Category: GIS and Remote Sensing

The geopolitical realities of remotely sensed data became plain in the aftermath of the 11^th September attack by terrorists on the United States.  The US National Imagery and Mapping Agency (formerly the Defense Mapping Agency of the Department of Defense) placed a moratorium on release of digital elevation data derived from NASA’s February 2000 Shuttle Radar Topography Mission, “in the interests of national security”.  The SRTM, which used radar interferometry from dual antennae on a 60 metre long boom, was intended to satisfy the huge demand from Earth scientists for digital elevation models of the continents for a large range of applications, ranging from accurate hydrological mapping to sophisticated mathematical analyses of landforms.  An accurate, high-resolution DEM is central to rapid topographic mapping of those many parts of the world where published scales do not exceed 1:250 000.  NIMA also maintains the classified DTED Level-1C global elevation data set, derived from a variety of sources, including clandestine aerial and satellite photography, and which has a resolution as precise as 30 metres.  SRTM data are reported to be more revealing.  At the heart of cruise-missile guidance and the real-time imaging radar used for navigation in low-flying, all-weather military aircraft lies DTED Level-1C data.  Such facilities are not known to be in the possession of, or under development by any agencies other than the military of a small number of developed countries, for obvious economic reasons.  Oddly, elevation data as revealing as DTED Level-1C for the whole of the USA and its territories are still available freely from the US Geological Survey.  Anyone “targeting” installations, either for military or more innocent purposes, need look no further than the growing number of commercial image providers who sell satellite images with spatial resolutions as good as 1 metre.  Indeed, some such companies currently promote their wares through images of Manhattan Island in the aftermath of 9^th September, and there is a thriving business in selling aerial photographs of real estate with resolutions up to the 10 centimetre level.

Browsing through the archive of data from the Terra satellite, particularly those from the ASTER instrument (visit http://edcimswww.cr.usgs.gov/pub/imswelcome/ ), reveals a disproportionate focus on Afghanistan compared with much of the rest of the world.  The majority of Afghan images were captured before 11^th September, and the area is hardly a priority for scientific research.  ASTER produces stereographic images with a 15 metre resolution, suitable for producing high-quality digital elevation models that rival those of SRTM.  It would not be surprising to discover that US and British Special Forces engaged in Afghanistan not only carried large-scale topographic maps derived from ASTER images, but also commercial Ikonos 1-metre images, that are capable of pinpointing vehicles and concentrations of people.  Nor is it surprising that relief agencies, intent on delivering humanitarian supplies to emergencies of many different kinds in nearly unknown terrain, rarely if ever have such sophisticated navigational aids.

Interferometric radar and faults of the Mojave Desert

Though it requires considerable computing power and specialized software, the use of “before” and “after” radar data to detect small-scale subsidence or shifts in the horizontal plane, is a potentially powerful tool in neotectonics (see Radar analysis of Turkish earthquake, August 2001 Earth Pages).  Motion detection by such radar interferometry becomes even more useful as historic radar images accumulate.  The workhorse for radar interferometry is the European Space Agency ERS series of satellites, which produce synthetic aperture radar images about 150 km wide along the same track, orbit after orbit.  The system has operated since 1992, so there are rich possibilities for multitemporal use of the distance-measuring capacity inherent in radar imagery.  Means of assessing the regional build-up of strain in seismically active areas are important in earthquake prediction, and such synopses help understand the tectonics at work there.

