Alireza Yousefi Kebriya; Ali Khalili; Hasan Rezaei
Abstract
Introduction
Dust storms have emerged as one of the most significant environmental challenges in arid and semi-arid regions, and their frequency and intensity have notably increased in Ilam Province in recent years. These storms have had wide-ranging impacts on public health, urban infrastructure, agriculture, ...
Read More
Introduction
Dust storms have emerged as one of the most significant environmental challenges in arid and semi-arid regions, and their frequency and intensity have notably increased in Ilam Province in recent years. These storms have had wide-ranging impacts on public health, urban infrastructure, agriculture, and the sustainability of natural resources. The province’s geographical location along the borders of Iraq and Syria makes it particularly vulnerable to transboundary dust storms originating from desertified areas in neighboring countries. Accordingly, precise monitoring of the spatiotemporal dynamics of dust storms and identifying their sources are essential for developing effective mitigation strategies and reducing their adverse impacts.
Materials and methods
In this study, PM concentration data from 2020 to 2025 were collected from air quality monitoring stations in Mehran and Dehloran as ground-based observations. In parallel, satellite-based indices were utilized, including the Aerosol Optical Depth (AOD) from MODIS, the Absorbing Aerosol Index (AAI) from Sentinel-5P TROPOMI, the Normalized Difference Dust Index (NDDI), and the Dust Event Count Map (DECM). All datasets were processed and analyzed using Google Earth Engine. To track the transport pathways of dust plumes, the HYSPLIT model was applied with a 24-hour backward trajectory simulation. Additionally, MODIS True Color images were employed to visually validate the HYSPLIT model outputs.
Results and discussion
Analysis of the DECM index from 2020 to 2024 revealed an upward trend in the frequency of dust events in Ilam Province. In 2020, the lowest number of events was recorded, although even in that year, Dehloran and Abdanan experienced over 30 events. In 2021, the number rose to over 120 events in border regions, reaching a critical peak in 2022 with more than 200 dust events recorded in Mehran, Dehloran, Eyvan, and southern Ilam. Although the numbers slightly decreased to 182 and 172 in 2023 and 2024, respectively, the spatial concentration of dust activity remained in the border areas. The Absorbing Aerosol Index (AAI) extracted from Sentinel-5P data further confirmed the severity of the situation. In 2020, the mean AAI values in Mehran, Dehloran, and Abdanan were around 0.28, increasing to 0.32 in 2021, and exceeding 1.3 in 2022 -indicative of very unhealthy conditions for the general population. Despite slight declines in 2023 (0.87) and 2024 (0.86), values remained in the unhealthy range. MODIS-derived AOD data also played a key role in assessing dust intensity. In 2020, AOD levels surpassed 1 in border areas and exceeded 1.6 in some regions in 2021. The critical peak occurred in 2022, when AOD values reached over 1.85 in southern Ilam and western Dehloran. Even central parts of the province saw AOD values greater than 0.5 in the same year. In 2023 and 2024, the values were 1.3 and 1.18, respectively, remaining within hazardous levels. The NDDI index, which reflects dust deposition on surfaces, peaked in 2021 with values exceeding 0.9 in some border areas. In 2022, the index dropped to approximately 0.5, possibly indicating airborne dust with limited ground deposition. It reached its lowest point in 2023 (below 0.5), followed by a slight increase to 0.54 in 2024. The HYSPLIT model was used to simulate dust transport pathways for two critical events in 2025. On April 15, 2025, the model identified western Iraq as the main dust source. Simulations showed that the dust plume reached the Ilam border at 11:00 AM and Dehloran station by 12:00 PM. Vertical profiles indicated that dust particles initially traveled at 500 meters altitude and later descended into the boundary layer, corroborating the recorded AQI level of 500 in Dehloran. In the second event on May 25, 2025, the dust originated from the deserts of eastern Syria. The particles formed at an altitude of 2000 meters and traveled across Iraq, reaching Mehran station at 12:00 PM. The trajectory showed a gradual descent to 500 meters, leading to severe surface-level pollution. Trajectory frequency maps indicated that more than 90% of paths passed through Syria, confirming the combined influence of Iraqi and Syrian sources. This event also saw an AQI level of 500 in Mehran. Overall, the results underscore the spatial stabilization of dust hotspots in Ilam’s border regions and highlight the critical role of transboundary dust sources in Iraq and Syria, as well as the synoptic wind patterns that facilitate their transport.
