Fixing the Soil Health Tech Stack Now: An 8-Step Action Plan
By Rob Trice. This article was first published in AgFunder Network. It is reposted with permission.
This is an exciting time in soil health, with USDA unleashing $2.8 billion in funding for climate-smart commodity agriculture. The hope is that organizations receiving grants for soil health work will spend this money in a manner that best serves the public interest.
Working together to fix the soil health tech stack is part of this process. We can do this by harmonizing and standardizing soil measurement data and analysis. This will build a thriving marketplace that rewards farmers and ranchers for taking actions and delivering outcomes that result in healthier soils and carbon sequestration.
The concept of the soil health tech stack — the three-layer graphic below — and the need to bridge its layers were introduced in September 2021 in an article by my colleague, Seana Day.
Over several months in 2022, Farm Foundation and its partners, including The Mixing Bowl and TomKat Ranch, held three interwoven events to address the challenges of the soil health tech stack. Specifically, we looked to bridge soil data interoperability, calibration, or standardization, and ease the movement of digital information between the entities collecting, analyzing, and taking action with soil-centric data.
Soil Sampling Campaign: In May to June 2022, robust soil sampling took place at TomKat Ranch, an 1,800-acre regenerative cattle ranch in Pescadero, California. Point Blue Conservation Science, along with students from Skidmore College working with the non-profit The Soil Inventory Project collected over 1,000 soil samples. These were analyzed for total percent carbon by dry combustion at three separate analytical laboratories. Bulk density was measured at a subset of sampling locations to create a robust soil carbon data layer across five pastures where TomKat had applied treatment regimes.
The Soil Data Hack: The Purdue Open Agriculture Technology Systems (OATS) Center took the soil data results from the campaign to create a publicly available data set and combined it with TomKat’s historical soil data and other soil data sets. This was to establish a data foundation for a soil data hack that took place during the “Fixing the Healthy Soils Tech Stack” virtual conference from August 23 to 24, 2022.
All of the soil data was put into the MODUS data standard. MODUS defines data terminology, metadata and file transfer formats to expedite the exchange, merging, and analysis of soil and other agriculture testing data. It is used by some but not all soil labs today.
AgGateway’s Laboratory Data Standardization Working Group is upgrading MODUS to MODUS 2.0 and is a key proponent of its wide-range adoption by all labs analyzing soil. As AgGateway outlines here, the use of standardized soil data can help scale the efficiency of a low-margin business, decrease errors, improve lab turnaround times, and feed data to farm management information systems (FMIS) for analysis and recommended action.
The Soil Data Hack, a two-day hackathon-style event, was designed to make tangible progress toward fixing the soil health tech stack by encouraging participating developers to create open source code to help with the transfer and presentation of soil-related data in a common medium.
Fixing the Soil Health Tech Stack Conference: The virtual event was four hours long on both Tuesday, August 23rd and Wednesday, August 24th, 2022. The broad arc of the conference topics included an overview of the soil data hack, the concept of the soil health tech stack, and how to fix it.
Results From the Fixing the Soil Health Tech Stack Events
The soil sampling campaign helped us develop a clean data set in the MODUS data format. It also helped us better understand the disparity in lab analysis, as two labs got the same soil samples and came back with different results for both soil carbon and bulk density.
During the two days of the Soil Data Hack, the hackers:
- Took the MODUS-based soil data and turned it into a JSON format
- Fed the data into a business intelligence and data visualization system (Power BI)
- Pulled it into OpenTEAM‘s open source FarmOS FMIS
- Leveraged FarmOS to associate soil data to GPS lat/long
- Used RDF (a World-Wide Web Consortium standard data description and exchange format) to put soil data on the blockchain and make it available for Regen Network’s carbon credit program
- Linked the MODUS data to any HTML browser for visualization
- Used HTML to compare different soil data
The virtual conference revealed broad recognition for the need to fix the soil health tech stack. Videos from the virtual conference and hackathon are live on the Farm Foundation YouTube channel. The videos include the conference sessions and also the report-out from the hackathon.