In terms of seismicity and tectonics there is no better studied area than that extending from the Pacific coast of southern California across the San Andreas Fault and the Mojave Desert.  Radar interferometry provided by 25 pairs of ERS images from 1992 to 2000 produces a spectacular picture of the gradual development of ductile strain underlying this risky area (Peltzer, G. et al. 2001.  Transient strain accumulation and fault interaction in the Eastern California shear zone.  Geology, v. 29, p. 975-978).  Unsurprisingly, shear strain along the San Andreas fault system shows up well.  The Garlock Fault that marks the NW flank of the Mojave is apparently resting after 10 thousand years of motion that averaged 7 mm per year.  The authors focus on displacements associated with the diminutive, by Californian standards, Blackwater and Little Lake fault systems, which trend SE-NW to link the epicentres of the 1872 Owens Valley earthquake and that at Landers in 1992.  Within 10 km of these aligned faults are clear signs of a step in strain rate, that suggests that the lineament lies above a major, active ductile shear zone; perhaps the birth of a new fault system.  Should this system fail in a brittle fashion it is likely to result in an event with a magnitude greater than 7 on the Richter scale, and a surface break more than 100 km long.  Peltzer et al. have achieved a test of concept for interferometric radar’s use in seismic risk assessment, that can be deployed anywhere, given the computing resources.  Their work transcends after-the-event studies that do little to assist the victims of earthquakes.

The Earth Observing-1 (EO-1) satellite, launched by NASA in late 2000, carries two remote-sensing instruments that may become operational devices in the future, given a proven track record on EO-1 and, of course, sufficient funding.  One, the Advanced Land Imager (ALI) is a test bed for sensors earmarked for the follow-on to the current Landsat-7 Enhanced Thematic Mapper+ (ETM+).  As well as the existing ETM+ six bands, ALI covers three others close to existing bands.  Whether by design or good fortune, two of these help define the important VNIR broad absorption by ferric iron minerals, neglected in remote sensing since the early days of the Landsat Multispectral Scanner.  Like the ETM+, ALI also carries a panchromatic band that spans the visible range, and which is aimed at providing a means of sharpening detail in images.  On ALI, however, this band has an improved resolution of 10 metres as opposed to the current 15.

More innovatory is the Hyperion instrument, a hyperspectral device that spans the visible to short-wave infrared range with 242 bands that are 10 nanometre wide.  Hyperion is comparable with airborne hyperspectral devices, such as AVIRIS.  In the experiment it captures data swathes that 7.7 km wide, made up from 256 pixels with a resolution of 30 m.  After initial difficulties with allowing for atmospheric effects  on the data, newly calibrated Hyperion data closely mimic mineral spectra.

Early work on EO-1 data in many fields, including geology, has proved sufficiently promising that NASA has given the mission a year-long extension.  Although data are restricted to only a few target areas suggested by the investigators, the extension is good news.  It is a reassurance about continuity of the Landsat programme, and a tantalising indication that the ill-fated hyperspectral Lewis satellite may be resurrected.

Information from: http://eo1.gsfc.nasa.gov/

The destructive Kocaeli earthquake (magnitude 7.4) of August 17 1999 involved horizontal slip of up to 5 metres.  Although it is possible to measure strains precisely using GPS arrays, many stations are needed to fully grasp strain patterns.  Interferometric processing of before and after radar data (InSAR) presents an opportunity to examine seismic strains over very large areas.  Displacements associated with the Kocaeli earthquake on the North Anatolian Fault, recorded by InSAR, extended for up to 60 km either side of the fault (Mayer, L. and Lu, Z. 2001.  Elastic rebound following the Kocaeli earthquake, Turkey, recorded using synthetic aperture radar interferometry.  Geology, v. 29, p. 495-498). 

The fault runs parallel to the look direction of SAR beams from the ERS-2 satellite in its ascending orbits.  This fortuitous geometry charted relative motions in a horizontal sense on either flank of the major strike-slip fault system, with a precision of about 3 cm.  Interesting in its own right, the recorded strain helps understand how and where the elastic strain energy released by earthquakes was stored.  The key to energy storage is the rebound pattern associated with strain release during earthquakes, to which the InSAR results are an approximation.  This pattern depends theoretically on the displacement along the fault itself, the shear modulus of the rock involved and the depth to which faulting extends.  In the case of Kocaeli, faulting penetrated to between 6 and 15 km below the surface.  Because elastic strain builds up around active faults, it may be possible to use InSAR monitoring as a means of predicting the risk of future failures on dangerous faults, like the North Anatolian Fault.  Earthquake records show that successive failure migrates westwards along the Fault, getting ever closer to Istanbul.

