High-resolution Satellite-based Modeling and Assessment of Gross Primary Production over Agroecosystems for Precision Agriculture

Abstract

Sustainable agriculture stands at the forefront of the global sustainable development strategic goals (SDGs), targeting the needs of increasing production to meet the growing food demands and to ensure profitability for producers, while maintaining affordability for consumers. Central to this endeavor is the precise monitoring of crop health, growth, vegetation productivity, water usage, and yield forecasting—each vital for enhancing productivity and optimizing resource allocation. Yet, addressing this strategic SDG is challenging given the big gap between industry applications and academic research: commercial methods often lack scientific rigor, while academic models are marred by complex, impractical parametrizations. This dissertation bridges this gap by providing a potential solution with strong scientific rigor and high scalability potential for easy commercial applicability. Initially, this work examined the biophysical components and links that explain the changes in Gross Primary Production (GPP) and Transpiration(T)—key indicators of vegetation growth, vegetation carbon capture, crop productivity, and water use in crop ecosystems. The analysis of these variables highlighted their critical role in agroecosystems and the potential benefits of using the Vegetation Photosynthesis Model (VPM) and Vegetation Transpiration Model (VTM) to estimate GPP and T respectively. The study evaluated the efficacy of high spatial resolution (HSR) satellite imagery compared to moderate spatial resolution (MSR) data as input in estimating GPP and T across various high-value crops, underscoring the potential of VPM and VTM in precision agriculture. Then, the research investigated the VPM_GPP estimates as a viable proxy for yield forecasting, addressing the challenges inherent in crop models and the limitations of remote sensing-based approaches. Through extensive testing across multiple agroecosystems, the findings revealed the VPM model's capacity to deliver accurate, scalable, and actionable insights into crop monitoring, field productivity variability, and yield predictions. Finally, the dissertation extended beyond theoretical models to practical application, engaging with over 170 industry stakeholders to align the technology evaluated in this document with their needs. This interaction ensures that the proposed solution is not only scientifically sound but also of tangible value to the agricultural sector, facilitating the transition of academic advancements into commercial viability. In essence, this work not only contributes to the academic discourse on sustainable agriculture, reducing the gap in knowledge about a robust method to estimate GPP_VPM-HSR, T_VTM, and a simpler accurate approach for yield forecasting, but also paves the way for its real-world implementation, fostering agricultural resilience and food security.