Evaluating the Performance of Dez Dam Reservoir Operation in Flood Control Using Machine Learning Algorithms and Remote Sensing

Evaluating the Performance of Dez Dam Reservoir Operation in Flood Control Using Machine Learning Algorithms and Remote Sensing | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Evaluating the Performance of Dez Dam Reservoir Operation in Flood Control Using Machine Learning Algorithms and Remote Sensing Sajad Zareie, Mohammad Abaforoushan, Mohammad Farhadiyan This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7104599/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 5 You are reading this latest preprint version Abstract Floods are considered one of the most destructive natural phenomena, causing extensive damage to the various sectors, including agriculture, infrastructure, housing, and socio-economic activities. The watersheds of the Dez and Karun rivers in Khuzestan Province require comprehensive management and precise analyses due to the frequent occurrence of the floods and their economic consequences. In this study, machine learning algorithms and Landsat 5, 7, and 8 satellite imagery were employed to identify flood-prone areas and evaluate the extent of flood-induced damages. Flood hazard zonation maps were generated with satisfactory accuracy and validated against field observation data, achieving an overall accuracy of 75%. By integrating hazard maps with land-use maps, vulnerable assets including agricultural lands, orchards, and rural settlements were identified. Also, regression models analyzed the relationship between river discharge and flood extent with 82.6% accuracy, revealing that outflow discharge is a key determinant of flood severity and spatial distribution. The findings demonstrate that combining remote sensing technologies with machine learning methods provides a robust tool for flood risk assessment and effective crisis management. Based on the results, can be proposed mitigation strategies for flood-prone areas, flood-related insurance policy frameworks, and optimized natural resource management. Finally, we strongly recommend avoiding development in high-risk flood zones and implementing watershed-scale risk reduction measures. Damage extent Flood forecasting Support vector machine Linear regression Landsat Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 1. Introduction Floods are among the most widespread and destructive natural disasters, inflicting heavy annual losses in human societies and economic damage on communities. Furthermore, climate change and human activities have exacerbated the intensity and frequency of flood events (Merz et al. 2021 ). Catastrophic flood events arise from multiple causative factors, including both anthropogenic and natural sources (Rao 2001 ). Given the extensive and long-term impacts of flooding, prioritizing preventive risk reduction measures such as developing accurate flood prediction models and improving land use planning in vulnerable areas, is of critical importance (Bonakdari et al. 2020 ; Chapi et al. 2017 ; Tien Bui et al. 2018). The Dez river Basin, as one of Iran's most flood-prone areas, has consistently faced threats from sudden and devastating floods. Recent flood events in this region have further highlighted the critical need for comprehensive and accurate studies to assess flood risks and identify vulnerable zones (Samadi et al. 2019 ). Integration of GIS and Remote Sensing data with other datasets offers significant potential in modern technology for flood disaster identification, monitoring, and assessment (Pradhan et al., 2009). Data-driven models serve as crucial alternatives for establishing relationships between input and output data without requiring precise understanding of underlying physical processes (Wang et al., 2023). To evaluate flood impact on socio-economic damages in the study area, remote sensing methods were implemented to determine flood magnitude. Traditional flood risk assessment methods primarily rely on hydrological and hydraulic data. While these approaches provide valuable information, they suffer from several limitations, including high input data requirements, substantial costs, and time-consuming processes. Recent advances in remote sensing technologies have enabled access to high-precision satellite data with extensive coverage. These datasets, combined with other spatial data such as topographic maps, land use classifications, and river network maps, can be effectively utilized to produce high-accuracy flood hazard zonation maps (Radwan et al. 2019). Remote sensing and GIS technologies enable the collection and analysis of large volumes of data, including topographic, hydrological, and climatic variables, which are also highly useful for identifying flood-prone areas. Remote sensing and GIS play a significant role in generating flood susceptibility maps and enhancing our understanding of flood-vulnerable regions (Tehrany, Pradhan, and Jebur 2014). Kazemi et al. ( 2024 ) investigated flood susceptibility mapping using machine learning and remote sensing. They used the machine learning techniques of MARS, CART, TreeNet, and RF. Based on the results, the TreeNet technique demonstrated the most promising performance among the machine learning algorithms. Hajji et al. ( 2025 ) conducted a comparative evaluation of three machine learning algorithms of gradient boosting, AdaBoost, and random forest for flood prediction. Their analysis revealed that the random forest algorithm exhibited superior predictive accuracy, outperforming both the gradient boosting ensemble and individual implementations of gradient boosting and AdaBoost. Pham et al. ( 2025 ) evaluated flood susceptibility through an integrated approach combining machine learning and remote sensing techniques. Their results indicated that the Genetic Algorithm-optimized Artificial Neural Network model achieved optimal performance metrics (RMSE = 4.332 and MAE = 4.020). The derived flood susceptibility map revealed significant spatial variations, with particularly pronounced contrasts between urban and suburban regions. By leveraging advanced remote sensing and GIS technologies, and machine learning algorithms, Asare-Kyei et al. (2015) developed a comprehensive model for flood hazard zonation. Their findings demonstrated that up-to-date and accurate maps of high- and medium-risk flood zones can be utilized in managerial decision-making to mitigate the devastating impacts of floods. Tehrany et al. ( 2014 ) demonstrated that integrating topographic data, land use, precipitation, river networks, and other parameters within a GIS environment combined with RS data can generate highly accurate flood susceptibility maps. Furthermore, Kazemi et al. ( 2024 ) have shown that advanced TreeNet and CART models exhibit high performance in classifying flood-prone areas in the Karun River basin and can be effectively utilized to produce flood susceptibility maps with appropriate accuracy. Through the analysis of hydrological data, Vaghefi et al. ( 2019 ) concluded that the likelihood of flooding in various regions of Iran follows an increasing trend. And, their findings indicate that preventive approaches based on data analysis, particularly through the integration of remote sensing technology and machine learning algorithms, can significantly enhance flood risk mitigation efforts. Although floods cannot be entirely prevented, adopting an integrated flood management approach can minimize their adverse consequences (Dewan et al. 2007 ). Studies conducted by Gudiyangada Nachappa et al. (2020) and Shahabi et al. (2020) have contributed significantly to improving flood mitigation planning and enhancing awareness of at-risk areas. By integrating remote sensing technology with machine learning algorithms, these studies have introduced novel perspectives in flood management. Given the escalating frequency of flood events and their substantial human and economic impacts, the development of precise, innovative, and efficient methods for flood risk assessment and crisis management has become an undeniable necessity. Within this context, identifying flood-vulnerable zones has emerged as a key research challenge, receiving considerable attention in recent studies. The innovation of this research lies in the systematic integration of remote sensing and GIS technologies, along with the development of a simple linear regression machine learning algorithm in the JMP software environment, to assess flood damage levels. Additionally, SVM algorithm, one of the most accurate and efficient machine learning methods for classification (Cortes and Vapnik, 1995), was employed to determine flood damage levels in the study area. The significance of this study is its contribution to enhancing awareness of at-risk areas and providing actionable data to support decision-making in spatial planning, integrated resource management, and the mitigation of potential hazards to human communities, infrastructure, and the environment. 2. Case study Previous studies indicate that flood occurrences in various regions of Iran have caused significant damage to the agricultural and livestock sectors, severely affecting the livelihoods of many rural communities. Consequently, preventive measures to mitigate these risks must be prioritized as a critical issue (Vaghefi et al. 2019 ). Accordingly, the Dez river basin up to its confluence with the Karun river in the Band-e Qir region was selected as the study area due to the occurrence of a major flood in 2020. The Dez river is one of the most important rivers in southwestern Iran, formed by the confluence of two main tributaries, the Sezar and Bakhtiari rivers, in Lorestan Province. The merging of these two branches creates the perennial Dez river, whose water supply is primarily derived from snow and rainfall in the Zagros Mountain range. The Dez River, with a total length of 415 km (Mahmoodi et al. 2021 ), flows into Khuzestan Province and passes through the city of Dezful, and finally joins the Karun River at Band-e Qir. In this study, the research area is located between 32°26′44″ N and 31°39′29″ N latitudes, 48°15′39″ E and 48°52′54″ E longitudes (Fig. 1 ). Historically, this region has been vulnerable to flooding and overflow during the monsoon rainfall season. 