Selim Şengül

Rainfall-induced soil erosion (SE) involves a hydrodynamic transition from distributed sheet flow to concentrated rill flow over an evolving mobile bed. To quantify the hydrodynamic mechanisms and flow–sediment interactions governing the interrill–rill transition in soils with contrasting particle-size distributions, 50 rainfall simulation experiments were conducted using sandy and loamy sand soils under five rainfall intensities (40–90 mm h−1) and five slope gradients (7.5%–17.5%). The sandy soil exhibited a transport-controlled response, with limited SE despite increasing hydrodynamic forcing, whereas the loamy sand soil developed hydraulically connected rills, resulting in an energy-controlled erosion regime. Flow remained laminar throughout all experiments, and supercritical flow occurred only under the highest rainfall intensity–slope combinations in the loamy sand soil. These findings suggest that supercritical flow accompanies advanced rill development but is not required for rill formation. Rainfall intensity dominated SE when all erosion regimes were analyzed together, whereas slope gradient became the dominant control in regime-specific analyses. Moreover, separating the interrill and rill regimes improved model performance for both soils, particularly for the loamy sand soil, where Nash–Sutcliffe efficiency increased from 0.59 to 0.97. Among the hydrodynamic parameters, unit stream power (P) emerged as the most effective predictor of SE. Receiver operating characteristic analysis identified critical P thresholds of 0.003 284 and 0.004 609 m s−1 for the sandy and loamy sand soils, respectively. Unlike the sandy soil, where particle motion and rill formation were separated by a broad energy interval, these processes occurred over a much narrower energy range in the loamy sand soil.

Rainfall simulators have been used for many years in laboratory settings to simulate natural rainfall. This natural precipitation produced in a laboratory environment is particularly important for hydrological research in controlled laboratory environments where its natural variability can be examined in detail. This study examines the spatial variability of rainfall's fundamental parameters, intensity, uniformity, and drop diameter, in a nine-subarea arrangement within a single-channel rainfall simulator. To observe how the distribution patterns change depending on rainfall intensity within the simulator's operating range of 30-140 mmh⁻¹, we conducted simulations at four different intensities (40, 70, 80, and 100 mmh⁻¹). Within the channel area, I (rainfall intensity) showed local variations. Similarly, Christiansen's uniformity coefficients (CU) ranged from 69% to 95%. However, local CU values in some sub-regions were significantly lower than these averages. D50 (median drop diameter) varied between 1.32 and 1.77 mm for different densities and sub-areas. Generally, it was observed that homogeneity improved as rainfall intensity increased, but there were still significant differences within the area. The size of the raindrops varies depending on where they fall, with larger drops generally observed in the middle of the area (A3, A4, A5). These results contradict the common assumption of uniform rainfall intensity within the channel area and highlight the necessity of sub-area-based analysis for the accurate interpretation of simulated rainfall experiments.

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.

This study provides a comprehensive analysis of the significance of rainfall simulator (RS) in hydrological research and investigates the effects of spatial variations on rainfall parameters. Rainfall simulators enable detailed examination of environmental variables under controlled laboratory conditions without relying on natural rainfall. However, the assumption of homogeneity in parameters such as rainfall intensity, uniformity, and drop size can lead to the neglect of spatial variations within the study area, thereby limiting the accuracy of the results. This limitation is particularly critical in studies focused on erosion, drainage, and infiltration, where spatial variations play a key role and may lead to misleading conclusions. In this study, performance parameters were evaluated across nine sub-regions along the simulator channel under four different rainfall intensities (40, 70, 80, and 100 mmh⁻¹). The effects of rainfall intensity on spatial uniformity and drop size were thoroughly analyzed. The findings reveal significant spatial variations in rainfall distribution. Notably, higher rainfall intensities were recorded in the middle regions resulting in higher uniformity values in these areas. Although the evaluation of uniformity coefficients for the entire area under 40 mmh⁻¹ and 70 mmh⁻¹ rainfall intensities yielded debatable results, sub-area analyses indicated that this uniformity did not hold true for the majority of the channel. Overall, a predominantly uniform rainfall distribution (>80%) was observed. Regarding drop sizes, spatial differences were identified, with a slight increase in drop size as rainfall intensity increased. These findings emphasize that treating rainfall parameters as a single fixed value for the entire study area may fail to fully capture the dynamic nature of natural rainfall. Considering spatial variations is essential for achieving more reliable and accurate results

