1U.S. Geological Survey (Emeritus) and Northern Arizona University, 3312 N. Patterson Blvd., Flagstaff, AZ 86004; email-gschaber@flagmail.wr.usgs.gov

2U.S. Geological Survey (Emeritus) and Northern Arizona University, 189 Wilson Canyon Rd., Sedona, AZ 86336

Four Shuttle Imaging Radar (SIR) missions flown between November 1981 and October 1994 enable us to demonstrate important roles that multifrequency and polarimetric SAR play in geologic mapping and paleoenvironmental studies in desert regions. Bir Safsaf, within the hyperarid core of the Sahara in southern Egypt, was recognized following the SIR-A and SIR-B missions in the 1980's as one of the key localities in Northeast Africa where penetration of blow sand and sandy alluvium by L-band (25-cm wavelength) radar signals delineates significant underlying geologic features. Characterized by sand-blanketed outcrops of sandstone, highly fractured granite, migmatite gneiss, and fully aggraded paleodrainage channels, Bir Safsaf was targeted as a focal point during the two SRL (SIR-C/X-SAR) missions to assess the role of multifrequency and polarimetric SAR in optimizing sand penetration and the delineation of underlying geology. In this summary report we focus on the use of SIR-C images to map sand-blanketed geologic contacts and structural features in the granitic basement rocks just north of Bir Safsaf. Radar depiction of fracture patterns in sand-blanketed granitic rocks there and near Dungul Oasis, 150 km southwest of Aswan in southern Egypt was investigated in the field and is attributed to the presence of clay weathered from the granites, in the fractures.

The detectability of most geologic features and structures underlying pervasive blankets of sand in desert regions such as the NE Sahara depends primarily on radar frequency, the thickness of the cover of blow-sand, and the nature and complexity of the underlying geology. The overall ranking of the utility of the SIR-C/X-SAR frequency bands and polarizations for geologic mapping below centimeter to meter deep blow sand at the Bir Safsaf site is found, in order of decreasing priority, to be: (1) LHV, (2) LHH(VV), (3) CHV, (4) CHH(VV) and (5) XVV.

Bir Safsaf in southern Egypt (lat 22°45' N., long 29°15' E.), is a small and inconspicuous oasis of grasses and a few palms that lies within the "hyperarid core" of the eastern Sahara--one of the driest and most desolate places on earth (Figs. 1-2). This is one of the localities in the Eastern Sahara where sand-buried paleodrainage channels of middle Tertiary to Quaternary age were earlier discovered by the authors using L-band data from SIR-A (McCauley et al., 1982, 1986a,b; Breed et al., 1983). Schaber et al. (1986) showed from field investigations and laboratory measurements of radar-scattering and dielectric properties that the dry sand sheet, blow sand, and sandy alluvium around Bir Safsaf are characterized by unusually low microwave attenuation at L-band frequencies. Thus, this area is ideally suited for subsurface imaging to depths as much as 1 to 3 m using L-band signals. As a result of these earlier SIR investigations, Bir Safsaf was selected as a focal point for SIR-C/X-SAR to assess the importance of multifrequency and polarimetric SAR sensors in optimizing "radar imaging depth" in sandy sediments, and in mapping underlying geologic features of potential economic importance (Schaber et al., 1995, 1997).

The reader is referred to Schaber et al. (1997) for a more detailed description of the SAR responses to the basement complex, and to other important SIR-C/X-SAR observations at Bir Safsaf not discussed here, including (1) selective portrayal on X-band images of wind-formed bedrock grooving, (2) mapping of sand-buried paleodrainages at X-, C-, and L-bands, and (3) detection of near-surface ground water associated with the location of the oasis itself at L-band and a steep antenna depression angle.

In this summary report we focus on the application of the multifrequency and polarimetric SIR-C/X-SAR images to delineate Precambrian age granites and granite gneisses which, being blanketed by blow sand, are not delineated on images acquired by conventional visible/near visible sensors (such as Landsat TM) (Figs. 3-5). This report also summarizes results of our most recent field expeditions in southern (Upper) Egypt near Dungul Oasis and northwest of Abu Simbel between November 23 and Dec. 2, 1996, and at Bir Safsaf between Feb. 28 and March 6, 1997 (just prior to the final SIR-C/X-SAR Team Meeting in Florence, Italy on March 11-14, 1997).

