This study assessed the efficacy of improving the accuracy of reservoir water level prediction models by employing automated machine learning models and efficient cross-validation methods for time-series data. Considering the inherent complexity and non-linearity of time-series data related to reservoir water levels, we proposed an optimized approach for model selection and training. The performance of twelve models was evaluated for the Obong Reservoir in Gangneung, Gangwon Province, using the TPOT (Tree-based Pipeline Optimization Tool) and four cross-validation methods, which led to the determination of the optimal pipeline model. The pipeline model consisting of Extra Tree, Stacking Ridge Regression, and Simple Ridge Regression showed outstanding predictive performance for both training and test data, with an R² (Coefficient of determination) and NSE (Nash-Sutcliffe Efficiency) exceeding 0.93. On the other hand, for predictions of water levels 12 hours later, the pipeline model selected through time-series split cross-validation accurately captured the change pattern of time-series water level data during the test period, with an NSE exceeding 0.99. The methodology proposed in this study is expected to greatly contribute to the efficient generation of reservoir water level predictions in regions with high rainfall variability.

Unlike artificially created homogeneous materials, the process of calculating the elastic modulus of natural soil involves the possibility of errors. Because the stress-strain behavior of soil is nonlinear, the secant modulus of elasticity is often used based on 1/2 of the stress at failure. Since soil has the property of changing its elastic modulus depending on the confining pressure, numerical analysis models that analyze its behavior inevitably include complex elements. The hyperbolic model, which relatively accurately simulates the behavior immediately after loading in soft ground, assumes that the stress-strain curve of the consolidated undrained triaxial test is hyperbolic and requires the slope of the tangent line at the starting point. However, the slope of the initial tangent in the stress-strain curve obtained from an actual triaxial test is difficult to have regularity according to changes in confining pressure. Additionally, due to the characteristics of a hyperbola, even small changes in related factors cause large changes in the hyperbola. Therefore, there is a lot of randomness in the process of calculating model parameters from the triaxial test results, which causes large differences in the results. Therefore, the method of calculating the initial elastic modulus by the consolidation test presented in this study is also used to verify the method by the triaxial test. It can be applied. However, since this study was applied to only one sample showing typical consolidation characteristics, it is necessary to check samples with various physical properties in the future.

As the agricultural land use shifts from paddy to upland, ensuring reservoir water supply stability for upland crop irrigation becomes essential. The objectives of this study were to estimate the irrigation water requirements considering the upland irrigation scenario and to evaluate the reliability of the water supply from the agricultural reservoir using resilience indexes. Two study sites, Sinheung and Hwajeong, were selected, and soybean and red peppers, the most water-intensive crops, were selected as study crops, respectively. For the irrigation scenario, two irrigation methods of traditional scheduling (which irrigates all sites at once) and rotational scheduling (which distributes irrigation by districts), along with the upland conversion rate, were considered. The net irrigation requirement was estimated through a water balance analysis. The stability of the reservoir was evaluated using resilience indexes based on the simulated 10-years reservoir water levels and drought criterion. Overall, the water supply of the reservoir was evaluated as stable during the simulated 10 years, except for the one year. Compared to the two irrigation methods, rotational scheduling resulted in lower irrigation water usage in both sites, with reductions of 1.6%, and 0.3%, respectively. As the upland conversion rate increases, the water deficit could be intensified in Hwajeong with a conversion rate exceeding 50%, showing the number of deficit(ND) over the one and a rapid increase in the deficit ratio(DR). It was confirmed that the reservoir operation criteria can be enhanced by incorporating resilience indicators along with crop growth information, thus, this will be a further study.

Plastic greenhouses are simple structures consisting of lightweight materials such as steel pipes and polyvinyl chloride. However, serious damage occurs due to heavy winds and typhoon every year. To prevent a collapse of structural members, the Ministry of Agriculture and Rural Development has distributed plans and specifications for disaster-resistant standards. Despite these efforts, more than 50% of greenhouses still do not satisfy the disaster-resistant standards. Among the greenhouses that do not meet these standards, 85% are single-span greenhouses proposed 20 years ago. Consequently, there is a need to evaluate the safety of wind loads for the single-span greenhouse. Unfortunately, there are no design specifications for the greenhouses under wind loads. Therefore, a Korean design standard (KDS) has been utilized. KDS is defined with reference to wind speeds occurring once every 500 years, raising concerns about potential overdesign when considering the durability of plastic greenhouses. To address this, the modified wind load, considering the durability of the plastic greenhouse, was calculated, and a safety evaluation was conducted for sigle G-type plastic greenhouse. It was observed that the moment acting on the windward surface was substantial, and there was a risk of the foundation being pulled out if the basic wind speed exceeded 32 m/s. In terms of the combination strength ratio, it was less than 1.0 only on the leeward side when the basic wind speed was 24 m/s and 26 m/s. However, in all other cases, it exceeded 1.0, indicating an unsafe condition and highlighting the necessity for reinforcement.

Extreme rainfall will become intense due to climate change, increasing inundation risk to agricultural land. Hydrological and hydraulic simulations for the entire watershed were conducted to analyze the impact of climate change. Rainfall data was collected based on past weather observation and SSP (Shared Socio-economic Pathway)5-8.5 climate change scenarios. Simulation for flood volume, reservoir operation, river level, and inundation of agricultural land was conducted through K-HAS (KRC Hydraulics & Hydrology Analysis System) and HEC-RAS (Hydrologic Engineering Center - River Analysis System). Various scenarios were selected, encompassing different periods of rainfall data, including the observed period (1973-2022), near-term future (2021-2050), mid-term future (2051-2080), and long-term future (2081-2100), in addition to probabilistic precipitation events with return periods of 20 years and 100 years. The inundation area of the Aho-Buin district was visualized through GIS (Geographic Information System) based on the results of the flooding analysis. The probabilistic precipitation of climate change scenarios was calculated higher than that of past observations, which affected the increase in reservoir inflow, river level, inundation time, and inundation area. The inundation area and inundation time were higher in the 100-year frequency. Inundation risk was high in the order of long-term future, near-term future, mid-term future, and observed period. It was also shown that the Aho and Buin districts were vulnerable to inundation. These results are expected to be used as fundamental data for assessing the risk of flooding for agricultural land and downstream watersheds under climate change, guiding drainage improvement projects, and making flood risk maps.