Water scarcity is a pressing issue, intensified by factors such as population growth and industrialization. Hence, it is crucial to monitor, conserve and analyse groundwater resources, which are essential sources of clean and usable water. This study examines changes in groundwater levels (GWLs) in northeastern Turkey's mountainous and snow‐covered area. The primary objective is to assess the effectiveness of integrated machine learning models, specifically, the extreme learning machine (ELM) technique combined with signal decomposition techniques such as ensemble empirical mode decomposition (EEMD), variational mode decomposition (VMD) and empirical wavelet transform (EWT) for monthly GWL prediction models. Seventy percent of the accessible data is allocated for training, while 30% is designated for testing. A correlation matrix involving precipitation, temperature, relative humidity and GWL parameters is generated with inputs that possess significant correlations being selected, such as GWL_t−1, GWL_t−2, RH_t, RH_t−1 and RH_t−2. To evaluate model results, various metrics, including mean squared error, mean absolute error, mean absolute percentage error, mean bias error, bias factor, determination coefficient, Nash‐Sutcliffe efficiency, as well as tools such as box plots, Taylor diagrams and radar charts, are utilized to compare outcomes during the interpretation phase. The results of the analyses show that applying data decomposition methods such as EEMD, VMD and EWT significantly improves the performance of the ELM algorithm in predicting GWLs. VMD‐ELM is the most accurate for GWL forecasting among the approaches examined. The R² values of the most successful models established in the two wells are 0.993 and 0.905. The outcomes of this research hold significance for decision‐makers and policymakers as it offers informative insights into aquifer surveillance, irrigation strategizing and efficient administration of water resources.

Considering the importance of limited natural resources, accurately recording and evaluating temperature data is critical. The daily average temperature values obtained for the years 2019–2021 of eight highly correlated meteorological stations, characterized by mountainous and cold climate features in the northeast of Turkey, were analyzed by an artificial neural network (ANN), support vector regression (SVR), and regression tree (RT) methods. Output values produced by different machine learning methods compared with different statistical evaluation criteria and the Taylor diagram. ANN6, ANN12, medium gaussian SVR, and linear SVR were chosen as the most suitable methods, especially due to their success in estimating data at high (> 15 ℃) and low (< 0 ℃) temperatures. All the methodologies and network architectures used produced successful results (NSE-R² > 0.90). Some deviations have been observed in the estimation results due to the decrease in the amount of heat emitted from the ground due to fresh snow, especially in the −1 ~ 5 ℃ range, where snowfall begins, in the mountainous areas characterized by heavy snowfall. In models with low neuron numbers (ANN1,2,3) in ANN architecture, the increase in the number of layers does not affect the results. However, the increase in the number of layers in models with high neuron counts positively affects the accuracy of the estimation.

Predicting the runoff from snowpack accumulated in mountainous basins during the melting periods is very important in terms of assessing issues such as water supply and flood control. In this study, the Hydrological Engineering Center–Hydrological Modeling System (HEC-HMS) was used to simulate snowmelt runoff in the Kırkgöze–Çipak Basin that has a complex topography where altitude differences range from 1823 m to 3140 m above the sea level. The Kırkgöze–Çipak Basin, located in eastern Turkey, is a basin where snowfall is highly effective during the cold season. There are three automatic meteorology and snow observation stations and three stream gauge stations in the basin, which are operated especially for the calibration and validation of hydrological parameters at different altitudes and exposures. In this study, the parameters affecting snow accumulation–melting and runoff were investigated using the simulations on an hourly basis carried out over a three-year period for temporal and spatial distribution at the basin scale. Different from previous studies focusing on the rate of snowmelt, the temperature index method, which is calculated with physically-based parameters (R² = 0.77~0.99), was integrated into the runoff simulations (R² = 0.84) in the basin. The snowmelt-dominated basin is considered to be the source of the headwaters of the Euphrates River.