The Western Desert of Egypt south of 24° N. latitude is characterized by a monotonous, nearly horizontal sand plain surface underlain by gently dipping sandstones and shales ranging from Paleozoic to Cretaceous in age. Erosion has removed most of the younger sedimentary rocks and exposed a basal quartzitic sandstone and conglomerate that erodes to a lag of quartz pebbles. The sedimentary rocks unconformably overly Precambrian basement crystalline rocks that are a part of the North African Shield. A large number of east-west trending normal faults extend in places for several hundred kilometers and locally account for basement uplift (Issawi, 1981; Jas et al., 1988). Sporadic outcrops of basement rocks include a major one just north of Bir Safsaf (Fig. 2). The surface of the Atmur el Kibeish region that surrounds Bir Safsaf is dominated by the northern extension of the Selima Sand Sheet, which covers about 105 km2 in southern Egypt and NW Sudan (Bagnold, 1931; Haynes, 1982; Breed et al., 1987). The sand sheet is a resurfacing product of wind erosion and deposition as eolian processes replaced fluvial processes in the late Quaternary (Breed et al., 1987). The ubiquity of the cover sands gives the plain around Bir Safsaf its monotonous appearance both on the ground and on Landsat TM images (Fig. 4a). On the ground, the sand sheet appears as an almost featureless, flat, to slightly rolling sand plain. It is overridden in a few areas by barchan dunes, but most of the wind-blown sediments are in the form of sand sheets and drifts. These deposits consist of loose grains, almost entirely quartz; they range in size from silt particles to small gravel (0.2 to 2 cm) but are mostly sand and granules (0.06-4.0 mm), which are typically coated with iron-oxide and clay minerals (Schaber et al, 1986; Davis et al., 1993). In the area around Bir Safsaf, surficial materials include windblown mineral components (Fe+, Mg, and clay minerals) derived from extensive exposures of Precambrian basement crystalline rocks and their weathering products. Computer processed Landsat TM images show that this clay-rich material has accumulated as a veneer on the surface (Fig. 9 in Davis et al., 1993.)

Because rainfall is now less than 1 mm/yr. and plant life is virtually nonexistent (except in a few small, isolated ground water oases such as Bir Safsaf), surface materials are finely laminated by wind and are little affected by soil moisture, vegetation, or bioturbation. Despite the present-day lack of rainfall, some ground water is still present at depths of 1.5 m to 3 m in the alluvium at scattered localities around Bir Safsaf (See Effects of Incidence Angle in Schaber et al., 1997).

The major outcropping (1,200 km2) of mostly Precambrian igneous and metamorphic rocks centered just north of Bir Safsaf includes primarily red granites, granodiorite, and migmatitic gneisses (Jas et al., 1988) (Fig. 2). Individual outcrops rarely rise more than a meter to several meters above the surrounding plain and commonly form low, rounded hills thinly veneered by lag gravel and windblown sand (Fig. 3). The outcrops are difficult to impossible to distinguish on Landsat TM data (Fig. 4). The absolute age(s) of the sand-mantled basement rocks at Bir Safsaf has not yet been well documented, and only reconnaissance geologic mapping has been completed (Schandelmeier et al., 1983; Schandelmeier and Darbyshire, 1984; Jas et al., 1988) (Fig. 2). Trenching in the low hills north of the oasis in 1984 (Fig. 8 in Schaber et al., 1997) revealed decomposition of the metamorphic rocks to a soft, mica and clay-rich (montmorillonitic) saprolite. The deep tropical weathering that produced the saprolite probably occurred during the Tertiary. Much of the saprolite cap has been removed by erosion except in low places and where clays have infiltrated into fractures in the beveled bedrock surface. We have suggested that the presence of these weathered products may play an important role in determining what the radar "sees" below the sand-covered basement rocks (Schaber et al., 1997).

