Guide to Magellan Image Interpretation

 

Chapter 3. The Non-SAR Experiments

Jeffrey J. Plaut

 

Introduction

[19] The Magellan radar system operated simultaneously in three modes: synthetic-aperture radar (SAR), altimetry, and radiometry. Data obtained from the altimeter and radiometer modes are extremely important components of many studies of the surface of Venus. This chapter describes the altimetry and radiometry systems and the resulting data sets. Applications that integrate the various data sets, including SAR image data, are presented at the end of the chapter.

 

Altimetry and Derived Data

The Altimetry System

The altimeter, with a nadir-looking geometry, is designed to precisely determine the distance between the spacecraft and a patch of surface approximately 10 to 30 km in diameter (the size of the altimeter "footprint"). Knowledge of the spacecraft's orbital position relative to the center of Venus is combined with the spacecraft-to-surface distance determined by the round-trip travel time of the altimeter signals to produce a topographic map of the surface. The time delays and Doppler-frequency shifts of the echoes are used to isolate portions of the echoes and associate them uniquely with areas of the surface. In addition, the altimeter echoes are analyzed for their intensity and time dispersion to estimate the reflectivity and rms slope for each altimeter footprint.

 

Altimetry Data

A typical Magellan altimetry orbit consists of about 1000 footprints. The resolution varies with spacecraft altitude, and hence with surface latitude, as shown in Table 3-1. The altimetry experiment was designed to produce a net of.....

 


Table 3-1. Altimeter footprint dimensions

Latitude

Cross-track dimension (E-W), km

Along-track dimension (N-S), km

.

80°, -60°

27

15

70°, -50°

24

11

60°, 40°

21

8

50°, -30°

19

11

40°, -20°

16

8

30°, -10°

14

9

20°, 0°

13

9

10°

12

8


 

 

.....measurements with spacing between the footprints approximately equal to the size of the footprints themselves.

Processing of Magellan altimetry data involved generation of an echo "profile" for each footprint [Pettengill et al., 1991; Ford and Pettengill, 1992]. This profile represents the strength of the received echo as a function of time delay (Figure 3-1). For relatively flat, uniform areas of Venus, most of the echo power is expected to come in a short period of time. The profile for these areas therefore displays a strong peak and a rapidly diminishing "tail" (Figure 3-l(a)). The specular (mirror-like) reflection from the nadir dominates the echo profile, while reflections from points off nadir (at larger incidence angles) make only small contributions to the tail of the echo.

If the surface within the footprint contains topographic undulations (slopes of a few degrees or more) at scales larger than the 12-cm wavelength, echoes from favorably oriented slopes off nadir may contribute strong specular reflections. In this situation, the peak of the echo profile is weaker because some surfaces at the nadir are tilted away from the spacecraft,....

 


[
20]

Figure 3-1. Altimeter echo profiles showing echo power vs. time for (a) a flat surface (12.40°N, 187.76°E; rms slope = 1.8 deg), (b) an undulating surface (8.87°N, 188.06°E; rms slope = 6.7 deg), and (c) an area with multiple <<peak>> contributions (8.35°N, 188.10°E; Sapas Mons summit).

Figure 3-1. Altimeter echo profiles showing echo power vs. time for (a) a flat surface (12.40°N, 187.76°E; rms slope = 1.8 deg), (b) an undulating surface (8.87°N, 188.06°E; rms slope = 6.7 deg), and (c) an area with multiple "peak" contributions (8.35°N, 188.10°E; Sapas Mons summit).

 

.....and the tail is stronger than that from a flat surface (Figure 3-1(b)). The echo profile is even more complex when the topography within or near the footprint varies rapidly at large (km) scales. When steep cliffs, valleys, or mountains are illuminated by the altimeter, several strong echoes may return to the spacecraft at different times, creating an ambiguity in the height determination for that footprint (Figure 3- l (c)) [Plaut, 1992].

Before Magellan arrived at Venus, a library of simulate altimeter echo profiles or "templates" was developed [Ford and Pettengill, 1992]. Each template provides the predicted response of a surface, based on three parameters: the distance from the spacecraft, the root-mean-square (rms) slope of surface undulations at scales greater than the wavelength, and the intrinsic Fresnel reflectivity of the surface. In data processing, the three parameters are found by matching the best template to the observed echo profile. In simple terms, the distance is determined by the position of the strongest pea in the profile, the rms slope is determined by the dispersion of the echo in time (the shape of the tail), and the reflectivity is determined by the overall strength of the echo.

