양미혜 Yang Mi-hye , 남원호 Nam Won-ho , 권영숙 Kwon Young-sook

DOI:10.5389/KSAE.2026.68.1.001 JKWST Vol.68(No.1) 1-12, 2026

---

Agricultural reservoirs supply water to approximately 63.7% of irrigated paddy fields, making stable supply essential for maintaining crop productivity. Water supply reliability is a key indicator of water resource stability; however, in the absence of standardized evaluation criteria, drought frequency has traditionally been used as a proxy to assess reservoir performance. Most reservoirs were constructed between the 1940s and 1970s, necessitating a reassessment of their current reliability under evolving management practices and changing water demand. This study quantified water supply reliability using two metrics: the Water Utilization Safety Rate (WUSR) and the Water Supply Reliability Rate (WSRR), incorporating the structure and dynamics of rural irrigation systems. Water demand and shortages were estimated by comparing management-level ponding depths across irrigated areas using the EPA Storm Water Management Model (EPA-SWMM), which simulates hydraulic delivery through canal networks. Results were validated against the K-HAS model, a conventional water balance method widely used in Korea. While K-HAS indicated consistently high reliability, the SWMM-based analysis revealed equal or lower grades, reflecting reduced actual supply reliability. Yongseong Reservoir showed the lowest WUSR, mainly due to overestimated demand influenced by field conditions and SWMM's limitations. In general, SWMM projected higher water demand than K-HAS, resulting in noticeable discrepancies. This study emphasizes the importance of accounting for irrigation network complexity in reservoir reliability assessments. Improving SWMM modeling accuracy and establishing standardized evaluation criteria will enhance future assessments and support more effective agricultural water resource management.

허윤규 Hur Yoon-kyu , 김락우 Kim Rack-woo , 이승헌 Lee Seung-hun , 김찬민 Kim Chan-min , 석희웅 Seok Hee-woong , 안수빈 Ahn Su-been , 이선형 Lee Sun-houng , 임성윤 Lim Seong-yoon

DOI:10.5389/KSAE.2026.68.1.013 JKWST Vol.68(No.1) 13-28, 2026

---

Ensuring the structural safety of agricultural facilities is increasingly important as extreme weather events intensify under climate change. This study estimates wind pressure coefficients for open-type rain-shelter structures in Korea through wind-tunnel tests and analyzes their distributions to support design. Tests on single- and multi-span models evaluated roof-section responses and guided the proposal of design coefficients. Under 0° (frontal) wind, pressure peaked at the windward eaves; under 90° (crosswind), the strongest contrasts appeared near the span center with recurrent local fluctuations at the side edges. Grouping pressure-tap data by roof section and averaging them yielded representative coefficients that capture these recurring patterns. Treating structural type and wind direction as independent factors produced coefficients more suitable for practical design and risk reduction. Future work will extend the scope of structures and wind conditions to provide a broader basis for standardizing wind-resistant design guidelines for open-type agricultural facilities.

정재영 Jung Jaeyoung , 김솔희 Kim Solhee , 권령섭 Kwon Ryoungseob , 유웨이궈 Yu Weiguo , 탕펑페이 Tang Pengfei , 이경도 Lee Kyungdo , 김태곤 Kim Taegon

DOI:10.5389/KSAE.2026.68.1.029 JKWST Vol.68(No.1) 29-41, 2026

---

Global food security is increasingly threatened by climate change, population growth, and market instability. Reliable monitoring of crop conditions in major production regions is essential for food supply planning. This study develops a web-based system to monitor maize and soybean in the U.S. Corn Belt, a key source of Korea's import crops. The system integrates four core processes: crop classification, growth monitoring, anomaly detection, and visualization. Multi-spectral satellite imagery forms the primary data source. Deep learning-based classification quantifies spatial distribution and annual changes in maize and soybean cultivation. Growth dynamics are assessed using time-series analysis of the NIRv index, enabling real-time comparison with long-term averages and the previous year. A Z-score-based method detects abnormal growth, allowing early identification of crop stress from drought, flooding, or pests. Results are delivered through a web-based dashboard with intuitive visualization for users. The platform can be extended to other countries, crops, and datasets. It provides timely, quantitative, and accessible crop information. This study demonstrates the feasibility of an independent crop monitoring system, reducing reliance on overseas platforms such as USDA NASS and EU MARS. The system supports evidence-based food import strategies and strengthens national food security through scientific analysis.

