Unveiling advanced data-informed predictive approaches for characterizing spatial variability in agriculture

dc.contributor.authorHernandez, Carlos
dc.date.accessioned2025-04-13T17:17:03Z
dc.date.graduationmonthMay
dc.date.issued2025
dc.description.abstractEmerging technologies such as spatial statistics, machine learning, and remote sensing have been shown to play a crucial role in digital agriculture, offering promising tools that enhance data-driven decision-making and contribute to the sustainability of farming systems. However, despite the growing volume of data collected each season, information on yield, environmental conditions (e.g., soil data), and crop variables are often incomplete or become too costly to obtain. In response to these challenges, this dissertation explores the development of predictive models that can support agricultural applications. The study is structured into three chapters. The first chapter introduces the expanding capabilities of digital technologies for data collection and predictive modeling, highlighting their potential for developing digital agricultural products at an on-farm level. The second chapter presents the design and implementation of a robust framework for generating spatial predictions of soybean seed quality parameters at the farm level. Recognizing that on-farm data often requires preprocessing before it can be effectively used in predictive modeling, the third chapter proposes a hierarchical Bayesian approach to address spatial misalignment issues. This method is designed to address the issue of missing values that are a result of the merging of datasets from disparate geographic locations. This approach facilitates the process of both data imputation and the integration of continuous spatial information with prior knowledge. Overall, this dissertation employs a combination of statistical techniques, machine learning algorithms, and spatial data management strategies to advance the design and implementation of agricultural predictive models. These models facilitate the spatial characterization of a broad range of agriculturally relevant variables, thereby contributing to more precise and informed decision-making in the field.
dc.description.advisorIgnacio Ciampitti
dc.description.degreeMaster of Science
dc.description.departmentDepartment of Agronomy
dc.description.levelMasters
dc.description.sponsorshipThe projects in this thesis were funded by the United Soybean Board, the North Central Soybean Research Program, the Iowa Soybean Association, and the United States Department of Agriculture - Natural Resources Conservation Service.
dc.identifier.urihttps://hdl.handle.net/2097/44879
dc.language.isoen_US
dc.subjectSpatial Variability
dc.subjectMachine Learning
dc.subjectPredictive modeling
dc.subjectSatellite
dc.subjectSoybean Quality
dc.subjectBayesian
dc.subjectGeostatistics
dc.titleUnveiling advanced data-informed predictive approaches for characterizing spatial variability in agriculture
dc.typeThesis
local.embargo.terms2026-05-30

Files

License bundle

Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.65 KB
Format:
Item-specific license agreed upon to submission
Description: