The Landsat suite of satellite sensors has been the most successful remote sensing effort dedicated to Earth observations.1 Born of civilian rather than military needs, the Landsat suite has provided 40 years of standardized, moderate-spatial-resolution, multispectral images of the world.2 No other data sets allow assessment of the changing human condition so effectively. No other data sets can match Landsat’s comprehensive record of Earth and its resources.
Consensus exists among government, commercial, and research users about the need for a sustained land imaging program. Sustainability can be achieved by developing an operational observing system whereby satellites will be designed and launched to provide a continuous stream of land images and data, similar to the policy articulated in the 2010 National Space Policy. 3 Compared to other spaceborne moderate-resolution sensors, long-term continuity has distinguished the Landsat sensor suite.
The committee’s definition of an “operational” program preserves continuity as the main goal: design the satellite system and launch schedule to provide a continuous stream of land images and data, implicitly requiring strategies to contend with future instrument or launch failures. The committee’s interpretation is that this goal does not require a “hot spare” that is already in orbit, but rather a commitment to adopt risk mitigation strategies that could range from instruments ready to launch to securing agreements with international partners for data access.
Surveys4 generally show that users want, aside from continuity, frequent, moderate-resolution imagery, and that concerns about orbits, calibration, and shape of spectral bands are secondary. Long-term stability and the ability to integrate with other sensors enable detection and analysis of rates of change. Fine-resolution land data—less than 5-m spatial resolution—appear to have a commercial market. At the coarse end of the scale, imagery at 250- to 1,100-m spatial resolution with daily worldwide coverage is used for regional-to global-scale science and operational weather prediction, oceanography, and snow-cover mapping. Between these scales, history and
1 J.R. Irons, J.L. Dwyer, and J.A. Barsi, The next Landsat satellite: The Landsat Data Continuity Mission, Remote Sensing of Environment 122:11-21, 2012.
2 T.R. Loveland and J.W. Dwyer, Landsat: Building a strong future, Remote Sensing of Environment 122:22-29, 2012.
3 National Space Policy of the United States of America, June 28, 2010, available at http://www.whitehouse.gov/sites/default/files/national_space_policy_6-28-10.pdf.
4 For example, K. Green, J. Plasker, G. Nelson, and D. Lauer, Report to the White House Office of Science and Technology Policy Future Land Imaging Working Group on the American Society for Photogrammetry and Remote Sensing survey on the future of land imaging, Photogrammetric Engineering and Remote Sensing 73:5-9, 2007.
TABLE 2.1 Landsat Satellite Characteristics
|System||Sensors||Radiometric Resolution||Temporal Resolution||Orbit Altitude||Launch Date||Decommission or Expiration Date|
|Landsat 1 (ERTS-A)||July 23, 1972||January 6, 1978|
|Landsat 2 (ERTS-B)||RBV and MSS||6 bits||18 days||900 km||January 22, 1975||February 5, 1982|
|Landsat 3||March 5, 1978||March 31, 1983|
|Landsat 4||July 16, 1982||June 15, 2001|
|MSS and TM||8 bits||16 days||705 km||TM: November 2011|
|Landsat 5||March 1, 1984||MSS: January 6, 2013|
|Landsat 6||ETM||8 bits||16 days||705 km||Launch failed October 5, 1993|
|Landsat 7||ETM+||8 bits||16 days||705 km||April 15, 1999|
|Landsat 8 (LDCM)||OLI and TIRS||12 bits||16 days||705 km||February 11, 2013|
NOTE: ETM, Enhanced Thematic Mapper; MSS, Multispectral Scanning System; OLI, Operational Land Imager; RBV, Return Beam Vidicon; TIRS, Thermal Infrared Sensor; TM, Thematic Mapper.
SOURCE: NASA Goddard Space Flight Center, see http://landsat.gsfc.nasa.gov.
user surveys have shown that Landsat data at moderate resolutions (15 to 100 m, at 8-to 16-day frequency) have significant intrinsic value for a broad range of federal and nonfederal scientific and operational uses but little promise for commercialization.
