A cutting-edge satellite-based alert system could help policymakers and conservationists put a dent in illegal logging by notifying users in real time of new bald patches in the world's rainforests.
The system, known as the Global Land Analysis and Discovery alert system, was developed by the Department of Geographical Sciences at the University of Maryland and Google, and uses an algorithm to analyze satellite images of tropical rainforests in Peru, the Republic of Congo and Indonesian Borneo in more precise detail than ever before. The goal is to provide high-resolution tree loss data across the most vulnerable swaths of forests, potentially helping researchers and officials catch illegal logging before too much damage is done.
Healthy trees absorb vast amounts of carbon dioxide. So while illegal logging devastates biodiversity and robs local communities of their economic benefits, deforestation also accelerates climate change, since forest degradation is the second-largest contributor of global carbon emissions.
"You can't solve the problem of climate change if you don't solve the problem of deforestation," said William Boyd, the senior advisor and project lead for the Governors' Climate and Forests Task Force, an international collaboration developing regulatory frameworks to reduce emissions from deforestation. "I think more information and the ability in real time to make visible these episodes of deforestation are quite important. They're a foundation for taking real action and getting an understanding of deforestation in a particular place," he added.
The weekly alerts are presented publicly by Global Forest Watch, an online forest monitoring and alert system originally launched by the World Resources Institute in 1997. Back then, satellite images were expensive and difficult to analyze. The project was retooled in 2014 when Google partnered with WRI, bringing with them "the power of large scale cloud computing," said James Anderson, communications manager for the forest team at WRI. The new system is now faster and more precise than ever.
"The USGS (U.S. Geological Service) and NASA had released the entire Landsat archives and the real challenge was, no one had the bandwidth to download and analyze all those images. Then Google came along," Anderson said.
A 2012 World Bank study estimated an area of forest the size of a football field is clearcut by illegal loggers every two seconds. And according to a report done for REDD+, or Reducing Emissions from Deforestation and Degradation, agriculture drives 80 percent of deforestation worldwide, with commercial agriculture, such as cattle ranching, soybean farming and oil palm plantations, playing a growing role in clear-cut logging.
Researchers say they chose these three countries because they represent a wide range of deforestation conditions. According to a methodology paper linked with the project, the Republic of Congo, for example, "is a high forest, low deforestation rate country with selective logging and smallholder agriculture being the primary drivers of forest loss." Indonesia, on the other hand, is a high deforestation rate country and Peru falls in the middle, although it has been trending toward, "increasingly high rates of forest conversion consisting of smallholder agriculture, artisanal gold mining and industrial agriculture, mainly palm oil."
The platform and the paper were funded by several organizations tied to forestry, such as Norway's International Climate and Forest Initiative and the United States Interagency SilvaCarbon program, a cooperative effort that aims to improve access to forest data.
Illegal logging has been difficult to prevent or even catch because most of it occurs deep in rainforest, hidden from forest law enforcement officials.
Prior to the GLAD alerts, conservationists and governments had to wait weeks or even months in order to obtain cloud-free satellite images from the U.S. Geological Survey clear enough for them to pick out tree cover loss, stymying their efforts to combat deforestation.
Mikaela Weisse, the data and website research assistant at Global Forest Watch, said the system downloads satellite images from USGS satellites to a program running an algorithm that compares each pixel in the new image to the previous four years of images. If it senses a significant difference in the pixel's patterns, an alert is triggered. Those alerts are compiled every week and then made available on Global Forest Watch at high resolution.
The images are now clear enough to detect new roads and pathways beginning to snake through uncut primary forests, usually a sign that logging or illegal gold mining is starting, but the alerts also highlight what are likely new patches of cleared forest in bright pink. Previously, images were not detailed enough to catch loss from logging or mining.
"It's one thing to have an NGO know where deforestation is happening and be able to call it out. I think it's another thing for law enforcement officials to use that and take action through the legal process to actually stop this deforestation from happening," Weisse said.
Weisse said they've been working with groups in each of the three test countries to slow illegal activities. In the Republic of Congo they're engaging with groups advancing the Forest Law Enforcement, Governance and Trade Action Plan, an initiative to reduce illegal logging. They said their partnerships are most advanced in Peru, where they've been working with the Ministry of the Environment, which hopes to use the data to monitor illegal activities inside protected areas and community lands. And in Indonesia they've been working with NGOs and companies interested in producing products that come from a deforestation-free supply chains, Weisse said.
"There is no guarantee the alerts will be taken seriously, but their fine resolution and and high accuracy means that the data does a better job of showing areas of clearing than other existing alert products," Weisse said.
The GLAD alert system, is "blowing our minds," said Matt Finer, a research specialist at the Amazon Conservation Association, a nonprofit using the data to investigate deforestation hotspots across Peru. Because tropical rainforests tend to be coated in mist, previous projects could take months before Finer and his team collected enough usable images. Then they had to painstakingly sift through dozens of satellite images, trying to calculate the amount of forest loss.
"These new alerts are so good they're kind of scary. It just let's us focus on the next phase of the research, which is trying to understand the patterns and drivers of deforestation," Finer said.