Conclusions
The findings demonstrate a notable increase in the frequency and intensity of dust storms in Ilam Province in recent years, with a clear spatial concentration in border areas. Transboundary sources, particularly desert regions in Iraq and Syria, have significantly contributed to the worsening dust pollution. The integration of satellite indices with the HYSPLIT model enabled the precise identification of dust origins, transport paths, and intensity. Consequently, implementing control strategies such as the restoration of drought-resistant vegetation, soil stabilization, land moistening, establishment of greenbelts along the borders, enhancement of regional cooperation with neighboring countries, and deployment of satellite-based early warning systems is essential. Without such interventions, the current trajectory may lead to a chronic crisis and exacerbate environmental, social, and economic vulnerabilities in the region.
Sina Nabizadeh; Ali asghar Naghipour; Ataollah Ebrahimi; Hamidreza Keshtkar; Elham Ghehsareh
Abstract
Introduction
Land use/land cover (LULC) maps are among the key tools for natural resource management, regional planning, and achieving sustainable development; therefore, the need for their accurate and up-to-date monitoring is increasingly emphasized. Continuous changes in land use driven by natural ...
Read More
Introduction
Land use/land cover (LULC) maps are among the key tools for natural resource management, regional planning, and achieving sustainable development; therefore, the need for their accurate and up-to-date monitoring is increasingly emphasized. Continuous changes in land use driven by natural and human-induced factors have significantly affected local and global ecosystems. In recent decades, advances in remote sensing technologies and machine learning algorithms have led to major improvements in the extraction and classification of spatial data. The Google Earth Engine (GEE) platform, as a powerful cloud-based infrastructure for processing large-scale spatiotemporal data, provides an efficient framework for producing accurate and updated maps. Within this context, the present study focuses on the integration of multi-temporal satellite images, the use of auxiliary data, and the comparison of three machine learning algorithms over a large and heterogeneous watershed (Karun 1), aiming to improve classification accuracy and enhance the capability for long-term monitoring of LULC changes.
Materials and methods
To assess LULC changes in the Karun 1 watershed, Landsat 7 ETM+ (2002) and Landsat 8 OLI (2024) images with cloud cover less than 10% and considering long-term mean precipitation were retrieved and processed as surface reflectance products in the GEE platform. Composite images were generated from nine Landsat scenes during the peak growing season (May to July) using a median filter and were then clipped to the watershed boundary. A total of 1,920 training samples representing seven LULC classes based on the Anderson classification scheme were collected using field survey data, aerial photographs, and Google Earth imagery. The reference dataset was randomly split into training (60%) and evaluation (40%) subsets Auxiliary variables (such as NDVI, NDBI, NDWI, and a DSM) were derived and included alongside original spectral bands. Classification experiments were implemented in GEE using three supervised algorithms: CART, RF, and SVM. Model hyper parameters and training procedures were configured to ensure reproducibility and consistency across methods.
Results and discussion
The results showed that the CART, RF, and SVM algorithms produced classified maps with excellent accuracy. The incorporation of vegetation indices and auxiliary data improved both the overall accuracy and the Kappa coefficient for both study years. The highest overall accuracy and Kappa coefficient were achieved by the SVM algorithm, with values of 93% and 91.5% in 2002, and 93% and 92% in 2024, respectively. According to the results of all three algorithms, rangelands constitute the largest proportion of the watershed area (on average about 40%), followed by forests (approximately 27%). The temporal analysis indicated a decreasing trend in the area of rangelands and forests, as well as a notable reduction in water bodies, particularly based on the SVM results.
Conclusions
The results of this study demonstrated that applying machine learning algorithms within the Google Earth Engine platform enables the production of accurate land use/land cover maps and the effective monitoring of environmental changes over a large and heterogeneous watershed. The obtained results can be used as an efficient tool for land use planning, natural resource management, monitoring vegetation degradation, and controlling land use changes in the study area. However, limitations related to the spatial resolution of Landsat imagery and pixel-mixing errors, particularly along the boundaries between LULC classes, are considered among the main challenges of this study. Accordingly, it is recommended that future research utilize higher spatial resolution data, such as Sentinel imagery, to improve classification accuracy, especially for vegetation cover mapping.
Shadi Jalilian; Shaban Shataee Jouibary; Mohammad Hadi Moayeri; Amir Saddodin
Abstract
IntroductionLandslides, as one of the most destructive natural disasters, cause significant annual human and financial losses worldwide, particularly in Golestan Province due to its specific topographic and climatic conditions. Although numerous studies have been conducted in this region, most have relied ...