Spurred by photographs of Earth from the Apollo missions in the 1960s, the Department of the Interior and the U.S. Department of Agriculture (USDA) envisioned a program to provide unclassified remotely sensed data in support of resource studies and planning.5 NASA launched the first Earth Resources Technology Satellite (ERTS) (subsequently renamed Landsat 1) in July 1972, and since then a total of seven successful missions have collected more than 2 million images of Earth spanning a 40-year period. While the technology used to capture Landsat data has evolved over its 40-year life span, each new Landsat system has been designed so that many of the imagery products are backward compatible. More important than each system’s innovation and science is the Landsat suite’s combined continuity of observations, which bring overwhelming value to each new Landsat system.
Tables 2.1 and 2.2 summarize the technical characteristics of the seven successful Landsat systems in the common categories of spectral, radiometric, spatial, and temporal resolutions. All systems have had the same swath width, 185 km. Over the span of the Landsat systems, spectral resolution has increased from 4 to 11 bands, with some changes in the shape of the spectral response functions, and the spatial resolution of those bands has narrowed from 80 to 15, 30, and 100 m.6 Radiometric resolution has increased from 6 bits on Landsats 1 through 3, to 8 bits on Landsats 4 through 7, and to 12 bits on Landsat 8. Landsats 1 through 3 had an 18-day repeat cycle, which was shortened to 16 days on subsequent missions. Temporal resolution has sporadically increased from a 16-day revisit to an 8-day revisit only when and where two Landsat systems were operating simultaneously, which has, unfortunately, been rare over the past 20 years because of Landsat 5’s inability to store data onboard and Landsat 7’s scan line corrector failure in 2003.
While unique, Landsat is only one of many multispectral Earth observing sensing systems. Commercial providers such as DigitalGlobe, Inc.,7 offer finer-spatial-resolution multispectral imagery for sale, but it is costly
5 D.T. Lauer, S.A. Morain, and V.V. Salomanson, The Landsat program: Its origins, evolution, and impacts, Photogrammetric Engineering and Remote Sensing 63:821-838, 1997.
6 Spatial resolution refers to the distance between distinguishable features in an image, whereas the pixel size in images delivered is often –resampled. Note that the spatial resolution of the thermal band decreased from 120 to 60 m with Landsat 7, but reverted to 100 m on Landsat 8.
7 DigitalGlobe acquired GeoEye in January 2013.
TABLE 2.2 Landsat Sensor Characteristics
|Band Identifier||Spectral Range (μm)||Spatial Resolution (m)||Notes|
|RBV 1||0.475-0.575||80||Landsats 1 and 2|
|RBV pan||0.505-0.750||38||Landsat 3|
|MSS 4||0.5-0.6||68×83 resampled to 57×79|
|MSS 8||10.4-12.6||Landsat 3 only|
|ETM 1-7||same as TM|
|ETM+ 1-5||same as ETM||ETM+ also has enhanced calibration|
|ETM+ 7-8||same as ETM|
|OLI 1||0.433-0.453||30||With 12-bit quantization, dynamic range of the OLI does not saturate over clouds or snow|
and license restricted, and the systems do not have the large synoptic geographic footprint of Landsat data. In the United States, the USDA National Agriculture Imagery Program subcontracts for suborbital aerial photography every 2-3 years and provides 1-m resolution imagery to the public domain at no cost. The images in Microsoft’s Bing Map and Google Earth are acquired from Landsat, aircraft, and commercial satellites. Systems such as NASA’s Moderate Resolution Imaging Spectroradiometer have daily temporal resolution and many more bands (36) but at a much coarser 250-to 1,000-m spatial resolution. Other governments and organizations outside the United States
(e.g., France, China, India, and Korea) collect moderate-resolution multispectral imagery. However, none of these other systems have the unique combined characteristics of Landsat because their data are often difficult to access, and some providers charge for or restrict their coverage or the use of their products.
A clear separation has developed between government and commercial sources of imagery, with the commercial sector providing the fine-resolution imagery (<5 m). While commercial products are mainly finer resolution than those provided by the Landsat system and are often not as comprehensive in coverage, they augment the operational capabilities available today and can enable focused studies that are impossible to undertake with Landsat-quality data. Formal agreements between the U.S. government and commercial remote sensing data providers would encourage the development and improvement of capabilities in the commercial remote sensing sector and likely increase the pool of experts in remote sensing.
