Sürdürülebilir Kalkınma Amaçları
Yayınlar
- ArticleAçık Erişim2025Advanced Soft Computing Techniques for Monthly Streamflow Prediction in Seasonal Rivers
Achite, Mohammed; Katipoğlu, Okan Mert; Kartal, Veysi; Sarıgöl, Metin; Jehanzaib, Muhammad; Gül, Enes
The rising incidence of droughts in specific global regions in recent years, primarily attributed to global warming, has markedly increased the demand for reliable and accurate streamflow estimation. Streamflow estimation is essential for the effective management and utilization of water resources, as well as for the design of hydraulic infrastructure. Furthermore, research on streamflow estimation has gained heightened importance because water is essential not only for the survival of all living organisms but also for determining the quality of life on Earth. In this study, advanced soft computing techniques, including long short-term memory (LSTM), convolutional neural network–recurrent neural network (CNN-RNN), and group method of data handling (GMDH) algorithms, were employed to forecast monthly streamflow time series at two different stations in the Wadi Mina basin. The performance of each technique was evaluated using statistical criteria such as mean square error (MSE), mean bias error (MBE), mean absolute error (MAE), and the correlation coefficient (R). The results of this study demonstrated that the GMDH algorithm produced the most accurate forecasts at the Sidi AEK Djillali station, with metrics of MSE: 0.132, MAE: 0.185, MBE: −0.008, and R: 0.636. Similarly, the CNN-RNN algorithm achieved the best performance at the Kef Mehboula station, with metrics of MSE: 0.298, MAE: 0.335, MBE: −0.018, and R: 0.597.
RDL Kalıcı Kaydıİncele - Conference ObjectYalnız künye2025Modeling of Meteorological Drought Under Climate Variability Using the Standardizes Precipitation Evaporation Index (SPEI) and Machines Learning Models for Wadi Ouahrane Basin, Algeria
Achite, Mohammed; Emami, Somayeh; Katipoğlu, Okan Mert; Toubal, Abderrezak Kamel; Zhang, Z.; Chenchouni, H.; Sarikaya, M.A.; Ciner, A.; Knight, J.; Gentilucci, M.
Machine learning algorithms are an important tool in capturing complex relationships and modeling high-dimensional climate data and drought severities. The main innovation of this study is the evaluation of the accuracy and applicability of augmented tree and kernel-based machine learning algorithms in standardized precipitation evapotranspiration index (SPEI) prediction. Rainfall and temperature data obtained from the ONM Department of Chlef City in the Wadi Ouahrane basin (270 km2), from 1973 to 2017 were employed to predict SPEI at a monthly time scale. In the model’s setup, two different models, Gradient Boosted Trees (GBT) and Support vector machine (SVM), were employed. The determination coefficients (R2), root mean square error (RMSE), mean absolute error (MAE) were employed to evaluate the model performances. In terms of accuracy and efficiency, the SVM model (R2 = 0.957, RMSE = 0.216, and MAE = 0.142) exhibits superiority over the GBT (R2 = 0.870, RMSE = 0.369, and MAE = 0.293) for SPEI estimation. Therefore, SVM model has been proposed for drought prediction at various time scales. The results of the study contribute to the understanding of drought dynamics and decision-making by policy makers in water resources management.
RDL Kalıcı Kaydıİncele - ArticleAçık Erişim2025Nature’s Guidance: Employing Bio-inspired Algorithm and Data-Driven Model for Simulating Monthly Maximum and Average Temperature Time Series in the Middle Black Sea Region of Türkiye
Katipoğlu, Okan Mert; Terzioğlu, Zeynep Özge; Zerouali, Bilel
This study compares the performance of various models in predicting monthly maximum and average temperatures across three distinct regions: Samsun, Amasya, and Çorum. The evaluated models include Artificial Neural Network (ANN), Shuffled Frog Leaping Algorithm coupled with ANN (SFLA-ANN), Firefly Algorithm coupled with ANN (FFA-ANN), and Genetic Algorithm coupled with ANN (GA-ANN). In setting up the models, the dataset was divided into 70% for training and 30% for testing, and the outputs of the models were evaluated using various graphical and statistical indicators. The model with the smallest root mean square error (RMSE) value was selected for the maximum and average temperature predictions. Accordingly, for maximum and average temperature predictions, SFLA-ANN (RMSE of 2.93) and GA-ANN (RMSE of 3.55) in Samsun, GA-ANN (RMSE of 2.91) and GA-ANN (RMSE of 2.50) in Amasya and GA-ANN (RMSE of 2.97) and GA-ANN (RMSE of 2.50) in Çorum performed better than the other models, respectively. In addition, for the maximum temperature prediction with the highest accuracy, the R2 value of the SFLA-ANN model in Samsun was 0.89. In contrast, the R2 values of the GA-ANN model in Amasya and Çorum were determined as 0.91 and 0.91, respectively. Similarly, it was observed that the R2 values of the GA-ANN model for the average temperature prediction with the highest accuracy at Samsun, Amasya and Çorum stations were 0.78, 0.92 and 0.92, respectively. Overall, the GA-ANN consistently demonstrated superior performance in predicting both maximum and average temperatures across all three regions, as evidenced by its consistently low RMSE values. These findings provide valuable insights into selecting effective models for temperature prediction tasks in different geographical regions.
