FarmLab Digital Soil Mapping Software

Overview

FarmLab Digital Soil Mapping Software is a low-cost measurement and visualization tool designed for agronomists and farmers who need rapid, accurate insights into soil conditions at the paddock scale. The platform combines a hybrid soil sampling methodology with digital mapping capabilities, enabling users to collect samples efficiently in the field and map results directly back to specific farm locations.

By integrating ground-truth sample data with remotely sensed environmental data, FarmLab generates detailed maps for soil carbon, soil nutrients, and other key soil properties. These maps help farmers and agronomists identify yield-limiting factors across large areas, monitor soil carbon levels, and assess crop susceptibility to disease. The software also supports soil carbon measurement projects and enables accurate stratification of farms that have not yet been sampled, optimizing future sampling strategies.

Technical specifications

Key features:

• Hybrid sampling and mapping approach that combines physical soil sample collection with remotely sensed data to produce accurate, paddock-scale soil maps
• Soil carbon mapping validated to an R-squared of 0.85 for the 0–30cm depth range using training data from 15 locations across 1,400 hectares
• Multi-property mapping capable of generating maps for soil carbon, nutrients, and other soil properties beyond carbon alone
• Stratification capability that identifies areas of variance within a field, allowing future sampling strategies to reduce the number of sample points needed while maintaining accuracy
• Scalable methodology demonstrated across more than 2,000 sample sites from client data
• Visualization tools that present soil property maps in an accessible format for farmer decision-making

How it works:

The software uses a baseline sampling approach where initial samples are collected at regular intervals across a farm area. These baseline points are then supplemented with remotely sensed data that correlates with target soil properties. Machine learning models are trained on this combined dataset to interpolate soil properties across the broader area. The resulting maps accurately capture variance in soil conditions, which can be used to stratify future sampling efforts and reduce costs.

Technology readiness level

FarmLab's hybrid soil mapping methodology has been validated through an initial desktop experiment spanning over 2,000 sample sites, achieving strong correlation results for soil carbon mapping. The approach has been successfully applied to map multiple soil properties and nutrients beyond carbon.

Future validation is planned to expand the method's applicability across cropping and cotton farms, with a structured protocol involving baseline sampling at one sample per 20 hectares and a subset at one sample per 100 hectares determined through stratification. This next phase will further validate model accuracy and refine the selection of optimal remotely sensed data for specific soil properties. The technology is positioned for continued development and broader commercial application in precision agriculture.