فهرست مطالب حمیدرضا پورقاسمی

Given the scarcity of data on river basins nationwide, numerous researchers turn to spatial analysis within a Geographic Information System (GIS) setting for hydrological studies and flood investigations. On that basis, identifying the most important factors influencing flood occurrence and severity, as well as building their sensitivity maps can be one of the most important solutions for flood reduction. Therefore, the objective of this study is to prepare a flood risk map in Bushehr province using machine learning techniques and to identify important factors affecting flood hazards.

In this study conducted in Bushehr province, we aimed to compare the effectiveness of three machine learning models: Support Vector Machine (SVM), Random Forest (RF), and Generalized Additive Model (GAM). Initially, layers of information influencing flood occurrence in the study area were identified. Each prepared map served as input for the models. Various layers such as slope, slope direction, elevation, distance to river, drainage density, lithology, land use, topographic wetness index, and vegetation cover index were prepared using ArcGIS and SAGA-GIS software, crucial for analyzing flood patterns. Using data from 925 flood locations, points were divided into two sets: 70% (645 points) for modeling and 30% (280 points) for evaluation. The effectiveness of the models was validated using Receiver Operating Characteristic (ROC) analysis.

The results indicated that among the ten main factors, height, rainfall, and lithology were the most important factors affecting flood occurrence, while slope and distance from the river had the least impact. Evaluating model accuracy using ROC revealed very good accuracy for the SVM model (0.86), generalized additive model (0.85), and RF model (0.88). Flood sensitivity analysis showed RF and GAM methods identified the highest area in the low susceptibility class, while the SVM method identified the highest area in the medium susceptibility class. Results indicated that 37.32% of the study area had low sensitivity, 26.01% had medium sensitivity, 12.42% had high sensitivity, and 24.42% were very sensitive to flood hazards. Also, two other models have had very good accuracy for flood modeling in the studied area. The ROC related to the RF model, SVM, and generalized collective model showed an accuracy of 88.5 for the RF model, 86% accuracy for the SVM model, and 85% accuracy for the generalized collective model.

This study concludes that integrating machine learning models, namely SVM, RF, and GAM, with GIS analysis holds tremendous potential for advancing our understanding of flood patterns in Bushehr province. Leveraging these tools allows for a deeper comprehension of flood dynamics, aiding informed decision-making and effective mitigation strategies. This approach marks a significant leap forward in proactively addressing flood challenges and fostering resilient flood management practices in Bushehr province.

The aim of this study is to identify the areas prone to gully erosion and landslide as a two-hazard map in the form of a single map in Gorganrood watershed. In addition, the use of two machine learning models such as RF and SVM to establishing the spatial relationship between these hazards and the GEFs and any hazard susceptibility mapping separately. In addition, the validation of the hazards susceptibility maps was conducted based on the ROC curve method, and the best model was chosen with the highest predictive performances. Finally, by combining the susceptibility landslides and gully maps, the two-hazard probability maps were produced, which were a combination of different models and the best model. The results showed a RF model with (AUC = 82.9) for landslide and (AUC = 96.9) for gully erosion have higher accuracy compared to the SVM model with a value of (AUC = 0.76) for landslide and (AUC = 93.9) for gully erosion. Finally, a single and comprehensive map was obtained by combining of the each hazard susceptibility map for identification the areas prone to both hazards based on both models. The final two-hazard map can be used as a valuable tool for sustainable land use planning in multi-hazard prone areas.

Piping erosion, gully erosion, and mass movement have complexity that the best way to comprehensively identify them is to use new statistical tools. The purpose of this study is spatial analysis of soil landforms based on spatial statistics and field studies. In this research, a digital elevation model was prepared. Then, the spatial pattern and interaction of the three soil landforms were investigated and the spatial effect of slope degree was analyzed on the spatial distributions of them. The results showed the pair correlation function (g(r)) was clustered for both piping erosion and gully erosion. However, it was dispersed for mass movements up to spatial distance of 140 m. The bivariate statistic (g12(r)) showed there is a significantly positive correlation between gully erosion and piping erosion. The correlation between mass movements with each of gullies and piping erosions were positive at the near distances. In addition, increasing the slope had a large effect on the spatial distribution of mass movements and a small effect on the spatial distribution of piping erosion and gully erosion. Therefore, a comprehensive approach to a landscape using quantitative statistics can lead to better identification of processes affecting them, and then, reduction of natural hazards.

