We are thrilled to announce our recent investment in EarthOptics, an innovative company developing next-generation technologies that give growers revolutionary insights into their soil. EarthOptics has developed a proprietary machine learning platform that allows them to train a model on a few soil samples and then use that model to interpret large land areas. The goal? To enable farmers to have a rich understanding of their soil at an affordable price.

“Soil management until the last couple years has been significantly underinvested in, mostly because it’s historically been time consuming, laborious and expensive. We want to bring all the soil data together in a single cloud based system and use machine learning to solve a number of agricultural problems.”

- Lars Dyrud, CEO of EarthOptics

The Time for Soil Health is Now

Maintaining healthy soils is one of, if not the most crucial, undertaking for farmers. The longevity of their farmland and thereby their business depends on it. Degraded soils can lead to erosion, reduced water and nutrient holding capacity and loss of soil organic matter.

Soil degradation is not only economically harmful for the farmer--it also can have disastrous consequences for the environment and food security. Globally, soil biodiversity has been estimated to contribute between $1.5 and 13 trillion annually to the value of ecosystem services and regulation of climate. The loss of healthy soils reduces agricultural yields and could result in a food production loss of 25 percent by 2050. According to John Crawford, Director of the Sustainable Systems Program in Rothamsted Research in England, If we keep treating our soils poorly we will have to convert about 70 percent of the earth’s surface to agriculture to meet demand for food by 2050, up from about 40 percent now.

There is a lot of data that would be immensely helpful for farmers to help prevent soil degradation. However the only way to get an accurate measurement for soil properties is to stick a probe in the ground and send the sample to a lab to be analyzed. But soils are incredibly variable and within a field one soil sample can have completely different microbiome characteristics and even soil type than a sample 30 feet away. Collecting soil samples and having them analyzed is costly so getting a dense rich map over thousands of acres would be impractical for most agriculture applications.

The Best of Both Worlds

It is this challenge that EarthOptics has set out to solve. Instead of completely eliminating soil samples, they take a few soil samples and treat them as training data for a machine learning platform to interpret measurements coming from vehicle or tractor based sensors and other satellite data sets. The result is a low cost, very detailed map that's feasible to create over thousands of acres.

The first step in the deployment of any EarthOptics product involves collecting mapping data using satellites as well as the company’s own measurement platform called GroundOwl, a sensor suite which consists of electromagnetic induction and a ground penetrating radar sensor. The tool can be mounted on a UTV and used at-speed during a farmer's normal operations.

The second step is to take field measurements which become training data for their machine learning platform. The training datasets are then used to interpret the data that was collected by the vehicle near where the probes were sampled and build a model for decoding the ground penetrating radar and EMI signals. Once the model has been created it is run on all the vehicle sensor data and any other satellite and elevation data they have available. Then gradient techniques are applied to create a final 3-D map that is as accurate as if a farmer had done about five times as many samples on the field.

Cracking Soil Compaction

The EarthOptics team decided to first apply their technology to soil compaction since it is one of the most expensive soil issues for farmers in terms of necessary amendments and resulting yield loss. The company decided to start with a compaction mapping and tillage prescription product after determining that compaction was under-invested in considering it’s one of the most expensive soil issues for farmers in terms of necessary amendments and resulting yield loss.

Soil compaction can lead to poor plant growth. Roots in compacted soils must work harder to grow meaning there will be fewer or weaker roots and plants can uptake fewer nutrients and less water. Additionally, it is difficult for water to percolate through highly compacted soils, causing water to build up which can literally suffocate roots from lack of airflow.

On the other hand, excessive tilling, in addition to being expensive for farmers, is the number one factor in reduced carbon sequestration capacity. According to Goldman Sachs, soil compaction costs US growers an estimated ~$10-20bn annually. EarthOptics estimates that $1.5bn is spent annually on tilling areas of the farm that are not compacted and don’t need to be tilled.

EarthOptic’s TillMapper creates a map of the grower’s field that showcases how compact the soil is at different depths. The customized tillage prescription can allow for variable tilling (tilling at variable depths) and spot tilling (tilling only on specific parts of the field that are compacted) and can also be adjusted according to grower preferences for compaction threshold and specific crop. The company is working with a large equipment manufacturer to implement their soil compaction mapping to the company’s variable depth tillage implement.

Advancing Carbon Verification

Following the release of their soil compaction product, EarthOptics premiered C-Mapper, a carbon mapping product which produces a visualization of carbon levels by depth across a field.

There has been an extraordinary amount of interest recently in soil carbon sequestration as a tool to mitigate climate change with numerous carbon markets emerging to facilitate the sale of carbon credits from farmers to corporations looking to offset their emissions. The global carbon market is estimated to be around $272 billion annually and growing, according to S&P Global.

But one of the most significant obstacles these carbon markets face to widespread adoption is the cost of verification. Verifying carbon sequestration using traditional means could cost about $15/acre--about as much as farmers get paid for a carbon credit. EarthOptics can create a carbon field map for $3/acre that offers the same level of accuracy.

EarthOptics is working with carbon marketplaces to provide a cost effective solution for carbon verification. EarthOptics has gone from doing about 30 fields with them in the spring of 2021 to now being utilized on nearly 1000 fields. According to Dyrud, C-Mapper grew more rapidly from a customer adoption standpoint than any startup product he has been involved with. Within months of C-Mapper launch, EarthOptics was working with several large agriculture companies.

“Scientists estimate that farm soils could store over 60 billion tons of additional carbon. EarthOptics’ machine-learning technologies will be at the center of helping farmers gain control over returning carbon to the soil, by accurately mapping both soil carbon and agricultural practices like tillage. We need to know what's working and where it's working for sustainability efforts to be successful at scale."

- Lars Dyrud, CEO of EarthOptics

Building out the Soil Cloud

The US soil sampling and testing market is estimated to be about $6 billion annually. This is in addition to the $10-20 billion in yield losses to US growers due to soil compaction, the $1.5 billion spent annually by US growers on unnecessary tillage and the ~$232 billion and growing global carbon market.

The adaptable nature of EarthOptic’s platform means the opportunity to add new capacities and measurements to their “soil cloud” is essentially limitless. The company is launching soil moisture mapping and soil nutrient fertility mapping, which will provide farmers with a single source for multiple levels of field data. In addition, trials are being completed on a new StabilityMapper™ program, which will target the construction industries. In effect, it will provide site developers, construction managers, municipalities and contractors with information outlining soil stability and density before final site selection, undercutting, laying a foundation or forming a roadbed.

S2G participated in EarthOptic’s recent $10.3 million funding round alongside lead investor Leaps by Bayer, the impact investment arm of Bayer, and previous investors FHB Ventures, Middleland Capital’s VTC Ventures and Route 66 Ventures. This funding will enable further growth and expansion across more farmland of the company’s patent-pending soil mapping and machine learning technologies.

Welcome EarthOptics to the S2G Ventures Community! We look forward to working with you as you enable farmers to access precise and affordable soil insights that maximize their ROI, soil health and environmental benefits.