Comparing Predictive Soil Property Modeling for Agricultural Management Zones

This project focuses on developing and evaluating management zones for precision agriculture in the Coastal Plain of the southeastern United States. Building on findings from the Sampling Type Analysis, this work compares geostatistical approaches and machine-learning-based models for predicting spatial variability in key soil properties relevant to agricultural management.

The analysis evaluates how predictive performance and resulting management zones vary by modeling approach and soil sampling methodology, with particular attention to dynamic soil properties that influence nutrient management. By assessing the strengths and limitations of each method within a low-relief landscape, this research aims to identify optimal strategies for creating accurate, practical management zones that support efficient resource use and informed decision-making in Coastal Plain agricultural systems.