Gully erosion is the most severe type of water erosion and is a major land degradation process. Gully erosion susceptibility mapping (GESM)’s efficiency and interpretability remains a challenge, especially in complex terrain areas. In this study, a WoE-MLC model was used to solve the above problem, which combines machine learning classification algorithms and the statistical weight of evidence (WoE) model in the Loess Plateau. The three machine learning (ML) algorithms utilized in this research were random forest (RF), gradient boosted decision trees (GBDT), and extreme gradient boosting (XGBoost). The results showed that: (1) GESM were well predicted by combining both machine learning regression models and WoE-MLC models, with the area under the curve (AUC) values both greater than 0.92, and the latter was more computationally efficient and interpretable; (2) The XGBoost algorithm was more efficient in GESM than the other two algorithms, with the strongest generalization ability and best performance in avoiding overfitting (averaged AUC = 0.947), followed by the RF algorithm (averaged AUC = 0.944), and GBDT algorithm (averaged AUC = 0.938); and (3) slope gradient, land use, and altitude were the main factors for GESM. This study may provide a possible method for gully erosion susceptibility mapping at large scale.

Soil erosion triggered by water and wind pose a great threat to the sustainable development of Pakistan. In this study, a combination of geographic information systems (GISs) and machine learning approaches were used to predict soil water erosion rates. The Revised Wind Erosion Equation (RWEQ) model was used to evaluate soil wind erosion, map erosion factors, and analyze the soil erosion rates for each land use type. Finally, the maps of soil water and wind erosion were spatially integrated to identify erosion risk regions and recommend land use management in Pakistan. According to our estimates, the Potohar Plateau and its surrounding regions were mostly impacted by water erosion and have a soil erosion rate of 2500–5000 t·km⁻²·a⁻¹; on the other hand, wind erosion predominated the Kharan Desert and the Thar Desert, with a soil erosion rate exceeding 15,000 t·km⁻²·a⁻¹. The Sulaiman and Kirthar Mountain Ranges were susceptible to wind–water compound erosion, which was more than 8000 t·km⁻²·a⁻¹. This study offers new perspectives on the geographic pattern of individual and integrated water–wind erosion threats in Pakistan and provides high-precision data and a scientific foundation for designing rational soil and water conservation practices.

Gully erosion is considered to be a highly destructive form of soil erosion, often leading to the occurrence of natural calamities like landslides and mudslides. Remote sensing images have been extensively utilized in gully erosion research, and the suitability of extracting gully morphology parameters in various topographic regions needs to be clarified. Based on field measurements, this paper focuses on two widely used high-resolution remote sensing images: Unmanned Aerial Vehicle (UAV) and Google Earth (GE) imagery. It systematically examines the accuracy of gully morphological characteristic extraction using remote sensing in two regions with different terrain characteristics. The results show the following: (1) Compared to interpreting wide gullies with unclear shoulder lines, centimeter-level UAV imagery is more suitable for interpreting narrow gullies with clear shoulder lines. Conversely, the interpretability of sub-meter-level GE imagery is exactly the opposite. (2) The error in interpreting gully head points (GHPs) based on UAV images is less than 1 m, while the errors in gully length (GL), width (GW), perimeter (GP) and area (GA) are all below 3%, and these errors are hardly affected by gully morphology. (3) The error of GHPs based on GE images is concentrated within the range of 1–3 m. Meanwhile, the errors associated with GL, GP and GA are less than 10%. Conversely, the error of GW exceeds 11%. Furthermore, the aforementioned errors tend to increase as the gully width decreases and the complexity of the gully shoulder line increases. These findings shed light on the suitability of two commonly used remote sensing images for gully morphology extraction and provide valuable guidance for image selection in future research endeavors in this field.

Gully erosion susceptibility (GES) maps are essential for managing land resources and erosion control. Choosing the optimal horizontal resolution in GES mapping is a challenge. In this study, the optimal resolution for GES mapping in a complex loess hilly area on the Chinese Loess Plateau was tested using two machine learning algorithms. Unmanned aerial vehicle (UAV) images with a 9 cm resolution and GNSS RTK field-measured data were employed as base datasets, and 11 factors were used in the machine learning models. A series of horizontal resolutions, from 0.5–30 m, was used to determine which was the optimal level and how the resolution influenced the GES mapping. The results showed that the optimal resolution for GES mapping was 2.5–5 m in the loess hilly area, for both the random forest (RF) and extreme gradient-boosting (XGBoost) machine learning algorithms employed in this study. High resolutions overestimated the probability of gully erosion in stable regions, and it became difficult to identify gully and non-gully regions at too-coarse resolutions. The variable importance for GES mapping changed with the resolution and varied among variables. Slope gradient, land use, and contributing area were, in general, the three most critical factors. Land use remained an important factor at all the tested resolution levels. The importance of the slope gradient was underestimated at coarse resolutions (10–30 m), and the importance of the contributing area was underestimated at resolutions that were comparatively fine (0.5–1 m). This study provides an essential reference for selecting the optimal resolution for gully mapping, and thus, offers support for approaches attempting to map gullies using UAV.

In this paper, the way topographic spatial information changes with resolution was investigated using semi-variograms and an Independent Structures Model (ISM) to identify the mechanisms involved in changes of topographic parameters as resolution becomes coarser or finer. A typical Loess Hilly area in the Loess Plateau of China was taken as the study area. DEMs with resolutions of 2.5 m and 25 m were derived from topographic maps with map scales of 1:10,000 using ANUDEM software. The ISM, in which the semi-variogram was modeled as the sum of component semi-variograms, was used to model the measured semi-variogram of the elevation surface. Components were modeled using an analytic ISM model and corresponding landscape components identified using Kriging and filter bank analyses. The change in the spatial components as resolution became coarser was investigated by modeling upscaling as a low pass linear filter and applying a general result to obtain an analytic model for the scaling process in terms of semi-variance. This investigation demonstrated how topographic structures could be effectively characterised over varying scales using the ISM model for the semi-variogram. The loss of information in the short range components with resolution is a major driver for the observed change in derived topographic parameters such as slope. This paper has helped to quantify how information is distributed among scale components and how it is lost in natural terrain surfaces as resolution becomes coarser. It is a basis for further applications in the field of geomorphometry.

Ephemeral gully (EG) erosion is an important type of water erosion. Understanding the spatial distribution of EGs and other influencing factors at a regional scale is crucial for developing effective soil and water management strategies. Unfortunately, this area has not been sufficiently studied. The present study visually interpreted the EGs based on Google Earth images in 137 small watersheds uniformly distributed in the Loess Plateau, compared them with measured results, and analyzed the factors influencing EG formation and density using GeoDetector. The results showed that visually interpreting EGs from Google Earth images was suitable for EG regional studies. Out of the 137 small watersheds, 33.6% had EG occurrence with an average density of 3.41 km/km². Rainfall (R) and slope gradient (S) were the primary factors influencing the formation of EGs, while the area proportion of sloping farmland (APSF) and soil erodibility (K) were the main factors affecting EG density. The interaction of dual factors had a greater influence compared to single factors, with the interaction between S and Normalized Difference Vegetation Index (NDVI) having the greatest impact on EG formation and the interaction between K and NDVI on EG density. Although natural forces significantly influence whether EGs can form in a specific area, human activities greatly affect the density of the gullies that develop. This underscores the importance of proper land management in controlling gully erosion. These findings could provide theoretical support for EG prediction models and a scientific basis for soil and water loss control strategies at the regional scale.