Originally published on LinkedIn by Camilla Roberts, Chief Commercial Officer at Food Agility CRC
Carbon trading is booming. The global market is valued at $277b[1] and the value of Australian Carbon Credit Units (ACCUs) is expected to jump from $20 to $50 by 2030.[2] And you know it has gone mainstream when banks build their own market: NAB have just announced its own international carbon market in a global collaboration with Natwest, CIBC and Itau Unibanco.
But how do we make sure that farmers, who will be sequestering carbon on behalf of the country, benefit from their good land stewardship?
You might think the answer is obvious: let them sell carbon credits.
But it’s far from that simple. Right now, the cost to participate in the carbon market can be so high it prohibits many farmers from fully participating in the market. One large farm operation has estimated it would need $30m to baseline their soil carbon to the Emissions Reduction Fund standard. In that instance the current price of carbon wouldn’t pay for the baseline test costs, let alone subsequent tests to prove carbon levels are the same or increasing. This poor business case provides little incentive for our farmers to meet their and Australia's sequestration potential.
So, what can we do? The answer lies in data: making better use of it, leveraging it to improve the market, and giving farmers the confidence to share it in the first place.
The Federal Government has formally recognised that the cost of measuring carbon is stalling Australia’s ability to sequester carbon and lower its GHG emissions. The pressure is mounting leading up to COP26 (UN Climate Change Conference) in November where Australia needs to improve its GHG reduction commitments. So, the Federal Government intends to launch a soil measurement technology program to help find a cheaper solution to measuring soil carbon.
But what will all the tech companies need to improve measurement accuracy and modelling? Farm data.
Good farm data is essential to ground truth these new technologies and the underlying data models. The Federal Government also need that data to understand what Australia’s potential is for sequestration and which management practices are the most effective in sequestering carbon, and so soon the government will offer to rebate soil test costs in exchange for a farm’s soil data.
While almost every other industry with public commitments to Net Zero has to rely on buying credits to reduce their carbon emissions, farmers are in the unique position of being able to sequester their own carbon in soil and vegetation on their own land. They can accelerate the process through specific land management choices, and then choose how much to retain to offset their own emissions and how much to sell for income.
But the benefits go far beyond selling credits. Good soil carbon levels are a proxy for good soil health, which is in turn a proxy for the sustainability of a farming business (healthier land = better and more reliable production). Net Zero and low carbon footprints are also strong value propositions to buyers of farm produce and very soon likely to be minimum requirements to access international markets. Already, the EU is on a path that will favour Net Zero imports. Farmers that can demonstrate their environmental credentials with verified data are likely to have greater access to premium prices and retain their international trade partners.
The biggest catch with selling carbon is the weak business case and uncertainty: in many cases the costs (including risks) still outweigh the returns, the trade-offs are unclear and the time horizons are long.
There are currently 34 approved methods registered under the Emission Reduction Fund (ERF) to reduce GHGs. Most methods are administratively burdensome, require intermediaries who take 25-50% of the carbon income, require costly measurement techniques and include permanence requirements of 25 – 100 years. That’s an enormous cost and risk for any business to take on. If you have sequestered carbon by planting native trees and a wildfire rips through them, what is the cost to you to replant so you can meet your permanence requirements? Are you still making a net profit? What’s the probability of being able to maintain that profit margin for up to 100 years?
While carbon can also be sold privately, it carries greater reputation and quality risks. Earlier this year, a NSW farm sold $500,000 worth of carbon credits to Microsoft. But the deal has been roundly criticised for the lack of integrity in the methodology used to calculate sequestration, demonstrating that private sales can carry reputational (and therefore future revenue) risks.
There are also important trade-off decisions which are difficult to make when carbon predictions still have low accuracy. For example, the choice between locking off parcels of land to sequester versus keeping that for traditional production activities, or retaining carbon to offset a farm’s own emissions versus selling credits, or changing management practices which could risk reducing productivity – the backbone of the underlying farm business.
Farmers are anxious about sharing their carbon data – and for good reason. Many consider a farm’s carbon data to be equivalent to the DNA of the farm, as carbon is a proxy for soil health and soil health is a proxy for yield, income and land valuation. Increasingly, good data about a farm’s assets (including soil) and production is favoured in property sales, so naturally who has your data (and whether they can be trusted) is important to know because it could impact your business valuation. Farmers who share their carbon data therefore need to know: who holds it, for what purpose, does it get deleted and how it is protected?
