Sources/Usage: Public Domain. View Media Details

As ecosystems in the high latitudes green or brown because of climate variability, Landsat’s rich archive of data—reprocessed in recent years to align and bring consistency to millions of observations that were collected using differing technologies from the 1970s, 80s and 90s—is revealing those transitions in ways no other satellite system can, said Woodcock, a renowned remote-sensing expert in Boston University’s Department of Earth and Environment.

It’s capturing more near-real time changes as well. When gypsy moths began defoliating wide swaths of forest in New England in the last few years, Woodcock said Landsat time series provided significant detail about the magnitude and location of the damage, and in a timely fashion.

“Time series data allow observations over a longer period of time to be able to detect subtle trends,” he said. “It takes a really high-quality dataset consistently processed through time to be able to recognize that and to document it.”

Landsat is that dataset. But its time series contributions are really only beginning to be unleashed, Woodcock said.

Open data proved a ground-breaking moment in the history of remote sensing. A mere 15 years ago, maybe 7 or 8 percent of Landsat images ever acquired had even been processed. Before Landsat data became freely available in 2008, it cost users hundreds or thousands of dollars to access the imagery. Decisions on which images to use were based largely on clarity and finances. And once those decisions were made, each user needed their images processed to meet their own specifications, typically making the comparison of two images of the same patch of ground difficult.

Now the smoothing out and correcting of small but important differences between today’s observations and those acquired by past Landsat sensors has resulted in a collection of analysis ready data (ARD) that spans virtually the entire Landsat archive and brings consistency across time, space, and those different sensors.

The creation of a first and then second collection of Landsat data, in which all the archive was processed each time using the same methods, “has been huge,” Woodcock said. But equally important with that consistent processing are the leaps in technology that have improved processing abilities—moves that Woodcock said support users’ ability to study large areas now much more effectively and efficiently.

Improvements in registering the images mean they line up better through time. And automation has made it possible to find the noise of clouds, shadows, and snow in those images and screen them. It’s enabled atmospheric corrections as well.

“All of those are huge,” Woodcock said. “So, EROS deserves a lot of credit, and USGS, too, because the quality of processing has gone way up in addition to making the data freely available.”

Having produced ARD for the United States, Woodcock said the next step is to convince the USGS to go global with it. As a member of the Landsat Science Team, he and others on the team have been pushing the USGS to do that, though without much success yet.

“On the (Landsat) Science Team, we’re sort of all about advancing the science, but also about advocating for the broader community of users,” he said. “And this one to date, to be honest, we’re failing the community. But we’ll keep after it.”

The pending release of the second collection of Landsat data, called Collection 2, goes beyond Collection 1 in that it will deliver on-demand global surface reflectance and surface temperature from 1984 to the present. It also adds more scenes to the highest-quality Tier-1 inventory and is available in a cloud-friendly format that allows users to work with the Landsat archive without downloading scenes.

By taking out differences in sun angles, aerosol scattering, thin clouds, and other atmospheric effects, surface reflectance provides a consistent characterization of the Earth’s surface over time and makes it easier to document change. And global surface temperature will enable scientists to track three decades of temperature change for every 30-by-30-meter plot of ground on the planet.

If Collection 2 can be integrated with the European Space Agency’s Landsat-like Sentinel-2 data and data from other similar satellite platforms, the potential is enormous, Woodcock said. The days of using time-tested paradigms for the image classification of such things as forests, crops, and grasslands, or for looking for change based on two images, would give way to tracking that change in near-real time using thousands and millions of images.

“For time series analysis, we’re using lots of observations. And we’ve only been at this for maybe 10 years,” Woodcock said. “Ten years sounds like a long time, but relative to the sophistication of image classification and these two-date change detection methods, I think we’re still kind of in our infancy. There’s a lot of progress to come.”

The possibilities are great, he said. The more data available, the closer we come to near-real time looks at changing landscapes, and the faster people will get information about the land they manage and care about, he said. That information, Woodcock said, will lead to such things as better land and pest management, reduced illegal logging in the Tropics, and more.

“Let’s be clear,” he said. “I think what we’re all after is the ability to sort through all available data streams to monitor environmental change in as close to near real time as possible. Go all the way down the list. The possibilities are sitting there, and they’re ripe. I’m hoping the USGS will take the leadership role in this, but if not, somebody else is going to do it. It’s inevitable.