Processed Magellan altimetry data are available in two forms: the Altimetry-Radiometry Composite Data Record (ARCDR), and the Global Data Record (GxDR). The ARCD is an orbit-by-orbit data set that contains all of the derived parameters (e.g., radius, rms slope, reflectivity, and emissivity for each altimeter and radiometer footprint. The latitude and longitude of each footprint are provided in the ARCDR, so specific features seen in the image data can be identified in the ARCDR by their map coordinates. While the ARCDR is a huge table of data, the GxDR takes the more familiar form of an image map. All measurements of planetary radius reported in the ARCDR are compiled in image form, using filtering an interpolation techniques, to form the GTDR, a global topographic map of the planet (Figure 3-2). Reflectivity, rms slope, and emissivity data are shown on the GREDR, GSDR, and GEDR, respectively (Figures 3-3 to 3-5). The GxDR maps are presented on CD-ROM in several map projections: full global sinusoidal equal-area projection; a Mercator projection showing latitudes between 65° north and south; an polar stereographic projections centered on the north and south poles, extending to 47° north and south latitude, respectively. The GTDR data sets can be used to study regional topographic patterns and are particularly useful when combined with SAR image data to show the relationships between topography and morphologic features in the images.

 

Reflectivity and RMS Slope Data

As described in the previous section, processing of the altimetry data involved the simultaneous estimation of planetary radius, Fresnel reflectivity, and surface rms slope.

 


[
21]

Figure 3-2. Global topography from Magellan altimetry (GTDR).

Figure 3-2. Global topography from Magellan altimetry (GTDR).


[
22]

Figure 3-3. Global reflectivity from Magellan altimetry (GREDR).

Figure 3-3. Global reflectivity from Magellan altimetry (GREDR).


[
23]

Figure 3-4. Global rms slope from Magellan altimetry (GSDR).

Figure 3-4. Global rms slope from Magellan altimetry (GSDR).


[
24]

Figure 3-5. Global emissivity from Magellan radiometry (GEDR).

Figure 3-5. Global emissivity from Magellan radiometry (GEDR).

 

[25] Through the template matching procedure, values of reflectivity and rms slope are obtained for each altimeter footprint.

Hagfors [1970] derived an expression that accurately predicts the intensity of radar backscatter from many planetary surfaces, as a function of incidence angle, Fresnel reflectivity, and a large-scale roughness parameter C:

 

mathematical equation

 

where

 

Greek letter sigma superscript 0= specific radar backscatter cross section

Greek letter theta= incidence angle, deg

Greek letter rho= Fresnel reflection coefficient (reflectivity)

C= roughness parameter, equivalent to 1/(rms slope in rad)2

 

This scattering function has proven to be appropriate for much of the surface of Venus, at incidence angles less than about 15 deg. Although recent studies [e.g., Tyler et al., 1992] have shown that other functional forms may better describe the scattering behavior of much of Venus, the Hagfors equation applied to Magellan altimetry data can be used to quantify the reflectivity and slope properties of the surface. All altimeter echo profiles have been matched to templates that were derived from the Hagfors scattering function.

Fresnel reflectivity is a measure of the efficiency of a surface in reflecting electromagnetic radiation. A reflectivity of 1.0 represents perfect, total reflection. The typical plains surface of Venus has a reflectivity of about 0.1. Reflectivity (Greek letter rho) can be related to the dielectric constant (e) of the surface:

 

 

A refinement to the determination of reflectivity can be made when a nadir-looking altimeter measurement is accompanied by a side-looking SAR backscatter measurement. Since both of these Magellan datasets are nearly global in coverage, this refinement or "correction" to the reflectivity is possible for almost every altimeter footprint. As described by Pettengill et al. [1988], subwavelength roughness, which dominates the scattering at large (SAR) incidence angles, tends to randomly scatter a nadir-pointing signal, diminishing the power of the echo detected by the altimeter. If this scattering effect is not taken into account, a rough surface will appear to have a reflectivity lower than that of a smooth surface composed of the same material. By estimating the areal coverage of random scatterers from the SAR backscatter crosssections, the reflectivity derived from the template matching procedure can be "corrected" (increased) to a value that is more likely to represent the material properties of the surface, independent of roughness. In the ARCDR files, the uncorrected reflectivity value is reported, along with the correction factor to be applied.