3. Materials and methods In this study, discharge data from the Dez dam outlet spanning the period 1998–2023 were utilized to conduct preliminary analyses for identifying flood event dates using Microsoft Excel. To assess downstream damage extent, Landsat 5, 7, and 8 satellite imagery were employed for flood monitoring and inundation mapping. Figure 2 illustrates the overall research framework, which integrates remote sensing techniques and machine learning algorithms to predict and assess flood-induced damage extent in the study area. Landsat 5, 7 and 8 images This study utilized Landsat 5, 7, and 8 satellite data obtained through the USGS website. Landsat 5 satellite carries seven spectral bands with 30 m resolution in the visible, near-infrared, and mid-infrared regions, plus one thermal band at 120 m resolution (Chander et al. 2009 ). Landsat 7 features eight spectral bands, including one 15 m panchromatic band, six 30 m multispectral bands, and one 60 m thermal band (Markham and Helder 2012 ). Landsat 8 is equipped with two advanced sensors: the Operational Land Imager (OLI) providing data in nine spectral bands (30 m resolution, with 15 m resolution for the panchromatic band), and the Thermal Infrared Sensor (TIRS) acquiring thermal data in two bands at 100 m resolution (Roy et al. 2014 ). The green, red, and near-infrared bands were employed to estimate flood damage identification parameters in the study area. Additionally, converting the bands to RGB format generates a standard color composite image, enabling enhanced discrimination of various land surface features (Fig. 3 ). ArcMap10.8 software was utilized for temporal analysis of flood event data. Statistical data Statistical analyses and simple linear regression machine learning modeling were performed using JMP Pro 16 software. Additionally, satellite image processing and classification were conducted using the SVM algorithm in ArcMap 10.8 software. Support Vector Machine (SVM) model SVM algorithm is a classification technique recognized as one of the most effective methods for categorization, prediction, and uncertainty detection. Unlike clustering algorithms, SVM belongs to supervised learning methods and operates through two distinct phases of training and testing (Statnikov et al. 2013 ). Since satellite images cannot be directly input into SVM for training and testing, the input raster must undergo preprocessing steps to modify the image for further processing. In this study, preprocessing for SVM implementation involved converting bands into a single RGB composite image, as illustrated in Fig. 3 . Effective classification was achieved by utilizing prominent colors in the multispectral image, where each ground feature is characterized by a distinct color representation. For land feature classification, SVM was applied based on ground surface color information, with color features serving as indicators of land use/cover types. To evaluate classification accuracy and generate flood susceptibility maps, eight distinct land classes were established, representing vegetation, soil, urban areas, and waterlogged zones. In the second phase, image processing converted the data from RGB color space, enabling the selection of training data (training and testing phases) for SVM classification. Monte Carlo method Monte Carlo method is a simulation process that employs theoretical models to create an artificial environment for estimating or predicting real-world system behavior. This synthetic environment may reflect either physical or virtual space, wherein researchers attempt to replicate and model the characteristics and responses of real systems. Depending on simulation objectives and implementation constraints, four following distinct simulation types can be identified: generative simulation, analytical/technical simulation, strategic simulation, and intuitive/cognitive simulation (Kawrakow, 2000 ). Generative simulation is typically employed when direct access to study variable data is unavailable or when obtained data are insufficient for analysis. This simulation approach is also applicable for predictive modeling using regression models with random components. Monte Carlo method, which falls under this category, simulates the final function through random number generation from probability distributions of variables (Milner, 1971 ). Simple Linear Regression (SLR) model SLR is a method that estimates the values of each independent variable from the predictor variable. Regression analysis enables the prediction of changes in the dependent variable through independent variables and determines the contribution of each independent variable in explaining the dependent variable. Regression is closely related to the correlation coefficient, meaning that calculating the correlation coefficient is necessary to perform regression. The stronger the correlation coefficient between variables, the more accurate the prediction will be. In a simple linear regression model, only two variables can be included. In this study, the output discharge and flood depth data are retrieved in the JMP Pro software environment and then fitted with theoretical regression functions. SLR model is expressed as follows: Y = α + βx (1) where, parameters Y , X , and αβ represent the dependent variable, independent variable, and regression coefficients, respectively. SLR model was implemented by calculating the model's slope and intercept. Using linear regression, the correlation and direction of the relationship between the two variables can be estimated. Main channel To estimate the damage level inflicted on agricultural lands, the outflow discharges of the Dez Dam under normal operational conditions were examined. The catchment area of the Dez river was calculated as 13.32 km² using ArcMap software and the SVM algorithm. In this study, the Dez river under normal conditions was considered the reference, and all surface waters resulting from flood events (excluding the reference river) were identified and analyzed as the damage area (Fig. 4 ). Statistical Data Evaluation The performance of the model was evaluated using the Mean Square Error (MSE), its standardized value (RMSE), and the correlation coefficient (R²) between the calculated and observed values. A lower MSE and RMSE (along with a higher R² value), indicate greater statistical accuracy of the applied model. The calculation equations for the above mentioned methods are as follows: \(\:MSE=\frac{1}{N}\sum\:_{i=1}^{N}({y}_{exp}^{i}-{{y}_{pre}^{i})}^{2}\) (2) \(\:RMSE=\sqrt{\frac{1}{N}\sum\:_{i=1}^{N}({y}_{exp}^{i}-{{y}_{pre}^{i})}^{2}}\) (3) \(\:{R}^{2}=1-\frac{\sum\:_{i=1}^{N}({y}_{exp}^{i}-{y}_{pre}^{i}{)}^{2}}{\sum\:_{i=1}^{N}({y}_{exp}^{i}-{y)}^{2}}\) (4) where, \(\:{y}_{exp}^{i}\) , \(\:{y}_{pre}^{i}\:\) and N represent the estimated value, measured value of the target variable, and number of data points, respectively. 4. Results and Discussion This study examined historical records to identify destructive flood hydrographs and assess damage levels, enabling flood prediction and preparedness for potential future events. Notably, approximately ten destructive flood hydrographs were identified, however, persistent cloud cover and/or sensor technical limitations prevented the utilization of their corresponding satellite imagery (Fig. 5 ). Among the most significant observations were the substantial discharge fluctuations, exhibiting considerable increases in some periods and marked decreases in others. These variations reflect major changes in precipitation patterns, runoff volumes, and emergency spillway gate operations within the study area. During these time intervals, several high-discharge events occurred, indicative of severe and destructive flooding. These events were systematically categorized and rigorously analyzed. The frequent occurrence of flood events during this period demonstrates that flooding constitutes a recurrent phenomenon in the region. Consequently, after thorough investigation of causative factors, the most practical mitigation strategies were implemented. Analysis of discharge variations over the available 25-year period (Fig. 5 ) enabled identification of destructive flood events, prompting detailed examination of Landsat 5, 7, and 8 satellite imagery. Following extensive analysis of the study area using remote sensing and SVM, results demonstrate (as evident in Fig. 6 ) that the predominant land use surrounding the research site consists primarily of agricultural lands. These areas exhibit high susceptibility to flood disasters, which may lead to substantial socioeconomic consequences in coming years (Sin-ampol et al. 2020). Accordingly, to demonstrate the significance of the subject, the study area was classified into eight distinct categories: urban areas, bare soil, rice paddies, wheat fields, barley fields, grasslands, shallow water, and deep water. Each panel (a-d) in Fig. 6 represents a distinct but methodologically consistent classification of the study area. The color scheme denotes land cover categories as follows: green indicates vegetation, red represents bare soil and uncultivated land, while blue or white corresponds to water bodies. By referencing the accompanying legend, the land cover classification of individual pixels can be accurately determined. Temporal comparison of these classified images facilitates systematic analysis of flood event occurrences. Each of the images (a, b, c, and d) in Fig. 6 corresponds to the dates of March 25, 1999; January 20, 2002; April 29, 2003; and April 17, 2019, respectively, representing the classification results of the study area obtained by SVM. In these images, each color denotes a specific class: green indicates vegetation, red represents bare soil and barren land, while blue or white signifies water. By comparing the classified images at different time points, the occurrence of floods over time can be analyzed. The comprehensive data presented in Table 1 outline the complete record of documented flood events in the study area. Based on Table 1 , it can be concluded that over time, the extent of inundation and flood-related damage has increased. This trend may indicate climate change, population growth, residential expansion in riverine areas, or alterations in water resource management. The reservoir inflow volume, resulting from rainfall and runoff, directly influences the volume of water released through the spillways. As inflow increases, pressure on the dam rises, necessitating greater water discharge. This situation can lead to higher outflow rates, potentially triggering extensive flooding and increased damage. Consequently, there is a direct correlation between outflow discharge, spillway operation duration, and the extent of flood-induced damage. Table 1 The most destructive floods area calculated using SVM. Date of flood occurrence Inundation area (Km 2 ) Extent of the main channel (Km 2 ) Flood damage area (Km 2 ) Mean discharge outflow from the dam (MCM) Number of days with dam spillway gates open 1999 41.18 32.13 9.05 149.53 10 2001 49.52 32.13 17.39 86.36 4 2002 57.86 32.13 25.73 102.89 9 2003 89.9 32.13 57.77 127.20 9 2016 26786 32.13 235.73 135.75 6 2018 311.49 32.13 279.36 90.33 14 2019 355.12 32.13 322.99 138.99 29 Synthetic data generation using Monte Carlo Simulation Due to limitations in accessing field data and insufficient data volume for precise statistical analysis, the Monte Carlo simulation method was employed to generate synthetic data. By defining appropriate statistical distributions for key variables, a set of synthetic data was produced to enable more accurate prediction of system behavior under various conditions. The resulting data served as a reliable substitute for real-world data in the prediction process. Following Monte Carlo simulation, which generated one thousand paired datasets of outflow discharge and corresponding flood damage, a linear regression model was reapplied to the synthetic data to examine the new statistical relationship between the variables (Fig. 7 ). This model, developed from combined datasets incorporating uncertainty and statistical noise, established a generalized relationship between dam outflow discharge and floodplain damage. The analysis not only provided enhanced understanding of damage trends under varying conditions but also highlighted the significant role of statistical simulation in improving predictive capabilities for flood management decision-making. Prediction using SLR model Timely and comprehensive flood reporting by disaster management specialists is essential for identifying and locating flood-affected areas, as well as for implementing mitigation measures to combat flood-related damages. To this end, cumulative flood extent mapping was conducted, and the collected data will facilitate more precise analysis of current conditions and necessary predictions for crisis management (Dewan, 2013 ). In this study, SLR model was employed to examine the relationship between dam outflow discharge and flood inundation extent. The model results presented in Fig. 8 revealed a statistically significant direct correlation between these two variables ( P-Value < 0.05 ). Furthermore, the predictive performance for future values can be calculated using SLR algorithm. The derived regression equation with calculated coefficients is as follows: \(\:Y=\:36.22365+0.068095*Discharge\) (5) Figure 9 presents the linear equation for Eq. 5, demonstrating that discharge rate is one of the most significant factors affecting flood-damaged area extent. The derived regression coefficient of approximately 0.068 indicates that for each unit increase (1 m³/s) in outflow discharge, the flood inundation area increases by an average of 0.068 km². These results confirm that dam outflow discharge is a primary determinant of flood damage extent in riverine areas. Regarding