The slipping of Landsat 9’s original launch readiness date from December 2020 to September 2021 brings with it uncertainty to the Landsat user community, but no real panic at this point.
There are many operational applications of Landsat data—the U.S. Department of Agriculture’s cropland surveys, for example, and burn severity mapping done by the Department of Interior and the U.S. Forest Service—that rely on as much near-real time imagery as they can get. So, for cropland surveys in particular, there will be no Landsat 9 coverage in 2021 for the Northern Hemisphere growing season.
That said, Landsat 7 still circles the planet with its much more accomplished partner, Landsat 8. The two satellites continue to provide eight-day coverage of Earth’s land surface and with it, viable science data, even though Landsat 7 is an aging observatory with a broken scan line corrector and other single point failures on the spacecraft. Yet even if Landsat 7 falters, the science community has Europe’s Sentinel-2 satellites, with their ability to acquire Landsat-scale data, as well as commercial remote sensing options.
No one disputes that Landsat 7’s remaining lifetime is short. Its orbit is decaying. When its local solar time—the time when it crosses the equator on each of its orbital passes—reaches 9:15 a.m. Mean Local Time (MLT), the illumination levels will no longer be consistent with prior Landsat sensors whose past MLTs varied between 9:15 a.m. and the nominal 10 a.m. overpass throughout their missions’ lifetimes.
As a result, Landsat 7 data will need to be used with caution until scientifically accepted corrections for orbital drift can be applied—if possible. By the best NASA and USGS calculations, Landsat 7 will be at 9:23 a.m. on April 1, 2021, and will degrade to 8:52 a.m. on April 1, 2022. It won’t approach that critical 9:15 a.m. juncture until July 2021.
As for Sentinel-2 satellites, they don’t have the thermal infrared spectral bands that Landsat 7 and 8 do. But Europe’s satellites will still play a role in the Landsat user community’s strategies in 2021, especially as scientists continue to work on harmonizing and even fusing Landsat and Sentinel-2 data.
With all that in mind, seven USGS-NASA Landsat Science Team members offered their insights into how the delay of Landsat 9’s launch will affect the work they do in 2021.
Dr. Nima Pahlevan, remote sensing scientist with Science Systems and Applications Inc. and a contractor within the Terrestrial Information Systems Lab at NASA Goddard Space Flight Center
With the enhanced ground characterization and improvements in data quality expected with Landsat 9, “it is very unfortunate” that the aquatic science community and water resource specialists and managers have to wait another year for the suite of observations and corresponding products that will come with Landsat 8 and Landsat 9 flying together, Pahlevan said.
While he believes extending the Landsat 7 operations in order to collect imagery even earlier than a 9:15 a.m. MLT equatorial crossing will be beneficial for the land community, Pahlevan said Landsat 7 is not considered a critical contributor to the aquatic science community.
“The lower sun elevation angles reduce the water-leaving signal, rendering the utility of Landsat 7 data even more challenging for our community,” he said.
Sentinel-2 data, on the other hand, are essential to better understand the temporal variability in aquatic ecosystems. That said, even with Sentinel-2A and 2B contributing to operational water quality studies, Pahlevan expects challenges in 2021.
That’s because aquatic ecosystems situated in areas with frequent cloud coverage may not be well captured with Landsat 8, and even Sentinel-2A/B, he said. For example, he cited “one of our study areas, New York harbor, where our investigations on impacts of COVID-19 on water quality during the lockdown were hampered by clouds in the April-May timeframe.”
Dr. David Roy, Professor, Department of Geography, Michigan State University
In one breath, Roy called the delay of Landsat 9’s launch “unfortunate,” and noted that the “first light imagery” for vegetation and agricultural monitoring in the Southern Hemisphere when Landsat 9 goes up in September 2021 won’t become available until the summer of 2022.
“This,” he said, “will be an issue.”
But like a lot of his counterparts on the Landsat Science Team, Roy also insisted that Landsat 9’s eventual arrival will be worth the wait, and in half-jest added, “it will give me more time to get ready for the Landsat 9 data.”
Like other scientists, he has taken to supplementing Landsat imagery with Sentinel-2 data for work he’s doing on mapping burned areas. Roy likes the idea of continuing to acquire Landsat 7 data for that work, even with the satellite’s ongoing orbital drift.
