Journal of the Korean Society of Agricultural Engineers publishes peer-reviewed research articles in engineering to help better understand and thus solve problems in agriculture, environment, food and other biological systems. Journal of the Korean Society of Agricultural Engineers presents cutting-edge research on a broad range of topics including irrigation and drainage, soil and water conservation, rural planning and development, agricultural structure & environmental control, rural environment & natural resources management, and more.
This study examines the multidimensional impacts of solar power deployment in rural areas on local communities using a system dynamics approach. A regional model was developed for Jeollanam-do, South Korea, integrating population dynamics, economic activity, land use, and carbon emissions. The simulation results indicate that while large-scale solar expansion significantly reduces net carbon emissions and increases regional income (GRDP), it accelerates population decline by diminishing rural landscape attractiveness. Notably, agrophotovoltaic (agroPV) systems were found to provide the most balanced outcome, mitigating population loss while achieving moderate economic benefits and substantial carbon reduction. These findings suggest that integrating solar power with agricultural land conservation, such as through agroPV, is essential for the long-term sustainability of rural communities.
Indirect Estimation of Sediment Deposition in Agricultural Reservoir Using Water Level Fluctuation Pattern Analysis
이정은 Lee Jeongeun , 김시호 Kim Siho , 황세운 Hwang Syewoon
This study proposes a novel methodology for indirectly estimating sediment deposition in aging agricultural reservoirs by analyzing water level fluctuation patterns. The approach was applied to Waryong Reservoir in Sacheon, Gyeongsangnam-do, using long-term hydrological observation data collected from 1997 to 2024. To classify rainfall events with similar rainfall and inflow characteristics, k - means clustering was employed, followed by Dynamic Time Warping (DTW) to group events exhibiting similar water level variation patterns. Events with comparable hourly rainfall distributions were then selected to calculate the Storage Response Index (SRI), enabling the identification of temporal differences in water level response patterns and the estimation of sediment deposition for specific storage ranges. Results indicate that approximately 48,930 m³ of sediment has accumulated in the 50-60% storage rate range, accounting for about 4.2% of the original design capacity. Comparison with water balance-based sediment estimation revealed a discrepancy of approximately 34,460 m³ within the same range. The proposed approach offers the advantage of estimating sediment volume without requiring detailed stage-storage survey data or precise bathymetric measurements. However, limitations a rise from uncertainties in inflow and withdrawal data, as well as the limited number of suitable events. Future research should focus on applying this method to multiple reservoirs and validating its performance through quantitative comparisons with field measurements.
Development of a FarmMap-based Facility Segmentation Model Using YOLOv12-seg
FarmMap is a digital agricultural land map based on high-resolution aerial imagery, containing boundary and attribute information (paddy fields, upland fields, orchards, facilities, etc.) of cultivated farmland, and is utilized in various administrative and research fields such as direct payment inspections and disaster insurance assessments. Since 2022, the transition to a nationwide update system has expanded the scope of work, revealing limitations of the traditional visual interpretation approach, including reduced accuracy, inconsistent quality, and increased time and cost. To address these challenges, this study developed a deep learning-based model using the latest YOLOv12-seg architecture to automatically detect and segment facilities within agricultural land. A dataset of 30,000 image-label pairs was constructed by combining FarmMap and aerial orthoimages, and an effective pre-processing method was proposed using clipping and cropping techniques. Performance comparison across different YOLOv12-seg model sizes (Nano, Small, Medium, Large, Extra-Large) revealed that the Medium model achieved the best performance with precision of 0.921, recall of 0.891, and mAP@50 of 0.946. Notably, by combining deep learning inference results with worker verification in a collaborative workflow, mAP@50-95 improved significantly from 0.812 to 0.961, demonstrating that both accuracy and efficiency in FarmMap update operations can be simultaneously achieved.
Soybean Irrigation Requirement during Heatwaves as Affected by Paddy-Soil Physical Properties
한경화 Han Kyunghwa , 고재준 Gou Jaejun , 송인홍 Song Inhong
Heatwaves intensify evaporative demand and can expose soybean cultivated on paddy soils to short-term water deficits. We aimed to (i) estimate soybean effective rooting depth (Zₑ) across representative topographic settings and soil textures of Korean paddy fields using physical constraints of the compacted layer, and (ii) quantify irrigation requirements (IR) during heatwaves. Zₑ was inferred from thresholds of plowpan bulk density and cone penetrometer resistance, yielding a range of 17-80 cm (mean 39 cm). By topography, Zₑ ranked fluvial plains > local valley-alluvial fans > Pleistocene terraces > reclaimed fluvial-marine plains; by texture, loam > silt loam > silty clay loam > sandy loam. Using 2025 weather records from Suwon, Jeonju, and Jinju, mean daily crop evapotranspiration (ETc) during 10 consecutive dry days averaged 4.8-8.1 mm day⁻¹. Assuming no rainfall for 10 days, we evaluated IR under two ETc scenarios (5 and 8 mm day⁻¹). IR declined significantly with increasing Zₑ. At ETc = 5 mm day⁻¹, IR = 0 mm when Zₑ = 56 cm, indicating soil water storage alone met crop demand. At ETc = 8 mm day⁻¹, IR = 25 mm was still needed even with Zₑ = 80 cm. At the mean Zₑ = 39 cm, IR rose 3.5-fold from 15 mm (5 mm day⁻¹) to 53 mm (8 mm day⁻¹). These results demonstrate that heatwave irrigation decisions are a function of Zₑ and ETc, supporting site-specific IR establishment considering paddy-soil physical properties.