Ten SIR-C/X-SAR image data takes (d.t.) of Bir Safsaf were acquired during SRL-1 and SRL-2 (Table 1) (Schaber et al., 1997). The surface footprints of the SAR data takes are of different swath widths over Bir Safsaf because of site targeting constraints, but are tightly nested. Three additional SRL-2 d.t.'s that included Bir Safsaf were acquired in the interferometric mode at identical incidence angles for the purpose of obtaining topographic information (Moreira et al., 1995)(Table 1). These interferometric data were not processed because the terrain at Bir Safsaf is virtually flat. External radar interference (ground-based and airborne) is a problem on many of the CHV and LHH images acquired of this site during SRL-1 and SRL-2.

The SAR images of different frequencies and polarization modes reveal important differences in geologic content (Figs. 4, 5, 6a-e). Schaber et al. (1997). suggest that, not unexpectedly, the XVV images of Bir Safsaf and vicinity contribute less overall geologic information from below the cover of blow sand than do the C- and L-band images that were acquired simultaneously. However, the X-band images are considerably better than either the C- band or the L-band images in portraying remarkably well the linear, wind-fluting pattern developed on the bedrock presently exposed to the wind, or buried to depths of 1 cm to 20 cm beneath the cover of blow sand (Fig. 6a). This enhancement of the eolian grooving at X-band was attributed by Schaber et al. (1997) to the phenomenon called "Bragg scattering" (Ulaby et al., 1982), and was confirmed during our field expedition to Bir Safsaf in 1997 (see Field Work at Bir Safsaf in 1997).

The co-polarized CHH image and CVV image differ little in overall content, as do the LHH images and LVV images. Although there are rare but subtle differences between the CHH(VV) and LHH(VV) images, characteristics attributed here to either CHH or LHH images are attributed as well to the CVV and LVV images, respectively, acquired during the same data take. The CHV images from SRL-1 data take 114.4 are superior to the co-polarized (HH, VV) C-band images for mapping fractures and delineating geologic contacts-- possibly because of the predominance of blocky facets ( e.g., multiple signal bounce) in the highly fractured granitic rocks (Fig. 6b,c). The LHH and LHV images from the same d.t. (Fig. 6d,e) are, however, superior to the XVV, CHH, and the CHV images with regard to the enhancement of bedrock geologic detail observable from beneath a centimeter to a meter or so of sand cover. This enhancement is especially apparent within the highly fractured granites and granite gneisses comprising the basement rocks near Bir Safsaf (See Portrayal of Crystalline Basement Rocks).

Differences in the geologic interpretability of SIR-C/X-SAR images of specific frequency and polarization result primarily from frequency and/or polarization-dependent surface scattering and volume scattering from within and below the overprint of eolian blow sand. On the basis of detailed field documentation carried out by the authors between 1982 and 1997, Schaber et al. (1997) interpreted these differences in SIR-C/X-SAR backscatter response as dominantly the result of: (1) increased "radar imaging depth" (to detectable subsurface dielectric inhomogeneities or a substrate horizon) with decreasing SAR frequency (i.e., increasing wavelength), and the sensitivities of SAR signals of different frequencies and polarizations to different scales and geometries of micro-roughness both on the surface and within the shallow subsurface, (2) the presence of radar-attenuating clays (derived from the extreme weathering (saprolitization) of the Precambrian rocks under earlier wetter conditions and blow sand that are preserved in the many fractures that characterize the basement rocks--and thus enhance their detection on the SAR images especially at L-band, and (3) the existence on the surface, and below the blanket of blow sand, of secondary deposits (such as iron oxide, calcite, gypsum, halite, etc.) with high dielectric permittivity.

Portrayal of the Precambrian Basement Rocks at Bir Safsaf on the SIR-C/X-SAR Images --The value of using multifrequency SAR images for mapping surface and shallow subsurface geology in sand-mantled desert terrains is best documented by the granitic stocks and fractured migmatites that dominate basement rocks just north of Bir Safsaf. The structures and geologic boundaries within these basement rocks are portrayed with increasing effectiveness by the X-, C-, and L-band SARs, respectively (Figs. 2-6). Schaber et al. (1997) showed that this portrayal is effectively illustrated for the reader through the use of simple lineament (structure) maps of the Safsaf site compiled from SRL-1 d.t. 114.4 (Fig. 7)(see Fig. 6 in Schaber et al., 1997 for the individual SRL-1 DT. 114.4 images used to compile the maps).

