spatial modeling

Characterized by fragile ecosystems, dry regions are prone to frequent damage and wind erosion, making it crucial to develop effective strategies for accurately assessing wind erosion. Therefore, this study sought to model and prepare a wind erosion risk map using remote sensing and machine learning approaches in Sabzevar City. Throughout the past decades, this region has continuously suffered from land degradation, water and wind erosion, land use conversion, and groundwater depletion. Thus, this study identified the most important parameters contributing to wind erosion in the study area after reviewing the sources of wind erosion and its evaluation models. Moreover, a distance measurement index was considered for each individual parameter. Also, machine learning methods were used to model and prepare the wind erosion map of Sabzevar city using the remote sensing indicators prepared and auxiliary points obtained by reviewing previous studies. Finally, to reduce the uncertainty of the results, the combined modeling method was used to prepare the wind erosion map. 

fifty ground data points were collected through performing field visits and reviewing the previous studies conducted in the study area. Accordingly, eight factors, including wind speed, land use, soil salinity, soil moisture, soil texture, vegetation density, precipitation, and digital elevation model (DEM) were selected based on the review of the related literature review. Then, to model wind erosion in the study area, the values found for each of the aforementioned factors were correlated with the field data. Moreover, four machine learning techniques, including SVM, GBM, GLM, and RF were used to predict wind erosion risk in the study area. In addition, the kappa coefficient, ROC curve, and True Skill Statistics (TSS) were used to assess model performance. Finally, a combined model was employed in the SDM statistical package to minimize the uncertainty of regional modeling. 

The SVM model demonstrated the best performance with AUC = 0.95, TSS = 0.97, and kappa = 0.87. The results indicated that the eastern regions of the study area were affected the most by severe wind erosion. Moreover, the combined model revealed that 44% of the area fell within the low wind erosion class (5340 square kilometers), 16% in the medium class (2007 square kilometers), 15% in the severe class (1916 square kilometers), and 25% in the very severe class (3122 square kilometers). The results of the significance of the variables showed that wind speed, precipitation, vegetation density, and surface moisture (NDMI) were the most important factors that contributed to the region’s wind erosion, respectively. Furthermore, land use change, soil salinity, and soil surface texture were found to have the least contribution to wind erosion in the study area, respectively. These findings can help the relevant decision-makers to set managerial policies and practical plans to avoid the risks of wind erosion. 

effective planning, optimal management, and corrective measures can be made and implemented to prevent land degradation and wind erosion in the affected areas by using the combined model with the least uncertainty. Generally, modeling wind erosion and preparing its map play a fundamental role in environmental studies. In this study, four models were applied in the form of an SDM statistical package, where the accuracy of the evaluation results clearly identified the best model. The difference in the results obtained from the performance of the models confirmed the uncertainty between them. Therefore, it appears that using a combined approach can be a suitable solution to reduce the uncertainty of modeling. The results of the study also suggested that the western areas of Sabzevar City were the most important centers of wind erosion production, where the surface of the hills was loamy and sandy. Thus, to avoid wind erosion in the study area, some corrective and biological measures are suggested, including the cultivation and development of salinity-resistant plant community cover, the cultivation of sand-loving plants, and the construction of windbreaks for sand dunes. While the measures performed so far in combating dust storms and wind erosion in Sabzevar city have proved to be ineffective, the intensified occurrence of dust in the region throughout the recent years and the irreparable consequences it has brought about make it necessary to devise and implement a comprehensive plan to deal with those phenomena. In this regard, the wind erosion risk intensity maps prepared in this study can be an efficient and appropriate tool for managing and reducing the effects of wind erosion and land destruction.

Identifying plant species distribution and potential habitats which are under degraded in the rangelands ecosystem is an essential challenge in natural resource science. Performance these studies will support rangeland conservation, restoration, and management measures. In this study, the potential habitat of Trigonella elliptica in rangeland of Yazd province was modeled using one of the advanced machine learning models (Random Forest algorithm). We used 11 variables including land use, soil salinity index, rainfall, minimum and maximum temperature, evaporation, elevation, aspect and degree of slope, distance from river, and topographic wetness index, as well as 103 presence points of T. elliptica to improve the model. 70 % of the T. elliptica presence points were randomly selected for model training and 30% of them for model testing. In order to evaluate the model and the importance of environmental variables were used the area under the receiver operator characteristic (ROC) curve and the Jackknife methods respectively. The evaluation results of the model using the ROC curve (AUC> 0.8) showed a very good performance. Error statistics including Accuracy, Precision, Bias, Probability of Detection and False Alarm Ratio showed 0.9, 0.79, 1, 0.93 and 0.04, respectively, which demonstrate the good performance of the model to prediction. In addition, the results of determining the importance of variables showed that the slope degree and following it, elevation and topographic wetness index are more important than other variables in determining the potential habitat of T. elliptica. The map obtained from the prediction of the potential habitat of T. elliptica can be very useful as accurate information in the rangeland management in order to Reclamation the destroyed habitats of this rangeland plant in Yazd province and be highly regarded by rangeland managers.

