Sinkholes and landslides occur due to soil collapse in different slope types, often triggered by heavy rainfall, presenting challenges in the semi-arid Golestan province, Iran. This study primarily focuses on the detection of these phenomena. Recent advancements in unmanned aerial vehicle (UAV) image acquisition and the incorporation of deep learning (DL) algorithms have enabled the creation of semi-automated methods for highly detailed soil landform detection across large areas. In this study, we explored the efficacy of six state-of-the-art deep learning segmentation algorithms—DeepLab-v3+, Link-Net, MA-Net, PSP-Net, ResU-Net, and SQ-Net—applied to UAV-derived datasets for mapping landslides and sinkholes. Our most promising outcomes demonstrated the successful mapping of landslides with an F1-Score of 0.95% and sinkholes with an F1-Score of 89% in a challenging environment. ResUNet exhibited an outstanding Precision of 0.97 and Recall of 0.92, culminating in the highest F1-Score of 0.95, indicating the best landslide detection model. MA-Net and SQ-Net resulted in the highest F1-Score for sinkhole detection. Our study underscores the significant impact of DL segmentation algorithm selection on the accuracy of landslide and sinkhole detection tasks. By leveraging DL segmentation algorithms, the accuracy of both landslide and sinkhole detection tasks can be significantly improved, promoting better hazard management and enhancing the safety of the affected areas.

Introduction Landslides and sinkholes damage social, economic, and natural infrastructure. These processes have direct and indirect impacts on important infrastructure, including residential areas, and influence land use change and migration from rural to urban areas. Sinkholes and landslides occur when parts of a soil collapse mainly in more gentle or steeper slopes, which are often triggered by intensive rainfall. One of the main goals in sustainable land management is the identification and control of natural disasters, which on the one hand leads to the quantitative and qualitative improvement of production in the long term, and on the other hand, maintains the quality of the soil and prevents soil degradation. In order to manage better and more stable, it seems necessary to know how to change and identify different forms of erosion such as sinkholes and landslides. Sinkholes and landslides occur when parts of a soil collapse mainly in more gentle or steeper slopes, which are often triggered by intensive rainfall. Materials and methods Recent advances in acquiring images from unmanned aerial vehicles (UAV) (UAV) and deep learning (DL) methods inherited from computer vision have made it feasible to propose semi-automated soil landform detection methodologies for large areas at an unprecedented spatial resolution. In this study, we evaluate the potential of two cutting-edge DL deep learning segmentation models, the vanilla U-Net model, and the Attention Deep Supervision Multi-Scale U-Net model, applied to UAV-derived products, to map landslides and sinkholes in a semi-arid environment, the “Golestan Province” (north-east Iran). Results and discussion Landslides: The performance of the U-Net model shows that it has fewer false positives, but at the same time, it has missed many landslide cells. Meanwhile, the ADSMS U-Net model has performed better in detecting landslide cells, but it attributed many cases to incorrect predictions (which is explained by the low accuracy score). The best F1 score achieved for the ADSMS U-Net model is 0.68. Sinkholes: For all band combinations, the performances of ADSMS U-Net are better than those of the traditional U-Net model. The best overall scores by ADSMS U-Net were obtained when trained on the ALL data. Regarding the effectiveness of the various combinations evaluated in this study, we can observe the contradictory behaviors of the models. The traditional U-Net achieves the best performance using the RGB optical combination, while the ADSMS U-Net can leverage topographic derivative information and optical data, showing the best results with the ALL combination. Moreover, it is evident that the DSHC data alone provides the worst results for both models. In overall, the results show that the ability of ADSMS U-Net to predict landslides is closer to the ground reality compared to U-Net. This model identifies most of the landslides in the test sections. Also, for all combinations of sinkhole bands, ADSMS U-Net performs better than the U-Net model. The best overall scores were obtained by ADSMS U-Net when trained on ALL data. Conclusions Since this kind of soil erosion is the main origin of some major soil erosion including gully initiation and extension, applying new technology namely, UAV and deep learning is highly important and recommended. Our framework can successfully map landslides in a challenging environment (with an F1-score of 69 %), and topographical derivates from UAV-derived DSM decrease the capacity of mapping sinkholes and landslides of the models calibrated with optical data. Future research could explore the use of such an approach to map landslides and sinkholes over time to assess time-based changes in the formation and spread of natural hazards.