The rms slope parameter is used to quantify the topographic undulation of the terrain, at scales ranging from tens of centimeters to a few kilometers. Magellan altimeter-derived rms slopes for a typical plains surface are in the range of 1 to 3 deg. Tectonic terrains that are severely fractured show values as high as 10 deg. Measurements of terrain slope characteristics of terrestrial test sites have shown that while the rms slope values derived from radar data can serve as a good relative measure of slopes, the absolute magnitude of the values may not accurately describe the actual slopes of the surface. The model from which the rms slopes are derived assumes that the slopes are random in orientation. The template matching algorithm may break down in cases where an organized pattern of slopes affects the scattering. Such a breakdown was observed in Magellan data on surfaces within certain impact parabolic features that showed a strongly asymmetric backscatter behavior between the right- and left-looking SAR data [Plaut et al., 1992].

 

Radiometry and Derived Emissivity

The radiometry experiment employed the high gain antenna in a passive mode to detect thermal radiation emitted by the surface of Venus at radio wavelengths [Pettengill et al., 1992]. Surface emissivity, an electrical property essentially independent of temperature, can be derived from the radiometry measurements to aid in the interpretation of the relative influence of dielectric constant and roughness on observations acquired by Magellan's SAR, altimeter, and the radiometer itself.

All objects emit radiation across the broad range of the electromagnetic spectrum, including the radio frequencies. The amount of energy emitted, or radiance, depends on the object's temperature and emissivity, as expressed in Stefan's law:

 

mathematical equation

 

where

R = total radiance, Wm-2

 

Greek letter sigma= 5.67 x 10-8 Wm-2K-4 (Stefan-Boltzmann constant)

e = emissivity

T = temperature, K

[26] Emissivity, which ranges from 0 to 1, is a measure of how well an object approximates a perfect blackbody radiator. A good emitter is required by thermal balance to be a good absorber, and hence a poor reflector. Ideally, then, the sum of the emissivity and the Fresnel reflectivity should equal one. This applies when measurements are integrated over the full upper hemisphere, which is not the case for Magellan emission and reflection measurements. In practice, the Magellan emissivity and reflectivity estimates track one another in the expected fashion, although their sum typically ranges between 0.90 and 0.95.

To solve the Stefan equation for emissivity, the radiance and temperature must be known. Magellan's antenna acquired the radiance measurements, while the temperature for any given area of the surface was assumed to be a well-behaved function of surface elevation. Observations from Earth and by previous spacecraft indicated that temperatures at the surface of Venus are essentially constant at a given elevation, regardless of time of the Venusian day, or other atmospheric factors. Surface elevations derived from Magellan altimetry data are thus converted to physical temperature estimates, and used in the above equation to solve for emissivity. The actual procedure is slightly more complex, incorporating factors such as the absorption and emission by the atmosphere, reflection of ambient "warm sky" radiation from the surface, and the detailed response of the high gain antenna [Pettengill et al., 1992].

Because the radiometry measurements were interleaved with the SAR observations, the two data sets have the same geometry; incidence angles of the SAR data are equivalent to the emission angles of the radiometry data (see Table 4-1). The radiometry footprint size varies with spacecraft altitude, and hence with latitude (Table 3-2). Parameters associated with each radiometry footprint are reported in the ARCDR files on CD-ROM. The full global ensemble of emissivity determinations is compiled on the GEDR (Figure 3-5), which may be found in digital form on the GxDR CD-ROM.

The emissivity of natural surfaces is controlled primarily by the dielectric constant. Material of high dielectric constant is generally a good reflector and hence a poor emitter. Emission from a surface that is smooth at the wavelength scale is easily modeled, given a value of dielectric constant. Conversely, a measurement of emissivity can be converted to a value of dielectric constant, if the surface is assumed to be smooth. The mean value of surface emissivity observed by Magellan is 0.845, which corresponds to a smooth-surface dielectric constant of about 4.0 [Pettengill et al., 1992]. A rougher surface tends to have a higher emissivity when observed at the horizontal polarization used by Magellan.

 


Table 3-2. Radiometer footprint dimensions (Cycle 1)

Latitude

Cross-track dimension (E-W), km

Along-track dimension (N-S), km

.

10°

23

15

23°, -3°

29

19

46°,-26°

36

27

62°, -42°

51

41

83°,-63°

77

64

90°, -67°

91

77


 

Many surfaces known to be rough from SAR observations- such as tesserae, rough lava flows, and crater ejecta blanketshave emissivities 0.05 to 0.10 higher than surrounding plains Magellan emissivity also shows the expected behavior as a function of emission angle: emissivities at large emission angles (low latitudes) average about 0.05 less than those at small emission angles (high latitudes).

