Search

Sinkholes and landslides occur when parts of a soil collapse mainly in more gentle or steeper slopes respectively, both often triggered by intensive rainfall. These processes often cause problems in the hilly regions in the “Golestan province” of Iran, and their detection is the essential aim for this research. The production of soil landforms maps is typically based on visual interpretation of aerial and satellite images eventually supported by field surveys. Recent advances in the acquisition of images from “unmanned aerial vehicles (UAV)” and of “deep learning (DL)” methods inherited from computer vision have made it feasible to propose semi-automated soil landforms detection methodologies for large areas at an unprecedented spatial resolution. In this study, we evaluate the potential of two cutting-edge DL 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 show that our framework can successfully map landslides in a challenging environment (with an F1-score of 69%), and that topographical derivates from “UAV-derived DSM” decrease the capacity of mapping sinkholes of the models calibrated with optical data.

Refereed/Peer-reviewed The study and mapping of evidences of soil piping using non-destructive techniques is a key issue in quantitative geomorphology. This paper attempts to combine Aerial mapping systems (Unmanned Aerial Vehicles (UAV)), soil physical and chemical attributes, and near-surface geophysical survey tools (ground penetrating radar and electrical resistivity tomography) to provide a comprehensive understanding of geometric features of soil piping in arid and semi-arid regions. The UAV Mapping System was applied to prepare ortho-photos, topography, analytical hill shading, drainage density, and land use maps of two different piping sites (site 1: rangelands, and site 2: agricultural lands) in Sarakhs plain, Razavi Khorasan Province, north eastern Iran. The physical and chemical soil attributes were analyzed in six soil profiles to check the hypothesis that these soil attributes control the occurrence of piping-related features (i.e., sinkhole, blind gully, gully), and to test if there are any differences in soil properties between the two land use types. The near surface geophysical tools were used to determine the approximate size of soil pipes, and to simulate their internal structure. The results of the derived UAV’s independent variables confirmed that typical soil erosion features related to piping (i.e., sinkhole, blind gully, gully) developed exactly in adjustment with subsurface processes (i.e., drainage density). The quantitative algorithms of pedology revealed significant differences of Na+ and SAR for soil profiles with and without piping erosion, but these soil properties themselves are not enough to explain piping development at the two study sites. The volume of GPR line surveys was 1701.5 m3 in site 1 and 1203 m3 in site 2. The potential distribution of subsurface pipes revealed by applying GPR were more extensive than those deduced from soil surface observations: three-dimensional pipe number density (number of soil pipes per unit soil volume; # m−3) of potential pipes and collapsed cavities simulated with a mean pipe length and pipe depth of 143.4 cm and 88 cm in rangelands and 175.75 cm and 79.14 cm in agricultural lands, respectively. The maximum density of pipes or pipe roof collapses occurred within soil depth boundaries of 0–50 cm at site 1 and 30–100 cm at site 2. The ratio of surface erosion features to subsurface potential piping in rangeland and agricultural lands was ca. 26.1 % and 10.4 %, respectively. The electrical resistivity tomography (ERT) measurements indicated higher resistivity values in piping-prone areas, where pipes were initiated at the bedrock interface. Integrating all data obtained by different techniques in this study allowed to better understand the extent of piping erosion in both surface and subsurface soil horizons.

Among the land degradation processes, gully erosion is the one that poses more environmental and societal challenges in arid regions. Predicting spatio-temporal gully development in a region under changing conditions is important to adopt proper mitigation measures. Here we investigate the Ghapan-Olya watershed in Golestan province in Iran, which is impacted by many erosional landforms, including gully phenomena. We applied the pixel-based distributed LANDPLANER model, to predict the location of rainfall induced gullies. Different input maps were prepared using UAV data, the region's soil, and seasonal land use information. Our modelling results from the application of topographic thresholds and the erosion index calculation were compared with field observations in locations with and without gullies and the application of quantitative metrics such as sensitivity, specificity, fallout, precision, and recall. Our study revealed that the spatial density of the gully’s location is more repetitious in the areas with altitude about 200-300m, steeper slope (between 15 to 30 degrees), and lower average accumulation value (

