How to help High Integrity Environmental Projects Stand Out

Hi there, my name is Connor McCormick. I’m part of the team building Gaia. What brings me to Digital Gaia is an 8-year obsession with the problem of externality pricing. I’ve been trying to answer one fundamental question: How might we make it profitable for companies to do good, like by planting trees and sequestering CO₂?

Years of investigation has left me with a belief that would have surprised me at the outset: the lack of funding for projects does not come from a lack of funds! There are many governments, organizations, and generous individuals that want to support highly beneficial projects. What holds them back is that it’s difficult to distinguish high quality projects from the sea of mediocre ones.

This is why we’re building Gaia: to help high integrity projects stand out, so that projects of all sizes can compete on an even playing field. Imagine a world where the success of a project is not determined by its size, but by its actual impact. This is what we will be discussing in this series of articles.

This Series

This series will unfold over several installments, each providing an in-depth look at different aspects of the Gaia platform and its implications for the ecosystem of sustainability projects and for science in general. We’ll cover topics such as how Gaia makes it easier for Sustainability Officers to find and vet projects, what details are important in evaluating a project, how the underlying technology (called federated inference) creates a rigorous, scientifically accurate view of a project, and how anyone from any background can contribute to the network under this new paradigm of scientific discovery.

• Part 1: How to help High Integrity Environmental Projects Stand Out (you’re here!)
• Part 2: How to Evaluate a Project using Independent Assessments
• Part 3: The Evolution of a Project as it Grows
• Part 4: Indigenous Knowledge and Scientific Expertise in one Network of Collective Intelligence

Part 1: The Challenge and The Solution

The Challenge

The challenge a Sustainability Officer faces while selecting projects for their portfolio is ensuring that the projects are genuinely beneficial. This can tend to result in a preference for large scale projects due to their ability to afford experts and collect data, leaving smaller projects behind. Gaia aims to level the playing field by providing a detailed and transparent evaluation of projects, regardless of their size.

How might we build a world where high integrity environmental projects of all sizes can compete on an even playing field? That’s the story we’ll tell today.

Let’s imagine that a Sustainability Officer is browsing a marketplace for ecocredits (in this case, https://regen.network), looking for a good project to support.

There are many projects on offer, but let’s imagine that she selects the Savimbo Putumayo project, to explore it more (Psst, Savimbo Putumayo is a real project in the Colombian rainforest which you can support here!)

You can see that there’s lots of information that’s helpfully provided to her by the marketplace. She can see the token class that’s available to purchase, the history of credit issuance, the impact and co-benefits of the project, and the underlying methodology of the project.

However, how is she supposed to assess the current state of the project? This is where the Gaia Assessment comes in.

The Gaia Assessment is a window into the current state of the project. It’s kind of like a LinkedIn profile for regenerative projects.

On this page, the impact emblem displays the overall impact score for the project, as well as a summary of the various different badges that the Savimbo Putumayo project has earned.

For example, by hovering over the one of the hexagons, she can see that this project has Carbon Sequestration highlighted, indicated it’s something that the system, who we call “Gaia”, believes the project is doing.

If she clicks on this emblem she’ll open the Gaia Rationale page for this project. This page will tell her all about Gaia’s beliefs about the Savimbo Putumayo project, we’ll explore each of these sections step by step.

(Guess what? All of the data and graphs you see on this page are real! As before, if you find yourself wanting to support the Savimbo Putumayo project you can do so here)

The Savimbo Putumayo project has 10 different lots, and each of them has two highlighted variables that Gaia thinks are important: Carbon Stock and Plant Count.

Carbon Stock is the amount of CO₂ sequestered by this lot. Plant Count is the number of living plants on the lot.

By clicking one of these badges she can see a high level overview of the sort of evidence that has gone into informing these highlights.

As she scrolls down, she’s greeted with a graph capturing the number of plants that Gaia believes are currently alive on in this lot (197) and the number that are expected to be alive by the time the project reaches maturity.

The further into the future that Gaia tries to look, the less confident she is about what will happen as you can see from the yellow area getting much wider. This is because Gaia uses a type of machine learning called Active Inference to fuse many different data inputs so she can provide a current understanding and future prediction of a project.

But what kinds of information does she pay attention to? I’m glad you asked.

It turns out that the most relevant variable to figuring out the plant count* is Wildfire Risk. This is because if a wildfire goes through it can completely wipe out the project.