RDL Kalıcı Kaydıİncele - ArticleAçık Erişim2025Enhancing streamflow drought prediction: integrating wavelet decomposition with deep learning and quantile regression neural network models
Mohammadi, Babak; Abdallah, Mohammed; Oucheikh, Rachid; Katipoğlu, Okan Mert; Cheraghalizadeh, Majid
Drought is a significant natural hazard that severely challenges water resource management and agricultural sustainability. This study aims to propose a novel approach for predicting streamflow drought indices (SDI-3, SDI-6, and SDI-12) in humid continental (Stockholm) and semi-arid (ELdiem) climates at different time-steps. The approach utilizes a Quantile Regression Neural Network (QRNN) coupled with wavelet decomposition (WD) techniques. Six mother wavelets (haar, sym8, coif5, bior6.8, demy, and db10) were used to decompose the SDI time series into different frequency bands, helping to identify patterns and trends in drought signals. The QRNN was compared with a tree-based machine learning (ML) model and two deep learning models: Long Short-Term Memory (LSTM) and Convolutional Neural Network (CNN). Results from stand-alone models showed that the LSTM model outperformed others in predicting SDI-3, while the QRNN model performed best in predicting SDI-6 and SDI-12 in both study regions. In the Stockholm station, the hybrid models achieved acceptable accuracy with bior6.8-LSTM2 (Nash–Sutcliffe efficiency (NSE) = 0.927), bior6.8-QRNN2 (NSE = 0.962), and demy-QRNN2 (NSE = 0.984) performing best for SDI-3, SDI-6, and SDI-12 predictions during the test phase, respectively. For the ELdiem station, the db10-QRNN3 (NSE = 0.926), demy-QRNN3 (NSE = 0.934), and demy-QRNN2 (NSE = 0.981) models demonstrated superior performance during the test phase in predicting SDI-3, SDI-6, and SDI-12, highlighting the robust capability of hybrid models across two case studies. The results indicate that combining WD with ML models can produce more accurate hydrological drought predictions than traditional models.
RDL Kalıcı Kaydıİncele - ArticleYalnız künye2025Developing sediment concentration prediction in the Euphrates River catchment, Türkiye, with a honey badger and coati optimization-based hybrid algorithm
Saroughi, Mohsen; Katipoğlu, Okan Mert; Kartal, Veysi; Simsek, Oguz; Kilinc, Huseyin Cagan; Pande, Chaitanya Baliram
Estimation of sediment concentration (SC) is of vital importance in terms of siltation and economic life of dams, lakes and aqueducts, reservoir operations, design of water resource structures, monitoring and control of water pollution, and flood management. Direct measurement of SC is a challenging and expensive task. For these reasons, it was used to estimate the SC values at a station in the Euphrates River. New hybrid models were established by combining the CatBoost regressor (CBR) and artificial neural network (ANN) models with the honey badger optimization algorithm (HBA) and coati optimization algorithm (COA). The performance of the new model was compared with stand-alone model of ANN and CBR, and their accuracy was evaluated. In the setup of the models, 4 different input combinations of lagged sediment and discharge values for up to 3 months were evaluated. It is noteworthy that as the number of input variables, i.e., lagged data input, increases, the prediction accuracy of the models generally increases. HBA and COA algorithms often improve the accuracy of sediment prediction by optimizing the parameters of the single machine learning model. In addition, according to the AIC performance metric, the HBA algorithm is generally slightly better capable of optimization than the COA. The best model outputs were obtained according to the HBA-CBR hybrid approach of scenario 4 (RMSE = 59.78, AIC = 785.03, R2 = 0.32, PBIAS = 0.016, SI = 0.48, and MBE = − 2.05 in test phase), which consists of discharge and sediment with a delay of up to 3 months. The results of the study are valuable for decision-makers and planners in terms of practical reservoir and flood management and protection of coasts and river beds.
