Rolf Schmitz is co-founder and co-CEO of CollectiveCrunch, a platform that is changing the understanding of forests worldwide by providing the most accurate, scalable and timely analytics globally and enabling sustainable forestry and bringing transparency to carbon trading markets.
Rolf is an engineer by training and holds an MBA from Manchester Business School. He has deep experience in global business development and sales, having built teams in Asia, the US and Europe.
Could you share the genesis story behind CollectiveCrunch?
We are involved in managing large amounts of data and extracting insights from it. Our initial idea when we started CollectiveCrunch was to combine climate data with business processes, as we felt it was an overlooked aspect of climate change.
Initially, I pursued logistics and energy. We have built a product that predicts the generation of energy from wind farms, which is critical to maintaining the stability of energy grids. The product is active on Fingrid, Finland’s national grid. However, we found crowded logistics and energy markets where it would be difficult for a small company to build a leadership role.
Through a friend of Jarkko, one of our co-founders, we became aware of the challenges in creating and maintaining forest inventories. We thought there was a shockingly low level of technical sophistication. As a result, inventories were expensive, inaccurate, and only done every 5-10 years. The importance of forests in climate change mitigation, ecosystem services and nature-based solutions was clear at the time. That’s how CollectiveCrunch became a “Forestry Artificial Intelligence Company.” On a personal level, we all grew up in the countryside, so we had a natural affinity for forests. That’s how we ended up building AI models for forests.
What types of instruments and cameras are used to monitor a forest?
Our approach is not to specialize in any one sensory method, but to combine all the relevant data sources we can get our hands on. Any sensory method has strengths and weaknesses; combining data sources allows us to counter weaknesses. For example, optical images are very useful, but are not available from satellites when there is cloud cover. In our business, satellite data is important, but so are LIDAR scans as they are available. From a business model perspective, we do not engage in data acquisition such as flying drones or renting aircraft to scan areas.
Apart from the array of satellite-based sensory data, LIDAR is a very important tool or method. High-resolution optical imagery from surface surveys is less prominent than LIDAR, but also used. One tool that is surprisingly still in use is the old 19th century method of hand-picked samples. With a lot of stats involved, I’d still call it a tool.
Is the system capable of being trained for different localized ecosystems to identify pathogen infections, anomalies and disorders or other types of tree diseases?
There is adaptation for different regional ecosystems, including change detection. Tree species, growth patterns and forest management practices vary greatly from region to region. The same applies to data acquisition methods and practices. So not only the trees but also the training data are different.
What kind of useful information can be obtained from this information?
- Grouped under the term “change detection”, you have the detection of storm damage, the identification of pest outbreaks and other negative impacts that require intervention to enable field intervention and limit the impact of the damage in question.
- Carbon inventories bring transparency to carbon projects and facilitate decisions on the evaluation and purchase of such projects and credits.
- In reforestation projects, the viability of newly planted trees depends on the right amount of soil moisture. Detection of excessive dryness or moisture can trigger intervention to prevent these young trees from failing.
- Forest inventories from commercial forestry inform decisions such as thinning (which stimulates growth) and crop optimization. Species detection makes the supply chain more efficient and increases margins. Together, this allows the industry to use forest resources more efficiently. This is crucial as much of the commercial forest is essential for supporting rural communities and driving the adoption of circular products and packaging.
- Monitoring biodiversity can trigger intervention if an area is suffering from degradation. Biodiversity is crucial for our forests to become more resilient as we move through this phase of accelerating climate change.
How does analytics benefit sustainable forest ownership?
Several benefits came into play. First, commercial forestry is constantly adopting new measures to become more sustainable. Many of these require better and deeper analysis. For example: logging, where an area of forest is 100% cut down, has a strong impact on the local ecosystem. It is done for reasons of efficiency – many sustainable products, such as fibre-based packaging, could not compete with less sustainable alternatives if the forest industry became less efficient. The industry is exploring alternatives where only the largest trees in each area are cut. It is much more durable, but from a logistics and cost perspective it is a very serious challenge. And it can only be done with state-of-the-art analytics.
Biodiversity is essential for forest resilience. Monitoring biodiversity and activating interventions where necessary is crucial for the short- and long-term viability of the forest.
For carbon capture projects, how does the system verify that a project reduces greenhouse gas emissions as advertised?
The system achieves a certain accuracy for the forest inventory in question, which is verifiable. Most of the greenwashing does not happen at the level of analysis, but in the way projects are structured. Forest carbon projects that aim to avoid deforestation suffer mainly from two problems:
- Baselines: This is the set of assumptions that project what would happen without the intervention. The intervention is then calculated as “additionality” above the baseline. Today’s benchmarks do not come from a data-driven analysis, but are often raw averages. Moreover, the baseline is calculated by the project managers themselves, who are in a conflict of interest: the lower the baseline, the more credits are created.
- Spillover: The phenomenon whereby positive things that happen within the project’s defined areas (such as reduced logging) are counterbalanced by what happens outside the project’s defined area. Very often, such areas are not tracked, so the project receives credits while the benefit is lost to the surrounding forests.
The fundamental problem here is that there is a lack of data-driven analysis to independently track what is happening. It is possible today, we can do it at scale, but there is very slow adoption of the latest technology in this area. In short, the problem is not the analysis, but what the credit calculation is based on.
Do you have any case studies you can share with customers using this system?
- ENCE, the largest forest owner in Spain, uses our system.
- Our first and biggest customer is Metsähallitus (Finnish State Forest).
- Our partner Forliance, one of the largest and most respected carbon project managers globally, is working with us on one of the largest carbon projects in Colombia.
- 7 of the top 10 forestry countries in the Nordic European countries are our customers. The latest addition is the Metsä Group, one of Finland’s “big 3”.
What is your vision for the future of forest conservation?
Our vision is data-driven with fact-based analytics in nature-based solutions. It is very clear that we need to move quickly to mitigate climate change. Currently, the vast number of forests on the globe are inventoried every 5-10 years. We should reduce this to monthly tracking to understand what is happening. In addition, we need to watch biodiversity. Without biodiversity, we lose the resilience of our forests in the midst of a climate crisis.
Is there anything else you’d like to share about CollectiveCrunch?
Yes: we can do this at scale. We currently cover 20 million hectares, about 50 million acres of forest. We do this with better precision than the conventional methods we replace. This is real and enables transparency in carbon trading markets.
Thanks for the great interview, readers who want to know more should visit CollectiveCrunch.