Browsing by Subject "Regenerative agriculture"

Now showing 1 - 1 of 1

Mapping knowledge domains of regenerative agriculture with a focus on on-farm nitrogen fertilization experimentation and response surface regression
(2025) Abdipourchenarestansofla, Morteza; Piepho, Hans-Peter
In the face of growing environmental concerns and the global demand for sustainable agriculture, achieving balanced nitrogen (N) management is critical for both maximizing crop productivity and maintaining environmental health. This dissertation proposes an innovative framework to address this challenge within the scope of regenerative agriculture, which emphasizes sustainable farming practices. Regenerative agriculture aims to reduce chemical inputs while maintaining yield levels yet implementing these practices at scale is complex due to the intricate interactions between biological, environmental, and technological factors on farms. This research tackles these challenges by introducing a Knowledge Domain Mapping (KDM)-based framework, integrating advanced technologies—including remote sensing, Internet of Things (IoT) telemetry, geospatial sciences, statistical modeling, machine learning, and cloud computing—to create a holistic and scalable system for optimizing nitrogen applications. Central to this research is the accurate estimation and spatial allocation of the Economic Optimum Nitrogen Rate (EONR), a crucial element for reducing nitrogen use and enhancing yield. A key contribution of this study is the development of a robust Response Surface Model (RSM) that leverages multispectral indices (MSIs) from Sentinel-2 imagery, historical IoT telemetry data, and on-machine nitrogen sensors. This RSM approach allows for precise EONR predictions tailored to field-specific conditions, reducing the need for traditional plot-based trials and achieving an average prediction error of just 14.5%. Applied to a 7-hectare winter wheat field, the model successfully predicted EONR values ranging from 43 kg/ha to 75 kg/ha across zones, showcasing the adaptability and accuracy of RSM for field-specific nitrogen recommendations. This precisionfocused approach exemplifies the study’s goal of minimizing environmental impacts while ensuring sustainable yield improvements. Beyond the initial field-level implementation, this research examines the generalizability of the RSM framework using two modeling strategies: a single RSM across fields and a weighted average model that aggregates individual field-specific RSMs. The weighted model demonstrated superior adaptability and high prediction accuracy, with a root mean square error (RMSE) of 11 kg N/ha for the EONR, highlighting the framework’s potential for broader application across different agricultural settings. Such generalizability supports the framework’s adoption in diverse farming environments, enabling precise and informed nitrogen management at scale. To facilitate widespread adoption and practical application, the dissertation also introduces a cloud-based infrastructure that integrates the KDM framework with real-time IoT data and satellite imagery. Leveraging cloud services like Amazon Web Services (AWS) Batch for job orchestration, Amazon S3 for scalable data storage, and RDS Postgres for structured data management, this8 infrastructure allows for seamless handling of both real-time and historical data. Spatial interpolation techniques, such as Kriging, enhance the model’s capability to generate real-time nitrogen prescription maps, enabling precise nutrient management for large-scale agricultural operations. Automated data quality control and data harmonization embedded within this cloud architecture provide a strong foundation for managing increasing data volumes and diverse field conditions, making the system cost-effective, adaptable, and efficient for modern agriculture. In summary, this dissertation maps regenerative agriculture via a comprehensive roadmap for translating regenerative agriculture principles into practical, operational nitrogen management practices. Through KDM an interdisciplinary approach is mapped by the integration of advanced modeling, data processing, and cloud technologies. This framework enables sustainable crop management and aligns with global goals for environmentally responsible food production. The innovations introduced here support a scalable, data-driven approach to agricultural sustainability, bridging scientific research with real-world applications to meet the evolving demands of modern agriculture.