The week of April 17, 2016

Can satellites help save the world’s rainforests?

By Nithin Coca

The rainforests are the world’s lungs. They provide the oxygen we breathe, suck up carbon from the atmosphere, and host an incredible amount of biodiversity. Yet over the past half-century, they’ve been destroyed at a rapid clip. And it’s only getting worse. Just last year, fires precipitated by a long dry season and the effects of El Niño burned 2 million hectares of Indonesian rainforest, emitting a dangerous haze that killed at least 19 people, while shortening the life spans of countless others among the 43 million impacted around Southeast Asia.

“At one point, the daily emissions were near the same level as the daily emissions of the United States. This was very shocking news for us as Indonesians,” said Teguh Surya, a forest campaigner with Greenpeace Indonesia. “Singapore, Malaysia, southern Thailand, and the Philippines were all affected.”

For many experts and activists, the fires came as no surprise. Years of rampant deforestation, the draining of wetlands, and the expansion of palm oil, timber, and paper pulp plantations had left the landscape scarred, unnaturally dry, and primed for disaster. Indonesia consumes little of the palm oil or paper pulp produced on those plantations; almost all of it goes to international markets, including the United States. “Much of the vital industry supporting global markets are in my country,” said Surya. “Raw materials—wood, timber, pulp and paper, palm oil, all from Indonesia.” Palm oil, for example, is now the top consumed oil in the U.S., due to health concerns about hydrogenated oils.

Besides simply destroying the forests and sickening the population, the fires emitted as much greenhouse gases into the atmosphere as Brazil does in a year. And typically, emissions from agriculture and land-use changes accounts for about 25 percent of global CO2 output. It’s a category that is “low-hanging fruit” for reduction. Quite simply, if we stop burning and clearing forests, we could make a big, immediate dent in the global production of carbon dioxide.

While the fires burned, satellites overhead were gathering data. Thanks to those eyes in the skies, we have a nearly day-to-day account of what happened in Indonesia, from when the hotspots, or potential fires, first emerged to when dangerous haze began forcing schools to close in Malaysia and Singapore.

Leading this tracking effort was the World Resources Institute (WRI). Its freely available Global Forest Watch tool incorporates a diverse range of satellite data and presents it in a user-friendly way. “In the last five years, the advances in cloud computing technology made it possible to have algorithms set up to analyze huge amounts of landsat data, which was not possible with computing power until recently,” said Susan Minnemeyer, mapping and data manager for WRI’s Forests Program.

“Cloud computing technology made it possible to have algorithms set up to analyze huge amounts of landsat data, which was not possible with computing power until recently.”

In the coming years, satellites are going to get a lot better, and it’s happening fast. Earlier this month, WRI announced a new tool, covering Peru, the Democratic Republic of the Congo, and Indonesia, that automatically analyzes satellite imagery and sends weekly updates about potential deforestation across all three continents—analysis that previously took months, if not years.

“This data set is really important,” said Mikaela Weisse, data and website research assistant for WRI’s Global Forest Watch. “This is the first opportunity we’ve had to show that technology can actually stop deforestation as it’s happening, and really be able to improve forest monitoring.” WRI hopes to expand the tool to all of the tropics soon, and the entire globe by the end of this year.

A few factors are driving this change. Primarily, there are more satellites in orbit with better monitoring capabilities. That means more satellites going over a given piece of land and taking better images more frequently. That addresses one of the biggest challenges when trying to monitor, for example, the tropics: cloud cover.

“In the tropics, the cloud cover is going to be close to 90 percent of time,” said Emiliano Kargieman, CEO of Satellogic. “That means if you have one satellite that goes over one point in Indonesia every three days, it will take you a month to get a good image.” That’s obviously unhelpful for timely monitoring, and the company’s building a real-time analytics engine based on its proprietary constellation of high-resolution satellites. “The advantage of having a large constellation of satellites is that you’re going over these spots several times a day, so when there’s an opening in the clouds, we can take a picture.”

The latest satellites also have much more sophisticated technology on board and are able to capture a wide spectrum of data the reveals details hidden to the human eye. “We can look at landscapes in a way where we can see some of the chemical processes. That becomes very helpful for analyzing change tracking, or processes that take days or week to develop,” said Kargieman. Satellogic says their technology can allow them to see down to the level of individual trees, even capturing minute-level changes invisible to humans.

Of course, all this information gathered creates a data problem: how to make useful the massive amounts of imagery captured at increasingly smaller and smaller time intervals. Enter cloud computing and companies like Descartes Labs, which are applying deep learning to multispectral data to make a living, breathing atlas of the world.

“I’d like to look at understanding the environment, including snowpack, water, forests, how cities grow into forests. And not looking at any of these factors in isolation, but actually looking at whole data sets together to see what correspondence we can see,” said Mark Johnson, CEO of Descartes Labs. “As we develop these data sets, we’re going to notice that human activity affects the environment in various ways, some of which we know, some of which we intuitively believe, and some of which will be a surprise.”

“We’re going to notice that human activity affects the environment in various ways, some of which we know, some of which we intuitively believe, and some of which will be a surprise.”

Better understanding and better data is one thing. But using this knowledge to inform policy and practice is another. Will all these better, more numerous, and more informative satellites make any difference? The fires were expected and predictable—the product of a certain kind of economic thinking, driven by market forces. Similarly, illegal deforestation is no secret: It’s been known about for years, yet it continues, not because of lack of data, but because of rampant corruption, lack of effective local governance, and willful negligence by companies all along the supply chain.

Satellites alone cannot solve the problem. “There’s a lot of information, but that doesn’t alone create action, and that’s a big challenge,” said Weisse. “We really need to get these tools into the hands of those who make decisions about landscapes.” That’s not easy, but simple tools, such as an accessible, easy-to-understand website, can help. According to Weisse, traffic to WRI’s Global Forest Watch is still mostly from the United States, but an increasing amount is from tropical countries. The recent launch of an Indonesian-language version, a joint venture with Greenpeace, helps get this data to those who can best use it.

Satellites can also provide data about supply chains, helping companies track the effects their suppliers have on the environment. “Ultimately in 10 years, we’ll have imagery of the entire earth all of the time, which opens up the possibility to track supply chains,” said Jeff Stein, vice president of business development at Orbital Insights. WRI has already partnered with Unilever to monitor for deforestation surrounding the global giant’s suppliers.

Similar satellite monitoring can also help provide accountability for carbon emissions, removing one barrier to meaningful climate-change agreements: a lack of trust. If countries were self-reporting emissions, what would hold back, say, China, with its notorious record of fudging economic data, from fudging climate data? But now that satellites can monitor emissions from every country from space, that issue becomes moot. Even notoriously unreliable carbon credits could be monitored in real time.

“We can now give credit where credit is due to countries preserving their forests, and provide transparency about countries contributing to global warming,” said Stein.

What it all amounts to is a better understanding of who’s doing what to the earth, where, and when. If a major hurdle to mitigating rainforest destruction has been a lack of data, satellites can make the picture clearer; similarly, those circling observers help clarify the view of global carbon emissions. Granted, an improved understanding of where we stand and why doesn’t guarantee action, but at least it’s a start—one made possible by all those eyes watching high overhead.

Illustration via J. Longo