future damage predictions, when the research findings were applied to the study area, the likelihood of more severe flooding events in the future became evident. This necessitates meticulous planning and the implementation of effective preventive measures in water resource and infrastructure management (Maiwald and Schwarz, 2012). As shown in Table 2 , the damage prediction model demonstrated strong performance, with an MSE of 3040.083 and RMSE of 55.136. These metrics indicate that the mean squared error (MSE) remains at a relatively low level, while the root mean squared error (RMSE) remains below 60 units. These values suggest substantial agreement between model predictions and actual observations, with prediction errors fluctuating within a relatively narrow range. Furthermore, the model achieved a coefficient of determination (R²) of 0.628, indicating that approximately 62.8% of the variation in damage values can be explained by the input variables. This moderate R² value, significantly above the random noise threshold (typically considered around 0.5), confirms that the model successfully captures a substantial portion of the actual damage variability. The combination of relatively low MSE and RMSE values along with a positive moderate R² demonstrates the model's satisfactory accuracy and reliability for damage prediction. These results confirm that the proposed framework performs adequately for predictive applications under the studied conditions. As illustrated in Fig. 10 , the sensitivity analysis reveals that dam outflow discharge exerts the most significant influence on damage extent prediction. The random, patternless distribution of errors around the zero axis indicates the absence of systematic bias or structural flaws in the model. Additionally, the actual vs. predicted values plot demonstrates strong correlation between model outputs and observed data. This visual analysis, combined with the statistical metrics (MSE, RMSE, and R²), provides compelling evidence for the model's reliability in predicting flood-inundated areas. Table 2 Accuracy of simple linear regression (SLR) model. MSE RMSE R² 3040.083 55.136 0.628 Classification Accuracy Assessment The kappa coefficient, as one of the classical performance metrics in remote sensing classification, is employed to evaluate the agreement between the classified data and reference data (Table 3 ). This coefficient adjusts the obtained accuracy by accounting for the probability of random agreement, thereby weighting the accuracy measures (Foody, 2020 ). However, Foody ( 2020 ) argues that the kappa coefficient may potentially misrepresent classification accuracy due to its consideration of random agreement, and could consequently be misleading when comparing different classification methods. Table 3 Confusion matrix for SVM classification. Year City Wheat Barley Rice Plain Land Water Deep water Total Kappa 1999 0 0 4 1 6 5 3 1 20 0.69419 2002 0 0 3 0 0 7 5 5 20 0.66887 2003 2 1 1 6 2 5 3 0 20 0.69697 2019 0 0 5 4 1 7 2 1 20 0.75460 Table 3 presents the results of confusion matrix over different years, indicating that the employed SVM model achieved acceptable accuracy in detecting and distinguishing various classes. The obtained kappa coefficient values (0.694, 0.669, 0.697, and 0.755) suggest a moderate to good agreement between the model's predictions and the ground truth data. On the other hand, the distribution of samples across different classes remained relatively consistent in terms of structural diversity over the years. However, certain classes, such as the fourth, sixth, and seventh classes, accounted for a larger proportion of samples in most years. This pattern may reflect changes in land use, hydrological patterns, or the expansion of urban and flood-prone areas during the study period. Overall, the analysis of kappa statistics and class composition demonstrates that the selected model successfully differentiated various classes with reasonable accuracy across different time intervals. Thus, it can serve as an effective tool for spatiotemporal change analysis, particularly in water resources and environmental hazard studies. In conclusion, the results confirm that SVM model is suitable for classifying such datasets and can extract reliable classifications. 5. Conclusion Floods, as one of the most significant natural hazards, pose a serious threat to lives and property, particularly in riverine areas. In this study, remote sensing and GIS technologies were employed to rapidly analyze and identify flood-vulnerable zones. The results demonstrate that machine learning algorithms, particularly linear regression and support vector machine (SVM) models can play a pivotal role in predicting flood-prone areas through the analysis of environmental data and satellite imagery. SVM algorithm, with its high capability in detecting complex boundaries between different classes, exhibited remarkable accuracy in image classification and analysis. Furthermore, the analyses revealed that increased outflow discharge from dams is directly correlated with the expansion of flood-affected areas and the occurrence of overflow events. The flood susceptibility map generated in this study, along with the predictive algorithm for water inundation levels, can serve as an effective tool for disaster management authorities, decision-makers, and water resource engineers. These tools not only enhance preparedness for future hazards but also facilitate the implementation of risk mitigation strategies and effective flood management planning. The findings of this study highlight the potential to support policy-making and operational measures aimed at reducing the impacts of current and future flood events. Declarations Author contributions Zareie Sajad: Conceptualization, Methodology, Validation, Formal Analysis, Writing, Review and Supervision, Final Approval for Manuscript Submission. Abaforoushan Mohammad: Conceptualization, Methodology, Formal Analysis, Writing, Visualization. Farhadiyan Mohammad: Conceptualization, Visualization, Writing – Review and Editing. The authors approved the final manuscript. Funding: No funding was obtained to conduct this study. Data availability: The datasets generated and analyzed in this study are available from the corresponding author upon reasonable request. Ethics approval: Not applicable. Consent to participate: Not applicable. Consent for publication: Not applicable. Competing interests: The authors declare no competing interests. References Asare-Kyei, Daniel, Gerald Forkuor, and Valentijn Venus. 2015. “Modeling Flood Hazard Zones at the Sub-District Level with the Rational Model Integrated with GIS and Remote Sensing Approaches.” Water 7(7):3531–64. doi: 10.3390/w7073531. Bonakdari, Hossein, Amir Hossein Zaji, Keyvan Soltani, and Bahram Gharabaghi. 2020. “Improving the Accuracy of a Remotely-Sensed Flood Warning System Using a Multi-Objective Pre-Processing Method for Signal Defects Detection and Elimination.” Comptes Rendus. Géoscience 352(1):73–86. Chander, Gyanesh, Brian L. Markham, and Dennis L. Helder. 2009. “Summary of Current Radiometric Calibration Coefficients for Landsat MSS, TM, ETM+, and EO-1 ALI Sensors.” Remote Sensing of Environment 113(5):893–903. doi: https://doi.org/10.1016/j.rse.2009.01.007. Chapi, Kamran, Vijay P. Singh, Ataollah Shirzadi, Himan Shahabi, Dieu Tien Bui, Binh Thai Pham, and Khabat Khosravi. 2017. “A Novel Hybrid Artificial Intelligence Approach for Flood Susceptibility Assessment.” Environmental Modelling & Software 95:229–45. doi: https://doi.org/10.1016/j.envsoft.2017.06.012. Cortes, Corinna, and Vladimir Vapnik. 1995. “Support-Vector Networks.” Machine Learning 20(3):273–97. doi: 10.1007/BF00994018. Dewan, Ashraf M. 2013. “Hazards, Risk, and Vulnerability.” Pp. 35–74 in Floods in a Megacity: Geospatial Techniques in Assessing Hazards, Risk and Vulnerability, edited by A. Dewan. Dordrecht: Springer Netherlands. Dewan, Ashraf M., M. Monirul Islam, T. Kumamoto, and M. Nishigaki. 2007. “Evaluating Flood Hazard for Land-Use Planning in Greater Dhaka of Bangladesh Using Remote Sensing and GIS Techniques.” Water Resources Management 21(9):1601–12. doi: 10.1007/s11269-006-9116-1. Foody, Giles M. 2020. “Explaining the Unsuitability of the Kappa Coefficient in the Assessment and Comparison of the Accuracy of Thematic Maps Obtained by Image Classification.” Remote Sensing of Environment 239:111630. doi: https://doi.org/10.1016/j.rse.2019.111630. Gudiyangada Nachappa, Thimmaiah, Sepideh Tavakkoli Piralilou, Khalil Gholamnia, Omid Ghorbanzadeh, Omid Rahmati, and Thomas Blaschke. 2020. “Flood Susceptibility Mapping with Machine Learning, Multi-Criteria Decision Analysis and Ensemble Using Dempster Shafer Theory.” Journal of Hydrology 590:125275. doi: https://doi.org/10.1016/j.jhydrol.2020.125275. Hajji S, Krimissa S, Abdelrahman K, Boudhar A, Elaloui A, Ismaili M, El Bouzekraoui M, Chikh Essbiti M, Kahal AY, Mondal BK and Namous M (2025) Enhancing flood prediction through remote sensing, machine learning, and Google Earth Engine. Front. Water. 7:1514047. doi: 10.3389/frwa.2025.1514047. Kawrakow, I. 2000. “Accurate Condensed History Monte Carlo Simulation of Electron Transport. II. Application to Ion Chamber Response Simulations.” Medical Physics 27(3):499–513. Kazemi, M., Mohammadi, F., Nafooti, M.H., Behvar K., Kariminejad N. (2024). Flood susceptibility mapping using machine learning and remote sensing data in the Southern Karun Basin, Iran. Appl Geomat 16, 731–750. https://doi.org/10.1007/s12518-024-00582-7. Mahmoodi, Nariman, Khaled Osati, Ali Salajegheh, and Mohsen Mohseni Saravi. (2021). “Trend in River Water Quality: Tracking the Overall Impacts of Climate Change and Human Activities on Water Quality in the Dez River Basin.” Journal of Water and Health 19(1):159–73. Maiwald, H., and J. Schwarz. 2012. “Damage and Loss Prediction Model Considering Inundation Level, Flow Velocity and Vulnerability of Building Types.” WIT Transactions on Ecology and the Environment 159:53–65. Markham, Brian L., and Dennis L. Helder. 2012. “Forty-Year Calibrated Record of Earth-Reflected Radiance from Landsat: A Review.” Remote Sensing of Environment 122:30–40. doi: https://doi.org/10.1016/j.rse.2011.06.026. Merz, Bruno, Günter Blöschl, Sergiy Vorogushyn, Francesco Dottori, Jeroen C. J. H. Aerts, Paul Bates, Miriam Bertola, Matthias Kemter, Heidi Kreibich, Upmanu Lall, and Elena Macdonald. 2021. “Causes, Impacts and Patterns of Disastrous River Floods.” Nature Reviews Earth & Environment 2(9):592–609. doi: 10.1038/s43017-021-00195-3. Milner, Robin. 1971. “An Algebraic Definition of Simulation between Programs.” Pradhan Biswajeet, Shafiee M, Pirasteh Saied, (2009). Maximum flood prone area mapping using RADARSAT images and GIS: Kelantan river basin. International journal of geoinformatics. 5 (2): 11. Radwan, Farid, A. A. Alazba, and Amr Mossad. 2019. “Flood Risk Assessment and Mapping Using AHP in Arid and Semiarid Regions.” Acta Geophysica 67(1):215–29. doi: 10.1007/s11600-018-0233-z. Rao, G. Nageswara. 2001. “Occurrence of Heavy Rainfall around the Confluence Line in Monsoon Disturbances and Its Importance in Causing Floods.” Journal of Earth System Science 110(1):87–94. doi: 10.1007/BF02702232. Roy, D. P., M. A. Wulder, T. R. Loveland, Woodcock C.E., R. G. Allen, M. C. Anderson, D. Helder, J. R. Irons, D. M. Johnson, R. Kennedy, T. A. Scambos, C. B. Schaaf, J. R. Schott, Y. Sheng, E. F. Vermote, A. S. Belward, R. Bindschadler, W. B. Cohen, F. Gao, J. D. Hipple, P. Hostert, J. Huntington, C. O. Justice, A. Kilic, V. Kovalskyy, Z. P. Lee, L. Lymburner, J. G. Masek, J. McCorkel, Y. Shuai, R. Trezza, J. Vogelmann, R. H. Wynne, and Z. Zhu. 2014. “Landsat-8: Science and Product Vision for Terrestrial Global Change Research.” Remote Sensing of Environment 145:154–72. doi: https://doi.org/10.1016/j.rse.2014.02.001. Samadi, A., S. S. Sadrolashrafi, and M. K. Kholghi. 