RDL Kalıcı Kaydıİncele - ArticleYalnız künye2025Machine learning-driven flood susceptibility mapping: comparing model performances and feature influences in a coastal watershed of the Eastern mediterranean
Abdo, Hazem Ghassan; Richi, Sahar Mohammed; Bindajam, Ahmed Ali; Zerouali, Bilel; Katipoğlu, Okan Mert; Prasad, Pankaj; Alharbi, Maged Muteb; Ramadan, Mona S.; Mallick, Javed; Ghribi, Mounir
Flood susceptibility mapping is crucial for effective disaster management and mitigation in regions prone to extreme weather events. Despite advances in the use of conventional models for flood mapping, coastal basins in the Eastern Mediterranean still lack studies that employ machine learning (ML) techniques to accurately estimate flood risk. Most research focuses on quantitative prediction without systematically elucidating the influencing environmental factors, limiting the effectiveness of the results in applied contexts. Hence, the need to develop machine learning-based explanatory models that not only increase prediction accuracy but also reveal the most important determinants of flooding in complex coastal environments. This study assesses flood susceptibility using three ML algorithms, including, Light Gradient Boosting Machine (LGBM), Artificial neural networks (ANN), and eXtreme Gradient Boosting (XGBoost), with feature importance analysis conducted using SHapley Additive exPlanations (SHAP). The XGBoost model demonstrated superior performance in identifying flood susceptibility areas, compared to ANN, and LGBM. Feature importance analysis revealed that Aspect, Distance from River, and Slope were the most influential features, with Aspect being the most dominant predictor of flood susceptibility. The methodological integration of XGBoost with SHAP not only achieved high performance, but also provided interpretable insights into environmental drivers of flood risk. This combination represents a significant advancement over prior approaches, as it bridges predictive power with explainability, thereby enhancing the reliability of flood susceptibility mapping. The findings contribute to improved flood risk management strategies and lay the foundation for future research on interpretable and robust ML models for flood prediction.
RDL Kalıcı Kaydıİncele - ArticleYalnız künye2025Deep learning and adaptive boosting for hydroelectric power prediction using hydro-meteorological data: Insights and feature importance analysis
Karakoyun, Yakup; Katipoğlu, Okan Mert; Dogan, Ahmet
This study explores the use of modern artificial intelligence (AI) models such as Adaptive Boosting (AdaBoost), Autoencoder Based Regression Model (Autoencoder), Deep Neural Network (DNN), Echo State Network (ESN), Light Gradient Boosting Machine (LGBM) Residual Neural Network (ResNet) and feature importance techniques to predict energy production. Daily rainfall, streamflow, temperature and energy production data between 2016 and 2019 were used in the study. According to the feature importance analysis performed using the LGBM model, streamflow, temperature, and precipitation were identified as the most critical parameters influencing energy production. This structured AI-driven approach provided robust model evaluation and reliable performance measurements. As revealed by the analysis, the LGBM model showed the most accurate result with the values of root mean square error (RMSE: 98.27), mean absolute error (MAE: 73.93), Akaike information criterion (AIC: 17,100), Nash–Sutcliffe efficiency coefficient (NSE: 0.69), Kling–Gupta efficiency (KGE: 0.77), coefficient of determination (R2: 0.69), mean bias error (MBE: −5.23), bias factor (BF: 1.02), and percent bias (PBIAS: 1.51), while the DNN model came in with the second-best results. These findings emphasize both the effectiveness of deep learning and ensemble AI models and their practical application in improving hydroelectric power generation forecasting. Furthermore, the study underscores the importance of accurately monitoring and managing streamflow for the optimization of hydroelectric systems.
RDL Kalıcı Kaydıİncele - ArticleYalnız künye2025Harnessing artificial ıntelligence for streamflow predictions under climate change scenarios in arid region
Abbaszadeh, Maryam; Bazrafshan, Ommolbanin; Katipoğlu, Okan Mert; Jamshid, Sajad
Understanding the streamflow fluctuation of the Minab River basin in the south of Iran is of great importance due to the presence of the Esteghlal Reservoir Dam at its outlet and agricultural activities that produce food for the residents and support neighboring cities. Water resources of this basin have been affected by climate change and it is expected to observe more effects in the future. Therefore, this study was conducted to evaluate the impact of climate change on river flow levels in the Minab River basin, comparing the future period (2020–2050) with the baseline period (1985–2014). Using the CanESM5 model, future climate scenarios SSP1-2.6, SSP3-7.0, and SSP5-8.5 were simulated to project changes in precipitation and maximum, and minimum temperatures. For streamflow simulation, five machine learning models including Generalized Linear Regression (GLR), K-Nearest Neighbors (KNN), Support Vector Machines (SVM), Decision Trees (DT), and Random Forest (RF) were employed, with KNN showing superior performance. The findings indicate max and min temperatures could rise by 4.9℃ and 5.2℃ respectively under the SSP3-7.0 scenario until 2050. Mean annual streamflow will experience an increase of 33% and the highest values will probably be observed in fall, which can be the trigger for more floods and decreasing streamflow during warmer seasons can cause many challenges in the availability of water resources. These insights are crucial for water resource management by local authorities and organizations.
