Soil Texture and pH Mapping Using Remote Sensing and Support Sampling

Abstract

Soil pH and texture are valuable information for agriculture, supporting the achievement of high productivity and low environmental impact, which is the basis for sustainable agricultural production. In this study, we present novel soil mapping techniques that integrate high-spatial-resolution satellite and ground data, surpassing traditional methods in precision and reliability. By synergizing remote sensing data, including polarimetric synthetic aperture and multispectral imagery, with climate and terrain information, alongside coarse-resolution soil data, we achieved high accuracy, with an average error of less than 6%, in predicting soil pH and texture parameters. Notably, the approach allows for detailed mapping at the pixel level, revealing nuanced variability within 10×10 m field pixels. Considering the accuracy, the method establishes itself as a benchmark for field management guidelines integrating a precision sampling approach, offering actual and high spatial resolution information crucial for sustainable agricultural practices. This holistic approach allows new opportunities to revolutionize soil management practices, facilitating variable rate applications, soil moisture, and fertilization mapping and ultimately enhancing agri-environmental sustainability.

Keywords

• Machine learning (ML)
• precision agriculture
• remote sensing
• soil health
• soil mapping