The identification of landslide-prone areas is an essential step in landslide hazard assessment and mitigation of landslide-related losses. In this study, we applied two novel deep learning algorithms, the recurrent neural network (RNN) and convolutional neural network (CNN), for national-scale landslide susceptibility mapping of Iran. We prepared a dataset comprising 4069 historical landslide locations and 11 conditioning factors (altitude, slope degree, profile curvature, distance to river, aspect, plan curvature, distance to road, distance to fault, rainfall, geology and land-sue) to construct a geospatial database and divided the data into the training and the testing dataset. We then developed RNN and CNN algorithms to generate landslide susceptibility maps of Iran using the training dataset. We calculated the receiver operating characteristic (ROC) curve and used the area under the curve (AUC) for the quantitative evaluation of the landslide susceptibility maps using the testing dataset. Better performance in both the training and testing phases was provided by the RNN algorithm (AUC ​= ​0.88) than by the CNN algorithm (AUC ​= ​0.85). Finally, we calculated areas of susceptibility for each province and found that 6% and 14% of the land area of Iran is very highly and highly susceptible to future landslide events, respectively, with the highest susceptibility in Chaharmahal and Bakhtiari Province (33.8%). About 31% of cities of Iran are located in areas with high and very high landslide susceptibility. The results of the present study will be useful for the development of landslide hazard mitigation strategies.

This investigation assessed the efficacy of 10 widely used machine learning algorithms (MLA) comprising the least absolute shrinkage and selection operator (LASSO), generalized linear model (GLM), stepwise generalized linear model (SGLM), elastic net (ENET), partial least square (PLS), ridge regression, support vector machine (SVM), classification and regression trees (CART), bagged CART, and random forest (RF) for gully erosion susceptibility mapping (GESM) in Iran. The location of 462 previously existing gully erosion sites were mapped through widespread field investigations, of which 70% (323) and 30% (139) of observations were arbitrarily divided for algorithm calibration and validation. Twelve controlling factors for gully erosion, namely, soil texture, annual mean rainfall, digital elevation model (DEM), drainage density, slope, lithology, topographic wetness index (TWI), distance from rivers, aspect, distance from roads, plan curvature, and profile curvature were ranked in terms of their importance using each MLA. The MLA were compared using a training dataset for gully erosion and statistical measures such as RMSE (root mean square error), MAE (mean absolute error), and R-squared. Based on the comparisons among MLA, the RF algorithm exhibited the minimum RMSE and MAE and the maximum value of R-squared, and was therefore selected as the best model. The variable importance evaluation using the RF model revealed that distance from rivers had the highest significance in influencing the occurrence of gully erosion whereas plan curvature had the least importance. According to the GESM generated using RF, most of the study area is predicted to have a low (53.72%) or moderate (29.65%) susceptibility to gully erosion, whereas only a small area is identified to have a high (12.56%) or very high (4.07%) susceptibility. The outcome generated by RF model is validated using the ROC (Receiver Operating Characteristics) curve approach, which returned an area under the curve (AUC) of 0.985, proving the excellent forecasting ability of the model. The GESM prepared using the RF algorithm can aid decision-makers in targeting remedial actions for minimizing the damage caused by gully erosion.