In the same way that topographic contours can be transformed to models of continuous elevation change using surface fitting, measurements of gravitational and magnetic field potentials, at points on the ground or along aerial survey lines, are sources of imagery.  Expressed as contours joining points with the same value, spatial distributed data are notoriously difficult to interpret, however much information they contain.  Not only do contours simplify the data by dividing them into arbitrary steps, how we interpret contour maps depends on how we perceive them.  Our eyes evolved to extract information distributed as a continuum across our field of view, and our visual cortex developed many tricks to innately interpret clues to shape, perspective and distance, to extend the limits of stereoscopic vision (we see objects in true 3-D only if they are closer than about 400 metres).  Our innate abilities “interpret” contours in terms of the spacing between them; the closer they are together the darker we perceive the area of steep gradient.  In other words we have to convert an image that is the “negative” of the first derivative to an understanding of the actual shape represented by contours!  Unsurprisingly, we have to learn to “read” maps, and that is a great deal more difficult for those showing potential-field intensity than for topographic elevation.  Cartographers long ago latched onto our use of shadows as clues to shape, and designed maps with shading as if the Sun was shining from the top of the sheet.  They also use different colours as a second clue to what is high and low.  Combining the two aids helps transform images of geographic variables – basically bland shifts from high to low – into visually stunning, and therefore more easily interpreted pictures.  Surface modelling of elevation and geophysical data, with such graphic tricks, literally throws hidden, and often unsuspected features into sharp relief.

These techniques have revitalized desktop interpretation of the world, especially using results of geophysical surveys.  However, in the same way that detail of a terrain blurs and loses information as resolving power falls, low-resolution data of other kinds obscure buried features, or give ambiguous hints to what they are and where they go.  Reducing the spacing of aerial surveys, and the height from which they are acquired, increases the resolving power of the technique.  Stunning examples of the state of this particular art appear in recent work by the US Geological Survey (Grauch, V.J.S. 2001.  High-resolution aeromagnetic data, a new tool for mapping intrabasinal faults: example from the Albuquerque basin, New Mexico.  Geology, v. 29, p. 367-370.  See also http://rmmcweb.cr.usgs.gov/public/mrgb/airborne.html ). 

Grauch worked on an area in which superficial materials and rapid rounding of topography result in poor surface expression of all but the largest faults.  By using aeromagnetic images modelled from survey lines spaced at 100 to 150 metres, he picked out not only hidden faults, but also the magnetic signatures of pipelines, water tanks and buildings.

up with playing Solitaire or Hearts between those moments of productive inspiration?  NASA Ames Research Center has set up a cottage industry (unpaid) to help Mars specialists there build a catalogue of impact craters on the Martian surface.  As those flyers tucked under your windscreen wipers say, “No experience needed”.

Probably the most important scientific breakthrough from studies of the Moon since the 1960s has been the discovery that its pocked surface resulted from impacts by chunks of interplanetary debris.  The rate of impact and the size of the colliding bodies, and therefore the energy that they delivered, has varied since the Moon formed.  The lunar cratering record, backed up by accurate dates of its products, is a detailed chronology of how impacts influenced Earth’s evolution – vital, since signs of impacts rapidly become masked by our planet’s vitality.

Mars, on which NASA scientists and many more besides focus their undivided attention, is also cratered as a result of the same kind of process.  Counting craters, measuring their diameters (a proxy for the energy involved in their formation) and looking for their age relative to one another and other features of the Martian scene is an excellent means of assessing aspects of the Red Planet’s evolution.  But Mars is a great deal bigger than the Moon, and the sheer tedium of doing the work has become a burden.  Those geologists who compiled the lunar record have moved on, and few relish the task as a profession, hence Ames’ appeal for public participation.