Fire is one of the destructive phenomena that have devastated a significant portion of forests and grasslands in Gilan Province in recent years. This study aimed to investigate the temporal and spatial relationship between climatic variables and wildfires in Gilan Province. The wildfire variables included the number and extent of wildfires, and the climatic variables consisted of seven parameters over the past 26 years (2001-2026). Pearson correlation and regression analysis were utilized to examine the temporal relationship. The relative importance of climatic variables in wildfire occurrence was determined using Mean Decrease Gini (MDG) and Mean Decrease Accuracy (MDA) statistics. For modeling and generating probability maps of wildfire occurrence, 70% of wildfire locations and various machine learning models (Logistic Regression, Random Forest, Support Vector Machine, and SVM-RF Hybrid) were employed using the R programming language. Model validation was conducted using 30% of wildfire locations and the Area Under the Curve (AUC) metric. The temporal results showed that during the 26-year period (2001-2026), a significant negative correlation was observed between the number of wildfires and the average seasonal precipitation, while positive correlations were found between the number of wildfires and the average seasonal wind speed and maximum wind speed at a 95% confidence level. Furthermore, a significant negative correlation was observed between the extent of wildfires and the average seasonal precipitation at a 95% confidence level. The spatial relationship analysis indicated that the average maximum temperature, average seasonal precipitation, and average relative humidity had the highest importance in wildfire occurrence within the geographical extent of Gilan Province. Model validation results revealed that the Random Forest model (AUC: 0.82) and the SVM-RF Hybrid model (AUC: 0.79) outperformed others in predicting the occurrence of wildfires. Therefore, predicting wildfires resulting from climatic factors in the forests and grasslands of Gilan Province using the aforementioned maps is feasible and can significantly aid natural resource managers in implementing protective measures in high-risk wildfire areas. Hence, it is imperative that proactive measures be taken by the Natural Resources and Watershed Management Organization of the province to prevent future wildfires with greater sensitivity.

Mass movements are a type of morphodynamical phenomena that are usu ally related to various factors and occur on slopes in mountainous areas. Every year, damages caused by landslides lead to financial loss and death over the world (Fathi et al., 2018). According to previous studies, landslides cause 17% of the world's natural disasters. Mortality rates from 1903 to 2004 for different continents, including Asia, United States, Europe, Africa, and Australia were 29%, 39%, 30%, 1%, and 1%, respectively (Kohorest et al., 2005). Over the last three decades, several research studies have been conducted on landslide susceptibility mapping using different methods for developing their classification. All mapping methods are classified into five different groups including landslide distribution analysis, qualitative, statistical, deterministic, and frequency analyses (Vanvestern, 2003). Iran is prone to landslide phenomena due to natural conditions such as mountain topography, high tectonic and seismicity activity, geological and climatic diversity. Environmental factors affecting the occurrence of landslides are slope degree, aspect, plan curvature, elevation, land use, lithology, distance from the road, river and fault, topographic moisture index, slope length index or sediment transport and vegetation cover (Silakhori et al., 1400).

Firstly, 134 points were identified using the Iranian landslide database and field survey in order to prepare a landslide susceptibility map using Bayesian theory. In the present study, 12 factors were used including elevation, slope and aspect, plan curvature, distance from the fault, road, and river, land use, geology, sediment transport index (STI), topographic wetness index (TWI), and vegetation cover. Then, topographic (1:50000), geology (1:100000), soil maps, and satellite imagery (Indian Remote Sensing) for 2012, were prepared and classified in ArcMap and ENVI environments. In order to evaluate Bayesian theory in landslide risk analysis, the relative performance curve of the relative efficiency of variables (ROC) was applied. This index is used to determine the accuracy and efficiency of the model (Egan, 1975; Williams et al., 1999). The area under the ROC curve represents the predicted value by describing its ability which accurately estimates the occurred events (landslide occurrence) and its non-occurred events (non-landslide occurrence). Therefore, the area under the curve is used as the model accuracy assessment. In the present study, 134 points of the landslide’s phenomena were used for modeling (70%) and accuracy assessment (30%) (Pourghasemi et al., 2013).

The results of the factors affecting the occurrence of landslides using Bayesian theory in the study area showed that most of the landslides occurred in the class of 15-30 degrees with a weight of 1.28, which is also reported by Eracanoglu and Gokceolu (2004). The reason is that human intervention on these slopes causes more susceptibility (Yalcin et al., 2011). The study of aspect shows that most of the landslides occurred in the west and south directions with a weight of 2.21 and 2.41, respectively, which is confirmed by the results of Shams and Alizadeh (2019). The results of plan curvature showed that most of the landslides happened with a weight of 1.1 in convex slopes, which is close to the results of Pourghasemi (2013). Convex slopes usually have the highest landslides which is also reported in previous studies (Vanvewsten et al., 2003; Jaaferi et al., 2014). The altitude of 500-1000m covered by sandstones, chile and siltstones showed a significant relationship with a high number of landslides (Ayalew & Yamagishi, 2005). Among different land uses, rangeland (weight of 3.48) indicated the most significant relationship in landslide occurrence, which is reported by Shams and Alizadeh (2019).  
The results of distance from roads, rivers, and faults showed that most landslides occurred at distances of more than 0-100 meters which confirms the results of previous studies (Pourghasemi & Mohammadi 2016). Also, the relationship between vegetation cover and landslides showed that the highest percentage occurred in class 0.5-0.3 with a weight of 3.56. In addition, the highest slippage for topographic wetness index and sediment transport index (river capacity) occurring were related to 6.39- 11.29 and 11.84-19.6 classes (weights of 1.009 and 1.17), respectively. The area under the curve was calculated at 85.56% in the model validation for the Bayesian model, which was classified as a very good performance. Therefore, the results of our study can play an important role in the management and planning of the Talar watershed.