The commonly radar-bright migmatite gneisses at Bir Safsaf are criss-crossed with radar-dark fractures (Figs. 4, 5, 6c-e). Clays derived from the weathering of the granitoids and migmatites may be the key to understanding why the radar "sees" all of the extremely fine igneous and metamorphic structures (Schaber et al., 1997). The chemical weathering process turns the feldspars into clays and leaves the quartz behind, preserving the original rock texture. Thus, in the various basement rocks where feldspars and mafic minerals are present, clays would be abundant (See analytical results described below in PORTRAYAL OF BASEMENT ROCKS NEAR DUNGUL OASIS) . These clays (mostly Smectite and Montmorillonite) would subsequently be concentrated within the granite stocks (see below) and preserved, along with blow sand, in various topographic traps (i.e., fractures) beneath the weathering cap and possibly extending down from the cap. Because clays absorb, or attenuate, radar signals, these fracture zones appear "dark" (weakly backscattering) in contrast to the surrounding granites and granite gneisses, which being fresher and roughened by exfoliation, are radar-bright by comparison to the clay-filled fractures.

Three ovoid-shaped granitoid stocks are recognized using dominantly the CHV- and L-band images from d.t. 114.4 (Fig. 5c-e). The 10-km-long by 17.5-km-wide stock immediately west of Bir Safsaf (see Fig. 6c ) appears radar-dark, or dark-mottled, on all three SAR images (i.e., X-, C-, and L) (Fig. 6). This stock (hereafter referred to as stock A) has been truncated by a major east-west fault and secondary north-south faults (Figs. 1-2, 4-5). Stock A is mantled on its west side by active sand (sand sheet and barchan dunes) that is presently being transported across the stock from north to south (Fig. 4b). Stock B, just north of stock A, is 10 km wide by 20 km long (Fig. 6c) and is recognizable on the X-band image (Fig. 6a) only as an indistinct, semi-ovoidal structure--not a distinct geologic feature. Stock B, however, is increasingly better discriminated on the CHH, LHH, CHV, and LHV images, respectively (Figs. 6b-e). The improved discrimination of stock B with increasing wavelength is mainly a result of increased radar contrast of this radar-dark geologic unit with surrounding crystalline rocks that are generally more complexly fractured than is stock B.

A third stock, C (9.3 km by 12.4 km), just east of stock A, is poorly delineated on the X-SAR image (Fig. 6a). Similar to stocks A and B, stock C is increasingly better discriminated as a distinct, radar-dark, ovoid-shaped feature on the CHH, LHH, CHV, and LHV images, respectively (Fig. 6). None of these ovoidal features has been previously documented from surface reconnaissance mapping or analysis of conventional, visible-wavelength images (Jas et al., 1988) (Fig. 1).

In contact with the north side of stock B is a highly fractured gneiss that is not identifiable on the XVV images but is portrayed with increasing effectiveness on CHH, CHV, LHH, and LHV images, respectively (Figs. 4-7). The increased mappability of the many fractures in this gneiss on the CHV images, compared to the CHH or CVV images, is especially dramatic. As shown in Figure 6, the fractured terrain has increasingly greater contrast in radar backscatter, compared to the surrounding terrains, on the XVV, CHH, , LHH, and LHV images, respectively.

Multifrequency SAR Backscatter From Selected Surfaces--In order to "quantify" the visual observation of increasing geologic diversity with decreasing SAR frequency using the SAR images of Bir Safsaf, we have in Figure 8 plotted the C- and L-band radar backscatter coefficient values against the X-band backscatter coefficient values (± 2.3 dB) for eleven geologic surfaces (See Fig. 4 in Schaber et al, 1997 for location and description of the eleven sites selected). Figure 8 emphasizes the increasingly larger deviation in backscatter from the eleven sites at C- and L-band, respectively, compared to the X-SAR backscatter values for these sites. Relative to the L-band data, the X-band data commonly have uniform backscatter across substantially different geologic terrains, thus suggesting that the backscatter returns at X-band are being masked to a large degree by the pervasive cover of blow sand where it exceeds several tens of centimeters in depth. Therefore, one can consider the deviations in radar backscatter values at C- and L-band (as compared to the X-band data) as evidence of SAR frequency-dependent "geologic diversity" below the surface blow sand cover that is best portrayed using the lower frequency (or larger wavelength) SAR sensors.

