What is the Soil Organic Carbon % Layer

Background

The Soil Organic Carbon % (SOC %) layer was developed by FarmLab using a combination of advanced machine learning techniques, historical soil data, and environmental factors as part of the National Soil Carbon Innovation Challenge. It is currently only available in Australia, however you can find out more about the background to the model and the challenge here: https://getfarmlab.com/advancing-soil-carbon-quantification-with-remote-sensing-and-machine-learning/

Here's a brief outline of the process to develop the model is as follows:

1. Data Collection: Comprehensive soil samples and historical SOC data were collected from various geographic locations. This data includes information from undisturbed lands, such as natural forests and grasslands.

2. Environmental Factors: Key environmental variables such as climate, vegetation type, land use history, and topography were incorporated to enhance the model's accuracy.

3. Machine Learning Algorithms: State-of-the-art machine learning algorithms were applied to analyze the collected data. These algorithms identified patterns and correlations between SOC levels and environmental factors.

4. Model Training: The model was trained on a large dataset to learn the baseline SOC % under different environmental conditions. The training process involved validating the model's predictions against known SOC levels in undisturbed soils.

5. Model Validation: The final model was validated using independent datasets to ensure its accuracy and reliability.

Selecting Model Dates

As the model is constantly being updated, you will find more than a single date exists for the layer. The current dates reflect the versions (one from December 23 and the other from November 23). Further iterations of the model may become available. 

It is recommended users use the latest model, as it contains more current data. Earlier models used additional 0-10cm soil carbon % results as part of their training data from parts of NSW and QLD, and as a result the model tends to over estimate 0-30cm values across those regions. 

Interpreting the SOC % Model Results

When you access the SOC % model, you will receive an estimated percentage of soil organic carbon for the specified location. This reflects the estimate of SOC % in whole soil for that specific 10 x 10m pixel. 

The results displayed are an average estimate over a 3 year cycle, and are not intended to reflect current carbon stocks beyond the model creation date.

Practical Applications

The "Zero Point" SOC % Model can be used in various practical applications, including:

• Stratification and Sample Plan Design for Carbon Offset projects. The model offers users the ability to accurately stratify, capturing the carbon variation across a project area. For more on how to do this, see the 'Range Stratification' knowledgebase article. 

• Soil carbon modeling. Using the modeling features in FarmLab, users can now use the SOC % model as an input to model more current soil carbon samples for an area - drastically improving the spatial model for that farm or project. 

• Estimating management practice changes on soil carbon: Using current soil test results, users can compare the effect of land management between the SOC % model and their results over time.

Conclusion

The SOC % Model is a powerful tool for understanding and managing soil organic carbon levels. By providing a reliable baseline, it helps users make informed decisions to promote soil health and sustainability. We hope this guide has provided you with a clear understanding of the model's capabilities and applications. If you have any questions or need further assistance, please don't hesitate to contact support@getfarmlab.com.