Iran is located in an arid and semi-arid climate with an average rainfall of about 239.3 mm during five years ago. There are many challenges for decision-makers in terms of supplying water for agriculture, industry, and drinking. On the other hand, lack of comprehensive decision-makers, along with lack of awareness of beneficiaries of the consequences of excessive use of groundwater resources, has led to land subsidence in most plains of the country. This issue, due to the nature of land cognition and its relation to the use of groundwater, has been considered by researchers in the fields of water resources management and geology. Land subsidence of Iran has been studied worldwide and the effects of groundwater abstraction have been determined at a high level compared to other factors. Hamedan province in the western part of Iran is not far from overuse of water and the effects of land subsidence, due to the extensive agricultural lands and agricultural hubs. In this regard, studying Kaboudrahang aquifer subsidence with its sinkholes (18 massive sinkholes) was selected as the study site in the northwest of Hamadan province. The hydrograph of the plain aquifer unit shows a profound drop in the groundwater level of the aquifer (40 meters) between the years 1988 and 2018.

Regarding plain subsidence, some researches have been done. Accordingly, the radar interference technique in ENVI software was used to determine the amount of land subsidence and finally its output expressed in GIS software. To use this method, 11 Sentinel satellite images from 2014 to 2019 were selected and groundwater changes in 49 observation wells were studied. To investigate land subsidence and groundwater changes, the relation between land subsidence rate and groundwater level drop was examined through a regression analysis, as well as several observation wells that have a reduction in the date of the last satellite image relative to the water level. A regression analysis was selected between the water level drop and land subsidence rate and between the water level drop and the cumulative land subsidence obtained by radar interferometry during the period. To investigate in more detail, the effect of groundwater abstraction in each observation well and the relationship between land subsidence and its changes over time have also been investigated. On the other hand, land subsidence is delayed according to some research on groundwater exploitation, which varies due to the different hydraulic conductivity of the aquifer along with the different geological structures of the plain. Lag regression was used as a simple tool to determine the lag time between groundwater exploitation and land subsidence.

The study results showed that during the years 2014 to 2019, the land subsidence of the plain was 124.7 mm and an average rate was 13 mm every six months. Positive and negative values were observed in the land subsidence rate map in which positive values indicate upward radar visibility and negative values are related to radar visibility or subsidence. In general, across the plain, there is no equal correlation between subsidence rate and groundwater level decline. Its main reason can be found in the difference of geological structures and diverse soil texture throughout the plain, which has led to different responses to groundwater exploitation. Wells that had a drop in water level between the first and last satellite image show a decreasing trend of land subsidence rate and cumulative land subsidence. This indicates a decrease in the power of land subsidence in the utilization of water over time. The reason is that groundwater passed through subsidence-sensitive layers, after which the effect on land subsidence is reduced. The lag time of subsidence and operation of groundwater was determined as an average of 1.5 years.

Based on the analysis of this research, more significant impact of land subsidence on agricultural lands can be identified due to the use of groundwater. On the other hand, there is need for proposals and solutions to reduce the rate of land subsidence, and so it is recommended to change the cultivation pattern, to control surface water with artificial recharge, and to create deep plowing in agricultural areas to increase permeability.

Frequency analysis of daily rainfall or return period of rainfall and flooding events is very important considering the behavioral complexity in water resources management; because ignoring it can lead to urban destructive floods. In the present research, three distribution functions of Pearson, Beta, and Gamma were compared to investigate and select the most appropriate distribution function for the precipitation data acquired from meteorology stations and CHIRPS satellite in seven stations in the watershed of Bustan Dam. Statistical analyses showed that satellite data were ineffective to estimate daily precipitation due to high errors in RMSE, MAD, and NASH. Meteorological data were used to spot the best distribution. Google Earth Engine and Python programming language were used. Then, the selected distribution function was used to determine the maximum daily rainfall, frequency probability, and return period of 2, 10, 50, 100, and 200 years. The results of the goodness of fit test, Error Sum of Squares, Bayesian Information Criterion, Akaike Information Criteria well as Kullback-Leibler Divergence showed that in five stations of Kalaleh, Qarnaq, Golestan National Park, Golestan Dam, and Glidagh, the Pearson function is the most suitable distribution function. Also, in the other two stations (Gonbad and Tamar), the Beta function was recognized as a suitable function. However, Gamma distribution in the study area is not efficient. So, it can be concluded that heavy and irregular rainfall can be effective in choosing the best distribution function at each station. Therefore, it is recommended to consider the maximum possible rainfall and as a result of the possible occurrence of floods with principled and accurate management to prevent human and financial losses in susceptible areas, especially in the study area.