2019. “Development and Testing of a Rainfall-Runoff Model for Flood Simulation in Dry Mountain Catchments: A Case Study for the Dez River Basin.” Physics and Chemistry of the Earth, Parts A/B/C 109:9–25. doi: https://doi.org/10.1016/j.pce.2018.07.003. Shahabi, Himan, Ataollah Shirzadi, Kayvan Ghaderi, Ebrahim Omidvar, Nadhir Al-Ansari, John J. Clague, Marten Geertsema, Khabat Khosravi, Ata Amini, Sepideh Bahrami, Omid Rahmati, Kyoumars Habibi, Ayub Mohammadi, Hoang Nguyen, Assefa M. Melesse, Baharin Bin Ahmad, and Anuar Ahmad. 2020. “Flood Detection and Susceptibility Mapping Using Sentinel-1 Remote Sensing Data and a Machine Learning Approach: Hybrid Intelligence of Bagging Ensemble Based on K-Nearest Neighbor Classifier.” Remote Sensing 12(2). doi: 10.3390/rs12020266. Sin-ampol, Phaothai, Tawee Chaipimonplin, and Supawadee Songka. 2020. “Local Community Engagement for Adaptation to Future Challenges in the Pilot Flood Detention Area of Thailand.” Pp. 203–28 in External Interventions for Disaster Risk Reduction: Impacts on Local Communities, edited by I. Chowdhooree and S. M. Ghani. Singapore: Springer Singapore. Statnikov, Alexander, Constantin F. Aliferis, Douglas P. Hardin, and Isabelle Guyon. 2013. Gentle Introduction to Support Vector Machines in Biomedicine, A-Volume 2: Case Studies and Benchmarks. World Scientific Publishing Company. Tehrany, Mahyat Shafapour, Biswajeet Pradhan, and Mustafa Neamah Jebur. (2014). “Flood Susceptibility Mapping Using a Novel Ensemble Weights-of-Evidence and Support Vector Machine Models in GIS.” Journal of Hydrology 512:332–43. doi: https://doi.org/10.1016/j.jhydrol.2014.03.008. Tien Bui, Dieu, Khabat Khosravi, Shaojun Li, Himan Shahabi, Mahdi Panahi, Vijay P. Singh, Kamran Chapi, Ataollah Shirzadi, Somayeh Panahi, and Wei Chen. 2018. “New Hybrids of Anfis with Several Optimization Algorithms for Flood Susceptibility Modeling.” Water 10(9):1210. Vaghefi, Saeid Ashraf, Malihe Keykhai, Farshid Jahanbakhshi, Jaleh Sheikholeslami, Azadeh Ahmadi, Hong Yang, and Karim C. Abbaspour. 2019. “The Future of Extreme Climate in Iran.” Scientific Reports 9(1):1464. doi: 10.1038/s41598-018-38071-8. Van Pham, T., Bui, D.X., Do, T.A.T. et al. Assessing flood susceptibility in Hanoi using machine learning and remote sensing: implications for urban health and resilience. Nat Hazards 121, 10149–10170 (2025). https://doi.org/10.1007/s11069-025-07211-5. Wang Wie, Gao Jie, Liu Zheng, Li Chuangi. 2023. A hybrid rainfall-runoff model: integrating initial loss and LSTM for improved forecasting. 11. https://doi.org/10.3389/fenvs.2023.1261239. Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 15 Oct, 2025 Reviewers invited by journal 12 Sep, 2025 Editor invited by journal 11 Sep, 2025 Editor assigned by journal 12 Jul, 2025 First submitted to journal 11 Jul, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7104599","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":513986340,"identity":"1fe82eb8-7e27-4101-b087-07e19e7d087b","order_by":0,"name":"Sajad Zareie","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAxklEQVRIiWNgGAWjYPACCQZ+GJONaC2SDSRqYWAwOECsSvkG9oefKyos5IzPH34mwVBjx8AnTUAz0PBkyTNnJIzNbqSZSTAcS2Zg40sgoIWB4YBkY5tE4rYbPGwSDGwHGNh4CDqMsfln4z+J+s39Z4Ba/hGhheEAM5tkY4NEggFDDpsEYxsRWgwOs7FZNhyTMJxxI83YIrEvmYeww9rbH99sqKmT5+8//PDGh292cvI9hBzGjMxJYGAg6JNRMApGwSgYBUQAAOu5M87C4o1SAAAAAElFTkSuQmCC","orcid":"","institution":"Shahid Chamran University of Ahvaz","correspondingAuthor":true,"prefix":"","firstName":"Sajad","middleName":"","lastName":"Zareie","suffix":""},{"id":513986341,"identity":"652a5c2b-81a7-4d29-8434-f08b1c9a17cc","order_by":1,"name":"Mohammad Abaforoushan","email":"","orcid":"","institution":"Shahid Chamran University of Ahvaz","correspondingAuthor":false,"prefix":"","firstName":"Mohammad","middleName":"","lastName":"Abaforoushan","suffix":""},{"id":513986342,"identity":"1f1d3ea9-a35a-4298-b744-4a7ab05648ca","order_by":2,"name":"Mohammad Farhadiyan","email":"","orcid":"","institution":"Shahid Chamran University of Ahvaz","correspondingAuthor":false,"prefix":"","firstName":"Mohammad","middleName":"","lastName":"Farhadiyan","suffix":""}],"badges":[],"createdAt":"2025-07-11 21:41:41","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7104599/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7104599/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":91701187,"identity":"4b8b845c-f34b-4f52-9c8c-7f7806762586","added_by":"auto","created_at":"2025-09-19 10:37:34","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":530391,"visible":true,"origin":"","legend":"\u003cp\u003eLocation of Study area.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7104599/v1/246e2e8b66e47de13f65417c.png"},{"id":91700840,"identity":"082beb67-e808-42c6-ae64-8ce8ed27e195","added_by":"auto","created_at":"2025-09-19 10:29:34","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":113000,"visible":true,"origin":"","legend":"\u003cp\u003eFlowchart of the study methodology.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7104599/v1/34b68f9d40c7f248981f5a43.png"},{"id":91700838,"identity":"22874450-7c42-4d1a-83bb-3fe95c0e02f5","added_by":"auto","created_at":"2025-09-19 10:29:34","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":283148,"visible":true,"origin":"","legend":"\u003cp\u003eRGB Composite of the Study Area.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7104599/v1/afb4c5fcc585e029194ca37f.png"},{"id":91701189,"identity":"1a3c468f-6d2c-4f8d-8081-5680dc914291","added_by":"auto","created_at":"2025-09-19 10:37:34","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":282295,"visible":true,"origin":"","legend":"\u003cp\u003eThe status and extent of Dez river under normal conditions.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7104599/v1/c201b9a7132a4e21c96fa486.png"},{"id":91701188,"identity":"ce4106ab-b5f9-4057-8f2d-07bdcc99db4d","added_by":"auto","created_at":"2025-09-19 10:37:34","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":28886,"visible":true,"origin":"","legend":"\u003cp\u003eWater discharge outflow from Dez Dam reservoir.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-7104599/v1/48ac093bdb4a02744d15a97b.png"},{"id":91700844,"identity":"7dfb4e4c-86da-47b7-9352-6257c023a58f","added_by":"auto","created_at":"2025-09-19 10:29:34","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":345270,"visible":true,"origin":"","legend":"\u003cp\u003eThe study area classified by SVM.\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-7104599/v1/718fcd23d2944b33d324dc30.png"},{"id":91703177,"identity":"4f688265-4bd1-4d50-b892-a193364a21cb","added_by":"auto","created_at":"2025-09-19 10:53:37","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":192402,"visible":true,"origin":"","legend":"\u003cp\u003eA comparison between original and Monte Carlo-Simulated data of discharge and floodplain.\u003c/p\u003e","description":"","filename":"7.png","url":"https://assets-eu.researchsquare.com/files/rs-7104599/v1/5a2a46f7aeb572481ae87db0.png"},{"id":91702273,"identity":"ebdbfa46-f68c-47d7-8284-517227902ec8","added_by":"auto","created_at":"2025-09-19 10:45:34","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":138973,"visible":true,"origin":"","legend":"\u003cp\u003eCorrelation between independent and dependent data.\u003c/p\u003e","description":"","filename":"8.png","url":"https://assets-eu.researchsquare.com/files/rs-7104599/v1/d8e2b36938a6f47709b5c4b3.png"},{"id":91701192,"identity":"3ad3e516-529c-4804-af5c-107433da357f","added_by":"auto","created_at":"2025-09-19 10:37:34","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":83146,"visible":true,"origin":"","legend":"\u003cp\u003eSimple linear regression (SLR) model.\u003c/p\u003e","description":"","filename":"9.png","url":"https://assets-eu.researchsquare.com/files/rs-7104599/v1/484fdb41ba448a06230a6c16.png"},{"id":91703190,"identity":"d8cc72b1-42d1-43e2-96af-664620786527","added_by":"auto","created_at":"2025-09-19 10:53:39","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":162171,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of actual and predicted floodplain values with residual plot.\u003c/p\u003e","description":"","filename":"10.png","url":"https://assets-eu.researchsquare.com/files/rs-7104599/v1/133fb105a8f3ca03a7faf491.png"},{"id":91817444,"identity":"43f97d70-64c3-423d-ba60-dbcb89a92669","added_by":"auto","created_at":"2025-09-22 06:56:22","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2743067,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7104599/v1/f19122de-6577-4393-ad29-971e71dcc848.pdf"}],"financialInterests":"","formattedTitle":"Evaluating the Performance of Dez Dam Reservoir Operation in Flood Control Using Machine Learning Algorithms and Remote Sensing","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eFloods are among the most widespread and destructive natural disasters, inflicting heavy annual losses in human societies and economic damage on communities. Furthermore, climate change and human activities have exacerbated the intensity and frequency of flood events (Merz et al. \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Catastrophic flood events arise from multiple causative factors, including both anthropogenic and natural sources (Rao \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2001\u003c/span\u003e). Given the extensive and long-term impacts of flooding, prioritizing preventive risk reduction measures such as developing accurate flood prediction models and improving land use planning in vulnerable areas, is of critical importance (Bonakdari et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Chapi et al. \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Tien Bui et al. 2018). The Dez river Basin, as one of Iran's most flood-prone areas, has consistently faced threats from sudden and devastating floods. Recent flood events in this region have further highlighted the critical need for comprehensive and accurate studies to assess flood risks and identify vulnerable zones (Samadi et al. \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Integration of GIS and Remote Sensing data with other datasets offers significant potential in modern technology for flood disaster identification, monitoring, and assessment (Pradhan et al., 2009). Data-driven models serve as crucial alternatives for establishing relationships between input and output data without requiring precise understanding of underlying physical processes (Wang et al., 2023). To evaluate flood impact on socio-economic damages in the study area, remote sensing methods were implemented to determine flood magnitude. Traditional flood risk assessment methods primarily rely on hydrological and hydraulic data. While these approaches provide valuable information, they suffer from several limitations, including high input data requirements, substantial costs, and time-consuming processes. Recent advances in remote sensing technologies have enabled access to high-precision satellite data with extensive coverage. These datasets, combined with other spatial data such as topographic maps, land use classifications, and river network maps, can be effectively utilized to produce high-accuracy flood hazard zonation maps (Radwan et al. 2019). Remote sensing and GIS technologies enable the collection and analysis of large volumes of data, including topographic, hydrological, and climatic variables, which are also highly useful for identifying flood-prone areas. Remote sensing and GIS play a significant role in generating flood susceptibility maps and enhancing our understanding of flood-vulnerable regions (Tehrany, Pradhan, and Jebur 2014).