Landslides influence the capacity for safe and sustainable development of mountainous environments. This study explores the spatial distribution of and the interactions between landslides that are mapped using global positioning system (GPS) and extensive field surveys in Mazandaran Province, Iran. Point-pattern assessment is undertaken using several univariate summary statistical functions, including pair correlation, spherical-contact distribution, nearest-neighbor analysis, and O-ring analysis, as well as bivariate summary statistics, and a mark-correlation function. The maximum entropy method was applied to prioritize the factors controlling the incidence of landslides and the landslides susceptibility map. The validation processes were considered for separated 30% data applying the ROC curves, fourfold plot, and Cohen’s kappa index. The results show that pair correlation and O-ring analyses satisfactorily predicted landslides at scales from 1 to 150 ​m. At smaller scales, from 150 to 400 ​m, landslides were randomly distributed. The nearest-neighbor distribution function show that the highest distance to the nearest landslide occurred in the 355 ​m. The spherical-contact distribution revealed that the patterns were random up to a spatial scale of 80 ​m. The bivariate correlation functions revealed that landslides were positively linked to several linear features (including faults, roads, and rivers) at all spatial scales. The mark-correlation function showed that aggregated fields of landslides were positively correlated with measures of land use, lithology, drainage density, plan curvature, and aspect, when the numbers of landslides in the groups were greater than the overall average aggregation. The results of analysis of factor importance have showed that elevation (topography map scale: 1:25,000), distance to roads, and distance to rivers are the most important factors in the occurrence of landslides. The susceptibility model of landslides indicates an excellent accuracy, i.e., the AUC value of landslides was 0.860. The susceptibility map of landslides analyzed has shown that 35% of the area is low susceptible to landslides.

The aim of the current study was to suggest a multi-hazard probability assessment in Fars Province, Shiraz City, and its four strategic watersheds. At first, we construct maps depicting the most effective factors on floods (12 factors), forest fires (10 factors), and landslides (10 factors), and used the Boruta algorithm to prioritize the impact of each respective factor on the occurrence of each hazard. Subsequently, flood, landslides, and forest fire susceptibility maps prepared using a Random Forest (RF) model in the R statistical software. Results indicate that 42.83% of the study area are not susceptible to any hazards, while 2.67% of the area is at risk of all three hazards. The results of the multi-hazard map in Shiraz City indicate that 25% of Shiraz city is very susceptible to flooding, while 16% is very susceptible to landslide occurrences. For four strategic watersheds, it is notable that in the Dorodzan Watershed, landslides and floods are the most important hazards; whereas, flood occurrences cover the largest area of the Maharlou Watershed. In contrast, the Tashk-Bakhtegan Watershed is so sensible to floods and landslides, respectively. Finally, in the Ghareaghaj Watershed, forest fire ranks as the strongest hazard, followed by floods. The validation results indicate an AUC of 0.834, 0.939, and 0.943 for the flood, landslide, and forest fire susceptibility maps, respectively. Also, other accuracy measures including, specificity, sensitivity, TSS, CCI, and Gini coefficient confirmed results of the AUC values. These results allow us to forecast the spatial behavior of such multi-hazard events, and researchers and stakeholders alike can apply them to evaluate hazards under various mitigation scenarios.

The instability of soil aggregates is one of the most important land degradation processes in different parts of Iran, particularly in arid and semi-arid regions with loess-derived soils. In this study, we modeled soil stability (measured as soil bulk density (BD) of the topsoil layer) using two algorithms: random forest (RF) and support vector regression (SVR). In this study, we used nine environmental variables, driven by unmanned aerial vehicles (UAV) and soil samples collected from a depth of 0–20 cm to map BD. The RF algorithm was applied to prioritize the factors affecting the BD. The Results were assessed using RMSE, MSE, and R2. The results of the variable importance analysis showed that distance to a river, distance to road, and altitude are the most important factors affecting soil stability. The prediction model for mapping soil BD has reasonable accuracy, i.e., the RMSE values of the training set of RF and SVR models were 0.08 (R2 = 0.95) and 0.16 (R2 = 0.21), respectively. We recognized that the methodology of this study could easily produce a soil property probability map (BD in this study) in the topsoil layer.