RDL Kalıcı Kaydıİncele - ArticleYalnız künye2025Validation of the EUMETSAT CMSAF global interpolated gridded rainfall estimation (GIRAFE v1) dataset for Türkiye
Pekin, Muhammet Ali; Katipoğlu, Okan Mert; Şahin Demirbağ, Nurdan
GIRAFE V1 is a satellite-based global precipitation dataset developed by EUMETSAT under the CMSAF program and released in 2024. This dataset was designed to support various applications such as climate analysis, hydrological modeling, and the investigation of long-term precipitation trends. It provides precipitation data with a spatial resolution of 1 × 1 degree and temporal resolutions at daily and monthly scales, covering the period from 2002 to 2022. In this study, the verification of the GIRAFE V1 data set was made according to the correlation coefficient (CC), Nash–Sutcliffe Efficiency, Kling-Gupta Efficiency, Root Mean Squared Error and Mean Error values by using the monthly total precipitation data of 79 observation points in Türkiye. Validation is handled both for each grid separately and for the entire data set. According to the findings obtained, it was seen that the data set was consistent at the rate of CC = 0.73 for the whole of Türkiye and for the 2002–2022 period. While the data set was CC = 0.79 in summer, it decreased to CC = 0.69 in winter and spring. After all, GIRAFE V1; It is emphasized that it can be selected as an alternative data set to traditional precipitation observation data, other satellite-based precipitation data and reanalysis precipitation data. The results of the study contribute to the use of an alternative dataset for flood, drought and environmental management applications for decision makers and planners.
RDL Kalıcı Kaydıİncele - ArticleYalnız künye2025Daily prediction of Urmia Lake water level using remote sensing data and honey badger optimization-based data-driven models
Saroughi, Mohsen; Katipoğlu, Okan Mert; Aktürk, Gaye; Gul, Enes; Simsek, Oguz; Citakoglu, Hatice
Artificial neural networks (ANNs), support vector regression (SVR) and CatBoost regression (CBR) machine learning methods have been combined with the honey badger optimization algorithm (HBA) and metaheuristic optimization algorithm to accurately and reliably predict lake water level (LWL), which is of great importance for the management and planning of water resources. In this study, meteorological and hydrological parameters, including temperature (T), precipitation (P), date (D), surface soil moisture (SSW), root zone moisture (RZW) and water level (WL), were employed as input data for predicting the LWL of Urmia Lake. The input data were employed to develop six different prediction scenarios. This study not only examined the impact of meteorological and hydrological parameters on LWL prediction but also compared the performance of individual models and hybrid models. The Akaike information criterion (AIC) index was used to ascertain the optimal machine learning model and to evaluate the six prediction scenarios. The results of the study indicate that, according to the AIC index, the data regarding the water level (WL) were significant in the prediction models. However, it should be noted that satisfactory results could also be obtained without using the WL data in certain scenarios. In scenario 4 (input data: D, T, P, SSW, RZW), where the WL variable was not included, the HBA-CBR hybrid model was the best model with the lowest AIC value (Train: -63,735, Test:-4693). In prediction scenario 6 (input data: D, T, P, SSW, RZW, WL), which included the WL data, the HBA-SVR hybrid model demonstrated high performance with the lowest AIC value (Train: -102,358, Test:-27,233). Accordingly, it was recommended to use lagged WL values as input in WL prediction because the prediction accuracy of the models significantly improved. Furthermore, hybrid models were found to perform better than individual models due to their more consistent results.
