Hadi Shokati

Soil erosion models are important tools for soil conservation planning. Although these models are generally well-tested against plot and field data for in-field soil management, challenges arise when scaling up to the landscape level, where sediment trapping along landscape features becomes increasingly critical. At this scale, a separate analysis of model performance for representing erosion, sediment transport, and deposition processes is both challenging and often lacking. Here, we assessed the capacity of the spatially distributed erosion and sediment transport model WaTEM/SEDEM to simulate sediment yields in six highly instrumented micro-scale watersheds ranging from 0.8–7.8 ha, monitored over eight years from 1994–2001, in Southern Germany. The watersheds were composed of two groups: four field-dominated watersheds characterised by arable land with minimal landscape structures, and two structure-dominated watersheds featuring a combination of arable land and linear landscape structures (mainly grassed waterways along thalwegs) that minimise sediment connectivity. Arable fields in both watershed groups were managed for soil conservation, including no-till and optimised crop rotations. A Generalised Likelihood Uncertainty Estimation (GLUE) framework was employed to account for measurement and model uncertainties across multiple spatiotemporal scales. Our results show that while WaTEM/SEDEM captured the magnitude of the very low measured sediment yields in the monitored watersheds, the model did not meet our pre-defined limits of acceptability when operating on annual time steps. Model performance improved substantially when outputs were averaged over the eight-year monitoring period, with mean absolute errors of 0.14 t ha⁻¹ yr⁻¹ for field-dominated and 0.29 t ha⁻¹ yr⁻¹ for structure-dominated watersheds. Our findings demonstrate that WaTEM/SEDEM can represent the influence of soil conservation practices on reducing soil erosion and sediment yield in our study area. However, the model is fit for long-term conservation planning at larger spatial scales and not for precise annual predictions for individual micro-scale watersheds with specific conservation practices even if high-resolution, high-quality input data are available for parameterisation.

Flooding is a major natural hazard that requires a rapid response to minimize the loss of life and property and to facilitate damage assessment. Aerial imagery, especially images from unmanned aerial vehicles (UAVs) and helicopters, plays a crucial role in identifying areas affected by flooding. Therefore, developing an efficient model for rapid flood mapping is essential. In this study, we present two segmentation approaches for the mapping of flood-affected areas: (1) a fine-tuned Segment Anything Model (SAM), comparing the performance of point prompts versus bounding box prompts, and (2) a U-Net model with ResNet-50 and ResNet-101 as pre-trained backbones. Our results showed that the fine-tuned SAM performed best in segmenting floods with point prompts (Accuracy: 0.96, IoU: 0.90), while bounding box prompts led to a significant drop (Accuracy: 0.82, IoU: 0.67). This is because flood images often cover the image from edge to edge, making bounding box prompts less effective at capturing boundary details. For the U-Net model, the ResNet-50 backbone yielded an accuracy of 0.87 and an IoU of 0.72. Performance improved slightly with the ResNet-101 backbone, achieving an accuracy of 0.88 and an IoU of 0.74. This improvement can be attributed to the deeper architecture of ResNet-101, which allows more complex and detailed features to be extracted, improving U-Net’s ability to segment flood-affected areas accurately. The results of this study will help emergency response teams identify flood-affected areas more quickly and accurately. In addition, these models could serve as valuable tools for insurance companies when assessing damage. Moreover, the segmented flood images generated by these models can serve as training data for other machine learning models, creating a pipeline for more advanced flood analysis and prediction systems.

Soil stoichiometry governs nutrient cycling to ensure optimal ecosystem functionality. Although the soil carbon‑nitrogen‑phosphorus (C:N:P) stoichiometry and ecosystem functioning are closely related, much less is known about how environmental predictors regulate the spatial distribution of these ratios in temperate regions. Specifically, the statistical relationships of soil properties (such as soil pH and clay content), climate variables (like precipitation and temperature), and topographic features (i.e., slope and aspect) on C:N:P stoichiometric patterns at regional scales remain poorly understood. In our study, we combined Cubist machine learning for spatial predictions with state-of-the-art statistical approaches—generalized additive models and structural equation modeling—to disentangle the quantitative relationships between environmental predictors and soil C:N, C:P, and N:P ratios across Germany. The relative importance analysis of environmental predictors shows that soil pH is the major predictor of stoichiometric ratios, acting through its fundamental control on nutrient availability. Higher soil pH corresponded to lower stoichiometric ratios and vice versa. Latitude emerged as another important predictor due to its effect on temperature, which plays a crucial role in these ratios, such that increasing latitude corresponds to lower ratios. As expected, wall-to-wall spatial distribution maps of the stoichiometric ratios showed varying patterns due to different environmental predictor influences. These findings enhance our understanding of environmental-stoichiometric interactions and offer valuable insights needed for sustainable soil management in temperate regions.