Read More
IntroductionLandslides, as one of the most destructive natural disasters, cause significant annual human and financial losses worldwide, particularly in Golestan Province due to its specific topographic and climatic conditions. Although numerous studies have been conducted in this region, most have relied on classical and traditional statistical methods, leaving a notable gap in the application of advanced deep learning algorithms. This research aimed to address this gap with three main innovations: 1) employing deep learning models (CNN and RNN) for the first time in the region, 2) conducting a systematic comparison with classical machine learning models (RF and SVM), and 3) focusing on a forest watershed with a sensitive and complex ecosystem (Hyrcanian). Materials and methodsThe study area was Forest Watershed 85, covering approximately 39,288 hectares in Golestan Province. Based on previous studies and analysis of regional conditions, ten factors influencing landslide occurrence were selected: elevation, slope percentage, slope aspect, Topographic Wetness Index (TWI), Normalized Difference Vegetation Index (NDVI), land use, distance to roads, distance to rivers, distance to faults, and average annual rainfall. These factors were prepared as raster layers with a 30x30 meter cell size in a Geographic Information System (ArcGIS) environment. Landslide point data (247 points) were obtained from the General Department of Natural Resources and Watershed Management of Golestan Province. To balance the dataset, an equal number of points (247) were randomly selected from areas without landslide phenomena. Thus, a balanced dataset comprising 494 points was prepared for modeling. The data were split into a 70:30 ratio (346 points for training and 148 points for validation). Four advanced models, including two machine learning models (Random Forest, RF, and Support Vector Machine, SVM) and two deep learning models (Convolutional Neural Network, CNN, and Recurrent Neural Network, RNN), were implemented and trained in the Python environment using the Scikit-learn and TensorFlow libraries. The performance of the models was quantitatively evaluated using the Receiver Operating Characteristic (ROC) curve and the Area Under the Curve (AUC) index. Furthermore, the validation of the generated susceptibility maps in high-risk areas (high and very high-risk classes) was performed using 30% of the unused samples, assessed by the overall accuracy and Kappa coefficient metrics. Results and discussionIn this research, the performance of four machine learning and deep learning algorithms was evaluated for landslide susceptibility zoning in Forest Watershed 85 of Golestan Province. The results demonstrated that the deep learning algorithm CNN, with an AUC score of 0.910, an overall accuracy of 87.72%, and a Kappa coefficient of 0.899 for high-risk classes (high and very high risk), was identified as the most efficient model. Variable importance analysis using the superior model (CNN) revealed that the factors Distance to Fault and Distance to River were respectively the most significant contributors to landslide occurrence in the study area. This finding is entirely consistent with the geological and geomorphological expectations of the region. The final landslide susceptibility map generated by the CNN model indicated that 31.46% of the watershed area (approximately equivalent to 12,360 hectares) is classified within the very high-risk category. ConclusionThe findings of this research clearly confirm the superiority and high potential of deep learning algorithms, particularly the Convolutional Neural Network (CNN) architecture, for producing high-accuracy landslide susceptibility maps compared to machine learning algorithms. The generated map can serve as a reliable and powerful scientific tool for managers and planners, enabling them to prioritize preventive measures, risk management, and safe land-use planning by focusing on high-risk areas. For future studies, integrating these models with optimization algorithms, utilizing higher-resolution data to overcome the limitations of the current data, and developing hybrid frameworks that enhance both accuracy and interpretability are recommended.
Amin Salehpour Jam; Jamal Mosaffaie
Abstract
Introduction
The country's watersheds are dynamic ecosystems whose health has been affected by civil, economic, and social developments. This is while, in the current situation, in addition to human and management factors, climate change has also had undesirable consequences in these areas. The decline ...
Read More
Introduction
The country's watersheds are dynamic ecosystems whose health has been affected by civil, economic, and social developments. This is while, in the current situation, in addition to human and management factors, climate change has also had undesirable consequences in these areas. The decline in the health of the country's watersheds, on the one hand, has made it difficult to provide ecosystem services in these areas, and on the other hand, has led to the emergence of various environmental hazards such as desertification and land degradation, land subsidence, floods, landslides, and dust phenomena. In this regard, the country's Natural Resources and Watershed Management Organization implements a variety of biological, mechanical, biomechanical, and management measures to conserve water and soil and control floods in its watersheds. This is despite the fact that these measures are mostly reactive (therapeutic) and less attention has been paid to their preventive aspects. Accordingly, this study aims to identify and prioritize reactive and proactive solutions to improve the health state of the Kal-Aji watershed based on the DPSIR framework and non-parametric statistical tests.