Watershed sediment yield estimation is an important process in watershed management. Because of limited information on sediment yield in small watersheds, empirical methods are developed in different watersheds. Sediment connectivity is an important characteristic of watershed sediment yield and is compared for better estimation of sediment delivery ratio (SDR). This research aims to compare and evaluate the spatial changes of SDR values and connectivity index on a small watershed scale. This study was conducted in the Marcheshme watershed of Semnan province with an area of 2418 ha. At the outlet of this watershed, an earthen dam was built by the natural resources and watershed management organization of the province (1995). No conservation and management measures were carried out in the watershed. In 2013, the amount of sediment trapped behind the dam was measured by the research center of Semnan province. It should be noted that there isn’t any overflow of the earth dam between 1995 and 2013. The WaTEM/SEDEM model was calibrated to calculate the SDR using the measured amount of sediment, and the SDR values for the entire watershed were distributed. A distribution map of the sediment connectivity index was prepared. The highest value of the SDR map was close to 0.5, the lowest value was close to zero, and the sediment connectivity index varied from -6.71 to 3.04. SDR and IC indices show that the amount of sediment transport and delivery near the outlet is higher than the average value in the watershed area. In the upper part of the watershed, erosion is more than the one in other parts, although SDR and the possibility of sediment connection are less than the ones in other parts. These two indicators and distribution of them can be important for prioritizing areas in terms of soil erosion and sediment management. This study shows the up-to-date application of available data and develops their potential for further analysis.

Assessing habitat changes is an important step in the spatial prioritization of management efforts, aimed at reducing conflicts. We assessed landuse/cover change in areas with high risk of human-bear conflict in Fars province. In the first step, we predicted the conflict hotspots, using bear damage incidents, a suit of predictors, and the Ensembles of Small Models (ESMs) approach. In the second step, we assessed the trend of landuse/cove changes in a 30-years period in the areas with medium to high risk of conflict, using remote sensing techniques. Results of conflict risk modeling showed that proportion of suitable habitats, distance to village, density of forest patches, and corridor bottlenecks were the main predictors, contributing to bear damaging risk. A total of 3.75 and 6.91% of the landscape were identified as the areas with high and medium risk, respectively. Assessment of landuse/cove changes showed that in a period of 30 years, the extent of croplands and orchards has increased from 12,167 to 52,662 hectares. Such a substanitial landuse/cover changes can increase the risk of bear damages. The obtained results emphasize that inter-organizational planning is an emergency effort in mitigating human-bear conflicts.

Population growth, urbanization, and land use change have increased disastrous floods. Iran is also among the countries at high risk of floods. The latest examples of flood damage are the devastating floods of the spring of 2019 with significant mortality and financial losses in more than ten provinces of the country. The purpose of this study is to prepare an urban flood risk map of District 4 City Shiraz. The vulnerability of the region was made using PROMETHEE Ⅱ and COPRAS multi-criteria decision-making models and urban flood hazard zones were prepared by partial least squares regression (PLSR) and ridge regression (RR) models and a risk map was obtained by multiplying the vulnerability and hazard in ArcGIS software. The highest percentage of the study area in the PROMETHEE Ⅱ and COPRAS models belongs to the moderate class of vulnerability. The evaluation of the vulnerability models using Boolean logic and RMSE and MAPE statistics, showed that the COPRAS model provided better results than the PROMETHEE model. The results of partial least square regression (PLSR) and ridge regression (RR) models in flood risk modeling were analyzed by the Taylor diagram, which showed the superiority of the ridge regression (RR) model and the accuracy of this model in preparing urban flood hazard maps. The risk map of the study area indicated that 34% of the area (973 ha) is in the range of high and very high flood risk.

Slope instability and landslides are important hazards to human activities that often result in the loss of economic resources, property damage and facilities. These hazards occur in the natural or man-made slopes. In the current study, the maximum entropy model was used which is one of the progressive data mining models, in order to modelling landslide susceptibility map for Gorganrood watershed. In the first step, the landslide inventory map was prepared consiste of 351 landslides. 18 geo-environmental factors were selected as predictors, such as: Digital elevation model, slope percent, aspect, distance from fault, distance from river, distance from road, rainfall, landuse, drainage density, lithology, soil texture, plan curvature, profil curvature, lithological formation, Topographic wetness index, LS factor, stream power index, Relative Slope Position and Surface roughness index. Three different sample data sets (S1, S2, and S3) including 70% for training and 30% for validation were randomly prepared to evaluate the robustness of the model. The accuracy of the predictive model was evaluated by drawing receiver operating characteristic (ROC) curves and by calculating the area under the ROC curve (AUC). The ME model performed excellently both in the degree of fitting and in predictive performance (AUC values well above 0.8), which resulted in accurate predictions. Furthermore, In this study the importance of variables was evaluated by the model. Dem (digital elevation model) (32.4% importance), lithology (22.9% importance) and distance from fault (14.8% importance) were identified respectively the main controlling factor among all other variables.