No quantitative or automated approach can ever surpass the wholly qualitative radar photo geologic interpretation of SAR images for remote studies of desert geology, geomorphology, and paleoclimate. However, based on the results shown in Figure 8, automatic screening of multifrequency, multipolarization, SAR scenes for potential variations in geologic diversity with SAR frequency possibly could be effective using simple algorithms designed to detect unusual contrasts in backscatter with different frequencies. Such a procedure could help determine which SAR bands or band/polarization combinations might provide maximum geoscience information. Similar statistical approaches have been suggested using Landsat MSS and TM data (Chavez, Jr., et al., 1982; Sheffield, 1985).

Field Work at Bir Safsaf in 1997- Following the authors' participation in the Centennial Symposium of the Geological Survey of Egypt (GSE) in Cairo In November 1996, they were invited by the Director of the GSE (Dr. Gaber Naim), to participate with GSE geologists in a geologic reassessment of the basement rocks at Bir Safsaf (Fig. 3). The focus of this field expedition was to re-assess the economic mineral potential of these basement crystalline rocks using the sand penetration capabilities of long wavelength SAR sensors. Between Feb. 28 and March 6, 1997, the authors joined the GSE geologists in a special tent camp GSE set up for this purpose 60 km northeast of Bir Safsaf (Fig. 9). Although the focus of the planned field work was somewhat compromised by their provision of a bulldozer instead of the requested backhoe for excavating trenches, many important geologic features within the crystalline basement rocks were examined in the field. Major tasks focused on: (1) Sampling the gravels deposits of now dry stream valleys (wadis) that once cut through the basement rocks, for the purpose of determining the likelihood of any placer minerals of economic value, (2) exploring and sampling contact aureoles surround granite stocks B and C for determining any economic mineral potential, and (3) documenting the geologic, environmental, and physical attributes of both the surface and shallow subsurface (below the blow sand) responsible for unique SAR signal responses from selected geologic features.

A number of pits were excavated by hand because the bulldozer could not be transported off road to the remote areas close to Bir Safsaf. Samples were collected for mineral assay by EGSMA. Several of the samples appear to include minute gold flakes imbedded within massive quartz veins that cut the granite bodies. The results from these assays have not been completed by the GSE at this writing; however, the quartz dikes and veins appeared overall to be quite dry (hydrothermally)

Figure 10 shows the ground view of the radar-determined contact between stock B and the rougher granite terrain immediately to the east (see Fig. 6c). The eastern side of stock B is mantled by a flat sand sheet with only occasional rare nubbins of granite protruding through the sand. The terrain east of stock B is characterized by more abundant and more massive nubbins of a red granite unit which appears to pre-date stock B. The photograph shown in Figure 11 was taken on top of a low hill (5-8 m high) of massive gray granodiorite that forms the bright spot surrounded by a radar-dark aureole within in stock B (See Fig. 6d for location). The radar-dark aureole was found on the ground to be represented by a smooth and flat sand sheet (Fig. 12). Although this sand sheet could not be excavated to expose the underlying bedrock (because of the lack of a backhoe), isolated nubbins of granite gneiss encountered during our traverse across the aureole provided evidence as to the nature of the less resistant rocks underlying the radar-dark aureole. This result fits our general observation that the denser, more massive, exfoliated granites and granodiorites that comprise the basement complex in the vicinity of Bir Safsaf commonly form local, radar-bright, outcrop clusters that are commonly elevated slightly above the sand sheet. Conversely, the more easily weathered granite gneisses in the basement complex commonly have been eroded level with the sand sheet surface--and being blanketed by a substantial sand sheet thicker than 1-2 meters or more, are radar-dark even at L-band.

In late November 1996, just prior to the Centennial Symposium of the Geological Survey of Egypt in Cairo, the authors in collaboration with Egyptian geologists visited another exposure of granitic basement rocks located approximately 50 km southeast of the Dungul Oasis and 150 km southwest of Aswan in southern Egypt. The appearance of these basement rocks on the ground is quite similar to those at Bir Safsaf; that is, blow sand obscures the fractures and covers all but the highest nubbins of the most competent granite (albeit exfoliated and wind eroded) that commonly protrude 2 m or less above the sand blanket (Fig. 13). These basement rocks, like those at Bir Safsaf, were uplifted by tectonic events and subsequently stripped of the overlying sedimentary rocks by a major, pre-Nile river system referred to as the Qena System by Issawi and McCauley (1992). To the north, around Dungul Oasis, marine sedimentary rocks of early Tertiary age form the resistant cap rocks of the Limestone Plateau (see Issawi and McCauley, 1992).

