AI can be a game-changer for the world’s forests. Here’s how

Despite a wealth of statistics about the value of our natural resources, we are essentially flying blind when it comes to their current state and how to best protect them. To fight climate change more effectively, while also feeding a growing population, protecting our ocean and freshwater resources, and stemming a global loss of biodiversity, we must close that information gap, and do so in a hurry.

Agriculture is the world’s largest industry, employing more than 1 billion people and generating over $1.3 trillion worth of food annually. But food production is growing more difficult, as arable land continues to decline, extreme weather wreaks havoc on predictable growing seasons, and climate change lowers the nutritional value from what is harvested. Our other land-based natural resources, such as forests, are in a similar state – forests shoulder the needs of 1.6 billion people, yet reports indicate we are losing 18.7 million acres a year.

This represents more than just a loss of natural resources, ecosystems and biodiversity, though that’s bad enough. Our current agriculture and land-use decisions are directly contributing to climate change, accounting for nearly 25% of the world’s total greenhouse gas emissions. Agriculture is one of the biggest drivers of deforestation, which is one of the top three anthropogenic sources of greenhouse gases – followed closely by agriculture itself.

Managing these resources more effectively improves the well-being of everyone on the planet and also has the potential to improve the planet itself. This presents an incredible opportunity – better management of our ecosystems and land could help feed a growing population, while substantially lowering carbon emissions. However, to act on that opportunity, we need a clearer picture of the current state of the planet’s natural systems; how they are changing and what the most effective intervention strategies are. Increasingly, this information gap will be filled by AI-enabled solutions.

That is why Microsoft created AI for Earth last year. This program makes Microsoft’s AI expertise and technology available to those working in the areas of biodiversity, water, agriculture and climate change, so that they can bring their solutions to scale and advance sustainability across the globe.

Forest management is a good example of how technology-first approaches can quickly deliver results. Conducting a forest inventory hasn’t changed much from the statistical sampling approach first introduced in the Nordic countries in the early 1900s. Teams go out into the woods, armed with tape measures, pencils, and a pad of real paper, not a device. They measure the diameter, height and species of each tree in many small “sample plots” to estimate what’s in the forest as a whole.

However, there are start-ups devoted to finding better ways to do this inventory. SilviaTerra, an AI for Earth grantee, came up with a software-based approach after becoming frustrated with the status quo while at Yale School of Forestry. Now, their software can assess forests using satellite imagery and machine learning. The algorithm, powered by AI, greatly reduces the amount of fieldwork needed to accurately assess forests and is the vanguard of a new generation of “precision forestry”. Their goal is to build a data library and powerful AI tools that can provide an up-to-date map of US forests for the first time in history, with detailed information about each tree. This kind of information enables data-driven environmental management for biodiversity, carbon sequestration, and many other ecosystem services provided by forests.

Work by AI for Earth researchers at Columbia University sheds even more light on why accurate, detailed, and up-to-date information is important. Dr Maria Uriarte, an ecologist, and Dr Tian Zheng, a statistician, have been studying the impact of extreme weather on forests and their regrowth patterns, with an eye towards the impact this has on carbon sequestration abilities – shorter, younger and less dense forests are less effective than older, denser areas. She recently took a team to Puerto Rico to assess the damage to the forests following Hurricane Maria. Uriarte and Zheng, both affiliated with the Data Science Institute at Columbia, will eventually use the collected data, with the remote-sensing images and measurements, to come up with a detailed estimate of the loss from the storm. Without current baseline data and a forward-leaning view of what the forest inventory may be in the future, planners may undervalue forests, or countries may over-value sequestration abilities.

Projects like these are critically important to building a big-data view on a global scale. We believe the world needs many more organizations doing more work like this, which is why we are so excited about how quickly the portfolio of AI for Earth grantees has grown. Already, we have awarded 147 grants to individuals and organizations working in 47 countries.

This work has given us a bird’s-eye view of how powerful the combination of human ingenuity and AI is, and just how badly it is needed to make the kind of quick, effective and global progress this planet needs. The good news is that the technology already exists to collect and analyze the data pouring in from sensors, satellites, drones, citizen scientists, camera traps and many other sources. With the cloud and AI, we can process all this accurately and cost-effectively, to do things that were previously so difficult and expensive as to be impossible.

The progress we’ve seen on land-cover mapping is an illustrative example. In the United States, the best available data sets on land cover, at a resolution of 30 meters, were last updated nearly seven years ago. (Globally, the picture is much less complete.) In collaboration with the Chesapeake Bay Partnership, a small non-profit called the Chesapeake Conservancy set out to create a better model to provide the data needed to enable precision conservation within the 64,000 square-mile Chesapeake Bay watershed. Microsoft worked with the Conservancy’s data and with geospatial software giant Esri to build an algorithm capable of generating a land cover map of the entire United States in real time. While there is still much work to be done, we are excited by recent results in which specialized computing infrastructure in Azure produced a map of the entire US in less than 10 minutes for under $50.

AI can be a game-changer because taking actions are easier and more effective – and less vulnerable to politicization – if we know what is happening on Earth, when and where. The speed of innovation is one of the few things keeping pace with climate change. Harnessing the power of AI to monitor the impacts of our current land use practices and to model scenarios means that, perhaps for the first time, we can have the right information at our fingertips to more effectively and sustainably manage our lands, watersheds and ecosystems.

This article is part of the World Economic Forum’s Fourth Industrial Revolution for the Earth series, which explores how innovative technologies are beginning to transform the way we manage natural resources and address climate change and other environmental challenges caused by industrialization.