Figure 1. NASA's Dryden Flight Research Center and the Jet Propulsion Laboratory (JPL) operate the DC-8 Airborne Laboratory and the AIRSAR/TOPSAR system.

Figure 2. Two separate radar antennas allow topographic data to be collected in a single pass of the aircraft. The difference in arrival times of the return signal at the two antennas can be related to the relative height of the target because the positions of the antennas are known. During nominal operation, one antenna transmits while both receive. In "ping-pong" mode, however, both antennas transmit and receive. The phase noise increases in ping-pong mode, but the vertical resolution also increases roughly by a factor of 2. Only the ping-pong mode is capable of resolving the major topographic features in the Texas coastal zone.

Figure 3. TOPSAR digital terrain model (DTM) of Bolivar Peninsula acquired in June 1996 by NASA/JPL. Horizontal resolution is 10 m . Approximately 1 million data points describe this barrier spit, which has a total relief of only about 5 m. The view, looking northeast, shows the southwest portion of the peninsula, which is 16 km long.

Figure 4. This is a TOPSAR shaded-relief topographic image of the same portion of Bolivar Peninsula. Despite TOPSAR's low resolution and high RMS, many of the cultural and geographic features visible in the ALTM image are discernible in the TOPSAR DTM. Dunes, beach ridges, elevated roadbeds, and other features with as little as 1 m of relief are identifiable.

Figure 5. This contoured surface depicts the differences between the ALTM and TOPSAR elevations. We averaged ALTM elevations over 10 x 10 m squares and then subtracted the corresponding TOPSAR elevation. The contoured residuals show that TOPSAR elevations tend to be 0 to 1 m above the ALTM elevations along the Gulf shoreline but drop to 1 to 3 m below the ALTM in center of the peninsula. This down-range departure of the TPSAR from the ALTM is consistent throughout the study area and suggests a systematic error in the TPSAR.

Figure 6. This plot compares topographic measurements across an undeveloped portion of the peninsula (C-C'). (see fig. 5 for location). The beach profile extends from the waterline across the beach, foredune, back dunes, and the barrier flat. We conducted a rapid-static GPS survey to measure this beach profile, then sorted the ALTM data for laser points that fell within +/- 2 m of the transect line. The land surface is covered by a variety of grasses and low shrubs that grow up to 0.5 m in height. The scatter in ALTM topography therefore reflects height error, real topographic variations within +/- 2 m of the beach profile line, and random reflections of vegetation.