3

Enhancing a Sustained Land Imaging Program

Landsat has formed the cornerstone of the nation’s land imaging effort, but it has never constituted the totality of that effort. Although the findings and recommendations of the committee incontrovertibly point to the need for a continuation of the critical Landsat time series, it is crucial to recognize that many other spaceborne missions have contributed greatly to U.S. imaging capabilities. For example, the Shuttle Radar Topography Mission has provided agencies and scientific users alike with near-global digital elevation data, and airborne programs continue to support focused operational uses, local scientific research objectives, and technology development. These measurements do not replace the moderate-resolution imaging of the Landsat satellites; they instead complement and add value to the core observations. Many remote sensing applications can only be done by integrating multiple data sources, and researchers routinely interpret images in the context of several types of data.1 The committee sees a great benefit in defining the U.S. land imaging program more broadly, recognizing the substantial contributions from a diverse set of airborne and spaceborne assets. Some other types of remotely sensed data—which include finer spatial resolution, active technologies including both LiDAR and radar, and hyperspectral capability—are already being acquired by the U.S. government, the private sector, and other countries, and some could be considered for a future land imaging satellite (Table 3.1). Not all these capabilities would or could be provided directly by the U.S. government; commercial providers and international partners are essential and likely will be integral parts of the full Sustained and Enhanced Land Imaging Program (SELIP). The government would not necessarily archive all these data—indeed, not all would be available at no cost—but the data management function of SELIP could provide links to these complementary data sets.

FINE-RESOLUTION SPACEBORNE AND AIRBORNE IMAGERY

The Landsat satellites image Earth roughly weekly at moderate resolution (30 to 100 m),2 and the historical record of Landsat data stretches back 40 years at 18-day and then 16-day revisit times (8 days with two satellites working together). This extraordinarily rich data set has led to many important studies that now monitor and explain diverse phenomena occurring on Earth’s surface. Nonetheless, as understanding of these observed processes improves, both the scientific frontier and the utility of operational use have advanced such that the value of the

____________

1 T. Lillesand, R.W. Kiefer, and J. Chipman, Remote Sensing and Image Interpretation, 6th ed., John Wiley and Sons, 2007.

2 A 15-m resolution capability was added to Landsat 7 via the ETM+ instrument but only in the panchromatic band (often referred to as the black-and-white band).



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 27
3 Enhancing a Sustained Land Imaging Program Landsat has formed the cornerstone of the nation’s land imaging effort, but it has never constituted the totality of that effort. Although the findings and recommendations of the committee incontrovertibly point to the need for a continuation of the critical Landsat time series, it is crucial to recognize that many other spaceborne missions have contributed greatly to U.S. imaging capabilities. For example, the Shuttle Radar Topography Mission has provided agencies and scientific users alike with near-global digital elevation data, and airborne programs continue to support focused operational uses, local scientific research objectives, and technology development. These measurements do not replace the moderate-resolution imaging of the Landsat satellites; they instead complement and add value to the core observations. Many remote sensing applications can only be done by integrating multiple data sources, and researchers routinely interpret images in the context of several types of data. 1 The committee sees a great benefit in defining the U.S. land imaging program more broadly, recognizing the substantial contributions from a diverse set of airborne and spaceborne assets. Some other types of remotely sensed data—which include finer spatial resolution, active technologies including both LiDAR and radar, and hyperspectral capability—are already being acquired by the U.S. government, the private sector, and other countries, and some could be considered for a future land imaging satellite (Table 3.1). Not all these capabilities would or could be provided directly by the U.S. government; commercial providers and international partners are essential and likely will be integral parts of the full Sustained and Enhanced Land Imaging Program (SELIP). The government would not necessarily archive all these data—indeed, not all would be available at no cost—but the data management function of SELIP could provide links to these complementary data sets. Fine-Resolution Spaceborne and Airborne Imagery The Landsat satellites image Earth roughly weekly at moderate resolution (30 to 100 m), 2 and the historical record of Landsat data stretches back 40 years at 18-day and then 16-day revisit times (8 days with two satel- lites working together). This extraordinarily rich data set has led to many important studies that now monitor and explain diverse phenomena occurring on Earth’s surface. Nonetheless, as understanding of these observed processes improves, both the scientific frontier and the utility of operational use have advanced such that the value of the 1  T.Lillesand, R.W. Kiefer, and J. Chipman, Remote Sensing and Image Interpretation, 6th ed., John Wiley and Sons, 2007. 2  A 15-m resolution capability was added to Landsat 7 via the ETM+ instrument but only in the panchromatic band (often referred to as the black-and-white band). 27