Development of an Airflow Prediction Model Inside a Smart Farm Greenhouse Using Physics-Informed Neural Networks
최재원 Choi Jaewon , 이종혁 Lee Jonghyuk , 여선영 Yeo Sunyoung , 배정우 Bae Jungwoo , 김서윤 Kim Seoyun , 류현 Ryu Hyeon , 이상익 Lee Sangik
This study was conducted to address the computational limitations of conventional finite element-based Computational Fluid Dynamics (CFD) analysis, which, despite its high accuracy, faces exponential increases in mesh density and computation time as facility size expands and boundary conditions become more complex. To overcome these challenges, a Physics-Informed Neural Network (PINNs) was applied to predict airflow inside a naturally ventilated greenhouse (7.0 m in width, 4.0 m in height, and 24.0 m in length) located at Kyungpook National University Farm. The PINNs models were constructed separately for the front-view and top-view domains. The front-view model considered the roof slope and utilized a deep neural network with four hidden layers (128-128-128-64 nodes), while the top-view model was simplified into a rectangular geometry (7.0 m × 24.0 m) with three hidden layers (64 nodes each). Both models used 2,000 collocation points and employed the Adam optimizer with an adaptive learning rate scheduler to improve computational efficiency. In addition, the Navier-Stokes equations and boundary conditions were incorporated to predict the velocity and pressure distributions. When compared with the conventional finite element-based CFD results, it was confirmed that the PINNs could reproduce physically consistent flow characteristics without mesh generation. These findings demonstrate that PINNs effectively mitigates the limitations of CFD and can serve as a foundational technology for future three-dimensional smart-farm simulations and internal environmental control studies.
Time-Dependent Effects of Initial Compaction Conditions on the Ultimate Uplift Capacity of Continuous Pipe Foundations
이원영 Lee Wonyoung , 손영환 Son Younghwan , 조상범 Jo Sangbeom , 전지훈 Jeon Jihun , 김태진 Kim Taejin
Due to the increasing frequency of extreme weather events, the importance of protected horticulture is being emphasized. In single-span greenhouses, continuous pipe foundations are used to enhance uplift resistance. Although the uplift behavior of continuous pipe foundations has been investigated in several studies, time-dependent effects have not yet been considered. Since construction-induced soil disturbance requires a certain period of time to stabilize, especially during the early stages after installation, it is important to clarify the time-dependent variation in uplift resistance. Therefore, this study analyzed the time-dependent behavior of continuous pipe foundations through in-situ uplift load tests. Continuous pipe foundations were constructed with two initial compaction conditions (2-layer and 4-layer compaction), and ultimate uplift capacity was evaluated on the day of construction, and 8, 21, and 76 days after construction. As a result, the ultimate uplift capacity showed a clear increasing trend up to 21 days compared with the day of construction. In addition, an increase in field dry unit weight and a pronounced decreasing trend in the Dynamic Cone Penetration Index (DCPI) were observed up to 21 days. These findings quantify how uplift resistance changes over time under different initial compaction conditions. This study provides useful data for stable, rational foundation design and construction management of disaster-resilient greenhouses.
Role of the Collaborative Governance to Resolve Transboundary Water Transfer Conflicts in Agricultural Reservoir Watersheds
In recent years, South Korea has institutionalized integrated water management through the enactment of the Framework Act on Water Management, the establishment of River Basin Water Management Committees, and the formulation of comprehensive basin management plans. As a result, attention has shifted from traditional, infrastructure-centered approaches to collaborative governance-based water management. This study analyzes a case of inter-basin water transfer conflict involving an Damyang agricultural reservoir and explores the factors contributing to the success of collaborative governance through stakeholder interviews and survey data. The research focuses on the drivers of cooperation, the roles of key actors and institutions, the process of negotiation and consensus-building, and issues such as new intake facility construction and cost-sharing arrangements. The findings reveal that trust-building, interest-based negotiation, proactive leadership by local governments, stakeholder participation, and institutional support are critical to the success of collaborative governance. This case provides a practical model for resolving water-related conflicts and offers foundational insights for the development and implementation of integrated water management policies.