Soil aggregate stability and its spatial distribution can be considered as a good indicator for assessing the results of measures conducted for mitigation soil erosion. This study was conducted in two adjacent sites in Chahmari region, Kuzestan province. At one site afforestation and contour furrowing were conducted to control soil erosion and the adjacent site with no controlling measures was considered as control. A total of 150 soil samples were collected from the surface layer (0-5 cm) of two sites and mean weight diameter of aggregates (MWD) were measured using dry and wet sieving (MWDd and MWDw, respectively). Based on digital soil mapping (DSM) approach and to map MWD spatially, several environmental covariates were derived from a Landsat 8 image and a digital elevation model (DEM). Two machine learning algorithms including artificial neural networks (ANN) and regression trees (RT) were used to predict MWD with covariates as inputs. Results indicated a significant difference between MWDd in two sites, but no significant difference was found between MWDw. Correlation analysis revealed no correlation between MWDw and all terrain attributes derived from the DEM, but significant correlations were obtained between MWDd and some terrain attributes. Most covariates derived from Landsat images had significant correlation with both MWDw and MWDd. ANN and TR had relatively high and almost the same accuracy in predicting MWDw, but in predicting MWDd, ANN was superior to RT. In general, the findings showed good performance of DSM techniques in predicting and spatial mapping of MWD.

Gully erosion is one of the most important advanced (accelerated) forms of water erosion; so, identification and consideration of effective factors and its zoning is one of the important tools for controlling this phenomenon. The main purpose of the present study is spatial modeling and assessment of gully erosion using a probabilistic Weights-of-evidence (WofE) model and its different scenarios for selecting the location of gullies identified in Maharloo Watershed, Fars Province. For this aim, first, using extensive field observations, points of headcut, end, and also the boundary of the gullies (gully polygons) were identified and the gully inventory map for the study area was prepared. Then, thematic layers such as percent slope, slope aspect, plan curvature, elevation, Topographic Wetness Index, average annual precipitation, NDVI, land use, lithological units, distance from river, distance from road, drainage density, and some soil characteristics (percent silt, percent clay, EC and pH) were identified as effective factors on gully occurrence and their maps were prepared and classified in GIS environment. In the next step, using WofE algorithm, the relationship between effective factors and three different scenarios according to gully locations (head, end and gully polygons) were determined and the weight of each factor's classes was calculated. Ultimately, gully erosion zonation maps were prepared in the ArcGIS software environment. Evaluation of three gully erosion zoning maps using the ROC curve and 30% of unused gullies in the modeling process indicated that the accuracy of the models prepared based on three scenarios are 0.847, 0.861 and 0.792, respectively. Also, results of layer weighting using WofE model indicated that rainfall increasing, high drainage density, high silt percentage, low altitudes, southern directions, low distances from roads and river, Asmari, Aghajari, Razak, Gachsaran, Mishan and Sachun geological formations, as well as bare lands, have played an effective role in the occurrence of gully erosion in the study area. In order to control and prevent this kind of erosion in the Maharlo Watershed, the protective measures and watershed management were performed in the early stages.

Land-subsidence phenomenon is one of the geomorphologic hazards known in arid and semi-arid areas in recent years. The main objective of the present study is to provide a land-subsidence spatial modeling and its assessment using the generalized additive model data mining technique in Jiroft Plain, Kerman Province. Land-subsidence inventory map was prepared for the study area using extensive field surveys. Evidential Belief Function was used to investigate relationships between land-subsidence and 10 effective factors including slope percentage, aspect, elevation, lithological units, distance from river, piezometric wells data, land use, plan curvature, topographic wetness index, and distance from fault, and the weight of each factor class was determined. Results of the relationships between land-subsidence and effective factors indicated that most of the subsidences of the region are occurred in low slopes (0-2%), flat and smooth aspects, small distance from a river (>50m), large distance from a fault (>4000 m), also in the third group lithological units (rhythetic dyke and rhyodity and dark green tuff), shrubbery land use type, convex slopes with high topographic wetness index (>12), and receding in groundwater level. A land-subsidence susceptibility zonation map was created using the GAM in statistical R statistical software for the study area. Results of validation of generalized additive model evaluation using 30% of unused points in the modeling process, and the receiver operating characteristic (ROC) showed that the land-subsidence susceptibility map prepared by Generalized Additive Model had a high accuracy (81.20%). Therefore, the map might be helpful for optimizing management of water resources and preventing the re-occurrence of this phenomenon in the area. Results of ranking effective factors from the generalized additive model showed that elevation, land use, and aspect were in the highest orders in land-subsidence occurrences