Gully headcuts are due to erosion generated by concentrated overland flow, non-uniform infiltration, the presence of impermeable sub-surface soil layers, and a hydraulic gradient. It has been seen in an extensive range of continuous and categorical conditioning factors in several countries. Researchers fail to determine the factor that is the most important cause of GHs occurrence. This study determines the morphological and physio-chemical soil features, the locations with significant GHs, and the soil and geomorphic parameters in the loess area of Golestan Province, Iran. The spatial distribution of GHs was determined by field surveys and analysis of 1 m-resolution aerial photography collected with an UAV. Physical and chemical features of soil samples were determined through lab analysis. The relationships between features and GHs were ranked by statistical significance using chi-square tests. Functional trees - (FT), naive Bayes tree (NBTree), and random forest (RF) models were employed to generate GHs sensitivity maps. One-hundred and twenty-seven GHs were identified and randomly divided into a training dataset comprised of 70% (89) of the GHs, and a validation dataset encompassing 30% (38) of the GHs. Twenty-two GHs conditioning factors were input into the models. The results of the chi-square analysis confirmed land use (90.8%) had the greatest influence on GHs occurrence. The results of training indicated that the AUC value for the FT, NBTree, and RF models were 83.4%, 94.8%, and 96.5%, respectively. The confidence interval (CI) for these methods was 0.771–0.885, 0.905–0.975, and 0.926–0.987, respectively. The highest standard error (SE) was related to the FT model (0.0314), followed by NBTree (0.0151), and RF (0.0119). The validation results indicated an AUC of 0.888, 0.947, and 0.981 for these methods, respectively. Therefore, RF was found to be the most effective model for predicting and mapping GHs in the future.

Despite the importance of delineating spatial modelling of gully headcuts (GHs) in erosion-prone environments, assessments of the factors that control the occurrence of headcuts is lacking. To fill this gap in the research, we identified 129 GHS field surveys. These 129 cases were randomly divided into two groups: 90 GHs (70%) for model training and 39 GHs (30%) for model validation. Subsequently, new unmanned aerial vehicle (UAV) imagery is used to develop spatial modelling to predict the location of GHs at sites prone to soil erosion in Golestan Province, Iran. Mapping GHs enables evaluation of 4 machine-learning techniques (or ensembles) – best-first decision tree (BFTree), bagging best-first decision tree (Bag-BFTree), random-subspace best-first decision tree (RS-BFTree), and rotation-forest best-first decision tree (RF-BFTree) – for modelling GHs. We use the information-gain ratio method to analyze the relationships between GHS and 22 GH conditioning factors. The 4 ensemble outputs are validated using a receiver operating characteristic (ROC) curve. The areas under the curves (AUCs) for prediction rates of the ensemble methods applied to the training group are BFTree – 88.3%, Bag-BFTree – 92.7%, RS-BFTree – 95.7%, and RF-BFTree – 93.2%. The AUCs for the model-validation group cases, however, are 84.9%, 94.1%, 97.4%, and 9.18%, respectively. Therefore, RS-BFTree is, statistically, the most effective ensemble method for accurate modelling of GHs. Variable-importance analyses using information-gain ratio indicate that out of 22 GH-influential factors, land use, slope degree, and slope-length are of more importance in developing of GH occurrence. Finally, to address the need for detailed observations and highly accurate erosion data in the field, UAV image-acquisition technologies are demonstrated.

In the present study, gully erosion susceptibility was evaluated for the area of the Robat Turk Watershed in Iran. The assessment of gully erosion susceptibility was performed using four state-of-the-art data mining techniques: random forest (RF), credal decision trees (CDTree), kernel logistic regression (KLR), and best-first decision tree (BFTree). To the best of our knowledge, the KLR and CDTree algorithms have been rarely applied to gully erosion modeling. In the first step, from the 242 gully erosion locations that were identified, 70% (170 gullies) were selected as the training dataset, and the other 30% (72 gullies) were considered for the result validation process. In the next step, twelve gully erosion conditioning factors, including topographic, geomorphological, environmental, and hydrologic factors, were selected to estimate gully erosion susceptibility. The area under the ROC curve (AUC) was used to estimate the performance of the models. The results revealed that the RF model had the best performance (AUC = 0.893), followed by the KLR (AUC = 0.825), the CDTree (AUC = 0.808), and the BFTree (AUC = 0.789) models. Overall, the RF model performed significantly better than the others, which may support the application of this method to a transferable susceptibility model in other areas. Therefore, we suggest using the RF, KLR, and CDT models for gully erosion susceptibility mapping in other prone areas to assess their reproducibility.

