An experimental laboratory setup was developed and evaluated in order to investigate detachment of soil particles by raindrop splash impact. The soil under investigation was a silty loam Cambisol, which is typical for agricultural fields in Central Europe. The setup consisted of a rainfall simulator and soil samples packed into splash cups (a plastic cylinder with a surface area of 78.5 cm2) positioned in the center of sediment collectors with an outer diameter of 45 cm. A laboratory rainfall simulator was used to simulate rainfall with a prescribed intensity and kinetic energy. Photographs of the soil’s surface before and after the experiments were taken to create digital models of relief and to calculate changes in surface roughness and the rate of soil compaction. The corresponding amount of splashed soil ranged between 10 and 1500 g m−2 h−1. We observed a linear relationship between the rainfall kinetic energy and the amount of the detached soil particles. The threshold kinetic energy necessary to initiate the detachment process was 354 J m−2 h−1. No significant relationship between rainfall kinetic energy and splashed sediment particle-size distribution was observed. The splash erosion process exhibited high variability within each repetition, suggesting a sensitivity of the process to the actual soil surface microtopography.
An understanding of splash erosion is the basis to describe the impact of rain characteristics on soil disturbance. In typical splash cup experiments, splashed soil is collected, filtered, and weighed. As a way to collect additional data, our experiments have been supplemented by a photogrammetric approach. A total of three soils were tested across three sites, one in the Czech Republic and two in Austria, all equipped with rain gauges and disdrometers to measure rainfall parameters. The structure from motion multiview stereo (SfM-MVS) photogrammetric method was used to measure the raindrops impact on the soil surface. The images were processed using Agisoft PhotoScan, resulting in orthophotos and digital elevation models (DEMs) with a resolution of 0.1 mm/pix. The surface statistics included the mean surface height (whose standard deviation was used as a measure of surface roughness), slope, and other parameters. These parameters were evaluated depending on soil
Field observations and consecutive modelling of soil erosion events proved to be essential for understanding and predicting erosion and sediment transport. An experimental approach often utilizes a large variety of rainfall simulators. In this technical note a complex methodology is introduced, using a mobile rainfall simulator developed at the Czech Technical University in Prague. An experimental setup with two watered plots (16 + 1 m 2 ) was established, which enables simultaneous measurements in two scales and monitoring of surface runoff, flow velocity, infiltration, sediment subsurface flow, vegetation cover effect suspended solids and phosphorus transport, surface roughness and surface evolution under rainfall and other variables. The simulator is built on a trailer transportable by car with folding arm carrying four FullJet WSQ nozzles operating independently. The configuration and water pressure 0.7 bar leads to the total watered area 2.4 x 9.6 m. Average drop size (d50) reaches 1.75 mm for 0.7 bar pressure. Christiansen uniformity index CU reaches 85%. A selection of experimental results highlights both the advantages and the weaknesses of the presented experimental setup.
<p>Soil erosion is a global environmental problem. &#160;The rapid monitoring of the coverage changes in and spatial patterns of photosynthetic vegetation (PV) and non-photosynthetic vegetation (NPV) at regional scales can help improve the accuracy of soil erosion evaluations. &#160;Three deep learning semantic segmentation models, DeepLabV3+, PSPNet, and U-Net, are often used to extract features from unmanned aerial vehicle (UAV) images; &#160;however, their extraction processes are highly dependent on the assignment of massive data labels, which greatly limits their applicability. &#160;At the same time, numerous shadows are present in UAV images. &#160;It is not clear whether the shaded features can be further classified, nor how much accuracy can be achieved. &#160;This study took the Mu Us Desert in northern China as an example with which to explore the feasibility and efficiency of shadow-sensitive PV/NPV classification using the three models. &#160;Using the object-oriented classification technique alongside manual correction, 728 labels were produced for deep learning PV/NVP semantic segmentation. &#160;ResNet 50 was selected as the backbone network with which to train the sample data. &#160;Three models were used in the study; &#160;the overall accuracy (OA), the kappa coefficient, and the orthogonal statistic were applied to evaluate their accuracy and efficiency. &#160;The results showed that, for six characteristics, the three models achieved OAs of 88.3&#8211;91.9% and kappa coefficients of 0.81&#8211;0.87. &#160;The DeepLabV3+ model was superior, and its accuracy for PV and bare soil (BS) under light conditions exceeded 95%; &#160;for the three categories of PV/NPV/BS, it achieved an OA of 94.3% and a kappa coefficient of 0.90, performing slightly better (by ~2.6% (OA) and ~0.05 (kappa coefficient)) than the other two models. &#160;The DeepLabV3+ model and corresponding labels were tested in other sites for the same types of features: it achieved OAs of 93.9&#8211;95.9% and kappa coefficients of 0.88&#8211;0.92. &#160;Compared with traditional machine learning methods, such as random forest, the proposed method not only offers a marked improvement in classification accuracy but also realizes the semiautomatic extraction of PV/NPV areas. &#160;The results will be useful for land-use planning and land resource management in the areas.</p>
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.