Despite problems with earlier overpass times with Landsat 7—making that data more difficult to compare to those from earlier in the year or in past years—Roy said the drift could enable novel remote-sensing research and applications. Because Landsat 5 also experienced orbital decay and drift, lessons learned about the effects on Landsat 7 images as the satellite drifts compared to Landsat 8 images means “we hope to figure out a way to correct the orbital drift effects in the Landsat 5 image record,” he said.
Building a spare Landsat and keeping it ready to launch would be a good option, Roy said. “It would be great if environmental satellite missions such as Landsat could be given the same operational status as meteorological satellites so that Landsat continuity could be guaranteed,” he said.
Dr. Ted Scambos, senior research scientist, Earth Science Observation Center, University of Colorado, Boulder, CO
Scambos is a polar scientist focused on the high latitudes, specializing in glaciology, remote sensing of the poles, climate change effects on the cryosphere, and Antarctic history.
He doesn’t view the delay of Landsat 9’s launch as a problem for his user community at this point because Landsat 8 already acquires and processes far more high-quality images than any previous Landsat mission. The overlap of Landsat 8’s orbital path at the poles provides more frequent coverage of those parts of the planet, as does the addition of imagery by the Sentinel-2 mission.
“Short of Landsat 9 being delayed a few years, I don’t think it is a very big issue for polar science,” Scambos said, although coverage of lower-latitude glaciers would be impacted. “Our ability to see very short-term events, like the calving of an iceberg or breakup of sea ice ... that timing is a bit hampered by not having Landsat 9 together with Landsat 8 at the same time,” he said. “But we’ll get there eventually.”
Landsat 7 will still provide an opportunity to calibrate across a long series of Landsat satellites, all the way back to 1984, Scambos said. And even if the timing of Landsat 7 acquisitions falls below 9:15 MLT at the equator, that isn’t necessarily bad for glacier science, he said, since his community is already used to dealing with unusual sun directions and low sun elevation at the high altitudes. Landsat 7 could even offer a bonus.
“The early morning crossing time can help with some tropical glaciers,” Scambos said, “allowing us to get a glimpse of them before the clouds and mist move in.”
Dr. Martha Anderson, Research Physical Scientist with the USDA-Agricultural Research Service
The lack of Landsat 9 imagery in the summer of 2021 means Anderson and her colleagues will explore the value of earlier (Landsat 7) acquisition times to maintain current temporal sampling.
They have experimented with mapping evapotranspiration (ET) using thermal data acquired throughout the day from NASA’s ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS) mission. “Preliminary results suggest performance remains reasonable at 9 a.m. over many landscapes,” Anderson said. “While perhaps not optimal for this application, we will learn how to compensate for the earlier overpass.”
They have been using Landsat’s Thermal Infrared Sensor (TIRS) and Operational Land Imager (OLI) to derive near-real time ET estimates to inform irrigation decisions in California vineyards. The frequency of updates is critical for planning weekly irrigation applications but has been a limiting factor because decisions are sometimes made on weeks-old Landsat imagery, Anderson said.
“We had been looking forward to the full coverage that Landsat 8 and 9 would provide, and the lower latency associated with Landsat 9,” without the scan-line gaps that impact Landsat 7 images, she said. “Our stakeholders,” she added, “must now wait another growing season to see a more optimal system.”
For now, they have found Sentinel-2 data are a useful source of albedo and leaf area information for ET retrieval, though the lack of an on-board thermal sensor with Sentinel-2 means they need to pair that data with another platform.
“I have every confidence that we will overcome and have a successful launch,” Anderson said of Landsat 9. “I look forward to that great day.”
Dr. Curtis Woodcock, Professor, Department of Earth and Environment, Boston University
When it comes to the delay of Landsat 9’s launch, this is what Woodcocks fears the most: a Landsat 7 satellite limping along, its orbit decaying, its decommissioning looming, and the possibility that Landsat wouldn’t have two satellites in space acquiring images of the Earth’s landscapes.
“The question is, ‘Can we keep Landsat 7 limping along so we at least have something close to two satellites up until we get to Landsat 9?’” Woodcock said.