OCR for page 27
28 LANDSAT AND BEYOND TABLE 3.1  Observing Technology and Key Observables Associated with an Enhanced Program Observing Description of Data Technology (Sensor) Produced Key Observables Typical Applications Fine resolution Optical imagery with Land cover, building footprints, Urban planning, impervious surface optical, stereo submeter to 10-m transportation and utility mapping, transportation maintenance, resolution infrastructure, coastal margins, coastal zone management, wildlife habitat, land surface topography topography, three-dimensional buildings LiDAR LiDAR altimeter and Land surface topography, forest Geomorphology and natural hazards, ice bathymetric measurements canopy height and leaf area, sheet volume, forest productivity and health based on multiple returns built structures Hyperspectral Optical imagery with Physiological signatures of Land carbon fluxes, biodiversity, invasive imaging narrow spectral resolution vegetation, mineralogy, snow species, snow hydrology, mineral contiguous channels grain size, water pollution exploration, volcano gas monitoring SAR, InSAR Active microwave (radar) Surface deformation, forest Natural hazards, water management, climate data structure, soil moisture and impacts, deforestation thaw depth NOTE: SAR, synthetic aperture radar; InSAR, interferometric SAR. Landsat data stream can be greatly increased by exploiting newer technologies that observe the surface at finer resolution and incorporate more of the electromagnetic spectrum. Primary among these modalities is the ability to observe the surface at finer resolution than Landsat’s tens of meters. Power, orbit, and data rate constraints restrict the total volume of data that any satellite can deliver, so it is not possible today to image the full Earth simultaneously at fine scale and rapid repeat times. Remote sensing sensor suites thus require a trade-off between spatial and temporal resolution. At the coarse end of the spatial scale, current technological limits permit the entire globe to be observed daily at spatial resolutions of 0.25 to 1.1 km, as by, for example, the NASA Earth Observing System and NOAA’s Suomi NPP (National Polar-orbiting Partner- ship). Limiting temporal coverage to every 8 days, the globe can be observed at 15 to 100 m by Landsat 7 and Landsat 8 working together at moderate resolution with orbits offset by 8 days. Extending to finer resolutions, specific local areas of about 200 km2 can be observed every 2 to 3 days at 0.5 to 2.6 m by commercial programs like ­ igitalGlobe, or the entire Earth could be observed annually if customer demand justified such a strategy. D If the surface regions of interest are smaller still, airborne sensors can supply data at fine spatial resolutions and regular repeat times of hours to days. Aerial photography is a viable industry, with many companies providing fine-resolution panchromatic and multispectral images. Extending this to a national scale, the U.S. Department of Agriculture’s NAIP (National Agricultural Imagery Program) makes aerial imagery available to government agencies and to the public at no charge. Similarly, the aerial imagery in Microsoft’s Bing Maps is updated annually for the United States and Europe, and Google Earth provides a capability for other providers to upload imagery. A detailed summary of the spatial, spectral, radiometric, and temporal characteristics of all of the land remote sens- ing systems is not included in this report, but Figure 3.1 presents an abbreviated list of important moderate- and fine-resolution satellite remote sensing sensor systems from 1999 through 2015. Landsat represents the current optimal trade-off between resolution, frequency of coverage, and global access constraints. Yet it is clear from Figure 3.1 that there are relatively few existing or proposed moderate-resolution remote sensing systems that can fill this critical need. The French SPOT 5 (2002), the Indian ResourceSat-1 (2003) and ResourceSat-2 (2011), Landsat 8 (2013), and the proposed Sentinel-2b (2014) and ResourceSat-2A are the most important operational systems. The foreign systems may provide useful data for U.S. users as long as the demands on the system are not too great. Without a Landsat-like U.S. instrument, the broad use of moderate- resolution imaging data and the gains of the exploitation of that data will suffer. The committee believes that maintaining the availability of such data is necessary if Earth is to continue to be observed frequently and with