One of the most critical environmental problems globally is soil erosion, which leads to several problems in the field of economy and sustainable development. This study aims to use new methods in soil science studies of the paired sub-catchments to update service descriptions in Dehgin, Hormozgan Province. For this purpose, to select sampling sites of pedons and soil samples, the conditional Latin Hypercube sampling (cLHS) method was used using environmental covariates (digital elevation model), which derivate from unmanned aerial vehicles (UAV), phantom4 pro, and multirotor. The results showed two main land types in the study area, including hills, plateaus, and high terraces. Soil development in the region is mainly influenced by elevation, and the soils of the hill lands are directly related to the type of parent material and slope percentage. In the areas containing wooded pasture and areas with silt loam texture class, the aggregate stability values are higher, which are auxiliary covariates essential and influential parameters in the aggregate stability map. The RMSE and R2 values are 0.32 and 0.26, respectively, for the evaluation map criteria. The values of cation exchange capacity (CEC) and nitrogen in both control and sample sub-catchments are small and the maximum amounts of soil organic carbon in the sample sub-catchment are higher than in the control sub-catchment. The use of new methods including UAV, accurate sampling methods using high-resolution environmental covariates, and digital soil mapping can reduce the number of soil samples, increase the accuracy of output maps, and provide more accurate erosion estimation results, which has reduced time and cost compared to the old methods.

Land use change is an essential determinant factor in hydrological processes and it affects many of them. Simulation of land use changes, rise in residential and agricultural area and its effects are among the most important issues related to the watershed management. The objective of this research is land use change detection in the past and simulation of future land use using CLUE-s Model. At first, land use map of years 1991 and 2019 was prepared using satellite images and synthetic method. Then, land use change was calculated, and then land use map of year 2040 was simulated using CLUE-s model. The results indicated that the main land use changes in Doroodzan watershed between 1991 and 2019 was the conversion of natural resources to residential area, agriculture and orchards. Accuracy assessment of regression model to investigate condition factors of each land use types using ROC corves demonstrated acceptable accuracy of this method. Using the CLUE-s model, managers and planners can be informed about the future conditions of the region to take into account the problem of land use change in their planning.

Various reasons of surface water resources scarcity, and consequently, the excessive removal of groundwater resources has posed a serious threat to these resources. Reduction of groundwater surface, land subsidence and Sinkholes, are the harmful effects of large-scale exploitation and the lower recharge of groundwater resources which become a social challenge. Therefore, the sustainable management of these water resources is a vital necessity. Storage and recharge of groundwater aquifers is a potential solution to achieve future water supply goals. In this paper, the potential of infiltration areas to recharge groundwater was investigated by using the maximum entropy model (MaxENT). The Jackknife test was used to determine the most effective factors on ground water supply. Based on Jackknife graph, soil texture and lithology factors were the most important factors influencing the prediction of areas with potential infiltration. The map of the maximum entropy model was classified to four classes with low, medium, high and very high permeability. The greatest potential was observed in the middle part of the watershed, which is the plains. Areas with low slope gradient, sandy soil texture with quaternary sediments showed the highest potential. Study area included 0.02 % very high recharge potential, 3.1 %, 0.7 % and 96.18% with high, moderate and low recharge potential, respectively. The precision of prediction of the final map of the model results was evaluated using some indices with good to excellent grade. According to the results, the plain areas of the Marboreh watershed are suitable to groundwater recharge.

The mechanism of land subsidence and soil deformation is the dissipation of the excess pore water pressure and the compaction of the soil skeleton under the effect of natural or man-made factors. Land subsidence has occurred in more of 300 plains of Iran, and Semnan Plain is one of the most important areas that face to this phenomenon. The aim of current research was land subsidence susceptibility mapping using GLM and FDA models. At first acquiring information about the land subsidence occurred in the past was done. Locations of the land subsidence were specified using Landsat 8 satellite images and field surveys. The receiver operating characteristic (ROC) curve and area under the curve (AUC) were used to assess accuracy of the models. The AUC for the GLM and FDA models was 0.657 and 0.618 respectively. Results of accuracy assessment show that accuracy of these models is located into the moderate class for land subsidence susceptibility mapping in the Semnan Plain. The research results can provide a basis for the protection of natural resources and environment and also promote the sustainable development of economy and society.

