Watching, listening, and learning: how AI can reduce deforestation

Automation and machine learning are accelerating forest conservation efforts and improving effective decision making.

Maybe you’ve heard this expression. If a tree falls in the forest and no one is around to hear it, does it make a sound? With today’s technology, it is actually possible to ensure every fallen tree will not only be heard – but also measured, identified and accounted for. Artificial intelligence (AI) gives us access to information that would otherwise be unseen and unknown, and the advancements are coming fast. Here is our look at AI solutions that are disrupting the way our forests are managed and curbing global deforestation levels.

Value of a Forest

UN Sustainable Development Goal 15 calls for sustainable management of forests and the end of deforestation. There are several good reasons for this. Forests provide 75% of the world’s freshwater, and 1.6 billion people rely on forests for food. Forests are also incredibly important for climate change, as they are a carbon sink for over 2 billion tonnes of CO2 annually. In fact, 25% of greenhouse emissions are attributed to the clearing of forests for agriculture and other land use changes, making forest protection a very cost-effective way to combat climate change.

In the past 25 years, the rate of deforestation has slowed by 50% – but there is a long way to go. The world’s forests are continuing to shrink, and now forests only makeup 30% of the earth. Now several companies are looking to find the next forest protection solution, harnessing using the speed and capacity of AI technology.

Halting Illegal Logging with AI

Illegal logging is a major cause of deforestation around the world. It occurs when forests are cleared outside of legal allowances to meet demands for timber or agriculture. Rainforest Connection is using the open-source machine learning tool  Tensorflow to help forest communities improve their monitoring and protection by detecting illegal logging in real time. Solar-powered audio sensors are placed throughout the forest that can distinguish chainsaws from the forest’s canopy cacophony, and sends alerts to park rangers who can respond to the disturbance. Rainforest connection builds the sensors from discarded smartphones in the US and educates kids in the process.

Other solutions are looking to be even more proactive to halt illegal activity. Orbital Insights has been developing methods for predicting deforestation together with the World Resources Institute and Global Forest Watch, the open-source satellite monitoring of global forests. They have investigated the potential for deep learning AI to identify patterns, such as a new road being built and the development of oil palm plantations. Although the cohort notes some challenging barriers, they foresee the data-driven approach as key to reducing deforestation in the decades to come.

AI for Sustainable Monitoring Forest Management

Historically, forest management has been incredibly complex. In the past, large teams would spend hours mapping random samples of the forest with pen and paper to quantify the health of a forest. SilviaTerra instead uses satellite imagery and AI software to simplify and accelerate the process of creating a forest inventory. This allows forest managers, conservation groups, and landowners to make better decisions using measurable and comparable data. Finnish company TerraMonitor (formally Satelio) is also using satellite images to detect environmental changes throughout Nordic countries.

aiTree takes it a step further by using AI software to attempt carbon accounting and forest management, improving decisions on timber allocations and conserving natural areas for ecosystem services. These products have numerous implications for public entities looking to create policies around forest management, and for companies to access available resources to guide responsible practices.

Hope for Future Developments

There is tremendous room for more growth and scalability for applying AI to forest conservation. Many ideas are not receiving enough attention, and need proper resources to fully develop. For example, teams are looking at ways to identify tree species using drones, or using AI to assess storm damage felt by a forest. Major tech players such as Amazon, Google and Microsoft have tech accelerators aimed at getting new AI companies off the ground. Perhaps the strongest its Microsoft’s AI for Good Initiative, which funds and supports companies that are using AI for creating positive humanitarian and environmental change. At Sustainia we will be keeping an eye on these latest developments, and will work to shine a light on valuable new projects.