Accurate greenhouse gas (GHG) quantification from rice paddies is essential for climate change adaptation but is constrained by labor-intensive measurement methods. Silicate fertilizers have the potential to mitigate GHG emissions, yet predictive tools for evaluating their effectiveness remain underdeveloped. The objective of this study was to develop and validate machine learning (ML) models for predicting methane (CH4), nitrous oxide (N2O), and carbon dioxide (CO2) emissions from silicate-fertilized rice paddies using environmental covariates as input features. Field experiments were conducted over 2-year using three silicate fertilizers with different iron slag contents (i.e., 0.0
Japan is one of the most earthquake-prone countries in the world, where a significant portion of disaster-related losses can be attributed to structural collapses. Thus, the inspection of geo-structures related to these earthquakes is important. Although conventional inspections are conducted by two-dimensional analyses, the actual behaviors of the structures are affected three-dimensionally. This has led to an increase in the demand for three-dimensional analyses. A three-dimensional survey has recently been developed, using an unmanned aerial vehicle (UAV), which has made the rebuilding of the three-dimensional shape of structures easier. Despite the development of this technique, it has not been adopted for the design of geo-structures. In this study, for the purpose of incorporating the three-dimensional survey technique into the design of geo-structures, an earth-fill dam is surveyed three-dimensionally by a drone, and a three-dimensional finite element model (FEM model) of the dam is made based on the survey results and the newly developed modeling approach. The spatial distribution of the ground stiffness is estimated based on screw weight sounding (SWS) test results, and then applied to the 3-D FEM modeling. Gaussian process regression (GPR), which serves as a geostatistical interpolation framework to transform sparse field measurements into continuous three-dimensional material property distributions for the FEM model, is employed for the spatial interpolation of the ground stiffness. Finally, the FEM model is applied to a seismic response analysis, and the locations of high acceleration and high shear stress inside the earth-fill dam are identified.
River catchments are critical to rice paddies and other croplands because they provide reliable supplies of safe water for irrigation and support essential ecosystem functions. However, pollution from illegal discharges of untreated wastewater threatens irrigation water quality, soil health, and food safety, for instance, often contain potentially toxic heavy metals including Cu, Zn, Pb, As, Ni, Hg, Cr, and Cd which can accumulate in crops and other biological and environmental systems. This review presents an overview of advanced pollution-tracing technologies, including heavy metal isotope analysis, resin capsules, fluorescence fingerprinting, microbial source tracking, and machine learning models. A comparative analysis of detection limits, operational costs, scalability, and technical constraints highlights both the strengths and limitations of each methods and underscores the value of multi-tracer approaches and integrated data analysis. By synthesizing recent case studies, the review identifies four core strategies for watershed-scale pollution monitoring: (1) integrating microbial, chemical, and geospatial tools within multi-tracer frameworks to improve source attribution, (2) standardizing tracing methodologies and improving cross-method comparability, (3) developing cost-effective and field-deployable tracing tools, and (4) integrating tracer-derived data with remote sensing observations, machine learning, and other artificial intelligence models. This synthesis provides a foundation for developing high-resolution, reliable, and scalable river pollution monitoring systems that support the protection of soil health and crop safety.
This study develops a Dynamic Irrigation Management Transfer (Dynamic IMT) framework to reinterpret Taiwan’s irrigation governance and its adaptive evolution under changing socio-environmental conditions. Building on Sasaki Nobuo’s governance typology, the framework conceptualizes governance transformation along two interacting dimensions: authority allocation and functional integration. It explains how irrigation institutions adjust between decentralization and centralization in response to internal organizational change and external environmental pressure. Taiwan’s trajectory demonstrates a gradual transition from farmer-led Decentralized-Integrated (DI) governance to state-led Centralized-Integrated (CI) management through intermediate stages of Decentralized-Separated (DS) and Centralized-Separated (CS) systems. This evolution was driven by internal factors including aging membership, fiscal dependence, and institutional ambiguity, as well as by external pressures such as urbanization, industrial water competition, and intensifying climate extremes. The 2020 reform that incorporated all Irrigation Associations into the Ministry of Agriculture was not a political reversal but an adaptive response intended to restore coordination, accountability, and water security. In comparative perspective, Japan’s Land Improvement Districts sustain participatory governance through fiscal autonomy, while Korea’s Rural Community Corporation maintains a hybrid balance between centralized efficiency and local responsiveness. Taiwan’s integrated model reflects demographic and climatic constraints that made hybrid governance unsustainable. The Dynamic IMT framework extends traditional IMT theory by emphasizing institutional adaptability, feedback learning, and cyclical adjustment. Sustainable irrigation governance depends on adapting authority, integration, and participation to changing demographic and environmental conditions.
Given the increasing need for accurate agricultural land monitoring due to the discrepancy between official records and actual land use in South Korea, an urgent demand is for automated monitoring systems. This study aims to develop a system that combines object detection and segmentation capabilities in high-resolution images using the YOLO-Seg model for paddy parcel monitoring based on drone imagery. We propose a centroid-based tiling-movement technique together with a systematic analysis of optimal overlap rates (0–90