In Figure 14 are shown enlargements from the fully processed (7.5 m/pixel) CHH, CHV, LHH, and LHV enlargements of the Dungul basement site that were acquired during SRL-2 data take 50.4 . As described above at Bir Safsaf, the CHH image shows little evidence that basement rocks exist just below the thin cover of blow sand. The CHV image faintly shows the presence of highly fractured and radar-bright granites near Bir Safsaf. The LHH and LHV images nicely portray the outline of the basement outcrop at the Dungul site, contrasting the rough (radar-bright) granite outcrops and the smoother (radar-dark), more radar-attenuating, sand and clay-filled fractures associated with the granites. The fracture patterns at both the Bir Safsaf and Dungul sites are characteristic indicators of buried granitic terrain, but the fractures in the Safsaf basement rocks appear to be more abundant, and somewhat better defined on the CHV, LHH and LHV images than at Dungul (compare Fig. 6d and Fig. 14). Considering only the two basement rock sites investigated to date in southern Egypt, it appears that such sand-blanketed basement complexes can be most confidently identified and mapped using long wavelength SAR (e.g., L-band). Of course, lower frequency radars such as P-band (e.g. 440 MHz; 68 cm wavelength) would be even more effective in mapping basement rocks with economic potential in desert environments, especially where sand cover is an obstacle to conventional Landsat TM-type sensors.

In order to determine if the Dungul basement rocks, like those at Bir Safsaf, have been deeply weathered (saprolitized), we hand dug two pits--one alongside a granite outcrop, the other along a local fracture zone between granite outcrops (Fig. 15). The test pits verified the existence of a minimum of at least 40 cm of highly altered granite, including the presence of a fine-grained, yellow-brown material concentrated along local fractures. Samples collected in November 1996 from these yellow-brown fracture fillings from the pits were subjected to X-ray diffractometer analysis of the clay minerals. Analysis (Dr. Roderic Parnell Jr., Department of Geology, Northern Arizona University, Flagstaff Arizona) showed the presence of 80% smectite (montmorillonite-randomly interstratified smectite-illite, approximately 90% smectite layers), 16% kaolinite, 2% mica or illite, and 2% talc. These clays could only have been derived from in situ weathering of the granite during a much wetter climatic period than at present (Schaber et al., 1997; McCauley et al., 1997a,b). The bulk samples collected as grus (disintegrated granite) from the bedrock between the clay-filled fractures is composed of 47% quartz, 29% plagioclase, 23% K-feldspar, 1% amphibole (probably tremolite), 1% rhodochrosite, and 1% total clays.

The maximum radar imaging depths documented in dry blow sand at the Dungul basement site is similar to that observed at Bir Safsaf: 40 cm and 50 cm at C-band, and 150 cm and 200 cm at L-band, assuming a loss tangent of 0.007 and 0.005, respectively, for the blow sand. (see below).

Table 2 lists values for the radar imaging depth (m) and signal attenuation values (dB/m) estimated for X-, C-, and L-band signals in dry silica blow sand and three types of alluvium. These results are based on (1) the measured range of loss tangent values for these materials reported by Schaber et al. (1986), (2) empirically-derived algorithm for "radar imaging depth" based on examination of backhoe trenches excavated at Bir Safsaf following SIR-A and SIR-B, and (3) laboratory and field analyses of the SIR-C/X-SAR data (Schaber et al. 1997). Field observations at Bir Safsaf and the near Dungul in 1996 and 1997 substantiate the theoretical maximum radar imaging depths in such low loss materials of 0.10 to 0.3 m depth at X-band, 0.2 to 0.5 m depth at C-band, and 0.8 to 2.0 m depth at L-band (see Table 2).