OCR for page 27
ENHANCING A SUSTAINED LAND IMAGING PROGRAM 29 Selected Moderate- and Fine-Resolution Optical Remote Sensing Systems 1999 - 2015 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 1999 IKONOS 2 Landsat 7 Enhanced Thematic Mapper+ 100 m 2000 EROS A 2001 QuickBird 2002 SPOT 5 2003 ResourceSat-1 55 China-Brazil Earth Resources Satellite (CBERS-2) 260 2004 2005 CastoSat 1 Nominal Spatial Resolution 2006 EROS B Panchromatic KOMPSAT 2 VNIR and/or SWIR 2007 WorldView 1 CartoSat 2 Thermal infrared 2008 GeoEye-1 Hyperspectral CartoSat 2A RapidEye 2009 WorldView 2 2010 CartoSat 2B 2011 Pleiades 1 ResourceSat-2 55 SPOT 6 2012 Pleiades 2 KOMPSAT 3 2013 SPOT 7 Landsat 8 120 GeoEye-2 2014 WorldView 3 Sentinel 2b Hyper- CartoSat 2C spectral KOMPSAT 3A CartoSat-1A 2015 CartoSat 3 ResourceSat-2A 55 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Nominal Spatial Resolution (meters) FIGURE 3.1  Characteristics of selected moderate- and fine-resolution optical remote sensing systems, 1999-2015. The spatial resolution of each remote sensing system is portrayed with the following circles: panchromatic (pan) band in orange; VNIR and/or SWIR bands in green; thermal infrared bands in red; and hyperspectral bands in yellow. There are more fine-resolution systems available than moderate-resolution (although several systems are planned for 2014-2015). SOURCE: John Jensen, University of South Carolina.