Foreign data sources can supplement national imagery data sources and can function as data gap fillers if appropriate agreements are in place. Foreign imaging assets can be used for mitigation of risk—for example, if a U.S. satellite fails—but generally they are considered as complementary data sets. The history of obtaining remote sensing data from foreign agencies shows a few outstanding successes, like the European Space Agency’s Envisat (until its failure in April 2012). This committee recognizes both potential benefits and risks of relying on foreign land image data sets, with the risks mainly relating to data availability and the matching of requirements to sensor characteristics.
The committee did not attempt any systematic analysis of Landsat users and their requirements because multiple studies and reviews have already been carried out. For example, in 2007, the American Society for –Photogrammetry and Remote Sensing (ASPRS) conducted a survey of 1,295 Landsat users8 and reported on their characteristics and their data requirements. In 2011, the U.S. Geological Survey (USGS) published a study on users, uses, and the value of Landsat and other moderate-resolution data,9 and in 2012 the Landsat Advisory Group of the National –Geospatial Advisory Committee10 reviewed the cost savings accruing to 10 of the largest government operational uses of Landsat imagery. All of these studies comment on the broad range of uses of Landsat data, from agricultural monitoring, to water management, to forest pest detection, to national defense (Table 1.1). The studies also note that users of Landsat data are overwhelmingly government agencies, academic institutions, and nongovernmental organizations, with commercial entities constituting only a small fraction of users, about 18 percent.11 Additionally, almost half of the users employ Landsat data to support operational decision making, with the remainder performing scientific research.12
The ASPRS study found that the characteristics of Landsat imagery most valued by users in order of priority are its low cost, SWIR bands, existence of the archive, the thermal band, and its moderate spatial resolution. During the public meetings held to obtain information for this report, the most common user request for technical improvements in Landsat was for more frequent temporal resolution, primarily to support agricultural monitoring and to increase the probability of coverage in the face of intermittent cloud cover.
8 K. Green, J. Plasker, G. Nelson, and D. Lauer, Report to the White House Office of Science and Technology Policy Future Land Imaging Working Group on the American Society for Photogrammetry and Remote Sensing Survey on the Future of Land Imaging, Photogrammetric Engineering and Remote Sensing 73:5-10, 2007, available at http://www.asprs.org/a/publications/pers/2007journal/january/.
9 H.M. Miller, N.R. Sexton, L. Koontz, J. Loomis, S.R. Koontz, and C. Hermans, The Users, Uses, and Value of Landsat and Other Moderate- Resolution Satellite Imagery in the United States—Executive Report, U.S. Geological Survey Open-File Report 2011-1031, 2011, available at http://pubs.usgs.gov/of/2011/1031/pdf/OF11-1031.pdf.
11 Miller et al., 2011.
12 Green et al., 2007.
FIGURE 2.1 Increase in scenes delivered since USGS made Landsat imagery available on the Web in October 2008 at no cost. SOURCE: U.S. Geological Survey.
Landsat data were originally available at low cost.13 During the era of Landsat commercialization (see Chapter 1), Landsat imagery cost up to $4,400/scene. With the launch of Landsat 7, USGS lowered the cost to $600/scene, and in October 2008 made the entire Landsat imagery archive available on the Internet at no cost. Use of Landsat imagery increased rapidly (Figure 2.1). It has become ubiquitous as the moderate-resolution data set for both Google and Bing, is the foundation of Esri’s ChangeMatters14 website, and is employed in weather reporting by many television stations. Other examples of applications made possible by free and easy access to Landsat imagery include monitoring consumptive outdoor water usage, updating global land use or land cover maps, forest health monitoring, national agricultural commodities mapping, flood mitigation mapping, forest
13 Approximately $15 for photographic prints and $200 per data set. See http://remotesensing.usgs.gov/landsat_fees.php and National Oceanic and Atmospheric Administration, 1983, Landsat data users notes: [Sioux Falls, S.D.], National Oceanic and Atmospheric Administration [variously paged].
fragmentation detection, forest change detection, world agriculture supply and demand estimates, wildfire management, and coastal change analysis.
Recently, the Landsat Advisory Group of the National Geospatial Advisory Committee15 was asked by the Department of the Interior to investigate the feasibility of once again charging for Landsat data. The Group strongly advised that Landsat data should continue to be distributed at no cost. It found that charging for Landsat data would
• Severely restrict data use;
• Violate existing Office of Management and Budget guidelines, federal law, Office of Science and Technology Policy, and U.S. National Space Policy,
• Require statutory changes;
• Cost more than the amount of revenue generated by the charges;
• Create a circular payment basis for public agencies;
• Stifle the innovation and business activity that create jobs;
• Inhibit data analysis in scientific and technical analyses;
• Negatively impact international relations with respect to national, homeland, and food security; and
• Negatively impact foreign policy and U.S. standing as the leader in space technology.