The most noteworthy features in the emissivity data occur in two settings: near the top of the planet's highest mountains and in association with certain impact craters. Most of the blue and violet areas in Figure 3-5 (emissivities less the 0.7) correspond to areas above approximately the 6053.5-km radius contour. Emissivity values in these areas drop as low 0.3, implying dielectric constants as high as 80, which far exceed values measured for terrestrial rocks [Ulaby et al., 1990]. A class of materials that have the electrical properties necessary to enhance the dielectric constant to the observed levels are iron sulfides, including pyrrhotite and pyrite [Pettengill et al., 1988; Klose et al., 1992]. It is proposed that the unique atmospheric conditions at high elevations (lower temperature and pressure) promote chemical weathering of the soils or rock coatings to produce the high dielectric minerals.

An exciting discovery from early Magellan emissivity data was the parabolic-shaped, low-emissivity features associated with some impact craters [Arvidson et al., 1991; Campbell et al., 1992]. These features correspond with many of the green areas in Figure 3-5. They often appear dark in SAR image data, but the correlation between low emissivity and low SAR backscatter is often not strong, implying that dielectric differences are the primary cause of the lowered emissivity [Plaut and Arvidson, 19921. The favored model for the formation of these features involves interactions between the cloud of impact debris lofted into the atmosphere and the strong east-to-west winds known to exist at about 60 km altitude [Campbell et al., 1992]. Why the dielectric constants would be enhanced on the resulting fallout deposits has yet to be explained.

 

[27] Incorporating the Data Sets Into Image Analysis

Analysis of Magellan data is significantly enhanced when relationships can be discerned between features seen in the SAR data and in the non-SAR data sets. Examples in this section show the various methods for combining the data to increase interpretability.

SAR image brightness is dominated by small-scale roughness effects for most of the planet's surface. Topography can influence brightness where slopes are steep enough to significantly alter the local incidence angle. In many cases, however, the SAR data do not indicate the large-scale trends in topography. In these instances, it is useful to compare or combine the SAR data and the topographic information from the altimetry experiment. Figure 3-6 shows the Artemis region in SAR image data and gray-scale altimetry. The altimetry image is extracted from the GTDR file, with a displayed range of planetary radius values from 6047 km to 6058 km. In this representation, black areas represent the lowest elevations and white areas represent the highest elevations. The altimetry data convey much topographic information that cannot be gleaned from the SAR image, such as the complex form of the corona trough, the gentle topographic swell outboard of the trough, and the dramatic asymmetric scarp-and-valley features in the central, northern and northwestern portions of the corona.

Image processing techniques can be used to explore the relationships between morphologic features observed in the SAR images and topography. A common method for combining data sets is the color overlay, in which color-coded elevation data are "painted" over the SAR image. This can be accomplished by separating a color elevation image into its hue, saturation, and intensity components, substituting the SAR image for the intensity component, and recombining the components into a color representation. Topographic contours may be added to further emphasize differences in elevation. Figure 3-7 is a combined elevation-and-SAR image of a region east of Devana Chasma. The color coding and topographic contours show large variations in the elevation of parts of the tessera outliers, as well as the prominent highs associated with two radially fractured features.

Visualization of topography and SAR image morphology is also possible using three-dimensional perspective rendering. The elevation data are used to simulate a view from an oblique direction, with the SAR image superimposed on the topographic shapes. This technique is used to generate the dramatic animated video "flyover" segments. The vertical scale is commonly exaggerated to emphasize subtle topo-.....

 


Figure 3-6. SAR image (left) and topography (right) of Artemis Corona. Planetary radius values between 6047 km (black) and 6058 km (white) are displayed in the topography image.

Figure 3-6. SAR image (left) and topography (right) of Artemis Corona. Planetary radius values between 6047 km (black) and 6058 km (white) are displayed in the topography image.


[
28]

Figure 3-7. Combined SAR image and topography data for a region east of Devana Chasma. Center coordinates are 15°N, 292°E. Contour interval is 200 m.

Figure 3-7. Combined SAR image and topography data for a region east of Devana Chasma. Center coordinates are 15°N, 292°E. Contour interval is 200 m.

 

....-graphic features, particularly when the lateral dimension of the area is large. Magellan three-dimensional perspective views are typically generated with vertical exaggeration factors of 10 to 20. Figure 3-8 is a three-dimensional perspective view of Western Eistla Regio, showing the Guor Linea rift structure associated with the large volcanic edifices Sif Mons and Gula Mons, which loom on the horizon.

Because emissivity is governed mostly by the dielectric properties of the surface, it is often useful to combine emissivity and SAR data to explore the relative influences of roughness and dielectric effects on SAR backscatter cross sections. Many high-backscatter regions, particularly at high elevations, appear bright in the SAR data because of their high dielectric constants (i.e., low emissivities) and not because of increased roughness. Figure 3-9 shows the emissivity of the Atla Regio highlands overlain on SAR image data. The strongest emissivity anomaly (values as low as 0.34) occurs on Ozza Mons, in the upper right of the image. The highest elevations, however, show a return to higher emissivities as well as lower surface roughness, leading to very low SAR....