Accurate estimation of soil moisture content (SMC) is crucial for effective water management, enabling improved monitoring of water stress and a deeper understanding of hydrological processes. While satellite remote sensing provides broad coverage, its spatial resolution often limits its ability to capture small-scale variations in SMC, especially in landscapes with diverse land-cover types. Unmanned aerial vehicles (UAVs) equipped with hyperspectral sensors offer a promising solution to overcome this limitation. This study compares the effectiveness of Sentinel-2, Landsat-8/9 multispectral data and UAV hyperspectral data (from 339.6 nm to 1028.8 nm with spectral bands) in estimating SMC in a research farm consisting of bare soil, cropland and grassland. A DJI Matrice 100 UAV equipped with a hyperspectral spectrometer collected data on 14 field campaigns, synchronized with satellite overflights. Five machine-learning algorithms including extreme learning machines (ELMs), Gaussian process regression (GPR), partial least squares regression (PLSR), support vector regression (SVR) and artificial neural network (ANN) were used to estimate SMC, focusing on the influence of land cover on the accuracy of SMC estimation. The findings indicated that GPR outperformed the other models when using Landsat-8/9 and hyperspectral photography data, demonstrating a tight correlation with the observed SMC (R2 = 0.64 and 0.89, respectively). For Sentinel-2 data, ELM showed the highest correlation, with an R2 value of 0.46. In addition, a comparative analysis showed that the UAV hyperspectral data outperformed both satellite sources due to better spatial and spectral resolution. In addition, the Landsat-8/9 data outperformed the Sentinel-2 data in terms of SMC estimation accuracy. For the different land-cover types, all types of remote-sensing data showed the highest accuracy for bare soil compared to cropland and grassland. This research highlights the potential of integrating UAV-based spectroscopy and machine-learning techniques as complementary tools to satellite platforms for precise SMC monitoring. The findings contribute to the further development of remote-sensing methods and improve the understanding of SMC dynamics in heterogeneous landscapes, with significant implications for precision agriculture. By enhancing the SMC estimation accuracy at high spatial resolution, this approach can optimize irrigation practices, improve cropping strategies and contribute to sustainable agricultural practices, ultimately enabling better decision-making for farmers and land managers. However, its broader applicability depends on factors such as scalability and performance under different conditions.

Soil erosion (SE) by water threatens global agriculture by depleting fertile topsoil and causing economic costs. Conventional SE models struggle to capture the complex, non-linear interactions between SE drivers. Recently, machine learning has gained attention for SE modeling. However, machine learning requires large data sets for effective training and validation. In this study, we present Erosion-SAM, which fine-tunes the Segment Anything Model (SAM) for automatic segmentation of water erosion features in high-resolution remote sensing imagery. The data set comprised 405 manually segmented agricultural fields from erosion-prone areas obtained from the rain gauge-adjusted radar rainfall data (RADOLAN) for bare cropland, vegetated cropland, and grassland. Three approaches were evaluated: two pre-processing techniques— resizing and cropping — and an improved version of the resizing approach with user-defined prompts during the testing phase. All fine-tuned models outperformed the original SAM, with the prompt-based resizing method showing the highest accuracy, especially for grassland (recall: 0.90, precision: 0.82, dice coefficient: 0.86, IoU: 0.75). SAM performed better than the cropping approach only on bare cropland. This discrepancy is attributed to the tendency of SAM to overestimate SE by classifying a large proportion of fields as eroded, which increases recall by covering most of the eroded pixels. All three fine-tuned approaches showed strong correlations with the actual SE severity ratios, with the prompt-enhanced resizing approach achieving the highest R2 of 0.93. In summary, Erosion-SAM shows promising potential for automatically detecting SE features from remote sensing images. The generated data sets can be applied to machine learning-based SE modeling, providing accurate and consistent training data across different land cover types, and offering a reliable alternative to traditional SE models. In addition, erosion-SAM can make a valuable contribution to the precise monitoring of SE with high temporal resolution over large areas, and its results could benefit reinsurance and insurance-related risk solutions.

Forest fires diminish forests’ ecological services, including carbon sequestration, water retention, air cooling, and recreation, while polluting the environment and endangering habitats. Despite considerable economic advancements, firefighting strategies remain less than optimal. This paper introduces the Bi-layer Predictive Ensemble (BIPE), an innovative machine learning model designed to enhance the accuracy and generalization of forest fire susceptibility mapping. BIPE integrates model-centric and data-driven strategies, employing automated methods such as 10-fold cross-validation and meta-learning to improve stability and generalization. During its 10-fold cross-validation, BIPE demonstrated excellent performance, with the Area Under the Curve (AUC) values ranging from 0.990 to 0.996 and accuracy levels consistently high, around 97%, underscoring its robust class separation ability and strong generalization across different datasets. Our results confirm that BIPE outperforms traditional high-performance models like Support Vector Machine (SVM), Multilayer Perceptron (MLP), Extreme Gradient Boosting (XGBoost), Deep Neural Network (DNN), and Convolutional Neural Network (CNN), showcasing its practical effectiveness and reliability on the data of nonlinear, high-dimensional, and complex interactions. Additionally, our forest fire susceptibility maps offer valuable complementary information for German forest fire management authorities, enhancing their ability to assess and manage fire risks more effectively.