OCR for page 27
30 LANDSAT AND BEYOND moderate resolution; it believes, further, that the usefulness of these data can be enhanced significantly if fine- resolution data are also available. Why are there not more moderate-resolution remote sensing satellite systems available for use today? Many countries and private-market firms recognize that while moderate-resolution systems are of value, there is more com- mercial demand for finer-resolution panchromatic and multispectral data. Figure 3.1 shows that almost all major public and commercial remote sensing systems are migrating toward finer-spatial-resolution panchromatic and visible and near-infrared (VNIR)/shortwave infrared (SWIR) wavelength bands. Many important applications are not possible using only Landsat-like moderate-resolution data, driving a dramatic shift toward finer spatial resolution. Several important applications and data sources that require finer-scale data than Landsat 8 delivers include the following: • Land use/land cover. Land cover information is categorized by the map scale at which the information is provided.3 Remote sensor data with fine spatial resolution are required to extract high-level information about “landscape metrics.”4 Many city and county agencies throughout the United States and some federal agencies require access to land cover products at a spatial resolution finer than 2 m. • Building and property infrastructure. Almost all counties in the United States collect and store property owner- ship information in a digital system,5 including detailed information about each parcel’s dimensions and all building footprints (perimeters). This effort requires a tremendous amount of remote sensing data collection and processing of fine-resolution imagery throughout the United States every year. Numerous government agencies, including the U.S. Census Bureau, also require building infrastructure information. Fine-resolution imagery can be used to identify the location of new residential structures and the associated road network information. This geospatial information is then conflated with postal and other sources of geospatial data to obtain accurate address information. • Socioeconomic characteristics. The American Community Survey is an ongoing Census Bureau statistical survey that samples a very small percentage of the population every year.6 Local and regional organizations use fine-resolution imagery to predict the spatial distribution of population between censuses to identify new develop- ments or structures and to estimate the number of persons living in each dwelling unit based on building footprint and square footage estimates. • Transportation and utility infrastructure. Federal and state departments of transportation rely heavily on high-resolution stereoscopic aerial photography, satellite imagery, and LiDAR data to monitor transportation infra- structure, allowing them to inventory and characterize roadways, especially to identify deteriorating infrastructure.7 • Hydrology. While moderate-resolution remote sensing data can be used to identify general stream or river centerlines, fine-resolution stereoscopic data or LiDAR data are required to precisely map drainage networks and determine the topography of floodplains for preparing digital flood insurance rate maps 8 and hydrologic models. • Vegetation assessment. Moderate-resolution imagery is useful for monitoring vegetation type (e.g., forest, rangeland, wetland, agriculture), biomass, and functional health over relatively large geographic areas. Fine spatial- and spectral-resolution imagery and LiDAR data can be used to identify vegetation structure, predict watershed runoff, model urban heat islands, and describe agriculture and forest canopy biomass. Extensive remote sensing literature addresses scientific research and applications for vegetation studies based on the use of fine-resolution remote sensing data. • Disaster emergency response examples. The Department of Homeland Security has significant fine-res- olution data requirements, such as determining the boundary of disaster areas and vulnerable structures. 9 USGS 3  J.R. Anderson, E.E. Hardy, J.T. Roach, and R.E. Witmer, A Land Use and Land Cover Classification System for use with Remote Sensor Data, U.S. Geological Survey Professional Paper 964, 1976. 4  M. Herold, J. Scepan, and K.C. Clarke, The use of remote sensing and landscape metrics to describe structures and changes in urban land uses, Environment and Planning A 34:1443-1458, 2002. 5  National Research Council, National Land Parcel Data: A Vision for the Future, The National Academies Press, Washington, D.C., 2008. 6  U.S. Census Bureau, American Community Survey, available at http://www.census.gov/acs/www/. 7  U.S. Department of Transportation, National Consortia on Remote Sensing in Transportation (NCRST), 2012, available at http://www.rita. dot.gov/rdt/remote_sensing.html. 8  National Research Council, Elevation Data for Floodplain Mapping, The National Academies Press, Washington, D.C., 2007. 9  U.S. Government Accountability Office, Homeland Security: Actions Needed to Improve Response to Potential Terrorist Attacks and Natural Disasters Affecting Food and Agriculture, GAO-11-652, 2011, available at http://www.gao.gov/new.items/d11652.pdf.