The decision to extend Landsat 7’s life is welcome and one for which the Landsat Science Team advocated, he said. In the past, when Landsat 5 was under commercial management, the time it passed over the equator on each orbit drifted to 9:15 a.m. as well, or toward the end of acquiring useful data. Yet those observations are still being used effectively today, Woodcock said.
“Even down to 9 o’clock, they’re still going to produce observations that will be better than none at all,” he said. “People are clever. People will figure out ways to take advantage. They’ll still be useful.”
What’s important to remember is that the addition of Landsat 8 brought a measurable increase in the number and quality of images, Woodcock said. Landsat 9 will do the same. But the longer it takes for Landsat 9 to launch, “the worse, right?” he said. “You don’t want to wait to get to that any longer than you have to.”
Still, he said his research will continue despite the delay. “We’ll wish we had the Landsat 9 observations,” Woodcock said. But with the availability of Landsat 7 and Sentinel-2, “it won’t stop us from trying to do what we’re trying to do.”
Dr. Matt Hansen, Professor, Department of Geographical Sciences, University of Maryland
A delay in the launch of Landsat 9 is not a deal breaker when it comes to the work he does in crop monitoring and deforestation, Hansen said.
“We’ll miss the (Northern Hemisphere) growing season for corn and soybeans next year with Landsat 9,” he said. “Our current capability will continue with (Landsats) 7 and 8, and 7 is less critical to us than 8. We’ll just have to wait. I accept that. It is what it is.”
His team will do some experiments with its preprocessing calibration of Landsat 7 as it nears the 9:15 a.m. threshold “to see if it will behave predictably,” he said. If it does, they will use it.
His team operates a tropical deforestation alert that uses Landsat 7 and Landsat 8, and produces updates every day if it can get clear pixels. But Hansen also has added Sentinel-2 for that, “so I don’t think we’re going to be hurting there.”
They also do annual soybean assessments using Landsat, but Hansen doesn’t see the delay in Landsat 9 impacting that work, either, because they use the maps generated from Landsat data to guide fieldwork. “Our final area estimates are leveraging the maps, but the maps don’t have to be perfect,” he said. “The better the map, the better our estimate. But we still get good estimates, even with Landsat 8 alone.”
Like many of his colleagues, he wishes Landsat 9 was going up sooner, but he understands the reasons for the delay. “Our whole society has just slowed down and is in a lower gear” because of the pandemic, Hansen said. “I think everyone understands that.”
Dr. Sean Healey, research ecologist with the U.S. Forest Service’s Rocky Mountain Research Station
The U.S. Forest Service uses Landsat in many monitoring applications, most of which will be negatively impacted by unavailability of Landsat 9 during the 2021 growing season.
The Forest Service’s Landscape Change Monitoring System (LCMS) provides up-to-date information to forest managers about loss and gain of tree cover after disturbances such as fires. The agency was counting on the availability of Landsat 9 in 2021 to offset the loss of Landsat 7, which has become unusable in the LCMS processing approach because of sun angle/overpass time problems, Healey said.
Similarly, Landsat is used in monitoring critical habitat for endangered and protected species. For example, the Forest Service is developing a system for monitoring populations and habitat of the Mexican spotted owl as part of a court order to resolve issues surrounding timber harvest and maintaining suitable habitat.
Landsat 7 has been useful in mapping historical changes to owl-friendly vegetation, but no longer serves that purpose, Healey said. So, now they are exploring the use of harmonized Landsat and Sentinel data to supplement both of those projects.
While he’s hoping the harmonized data will minimize the impact of Landsat 9’s delay launch, “adapting a monitoring system to a new data stream requires significant effort, and we must recognize that not all Forest Service applications will be able to prevent monitoring gaps in 2021, especially in cloudy areas where we need every available image to create a good composite view,” Healey said.
That said, Landsat 7 will still be important in things like active fire monitoring, even with its orbital drift, he said. The thermal signal of wildfire is strong regardless of the time of day. So, keeping Landsat 7 in place as a supplement to NASA’s MODerate Resolution Imaging Spectroradiometer (MODIS) will fill a key need in that important operational work, Healey said.