Weaving Together Existing Solutions
Through our events, we determined that, by weaving together existing solutions, we can make great strides in fixing the soil health tech stack. Specifically:
- A dynamic framework for monitoring soil health exists as part of Point Blue Conservation Science’s Range-C and forthcoming Crop-C Monitoring Projects.
- A solid data collection method exists with The Soil Inventory Project’s approach for in-field, distributed soil sampling and transfer of samples to soil labs for analysis.
- A lab sample prep standard operating protocol exists in what the Soil Health Institute has developed, and it will help minimize testing errors and variance between soil labs.
- A soil data standard exists in the form of MODUS 2.0 as maintained by AgGateway.
- Sovereign MODUS soil data exchange can occur through platforms like Purdue OATS Trellis to transfer data to XML or JSON. It can also ink to other interoperable farm data applications like OpenTEAM’s FarmOS and Ag Data Wallet, the USDA’s Producer Operational Data System or other FMIS programs. The key point to underscore is that data transfer tools exist to enable the data owner to manage what data is shared when and with whom.
- Large-scale open aggregated MODUS soil data sets can be made available for analysis through tools like the OpenTEAM Digital Farmer Coffee Shop.
Additional effort is now required to build out the interoperability between these tools and others.
The 8-Step Action Plan to Fix the Soil Health Tech Stack
1. Upgrade and invest in soil health testing infrastructure
It is widely acknowledged that the United States has the most robust soil lab infrastructure in the world. However, much of that infrastructure was established to measure soil type and soil chemical nutrient levels. With more focus on soil organic carbon and rising interest in microbial analysis, we need to upgrade soil lab capabilities to account for new demands.
In addition to physical measurement equipment, we should also add knowledge management infrastructure to enable the digitization of lab data in a way that will support levels of provenance, attribution, and sharing consent management.
We should also not overlook the need to train and staff an adequate number of lab personnel facile with the new testing equipment and digital data tools.
2. Expand the use of standardized soil field collection methods to create a national soil health inventory database
Non-profit project The Soil Inventory Project (TSIP) was awarded a $20 million USDA grant to fund climate-smart practice adoption on 120,000 acres nationwide, and apply their distributed inventory system to monitor soil health outcomes. This effort is part of their wider effort to create a distributed national soil health database using scientifically proven and affordable methods for collecting and analyzing soil data. With its USDA grant, TSIP is better positioned to help other organizations overcome the cost and burden of collecting large-scale soil data in the field.
We should also embrace the further development of new techniques and methods that can ease the time and resource burden of in-situ soil sampling. Additionally, with the establishment of a large, ground-truthed soil database, one day we may hopefully be able to undertake soil analysis simply through the use of airborne remote sensors.
Soil data should be aggregated in a national soil health inventory database (such as the The Soil Inventory Project is undertaking). That database should be an accessible versioned, searchable registry of measurement protocols enabling interoperability of results. Machine learning data sets may be fed input from multiple models (field measurement, lab measurement, and remote sensing) and a registry of measurement protocols will allow the comparison and calibration of analysis between measurement methods.
Of course, soil data should be shared securely in an anonymized and aggregated fashion to establish regional baselines along the lines of FAIR (Findable, Accessible, Interoperable & Reusable) data standards.
3. Embrace standardized frameworks for field monitoring of soils
To promote apples-to-apples monitoring of agricultural lands, common frameworks and stratification tools should be promoted to help farmers and ranchers select indicators, develop study areas, determine how many samples to take and when and how to ensure data quality. Point Blue and its partners have received a USDA grant to promote the adoption of the Range-C Monitoring Framework to assist farmers, ranchers, and researchers with these tasks. Future standards such as Crop-C will be versioned to adapt to upgrades in technology and new protocols referenced in a common shared registry.
4. Promote and adopt lab soil sample preparation standard operating procedures
To minimize the discrepancies that can result from different labs handling soil sample testing differently, we recommend an effort be made to promote all soil labs to abide by the Soil Health Institute’s soil sample preparation standard operating procedures that were developed as part of their “North American Project to Evaluate Soil Health Measurement.”
Labs can capture whether they have followed the protocols through tools made in Survey Stack or the Question Set Library. It will be helpful to have information related to the protocols in the metadata that travels with soil samples from the field through lab testing.