But the good news is (as you can see from the chart) that according to Sust Global the wildfire risk for this area is close to zero. So, despite being highly relevant, it’s not very impactful in the final prediction.

*according to our current model, as the model is updated these things change. We’ll get to the model later on

In this case, you can see that there is a long list of images that they’ve provided to corroborate the planting process.

Whereas wildfire risk can kill off the plants, you need them there in the first place for it to even matter. This is likely why Batch Size (meaning the number of plants that Savimbo says they planted) is the second most relevant variable.

In this case, you can see that there is a long list of images that they’ve provided to corroborate the planting process.

She can also see a historical and predicted view of the region’s weather.

Here, the Sentinel Satellite provides her with a sense of whether the trees are growing as expected. Early in a project’s lifecycle it’s not very informative (the trees are too small) but over time this becomes useful to check for progress.

Finally, the least relevant of the variables (in this case) are the number of trees claimed to be counted by Savimbo and by Digital Gaia.

In fact, it used to be these were much more relevant, but as new information came in the’ve become less important to the overall conclusion. This is good news! It means that the system is now leaning on many sources of information instead of heavily leaning on just a few.

Now that she’s seen all the variables that were important in informing that graph of current and future plant counts, she’s wondering what evidence sits behind those conclusions.

Here she can see the summary of the data sources:

1. Copernicus Satellite
2. Prach Sri (from Savimbo)
3. Matthew Jenkins (from Digital Gaia)

Including how much evidence they uploaded and how confident the system is in it.

But this clearly isn’t enough. It’s not sufficient to just know what information was important, but also what assumptions Gaia makes in order to come to these conclusions.

She’s in luck, because this is also transparently available to explore in the rationale.

Here she’s provided with what she needs to get a high level understanding of the underlying model (the assumptions and hypotheses) that Gaia used to come to her conclusion. Including links to the Github repo where all the code is available to see. The data that informed the process is not available because that is kept in a separate, private repository.

She can see the contributors, and even visit their Github profile to check their credentials. Here she sees that Matthew Jenkins, who was a significant contributor to the model, has a Ph.D. in Horticulture and Agronomy.

She can also see a high level diagram of the model assumptions. For example, this flowchart is saying that genetics influences plant size, which changes the state of the plant at time t, which changes the state of the plant at time t+1, etc.

If this is something she’s knowledgable in, it gives her a high level sense that the model is covering the factors that she believes are important.

If she’s not totally convinced, by clicking the Download button, she can easily pull up the underlying code itself. Allowing her to explore the underlying model and assumptions, and if she notices any opportunities to contribute — by fixing errors or adding on to the model — she can do so easily by submitting a pull request on github repo.

So now that she’s seen all this information, she’s feeling pretty confident that this is a high integrity project, both in terms of its plan and its execution, and she’s ready to go back to the marketplace buy some of the Savimbo Putumayo credits.

Summary

In this first installment of our series, we summarized how the challenges that Sustainability Officers face when looking for projects can make it hard for smaller projects to prove their impact and integrity.

We introduced Gaia, a platform aimed at leveling the playing field for environmental projects of all sizes. Gaia provides a comprehensive, transparent, and dynamic assessment of projects, making it easier for officers to make informed decisions.

We took a closer look at how Gaia works, using the real-life example of the Savimbo Putumayo project. We saw the various variables Gaia considers, how these variables interact, and how the system adjusts its predictions as new data comes in.

We explored the level of transparency Gaia offers, allowing anyone to dive into the underlying model assumptions, see the contributors, and even explore and contribute to the model’s code.

By the end, we saw how Gaia can instill confidence in Sustainability Officers about a project’s integrity, helping them make informed decisions and ultimately leading to a fairer and more effective market for sustainability projects.

As we wrap up this first part, here are some ways you can get involved:

• If you’re involved with a project, a funder, or a marketplace, reach out to us <connor at digitalgaia.earth>. We’re eager to collaborate and help bring more transparency and fairness to the world of sustainability projects.
• If you’re a scientist, we could use your expertise. Your insights can help us improve our models and make Gaia even better, please reach out!
• If you’re just interested in the concept and want to learn more, watch this video that provides a quick overview of Gaia and its potential impact.
• If you want to stay updated and receive the next installments of this series straight into your inbox, consider subscribing to this newsletter.
• And finally, we’re thinking of organizing a webinar to dive deeper into these topics and answer any questions you might have. Would that be of interest to you?

Coming soon! How to Evaluate a Project using Independent Assessments