To meet the requirements for continuity in the face of technological development and ongoing understanding of the land surface, the Sustained and Enhanced Land Imaging Program (SELIP) relies on well-defined users with clear scientific or operational requirements so that program goals are clearly articulated. Because users of land imaging data are widely spread across the government and private sector, current and future users groups will be diverse and broadly inclusive. Agreement on a set of core measurements simplifies the development of standardized sensors, data archiving, processing, and dissemination.
Although it will always be difficult to satisfy every user need, the committee found remarkable consistency in user requirements. The core scientific and operational requirement for the SELIP is the capture and distribution of global, moderate-resolution (30-100 m), multispectral data products, enhanced by a panchromatic band at finer resolution. The suite of applications for analyses of the data requires the full range of spectral capabilities—visible, near infrared, shortwave infrared, and thermal infrared—but there are no requirements to provide all measurements on the same platform, nor to continue to fly the same sensor, nor to restrict future systems to the current viewing angles and swath width. It is no coincidence that these requirements echo the present capability of the Landsat sensor suite, because assuring continuity of the ongoing data stream is the key aim for the future program.
The following requirements would satisfy a broad range of key federal and nonfederal users, both scientific and operational:
• Spatial resolution
— 30 m except in the thermal band, which would have coarser spatial resolution.
— Finer resolution (10-15 m), perhaps in a panchromatic band, was desired by some.
• Spectral requirements
— Visible and near-infrared region (VNIR, 0.4-1.1 µm).
— Shortwave infrared region (SWIR, 1.2-2.8 µm).
— Thermal infrared region (8-12 µm, with some interest in 3.5-4.0 µm).
— Calibration sufficient to allow backwards-compatible comparisons of future image products to earlier image collections.
— A larger dynamic range in the VNIR region to prevent saturation over snow and clouds; this requirement has been met in the Landsat 8 Operational Land Imager, with its 12-bit instead of 8-bit quantization.
15 National Geospatial Advisory Committee-Landsat Advisory Group Statement on Landsat Data Use and Charges, September 18, 2012, available at http://www.fgdc.gov/ngac/meetings/september-2012/ngac-landsat-cost-recovery-paper-FINAL.pdf.
• Coverage and repeat cycle
— Ability to acquire and make available imagery anywhere on Earth, except perhaps for areas very near the poles, at approximately weekly frequency. (This frequency is desired not necessarily to acquire weekly data but rather to acquire cloud-free images.) The 705-km Landsat orbit, at 98 degrees inclination, provides 16-day frequency.
— Increased temporal frequency could be achieved with a slightly larger swath and consequently slightly larger off-nadir view angles at the edge (the users queried did not object to this).
• Data management and distribution
— A free data policy, as is currently in place, provides huge benefits to the nation as well as the international user community by supplying imagery to operational programs critical to U.S. needs as well as spurring innovation in the private sector.
— The USGS data distribution system is successful and effective but could continue to make technological advances and to streamline methods for managing Landsat imagery and derived products.
This set of requirements could be met by implementing the system as a series of satellite platforms, possibly with smaller satellites, whereby all capabilities may not reside on a single spacecraft. Many applications do not require precise simultaneity of all spectral bands, so that satellites flying in formation with nodes adjusted so that multiple spectral bands are acquired within hours could suffice.
The top priorities for the Sustained and Enhanced Land Imaging Program (SELIP) should be to ensure that the core program provides for continuity of Landsat products and coverage on a secure and sustainable path.
The SELIP should take advantage of technological innovation in sensors, spacecraft, and data management and analysis to improve system performance, allow for new analyses that better exploit the data and meet future needs. Because future measurements will derive from both current and new technologies, new implementations of existing data products derived from a multispectral sensor should be cross-calibrateable with Landsat legacy products and be essentially interchangeable for scientific and operational purposes.
To better meet these primary goals, the committee recommends that the program should
• Systematically monitor users and uses of Landsat data so that the program can evolve with changing user requirements and
• Consider alternative implementations that continue to enable the collection of global, moderate-resolution data with the full range of spectral capabilities.