The aim of this study was to analyze the effectiveness of watershed management activities in a representative  sub-catchment (treated) and its comparison with a control on, and to identify the erosion sensitive areas in these sub-catchments areas of ​​58 hectares (control) and 83 hectares (treated) the using UAV photogrammetry. Based on a curve number and four other basic maps (elevation, slope, drainage direction, and accumulation density), three daily precipitation scenarios of 20, 40, and 60 mm were considered to model the LANDPLANER with two indices of Erosion Index and Topographic Threshold. The results were evaluated using the boxplot methods, foufold plot, and the area under the receiver operating characteristic curve. The Erosion index (IE) in the LANDPLANER model, considering the maximum daily precipitation of 60 mm, was 0.54821 for the control sub-catchment which was more than that of the treated one (0.15593). The modeling based on the topographic Threshold also showed that the numerical changes of the index between 0-2 in the trated sub-catchment were less than that of the control area(0-3). Therefore, it can be said that due to the appropriate protective operations that had taken place on the treated sub-catchment, this area is in a better condition than the control one, as the volume of soil erosion was less than that of the control sub-catchment. It is informative to know that all of the three precipitation scenarios predict less soil erosion in the treated sub-catchment. Therefore, simulation based on the rainfall and curve number scenarios with LANDPLANER physical model is recommended to evalute the effect of watershed management activities on the paired sub- catchments that are degraded due to soil erosion.

Landslides are natural hazards that destroy nature and often cause severe damage to natural resources and loss of life and human property (Intaravichin and Dasananda, 2010; Feyzzadeh et al., 2013). Despite significant advances in risk analysis and landslide risk management, this natural hazard remains a real threat to the lives and property of people in mountainous areas (Getz et al., 2011). Natural factors affecting landslides can be rainfall, lithology, slope, and the like (Talebi and Niazi, 1390). Among the natural factors affecting the occurrence of landslides, we can mention factors such as rainfall, lithology, slope, and the like (Talebi and Niazi, 1390). In general, Iran is considered a high-risk country due to favorable geographical conditions, lack of comprehensive environmental management, and non-compliance with environmental thresholds. Approximately 38 hazards have been identified and recorded in Iran, and due to the multiplicity, diversity, recurrence, and severity of natural hazards and environmental unrest are among the ten most prosperous countries in the world (Mohammadi et al., 2004). Considering landslide hazard zoning map is essential for a proper understanding of its mechanisms in recent research (Sun, 2014).

In order to prepare a landslide susceptibility map using Dempster-Schiffer theory, first, using the landslide database, the General Department of Natural Resources of Mazandaran province, and field studies, landslide points in the region were identified. During the field observations, the accuracy of the collected information was evaluated, and the position of the landslide points was checked using the Global Positioning System (GPS), and 26 new landslides were recorded. A total of 134 slide points were examined. In this research, 12 information layers including altitude map, slope direction, slope degree, slope shape, distance from fault, distance from road, land use, geology, distance from waterway, Slope length or sediment transport index (STI), topographic moisture index (TWI) and vegetation were prepared and classified using ArcMap and ENVI software using 1: 50,000 topographic maps, 1: 100,000 geology, geology and satellite imagery, IRS for 2012. In order to evaluate the Dempster-Schaefer theory in landslide risk analysis, the relative performance curve of the relative efficiency of variables (ROC) is used. ROC curve analysis is used to determine the accuracy and efficiency of the model (Egan, 1975; Soots, 1988; Williams et al., 1999). This curve is one of the most efficient methods in providing diagnostic properties, probability identification, and prediction systems that quantitatively estimate the modelchr('39')s accuracy (Soots, 1988). The area under the ROC curve represents the prediction value of the system by describing its ability to accurately estimate its occurrence events (landslide occurrence) and its non-occurrence events (non-landslide occurrence). Therefore, the area of ​​the curve is used as a measure of the accuracy of the model. In this study, out of 134 slip points, several points were used for modeling, and part of it was used to evaluate the model (with a ratio of about 70 to 30%) (Matthew et al., 2007; Pourghasemi et al., 2009; Constantine et al., 2010).

The study of the slope factor shows that although the highest percentage of landslides is related to the class of 15-30 degrees with 52.127 percent, which is consistent with the views (Moradi et al., 2010). However, the highest weight is related to the class of 5-15 degrees, which can be attributed to road construction and unprincipled constructions that have caused the slope to be cut in the region. No landslides have been observed or recorded in areas with a slope above 50 degrees, which can be due to the rocky outcrop and the lack of soil layer in the high slopes. According to the results of this study, the western and southern slopes with 19.15% and 18.1%, respectively, had the highest percentage of landslides in the region, which with the results of studies (Pour Ghasemi et al., 2009) and (Lee, 2007) corresponds. The study of altitude factors indicates that the altitude class of 1000-500 with 56.38% of landslides, the highest number of landslides, and the altitude class of 0-500 have the highest weight, also, with increasing altitude in the study area, the sensitivity to landslides decreases, which is consistent with the studies of Moradi et al. (2010) and Pourghasemi et al. (2011). At altitudes above 2000 m, no landslides have been observed or recorded, which indicates a decrease in sensitivity to landslides with increasing altitude (Moradi et al., 2010). The highest weight for the geological factor belongs to group 1, which includes deposits and unhardened alluvial deposits of the Quaternary period, while the highest percentage of landslides occurred in the units. The lithology of the Shemshak Formation is located at 37.23%, which is consistent with the views (Naji, 2006), (Shadfar et al., 2008), (Pourghasemi et al., 2011). The study of the linear element of the distance from the waterway has shown that because the highest percentage of landslides occurred at a distance of more than 400 meters at the rate of 31.915%. However, the highest weight in calculations by class is allocated 0-100 meters, which is in accordance with the opinions (Pourghasemi et al., 2009), (Dokota et al., 2013). Investigation of the distance from the road showed that the highest percentage of landslides occurred at a distance of more than 500 meters (30.85%); however, the highest weight in the calculations to the class 0-100 for the distance from the road is allocated which is consistent with the results of studies of Mazouki et al. (2011), Pourghasemi et al. (2009, 2012).