There are instances where sharing data makes good sense but it’s something that the farmer should provide informed consent for. Unfortunately, informed consent is hard to achieve in relation to agriculture data agreements because most farmers are not lawyers and as a result, we see high levels of anxiety and delays around data sharing.
Another valid concern is just how much data is needed anyway? While many aspects of carbon accounting are in development, the preference from innovators and scientists is (understandably) to access as much raw data as possible. The carbon market is changing rapidly and while the data needed today may not be needed tomorrow, organisations may still hold and use your farm’s data for a long time. So, only sharing what’s necessary is a great way to de-risk data sharing.
Finally, data in agriculture is notorious for being difficult to work with. Indeed, much of the value of a new data set is its ability to be integrated with other datasets for new insights.
How do we overcome these barriers and give farmers the confidence to share their farm data so we can accelerate good land practice, carbon sequestration and new revenues for farmers?
Solutions to these challenges can be thought of like an analogue farm security system: gates, cattle grids and pathways to market. This is how I conceptualise the data management techniques that unlock farm data for innovation while also balancing risks. It involves a series of mechanisms that allow farmers to share, control and standardise datasets so they can access new revenue streams without putting the farm’s DNA at risk.
1) Gates: an opening to share data outside the farm gate. A practical way to accelerate data sharing is using a template data sharing agreement that balances risks and rewards for all parties. Building trust and showing what ‘balance’ looks like in a data agreement is the name of the game, which is why Food Agility has developed a template agreement with a tier 1 legal firm that builds on the NFF’s Australian Farm Data Code and Food Agility’s Data Policy and is now being tested in national pilots.
2) Cattle grid: controlling the release of data through the farm gate. This is about being smart about which data assets can get through the gate. So if the gate is accidentally left open by someone else, there is a fall back security mechanismm. Federated learning and homomorphic encrypted analytics are two techniques we are exploring that limit the need to share all a farm’s data, while still unlocking insights that secure new revenue.
Federated learning is a machine learning technique that trains an algorithm across multiple decentralised local data sources without actually transferring any data. Homomorphic encryption can combine two separately owned data sets and secure new insights without either party seeing the others’ underlying data. It may sound complex, but this is already used commercially. When you go to the supermarket and use your credit card, the bank knows the total transaction value and location but not the individual items purchased and the supermarket knows the individual items but not your age and income. Homomorphic encrypted analytics allows the bank and the supermarket to match their separately owned and private datasets, to secure more insights about their customers without breaching privacy laws.
In agriculture, this could be applied to public soil datasets being matched with private soil datasets. When combined, they provide the necessary standard of quality and historical data needed to validate remote measurement of carbon without the farmer signing over all their data to a third party, just what’s needed could potentially be matched while still encrypted. Wouldn’t it be great if we could trust this encrypted matching to secure carbon credit revenue without needing to share all a farm’s DNA data to a third party?
3) Path to market: paving the pathway by making data interoperable. If data is interoperable, it is easier to share and can be combined with other data sets to create new insights and more accurate predictions and measurements. A simple way to think of interoperability is this: if the public soil data set uses the same definitions and parameters as a private soil dataset then they can be easily integrated, but if one dataset was in hectares and the other acres, additional work needs to be done to make them compatible. Food Agility works with experts in this field, following the FAIR principles (Findable, Accessible, Interoperable and Reusable) to ensure we leverage the value of every data set. We are currently working on a project to match private soil data sets with with Australia’s largest soil carbon dataset (Visualising Australasia’s Soil (VAS)), developed by the Soil CRC and Federation University’s.
It is possible for every farmer to benefit from the carbon market, enjoy new revenue streams, and greater business sustainability in the face of climate change. To deliver on that possibility, we must accelerate data sharing by lowering the cost of its collection and reducing the risks. We can reduce risks through approaches like Food Agility’s Data Sharing Agreement template that balances risk and return, minimising the data transferred to other parties, and making datasets more interoperable. In doing so, we will secure the data needed to ground-truth, cheaper more accurate predictions.
There is money to be made in carbon. Who gets to make it will depend on how well we manage farm data in these vital early stages of the market’s evolution.
Contact me if you're interested in leveraging the opportunities in carbon for agriculture and secure data sharing in agriculture.
[1] https://www.reuters.com/article/us-europe-carbon-idUSKBN29W1HR
[2] https://www.afr.com/companies/energy/carbon-offset-prices-reach-record-as-buyers-grow-20210707-p587js