OCR for page 27
ENHANCING A SUSTAINED LAND IMAGING PROGRAM 31 heavily relies on high-resolution remote sensing data when responding to emergencies, such as the Deepwater Horizon oil spill or Hurricane Sandy. Many other examples are described in the extensive literature on damage mapping. LiDAR (Light Detection and Ranging) Many remote sensing applications require elevation data in order to interpret spaceborne imaging data accu- rately. Topography is well known over much of Earth’s land surface at 5- to 10-m height accuracy and 30-m data postings, but this is inadequate for evaluating such things as water flow patterns, coastal erosion and storm susceptibility, or subtle geologic processes. Existing data typically yield only a single estimate of height for each resolution element in a digital image, whereas for many applications a profile of height is critical. For example, understanding the health and evolution of forested areas requires detailed knowledge of how the biomass is dis- tributed with respect to height. These data are currently best obtained using a profiling LiDAR instrument, which produces the finest-scale surface height measurements (at approximately centimeter accuracy) along with elevation profiles of urban and vegetated regions. Today LiDAR data from aircraft platforms yield detailed masspoint information (i.e., x,y location and z eleva- tion data) about the terrain and buildings, vegetation (trees, shrubs, grass), telephone poles, and roads, for example. The masspoint information can be processed to create digital surface models (DSMs) that contain information about terrain, vegetation, and building height. The vegetation and building height information can be removed from the DSM, creating a bare-earth digital terrain model (DTM), necessary for hydrologic modeling (Figure 3.2). Airborne LiDAR mapping of small areas and terrestrial LiDAR scanning of even smaller footprints form a thriving commercial industry. The LIST (LiDAR Surface Topography) mission to regularly map Earth’s surface at fine resolution (5 m spatial, 10 cm height) is among the recommended missions in the National Research Council report Earth Science and Applications from Space: National Imperatives for the Next Decade and Beyond.10 Launch of the LIST mission is more than a decade away, but SELIP could include access to currently available airborne data and a plan for eventual incorporation of satellite laser altimeter information. Synthetic Aperture Radar The Landsat instruments provide coverage of Earth’s surface in the visible, near-infrared, and thermal-infrared sections of the electromagnetic spectrum. As such, they are primarily sensitive to the chemical composition of the surface. Characteristics of surface shape or texture, including precise measurements of deformation, are best inferred from longer-wavelength sensors operating in the microwave bands, with wavelengths from 3 to 24 cm. In particular, radar remote sensing yields these descriptors of the surface while adding the ability to acquire data when optical measurements are not possible, such as at night or during periods of clouds and inclement weather. Thus SELIP can augment the Landsat series, so that descriptors of the surface invisible to optical instrumentation can be exploited for analysis and operational capability. Because both airborne and spaceborne synthetic aperture radar (SAR) instruments operate at rather long wave- lengths, they image Earth’s surface independent of most weather conditions, day or night, and provide essential capability at high latitudes and in areas with persistent cloud cover. Similarly, the longer wavelengths penetrate well into vegetation, dry soil, and dry snow. These data are sensitive to water content and surface roughness and convey important information about soil moisture. When these radar images are combined interferometrically, as described in the next paragraph, it is possible to map crustal deformation at millimeter levels so that distortions of the surface from natural hazards such as earthquakes and volcanoes, or even from variations in the flow of water or other fluids in the crust, can be visualized (Figure 3.3). Over forested areas, it is possible to map tree heights and canopy distributions, key parameters for measuring Earth’s biomass and its changes. The operating wavelength is generally chosen to maximize performance for specific objectives: short wavelengths for high-resolution imaging, 10 National Research Council, Earth Science and Applications from Space: National Imperatives for the Next Decade and Beyond, The National Academies Press, Washington, D.C., 2007.

OCR for page 27
32 LANDSAT AND BEYOND (a) (b) (c) (d) (e) FIGURE 3.2  LiDAR information extraction: (a) LiDAR- derived masspoints of the Monterey Bay, California, shoreline viewed obliquely; (b) LiDAR-derived bare-Earth color-coded digital terrain model (DTM). (c) LiDAR- derived digital surface model (DSM) of a 7 × 7 km tile collected over Denver, Colorado, containing trees, build- ings, and terrain. The LiDAR data have been shaded using LiDAR Analyst software; higher eleva­ions are in white t and lower elevations are in green. (d) LiDAR-derived DTM with all trees and buildings removed. The Bare Earth grid is automatically extracted from the LiDAR using LiDAR Analyst. (e) LiDAR-­ erived building footprints extracted by ­ d LiDAR Analyst as 3D Shapefiles. These files include geometric and descriptive attributes for each building such as maximum height above ground, roof type, and area. SOURCE: (a,b) Used with permission of John Copple and Sanborn Map Company. (c-e) R. Franklin, ­ iDAR advances and challenges, Imaging Notes 23, 2008, available at http://www.imagingnotes.com/go/ L article_free.php?mp_id=129. LiDAR Analyst is an Overwatch Textron Systems software product designed in 2004 as a plug- in for ArcGIS, ERDAS Imagine, Remote View, and ELT. Courtesy of Imaging Notes Magazine, Spring 2008, and Blueline Publishing, LLC, used with permission.