RDL Kalıcı Kaydıİncele - ArticleYalnız künye2025Investigating the Performance of Machine Learning Models in Estimating Monthly Dam Evaporation: A Case Study of Sidi Mhamed Ben Aouda Dam, Wadi Mina Basin, Algeria
Achite, Mohammed; Emami, Somayeh; Emami, Hojjat; Katipoğlu, Okan Mert; Pandey, Kusum; Hajimirzajan, Amir; Elshaboury, Nehal
Effective water resource management in arid and semi-arid regions requires accurate estimation of hydrological parameters such as evaporation. This study investigates the monthly evaporation of the Sidi-M’Hamed Ben Aouda dam basin in northwest Algeria using six machine learning models: random forest (RF), extra tree (ET), gradient boosting (GB), category boosting (CatBoost), light gradient boosting machine (LGBM), and multi-layer perceptron (MLP). Climatic inputs (temperature, relative humidity, wind speed, and sunshine hours) from 1978 to 2023 were used to train and test the models. The findings reveal that the ET model achieved the best balance between accuracy and computational speed, while the RF model provided the highest overall accuracy. GB had a faster runtime with slightly reduced accuracy, whereas CatBoost and MLP underperformed. This comparative analysis highlights the suitability of ensemble tree-based models, particularly RF and ET, for accurate and efficient evaporation prediction, supporting water resource planning in data-scarce and climate-sensitive regions.
RDL Kalıcı Kaydıİncele - ArticleYalnız künye2025Hydroclimatic drought trends and change point detection in the Yeşilirmak basin: a comprehensive evaluation using SPI, SPEI, DMI indices and ITA, SQMK, IPTA approaches
Öllükçü, Meva Nur; Katipoğlu, Okan Mert
Drought was defined as water scarcity resulting from below-normal precipitation in a specific region over an extended period. Analysis of droughts was of vital importance in terms of water resources planning and management, hydropower generation, irrigation planning, agricultural productivity, and reservoir planning. In this study, drought status was evaluated according to the Standard Precipitation Index (SPI), Standardized Precipitation Evapotranspiration Index (SPEI), and De Martonne Index (DMI) values calculated between 1979 and 2022 for seasonal and annual periods for Amasya, Çorum, Samsun, and Tokat meteorological stations located in the Yeşilırmak Basin, and station-based drought percentages were expressed. In addition, trend and change point analyses were conducted for each drought index according to Şen’s Innovative Trend Analysis (ITA), Sequential Mann-Kendall (SQMK), and Innovative Polygon Trend Analysis (IPTA) methods. The analysis identified the trends and onset years of droughts in the region, revealing that severe droughts were dominant at a 15% level. ITA results indicated a significant decreasing trend of 5–10% in spring, summer, autumn, and annual periods, alongside an increasing drought trend. SQMK analyses confirmed a general decreasing trend at the stations, while IPTA highlighted the dominance of increasing drought intensities in SPI, SPEI, and DMI trends. These findings underscored the need for policymakers to take proactive measures to mitigate drought risks in the Yeşilırmak Basin.
RDL Kalıcı Kaydıİncele - ArticleYalnız künye2025Application of Time Series Methodologies for Robust Forecasting of Atmospheric Pollutant Concentrations: PM10, SO2, NO2, NOx, and O3 in an Urban Environment
Katipoğlu, Okan Mert; Elshaboury, Nehal; Kartal, Veysi; Ertugay, Neşe; Kilinc, Huseyin Cagan; Şenocak, Serkan; Pande, Chaitanya Baliram
In recent years, the concentration of air pollutants has risen significantly due to urbanization and increased transportation. Erzincan province, due to its geographical location, is one of the region's most severely impacted by air pollution. Consequently, it is essential to accurately estimate air quality parameters to effectively analyze and manage the associated health risks. This study utilized a time series methodology to examine hourly air quality parameters, including particulate matter (PM10), sulfur dioxide (SO2), nitrogen dioxide (NO2), nitrogen oxides (NOx), and ozone (O3) throughout the year 2022 in Erzincan province. Various predictive techniques, such as feed-forward neural networks, group method of data handling, long short-term memory (LSTM), and least squares boosting tree methods, were employed for the analysis. For estimating air quality parameters, values delayed by up to 4 h—determined as the effective time period through correlation analysis—were utilized. Prediction performance was evaluated using eight different statistical parameters and visualization techniques. Notably, air quality parameters can be effectively estimated using historical data, with the LSTM model demonstrating superior performance compared to other models in this estimation. Furthermore, the most accurate predictions for O3 values were achieved using the LSTM algorithm, yielding an R2 of 0.93 and a root mean square error (RMSE) of 6.02. The findings of this study can aid policymakers in developing water resource management strategies, pollution control policies, and measures to combat climate change.
