The accurate monitoring and measuring of the impact of forest restoration projects are an integral part of ensuring their integrity and validity. One example of a company at the forefront of gathering accurate data is Space Intelligence and we interviewed Prof. Ed Mitchard BA PhD SMIEEE, its Co-Founder and Chief Scientist.
Space Intelligence is a company founded in 2018, an innovative data provider for nature-based projects with expertise in satellite data analytics, artificial intelligence, and forest ecology. Mr Mitchard told us about his journey of founding Space Intelligence and stressed on the importance of REDD+ projects and their high impact on suppressing deforestation.
Space Intelligence’s innovation was also recently validated by one of the most influential tech majors – Apple. The company chose its product offerings Carbon and Habitat Mapper to construct habitat and forest carbon maps of Apple’s Restore Fund project areas.
Prof. Mitchard, what is your background and what brought you to Space Intelligence?
At school I was horrified by global deforestation and how it was accelerating and not decreasing with time. I was lucky enough to win a competition to be part of filming a documentary about tropical forests and deforestation in the Brazilian Amazon when I was 15 – that was sufficient to make me certain I wanted to dedicate my career to helping solve this problem.
During my degree, I also became fascinated with satellite imagery and could see that the full potential of satellite data was not being used to map forests and forest loss. I won funding to do a PhD at the University of Edinburgh, where I developed new methods of using a type of satellite called a radar satellite, which can see through clouds and through the top of trees to see the structure of forests, to produce maps of the carbon storage of forests.
I developed and tested my methods with fieldwork in central Africa, with long field campaigns in Cameroon, Gabon and Uganda. This led not just to scientific advances, but it was fascinating and inspirational to meet people who lived in and around these threatened forests.
This led me to understand better that small amounts of funding from rich developed countries could make a massive difference in these people’s lives, for example by converting them from slash-and-burn and subsistence agriculture to growing agroforestry cash crops (like cacao), or funding education.
That small amounts of capital could help people, while also reducing the pressure on wildlife and forests, is obvious when you visit these places. I became very interested in how satellite data could help unlock the flow of finance by providing planning and monitoring data for these communities.
I continued working at the University of Edinburgh after my PhD, building up a research group that developed and tested new ways of mapping the carbon stored in tropical forests and peatlands, and how it is changing. I ended up publishing almost 100 peer-reviewed papers and winning over $8 million in grant funding, and I did some fascinating research.
But I was not moving the needle on stopping deforestation. So in 2018, I founded Space Intelligence with my colleague Dr Murray Collins, with the aim of bringing the latest high-integrity satellite forest monitoring technology, combined with our expertise brought about by years of fieldwork in tropical countries, to the forest carbon markets.
Could you describe Space Intelligence’s technology?
We use different types of satellite data, from optical, radar and LiDAR satellites, to create maps of landcover and carbon storage. These maps are very high accuracy because they are developed specifically for each project area given to us by our clients: ‘audit quality’, suitable for going through the validation and verification process of carbon standards.
Our maps involve looking at each 20m x 20m pixel about 200 times through a year from different types of satellite and over the full range of seasons. Our advanced cloud-based machine learning methods can obtain rich information from this dense stack of data, leading to high-quality outputs compared to using a single image mosaic. For example, grassland and cropland might look very similar at the peak of the growing season, but our multi-date imagery sees that cropland is suddenly harvested at some point and uses that to classify those pixels separately to grassland.
Our clients are typically developers of avoided deforestation, avoided degradation or forest restoration projects, mostly under Verra’s Verified Carbon Standard. They include large funders of carbon offset projects such as Apple and Everland, as well as high-quality project developers such as Green Gold Forestry (who have an avoided degradation project in Peru, using our CarbonMapper product to look for illegal degradation in and around their project area and track the recovery of forest from past logging).
Can you tell us a little bit about what makes the company’s approach to Digital MRV different from others in the field?
We use our deep knowledge of tropical forests and satellite data analysis to create custom maps for each area we’re asking to look at. These cover the baseline period before the present and annual Monitoring, Reporting and Verification data once a project starts.
We focus on two main products: HabitatMapper assesses the landcover type of each pixel (e.g. classes such as ‘intact forest’, ‘degraded forest’, ‘grassland’, ‘crops’, ‘water’ etc); CarbonMapper gives each pixel a value of the amount of carbon it contains, giving information on the differences in carbon storage within a forest class, and enabling that pixel to be tracked through time to see its growth in carbon, or losses due to degradation.
Others often use generic maps produced at a global scale, which cannot approach the accuracy needed to bring full integrity to the project monitoring or the carbon markets.
Space Intelligence was recently named as primary carbon data partner by Everland. What other partners of yours also benefit from your carbon data and how would you describe the development of interest in carbon credits from your perspective?
We see ever-growing interest in our services: 99% of tropical forests are not currently under REDD+ projects, and there is therefore great opportunity for increased REDD+ project coverage. With tropical deforestation currently over 1%/year, there is a great need! There are also large areas of the world available for restoration, but this must be done carefully to ensure the right species are planted, and those areas were cleared long ago, and not just cleared in order to be paid for restoration. In both cases, our data on land cover and carbon are needed.
We supply maps directly to clients in geographic data formats, but also provide access to a web platform. We are finding that clients are increasingly interested in web platform delivery, allowing them and their customers to directly interact with our high-quality mapping data without needing to download large files, or use specialist software. Our clients are also very interested in the prospect of Digital MRV, being able to use our web platform to provide data directly to auditors or Verra.
Recently there has been much controversy regarding the quality of carbon credits, particularly forest carbon credits, as well. What do you believe is the main reason for faulty credits and how does Space Intelligence contribute to mitigating the issue?
I have visited many REDD+ projects around the tropics and have only good experiences: projects seem to be raising incomes and opportunities for local people while protecting forests and enhancing biodiversity. And most projects do seem to be doing a good job of keeping deforestation rates low within their project boundaries.
But there are controversies regarding the setting of baselines. It is impossible to know what would happen to an area of forest if a project didn’t exist: nobody can predict the future accurately. But, given that, baselines should be conservative, and I think there is some evidence that some projects were not conservative. This should be corrected through time under the current approaches, but I think in future it is more likely baselines will be more conservative to start with.
In general the best solution to ensuring trust in the forest carbon markets is to base projects on high-quality datasets, including satellite-based maps such as those that Space Intelligence provides, and also to ensure those maps and other information about the projects is available on transparent and easy-to-use web platforms.
Carbon offset projects are beneficial to the planet, involve a massive amount of data collection, checking, and auditing, and have nothing to hide: the data that show that should be made public. Integrity and transparency go hand in hand in my view.
As you have literally the highest perspective, can you already see any positive effects of forest conservation and restoration efforts?
If one looks at REDD+ projects from satellites one normally sees a sea of green: within the project boundaries forests are protected. I have also seen project areas getting greener through time as restoration projects work. As part of our work with Space Intelligence I want such satellite images and maps to be much more widely available.
What is in the pipeline for Space Intelligence?
Just this week Apple announced that Space Intelligence’s Carbon and Habitat Mapper products were chosen to construct habitat and forest carbon maps of Apple’s Restore Fund project areas. This serves as a major validation point – that one of the biggest and most influential tech companies in the world has chosen to use Space Intelligence’s technology in this new initiative – and we’re thrilled to be a part of it.