This study explores the spatio-temporal dynamics of the Normalized Difference Vegetation Index (NDVI) to detect early signs of land degradation. Utilizing high-resolution NDVI data from the Google Earth Engine, spanning from 2004 to 2023 with a 30-meter resolution, this research analyzes monthly variations. To illustrate these dynamics, the study focuses on Sabzevar County, located in northeastern Iran, which extends over 7,217 km^2and is approximately 220 kilometers distant from Mashhad. Validation of the NDVI data was performed using field observations from strategically located vegetation plots. One square meter plots were systematically established along 100-meter transects (10 transects in total), where the vegetation coverage in each plot was quantitatively assessed by experts. Comprehensive statistical analysis incorporated Kendall's tie test, alongside measurements of autocorrelation, coefficient of variation, and standard deviation, using R software to assess the trends and intensities of NDVI changes. The findings revealed a critical breakpoint in 2020, with increases in all three statistical indices—autocorrelation 0.82, coefficient of variation 0.65, and standard deviation 0.58—indicative of accelerating degradation prior to this year. Furthermore, the intensity of NDVI changes varied significantly across the study area, ranging from 0.05 in central and northern regions to 0.76 in the western parts. This research underscores the value of integrating field data with remote sensing technology to provide a robust analytical tool for early detection of land degradation. This method enables precise, timely assessment and proactive management of vulnerable ecosystems, particularly in arid regions.

Soil erosion is a primary geomorphic process that may result in hazards and significant socioeconomic losses. These processes occur mainly through the surface and subsurface flows. We conducted a systematic literature review on the quantitative attribution analysis of soil erosion, presenting state-of-the-art erosion processes and demonstrating the relative importance of soil erosion as a natural hazard responsible for land degradation and desertification. This explains why a multidisciplinary approach is needed to understand how erosion occurs and what factors are involved. This justifies the multihazard analysis and the need to model the erosion processes. Knowledge of the quantitative elements of soil erosion measurement combined with the consideration of multiple risk assessments can help develop conceptual models of slope hydrology and soil erosion that can help decision-makers determine an appropriate early warning system design policy. Filling these gaps will guide us to increase our knowledge of surface and subsurface erosion, thereby helping us to better explore the changing landscape for improvement and develop strategies and effective soil erosion control techniques. However, more research is required to better explore the morphology and connectivity of soil erosion, their subsurface watershed, and behavior, as well as several challenges, opportunities, and strategies facing the analysis of soil erosion.

This study assessed the soil erosion dynamics in Iran's Maharloo watershed using remote sensing indices (Normalized Difference Vegetation Index (NDVI), Normalized Difference Salinity Index (NDSI), and Topsoil Grain Size Index (TGSI)) and machine learning models (RF, SVM, and BRT). Landsat 8 satellite images (2005–2024) were processed via the Google Earth Engine, with field validation ensuring accuracy. Among the indices, TGSI (R² = 0.86), NDSI (R² = 0.89), and NDVI (R² = 0.87) showed the strongest correlations with ground data (Rain, Soil and Vegetation). The RF outperformed the other models (AUC = 0.89), identifying the central and western regions as warning erosion zones. Breakpoint analysis revealed abrupt changes in NDVI and NDSI (2013), while early warning signals (autocorrelation, variance, and skewness) indicated an escalating erosion warning, particularly near wetlands and rainfed fields. Spatial trends highlighted significant NDVI declines (Kendall's τ = 0.69) in wetland peripheries and NDSI increased (τ = 0.52) in northern farmlands. These findings underscore the efficacy of integrating machine learning and remote sensing for erosion monitoring, providing actionable insights for land management and conservation strategies.