Commentary | The Rime of the Ancient (Land) Mariner
The March 4 SpaceNews had a good editorial on the Landsat program [“Enjoy Landsat 8 — While It Lasts,” page 18]. However, for those uninitiated in Landsat affairs, it was rather incomplete.
Welcome to a sad tale of many twists and turns told by two ancient Landsat mariners. We are still not certain who shot the Landsat albatross (several names come to mind) but we have had bad luck ever since. Continuation of Landsat past version 8 is only the most recent turn.
Planning for a U.S. civilian land imaging program began in the early 1960s with consideration being given to many alternative configurations. By the time the Earth Resources Technology Satellite (ERTS) was launched in 1972, it was effectively a late 1960s NASA smallsat. The bus was a recycled Nimbus satellite with two sensors onboard: a three-band return beam vidicon and an innovative multispectral scanner.
The originally proposed land observation concept, from prior science planning meetings, was to have three or four satellites in orbit at any given time to provide sufficient temporal repeatability to monitor seasonality of vegetation dynamics. However, the ERTS program was never very popular within NASA nor well understood by either the U.S. Geological Survey or the U.S. Department of Agriculture, where much of the early advocacy for a satellite land observatory originated. The basic goals of monitoring land dynamics at spatial resolution that captures human impacts and a temporal resolution that monitors vegetation seasonal progress were never fully achieved. ERTS, renamed Landsat in the mid-1970s, hobbled through three iterations using the original return beam vidicon- multispectral scanner design, demonstrating the capacity of such a system to collect global systematic, annual land observations over a decade. The multispectral scanner instrument on Landsat 3 had serious technical problems, requiring continuation of Landsat 2 operations well past its design life.
The second-generation Landsat observatory was first flown in 1982. Two instruments were flown on this generation: a new advanced multispectral scanner, the Thematic Mapper, and another near-copy of the Landsat 1-3 multispectral scanner sensor. The latter instrument was flown because many users claimed not to be able to use the higher dimensions (30-meter spatial resolution, seven spectral bands and 8-bit measurements) of the Thematic Mapper data stream. Mind you, this was in the days of seven-track tapes and computer mainframes. The replicate multispectral scanner data stream was flown on both Landsat 4 and 5. The design for the Thematic Mapper instrument had been proposed in the late 1960s but it took 20 years before the necessary technologies were sufficiently mature to build and operate this advanced sensor system.
The Thematic Mapper, like the ERTS-1 multispectral scanner sensor but more so, was quite a Rube Goldberg device, with complex optics and mechanical scan mirrors. In fact, the scan line corrector mirror on Landsat 7 failed four years after mission launch. Until Landsat 8 was launched a month ago, the United States continued to fly these complex electro-optical scanning sensors, including in the Earth Observing System Moderate Resolution Imaging Spectroradiometer instrument. The basic reason that the electro-optical scanning sensors survived within the Landsat mission for more than two decades was that the U.S. government, reacting to the curse, decided in 1982 to commercialize Landsat. Commercialization was also not a success for various reasons, resulting in the Landsat program, past Landsat 6, being returned to government control in 1992. Unfortunately the commercial enterprise was further decimated when the commercially designed Landsat 6, also a scanning device, failed to achieve orbit in 1993.
Remarkably, by the time Landsat 7 was flown in 1999, the United States had already accumulated nearly 27 years of repetitive global land observations with the Landsat observatories. Further, when the commercial Landsat operator abandoned Landsat 5 in 2001, the United States had two satellites in orbit at the same time providing eight-day repeat cycle observations. Today, we have a continuous archive of 40-plus years of Landsat observations, providing fundamental, critical information of the Earth’s land areas at a period of major land and climate changes. This archival record provides the foundations upon which our understanding of current and future treads in biospheric productivity is based.
Passage of the Land Remote Sensing Policy Act of 1992 was critical for the Landsat mission. We almost thought the curse had been broken. This 1992 law brought Landsat back under U.S. government management and provided a pathway forward to make Landsat operational. But no, this was a false hope, for the curse was not broken. The partnership of NASA, the National Oceanic and Atmospheric Administration and the Department of Defense established to manage Landsat soon fell apart. In the post-Berlin Wall era, faced with declining budgets, the Pentagon saw that its goals of higher-resolution “point and shoot” capability would not be realized.
One bright spot in the 1992 law was that NASA was directed to develop advanced technologies for future Landsat missions under its New Millennium Program. This directive, in effect, would let us catch up with the French SPOT satellites that had used simpler solid-state linear array sensors and optics dating back to 1986. The first New Millennium Program mission was referred to as Earth Observer (EO)-1 and produced the first high-quality smallsat approach to a Landsat-like mission.
However, as time approached to develop the Landsat 8 mission, the curse struck again. NASA ended up pretty much ignoring the EO-1 Advanced Land Imager solution and oversaw the development of a much larger third-generation, solid-state instrument called the Operational Land Imager. A second large instrument, the Thermal Infrared Sensor, was added late in the development cycle to more fully match Thematic Mapper heritage. The Landsat 8 Operational Land Imager and Thermal Infrared Sensor are great engineering achievements, but as a pathway forward to smaller, more affordable operational land observatories, Landsat 8 is missing — nay, it ignored — the downsizing lessons learned from EO-1.
Today we are facing perhaps the ultimate Landsat mission crisis. We now have a Landsat 8 observatory at a cost approaching $900 million, and the projected costs for a Landsat 9 clone copy are well in excess of a billion dollars. At these costs, especially in today’s economic environment, the Landsat program will most likely continue to miss an important goal that early visionaries of Landsat understood all too well: the need to have at least three or four satellites in orbit at any given time to allow users to study and better understand land surface dynamics. For example, the ability to monitor within growing season variations in crop health and/or conduct more frequent assessments of water use for crop irrigation could have a significant impact on deriving earlier and more accurate estimates of productivity.
Fortunately, there are companies that have demonstrated that they do know how to build dramatically lower-cost small satellite land observatories having stable Landsat-like performance. Perhaps the best known of these vendors is Surrey Satellite Technology Ltd., founded in England but with a growing manufacturing presence in the United States. By using lower-cost small satellites to deploy an observatory constellation, the Landsat mission would be capable of monitoring short-term land processes at human scales while reducing data gap risks associated with the current single-satellite Landsat approach.
The primary problem that the Landsat mission faces today is that smallsat solutions are not “trusted” in the United States for a multitude of reasons, most of which are political in nature and not based on technical issues. Consider that both NASA (with EO-1) and the aerospace industry have demonstrated that they know how to build reliable smallsat Earth observatories. Today we struggle not with technical constraints but with the lack of political will to move toward more-robust, yet lower-cost, 21st century solutions to Earth observations.
If the struggle continues to remain caught in these doldrums, not only will the albatross have died, but the ancient (land) mariner — Landsat — also will expire because of a lack of societal commitment and political will.
Samuel N. Goward is a professor in the Department of Geographical Sciences at the University of Maryland in College Park. Darrel L. Williams, chief scientist for Global Science & Technology Inc. of Greenbelt, Md., retired from NASA after serving as Landsat project scientist for nearly half of his 35-year career at NASA Goddard Space Flight Center.