Land suitability assessment, as an important process in modern agriculture, involves the evaluation of numerous aspects such as soil properties, climate, relief, hydrology and socio-economic aspects. The aim of this study was to evaluate the suitability of soils for wheat cultivation in the Gavshan region, Iran, as the country is facing the task of becoming self-sufficient in wheat. Various methods were used to evaluate the land, such as multi-criteria decision-making (MCDM), which is proving to be important for land use planning. MCDM and machine learning (ML) are useful for decision-making processes because they use complicated spatial data and methods that are widely available. Using a geomorphological map, seventy soil profiles were selected and described, and ten soil properties and wheat yields were determined. Three MCDM approaches, including the technique of preference ordering by similarity to the ideal solution (TOPSIS), gray relational analysis (GRA), and simple additive weighting (SAW), were used and evaluated. The criteria weights were extracted using Shannon’s entropy method. Random forest (RF) model and auxiliary variables (remote sensing data, terrain data, and geomorphological maps) were used to represent the land suitability values. Spatial autocorrelation analysis as a statistical method was applied to analyze the spatial variability of the spatial data. Slope, CEC (cation exchange capacity), and OC (organic carbon) were the most important factors for wheat cultivation. The spatial autocorrelation between the key criteria (slope, CEC, and OC) and wheat yield confirmed these results. These results also showed a significant correlation between the land suitability values of TOPSIS, GRA, and SAW and wheat yield (0.74, 0.72, and 0.57, respectively). The spatial distribution of land suitability values showed that the areas classified as good according to TOPSIS and GRA were larger than those classified as moderate and weak according to the SAW approach. These results were also confirmed by the autocorrelation of the MCDM techniques with wheat yield. In addition, the RF model showed its effectiveness in processing complex spatial data and improved the accuracy of land suitability assessment. In this study, by integrating advanced MCDM techniques and ML, an applicable land evaluation approach for wheat cultivation was proposed, which can improve the accuracy of land suitability and be useful for considering sustainability principles in land management.

Accurate spatiotemporal monitoring and modeling of soil moisture (SM) is of paramount importance for various applications ranging from food production to climate change adaptation. This study deals with modeling SM with the random forest (RF) algorithm using datasets comprising multispectral data from Sentinel-2, Landsat-8/9, and hyperspectral data from the CoSpectroCam sensor (CSC, licensed to AgriWatch BV, Enschede, The Netherlands) mounted on an unmanned aerial vehicle (UAV) in Iran. The model included nine bands from Landsat-8/9, 11 bands from Sentinel-2, and 1252 bands from the CSC (covering the wavelength range between 420 and 850 nm). The relative feature importance and band sensitivity to SM variations were analyzed. In addition, four indices, including the perpendicular index (PI), ratio index (RI), difference index (DI), and normalized difference index (NDI) were calculated from the different bands of the datasets, and their sensitivity to SM was evaluated. The results showed that the PI exhibited the highest sensitivity to SM changes in all datasets among the four indices considered. Comparisons of the performance of the datasets in SM estimation emphasized the superior performance of the UAV hyperspectral data (R2 = 0.87), while the Sentinel-2 and Landsat-8/9 data showed lower accuracy (R2 = 0.49 and 0.66, respectively). The robust performance of the CSC data is likely due to its superior spatial and spectral resolution as well as the application of preprocessing techniques such as noise reduction and smoothing filters. The lower accuracy of the multispectral data from Sentinel-2 and Landsat-8/9 can also be attributed to their relatively coarse spatial resolution compared to the CSC, which leads to pixel non-uniformities and impurities. Therefore, employing the CSC on a UAV proves to be a valuable technology, providing an effective link between satellite observations and ground measurements.

The estimation of soil moisture (SM) as an important variable in the hydrological cycle of nature is necessary for the optimal management of water and soil resources. One of the indirect methods to estimate SM is using visible imagery with unmanned aerial vehicles (UAVs). This study aims to evaluate the potential of visible UAV imagery for estimating SM in a bare soil field in Iran. In this study, M5 tree (M5P), random forest (RF), sequential minimal optimization regression (SMOreg), and multilayer perceptron (MLP) methods have been used for SM modeling from RGB (Red, Green and Blue) bands and brightness and intensity indices of aerial imagery. Three evaluation methods were used to assess the accuracy of the models, including the coefficient of determination (R2), root mean square error (RMSE), and mean absolute error (MAE). Four different indices, including difference index (DI), ratio index (RI), normalized difference index (NDI), and perpendicular index (PI), were used to estimate SM. The green and red bands pair were found to be the optimal bands for SM estimation. The findings showed that the PI index provided the most accurate SM estimates (R2 = 0.51). The RF model predicted SM most accurately among the machine learning models tested (R2 = 0.67). However, all models underestimated SM content in high-moisture areas and overestimated it in low-moisture areas, with the MLP model showing the most significant overestimation. All the indices were saturated beyond 25% SM. In general, this study highlighted the potential of aerial RGB imagery and associated indices for assessing SM levels within bare soil fields. However, it should be noted that the use of individual bands and indices alone is not sufficient to make an accurate estimate of SM.