The idea is that the basic information on crater occurrence, size and relative age – that’s based on relations between overlapping craters and degradation by Mars’ “weather” – can easily be gathered by interested, but untrained people.  The statistical work can then be done much more quickly.  If you fancy being a NASA “Clickworker”, then connect to http://clickworkers.arc.nasa.gov/top

Since inception on November 17, 2000, all clickworkers combined have contributed 340,070 crater-marking and 93,891 crater-classification entries.  It seems better by far than simply running the SETI distributed software to analyse radio frequencies for possible signs of intelligence out there.  You get to look at some magnificent high resolution images too.

The 15-channel imaging system aboard the first of NASA’s Earth Observing System constellation of satellites (Terra) began to demonstrate its potential in November.  The 9 month delay between Terra’s launch in December 1999 and the appearance of its first scientific data irked many potential users, already chewing carpets because of the 18 month delay in the launch.  However, wrangling between ASTER’s designers at ERSDAC, the Japanese space agency, and NASA was resolved by November 17^th.  If you are interested, the data can be accessed at the new EOS Data Gateway (edcimswww.cr.usgs.gov/imswelcome/).  Some 70 000 scenes are already “in the can”, but slow processing means that only a trickle of calibrated (Level 1b) data adds to the archive daily, so that cover is very patchy at present.  Nonetheless, quality of cloud-free scenes is excellent, and the new potential, especially for geological challenges, is dramatic.  It is worth noting that the data are in a somewhat difficult format, and can be had either on tape (tar format) or by ftp downloads (file format 125 Mb per scene).  The EOS Data Gateway does plan for release eventually on CDs..

Using satellite images for geological mapping and exploration, or for monitoring short-lived phenomena, such as volcanic eruptions, is now standard Earth sciences technology.  But it involves substantial costs for data and for the software needed for analysis, or so it was.  Access to the most recent images from the US Landsat-7 and French SPOT systems is now on-line using sophisticated browsing sites on the Web.  Both enable guest users, as well as those who have signed up for slightly more sophisticated services, to browse and download reduced-resolution JPEG versions of archived images, and to order data, if needs be.  For Landsat-7, go to http://landsat7.usgs.gov/ though this means going through several pages.  To jump straight to the Earth Observation System (EOS) Data Gateway try http://edcimswww.cr.usgs.gov:80/ims-bin/pub/nph-ims.cgi?endform=1&u=259015&sid=959259015%2D52891371&mode=SRCHFORM .  This currently opens a data search and order form.  Choose a search keyword first using the Data Set button, selecting Landsat-7 Level 1 data.  You can choose several options for the geographic search area, and simply enter a date range (e.g. 2000-01-01 and 2000-05-25 for this year’s archives.  Then Start Search.  Sometimes your search will take quite a while, dues to pressure on the server’s bandwidth.  The good news is, you can disconnect and go back later to the relevant page using Internet Explorer or Netscape History listing.  For SPOT, access is via the DALI server at http://www.spotimage.fr/home/proser/whatdali/daligst/daligst.htm or the Sirius server at http://sirius.spotimage.fr/anglais/Welcome.htm – the Sirius service is a little more complicated than DALI, but is set to become SPOT-Image’s standard browser.

Image quality in both cases is excellent, with the Landsat-7 browse images having a roughly 250 m resolution, and SPOT data showing at about 120 m (4 to 8 times better than similarly available data from meteorological satellites).  Use the right mouse button with cursor over the image and select Save Image As: assigning your own name instead of the default given by the server, e.g.  geology1.jpg.  You can then make some cosmetic changes to contrast and colour balance using graphics software such as MS PhotoEditor or Adobe PhotoShop.

Remember that SPOT data of whatever kind are covered by SPOT-Image copyright, but the USGS who distribute Landsat-7 data make no such claim.  Clearing copyright for publication and acknowledging sources is an important responsibility for uses in research or publications.