The quantitative characterizations of piping collapses (PCs) and gully headcuts (GHS) are an important part of the study of geomorphological processes under different types of climate such as arid, semiarid, and temperate and non-climatic conditions like land use changes and agricultural machinery. The present study aimed to evaluate summary statistics to characterize PCs and/or GHS structure by using numerical functions as well as understanding the interaction of gully head features and collapse pipes by applying univariate and bivariate summary statistics. To this end, a 703.36 ha area was first selected in the loess-covered hilly region of Golestan Province, eastern-north of Iran. Then, the maps of 345 PCs and 133 GHS were obtained by a detailed reconnaissance surveys in the field by using an unmanned aerial vehicle (UAV) photography. Finally, the univariate (L, g, and O-ring functions) and bivariate (L12, g12, O12, g22(r) _ g11(r), g12(r) _ g11(r), and g1,1+2 _ g2,1+2 functions) summary statistics were used to investigate the statistical analyses of PCs and GHS, separately. Based on the results of the univariate summary statistics, the PCs had a clustered distribution and the pattern of GHs was aggregate. Based on bivariate summary statistics, gully headcuts were positively related to collapse pipes. In addition, PCs had more aggregation compared to the GHS. The neighborhood density of PCs was relatively less frequent around GHs than PCs. The PCs were spatially located in more dense parts of the study area. It means that PCs are positively correlated with each other and generally multi PCs should be put together to form one gully headcut. In general, the proposed summary statistics lead to a better understanding of the studied soil erosion processes in the study area.

Loess-derived soils in semi-arid regions are valuable resources. These regions have erosional landforms (e.g., piping, gully heads, and gullies) complexly produced by geo-environmental forces. Identification of the connections between landform patterns and the underlying geomorphological mechanisms is essential for understanding how erosive rainfall creates blind gully heads from pipes and gullies from gully heads. The goal of this study was to clarify the complex spatial interactions between geomorphological processes resulting from extremely intense extreme rainfall events. Fieldwork was conducted to map all collapsed pipes (single and multiple sinkholes), closed depressions, and gully heads using photogrammetric drones in 2018 and 2019 in a 2700-hectare area of loess-derived soils. In 2018, 837 pipes and 283 gully head locations were identified. In 2019, these numbered 1034 and 549. Geomorphic transformations were described statistically and the erosional landforms were compared to land-degradation trends. Piping and head-cut types are reflections of specific geomorphic conditions in the hilly loess topography. A conceptual model is proposed to characterize the landform shifts in loess deposits in four dynamic states. The spatial processes and interactions of collapsed pipes and gully heads reveal the natural processes underlying their formation and provide insights that my help to identify solutions to curtail destructive activities and mitigate the forces driving erosion in regions similar to the study area.

Rangelands, as one of the most valuable national assets, are characterized by plant communities which contain complex functional interactions. A major objective of plant ecology is simplification of complex spatial interactions, especially in monitoring activities. In this respect, vegetation indicators have an important role for monitoring the desertification trend in these crucial communities. Furthermore, identifying the connection between plant patterns and their underlying ecological mechanisms, such as competition and facilitation, is essential to predict how climate and human changes alter plant community organization. In this study, the conceptual model was introduced to summarize vegetation transition shifts in various dynamics states in rangeland. Firstly, detailed field studies were carried out to investigate the pattern of Halocnemum strobilaceum (Ha. St.) shrubs in Sufikam plain by using spatstat package in R software. Secondly, statistical results were introduced to the monitoring model for assessing vegetation transition shifts. The results of statistical analysis of spatial data in 2009 revealed that the Ha. St. shrubs had dispersed distribution up to scales of 17 m and these pattern was a significant departure from random labeling at the scales of 0–7 m. In 2017, the pattern of Ha. St. shrubs showed a greater tendency to clustered pattern at the scales of 0–18 m and after that they had a random pattern at the scales of 18–50 m. According to the new proposed model, the dispersed distribution in 2009 indicated a healthy state for Ha. St. shrubs, while a more aggregated pattern in 2017 reflected intraspecific competition for scarce resources. Thus, it was concluded that the model is useful for rangeland managers to get a correct understanding of eradication driving forces and destructive activities in these areas.

