Predicting dam inflow is necessary for effective water management. This study created machine learning algorithms to predict the amount of inflow into the Soyang River Dam in South Korea, using weather and dam inflow data for 40 years. A total of six algorithms were used, as follows: decision tree (DT), multilayer perceptron (MLP), random forest (RF), gradient boosting (GB), recurrent neural network–long short-term memory (RNN–LSTM), and convolutional neural network–LSTM (CNN–LSTM). Among these models, the multilayer perceptron model showed the best results in predicting dam inflow, with the Nash–Sutcliffe efficiency (NSE) value of 0.812, root mean squared errors (RMSE) of 77.218 m³/s, mean absolute error (MAE) of 29.034 m³/s, correlation coefficient (R) of 0.924, and determination coefficient (R²) of 0.817. However, when the amount of dam inflow is below 100 m³/s, the ensemble models (random forest and gradient boosting models) performed better than MLP for the prediction of dam inflow. Therefore, two combined machine learning (CombML) models (RF_MLP and GB_MLP) were developed for the prediction of the dam inflow using the ensemble methods (RF and GB) at precipitation below 16 mm, and the MLP at precipitation above 16 mm. The precipitation of 16 mm is the average daily precipitation at the inflow of 100 m³/s or more. The results show the accuracy verification results of NSE 0.857, RMSE 68.417 m³/s, MAE 18.063 m³/s, R 0.927, and R² 0.859 in RF_MLP, and NSE 0.829, RMSE 73.918 m³/s, MAE 18.093 m³/s, R 0.912, and R² 0.831 in GB_MLP, which infers that the combination of the models predicts the dam inflow the most accurately. CombML algorithms showed that it is possible to predict inflow through inflow learning, considering flow characteristics such as flow regimes, by combining several machine learning algorithms.

Baseflow is influenced by incoming groundwater to aquifers and is closely related to watershed characteristics. Understanding baseflow characteristics is of great importance to river ecosystems and water management. Baseflow estimation typically depends on the observed streamflow in gauged watersheds, but accurate predictions of streamflow through modeling can also be useful in estimating baseflow. However, uncertainty occurs in the baseflow estimation process when modeling streamflow. Therefore, the purpose of this study is to compare the method that is proposed by Arnold and Allen (Scenario I) to an improved recession prediction method where the alpha factor (baseflow recession coefficient) is recalibrated and is applied to SWAT (Scenario II). Although the differences between the results (NSE, R², RMSE, MAE, d) of Scenarios I and II were small regarding streamflow and recession, the Scenario II method more accurately reflected the recession characteristics than the Scenario I method. Furthermore, the Scenario II method was better in baseflow prediction than for the Scenario I method proposed by Arnold and Allen. Therefore, these outputs pave the way and contribute to an efficient method for water management in watersheds.

Rainfall erosivity factor (R-factor) is one of the Universal Soil Loss Equation (USLE) input parameters that account for impacts of rainfall intensity in estimating soil loss. Although many studies have calculated the R-factor using various empirical methods or the USLE method, these methods are time-consuming and require specialized knowledge for the user. The purpose of this study is to develop machine learning models to predict the R-factor faster and more accurately than the previous methods. For this, this study calculated R-factor using 1-min interval rainfall data for improved accuracy of the target value. First, the monthly R-factors were calculated using the USLE calculation method to identify the characteristics of monthly rainfall-runoff induced erosion. In turn, machine learning models were developed to predict the R-factor using the monthly R-factors calculated at 50 sites in Korea as target values. The machine learning algorithms used for this study were Decision Tree, K-Nearest Neighbors, Multilayer Perceptron, Random Forest, Gradient Boosting, eXtreme Gradient Boost, and Deep Neural Network. As a result of the validation with 20% randomly selected data, the Deep Neural Network (DNN), among seven models, showed the greatest prediction accuracy results. The DNN developed in this study was tested for six sites in Korea to demonstrate trained model performance with Nash–Sutcliffe Efficiency (NSE) and the coefficient of determination (R²) of 0.87. This means that our findings show that DNN can be efficiently used to estimate monthly R-factor at the desired site with much less effort and time with total monthly precipitation, maximum daily precipitation, and maximum hourly precipitation data. It will be used not only to calculate soil erosion risk but also to establish soil conservation plans and identify areas at risk of soil disasters by calculating rainfall erosivity factors.

