Brian Markham was a fixture of Landsat calibration science for four decades. And in his retirement, he still advises the NASA Landsat calibration team.
It was in the Landsat Thematic Mapper (TM) era of the 1980s that Brian first started working on Landsat calibration under the tutelage of renowned NASA calibration scientist John Barker.
Brian and John worked together on an endeavor to quantify how the then-new Landsat TM sensors performed and how well their data related to phenomena on the ground. That project, known as the Landsat Image Data Quality Analysis, or LIDQA, is credited with beginning the standard of excellent calibration that is now a hallmark of the Landsat Program. Results of this work described the radiometric (think energy measurements) and geometric (think pixel geolocation) performance of the Landsat 4 and 5 sensors.
In 1993, as the development of Landsat 7 got underway, Brian was tapped to take on the role of Landsat Calibration Scientist.
An Integrated Approach to Landsat Calibration
It was in the Landsat 7 era that calibration scientists from NASA, the U.S. Geological Survey (USGS), and universities began to meet regularly as the Landsat Calibration Working Group. Initially, Brian and this team worked to reconstruct the calibration record of Landsat 5 which had languished during the Landsat commercialization period. With Landsat 7 returned to government control, and launched in 1999, the practice of continuous calibration trending was established.
As a result of this work, calibration and performance monitoring became recognized as an essential element of the Landsat science mission, both prior to and following launch. NASA took the lead role for calibration efforts pre-launch and during post-launch commissioning and USGS took the lead for calibration during satellite operations. Both organizations participate and collaborate on calibration throughout the entire mission—a pattern that remains today.
Using lessons learned from the efforts to calibrate Landsat 5 and 7, Brian, the Landsat science office, the calibration team, and outside experts developed a set of requirements, that ensured Landsat 8 could be rigorously calibrated. That standard carried over to Landsat 9, resulting in the best calibrated Landsat sensors to-date. (Additional documents that catalogued specific calibration testing requirements were titled “Special Calibration Test Requirements,” or SCTR—pronounced “scatter”—a source of levity for a team measuring light.)
Brian was also part of an international partnership that worked to cross-calibrate Landsat 8 and the European Space Agency’s Sentinel-2 satellites. This effort started prior to the launch of Sentinel-2A and continues to this day; the Sentinel-2 and Landsat calibration teams monitor each other’s calibration and routinely attend each other’s calibration meetings.
When you pose a calibration question to Brian, you can perceive him thinking through the instrument design, envisioning the path of light traveling through filters and bouncing off mirrors to land on detectors, of calibration lamps and uniformly reflective diffusor panels. Brian’s calibration acumen is appreciable, and as the head of Landsat calibration at NASA, his gentle kindness and quirky humor endeared him to his team and strengthened his effective leadership.
We spoke with Brian recently to get his take on why calibration is an essential part of the Landsat program and data continuity. Here are highlights from our conversation:
For a satellite like Landsat, what does calibration mean?
Calibration technically means adjusting parameters to make the data properly reflect physical units. We do this with our updates to the calibration parameter files.
The goal of calibration is to make sure that the measurements made by Landsat instruments accurately show how much radiance comes into the sensor and where on the ground it comes from.
Much of what our calibration team does is more technically “characterization.” Not all characterization leads to calibration updates, but characterization helps us better understand the data so we can be sure it yields meaningful information for data users.
Why do we carefully calibrate Landsat data?
I think in part it is our nature as scientists to make our data as good as we can, but it is also necessary for scientific usage of the data where one wants to be able to relate the data to physical units (e.g., radiance, reflectance or temperature).
You need good calibration to extract biophysical parameters from the data and to physically locate the data on the ground. There is also the need to have data consistent radiometrically and geometrically over time so you can assess how our planet is changing—that you can be confident that the change the data tells you is happening on the ground is actual biophysical change and not just data artifacts (i.e. bad data).
Landsat calibration is considered the gold standard for medium resolution satellite data. Can you tell us why?
It is the time and effort we put into it. Many commercial satellite systems emphasize high resolution, relative radiometry (destriping), and geometric accuracy, but it takes a lot of time and money to perform detailed characterizations of these instruments, and it is not cost effective for commercial systems that need to make a profit. Having a government-provided standard for them to tie their data to can be cost effective for them.
Why is calibration an important part of data continuity?
It all fundamentally comes down to this: if we’re using a system to quantify changes on the Earth, we must make sure the system we’re using is not changing.
Consistent data—provided by solid calibration—is needed to provide data continuity.
Are there significant calibration milestones that standout over your career?
I think my role on Landsat 8 was the most rewarding because it involved a new sensor.
Landsat 7’s ETM+ sensor was a derivative of a previous sensor, so I had less influence over it. Its onboard calibrators did not all behave well. They all degraded over time, and we did not have a lot of redundancy. So, when we got to Landsat 8’s OLI senor, we built in enough redundancy that we could do excellent calibration almost completely based on its onboard calibrators.
I was involved on Landsat 8 from requirements development all the way through to on-orbit performance evaluation and trending.
Seeing how well it performed on orbit was gratifying.
