Soil moisture construction is a very important way to reduce soil erosion and maintain crop growth in arid and semiarid regions of China. A microcatchment technique for soil erosion prevention widely used on soil slopes on the Loess Plateau are fish‐scale pits (FSPs). This technique is combined with afforestation techniques to store runoff, interrupt the dynamics of soil erosion, and help mitigate frequent water deficiencies encountered by forest trees. This study investigated the effects of FSPs on soil erosion dynamics including rill formation and runoff. A three‐dimensional (3D) laser scanner was used to evaluate the erosion resistance of a 15° slope treated with FSPs arranged in a triangular pattern and the hydraulic characteristics of the flow on this slope during several intermittent simulated rainfall events. The following results were obtained. (a) The FSP‐treated slope displayed eight progressive stages of erosion: splash erosion, sheet erosion, scouring by water streams, formation of the scour pits, rill erosion, down‐slope erosion, up‐slope erosion, and collapse of the pit walls. (b) As the rainfall duration increased, the runoff velocities at various locations on the slope fluctuated, but generally, the runoff velocities were significantly higher in the down‐slope positions than for the midslope and upslope positions. When the cumulative rainfall duration reached approximately 58 min and the total rainfall reached approximately 88.5 mm, the ability of the FSPs to intercept and store runoff rapidly decreased. In the first two rainfall events, both the runoff reduction benefits (RRB) and sediment reduction benefits (SRB) were positive, but following the third rainfall event, the SRB and RRB of the FSPs were negative. The accuracy of erosion parameters extracted using the 3D laser equipment and the ArcGIS software in comparison with the measured rill erosion parameters all less than 10%. The relative error between the measured sediment and the calculated sediment is within 5.66–22.13%. (c) During the rainfall process, the flow on the upslope was constantly in the laminar regime, but after the FSPs were completely filled with water, the laminar flow on the downslope transitioned into a turbulent flow. (d) As the cumulative rainfall duration increased, the degree of topographic relief and the number of rills increased, as did the measurements of surface roughness and flow resistance. In conclusion, the microcatchment methods reduced the rill erodibility and enhanced the soil's resistance to concentrated flow erosion.

Three micro-catchment measures that are named fish-scale pits (FSPs), artificial digging (AD), and contour plowing (CP) for soil erosion prevention are widely used in the Loess Plateau. To clarify the effectiveness of these measures in intercepting runoff and reducing erosion and the mechanism of water flow movement, intermittent simulated rainfall events was carried out in the 15° slopes with FSPs, AD, CP, and control slope (CK). The results demonstrated the following. (1) For cumulative rainfall <83 mm, three measures effectively intercepted runoff and reduced sediment compared with the CK. The runoff and sediment reduction effect of three measures gradually disappeared when cumulative rainfall increased to 83, 99, and 108 mm, and the sediment generation of the three measures successively exceeded that of the CK and was more than two times higher. (2) Laminar or transition flow occurred for the CK, and the flow pattern changed from subcritical to supercritical at 101 mm of cumulative rainfall. For three measures, the flow patterns became turbulent within a short time but remained subcritical. (3) A correlation analysis showed that the soil detachment rate, hydraulic shear stress, and stream power in the micro-catchment measures can be described using linear functions, which reduced the rill erodibility and enhanced the soil’s resistance to concentrated flow erosion. This research has important guiding significance on the rational and effective implementation of micro-catchment practices to prevent severe soil erosion and increase water storage for crop production on the Loess Plateau of China.

No abstract

Check dams are widely used on the Loess Plateau of China to control soil and water loss, develop agricultural land and improve watershed ecology. Detailed information on the spatial distribution of check dams and the area of dam land is critical for quantitatively evaluating hydrological and ecological effects, planning the construction of new dams and repairing damaged dams. Therefore, this research presents a method that integrates deep learning and geospatial analysis to facilitate the extraction of check dam areas in broad areas from high‐resolution Gaofen‐2 (GF‐2) multispectral imageries, including red (R), green (G), blue (B) and near‐infrared (NIR) bands and digital elevation model (DEM). First, we generated three datasets with different band combinations (RGB, RGB + NIR and RGB + DEM) using GF‐2 remote sensing images combined with Advanced Land Observing Satellite—the Phased Array type L‐band Synthetic Aperture Radar DEM data to determine the optimal data combination for dam area extraction. Next, four widely used semantic segmentation networks—Fully Convolutional Network (FCN), U‐Net, PSPNet and DeepLabv3+—were modified to support the arbitrary number of input channels and evaluated for check dam area extraction. Finally, the check dam candidate areas in the Yan River basin were extracted from DEM using geospatial analysis to optimize the dam area extraction results. The results showed that all deep learning (DL) models could extract dam areas quickly and accurately with mean intersection over union and overall accuracy values >85% and >98%, respectively. PSPNet had the best performance for testing datasets with different band combinations. We also found that the DL models in the RGB + DEM images had the best remote sensing image segmentation results, avoiding many miss‐classified pixels. The F1 scores in the RGB + DEM test dataset for FCN U‐Net, PSPNet and DeepLabv3+ reached up to 92%, or 3.3%, 2.3%, 1.9% and 3.1% higher than the corresponding values for RGB images. Potential check dam candidate regions (~3969 km2) were obtained for application analysis, which reduced the original area by 50%. The DL models for the RGB + DEM images were applied in the dam candidate regions, extracting 91.9 km2 of agricultural production dam land and 10.6 km2 of runoff and sediment silted dam land. The extraction results will facilitate quantitative analyses of check dams, improve the management of these structures and promote the efficiency of controlling soil losses.

No abstract