After World War II astronomers turned the military technology of radar to the study of the planets, using it as a way to determine their size, rotation, orbits, and surface features. The space age offered a means to deliver radar systems to the planets themselves, and NASA almost from its inception conceived of radar as a key component of planetary exploration.1 In the 1970s JPL developed a new type of radar for satellite remote sensing. Known as synthetic aperture radar, this technology provided fresh views of Earth and the planets.
As with a camera aperture, larger radar antennas provide higher image resolution. Synthetic aperture radar (SAR) takes its name and operating principle from the simulation of a large antenna, or aperture, by combining observations from many points along the antenna's flight path. An SAR records the amplitude and phase of radar echoes as a function of time, with each point on the planet's surface identified by a combination of time delay and Doppler shift. The SAR then integrates the data along the flight path to synthesize a large aperture (see figure 7.1). In effect, an SAR system substitutes software for hardware—that is, data processing for a physically large antenna.2
The smaller size of SARs was only one of several features that recommended them for use on satellites. Unlike real aperture radars, whose resolution is proportional to altitude, SAR resolution is independent of height and in fact varies directly with antenna size (a smaller antenna providing finer resolution). In practice, the altitude determines the antenna size required for signal transmission and acquisition, but SAR could still reduce satellite-borne radar resolution from tens of kilometers to tens of meters. Radars have the further advantage of collecting data at night and receiving pulses through clouds or through certain surface covers, such as sand or vegetation; the longer wavelengths than optical images also provide different information on the roughness and material properties of target surfaces.3
The ability to penetrate clouds had suggested, in particular, the use of radar on a mission to Venus, whose hazy atmosphere precluded optical observation of the surface. In 1960 JPL started a radar program with Venus in mind. A radar team under Walter E. Brown, Jr., first tested standard radars on rocket flights in conjunction with aircraft-borne radar in order to scale radar backscatter behavior with altitude. The radar team shifted to aircraft flights starting around 1970, and the detection of ocean waves by JPL's airborne SAR in 1971, as we know, led to the Seasat radar.4 The successful tests of the SAR also earned it a central place in initial planning for the Venus Orbiting Imaging Radar mission.5
JPL's shift to synthetic aperture radar came at the recommendation of radar scientists at the University of Michigan. JPL did not originate SAR; the concept dated back to the end of World War II, when the military began sponsoring highly classified programs. By the 1960s commercial geological survey were using airborne SAR systems, and reconnaissance aircraft no doubt were as well.6 JPL's contributions came not in inventing a new technology, but through innovative use of existing technology. The lab departed from current practice first by using longer wavelengths, around 23 centimeters instead of the military's 3-centimeter wavelengths. Unlike the military programs, which wanted the highest possible resolution (and hence smaller antennas and short wavelengths), JPL sought resolution sufficient for scientific purposes, balanced with other mission requirements. In particular, the dense atmosphere of Venus threatened to attenuate shorter wavelengths; at the same time, at longer wavelengths the little-understood Venusian magnetic field might affect signal polarization. So Brown's team settled on an intermediate
wavelength of 30 centimeters, later shortening it to 23 centimeters when they encountered interference from military aircraft radars.7
Seasat's radar resolution of 25 meters would not have excited military interests in advance, but the JPL program more than compensated in other ways. The longer wavelength proved crucial to detecting ocean waves and hence justifying Seasat: higher-resolution systems had been picking up short-wavelength capillary waves on the ocean surface that obscured long-wavelength swells, whereas the JPL radar just saw the swells. The opportunity to map direction and wavelength of wave patterns had piqued ocean-ographers' interest, and the longer wavelength also more easily mapped the internal waves that enabled submarine detection. As the first outfit to try putting an SAR on a satellite, JPL had also to adapt the system to the space environment, including vacuum, low temperatures, and especially vibration. Existing airborne SARs were generally mounted on elastic shock absorbers, but JPL's radar group instead built components rugged enough to handle the vibration of launch and spaceflight. Active radar systems also needed power that passive cameras did not; hence the need for the large solar arrays on Seasat, which would prove its undoing.8