SRTM and ASTER

There are several other web sites to watch.   In February 2000 NASA, the US National Image and Mapping Agency (NIMA), and the Italian and German Space Agencies deployed the Shuttle Radar Topographic Mission (SRTM) aboard a Space Shuttle flight.  The SRTM uses radar reflection received by two antennae separated by a long arm deployed from the Shuttle to estimate topographic elevation of the Earths surface, with a method known as radar interferometry.  The resulting data are in the form of a digital elevation model (DEM) with elevation values for cells 30 metres square.  A DEM is therefore a 3-D model of topography, and shows landforms in stunning detail, together with geological features that control them.  Because radar relies on energy transmitted from the spacecraft and radar waves can penetrate cloud, the SRTM produces data whatever the time or the weather.  The mission successfully captured the entire continental surface between 60°N and 60°S, and will revolutionize both geomorphology and geology.  From November 2001 the US Geological Survey and the German Space Agency (DLR) will release DEMs publicly and at low cost, but ‘tasters’ are available from the following web sites:  NASA – http://www.jpl.nasa.gov/srtm DLR – http://www.dlr.de/srtm .

For the next decade or so, the main Earth-oriented thrust by NASA is the Earth Observing System (EOS), which will be a constellation of satellites that orbit from pole to pole to give coverage of the entire surface.  On 18 December 2000 NASA launched the first of these, named Terra.  This satellite carries several payloads that produce images of various kinds, the most geologically important of which is the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), designed and built by the Japanese space agency (ERSDAC), but operated jointly with NASA.  ASTER captures images for several ranges of wavelength in the visible, reflected and emitted infrared, which are designed to highlight the spectral properties of common minerals.  So, ASTER is a geological remote-sensing system par excellence.  The system is working properly, but will begin to produce scientific data sometime in late 2000.  Like the SRTM, the plan for ASTER is eventually to produce full coverage of the continental surface during its lifetime.  For the moment, you can keep a watching brief by visiting NASA – http://terra.nasa.gov and ERSDAC – http://astweb.ersdac.or.jp

Using satellite images for geological mapping and exploration, or for monitoring short-lived phenomena, such as volcanic eruptions, is now standard Earth sciences technology.  But it involves substantial costs for data and for the software needed for analysis, or so it was.  Access to the most recent images from the US Landsat-7 and French SPOT systems is now on-line using sophisticated browsing sites on the Web.  Both enable guest users, as well as those who have signed up for slightly more sophisticated services, to browse and download reduced-resolution JPEG versions of archived images, and to order data, if needs be.  For Landsat-7, go to http://landsat7.usgs.gov/ though this means going through several pages.  To jump straight to the Earth Observation System (EOS) Data Gateway try http://edcimswww.cr.usgs.gov:80/ims-bin/pub/nph-ims.cgi?endform=1&u=259015&sid=959259015%2D52891371&mode=SRCHFORM .  This currently opens a data search and order form.  Choose a search keyword first using the Data Set button, selecting Landsat-7 Level 1 data.  You can choose several options for the geographic search area, and simply enter a date range (e.g. 2000-01-01 and 2000-05-25 for this year’s archives.  Then Start Search.  Sometimes your search will take quite a while, dues to pressure on the server’s bandwidth.  The good news is, you can disconnect and go back later to the relevant page using Internet Explorer or Netscape History listing.  For SPOT, access is via the DALI server at http://www.spotimage.fr/home/proser/whatdali/daligst/daligst.htm or the Sirius server at http://sirius.spotimage.fr/anglais/Welcome.htm – the Sirius service is a little more complicated than DALI, but is set to become SPOT-Image’s standard browser.

Image quality in both cases is excellent, with the Landsat-7 browse images having a roughly 250 m resolution, and SPOT data showing at about 120 m (4 to 8 times better than similarly available data from meteorological satellites).  Use the right mouse button with cursor over the image and select Save Image As: assigning your own name instead of the default given by the server, e.g.  geology1.jpg.  You can then make some cosmetic changes to contrast and colour balance using MS PhotoEditor or Adobe PhotoShop.

Remember that SPOT data of whatever kind are covered by SPOT-Image copyright, but the USGS who distribute Landsat-7 data make no such claim.  Clearing copyright for publication or acknowledging sources is an important responsibility for uses in research or publications.