Materials and methods
In the first stage, the drivers and pressures resulting in the health status of the Kal-Aji watershed and the related impacts were identified through a literature review, a visit to the watershed, interviews with experts from the departments of natural resources, environment, regional water, the Agricultural Jihad, the Agricultural and Natural Resources Engineering Organization of Golestan, faculty members of academic and research centers, and interviews with local communities. Then, a working group consisting of 26 stakeholders, local knowledgeable individuals, and experts knowledgeable about the issues and problems of the watershed began to determine solutions to improve the health of the Kal-Aji watershed and eliminate or modify the related adverse impacts. In the last stage, after forming the DPSIR table and identifying the various components, the importance of each of the variables categorized under the five DPSIR components was prioritized and determined. For this purpose, a Likert-scale questionnaire was used as a measurement tool. In this study, each variable was considered as an item, and the validity of the questionnaire was finally confirmed based on the opinions of experts. Also, Cronbach's alpha method was used to calculate the reliability of the measurement tool.
Results and discussion
In this study, eight drivers and 16 subsequent pressures were identified in creating six variables related to the health state component of the Kal-Aji watershed. In addition, seven variables related to the adverse impacts of the current watershed health state and 28 responses to improve the watershed health state were identified and introduced. The results show that D1 and D8 have the minimum and maximum links in creating subsequent pressures, respectively. Also, D5 with 8 links is the second driver with the maximum connections with the pressure component. In this study, the responses identified as resolving the problems listed under the four components of driver-pressure-state-impact are one of the following three: (1) a specific response to a specific problem, (2) a multi-objective response that resolves more than one problem, and (3) the existence of different responses for a specific problem. Accordingly, of the total responses, 35.7 percent (10), 46.4 percent (13), 14.3 percent (4), and 25.0 percent (7) are related to the components of driver, pressure, state, and impact, respectively. Considering the calculated values above 0.7 of Cronbach's alpha, all questionnaires have acceptable reliability (P and R) and goodness (D, S, and I) in this study. There is also a significant difference between the types of variables identified under the components of the DPSIR framework. The results showed that D5, D8, and D3, in order of importance, have been assigned the first three priorities from the experts' perspective. The results also showed that P1, P11, P6, and P3, respectively, have been assigned the top five priorities of pressures. These pressures are among the common pressures in the country's watersheds. In this study, S2 was prioritized as the most important indicator of the health status of the watershed. Consistent with the results of other studies, in addition to applying reactive water and soil conservation responses in watershed management to improve directly the state, it is necessary to pay attention to proactive responses that eliminate the various drivers and pressures identified as a result of the current state of the watershed's erosion and sediment production potential. Also, S3 and S1 were prioritized in the next order of importance. The ranking of the consequences resulting from the current state of watershed health showed that I3, I1, and I2 were assigned the first to third priority, respectively. Also, prioritizing the importance of various reactive and preventive responses showed that R20, R4, R6, R17, and R9, respectively, have been assigned the top five priorities in eliminating or modifying drivers and pressures, improving the health state of the Kal-Aji watershed, and eliminating or modifying the adverse impacts of its health state. In this regard, R4, R6, R17, and R9 are among the common responses considered in other studies conducted in other watersheds of the country.
Conclusions
In this study, the importance of various drivers and pressures resulting in the current state of watershed health and its subsequent impacts were identified and prioritized. Also, types of reactive and proactive responses to improve the watershed health and eliminate or reduce the associated adverse impacts were identified and prioritized. The results of the Friedman test indicated the presence of a significant difference between the importance of the types of variables identified under the DPSIR framework components. Accordingly, D5, P1, S2, and I3 were prioritized as the most important drivers, pressures, states, and impacts, respectively. Also, R20, R4, and R6 were assigned the first three priorities of responses, respectively. In this regard, it is strongly recommended to pay attention to (1) all solutions to address the problems identified under the components of the DPSIR framework, (2) the role of various stakeholders in the basin in the planning process to improve watershed health and reduce related risks, (3) the development of action plans related to the responses, and (4) the development of decision support systems and related databases.