Gully erosion is a significant geo-hydrological phenomenon occurring worldwide. Such phenomena contribute significantly to the soil erosion on a catchment and to its morphological shaping. In so doing, it impacts all the processes acting in a hillslope. Being one of the principal processes of soil erosion, gully erosion will play an important role in worsening the effects of climate and land use changes in the near future. Therefore, the prediction of the spatial and temporal occurrence of these phenomena is an interesting problem both for the scientific world and more in general for the society. In this work we model gully erosion on different study areas, using a distributed pixel-based model. The LANDPLANER model was used for the purpose, mainly due to its ability to deal with scenario-based analysis and due to the limited requirements of input data in its basic configuration, which include a DEM and a land cover map to derive the parameters of the runoff through the Curve Nuumber method. The twofold erosion modelling schema integrated in LANDPLANER, namely a quasi-static topographic threshold approach and a dynamic simplified erosion index, were tested in some study areas, considering the local morphological, climatological and Land Use/Land Cover conditions. We propose a framework to derive such dynamical model input data to better characterize the spatial and temporal occurrence of gully phenomena. Open source (SENTINEL-2) and commercial satellite (WorldView 3) data are used for the purpose and with the main objectives to perform long term, seasonal, event and scenario-based modelling analyses. Modelling results were validated using geomorphological and gully phenomena inventories, using different spatial criteria and different performance metrics. We maintain that the tested modelling and validation approaches can be easily replicated and applied in other different study areas to better characterize the spatial and temporal occurrence of gully erosion phenomena.

When the effect of two soil erosion features, in this study, gully headcuts (GHs) and pipe collapses (PCs), is investigated jointly, a new comprehensive concept related to their controlling factors can be found. The objective of this paper to evaluate susceptibility to these two features in the hilly region of the Golestan Province (NE Iran), which may be helpful in land management and the sustainable development of the region. The maps of the controlling factors of GHs and PCs were constructed and the random forest algorithm was used to prioritize the factors controlling the occurrence of these two soil erosion features. The GHs and PCs susceptibility maps were prepared using the random forest model in the R software and validated applying the receiver operating characteristic curves, fourfold plot, and Cohen's kappa index, whereas the susceptibility map of the two soil erosion features was prepared by overlapping these maps. The results of factor importance analysis have indicated that land use, slope, and silt content are the most important factors in the occurrence of PCs, whereas slope gradient, silt content, and distance to streams are the most important in GHs occurrence. This means that steep and uncultivated slopes in the study area are most susceptible to GHs and PCs. The susceptibility models of GHs and PCs have excellent accuracy, that is, the area under the receiver operating characteristic curve values of GHs and PCs models were 0.960 and 0.935, respectively. The susceptibility map of GHs and PCs analysed jointly has shown that 66% of the area is not susceptible to any of these soil erosion features, whereas 15% is susceptible to both of them. The places susceptible to the two types of soil erosion features (GHs and PCs) indicate the areas where there is a high probability of GHs retreat due to the PCs. This study has confirmed that it is feasible to forecast the spatial behaviour of GHs and PCs occurrence and development. This is particularly preferred when the developed methods are applied for the susceptibility evaluation under various mitigation scenarios.

Gully erosion is an important soil degradation process, which under climate changes is projected to increase. Therefore, better understating of factors controlling gully erosion and prediction of gully headcuts' (GHs) location is still highly relevant. This study aimed to examine the spatial distribution of GHs and to assess the importance of pedological (i.e. aggregate stability, organic matter, bulk density, silt, clay, and sand content) and topographical factors (i.e. altitude, slope length, gradient, and aspect) using summary statistics and the maximum entropy (MaxEnt) model. The study was conducted in the loess-covered region of NE Iran. The highly precise data of 287 GHs locations were obtained by extensive fieldwork and the interpretation of UAV images. The spatial distribution of GHs was evaluated by univariate pair correlation function and O-ring statistics. The spatial effect of GHs density controlling factors was assessed by the cumulative density correlation function Cm,K(r). Variable importance was analyzed using the MaxEnt model, which was also for the susceptibility modelling of GHs. The results of univariate tests showed the aggregated distribution of GHs. The Cm,K(r) analyses indicated that the areas characterized by higher values of bulk density, aggregate stability, and organic matter content have lower GHs density, whereas the areas with high silt content and higher slope gradient have higher GHs density. According to the MaxEnt, there is no one single factor responsible for GHs location, but rather the combination of topographical and pedological factors with the predominance of slope gradient (0.86) and silt content (0.57). The MaxEnt modelling of GHs susceptibility has revealed that the best accuracy (0.958) is given when all pedological and topographical factors are used in the model. The susceptibility maps prepared in the study can be used for soil conversation and land use planning and, consequently, for sustainable development in the region.