OCR for page 27
ENHANCING A SUSTAINED LAND IMAGING PROGRAM 33 FIGURE 3.3  A time series of interferometric synthetic aperture radar (InSAR) measurements reveal variable deformation pat- terns from the emplacement of a dike under the flank of the Fernandina volcano, Galapagos Islands. The patterns are similar in the first and last periods, but a faulting event on the caldera rim dramatically altered the shape of the deformation in the middle time period. These patterns are diagnostic of changes in activity within the volcano. The inset at right is the inferred shape of the magma chamber. SOURCE: Reprinted by permission from Macmillan Publishers Ltd: Nature, F. Amelung, S. Jonsson, P. Segall, and H. Zebker, Widespread uplift and ‘trapdoor’ faulting on Galapagos volcanoes observed with radar interferometry, Nature 407(6807):993-996, 2000, copyright 2000. moderate wavelengths for ocean observations, and longer wavelengths to maximize penetration into the surface cover and estimate forest biomass. In interferometric SAR (InSAR) mode, the use of multiple antenna positions—either two antennas on a single aircraft or satellite, or one antenna in a slightly displaced position on a series of separate flight lines or orbits— delivers detailed information about surface topography, a critical parameter of the Earth system supporting many different types of investigations. Time series of such data measure surface deformation at millimeter to centimeter accuracies, permitting monitoring of crustal deformation due to tectonic forces, 11 groundwater flow, or oil and gas extraction, among others. These key measurements extend the usefulness of land imaging far beyond multispectral imaging of the surface. The 2007 NRC decadal survey Earth Science and Applications from Space12 recommended a SAR mission, Deformation, Ecosystem Structure and Dynamics of Ice (DESDynI), but it has no target launch date.13 Other nations have provided most of the leadership and implementation of SAR missions, so an enhanced land imaging program would benefit from including mechanisms and funding to incorporate data from airborne SAR and international SAR missions before a U.S. mission might become operational. Table 3.2 lists the spaceborne systems that have provided the most data for SAR studies. These systems have been developed by several countries around the world and show increasing lifetime, coverage, and resolution over time. Three major civilian radar satellites currently in orbit, none of which is from the United States, are carrying out a variety of investigations of Earth, including studies of crustal deformation. 11  H.A. Zebker, P.A. Rosen, R.M. Goldstein, A. Gabriel, and C.L. Werner, On the derivation of coseismic displacement fields using d ­ ifferential radar interferometry: The Landers earthquake, Journal of Geophysical Research 99:19617-19634, 1994. 12  National Research Council, Earth Science and Applications from Space, 2007. 13  National Research Council, Earth Science and Applications from Space: A Midterm Assessment of NASA’s Implementation of the Decadal Survey, The National Academies Press, Washington, D.C., 2012.