\u003c/p\u003e\u003cp\u003eKazemi et al. (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) investigated flood susceptibility mapping using machine learning and remote sensing. They used the machine learning techniques of MARS, CART, TreeNet, and RF. Based on the results, the TreeNet technique demonstrated the most promising performance among the machine learning algorithms. Hajji et al. (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2025\u003c/span\u003e) conducted a comparative evaluation of three machine learning algorithms of gradient boosting, AdaBoost, and random forest for flood prediction. Their analysis revealed that the random forest algorithm exhibited superior predictive accuracy, outperforming both the gradient boosting ensemble and individual implementations of gradient boosting and AdaBoost. Pham et al. (\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2025\u003c/span\u003e) evaluated flood susceptibility through an integrated approach combining machine learning and remote sensing techniques. Their results indicated that the Genetic Algorithm-optimized Artificial Neural Network model achieved optimal performance metrics (RMSE\u0026thinsp;=\u0026thinsp;4.332 and MAE\u0026thinsp;=\u0026thinsp;4.020). The derived flood susceptibility map revealed significant spatial variations, with particularly pronounced contrasts between urban and suburban regions. By leveraging advanced remote sensing and GIS technologies, and machine learning algorithms, Asare-Kyei et al. (2015) developed a comprehensive model for flood hazard zonation. Their findings demonstrated that up-to-date and accurate maps of high- and medium-risk flood zones can be utilized in managerial decision-making to mitigate the devastating impacts of floods. Tehrany et al. (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2014\u003c/span\u003e) demonstrated that integrating topographic data, land use, precipitation, river networks, and other parameters within a GIS environment combined with RS data can generate highly accurate flood susceptibility maps. Furthermore, Kazemi et al. (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) have shown that advanced TreeNet and CART models exhibit high performance in classifying flood-prone areas in the Karun River basin and can be effectively utilized to produce flood susceptibility maps with appropriate accuracy. Through the analysis of hydrological data, Vaghefi et al. (\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) concluded that the likelihood of flooding in various regions of Iran follows an increasing trend. And, their findings indicate that preventive approaches based on data analysis, particularly through the integration of remote sensing technology and machine learning algorithms, can significantly enhance flood risk mitigation efforts. Although floods cannot be entirely prevented, adopting an integrated flood management approach can minimize their adverse consequences (Dewan et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2007\u003c/span\u003e). Studies conducted by Gudiyangada Nachappa et al. (2020) and Shahabi et al. (2020) have contributed significantly to improving flood mitigation planning and enhancing awareness of at-risk areas. By integrating remote sensing technology with machine learning algorithms, these studies have introduced novel perspectives in flood management. Given the escalating frequency of flood events and their substantial human and economic impacts, the development of precise, innovative, and efficient methods for flood risk assessment and crisis management has become an undeniable necessity. Within this context, identifying flood-vulnerable zones has emerged as a key research challenge, receiving considerable attention in recent studies.\u003c/p\u003e\u003cp\u003eThe innovation of this research lies in the systematic integration of remote sensing and GIS technologies, along with the development of a simple linear regression machine learning algorithm in the JMP software environment, to assess flood damage levels. Additionally, SVM algorithm, one of the most accurate and efficient machine learning methods for classification (Cortes and Vapnik, 1995), was employed to determine flood damage levels in the study area. The significance of this study is its contribution to enhancing awareness of at-risk areas and providing actionable data to support decision-making in spatial planning, integrated resource management, and the mitigation of potential hazards to human communities, infrastructure, and the environment.\u003c/p\u003e"},{"header":"2. Case study","content":"\u003cp\u003ePrevious studies indicate that flood occurrences in various regions of Iran have caused significant damage to the agricultural and livestock sectors, severely affecting the livelihoods of many rural communities. Consequently, preventive measures to mitigate these risks must be prioritized as a critical issue (Vaghefi et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Accordingly, the Dez river basin up to its confluence with the Karun river in the Band-e Qir region was selected as the study area due to the occurrence of a major flood in 2020. The Dez river is one of the most important rivers in southwestern Iran, formed by the confluence of two main tributaries, the Sezar and Bakhtiari rivers, in Lorestan Province. The merging of these two branches creates the perennial Dez river, whose water supply is primarily derived from snow and rainfall in the Zagros Mountain range. The Dez River, with a total length of 415 km (Mahmoodi et al. \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), flows into Khuzestan Province and passes through the city of Dezful, and finally joins the Karun River at Band-e Qir. In this study, the research area is located between 32\u0026deg;26\u0026prime;44\u0026Prime; N and 31\u0026deg;39\u0026prime;29\u0026Prime; N latitudes, 48\u0026deg;15\u0026prime;39\u0026Prime; E and 48\u0026deg;52\u0026prime;54\u0026Prime; E longitudes (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Historically, this region has been vulnerable to flooding and overflow during the monsoon rainfall season.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"3. Materials and methods","content":"\u003cp\u003eIn this study, discharge data from the Dez dam outlet spanning the period 1998\u0026ndash;2023 were utilized to conduct preliminary analyses for identifying flood event dates using Microsoft Excel. To assess downstream damage extent, Landsat 5, 7, and 8 satellite imagery were employed for flood monitoring and inundation mapping. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e illustrates the overall research framework, which integrates remote sensing techniques and machine learning algorithms to predict and assess flood-induced damage extent in the study area.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eLandsat 5, 7 and 8 images\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThis study utilized Landsat 5, 7, and 8 satellite data obtained through the USGS website. Landsat 5 satellite carries seven spectral bands with 30 m resolution in the visible, near-infrared, and mid-infrared regions, plus one thermal band at 120 m resolution (Chander et al. \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2009\u003c/span\u003e). Landsat 7 features eight spectral bands, including one 15 m panchromatic band, six 30 m multispectral bands, and one 60 m thermal band (Markham and Helder \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Landsat 8 is equipped with two advanced sensors: the Operational Land Imager (OLI) providing data in nine spectral bands (30 m resolution, with 15 m resolution for the panchromatic band), and the Thermal Infrared Sensor (TIRS) acquiring thermal data in two bands at 100 m resolution (Roy et al. \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe green, red, and near-infrared bands were employed to estimate flood damage identification parameters in the study area. Additionally, converting the bands to RGB format generates a standard color composite image, enabling enhanced discrimination of various land surface features (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). ArcMap10.8 software was utilized for temporal analysis of flood event data.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eStatistical data\u003c/b\u003e\u003c/p\u003e\u003cp\u003eStatistical analyses and simple linear regression machine learning modeling were performed using JMP Pro 16 software. Additionally, satellite image processing and classification were conducted using the SVM algorithm in ArcMap 10.8 software.\u003c/p\u003e\u003cp\u003e\u003cb\u003eSupport Vector Machine (SVM) model\u003c/b\u003e\u003c/p\u003e\u003cp\u003eSVM algorithm is a classification technique recognized as one of the most effective methods for categorization, prediction, and uncertainty detection. Unlike clustering algorithms, SVM belongs to supervised learning methods and operates through two distinct phases of training and testing (Statnikov et al. \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Since satellite images cannot be directly input into SVM for training and testing, the input raster must undergo preprocessing steps to modify the image for further processing. In this study, preprocessing for SVM implementation involved converting bands into a single RGB composite image, as illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. Effective classification was achieved by utilizing prominent colors in the multispectral image, where each ground feature is characterized by a distinct color representation. For land feature classification, SVM was applied based on ground surface color information, with color features serving as indicators of land use/cover types. To evaluate classification accuracy and generate flood susceptibility maps, eight distinct land classes were established, representing vegetation, soil, urban areas, and waterlogged zones. In the second phase, image processing converted the data from RGB color space, enabling the selection of training data (training and testing phases) for SVM classification.\u003c/p\u003e\u003cp\u003e\u003cb\u003eMonte Carlo method\u003c/b\u003e\u003c/p\u003e\u003cp\u003eMonte Carlo method is a simulation process that employs theoretical models to create an artificial environment for estimating or predicting real-world system behavior. This synthetic environment may reflect either physical or virtual space, wherein researchers attempt to replicate and model the characteristics and responses of real systems. Depending on simulation objectives and implementation constraints, four following distinct simulation types can be identified: generative simulation, analytical/technical simulation, strategic simulation, and intuitive/cognitive simulation (Kawrakow, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2000\u003c/span\u003e). Generative simulation is typically employed when direct access to study variable data is unavailable or when obtained data are insufficient for analysis. This simulation approach is also applicable for predictive modeling using regression models with random components. Monte Carlo method, which falls under this category, simulates the final function through random number generation from probability distributions of variables (Milner, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e1971\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003cb\u003eSimple Linear Regression (SLR) model\u003c/b\u003e\u003c/p\u003e\u003cp\u003eSLR is a method that estimates the values of each independent variable from the predictor variable. Regression analysis enables the prediction of changes in the dependent variable through independent variables and determines the contribution of each independent variable in explaining the dependent variable. Regression is closely related to the correlation coefficient, meaning that calculating the correlation coefficient is necessary to perform regression. The stronger the correlation coefficient between variables, the more accurate the prediction will be. In a simple linear regression model, only two variables can be included. In this study, the output discharge and flood depth data are retrieved in the JMP Pro software environment and then fitted with theoretical regression functions. SLR model is expressed as follows:\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"No\" id=\"Taba\" border=\"1\"\u003e\u003ccolgroup cols=\"2\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cem\u003eY\u0026thinsp;=\u0026thinsp;α\u0026thinsp;+\u0026thinsp;βx\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e(1)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003ewhere, parameters \u003cem\u003eY\u003c/em\u003e, \u003cem\u003eX\u003c/em\u003e, and \u003cem\u003eαβ\u003c/em\u003e represent the dependent variable, independent variable, and regression coefficients, respectively. SLR model was implemented by calculating the model's slope and intercept. Using linear regression, the correlation and direction of the relationship between the two variables can be estimated.