How Effective Are Foundation Models for Crop Type Mapping Using Hyperspectral Imaging? A Comparative Study of Machine Learning, Deep Learning, and Geospatial Foundation Models

Soil maps help to track carbon storage, water availability, and land health, often using visible surface features such as plants to reveal what is happening underground. Plants, soil, and water are known to constantly influence each other, yet mapping methods rarely account for how this shifts those signals across places and over time. This study shows why current mapping methods may fail in rapidly changing environments and suggests ways to improve their reliability.