validates the proposed method. This research also presented a colored point cloud library of bushes built for the work used by other studies in the field.

N!ow

Opening ParagraphN!ow is a belief concerning rain and cold which is held by the !Kung Bushmen in the region around Nyae Nyae in South West Africa.About ten inches of rain falls in an average year in this part of the Kalahari Desert and sinks into the deep sands. There is no run-off in streams and there are few water holes. The rain is sufficient to support a covering vegetation of grass, shrubs, and scrubby trees. The vegetation includes numerous edible roots, tubers, leaves, fruits, and nuts, for which the Afrikaans language provides the convenient word veldkos.