We should also allow for the Soil Health Institute SOPs to keep a registry of versions and that labs can capture the SOP version they are following when they are undertaking a soil analysis.
5. Promote the development of tools using MODUS to make it the standard format to harmonize soil data
We need to see MODUS adopted as the data output of soil labs. AgGateway is the leader in seeing the MODUS data format application in agriculture, and deploying infrastructure to make the definitions of its codes be machine-readable and machine-actionable. Right now they are in the process of promoting the 2.0 version and we should anticipate future versions will arise. Labs should make soil analysis results available in MODUS and available to clients online (in online data formats like CSV, JSON or FarmOS, not just in hard copy or PDF format).
During our two-day hackathon, we saw both a state government’s department of agriculture as well as a large digital agtech company commit to using the MODUS-based tools we developed. We should leverage these “early adopters” to refine the tools and then promote them heavily to gain adoption.
Adoption of standardized soil data tools should not be limited to the US. The MODUS data standard needs to become a global format for soil data. Additionally, we should anticipate the need for data standards for future soil analysis. For instance, FAO’s GLOSOLAN has already created a standard approach for soil spectroscopy that should be promoted for global harmonization. Internationally focused actors like CGIAR, OpenGEOHub, and LandPKS are potential allies to create a global soils ledger.
The simplest way to promote MODUS is to provide resources for the development of MODUS-based tools and to promote the use of MODUS-based tools amongst the developer community.
6. Address sovereignty of agriculture data
While technical solutions exist to maintain a farmer or rancher’s sovereignty over data and data sharing, many in agriculture are unaware of these solutions; most software solutions in agriculture do not use these tools today. We need a conversation to get over the constant boogeyman of “data privacy” so that we can provide those who choose to share data the confidence that those receiving data will use it in a transparent manner for appropriate purposes.
Conditional data use agreements and consent management processes controlled by the producer need to be more widely embraced. The conversation needs to include the benefit to the farmer/rancher of sharing data and also needs to address (beyond technical matters) the social and legal aspects of implementing a trusted solution.
7. Get soil health data on the balance sheet
We need a discussion on how to account for soil data on the balance sheets of farming and ranching operations. Is it possible to create a single score that encompasses the soil health of agricultural land (similar to a corn suitability rating, for instance)?
While there are emerging markets for soil carbon and other environmental marketplaces, financial recognition of healthy soils appears to be undervalued by land buyers, lenders, and insurers.
Research exists to show the long-term benefits of healthy agricultural soils. Tools do exist to score soil health, to hold environmental claims data and to enable benchmarking of data. However, where are market participants like banks and insurance companies in terms of adopting the use of soil health assessment tools in their financial products?
8. Adopt a common semantic infrastructure for soil health
An additional source of friction occurs when different layers of a tech stack use different semantics (e.g., variables, controlled vocabularies, etc.) because this introduces the need to map/translate among them.
We need to develop and adopt a semantic infrastructure (a common set of variable definitions and their associated controlled vocabularies, distributed using APIs) shared across the tech stack to make communications easier. AgGateway’s Agrisemantics Working Group is implementing such an infrastructure to distribute MODUS Codes, their definitions, and other semantic resources.
With the addition of the semantic infrastructure, the soil health tech stack now looks like this:
There was keen interest amongst the event participants to turn our talk into action and our refine our tools to fix the soil health tech stack. I welcome your ideas on how we can move forward collaboratively on these initiatives.
Rob Trice is founding partner of The Mixing Bowl and Better Food Ventures. Other contributions to and reviews of this article include: Aaron Ault (Purdue Open Ag Technology Center). Chelsea Carey (Point Blue Conservation Science), Kris Covey (The Soil Inventory Project & Skidmore College) Dorn Cox (OpenTEAM), Andres Ferreyra (Syngenta; AgGateway member/volunteer), Martha King (Farm Foundation), Wendy Millet (TomKat Ranch), Cristine Morgan (Soil Health Institute), Liz Rieke (Soil Health Institute), Drew Zabrocki (Semios).