Saeedreza Moazeni Noghondar; Ali Salajeghe; Shahram Khalighi Sigaroudi; Ali Golkarian
Abstract
Introduction
Mountainous regions, as the most sensitive and vital ecosystems, play a crucial role in providing fresh water, regulating climate, and preserving biodiversity. However, these areas are vulnerable to soil erosion, natural resource degradation, and climate change. Sustainable water and soil ...
Read More
Introduction
Mountainous regions, as the most sensitive and vital ecosystems, play a crucial role in providing fresh water, regulating climate, and preserving biodiversity. However, these areas are vulnerable to soil erosion, natural resource degradation, and climate change. Sustainable water and soil management in mountainous regions is particularly important, as soil quality is directly linked to essential ecosystem services. Natural factors such as slope and aspect influence the distribution of organic matter, soil aggregate stability, and water infiltration, while human activities like uncontrolled pasture exploitation, overgrazing, vegetation destruction, and land use changes contribute to soil degradation, reduced infiltration, and severe erosion. Management practices, including watershed structures and planting native species, can enhance soil water and nutrient retention, improving its physical and chemical properties. Paired watersheds serve as natural laboratories for monitoring and evaluating changes in soil, water, and ecosystem quality under the influence of human activities and climatic conditions. This study aims to investigate the impact of topographic variables, specifically slope aspect and the topographic wetness index (TWI), on the effectiveness of restoration measures in enhancing the physical and chemical properties of soil.
Materials and methods
This study was conducted in the paired watershed of Gonbad, located in Hamedan province, Iran. The area comprises two sub-watersheds: one subjected to watershed management operations (treated sub-watershed) and the other without such activities (control sub-watershed). To assess the effect of topography on soil properties, soil sampling was performed at various points in both the control and treated areas, based on two key topographic factors: slope aspect and TWI. For slope aspect analysis, three points were selected on north-facing slopes and three on south-facing slopes within each sub-basin. The TWI was calculated using the relevant equation, and each sub-basin was divided into three zones with varying moisture conditions (low, medium, and high TWI). Soil samples were collected at the end of the growing season from a depth of 0-15 cm while maintaining the soil structure. With three replicates, a total of 36 sampling points were established. Soil physical and chemical properties, including permeability, texture, porosity, aggregate stability, organic matter, pH, electrical conductivity, water holding capacity, and surface cover components, were measured. Restoration measures in the treated sub-watershed included biological measures (seeding of drought-resistant species such as Astragalus gossypinus and Bromus tomentellus) and managerial measures (complete grazing exclusion). To analyze the effects of restoration measures, slope aspect, and TWI on soil hydrological properties, statistical methods including analysis of variance (ANOVA) based on a nested design, and Pearson correlation were employed using SAS and R software.
Results and discussion
The findings demonstrated that restoration measures and topographic variations significantly improved soil properties. Comparison of the treated and control watersheds using the t-test revealed that the treated watershed exhibited lower bulk density (1.18±0.01 vs. 1.31±0.02 g/cm³), reduced bare soil percentage (13.06±1.38% vs. 32.5±1.61%), and higher steady-state infiltration rate (28.44±1.92 vs. 19.78±0.82 mm/h) (P<0.05). Additionally, soil porosity (51.13±0.73% vs. 41.66±1.14%) and aggregate stability (1.96±0.52 vs. 1.52±0.39 mm) were significantly greater in the treated watershed (P<0.05). Organic matter content was also higher in the treated watershed (2.15±0.62% vs. 1.5±0.38%) (P<0.05), indicating the positive influence of restoration on soil quality and erosion control. Slope aspect significantly affected certain soil properties. The t-test showed that north-facing slopes had greater aggregate stability (2.14±0.33 vs. 1.35±0.29 mm), higher organic matter (2.13±0.64% vs. 1.32±0.38%), and denser vegetation cover (60.39±3.18% vs. 48.17±3.2%) compared to south-facing slopes (P<0.05). These differences are linked to improved moisture conditions on north-facing slopes due to reduced solar radiation and denser vegetation, enhancing organic matter retention and erosion resistance. Analysis of TWI classes using ANOVA indicated that the TWI3 class had the highest organic matter (1.45±0.61%) and aggregate stability (1.96±0.52 mm), though these differences were not statistically significant (P>0.05). However, soil texture varied significantly with TWI: clay increased from 42.75±3.12% (TWI1) to 46.12±3.05% (TWI3), silt from 28.43±2.76% to 31.66±2.78%, and sand decreased from 28.82±5.53% to 22.22±5.68% (P<0.05), suggesting finer particle deposition in higher TWI zones (lower slopes). Principal component analysis (PCA) showed that TWI and vegetation were correlated with organic matter, aggregate stability, and infiltration (P<0.05), while bare soil and bulk density were associated with southern slopes and low TWI. Northern slopes and high TWI classes showed the greatest improvement in organic matter and vegetation. These findings confirm the key role of slope orientation, TWI, and biological interventions in improving soil and vegetation quality in semi-arid regions.