RDL Kalıcı Kaydıİncele - ArticleAçık Erişim2025Air temperature estimation and modeling using data driven techniques based on best subset regression model in Egypt
Elbeltagi, Ahmed; Vishwakarma, Dinesh Kumar; Katipoğlu, Okan Mert; Sushanth, Kallem; Heddam, Salim; Singh, Bhaskar Pratap; Shukla, Abhishek; Gautam, Vinay Kumar; Pande, Chaitanya Baliram; Hussain, Saddam
Air temperature plays a critical role in estimating agricultural water requirements, hydrological processes, and the climate change impacts. This study aims to identify the most accurate forecasting model for daily minimum (Tmin) and maximum (Tmax) temperatures in a semi-arid environment. Five machine learning models—linear regression (LR), additive regression (AR), support vector machine (SVM), random subspace (RSS), and M5 pruned (M5P)—were compared for Tmax and Tmin forecasting in Gharbia Governorate, Egypt, using data from 1979 to 2014. The dataset was divided into 75% for training and 25% for testing. Model input combinations were selected based on best subset regression analysis, result shows the best combination was Tmin(t−1), Tmin(t−3), Tmin(t−4), Tmin(t−5), Tmin(t−6), Tmin(t−7), Tmin(t−8) and Tmax (t−1), Tmax (t−2), Tmax (t−3), Tmax (t−4), Tmax (t−5), Tmax (t−6), Tmax (t−8) for daily minimum maximum air temperature forecasting, respectively. The M5P model outperformed the other models in predicting both Tmax and Tmin. For Tmin, the M5P model achieved the lowest root mean square error (RMSE) of 2.4881 °C, mean absolute error (MAE) of 1.9515, and relative absolute error (RAE) of 40.4887, alongside the highest Nash-Sutcliffe efficiency (NSE) of 0.8048 and Pearson correlation coefficient (PCC) of 0.8971. In Tmax forecasting, M5P showed a lower RMSE of 2.7696 °C, MAE of 1.9867, RAE of 29.5440, and higher NSE of 0.8720 and R² of 0.8720. These results suggest that M5P is a robust and precise model for temperature forecasting, significantly outperforming LR, AR, RSS, and SVM models. The findings provide valuable insights for improving decision-making in areas such as water resource management, irrigation systems, and agricultural productivity, offering a reliable tool for enhancing operational efficiency and sustainability in semi-arid regions. The Friedman ANOVA and Dunn’s test confirm significant differences among temperature forecasting models. Additive Regression overestimates, while Linear Regression and SVM align closely with actual values. Random Subspace and M5P exhibit high variability, with SVM differing significantly. For maximum temperature, Random Subspace and M5P perform similarly, while SVM remains distinct.
RDL Kalıcı Kaydıİncele - ArticleAçık Erişim2025Reconstructing Missing Streamflow Data in Mountainous, Snow-Dominated Regions Using Optimized Hybrid Machine Learning
Çırağ, Burak; Acar, Reşat; Katipoğlu, Okan Mert; Yağanoğlu, Mete; Şengül, Selim
Accurate streamflow estimation is pivotal for hydrological research and water resource management, particularly in mountainous, snow-dominated regions where missing data can significantly impair watershed modeling and water budget evaluations. This study presents an advanced hybrid machine learning framework to impute missing streamflow records at a strategically located observation station within a snow-abundant sub-basin of the Euphrates-Karasu Basin in Erzurum Province, Eastern Turkey. Leveraging daily meteorological and hydrological data collected over four decades (1983–2023), we developed and rigorously evaluated several predictive models using comprehensive performance metrics and rank-based analysis. In addition to conventional approaches-including a single Artificial Neural Network (ANN), CatBoost, and Deep Neural Network (DNN)-the study also compares innovative hybrid models in which ANN parameters were optimized using the Whale Optimization Algorithm (WOA) and the Hydrologic Cycle Algorithm (HCA). Notably, the hybrid models demonstrated superior predictive accuracy over traditional methods, thereby underscoring their potential to bridge data gaps in challenging terrains. The results affirm the efficacy of the proposed framework and its promise as a robust methodological foundation for future investigations in watershed modeling, drought assessment, and water budget analysis.