Land subsidence is a geo hazard that leads to slow or rapid decrease of ground level. Land subsidence presents a constant threat to the safety of surface infrastructures such as motorways, railways, power lines, and telecommunication cables. Arid and semi-arid countries like Iran are very susceptible to land subsidence phenomenon. Land subsidence has occurred in more of 300 plains of Iran, and Semnan Plain is one of the most important areas that face to this phenomenon. The purpose of this research was land subsidence susceptibility assessment using entropy model. In the first step, acquiring information about sixty-five land subsidence in the past was done. Twelve conditioning factors from different basic layers including topography, geology, land use, and ground water table were considered for modeling aims. The receiver operating characteristic (ROC) curve and area under the curve (AUC) were used to assess accuracy of the Entropy model. The AUC for the entropy model was 0.715 that showed this model is acceptable for land subsidence susceptibility mapping in the Semnan Plain. Results of this research can help to policy makers and planners for environment and urban management and planning.

Creating a gully is a reaction to the geomorphologic conditions, this type of erosion extends over a wide range of environments, and the threshold of topography is related to the slope and drainage level and Controls the position and expansion of gullies in different uses. The slop-area equation as a relationship between the slope and the upstream area of the watershed in any area will in part help predict gully erosion. The above relationship has been influenced by several factors including environmental conditions (climate, lithology, soil type), the type of land use, and the type of effective mechanisms for the creation and expansion of the gully (surface flow, sub-surface flow, dissolation, and piping), and the type of method used for extraction of slope - area is located. And this is the simplest formula used by two parameters to predict gully erosion. The results show that the value of the relationship between -0.233 and -0.205 was obtained. The equation for agricultural use was equal to Y = 5.7426X-0.205and for the pasture use equal to Y = 10.653X-0.233. In this equation, the power can be close but the coefficient of the equation is different. Gully erosion of the whole hinterland (farmland, agriculture) is expanding. The threshold of topography indicates a negative relationship between land area and slop in different kind of agricultural and rangeland. By reducing the slope, more area is needed for the development and expansion of the gully, and most of the gully is due to surface runoff.

Statistical modeling methods are based on multivariate regression methods and require the presence and absence location of data for the construction of the model. In most cases, there is no trustworthy absence data. Therefore, other methods that are based only on the presence of the phenomenon are used. Considering the importance of modeling - saving time and cost and the probable prediction of the process - in this paper three sampling methods, Bootstrap, cross-validation (CV) and subsampling, were investigated to estimate areas with groundwater recharge potential using the maximum entropy model in the Marboreh watershed. The information about percolation points in Marboreh watershed, which was gathered using the double ring method and soil sampling, included the location of the samples, soil texture, and percolation rate. Due to the extent of the catchment area and the cost of the sampling process, information from previous studies in the study area, which were gathered from the Regional Water Authority (RWALP) and Agricultural Research, Education and Promotion Organization of Lorestan province (AREPOLP), was also used. The ROC index was used validate model predictions. The validation index indicated that the bootstrap had the best performance (ROC=0.955%). The results showed that each factors in these three methods was somewhat different, which was more than other factors in the drainage density, land use and soil texture. Based on the results of performance index, there is a very slight difference between the three sampling methods, so that they can be differentiated in relation to their different strategies, and this difference in the outputs, is not related to the diversity of the phenomenon studied. In this paper, according to the results and assessments, the Bootstrap method is recommended for the modeling the groundwater recharge areas due to the small number of sampling data compared to the very large area of study. Due to the large extent of the study area, it is suggested that this simulation be performed for more precision at smaller extent large areas with further data to study similar studies. Despite the increase in the number of pixels of high infiltration areas in the Bootstrap sampling method, compared to the other two methods, the performance of the recharge zoning increased slightly.