\u003c/p\u003e\u003cp\u003e\u003cb\u003eMain channel\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTo estimate the damage level inflicted on agricultural lands, the outflow discharges of the Dez Dam under normal operational conditions were examined. The catchment area of the Dez river was calculated as 13.32 km\u0026sup2; using ArcMap software and the SVM algorithm. In this study, the Dez river under normal conditions was considered the reference, and all surface waters resulting from flood events (excluding the reference river) were identified and analyzed as the damage area (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eStatistical Data Evaluation\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe performance of the model was evaluated using the Mean Square Error (MSE), its standardized value (RMSE), and the correlation coefficient (R\u0026sup2;) between the calculated and observed values. A lower MSE and RMSE (along with a higher R\u0026sup2; value), indicate greater statistical accuracy of the applied model.\u003c/p\u003e\u003cp\u003eThe calculation equations for the above mentioned methods are as follows:\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"No\" id=\"Tabb\" border=\"1\"\u003e\u003ccolgroup cols=\"2\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:MSE=\\frac{1}{N}\\sum\\:_{i=1}^{N}({y}_{exp}^{i}-{{y}_{pre}^{i})}^{2}\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003e(2)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:RMSE=\\sqrt{\\frac{1}{N}\\sum\\:_{i=1}^{N}({y}_{exp}^{i}-{{y}_{pre}^{i})}^{2}}\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e(3)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{R}^{2}=1-\\frac{\\sum\\:_{i=1}^{N}({y}_{exp}^{i}-{y}_{pre}^{i}{)}^{2}}{\\sum\\:_{i=1}^{N}({y}_{exp}^{i}-{y)}^{2}}\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e(4)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003ewhere, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{y}_{exp}^{i}\\)\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{y}_{pre}^{i}\\:\\)\u003c/span\u003e\u003c/span\u003eand N represent the estimated value, measured value of the target variable, and number of data points, respectively.\u003c/p\u003e"},{"header":"4. Results and Discussion","content":"\u003cp\u003eThis study examined historical records to identify destructive flood hydrographs and assess damage levels, enabling flood prediction and preparedness for potential future events. Notably, approximately ten destructive flood hydrographs were identified, however, persistent cloud cover and/or sensor technical limitations prevented the utilization of their corresponding satellite imagery (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eAmong the most significant observations were the substantial discharge fluctuations, exhibiting considerable increases in some periods and marked decreases in others. These variations reflect major changes in precipitation patterns, runoff volumes, and emergency spillway gate operations within the study area. During these time intervals, several high-discharge events occurred, indicative of severe and destructive flooding. These events were systematically categorized and rigorously analyzed.\u003c/p\u003e\u003cp\u003eThe frequent occurrence of flood events during this period demonstrates that flooding constitutes a recurrent phenomenon in the region. Consequently, after thorough investigation of causative factors, the most practical mitigation strategies were implemented. Analysis of discharge variations over the available 25-year period (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e) enabled identification of destructive flood events, prompting detailed examination of Landsat 5, 7, and 8 satellite imagery.\u003c/p\u003e\u003cp\u003eFollowing extensive analysis of the study area using remote sensing and SVM, results demonstrate (as evident in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e) that the predominant land use surrounding the research site consists primarily of agricultural lands. These areas exhibit high susceptibility to flood disasters, which may lead to substantial socioeconomic consequences in coming years (Sin-ampol et al. 2020). Accordingly, to demonstrate the significance of the subject, the study area was classified into eight distinct categories: urban areas, bare soil, rice paddies, wheat fields, barley fields, grasslands, shallow water, and deep water.\u003c/p\u003e\u003cp\u003eEach panel (a-d) in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e represents a distinct but methodologically consistent classification of the study area. The color scheme denotes land cover categories as follows: green indicates vegetation, red represents bare soil and uncultivated land, while blue or white corresponds to water bodies. By referencing the accompanying legend, the land cover classification of individual pixels can be accurately determined. Temporal comparison of these classified images facilitates systematic analysis of flood event occurrences.\u003c/p\u003e\u003cp\u003eEach of the images (a, b, c, and d) in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e corresponds to the dates of March 25, 1999; January 20, 2002; April 29, 2003; and April 17, 2019, respectively, representing the classification results of the study area obtained by SVM. In these images, each color denotes a specific class: green indicates vegetation, red represents bare soil and barren land, while blue or white signifies water. By comparing the classified images at different time points, the occurrence of floods over time can be analyzed.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe comprehensive data presented in Table \u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e outline the complete record of documented flood events in the study area. Based on Table \u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, it can be concluded that over time, the extent of inundation and flood-related damage has increased. This trend may indicate climate change, population growth, residential expansion in riverine areas, or alterations in water resource management. The reservoir inflow volume, resulting from rainfall and runoff, directly influences the volume of water released through the spillways. As inflow increases, pressure on the dam rises, necessitating greater water discharge. This situation can lead to higher outflow rates, potentially triggering extensive flooding and increased damage. Consequently, there is a direct correlation between outflow discharge, spillway operation duration, and the extent of flood-induced damage.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eThe most destructive floods area calculated using SVM.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"6\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDate of flood occurrence\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eInundation area (Km\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eExtent of the main channel (Km\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eFlood damage area (Km\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eMean discharge outflow from the dam (MCM)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eNumber of days with dam spillway gates open\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e1999\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e41.18\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e32.13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e9.05\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e149.53\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e10\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e49.52\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e32.13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e17.39\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e86.36\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2002\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e57.86\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e32.13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e25.73\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e102.89\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e9\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2003\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e89.9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e32.13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e57.77\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e127.20\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e9\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2016\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e26786\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e32.13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e235.73\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e135.75\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e6\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2018\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e311.49\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e32.13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e279.36\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e90.33\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e14\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2019\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e355.12\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e32.13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e322.99\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e138.99\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e29\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eSynthetic data generation using Monte Carlo Simulation\u003c/b\u003e\u003c/p\u003e\u003cp\u003eDue to limitations in accessing field data and insufficient data volume for precise statistical analysis, the Monte Carlo simulation method was employed to generate synthetic data. By defining appropriate statistical distributions for key variables, a set of synthetic data was produced to enable more accurate prediction of system behavior under various conditions. The resulting data served as a reliable substitute for real-world data in the prediction process.\u003c/p\u003e\u003cp\u003eFollowing Monte Carlo simulation, which generated one thousand paired datasets of outflow discharge and corresponding flood damage, a linear regression model was reapplied to the synthetic data to examine the new statistical relationship between the variables (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThis model, developed from combined datasets incorporating uncertainty and statistical noise, established a generalized relationship between dam outflow discharge and floodplain damage. The analysis not only provided enhanced understanding of damage trends under varying conditions but also highlighted the significant role of statistical simulation in improving predictive capabilities for flood management decision-making.