Gully head erosion is a significant process in semi-arid regions contributing to land degradation and hillslope dynamics. Predicting when and where gully heads will expand under different seasonal scenarios is an essential step to take the appropriate mitigation measures. The LANDPLANER model was applied in this study area in order to analyze and investigate specific issues related to the triggering of gully head processes. The model is mainly designed to describe the dynamic response of slopes (or basins) under changing scenarios, including meteorological factors, vegetation or land use, and slope morphology. This model enabled us to predict the occurrence of gully heads under different seasonal conditions defined by Curve Number and rainfall scenarios. The LANDPLANER erosion modeling schema is twofold and integrates two different approaches. The first is empirical and exploits an erosion topographic threshold equation, while the second is aimed to define an erosion index calculated starting from the results of the hydrological model. These two indices predict where erosion processes can occur. To analyze the impact of land use changes on slope response to rainfall events, we simulated seasonal land use changes scenarios, modifying the Curve Number values associated to each specific land use classes. The twofold strategy was also adopted to validate the model, namely testing the model with synthetic dataset, and applying the model to real cases. The results indicated that the density of gully heads is more frequent in the South facing slopes with low elevation (around 200–300 m), steeper slope (within 20–40°), and higher average accumulation value (>100 m). Higher values of topographic threshold and erosion index (region 1: 13.1, region 2: 8.9) are obtained for the cold season in winter, where rainfall is intense and vegetation cover is lowest in the rangelands. Given to both observing receiver operating characteristic curve, the best and worst performances of gully head occurrence belong to the third Curve Number scenario (the lowest vegetation cover in terms of rangelands in winter) and the second Curve Number scenario (the lowest vegetation covers in terms of agricultural lands in summer) in turn. Further, the LANDPLANNER model effectively characterizes the status and trend of gully head formation in the study area.

Finding a digital elevation model (DEM) of suitable spatial resolution is vital to investigate piping erosion using aerial remote-sensing platforms like unmanned aerial vehicles (UAV). Previous studies have implied that the best spatial resolution is a DEM with the most detail. This study evaluates piping-affected areas with five DEMs (1, 5, 10, 20, and 30 m resolutions) with three trained machine-learning methods: support vector machine (SVM), maximum entropy (ME), and boosted regression tree (BRT). This method enables the identification of the specific impacts caused by changing pixel resolution to guide the selection of the most effective DEM. This study employs piping morphometry data to predict the locations of completely collapsed pipes. The performance of the methods for mapping of pipes was assessed against a piping inventory map. The results demonstrate that the finest resolution DEM is not always the most useful. Though 1 m-resolution DEMs show the most detail, the best performance was the 5 m-resolution DEM when tested for all three mapping models. The 5 m-resolution DEM-SVM combination was the best predictor of known piping sites (AUC = 81.0%). The 5-m DEM-ME was second most effective model (AUC = 75.8%). And 5-m DEM-BRT was third (AUC = 72.9%). Applying more DEM derivatives may increase confidence in the selection of the most appropriate resolution.

