Partnered content
The science is clear. We have seven years left to cut emissions at the scale required to have a chance of leaving a liveable planet for future generations. The most recent IPCC report was the latest in a line of increasingly urgent calls for action. The report confirms the critical role that the protection and restoration of nature plays to achieve that goal. In fact, the IPCC confirms that halting the destruction of intact ecosystems - forests in particular - constitutes the third most impactful climate solution we have at hand today, right after wind and solar energy.
The challenge is daunting, but there is hope. There has never been more business interest in the conservation and restoration of nature than there is today. Companies are taking unprecedented action via the growing voluntary carbon market, their supply chains and investment portfolios, as part of their commitments to transitioning their businesses to net-zero and nature-positive. Meanwhile reporting frameworks, regulators and investors are asking for better disclosure on the climate impact of actions taken to protect and restore nature. This is encouraging.
But there is a catch: our current carbon accounting systems are not fit for purpose. Without validation, accountability or measurability, a company's sustainability data can be compromised. This can raise questions about the credibility of a company's natural capital investments, be it via the carbon market or its supply chains. Current accounting systems do not capture the emission reductions and removals achieved by the above companies' attempts to protect and restore nature with the integrity, reliability, speed and scalability that the transition to net-zero and nature-positivity requires.
Recent advances in Earth Observation science have unlocked new solutions that are now being introduced to the market. Algorithmic in nature, they have the potential to transform carbon and natural capital accounting by directly estimating the carbon removed from our atmosphere by nature-based solutions (more on the meaning and importance of direct estimates below). Businesses, standards and certification bodies need to embrace these technological advances as important enablers towards a net-zero, nature-positive economy.
Here are three reflections on why this matters, and how companies working with algorithmic Earth Observation technologies can take action to infuse their solutions with trust for improved carbon accounting.
1. It matters because we need to see what the atmosphere sees
Satellite images have been used for many years to estimate CO2 emissions from land use conversion, for example when a piece of the Amazon rainforest is cleared to make space for agricultural land. However, these approaches only capture emissions from tree cover loss that triggers a change in land use classification. They do not directly estimate biomass loss. This has at least two major limitations: firstly, this approach can't measure forest degradation (e.g. small scale logging that does not trigger a change in land use classification under a strict definition of deforestation). As forest degradation can cause as much carbon emissions as deforestation, this means that these standard approaches may be blind to 50% of the problem. Secondly, they can't see the positive carbon impacts of forest growth and regrowth.
This means that companies can't show off the positive impacts of their investments with confidence. For a credible transition to net-zero, we need to power our carbon accounting systems with measurement solutions that overcome these shortcomings. To assess the climate impact of investments, we need monitoring solutions that see what the atmosphere sees, and that aren't limited by definitions of what counts as land use change, or deforestation, which suffer from human or political biases.
2. The potential of Algorithmic Earth Observation
Advances in algorithmic Earth Observation technology now make it possible to directly estimate the carbon stored in trees for any piece of land, and how this changes over time (rather than relying on land cover classification). The most advanced Earth Observation approaches do this by fusing data received from various spaceborne sensors, including laser, radar, optical and ancillary data. Advanced algorithmic approaches quantify these ecological processes with unprecedented quality while unlocking the cost-effectiveness and data consistency that allow investors to truly account for the carbon stored in nature as a result of their action. It is a much needed step-change in the quality of carbon accounting technologies.
3. The corporate responsibility of algorithmic data providers: build and test with scientific integrity
Unleashing the potential of algorithmic carbon measurement solutions necessarily raises questions about the algorithm's training and validation regime. The accuracy of an algorithm's estimates depends upon the quality of the training data it receives. Feeding an algorithm with high-quality, high-integrity training data is essential for its performance. But this is not enough. The key question is how well the algorithm performs against fully independent, high-quality instruments estimating the same unit at a comparable scale. In the context of carbon stored in trees, the gold standard are estimates delivered by high-quality airborne laser instruments.
Figure 1: Comparison between independent airborne LiDAR (ALS) and Chloris derived estimates
At Chloris Geospatial, we firmly believe in the need to equip the world with trustworthy algorithms that are fit for purpose, and built with the highest standards of scientific integrity. That's why we compared our machine-learning derived predictions of forest carbon stock and change against such airborne laser-derived measurements carried out by NASA and other academic research institutions, and made our validation results publicly available. The result? Comparable levels of accuracy in measuring carbon stock changes, but at a fraction of the cost of airborne-lidar, and with the ability to perform measurements repeatedly and with consistency in space and time. This insight matters as it proves that algorithmic solutions can deliver the required data quality in a scalable and cost-effective way, and thereby contributing to keeping more funds available for the actual conservation and restoration activities on the ground.
In conclusion, as we strive towards a sustainable future, we must prioritise this level of transparency, data quality and cost-effectiveness to be able to achieve a high integrity transition to net-zero and nature-positive economies for all.