OCR for page 27
34 LANDSAT AND BEYOND TABLE 3.2  Selected Spaceborne Synthetic Aperture Radar Systems Operational Band Wavelength Spatial Resolution System Country or Organization Lifetime (nominal) (cm) (m) SEASAT United States 1978 L-band 24 20 ERS-1/2 European Space Agency 1991-2010 C-band 6 20 JERS-1 Japan 1992-1998 L-band 24 20 SIR-C United States 1994 C/L-band 3/6/24 20 Radarsat-1 Canada 1995-present C-band 6 10 Envisat European Space Agency 2002-2012 C-band 6 20 ALOS-1 Japan 2006-2012 L-band 24 20 Radarsat-2 Canada 2007-present C-band 6 3 TerraSAR-X Germany 2007-present X-band 3 1-3 COSMO-SkyMed Italy 2007-present X-band 3 1-15 (multiple platforms) Hyperspectral Imagery Finally, it is important to recognize that while Landsat produces comprehensive coverage at several distinct wavelengths, additional and stronger characteristics about surface composition follow if the reflectance spectrum is known more completely. Imaging spectrometry acquires such data at hundreds of contiguous spectral bands simultaneously. Its value lies in its ability to provide a high-resolution reflectance spectrum for each pixel in the image. Many, although not all, surface materials have diagnostic absorption features that are only 20 to 40 nm wide. Therefore, spectral imaging systems that acquire data in 10-nm bands contiguously between 400 and 2,500 nm may be used to identify surface materials with diagnostic spectral absorption features. This feature is superior to multi­ spectral remote sensing systems that acquire data in wider, often discontinuous bands. The SELIP would benefit from exploring the advantages and practicality of adding hyperspectral analysis to the planned Landsat acquisitions. Earth’s surface consists mainly of soil, vegetation, snow, ice, and water as well as areas of built structures. Each of these constituents has properties with distinct spectral signatures, which, when measured by a hyperspectral imager, convey information about such properties as productivity, nutrient limitation, water stress in vegetation, soil mineralogy related to locations of natural resources, snow grain size and dust or soot content, and sediment and plankton abundance in water (Figure 3.4). NASA and the Department of Defense have operated airborne imag- ing spectrometers for more than two decades, and more recently, the National Science Foundation, commercial companies, and institutional laboratories have flown airborne instruments. For example, NASA flew the Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) sensor to collect multiple flight lines of hyperspectral data over the Deepwater Horizon oil spill in the Gulf of Mexico (Figure 3.5). Among the recommendations in the 2007 NRC decadal survey for a flight around 2020 is HyspIRI, which combines optical imaging spectrometry with multispectral thermal imagery. HyspIRI has no projected launch date. Such data are valuable for quantification of land surface composition (chemical composition of foliage, mineral- ogy, and other properties) and provide unique information on plant biodiversity and invasive plants. Hyperspectral imagery is extraordinarily flexible because complete spectral coverage (typically in the visible through shortwave infrared regions) is available. This allows specific regions of the spectrum to be selected for current and future data products. Imaging spectroscopy has benefited from technology improvement over the past decades, with improved optics that allow for smaller and less expensive instruments, enhanced downlink capabilities allowing exploita- tion of the entire spectrum, and uniform detector arrays increasing measurement accuracy, precision, and spatial registration. Several technology demonstration spectrometers have flown in Earth orbit, allowing the evaluation of spaceborne imaging spectroscopy data products, and a high-performance imaging spectrometer has flown to the Moon, demonstrating the key aspects of the capability in a prolonged spaceflight environment.

OCR for page 27
ENHANCING A SUSTAINED LAND IMAGING PROGRAM 35 (a) (b) FIGURE 3.4  (a) An AVIRIS hyperspectral data cube of Sullivan’s Island, South Carolina. The image on top is a color com- posite of just three of the available 224 bands (green, red, and near-infrared), and all of the bands are shown below in the depth of the data cube. (b) A comparison of the sensitivity of the 244 AVIRIS bands with the location of the nine Landsat 8 non-contiguous bands. SOURCE: J.R. Jensen and R.R. Jensen, Introductory Geographic Information Systems, Pearson Edu- cation, Upper Saddle River, N.J., page 91, Figure 3-28, 2013. (b) Courtesy of John R. Jensen, University of South Carolina. Table 3.3 identifies characteristics of the most important current and future hyperspectral data collection systems. Commercial and International Data Purchases Expanding SELIP to include additional satellites providing all of the above capabilities would be prohibitively expensive given current budget constraints. Yet the committee believes that it is important to enable access to these data types in a cost-effective way, so that the full value of Landsat-class data is realized, and to enable more advanced work as enhanced capabilities allow. Including these diverse data sources would be a way to bolster a U.S.

OCR for page 27
36 LANDSAT AND BEYOND FIGURE 3.5  Seven flightlines of AVIRIS data collected on May 17, 2010, overlaid on a Landsat Thematic Mapper image of the Gulf Coast and the Deepwater Horizon oil spill. Each of the flightlines can be used to construct a datacube, similar to Figure 3.4(a). SOURCE: Courtesy of NASA/JPL-Caltech/Dryden/USGS/University of California, Santa Barbara, available at http://photojournal.jpl.nasa.gov/catalog/?IDNumber=pia13167. TABLE 3.3  Characteristics of Selected Satellite and Airborne Hyperspectral Remote Sensing Systems Spectral Spectral Instantaneous Total Coverage Interval Number of Quantization Field of View Field of View Sensor Technology (nm) (nm) Bands (bits) (mrad) (°) AVIRIS/ Whiskbroom 400-2500/ 10/5 224/400 12 1.0 30 AVIRISng linear array/ 350-2500 (airborne) pushbroom Hyperion linear array 400-2500 10 220 11 (spaceborne) CASI-1500 Linear (1500) and 370-1050 2.2 288a 14 0.49 40 (airborne) area array CCD (1500-288) a The number of bands and the number of pixels in the across-track are programmable.