RDL Kalıcı Kaydıİncele - ArticleAçık Erişim2025Singular Spectrum Analysis for Noise Reduction and Feature Extraction in Hybrid Deep Learning Models: Integrating Meteorological Variables for Improved SGI Predictions
Koç, Erdal; Katipoğlu, Okan Mert
Within the scope of this study, a range of advanced machine learning and deep learning models—including Singular Spectrum Analysis (SSA), Adaptive Neuro-Fuzzy Inference System (ANFIS), Categorical Boosting (CatBoost), Convolutional Neural Network (CNN), Deep Autoencoder, Deep Neural Network (DNN), Gated Recurrent Unit (GRU), and Long Short-Term Memory (LSTM)—were employed to estimate the Standardized Groundwater Index (SGI) in Erzincan Province. SSA was utilized as a preprocessing technique to decompose input variables such as precipitation, relative humidity, temperature, and past SGI values into distinct components including trend, seasonality, cyclicality, and noise. These decomposed components were then fed into the artificial intelligence models to construct hybrid forecasting frameworks. The performance of each hybrid model was evaluated using multiple statistical indicators and visual analyses. The findings demonstrated that incorporating all SSA-derived subcomponents as inputs generally improved the monthly SGI prediction accuracy. However, for 12-month SGI predictions, the results were more variable, with both improvements and deteriorations observed depending on the model configuration. Additionally, the elimination of noise components was found to enhance both model generalization capability and overall prediction performance. Among the models tested, ANFIS emerged as the most effective in capturing GWD dynamics. To further investigate variable importance, Sobol sensitivity analysis was applied to the ANFIS outputs. The analysis revealed that previous SGI-1 values (t − 1) and relative humidity were the most influential inputs in predicting current SGI-1 (t) values.
RDL Kalıcı Kaydıİncele - ArticleAçık Erişim2025A novel integrated approach to predict the sodium absorption ratio (SAR) of groundwater sustainability using deep learning models and SHAP approach
Moharir, Kanak N.; Pande, Chaitanya Baliram; Chakrabortty, Rabin; Pramanik, Malay; Paneerselvam, Balamurugan; Katipoğlu, Okan Mert; Pal, Subodh Chandra; Scholz, Miklas; Yadav, Krishna Kumar; Mansour, Lamjed
Agriculture is a crucial factor in improving the country economic growth. With the massive impact of environment variation, supplying and measuring the quality of water for irrigation usage is a crucial task for water resource management authority. The demand for good quality water for both drinking and irrigation uses is increasing day by day in recent years. The main aim of the present study is to identify the suitability of surface water for irrigation uses in the Man River basin of Maharashtra, India using advanced techniques. These region first time is used such kind of models i.e. boosted tree, AdaBoost, decision tree, extremely randomized tree model, and feed-forward neural network (deep learning) models, and these models are better to analyze and understand the groundwater quality datasets in the saline area. The prediction results are evaluated with the use of different performance metrics, mean absolute error (MAE), mean absolute relative error (MARE), Nash Sutcliffe efficiency (NSE), root mean squared error (RMSE), and coefficient of determination (R2). The study identified that the boosted tree model is more appropriate for the SAR prediction value, this model with high accuracy compared with other models. The result of this study shows that boosted tree model is very suitable for prediction of SAR and also provided the accurate information for agriculture purposes. In the first scenario, the boosted tree model shows lower value of mean squared error (MSE) of 0.26 and higher R2 of 0.88, the second scenario shows lower value of MSE of 0.11 and higher value of R2 of 0.91. Overall, in both the scenarios, the result of boosted tree model is more favorable for SAR% prediction. This work shows that machine and deep learning models can improve and better prediction of the groundwater quality in the study area. It is essential to understand the water quality and improve the sustainable agriculture system and development. These advanced modeling methods help to stakeholders make better water management and irrigation decisions, boosting agricultural sustainability and productivity.
RDL Kalıcı Kaydıİncele - ArticleYalnız künye2025Analysis of drought patterns and categories based on the Standard Precipitation-Temperature Index (SPTI): application of fuzzy c-means clustering and principal component analysis
Berhail, Sabri; Katipoğlu, Okan Mert
This study deals with climate variability issues in the Kebir Rhumel basin, situated in northeastern Algeria. This region is experiencing significant changes in climate parameters, which will affect the region’s water resources, agriculture, and overall ecosystem health. The main goal of this study is to examine the severity and frequency of drought from 1981 to 2019. It used the Standard Precipitation-Temperature Index (SPTI), which was recently developed, and clustering methodologies such as c-means and k-means to examine the variation in rainfall patterns characterized by intervals of intense drought and heavy precipitation. These methods help assess the impact of these disturbances in water availability, which result in challenges for agricultural productivity and water management. Additionally, this study points out the significant trend in temperatures over the reported timeline, which aligns with global warming patterns; this intensifies the effect of drought and impacts the region’s water resources. The southeastern and western parts of the basin are areas that experience continuous extreme drought, and different levels of severe drought have been observed over decades. When comparing the results with existing methods, this research has revealed unique climatic trends in this basin area through clustering techniques applied to SPTI results. In particular, 45% of the stations indicate moderate wetness, 35% of the stations experience extreme wetness, and 20% of the stations are extremely dry. Each of these climatic patterns poses distinct challenges for water resource management. These findings demonstrate the complex interplay between rainfall variability, temperature increases, and drought severity. They underscore the urgent need for adaptive water management strategies and plans to address climate change impacts.