Transferring radars from airplanes to satellites likewise entailed a shift from recording data onboard to transmitting it. That was no easy feat. Synthesizing a large aperture imposed formidable demands in data processing— not only in the integration of radar returns along the flight path, but also in accounting for phase shifts (as the distance changed to each point with the spacecraft motion), planet curvature and rotation, spacecraft attitude drift, atmospheric distortion, and a host of other variables. The SAR data from Seasat emerged at the rate of 111 megabits per second, requiring real-time transmission to Earth instead of onboard processing or recording. On the ground it took an entire day to process about five minutes' worth of data, covering about 200,000 square kilometers; lab staff predicted that at that rate it would take them seventy-five years to process all the images if Seasat worked for three years as planned.9
As it turned out there were only three months of data to worry about, though even that taxed JPL's capability. The lab had initially pursued an alldigital data system for the Seasat SAR, with an onboard processor and digital downlinks and recorders at the ground stations, but technical obstacles and escalating costs shelved the digital system in favor of an existing analog alter-native.10 The radar team thence had to reduce the data on the ground with an optical correlator, in which the radar signal patterns, recorded on film, formed a Fresnel lens; together with corrective optics in the correlator, the lens focused light to reproduce each radar image of the surface (a bit like the production of a hologram). The optical correlator proved a bottleneck in the Seasat data flow, greatly slowing the processing of SAR images and spurring subsequent attempts to develop digital SAR processors.11
The Seasat SAR signaled the increasing centrality of data processing to mission success. The flood of data threatened to swamp scientists. How to select particular bits of information from this incessant stream? How to convert downlinked telemetry to user-friendly data formats? To solve these problems NASA and JPL established the Seasat Data Utilization Project, which converted raw telemetry to scientific data, correlated this with location data, and developed algorithms to extract geophysical information, such as wind or wave patterns. One NASA manager called it "the least glamorous project ongoing at JPL" but at the same time recognized its crucial role in converting and distributing the data, and thus in changing perceptions of Seasat from failure to success.12 Seasat thus demonstrated that space-based SAR could overcome the limited spatial and temporal coverage of airborne radars and provide synoptic, repetitive observations.13 Even before Seasat, JPL managers recognized the programmatic potential for earth and ocean science as well as planetary exploration.14 Venus mission planning continued to include an SAR, and JPL also parlayed the radar into a series of flights on the space shuttle, known as the Shuttle Imaging Radar (SIR). The first of these, SIR-A, flew on the second shuttle flight, in November 1981, and returned images of 10 million square kilometers, about 2 percent of Earth's surface.15 SIR-A confirmed the utility of SAR for both geology and oceanography and provided surprising uses in archeology and anthropology. Airborne SAR flights by JPL in 1977 and 1978 had revealed a system of canals buried under vegetation in Guatemalan rain forests; archeologists attributed the canals to ancient Mayan civilization and concluded that population pressure forced the Maya to develop intensive agriculture instead of primitive slash-and-burn farming.16 SIR-A images of the deep Sahara Desert then revealed prehistoric river channels buried under two meters of sand; the very dry, fine-grained sand of the region proved transparent to radar. The images sent JPL radar scientist Charles Elachi and geologist Ronald Blom on an expedition to Egypt with other geologists and archeologists, where they verified that major rivers had traversed the area in the Tertiary period, carving valleys through a now-featureless landscape in which rain falls at thirty- to fifty-year intervals. The expedition confirmed as well that early human settlements had congregated near these river systems.17
A subsequent flight, SIR-B in 1984, was hampered by antenna and electrical problems and met only 40 percent of its goals, though it did prove the technique of stereo imaging and onboard digital processing.18 The SIR series would continue in the 1990s, after an interruption by the Challenger disaster, and the similarly delayed Venus mission would fly with an SAR as Magellan in 1992. Synthetic aperture radar thus provided programmatic continuity through three decades. It had other implications as well. As with Voyager's flight computer, the complexity of the technology outran its developers. Tony Spear, the Seasat SAR manager, noted before launch that despite pre-flight testing, "we will fly Seasat-A SAR not completely understanding its performance." Spear added, "Some sensors now and some future sensors . . . are as complicated as whole, earlier spacecraft."19
Was this article helpful?