\u003c/p\u003e\u003cp\u003e\u003cb\u003ePrediction using SLR model\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTimely and comprehensive flood reporting by disaster management specialists is essential for identifying and locating flood-affected areas, as well as for implementing mitigation measures to combat flood-related damages. To this end, cumulative flood extent mapping was conducted, and the collected data will facilitate more precise analysis of current conditions and necessary predictions for crisis management (Dewan, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2013\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eIn this study, SLR model was employed to examine the relationship between dam outflow discharge and flood inundation extent. The model results presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e revealed a statistically significant direct correlation between these two variables (\u003cem\u003eP-Value\u0026thinsp;\u0026lt;\u0026thinsp;0.05\u003c/em\u003e). Furthermore, the predictive performance for future values can be calculated using SLR algorithm. The derived regression equation with calculated coefficients is as follows:\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"No\" id=\"Tabc\" border=\"1\"\u003e\u003ccolgroup cols=\"2\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:Y=\\:36.22365+0.068095*Discharge\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e(5)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003e presents the linear equation for Eq.\u0026nbsp;5, demonstrating that discharge rate is one of the most significant factors affecting flood-damaged area extent. The derived regression coefficient of approximately 0.068 indicates that for each unit increase (1 m\u0026sup3;/s) in outflow discharge, the flood inundation area increases by an average of 0.068 km\u0026sup2;. These results confirm that dam outflow discharge is a primary determinant of flood damage extent in riverine areas.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eRegarding future damage predictions, when the research findings were applied to the study area, the likelihood of more severe flooding events in the future became evident. This necessitates meticulous planning and the implementation of effective preventive measures in water resource and infrastructure management (Maiwald and Schwarz, 2012).\u003c/p\u003e\u003cp\u003eAs shown in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, the damage prediction model demonstrated strong performance, with an MSE of 3040.083 and RMSE of 55.136. These metrics indicate that the mean squared error (MSE) remains at a relatively low level, while the root mean squared error (RMSE) remains below 60 units. These values suggest substantial agreement between model predictions and actual observations, with prediction errors fluctuating within a relatively narrow range.\u003c/p\u003e\u003cp\u003eFurthermore, the model achieved a coefficient of determination (R\u0026sup2;) of 0.628, indicating that approximately 62.8% of the variation in damage values can be explained by the input variables. This moderate R\u0026sup2; value, significantly above the random noise threshold (typically considered around 0.5), confirms that the model successfully captures a substantial portion of the actual damage variability.\u003c/p\u003e\u003cp\u003eThe combination of relatively low MSE and RMSE values along with a positive moderate R\u0026sup2; demonstrates the model's satisfactory accuracy and reliability for damage prediction. These results confirm that the proposed framework performs adequately for predictive applications under the studied conditions.\u003c/p\u003e\u003cp\u003eAs illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003e, the sensitivity analysis reveals that dam outflow discharge exerts the most significant influence on damage extent prediction. The random, patternless distribution of errors around the zero axis indicates the absence of systematic bias or structural flaws in the model. Additionally, the actual vs. predicted values plot demonstrates strong correlation between model outputs and observed data. This visual analysis, combined with the statistical metrics (MSE, RMSE, and R\u0026sup2;), provides compelling evidence for the model's reliability in predicting flood-inundated areas.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eAccuracy of simple linear regression (SLR) model.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"3\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMSE\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eRMSE\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eR\u0026sup2;\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e3040.083\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e55.136\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.628\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eClassification Accuracy Assessment\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe kappa coefficient, as one of the classical performance metrics in remote sensing classification, is employed to evaluate the agreement between the classified data and reference data (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). This coefficient adjusts the obtained accuracy by accounting for the probability of random agreement, thereby weighting the accuracy measures (Foody, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eHowever, Foody (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) argues that the kappa coefficient may potentially misrepresent classification accuracy due to its consideration of random agreement, and could consequently be misleading when comparing different classification methods.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eConfusion matrix for SVM classification.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"11\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eYear\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCity\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eWheat\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eBarley\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eRice\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003ePlain\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eLand\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c8\"\u003e\u003cp\u003eWater\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c9\"\u003e\u003cp\u003eDeep water\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c10\"\u003e\u003cp\u003eTotal\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c11\"\u003e\u003cp\u003eKappa\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e1999\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e\u003cp\u003e20\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.69419\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2002\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e\u003cp\u003e5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e\u003cp\u003e20\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.66887\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2003\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e\u003cp\u003e20\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.69697\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2019\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e\u003cp\u003e20\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.75460\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e presents the results of confusion matrix over different years, indicating that the employed SVM model achieved acceptable accuracy in detecting and distinguishing various classes. The obtained kappa coefficient values (0.694, 0.669, 0.697, and 0.755) suggest a moderate to good agreement between the model's predictions and the ground truth data. On the other hand, the distribution of samples across different classes remained relatively consistent in terms of structural diversity over the years. However, certain classes, such as the fourth, sixth, and seventh classes, accounted for a larger proportion of samples in most years. This pattern may reflect changes in land use, hydrological patterns, or the expansion of urban and flood-prone areas during the study period.\u003c/p\u003e\u003cp\u003eOverall, the analysis of kappa statistics and class composition demonstrates that the selected model successfully differentiated various classes with reasonable accuracy across different time intervals. Thus, it can serve as an effective tool for spatiotemporal change analysis, particularly in water resources and environmental hazard studies. In conclusion, the results confirm that SVM model is suitable for classifying such datasets and can extract reliable classifications.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eFloods, as one of the most significant natural hazards, pose a serious threat to lives and property, particularly in riverine areas. In this study, remote sensing and GIS technologies were employed to rapidly analyze and identify flood-vulnerable zones. The results demonstrate that machine learning algorithms, particularly linear regression and support vector machine (SVM) models can play a pivotal role in predicting flood-prone areas through the analysis of environmental data and satellite imagery. SVM algorithm, with its high capability in detecting complex boundaries between different classes, exhibited remarkable accuracy in image classification and analysis.\u003c/p\u003e\u003cp\u003eFurthermore, the analyses revealed that increased outflow discharge from dams is directly correlated with the expansion of flood-affected areas and the occurrence of overflow events. The flood susceptibility map generated in this study, along with the predictive algorithm for water inundation levels, can serve as an effective tool for disaster management authorities, decision-makers, and water resource engineers. These tools not only enhance preparedness for future hazards but also facilitate the implementation of risk mitigation strategies and effective flood management planning. The findings of this study highlight the potential to support policy-making and operational measures aimed at reducing the impacts of current and future flood events.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAuthor contributions\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eZareie Sajad: Conceptualization, Methodology, Validation, Formal Analysis, Writing, Review and Supervision, Final Approval for Manuscript Submission. Abaforoushan Mohammad: Conceptualization, Methodology, Formal Analysis, Writing, Visualization. Farhadiyan Mohammad: Conceptualization, Visualization, Writing \u0026ndash; Review and Editing. The authors approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u0026nbsp;\u003c/strong\u003eNo funding was obtained to conduct this study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability:\u0026nbsp;\u003c/strong\u003eThe datasets generated and analyzed in this study are available from the corresponding author upon reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval:\u003c/strong\u003e Not applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to participate:\u003c/strong\u003e Not applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication:\u003c/strong\u003e Not applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests:\u0026nbsp;\u003c/strong\u003eThe authors declare no competing interests.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAsare-Kyei, Daniel, Gerald Forkuor, and Valentijn Venus. 2015. \u0026ldquo;Modeling Flood Hazard Zones at the Sub-District Level with the Rational Model Integrated with GIS and Remote Sensing Approaches.\u0026rdquo; Water 7(7):3531\u0026ndash;64. doi: 10.3390/w7073531.\u003c/li\u003e\n\u003cli\u003eBonakdari, Hossein, Amir Hossein Zaji, Keyvan Soltani, and Bahram Gharabaghi. 2020. \u0026ldquo;Improving the Accuracy of a Remotely-Sensed Flood Warning System Using a Multi-Objective Pre-Processing Method for Signal Defects Detection and Elimination.\u0026rdquo; Comptes Rendus. G\u0026eacute;oscience 352(1):73\u0026ndash;86.\u003c/li\u003e\n\u003cli\u003eChander, Gyanesh, Brian L. Markham, and Dennis L. Helder. 2009. \u0026ldquo;Summary of Current Radiometric Calibration Coefficients for Landsat MSS, TM, ETM+, and EO-1 ALI Sensors.\u0026rdquo; Remote Sensing of Environment 113(5):893\u0026ndash;903. doi: https://doi.org/10.1016/j.rse.2009.01.007.\u003c/li\u003e\n\u003cli\u003eChapi, Kamran, Vijay P. Singh, Ataollah Shirzadi, Himan Shahabi, Dieu Tien Bui, Binh Thai Pham, and Khabat Khosravi. 2017. \u0026ldquo;A Novel Hybrid Artificial Intelligence Approach for Flood Susceptibility Assessment.