김상민 Kim Sang Min

DOI:10.5389/KSAE.2026.68.1.043 JKWST Vol.68(No.1) 43-54, 2026

---

This study aimed to evaluate the applicability of deep learning-based time-series models for predicting reference evapotranspiration (ET₀), benchmarked against the FAO Penman-Monteith (FAO-PM) method. We developed and compared three representative Recurrent Neural Network (RNN) models―basic RNN, Gated Recurrent Unit (GRU), and Long Short-Term Memory (LSTM)―using daily meteorological data from 21 observation stations across the Nakdong River Basin in South Korea. The results consistently showed that LSTM and GRU models, which incorporate gate mechanisms, significantly outperformed the basic RNN model in prediction accuracy across all stations. The LSTM model demonstrated the best overall performance, achieving the lowest average Root Mean Square Error (RMSE) of 0.871 mm/day and the highest coefficient of determination (R²) of 0.767 during the test period. The GRU model's performance was nearly equivalent to LSTM’s, making it a computationally efficient alternative. While the relative superiority of the models was consistent, the absolute prediction error varied depending on the distinct climatic characteristics of each station. Accuracy was highest at stations with stable wind conditions, whereas errors increased in coastal areas with strong, variable winds. These findings demonstrate that LSTM and GRU are robust and reliable data-driven methodologies for accurately predicting ET₀ across diverse climate environments, highlighting their high potential as effective tools for agricultural water resource management.

박진아 Park Jin-a , 서효재 Seo Hyo-jae , 서일환 Seo Il-hwan

DOI:10.5389/KSAE.2026.68.1.055 JKWST Vol.68(No.1) 55-64, 2026

---

This study aimed to evaluate short-term impacts of rice-farming operations on air quality. A fixed monitoring station was installed in a rice field in the Nonsan Plain, Korea, measuring PM-10, PM-2.5, NH₃, NO₂, and SO₂ at 5-min intervals. The system was synchronized with 360° CCTV and farm logs, and data were processed through two-stage QA/QC. Event analysis was conducted by comparing pollutant concentrations before, during, and after each operation within ±24 h. During tillage, particle concentrations rose sharply (PM-10: 24 h mean 54.7 ug/m³, during-work 67.1 ug/m³, peak 104.2 ug/m³; PM-2.5: 24 h mean 35.9 ug/m³, during-work 48.5 ug/m³, peak 80.1 ug/m³,), accompanied by high NH₃ levels (during-work mean 119.6 ppb, peak 150.7 ppb). Field leveling under flooded conditions did not significantly elevate PM concentration but produced a delayed NH₃ peak about 12-15 h later (peak 156.5 ppb), indicating post-flood volatilization. Transplanting generated moderate increases in particles (PM-10 increased from 21.1 to 37.5 ug/m³, and PM-2.5 from 12.5 to 24.6 ug/m³) with minimal gaseous changes. Harvest mainly increased PM concentrations (PM-10 peaked at 57 ug/m³, and PM-2.5 at 30 ug/m³). These results clarify that tillage and harvest primarily affect air quality through resuspension of particles, while flooded field leveling drives NH₃ volatilization. The operation-specific responses provide a scientific basis for scheduling farm activities, implementing NH₃ abatement strategies, and reducing exposure risks in rice agroecosystems.

김용원 Kim Yongwon , 정지훈 Chung Jeehun , 김진욱 Kim Jinuk , 이용관 Lee Yonggwan , 장원진 Jang Wonjin , 김성준 Kim Seongjoon

DOI:10.5389/KSAE.2026.68.1.065 JKWST Vol.68(No.1) 65-76, 2026

---

This study assessed the impact of increased groundwater use on watershed hydrology in Han River basin (35,770.4 km2) by applying groundwater use data from the 1980s (1976-1985) and 2010s (2006-2015) to SWAT (Soil and Water Assessment Tool). The groundwater use data was established using groundwater survey yearbooks and applied to each subbasin using WUS tool of SWAT. SWAT was calibrated at 4 multi-purpose dams and 3 multi-functional weirs under 2010s weather and groundwater use conditions. The average calibration results at 4 dams and 3 weirs showed that 0.76 of R2, 0.68 of NSE, and 2.13 mm/day of RMSE, respectively. After, 1980s groundwater use data was applied to the calibrated SWAT under the same weather conditions. The impact of increased groundwater use revealed a decrease in the average streamflow from 58.0 m3/sec to 55.1 m3/sec, representing an approximate 5.29% reduction. The total runoff (TR), baseflow (BF), and groundwater recharge (GWR) decreased by 2.5%, 20.6%, and 20.7%, respectively. Seasonal analysis indicated that TR notably decreased during autumn and winter, while BF showed a maximum reduction of 25.6% during summer season.