Source: USGS Earth Resources and Observation Science (EROS) Center
The rapid emergence of smallsats begs the question: What role does a legacy satellite like Landsat play in the development and use of these relatively inexpensive orbiters?
Recent research suggests Landsat offers the same value it’s offered to generations of satellite developers: A gold standard for calibration that adds value and reliability to other satellite systems.
Rasmus Houborg is a scientist at South Dakota State University’s Geospatial Sciences Center of Excellence (GSCE) who has spent more than two years working with data from Planet Labs, a company with a constellation of approximately 175 active cubesats, which they refer to as “Doves.”
What Houborg has learned is that bootstrapping cubesat data to highly-calibrated instruments is necessary to turn raw cubesat readings into reliable data that can be used to produce timely, location-specific insights useable for end users.
“It needs a higher quality satellite,” Houborg said. “Landsat is a highly-calibrated system. You simply cannot generate high quality science products with cubesat systems alone.”
The GSCE is a partner of the USGS Earth Resources Observation and Science (EROS) Center, which houses the Landsat archive and serves as the satellite’s ground station and an international hub for calibration expertise.
Houborg and Matthew McCabe of King Abdullah University of Science and Technology in Saudi Arabia recently published an article in the journal Remote Sensing that outlined a method for enhancing the utility of Planet cubesat data.
The study looked at alfalfa plots in Saudi Arabia over a six-month period. The goal was to produce daily Normalized Difference Vegetation Index (NDVI) and Leaf Area Index (LAI) readings at a 3-meter resolution.
To do that, researchers developed a Cubesat Enabled Spatio-Temporal Enhancement Method (CESTEM), which optimizes the utility and quality of cubesat data by fusing them with data from Landsat-8 and NASA’s Moderate Resolution Imaging Spectroradiometer (MODIS).
“While daily imaging was available during the majority of the six-month study period, original (i.e., not CESTEM corrected) NDVI time series over a multicut alfalfa field indicated significant day-to-day variations not directly attributable to vegetation dynamics or phenology,” the study said.
CESTEM produced daily leaf-area index (LAI) readings that were consistent with Landsat’s radiometry, but at cubesat’s temporal and spatial resolution. The improvements cleared the noise from the Planet data, offering confidence in the day-by-day green-ups observed during the study period.
“You can really see the noise removal and the key importance of doing that for enhanced interoperability and utility of sensor data,” Houborg said.
Dennis Helder heads the EROS Calibration and Validation Center of Excellence (ECCOE), which works to maintain and improve Landsat’s consistency and reliability as a benchmark for other Earth-observing systems, to improve cross-calibration opportunities between remote sensing systems, and to share new measurements and expertise with the scientific community.
Helder said Houborg is doing some of the most interesting work on harmonizing cubesats and Landsat, work that aligns with the ECCOE mission of improving data reliability and interoperability.
Houborg will be a key member of the ECCOE workshop, which takes place one day after the Joint Agency Commercial Imagery Evaluation (JACIE) workshop, Sept. 17-19 in College Park, MD.
Houborg will also be a speaker at the JACIE workshop.
“The focus of the workshop is ‘how can we take the large heritage government satellites and cross-calibrate them to the smallsats that are flying, which are by design smaller, less expensive and without onboard calibration systems?’” Helder said. “Is there a way we can take what we know and apply it to what they’re doing, and is there a way they can take what they’re doing and couple it with our instruments, so the user will have a better experience with products from both systems?”
The fusion of Landsat and Planet data could offer the scientific community reliable data for work on a wide variety of issues, Helder said, from urbanization to deforestation or natural disasters. Planet Labs was recently added to the International Charter ‘Space and Natural Disasters,’ and will contribute imagery to responding agencies.
Daily readings are especially valuable for agriculture. A sensor that can be trusted to catch trouble spots within fields could help farmers decide where to put their resources. It could also offer more immediate verification of crop damage from hail for crop insurance purposes.
“Most people in agriculture would like to look at least once a week,” Helder said.
The possibility of interoperability of Planet and Landsat data is exciting for Greg Stensaas, as well. Stensaas is the manager of the Requirements, Capabilities & Analysis for Earth Observation (RCA-EO) project, which includes characterizing and understanding remote sensing systems and being the USGS JACIE co-chair.
The use of highly-calibrated systems to improve higher-resolution data for scientific research is picking up steam, Stensaas said.
“Between Landsat and Sentinel, there are going to be many papers written in the near term about using those datasets as benchmarks or data points that can cross-calibrate or cross-characterize other data sets to support science,” Stensaas said.
The explosion in cubesat data is sure to factor into upcoming discussions by the Architecture Study Team (AST) for Landsat 10. Landsat 9’s launch next year will preserve the continuity of the mission as those discussions begin.
“I think as we move into the assessment of potential opportunities for the next Landsat, one of the main pillars that will remain is a calibration based system for the rest of the world to compare to,” Stensaas said.
That system does more than simply improve data for dreamers and doers in the private sector. It can actually save money for developers. Astro Digital, a satellite constellation operator out of Santa Clara, CA, announced in early August that cross-calibration with Landsat and Sentinel will allow it to achieve its goals with half the satellites it initially expected to launch. Trimming the number of satellites from 10 to five will save the company several million dollars, CEO Chris Biddy said.
Stories like those are sure to come up during next month’s calibration gatherings in Maryland.
+ Learn more on the upcoming ECCOE workshop and the smallsat discussion
+ Learn more about the upcoming JACIE conference
The Pale Blue Dot Visualization Challenge—aimed at making Earth observation data accessible to everyone—has officially kicked off.