\u0026rdquo; Environmental Modelling \u0026amp; Software 95:229\u0026ndash;45. doi: https://doi.org/10.1016/j.envsoft.2017.06.012.\u003c/li\u003e\n\u003cli\u003eCortes, Corinna, and Vladimir Vapnik. 1995. \u0026ldquo;Support-Vector Networks.\u0026rdquo; Machine Learning 20(3):273\u0026ndash;97. doi: 10.1007/BF00994018.\u003c/li\u003e\n\u003cli\u003eDewan, Ashraf M. 2013. \u0026ldquo;Hazards, Risk, and Vulnerability.\u0026rdquo; Pp. 35\u0026ndash;74 in Floods in a Megacity: Geospatial Techniques in Assessing Hazards, Risk and Vulnerability, edited by A. Dewan. Dordrecht: Springer Netherlands.\u003c/li\u003e\n\u003cli\u003eDewan, Ashraf M., M. Monirul Islam, T. Kumamoto, and M. Nishigaki. 2007. \u0026ldquo;Evaluating Flood Hazard for Land-Use Planning in Greater Dhaka of Bangladesh Using Remote Sensing and GIS Techniques.\u0026rdquo; Water Resources Management 21(9):1601\u0026ndash;12. doi: 10.1007/s11269-006-9116-1.\u003c/li\u003e\n\u003cli\u003eFoody, Giles M. 2020. \u0026ldquo;Explaining the Unsuitability of the Kappa Coefficient in the Assessment and Comparison of the Accuracy of Thematic Maps Obtained by Image Classification.\u0026rdquo; Remote Sensing of Environment 239:111630. doi: https://doi.org/10.1016/j.rse.2019.111630.\u003c/li\u003e\n\u003cli\u003eGudiyangada Nachappa, Thimmaiah, Sepideh Tavakkoli Piralilou, Khalil Gholamnia, Omid Ghorbanzadeh, Omid Rahmati, and Thomas Blaschke. 2020. \u0026ldquo;Flood Susceptibility Mapping with Machine Learning, Multi-Criteria Decision Analysis and Ensemble Using Dempster Shafer Theory.\u0026rdquo; Journal of Hydrology 590:125275. doi: https://doi.org/10.1016/j.jhydrol.2020.125275.\u003c/li\u003e\n\u003cli\u003eHajji S, Krimissa S, Abdelrahman K, Boudhar A, Elaloui A, Ismaili M, El Bouzekraoui M, Chikh Essbiti M, Kahal AY, Mondal BK and Namous M (2025) Enhancing flood prediction through remote sensing, machine learning, and Google Earth Engine. Front. Water. 7:1514047. doi: 10.3389/frwa.2025.1514047.\u003c/li\u003e\n\u003cli\u003eKawrakow, I. 2000. \u0026ldquo;Accurate Condensed History Monte Carlo Simulation of Electron Transport. II. Application to Ion Chamber Response Simulations.\u0026rdquo; Medical Physics 27(3):499\u0026ndash;513.\u003c/li\u003e\n\u003cli\u003eKazemi, M., Mohammadi, F., Nafooti, M.H., Behvar K., Kariminejad N. (2024). Flood susceptibility mapping using machine learning and remote sensing data in the Southern Karun Basin, Iran. Appl Geomat 16, 731\u0026ndash;750. https://doi.org/10.1007/s12518-024-00582-7.\u003c/li\u003e\n\u003cli\u003eMahmoodi, Nariman, Khaled Osati, Ali Salajegheh, and Mohsen Mohseni Saravi. (2021). \u0026ldquo;Trend in River Water Quality: Tracking the Overall Impacts of Climate Change and Human Activities on Water Quality in the Dez River Basin.\u0026rdquo; Journal of Water and Health 19(1):159\u0026ndash;73.\u003c/li\u003e\n\u003cli\u003eMaiwald, H., and J. Schwarz. 2012. \u0026ldquo;Damage and Loss Prediction Model Considering Inundation Level, Flow Velocity and Vulnerability of Building Types.\u0026rdquo; WIT Transactions on Ecology and the Environment 159:53\u0026ndash;65.\u003c/li\u003e\n\u003cli\u003eMarkham, Brian L., and Dennis L. Helder. 2012. \u0026ldquo;Forty-Year Calibrated Record of Earth-Reflected Radiance from Landsat: A Review.\u0026rdquo; Remote Sensing of Environment 122:30\u0026ndash;40. doi: https://doi.org/10.1016/j.rse.2011.06.026.\u003c/li\u003e\n\u003cli\u003eMerz, Bruno, G\u0026uuml;nter Bl\u0026ouml;schl, Sergiy Vorogushyn, Francesco Dottori, Jeroen C. J. H. Aerts, Paul Bates, Miriam Bertola, Matthias Kemter, Heidi Kreibich, Upmanu Lall, and Elena Macdonald. 2021. \u0026ldquo;Causes, Impacts and Patterns of Disastrous River Floods.\u0026rdquo; Nature Reviews Earth \u0026amp; Environment 2(9):592\u0026ndash;609. doi: 10.1038/s43017-021-00195-3.\u003c/li\u003e\n\u003cli\u003eMilner, Robin. 1971. \u0026ldquo;An Algebraic Definition of Simulation between Programs.\u0026rdquo;\u003c/li\u003e\n\u003cli\u003ePradhan Biswajeet, Shafiee M, Pirasteh Saied, (2009). Maximum flood prone area mapping using RADARSAT images and GIS: Kelantan river basin. International journal of geoinformatics. 5 (2): 11.\u003c/li\u003e\n\u003cli\u003eRadwan, Farid, A. A. Alazba, and Amr Mossad. 2019. \u0026ldquo;Flood Risk Assessment and Mapping Using AHP in Arid and Semiarid Regions.\u0026rdquo; Acta Geophysica 67(1):215\u0026ndash;29. doi: 10.1007/s11600-018-0233-z.\u003c/li\u003e\n\u003cli\u003eRao, G. Nageswara. 2001. \u0026ldquo;Occurrence of Heavy Rainfall around the Confluence Line in Monsoon Disturbances and Its Importance in Causing Floods.\u0026rdquo; Journal of Earth System Science 110(1):87\u0026ndash;94. doi: 10.1007/BF02702232.\u003c/li\u003e\n\u003cli\u003eRoy, D. P., M. A. Wulder, T. R. Loveland, Woodcock C.E., R. G. Allen, M. C. Anderson, D. Helder, J. R. Irons, D. M. Johnson, R. Kennedy, T. A. Scambos, C. B. Schaaf, J. R. Schott, Y. Sheng, E. F. Vermote, A. S. Belward, R. Bindschadler, W. B. Cohen, F. Gao, J. D. Hipple, P. Hostert, J. Huntington, C. O. Justice, A. Kilic, V. Kovalskyy, Z. P. Lee, L. Lymburner, J. G. Masek, J. McCorkel, Y. Shuai, R. Trezza, J. Vogelmann, R. H. Wynne, and Z. Zhu. 2014. \u0026ldquo;Landsat-8: Science and Product Vision for Terrestrial Global Change Research.\u0026rdquo; Remote Sensing of Environment 145:154\u0026ndash;72. doi: https://doi.org/10.1016/j.rse.2014.02.001.\u003c/li\u003e\n\u003cli\u003eSamadi, A., S. S. Sadrolashrafi, and M. K. Kholghi. 2019. \u0026ldquo;Development and Testing of a Rainfall-Runoff Model for Flood Simulation in Dry Mountain Catchments: A Case Study for the Dez River Basin.\u0026rdquo; Physics and Chemistry of the Earth, Parts A/B/C 109:9\u0026ndash;25. doi: https://doi.org/10.1016/j.pce.2018.07.003.\u003c/li\u003e\n\u003cli\u003eShahabi, Himan, Ataollah Shirzadi, Kayvan Ghaderi, Ebrahim Omidvar, Nadhir Al-Ansari, John J. Clague, Marten Geertsema, Khabat Khosravi, Ata Amini, Sepideh Bahrami, Omid Rahmati, Kyoumars Habibi, Ayub Mohammadi, Hoang Nguyen, Assefa M. Melesse, Baharin Bin Ahmad, and Anuar Ahmad. 2020. \u0026ldquo;Flood Detection and Susceptibility Mapping Using Sentinel-1 Remote Sensing Data and a Machine Learning Approach: Hybrid Intelligence of Bagging Ensemble Based on K-Nearest Neighbor Classifier.\u0026rdquo; Remote Sensing 12(2). doi: 10.3390/rs12020266.\u003c/li\u003e\n\u003cli\u003eSin-ampol, Phaothai, Tawee Chaipimonplin, and Supawadee Songka. 2020. \u0026ldquo;Local Community Engagement for Adaptation to Future Challenges in the Pilot Flood Detention Area of Thailand.\u0026rdquo; Pp. 203\u0026ndash;28 in External Interventions for Disaster Risk Reduction: Impacts on Local Communities, edited by I. Chowdhooree and S. M. Ghani. Singapore: Springer Singapore.\u003c/li\u003e\n\u003cli\u003eStatnikov, Alexander, Constantin F. Aliferis, Douglas P. Hardin, and Isabelle Guyon. 2013. Gentle Introduction to Support Vector Machines in Biomedicine, A-Volume 2: Case Studies and Benchmarks. World Scientific Publishing Company.\u003c/li\u003e\n\u003cli\u003eTehrany, Mahyat Shafapour, Biswajeet Pradhan, and Mustafa Neamah Jebur. (2014). \u0026ldquo;Flood Susceptibility Mapping Using a Novel Ensemble Weights-of-Evidence and Support Vector Machine Models in GIS.\u0026rdquo; Journal of Hydrology 512:332\u0026ndash;43. doi: https://doi.org/10.1016/j.jhydrol.2014.03.008.\u003c/li\u003e\n\u003cli\u003eTien Bui, Dieu, Khabat Khosravi, Shaojun Li, Himan Shahabi, Mahdi Panahi, Vijay P. Singh, Kamran Chapi, Ataollah Shirzadi, Somayeh Panahi, and Wei Chen. 2018. \u0026ldquo;New Hybrids of Anfis with Several Optimization Algorithms for Flood Susceptibility Modeling.\u0026rdquo; Water 10(9):1210.\u003c/li\u003e\n\u003cli\u003eVaghefi, Saeid Ashraf, Malihe Keykhai, Farshid Jahanbakhshi, Jaleh Sheikholeslami, Azadeh Ahmadi, Hong Yang, and Karim C. Abbaspour. 2019. \u0026ldquo;The Future of Extreme Climate in Iran.\u0026rdquo; Scientific Reports 9(1):1464. doi: 10.1038/s41598-018-38071-8.\u003c/li\u003e\n\u003cli\u003eVan Pham, T., Bui, D.X., Do, T.A.T. et al. Assessing flood susceptibility in Hanoi using machine learning and remote sensing: implications for urban health and resilience. Nat Hazards 121, 10149\u0026ndash;10170 (2025). https://doi.org/10.1007/s11069-025-07211-5.\u003c/li\u003e\n\u003cli\u003eWang Wie, Gao Jie, Liu Zheng, Li Chuangi. 2023. A hybrid rainfall-runoff model: integrating initial loss and LSTM for improved forecasting. 11. https://doi.org/10.3389/fenvs.2023.1261239.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"natural-hazards","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"nhaz","sideBox":"Learn more about [Natural Hazards](https://www.springer.com/journal/11069)","snPcode":"11069","submissionUrl":"https://submission.nature.com/new-submission/11069/3","title":"Natural Hazards","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Damage extent, Flood forecasting, Support vector machine, Linear regression, Landsat","lastPublishedDoi":"10.21203/rs.3.rs-7104599/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7104599/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eFloods are considered one of the most destructive natural phenomena, causing extensive damage to the various sectors, including agriculture, infrastructure, housing, and socio-economic activities. The watersheds of the Dez and Karun rivers in Khuzestan Province require comprehensive management and precise analyses due to the frequent occurrence of the floods and their economic consequences. In this study, machine learning algorithms and Landsat 5, 7, and 8 satellite imagery were employed to identify flood-prone areas and evaluate the extent of flood-induced damages. Flood hazard zonation maps were generated with satisfactory accuracy and validated against field observation data, achieving an overall accuracy of 75%. By integrating hazard maps with land-use maps, vulnerable assets including agricultural lands, orchards, and rural settlements were identified. Also, regression models analyzed the relationship between river discharge and flood extent with 82.6% accuracy, revealing that outflow discharge is a key determinant of flood severity and spatial distribution. The findings demonstrate that combining remote sensing technologies with machine learning methods provides a robust tool for flood risk assessment and effective crisis management. Based on the results, can be proposed mitigation strategies for flood-prone areas, flood-related insurance policy frameworks, and optimized natural resource management. Finally, we strongly recommend avoiding development in high-risk flood zones and implementing watershed-scale risk reduction measures.\u003c/p\u003e","manuscriptTitle":"Evaluating the Performance of Dez Dam Reservoir Operation in Flood Control Using Machine Learning Algorithms and Remote Sensing","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-19 10:29:29","doi":"10.21203/rs.3.rs-7104599/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"","date":"2025-10-15T05:39:27+00:00","index":0,"fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-12T05:47:13+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"Natural Hazards","date":"2025-09-11T09:16:24+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-07-12T13:34:42+00:00","index":"","fulltext":""},{"type":"submitted","content":"Natural Hazards","date":"2025-07-11T17:41:12+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"natural-hazards","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"nhaz","sideBox":"Learn more about [Natural Hazards](https://www.springer.com/journal/11069)","snPcode":"11069","submissionUrl":"https://submission.nature.com/new-submission/11069/3","title":"Natural Hazards","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"e7e0ae75-5e6d-4eea-9483-54ae1abe212d","owner":[],"postedDate":"September 19th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-29T05:21:17+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-19 10:29:29","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7104599","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7104599","identity":"rs-7104599","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}