OCR for page 27
ENHANCING A SUSTAINED LAND IMAGING PROGRAM 37 national program that ensures continuity and compatibility with the U.S. Landsat archive. 14 In particular, this could serve as a way to increase temporal and spectral coverage relative to what a baseline U.S. system might provide. Not all data feeding the archive of the land imaging program need to be from U.S. spaceborne satellites. Other countries continue to invest in moderate-resolution satellite remote sensing systems, including SAR. In the optical domain, the European Sentinel-2, to launch in 2014, will collect all but the thermal infrared bands of Landsat, and in a wider swath for shorter revisit. The United States will have access to Sentinel-2 data under a free data policy and could complement that with data (also freely available) from a U.S.-funded thermal infrared-only small satellite. Other nations, such as India and Japan, operate land imaging programs that could potentially fill data gaps in moderate-resolution imagery. Data with fine spatial, and in some cases spectral, resolution are available commercially. This would be a comparatively low-cost way to augment a U.S. national program and ensure con- tinuity and compatibility with the U.S. archive. Information from these other sources, if properly integrated in the U.S. imaging program, could increase temporal, spatial, or spectral coverage relative to what a cost-constrained baseline U.S. system might afford. Research and Development Component To ENHANCE a Sustained Land Imaging Program The Sustained and Enhanced Land Imaging Program would include a research and development (R&D) com- ponent with the mission of developing and testing new data products based on the core data sets of the land imaging system. This type of supporting work advances the program with improvements in technology, and experience gained during R&D facilitates iterative improvements in the land imaging program itself. The R&D component also would include development of advanced measuring technologies as well as new measuring requirements that will drive continual improvements in the core land imaging capabilities. Collaborations between the responsible federal agencies, such as USGS and NASA, and private companies will be advantageous. Furthermore, improved collaborations between NASA and USGS may result in the development of new observing technology by the NASA Earth Science Technology Office. Close collaboration between USGS and NASA will also facilitate the transition between research and operations. R&D relevant to a national land imaging program is also being done at compa- nies such as Google and Microsoft. Findings The committee found as follows: • Continuity of moderate-resolution multispectral imagery with global land coverage at weekly frequency is a necessary component of a sustained and enhanced land imaging program, but it is not sufficient for monitoring the range of land surface properties that are critical for both scientific research and operational management. • Optical imagery with fine spatial resolution and data from LiDAR, SAR, and hyperspectral instruments provide distinct and synergistic information about Earth’s land surface. • Commercial companies and other countries are a significant source of land imagery that are not available from programs operated by the U.S. government. • Many important public and commercial applications require fine-resolution satellite and airborne remote sensor data that cannot be satisfied using moderate-resolution Landsat-8-type data alone. Recommendations The committee recommends that the Sustained and Enhanced Land Imaging Program integrate mea- surements from commercial partners, spaceborne sensors recommended by the 2007 NRC report Earth 14  It should be noted that there are some difficulties associated with merging some data from different platforms, e.g., domestic versus inter- national sources. Differences in calibration are often encountered and are routinely solved.

OCR for page 27
38 LANDSAT AND BEYOND Science and Applications from Space, and a variety of airborne sensors and acquisitions to enable analyses not possible using only moderate-resolution multispectral data. These measurements should include, but not be restricted to, the following: • Airborne and spaceborne fine-resolution remote sensing data from public and commercial sources that can be used for detailed land use and land cover, urban infrastructure, transportation, hydrology, and disaster response; • LiDAR data that can be used to extract precise digital surface and terrain models, building and vegetation height information, and vegetation canopy and its internal structure information; • Synthetic aperture radar (SAR) and interferometric SAR (InSAR) images at resolutions suitable for studies of deformation, elevations, and surface cover; and • Hyperspectral data collection and information extraction capabilities for hydrology, ecosystem health and biodiversity, and soil science and mineralogy.