Conclusion
The present research showed that restoration measures generally lead to improved soil quality, but the effectiveness of these measures is significantly affected by topographic characteristics, especially slope aspect and topographic wetness index (TWI). North-facing slopes and points with high TWI showed the greatest improvement in soil parameters, especially organic matter and soil aggregate stability. The findings of this research show the importance of considering topographic characteristics in planning and implementing restoration measures, and it is suggested that in future studies, more focus should be on investigating the factors affecting the reduction of some parameters and the long-term effects of restoration measures.
Ehsan Bazrafshan; Hossein Malekinezhad; Seyed Zeynalabedin Hosseini; mehdi sepehri
Abstract
Introduction
Flood risk management is one of the most significant environmental and developmental challenges in arid and semi-arid regions of Iran. The Fakhraabad watershed in Yazd is prone to sudden floods due to its unique climatic and topographical characteristics, which can result in substantial ...
Read More
Introduction
Flood risk management is one of the most significant environmental and developmental challenges in arid and semi-arid regions of Iran. The Fakhraabad watershed in Yazd is prone to sudden floods due to its unique climatic and topographical characteristics, which can result in substantial economic, social, and environmental damages. Since studies related to natural phenomena are often associated with complexity and uncertainty, employing multi-criteria decision-making methods capable of effectively managing these uncertainties can play a key role in improving the analysis and simulation of natural resource-related phenomena. Ultimately, this approach could lead to a reduction in the economic and human costs resulting from these events. Among these methods, the Analytic Hierarchy Process (AHP) has gained a special position due to its simple structure, transparency, and widespread use in natural resource studies and watershed management. This method provides a systematic way of weighting and prioritizing criteria and has offered reliable results in many studies. However, when data or expert judgments are uncertain, the use of complementary approaches can enhance the accuracy of results. In this regard, the IRNAHP method (Fuzzy Analytic Hierarchy Process with Interval Numbers) has a greater ability to model uncertain conditions and can be used as a complement to AHP. Comparing these two methods, while maintaining the position of AHP as a foundational tool, provides a suitable path for selecting a more precise approach in flood risk management.
Materials and methods
In this study, to prioritize the sub-basins of the Fakhraabad watershed in Yazd based on flood susceptibility, eight main criteria were used: Digital Elevation Model (DEM), slope, precipitation, fractal dimension, connectivity, Topographic Wetness Index (TWI), Topographic Control Index, and stream power. Data related to these criteria were extracted from hydrological sources, topographic maps, and climatic data, and processed in a Geographic Information System (GIS) environment. To weight and prioritize the criteria, two multi-criteria decision-making methods, AHP and IRNAHP, were applied. In the AHP method, pairwise comparisons were completed with expert opinions, and the relative weights of the criteria were calculated. In the IRNAHP method, fuzzy logic and interval numbers were used to consider the uncertainty in human judgments. The weights obtained from both methods were integrated in the GIS environment to produce flood susceptibility maps of the sub-basins. Finally, to validate the results obtained from the two methods, the output of the SWAT model was used as a reference for comparison to evaluate the accuracy and reliability of each method.
Results and discussion
Comparing the results of the two methods with the SWAT model output showed that both AHP and IRNAHP were able to provide an appropriate flood susceptibility zonation pattern. The AHP method, due to its simplicity, transparency, and widespread application in water resource studies, remains a valuable tool. However, as the number of pairwise comparisons increases and expert judgments become more uncertain, the likelihood of uncertainty in the weighting of the indicators also rises. In contrast, the IRNAHP method, utilizing fuzzy logic and interval numbers, was able to better manage this uncertainty and provide more accurate results. Compared to similar studies, the findings of this research also indicated that IRNAHP performed better than AHP in dealing with ambiguous data and environmental complexities.