 


[
29]

Figure 3.8. Simulated three-dimensional perspective view of Guor Linea (foreground) and Sif Mons and Gula Mons (left and right on the skyline, respectively). Vertical exaggeration of relief is 22.5 times.

Figure 3.8. Simulated three-dimensional perspective view of Guor Linea (foreground) and Sif Mons and Gula Mons (left and right on the skyline, respectively). Vertical exaggeration of relief is 22.5 times.

 

.....backscatter cross sections [the "dark summits" of Pettengill et al., 1992]. Interestingly, Maat Mons (left center) shows low emissivity values only on its southwest flanks. Klose et al. [1992] suggest that the irregular pattern of emissivity on Maat Mons could imply that recent volcanic resurfacing has not allowed sufficient time for the weathering of surface rocks to produce the high dielectric materials seen on other mountaintops.

 


[
30]

Figure 3-9. Emissivity of Atla Regio overlain in color on SAR image data.

Figure 3-9. Emissivity of Atla Regio overlain in color on SAR image data. Emissivity values range from 0.35 (violet) to 0.95 (red). Note the relatively high emissivity values at the "dark summit" of Ozza Mons (upper right center) and on all but the southwest flank of Maat Mons (left center).

 

References

- Arvidson, R. E., V. R. Baker, C. Elachi, R. S. Saunders, and J. A. Wood, 1991, "Magellan: Initial analysis of Venus surface modification," Science, v. 252, p. 270-275.

- Campbell, D. B., N. J. S. Stacy, W. I. Newman, R. E. Arvidson, E. M. Jones, G. S. Musser, A. Y. Roper, and C. Schaller, 1992, "Magellan observations of extended impact related features on the surface of Venus," J. Geophys. Res., v. 97, p. 16,249-16,277.

- Ford, P. G., and G. H. Pettengill, 1992, "Venus topography and kilometer-scale slopes," J. Geophys. Res., v. 97, p. 13,103-13,114.

- Hagfors, T., 1970, "Remote probing of the moon by infrared and microwave emissions and by radar," Radio Sci., v.2, p. 445 465.

- Klose, K. B., J. A. Wood, and A. Hashimoto, 1992, "Mineral equilibria and the high radar reflectivity of Venus mountaintops," J. Geophys. Res., v. 97, p. 16,353-16,369.

- Pettengill, G. H., P. G. Ford, and B. D. Chapman, 1988, "Venus: Surface electromagnetic properties," J. Geophys. Res., v. 93, p. 14,881-14,892.

- Pettengill, G. H., P. G. Ford, W. T. K. Johnson, R. K. Raney, and L. A. Soderblom, 1991, "Magellan: Radar performance and data products," Science, v. 252, p. 260-265.

- Pettengill, G. H., P. G. Ford, and R. J. Wilt, 1992, "Venus surface radiothermal emission as observed by Magellan,' J. Geophys. Res., v. 97, p. 13,091-13,102.

- Plaut, J. J., 1992, "'Problem' footprints in Magellan altimetry data" (abstract), papers presented to the International Colloquium on Venus, LPI Contract No. 789, p. 90-92.

- Plaut, J. J., and R. E. Arvidson, 1992, "Comparison of Goldstone and Magellan radar data in the equatorial plains of Venus," J. Geophys. Res., v. 97, p. 16,279-16,291.

- Plaut, J. J., R. S. Saunders, E. R. Stofan, R. L. Kirk, G. G. Schaber, L. A. Soderblom, P. G. Ford, G. H. Pettengill, D. B. Campbell, J. J. S. Stacy, R. E. Arvidson, and R. Greeley, 1992, "Anomalous scattering behavior of selected impact 'parabola' features: Magellan cycle-to-cycle comparisons" (abstract), papers presented to the International Colloquium on Venus, LPI Contract No. 789, p. 92-93.

- Tyler, G. L., R. A. Simpson, M. J. Maurer, and E. Holmann, 1992, "Scattering properties of the Venusian surface: Preliminary results from Magellan," J. Geophys. Res., v. 97, p. 13,115-13,139.

- Ulaby, F. T., T. H. Bengal, M. C. Dobson, J. R. East, J. B. Garvin, and D. L. Evans, 1990, "Microwave dielectric properties of dry rocks," IEEE Trans. Geosci. Rem. Sens., v. 28, p. 325-336.

 
previousindexnext