Background: Teleological reasoning is a cognitive bias purported to disrupt student ability to understand natural selection. Few studies have described pedagogical efforts to decrease student endorsement of teleological reasoning and measure the effects of this attenuation on the understanding and acceptance of evolution. This exploratory study examined the influence of explicit instructional activities directly challenging student endorsement of teleological explanations for evolutionary adaptations on their learning of natural selection over a semester-long undergraduate course in evolutionary medicine. In a convergent mixed methods design this study combined pre- and post-semester survey data (N = 83) on understanding natural selection, student endorsement of teleological reasoning, and acceptance of evolution, with thematic analysis of student reflective writing on their understanding and acceptance of natural selection and teleological reasoning. Results: Student endorsement of teleological reasoning decreased and understanding and acceptance of natural selection increased during a course on human evolution with teleological intervention (p ≤ 0.0001), compared to a control course. Endorsement of teleological reasoning was predictive of understanding of natural selection prior to the semester. Thematic analysis revealed that prior to the course students were largely unaware of the concept of teleological reasoning and their own tendency to think about evolution in a purpose-directed way, but perceived attenuation of their own teleological reasoning by the end of the semester. Conclusions: This exploratory study provides initial evidence that class activities to directly challenge student endorsement of unwarranted design teleological reasoning reduces the level and effects of teleological reasoning in an evolution course. Students were unaware of their high levels of endorsement of teleological reasoning upon entrance into the undergraduate human evolution course, which is consequential because teleological reasoning is a predicter of natural selection understanding. As a result of developed anti-teleological pedagogy, students had decreased unwarranted teleological reasoning and increased acceptance and understanding of natural selection over the course of the semester. The data presented show that students are receptive to explicit instructional challenges to their teleological reasoning and that attenuation of this bias is associated with gains in natural selection understanding and acceptance.

The negative changes in the hydrological cycle are increasing due to climate change and urbanization, resulting in deterioration of water quality and environmental issues. Although Low-Impact Development (LID) techniques studies have been conducted to solve this problem, the spatial distribution of LID facilities and sewer types has received less attention. In this study, it is proposed to analyze the effects of sewer type, the spatial distribution of LID facilities, and LID type on runoff and water quality using the Storm Water Management Model and to identify effective ways of improving the hydrological cycle and Non-Point Source (NPS) pollution associated with urbanization. As a result of the runoff reduction analysis, 68% of the rainfall was discharged at the outlet for separate sewers, 79% for combined sewers without storage tank, and 49% for combined sewers with storage tank. The LID scenario results showed the distributed LID application method has higher reduction efficiency of runoff and NPS pollution than the intensive application method. Moreover, intensive application of LID in downstream areas resulted in higher runoff reduction efficiency than the application of LID in upstream areas. It will be used not only in the hydrological cycle plan but also in NPS pollution management.

Changes in land use and climate can affect both the surface runoff and baseflow of streamflow. Baseflow significantly contributes to stream function in regions where climatic characteristics are seasonally distinct. Baseflow estimation depends on the observed streamflow in gauge water sheds, but baseflow estimations in data-poor or gauged watersheds depend solely on streamflow predicted from the hydrologic model. To accurately predict base runoff through the model, it is necessary to apply proper hydrological parameters. Accordingly, the objectives of this study are to (1) improve predictions of SWAT by applying the alpha factor estimated using BFLOW for calibration, and (2) evaluate streamflow and baseflow the effects of changes in land use and climate. The results show the alpha factor estimated using BFLOW in SWAT calibration improves the prediction for streamflow and recessions in the baseflow. In this study, streamflow increased due to land use change (impervious urban and agricultural areas), while baseflow decreased. The baseflow was more significant in the dry season than in the wet season, and the baseflow fluctuation was significant from February to May. Moreover, the changes in land use in the study area lead to differences in the seasonal characteristics observed for the temporal distribution of streamflow and baseflow.