The present study compares the performance of three statistical models including logistic regression, frequency ratio and maximum entropy in preparing the zoning map of landslide sensitivity on the watersheds of 24 and 25 (Malak-roud and Shen-roud) regions of Guilan province. For this purpose, all the roads of the two areas were survyed and the number of 844 landslides were detected. Then effective data layers on landslide formation were prepared and then zoning maps of the landslide sensitivity were extracted to the three methods. The results showed that the logistic regression model has an accuracy of 74.6 percent of the Area under ROC Curve in prediction of sensitive areas compared to the accuracy of 68.2 and 65.2 percent by the probabilistic frequency ratio model and Entropy model, respectively. Moreover, based on the logistic regression model, the distance and slope factors have the highest effect on the landslide of forest road trenches. According to this study, a total of 4.32 square kilometers (43.25%) of the landslide areas is categorized in very high and high susceptible groups which have occurred along 51.38 percent of the area. Despite the lack of consideration to superstructure conditions in the present study, a notable percentage of existing roads are exposed to the collapse of trenches and require adequate funding for repair operations. According to the logging ban as a new approach in forest utilization and lack of revenue regeneration of alternative conservation projects great dangers threaten the forest engineering installations and road structures.

Piping erosion is an influential landform of intensified water erosion that interact with each other in the loess-derived soilsof the Province of Golestan (NE Iran). This study attempts to emphasize the use of high-precision data through field surveys and UAV images, and then to establish a threshold to study the factors that affect piping erosion. The present study was carried out using photogrammetric drones in an area of ​​approximately 2700 hectares in the loess lands in the east of the Province of Golestan and identified areas affected by piping erosion. A total of 833 pipes were recorded using the GPS and the orthophoto images from the UAV. Topographic, hydrological, and biological factors were treated as independent variables and piping erosion as dependent variables through statistical analyzes. The topographic factors indicated that the maximum density of piping occurred at altitudes of between 300–350 m, at maximum slope of more than 30%, at slope length less than 5m, and in concave tracts and profile curvatures. Regarding the hydrological factors, the maximum piping density occurred at the maximum numerical value of the topographic wetness index (more than 12), and at the closest distance to waterways (less than 100m). Biological factors indicated that the maximum density of pipes is in the rangelands and where this erosional facies is closest to the roads. Therefore, it can be suggested that an accurate identification of the piping erosion is facilated through ground observations aided by high-precision areal photography and understanding causative factors developing them.

Landslide susceptibility maps are considered a backbone for decision-makers to suggest solitary or combined technical and regulatory measures. Such maps are also considered an invaluable tool for engineers, earth scientists, planners, and decision-makers to select the most suitable areas for agriculture, building, and other development activities. Hence, thanks to landslide susceptibility maps, addressing highly susceptible areas are feasible, so that over the course of further detailed studies on the imminent landslide occurrences in the future, landslide potential risk is mitigated.

In this study, two robust data mining models, namely random forest and maximum entropy were used to map landslide susceptibility across the Owghan Watershed in Golestan province. After preparing the landslide inventory map via extensive field surveys, interpreting Google Earth images, and the archived data acquired from different organizations, landslide points were split into two sets of training (70%) and validation (30%) by using the Mahalanobis distance technique. Further, drawing on the extensive literature review, fifteen factors including climatic, geological, tectonic, topo-hydrological, and anthropogenic drivers, as landslide-controlling factors were selected and sieved through the variance inflation factor test. Ultimately, by implementing the above-mentioned data mining techniques, the most important factors in the modeling process, as well as the highly susceptible locations in the study area, were introduced.

Evaluating the learning capability, both the random forest and maximum entropy models with the respective area under the receiver operating characteristic curve (AUROC) values of 0.923 and 0.91, showed almost identical fitting abilities. However, getting to the validation stage, it was found that the random forest with the AUROC value of 0.9 clearly outperforms maximum entropy (AUROC= 0.85) in terms of prediction power and generalization capacity. Hence, the random forest was suggested as a better-performing model for landslide susceptibility mapping in the Owghan watershed, compared to its counterpart. About 10% of the study area falls into high and very high landslide susceptibility zones. Furthermore, five landslide-controlling factors including rainfall, normalized difference vegetation index, height above the nearest drainage, lithological formation, and proximity to roads have been found to be the most significant factors contributing to landslide occurrence in the study area. Additionally, the results attest that announcing the Safiabad village as a landslide-prone area by the authorities is technically sound and evacuating the residents to a new place has been a right decision; however, some parts of the newly inhabited area shows landslide predisposing patterns which can lead to a higher susceptibility of the area to landslide occurrence in the future.

Scrutinizing the results of random forest model revealed that a combination of natural factors (intense rainfall, bare lands, susceptible lithological formations, and topo-hydrological mechanisms) and anthropogenic interferences (tillage parallel to slope length/perpendicular to contour lines and unprincipled road construction) are synergistically responsible for landslide occurrence in the Owghan Watershed. On the other hand, announcing the Safiabad village as a critical landslide-prone area seems to be a wise decision, although the newly inhabited place seems to be selected merely based on having a suitable slope steepness (i.e., almost flat) and being accessible through several connecting routes, while the enhanced conservation tillage methods have not been applied to the selected site and adjacent areas. The latter, according to our inferences, can trigger a crisis in a larger extent. Moreover, owing to the presence of other landslide predisposing factors in the new residential site, safe areas should be pointed out and announced by adopting a holistic view on the entire influential and predisposing conditions for landslide occurrence.