Conclusion
This study compared the two methods, AHP and IRNAHP, for analyzing the flood susceptibility of the sub-basins in the Fakhraabad watershed in Yazd. The results showed that in the AHP method, sub-basins 4, 3, 31, 27, and 29 had the highest flood susceptibility, while sub-basins 15, 22, 14, 21, and 6 showed the lowest susceptibility. In the IRNAHP method, sub-basins 4, 3, 31, 27, and 24 had the highest flood susceptibility, while sub-basins 15, 22, 14, 21, and 12 had the lowest. Due to its simple structure and high interpretability, AHP remains a reliable method for prioritizing criteria. However, when the data or expert opinions contain ambiguity and uncertainty, its accuracy decreases. In contrast, the IRNAHP method, by incorporating fuzzy logic and interval numbers, overcomes this limitation and provides more accurate results. Therefore, it can be concluded that IRNAHP, as a complementary approach to AHP, is a more efficient tool for flood risk management in vulnerable areas.
Mahmoudreza Tabatabaei; Amin Salehpour Jam; Jamal Mosaffaie
Abstract
IntroductionRapid population growth and economic development, especially in sensitive areas like watersheds, have increased the demand for water resources, leading to significant challenges. In watershed management, monitoring erosion and sedimentation processes is of paramount importance, as these processes ...
Read More
IntroductionRapid population growth and economic development, especially in sensitive areas like watersheds, have increased the demand for water resources, leading to significant challenges. In watershed management, monitoring erosion and sedimentation processes is of paramount importance, as these processes directly influence the quality and quantity of water resources. Awareness of suspended sediment loads in these basins can facilitate a deeper understanding of ecological processes and assist in the assessment of the environmental status of watersheds, which in turn improves the management and conservation of water resources. The word “transform” in your sentence means to change or significantly improve something. In this context, it suggests that the design and implementation of advanced systems will lead to a significant enhancement or alteration in the way research is conducted regarding watershed management, sediment, and erosion. By aggregating and analyzing hydrologic data, this system provides accurate and up-to-date information on the amount of suspended sediment in rivers, enabling continuous and comprehensive monitoring and evaluation of watersheds. This, in turn, facilitates the adoption of necessary measures for the optimal management of these resources.. Materials and methodsIn this research, a specialized spatial system for the management and analysis of hydrological data has been developed using the C# programming language and open-source spatial libraries. This system utilizes the SQLite database as its underlying data storage platform and employs Entity Framework 6 (EF6) and LINQ for facilitating data extraction and management. In the database design, sedimentation and flow discharge data have been comprehensively collected and stored. The system is capable of executing a variety of spatial and descriptive queries and analyses on this data. Furthermore, it provides the capability to perform statistical analyses and extract statistical summaries from the sedimentation data. These features enable researchers to process and analyze the data with ease, allowing for a more precise examination of results related to erosion and sedimentation processes. The designed system serves as an efficient tool, facilitating the monitoring and assessment of watershed conditions. Results and discussionThe results of this research can be summarized into two main sections: system design and construction, and statistical analysis of sedimentation data in the Aras watershed. In the first section, which focuses on the design and development of the spatial system, appropriate graphical interfaces for data storage, user interaction with the database, and data management were designed and coded with the help of modern techniques and advanced methods. By utilizing these tools, the data extraction and management process was streamlined, and the possibility of performing statistical analyses was enabled. As a result, users can process and analyze the data with greater accuracy and ease. In the second section, the statistical results obtained from the analysis of the sedimentation data of the Aras Basin (with 26,156 data points recorded up to 2017) indicate that, at the basin scale, the average daily suspended sediment discharge is 11,814.95 tons per day, the average concentration of suspended sediment is 4,185.68 milligrams per liter, and the average instantaneous flow discharge is 13.16 cubic meters per second. At the level of study units, the highest and lowest average suspended sediment discharge values were found in the study units of Jolfā-Dūzāl (code 1105) with a value of 10,312.33 tons per day, and Qara Ziyaldin (code 1108) with a value of 991.96 tons per day, respectively. Moreover, at the hydrometric station scale, the highest and lowest average daily suspended sediment discharge were observed at the Jolfā hydrometric station (code 807-19) with 571,697.82 tons per day, and Nāvar-Khrouji Neor (code 0195-19) with 2.82 tons per day, respectively. The extreme values of suspended sediment recorded at the Jolfā hydrometric station are significantly high (over 3 million tons per day), so these figures should be used with caution. According to the findings, the land in this area is located in marly zones that are highly susceptible to erosion, and the contribution of lateral erosion in producing sediment is significant in river sections that have developed meandering patterns. However, further studies on land-use changes and other factors influencing sediment generation in this area are required. ConclusionThis research was conducted with the aim of developing a national software infrastructure for managing suspended sediment data and flow discharge from hydrometric stations. The developed system possesses the capability to store, retrieve, and analyze spatial and temporal data, and it is specifically utilized for simulating suspended sediment in rivers. Additionally, advanced graphical interfaces have been created to significantly facilitate user interaction with the data. One of the notable features of this system is its use of open-source technologies in its design and development. This ensures that there are no legal restrictions or dependencies on foreign licenses for its use and development. This characteristic significantly enhances the commercialization potential of the system both domestically and in international markets. In addition to the design and development of this software system, the results obtained from its application in rapid and accurate statistical analysis of sediment data from the Aras watershed clearly demonstrate its unique capabilities in data management and analysis. Given the importance of accurately understanding erosion and sedimentation processes in the watersheds of the country, this system serves as an efficient and effective tool for obtaining precise and reliable information.