RDL Kalıcı Kaydıİncele - ArticleYalnız künye2025Mapping landslide susceptibility in the Eastern Mediterranean mountainous region: a machine learning perspective
Abdo, Hazem Ghassan; Richi, Sahar Mohammed; Prasad, Pankaj; Katipoğlu, Okan Mert; Halder, Bijay; Niknam, Arman; Hang, Hoang Thi; Alharbi, Maged Muteb; Mallick, Javed
Assessing landslide susceptibility is essential in urban planning and risk management. In the context of the Eastern Mediterranean region, there is a continuing need to compare the performance of machine learning (ML) algorithms in predicting landslide susceptibility, which can improve landslide risk management measures. Therefore, this study aims to evaluate and compare the predictive capabilities of three ML models: Multilayer perceptron (MLP), Light Gradient Boosting Machine (LGBM), and Extreme Gradient Boosting (XGBoost) models, in evaluating the susceptibility of various types of landslides and to refine the combination of causal factors. An evaluation of 19 conditioning factors, including topographical, geological, and environmental variables, was conducted to assess their effects on landslide susceptibility in different models in a geographic information system (GIS) environment. The results show that "Elevation" and "Slope" were consistently identified as the most influential factors in all models, with MLP demonstrating the greatest sensitivity to "Elevation." The study area was divided into five susceptibility categories: very low, low, moderate, high, and very high. According to the LGBM model, 24.27% of the area was classified as "very low" susceptibility, while the XGBoost and MLP models identified 25.69% and 27.28%, respectively. On the other hand, the "very high" susceptibility category covered 19.57%, 20.31%, and 19.78% of the area for the LGBM, XGBoost, and MLP models, respectively. The AUC-ROC approach has been utilized to evaluate, validate, and compare the performance of different ML models. Our study found AUC values for three MLTs. These findings suggest that all models demonstrate reasonable accuracy in identifying susceptible zones, and XGBoost demonstrated the best performance among the MLTs, with an AUC of 92.6% compared to the others. The insights gained from this study can inform targeted mitigation strategies to reduce landslide risks in Lebanon.
RDL Kalıcı Kaydıİncele - ArticleAçık Erişim2025Enhancing hydrological time series forecasting with a hybrid Bayesian-ConvLSTM model optimized by particle swarm optimization
Kilinc, Huseyin Cagan; Apak, Sina; Ergin, Mahmut Esad; Ozkan, Furkan; Katipoğlu, Okan Mert; Yurtsever, Adem
Hydrological time series forecasting often relies on addressing the inherent uncertainties and complex temporal dependencies embedded in the data. This study presents an innovative hybrid framework, the Bayesian-ConvLSTM-PSO model, specifically designed to tackle these challenges. The framework synergistically combines 1D convolutional neural networks (CNNs), a convolutional Bayesian network, multi-head attention, and long short-term memory (LSTM) networks, with parameters optimized through particle swarm optimization (PSO). The fusion of the convolutional Bayesian network and 1D convolutional neural networks enhances feature robustness by capturing both probabilistic uncertainties and spatial patterns effectively. The multi-head attention model further amplifies this by focusing on the most relevant features, improving the learning process and ensuring better representation of complex temporal dependencies. The proposed model is rigorously tested on daily streamflow data from three flow measurement stations (FMS): Ahullu (D14A014), Kızıllı (D14A080), and Erenkaya (D14A127). Experimental results reveal that the Bayesian-ConvLSTM-PSO model achieves significant performance gains across various evaluation metrics, including root mean square error (RMSE), mean absolute error (MAE), determination coefficient (R2), Kling–Gupta efficiency (KGE), and bias factor (BF). Notably, the model demonstrates exceptional accuracy with an R2 of 0.9950, a KGE of 0.9950, and a bias factor of 0.0003, surpassing the results of PSO-1D CNN-LSTM and benchmark models, such as DNN, DNN-LSTM, and 1D ConvLSTM. These compelling findings underscore the potential of the Bayesian-ConvLSTM-PSO framework as a robust and effective tool for applications in river engineering and hydrological time series forecasting.
RDL Kalıcı Kaydıİncele