This study is performed to determine the importance of metaheuristic approaches and their accuracy with respect to the spatial modeling process emphasising on flood phenomena. In this study, four ensemble metaheuristic approaches, including the combinations of ANFIS (Adaptive Neuro-Fuzzy Inference System) with the genetic algorithm (GA), simulated annealing (SA), imperialist competitive algorithm (ICA), and differential evolution (DE), are used for flood zoning in Dingnan County, China. Based on historical records, aerial photo interpretation, field surveys, and Google Earth, the location of 115 floods are recorded in the study area. Different categorical and continuous factors affecting floods, including the plan curvature, altitude, distance to rivers, slope degree, rainfall, land use, lithology, stream power index, topographic wetness index, soil type, aspect, normalized difference vegetation index, and profile curvature, are identified in the study area and entered in GIS software. Flood susceptibility maps (FSMs) are validated using the ROC curve. The results confirm that the AUCs of the four combined metaheuristic models are larger than 79%. The highest AUC value is obtained for the ANFIS–GA ensemble (0.903), followed by the ANFIS–SA (0.843), ANFIS–DE (0.812), and ANFIS–ICA (0.798). The RMSEs obtained for the training data of the different models are 0.2562 (ANFIS-DE), 0.3121 (ANFIS-GA), 0.3345 (ANFIS-ICA), and 0.3398 (ANFIS-SA). The results of this study show that the proposed ensemble approaches are useful for flood hazard management and land use planning in Dingnan County, China, and other places.

It is of fundamental importance to model the relationship between geo-environmental factors and piping erosion because of the environmental degradation attributed to soil loss. Methods that identify areas prone to piping erosion at the regional scale are limited. The main objective of this research is to develop a novel modeling approach by using three machine learning algorithms-mixture discriminant analysis (MDA), flexible discriminant analysis (FDA), and support vector machine (SVM) in addition to an unmanned aerial vehicle (UAV) images to map susceptibility to piping erosion in the loess-covered hilly region of Golestan Province, Northeast Iran. In this research, we have used 22 geo-environmental indices/factors and 345 identified pipes as predictors and dependent variables. The piping susceptibility maps were assessed by the area under the ROC curve (AUC). Validation of the results showed that the AUC for the three mentioned algorithms varied from 90.32% to 92.45%. We concluded that the proposed approach could efficiently produce a piping susceptibility map.

Point pattern analysis of collapsed pipes as a subsurface erosion in natural and manmade conditions plays an important role to understand their landforms/features, predisposing factors, and for prevention and forecasting purposes. Thus, the present study aimed to evaluate the spatial pattern and the associated land factors of piping erosion in loess-derived soils in a semiarid climate in Golestan Province, Iran by applying numerical summary statistics. To this end, a 105 ha area with homogeneous environmental conditions (gentle slope and loess-derived soil) was selected and the maps related to 101 collapsed pipes and their features were obtained by the aerial photos provided by an unmanned aerial vehicle (UAV) with a 50 cm resolution and field survey. The mean of spatial distances between pipe locations was 309 m, and the frequency of pipes per hectare was 0.97. Moreover, soil samples of the pipe locations were collected and physical-chemical soil properties were measured in laboratory. Approximately 95% simulation envelopes were selected using the 5th-lowest and 5th-highest values of 199 Monte Carlo simulations of the null model of homogeneous complete spatial randomness. Based on the results of univariate pair correlation function, the significant aggregation of 101 collapsed pipes was observed at a scale of 0–50 m in both rangeland and agricultural land use types. The bivariate pair correlation function, which is considered to be the most informative second-order summary characteristic, was used for analyzing the statistical correlations between collapsed pipes and linear phenomena including distance from drainage networks, ridges, and roads. Based on bivariate summary statistics, collapsed pipes had positively been affected by both the distribution of drainage networks and ridges. However, the negative statistical correlations occurred between pipes and roads at the scales of 1–50 m. Also, the correlations of soil characteristics (silt content, exchangeable sodium percentage (ESP), the weight of soil in a given volume (bulk density), soil electrical conductivity (EC), and organic matter (O.M.)) of neighboring collapsed pipes were evaluated by mark correlation function. Based on mark correlation function kmm(r), a significantly positive correlation was found between the pipes density and silt content, ESP, and bulk density, when they are more than overall average. In addition, less values of EC and O.M. were positively related to the aggregation of collapsed pipes. Similar to the results of summary statistics, the maps confirmed all statistical correlations. Consequently, the outcome of this study highlights the spatial pattern of collapsed pipes and their associations in the study area.