Nader Jandaghi; Mehdi Alibegli
Abstract
Introduction
Water resources management in arid and semi-arid regions is one of the fundamental challenges of the present century. Population growth, climate change, and increasing demand have put significant pressure on surface water resources. In this context, monthly runoff prediction serves as a ...
Read More
Introduction
Water resources management in arid and semi-arid regions is one of the fundamental challenges of the present century. Population growth, climate change, and increasing demand have put significant pressure on surface water resources. In this context, monthly runoff prediction serves as a strategic tool for reservoir planning, flood control, and sustainable watershed management. However, the complexity of hydrological processes and the nonlinear relationships among climatic variables make such predictions challenging. In recent years, hybrid models have emerged as a novel approach in hydrology that enhances prediction accuracy and stability by integrating the strengths of different models. The present study aims to evaluate the performance of linear and deep learning models and to introduce an optimal model for enhancing water resources management.
Materials and methods
In this study, monthly runoff in the Qarasu watershed, with an area of 1,624 km2, was modeled and predicted using data from three hydrometric stations (Pole-Touskestan, Naharkhoran, and Siyahab) over a common 36-year period (1985-2021). Three single models (SARIMA, BiLSTM, and GRU) and two hybrid models (BiLSTM-GRU and SARIMA-BiLSTM-GRU) were employed to model monthly runoff and forecast values for a 12-month horizon. In hybrid models, the SARIMA model is used to model the linear components of the time series, while BiLSTM and GRU models have the ability to identify and represent complex, non-linear patterns. Model performance was evaluated using RMSE, MAD, and MSE indices. The models were implemented using Python and commonly used libraries, including TensorFlow, Keras, numpy, pandas, matplotlib, scipy, and sklearn.
Results and discussion
The hybrid SARIMA-BiLSTM-GRU model, by integrating linear and nonlinear components, provided the most accurate monthly runoff predictions. The RMSE values of this model at the Pole-Touskestan, Naharkhoran, and Siyahab stations were estimated at 0.0295, 0.0173, and 0.1683 m³/s, respectively. The BiLSTM-GRU model ranked second, with RMSE values of 0.0326, 0.0226, and 0.3013 m³/s. Among the individual models, BiLSTM and GRU produced similar and relatively accurate results, while the linear SARIMA model, with RMSE values of 0.0851, 0.0230, and 0.3892 m³/s, showed the lowest performance. On average and based on the RMSE index, the SARIMA-BiLSTM-GRU hybrid model reduced prediction errors by 39.66% to 56.75% compared to the other models. The findings of this research confirm and reinforce the theoretical foundations regarding the superiority of complex hybrid models over simple, linear models.
Conclusions
This study demonstrated that hybrid models combining linear and deep learning approaches can significantly improve the accuracy and stability of monthly runoff predictions. The modeling and validation results indicated that the hybrid SARIMA-BiLSTM-GRU model, which effectively combines linear (SARIMA) and nonlinear (BiLSTM and GRU) components, was selected as the optimal model for this watershed, demonstrating the best performance in monthly runoff prediction. The results highlight the importance of applying hybrid approaches in water resources management and flood control. It is recommended to employ more advanced models and diverse input variables, such as precipitation and temperature, to further enhance the accuracy and reliability of predictions. Furthermore, given the observation of overfitting in the base LSTM and GRU models, it is recommended to employ more advanced architectures, such as models based on the Attention mechanism like the Transformer, to enable the model to intelligently focus on the key and influential time periods in the runoff process.
