Advertisement

SKIP ADVERTISEMENT

Green Column

Mapping the World's Problems

SAN FRANCISCO — Nearly a decade ago, an environmental group in Brazil grew concerned that government data and maps about Amazon deforestation were out of date and hard to view. The group, Imazon, decided to create its own monitoring tools, using information from satellites.

Imazon’s efforts caught the attention of Google, the search engine giant. Now, monthly reports on the Brazilian Amazon are produced through Google Earth Engine, a technology platform within the company. The partnership has made data processing faster and the information more accessible, according to Carlos Souza Jr., a senior researcher at Imazon.

“What Google could do is to make things easier,” Dr. Souza said.

Imazon was the first data-mapping project for Google Earth Engine, but many more are in the works, tackling problems as varied as overfishing and predicting the spread of disease. The platform uses satellite imagery to provide visualizations for environmental and health issues. The satellites offer a trove of data that dates back decades and also continually updates.

Future projects include mapping the locations of fishing vessels, which can help detect illegal fishing; tracing the spread of malaria; showing sea-ice melt; and depicting fires, such as the flaring of natural gas from North Dakota’s shale fields. The European Union’s Joint Research Center, working with Google Earth Engine, presented a paper last month on mapping the planet’s permanent and seasonal water sources. Other projects, to be presented at the United Nations’ conference in Japan this March on disaster risk reduction, include a tool that will help model and map rising sea levels as they affect areas that differ in elevation.

“There’s just a wonderful set of applications,” said Rebecca Moore, a com-puter scientist who leads Google Earth Engine. Inspired by Imazon, she and her team spent the first several years focused largely on mapping global forests to make sure the platform would run smoothly. But now applications are proliferating. She describes the program as “like an analytical engine for the earth, an environmental engine” that can take in an array of data feeds, such as satellite imagery, weather and time, and use them to gain insights.

The projects are done in partnership with scientists, who do the core work of building models and interpreting the data, while Google Earth Engine provides computing heft and help with presenting the data.

Image
An analysis of forest canopy damage in part of Brazil, produced by Imazon and Google Earth Engine.Credit...Google

One coming project uses satellite data to map where ships are traveling so that ships in no-fishing zones can be spotted. The satellites register signals from a tracking system that vessels over a certain size are required, for safety reasons, to carry. “We can see them leave port, where they go, how long they stay out,” said Jacqueline Savitz, a vice president for Oceana, an environmental group that is pursuing the project with Google Earth Engine and the environmental nonprofit organization SkyTruth. The groups demonstrated a two-year pilot in Sydney, Australia, last month. How much of the data will ultimately be made public is still under discussion, Ms. Savitz said, though the public could expect to find information such as the flags of fishing vessels in certain areas.

For scientists, one benefit of working with Google is the enormous computing power. To create maps of global deforestation between 2000 and 2012, done in conjunction with University of Maryland scientists, Google used nearly 700,000 satellite images and processed the data on 10,000 computers in parallel. “We had the result in a couple of days, whereas on a single computer it would have taken more than 15 years,” Ms. Moore said.

Google also has experience in presenting complex data clearly and simply, scientists said. In many fields, data mapping is not new, but local governmental and nongovernmental officials find existing maps and other tools difficult to gain access to or out of date.

“What is missing is the ability to update them regularly and the ability for nonexperts to run them,” said Hugh Sturrock, a spatial epidemiologist at the School of Medicine at the University of California, San Francisco. He is working with Google Earth Engine on a pilot project to map the spread of malaria. The project is to begin in a few months in the small African nation of Swaziland. The goal is to identify environmental factors such as proximity to water bodies, temperature and vegetation density that might aid the spread of the disease. With such information, researchers could potentially predict how it would spread and be better able to contain it. Dr. Sturrock hopes to extend the project to other nations, including the Solomon Islands and Vanuatu.

Not everyone is a fan of such methods. When research two years ago, led by University of Maryland academics using the Google Earth Engine platform, cited Indonesia as having the most rapid deforestation of any nation from 2000 to 2012, Indonesian officials pushed back. The map did not account for “the country’s temporary deforestation,” the secretary general of the forestry ministry, Hadi Daryanto, told The Jakarta Post. (Ms. Moore, of Google, dismissed the concerns, saying that additional scientific analysis that excluded tree plantations still showed Indonesia as having the world’s most rapid deforestation.)

But many scientists praise the data-mapping capabilities. “Being able to monitor by satellite provides a very, very useful tool,” said Rita Colwell, a distinguished university professor in the University of Maryland’s cell biology and molecular genetics department, who has not worked with Google Earth Engine, though she has worked with another part of the company. It can help the medical authorities to predict, rather than respond to, outbreaks, she said.

Satellite data will soon lead to a map of global sea ice, as well as a visualization of the world’s fires. Both projects should be ready this year, according to Randy Sargent, a scientist at Carnegie Mellon University who is working on them at Google Earth Engine.

“When you visualize it, you can get it at the gut level,” he said. “You can see it happening.”

A version of this article appears in print on   in The New York Times International Edition. Order Reprints | Today’s Paper | Subscribe

Advertisement

SKIP ADVERTISEMENT