Through compilation of SAR wavelength and polarization-independent lineament maps of local structures (Fig. 7) we have demonstrated that the use of SIR-C/X-SAR multifrequency and polarimetric radar images of Bir Safsaf significantly increases the potential for effective remote geologic mapping in desert areas. This conclusion is especially true where bedrock and other geologic features are covered by only a centimeter to a meter or so of blow sand. For the Safsaf site, an overall ranking of the utility of the SIR-C/X-SAR frequency bands and polarizations for geologic mapping below the blow sand cover would be, in order of decreasing priority: (1) LHV, (2) LHH(VV), (3) CHV, (4) CHH(VV) and (5) XVV (Schaber et al., 1997). A similar ranking would also apply to the XVV, CHH(HV) and LHH(HV) SRL images acquired of the Dungul basement complex. The lower overall ranking of SAR images acquired at higher frequencies reflects the increased radar imaging depth with decreasing frequency.

The justification for complete global mapping of desert regions of the world by multifrequency and polarimetric SAR sensors, especially the lower frequency SARs remains quite strong scientifically. It is within the vast regions of arid-to-hyperarid lands, which account for approximately 30 percent of the land surface, where much of the geologic record of paleoclimate change, as well as untapped natural resources lie preserved but hidden from conventional view below often thin deposits of blow sand.

Acknowledgments

Special thanks are due to the outstanding Astronaut Crew members of Space Ship Endeavour for the SRL missions 1 and 2 (STS- 59 and 68), and to the excellent SIR-C/X-SAR Project teams at NASA/JPL (California), DARA (X-SAR, Germany), and ASI (X-SAR, Italy) for their SAR sensor design, software development, mission planning and operations, and data processing and distribution to the project principal Investigators. The authors also thank Dr. Miriam Baltuck, SIR-C Program Scientist at NASA Headquarters. This SIR-C/X-SAR research was funded by NASA/JPL contract WO-8864 to the U.S. Geological Survey (for G. Schaber), by NASA contract 960529 to Northern Arizona University (Flagstaff, Arizona) (for G. Schaber), and by NASA/JPL contract 958452 to Northern Arizona University (for J. McCauley).

Bagnold, R.A., 1931, Journeys in the Libyan Desert 1929-1930, Geogr. J. 78:13-39.

Breed, C.S., Schaber, G.G., McCauley, J.F., Grolier, M.J., Haynes, C.V., Issawi, Bahay, and McHugh, W.P., 1983, Subsurface geology of the Western Desert in Egypt and Sudan revealed by Shuttle Imaging Radar (SIR-A), Spaceborne Imaging Radar Symposium (Jan. 17-20, 1983), JPL Pub. 83-11, Jet Propulsion Laboratory (Pasadena CA), pp. 10-12.

Breed, C.S., McCauley, J.F., and Davis, P.A., 1987, Sand sheets of the eastern Sahara and ripple blankets on Mars, in Desert sediments--Ancient and modern (L. Frostick and I. Reid, Eds.), Special Publication 35, Geological Society of London, pp. 337-359.

Chavez, Jr., Berlin, G.L., and Sowers, L.B., 1982, Statistical method for selecting Landsat MSS ratios, J. Applied Photographic Engineering 8:23-30.

Davis, P.A., Breed, C.S., McCauley, J.F., and Schaber, G.G., 1993, Surficial geology of the Safsaf region, south-central Egypt, derived from remote-sensing and field data, Remote Sensing of Environ. 46:183-203.

Haynes, C.V., 1982, Great Sand Sea and Selima sand sheet, eastern Sahara: Geochronology of desertification, Science 217:629-632.

Issawi, Bahay, 1981, Geology of the southwestern desert of Egypt, Annals of the Geological Survey of Egypt XI:56-66.

Issawi, Bahay and McCauley, J.F., 1992, The Cenozoic rivers of Egypt: The Nile problem, In Friedman, R., and Adams, B. (Eds.), Followers of Horus: Studies dedicated to M.A. Hoffman, pp. 1-18, Egyptian Studies Association, Publication No. 2, Oxbow Monograph 20, Oxbow Books, Oxford, Britain.

Jas, C., Klitsch, E., Schandelmeier, H., and Wycisk, P. (Eds.), 1988, Jebel Uweinat/NF 35; Geological map of NW-Sudan and SW-Egypt, Technische Fachhochschule, Berlin, scale 1:1,000,000.

McCauley, J.F., Schaber, G.G., Breed, C.S., Haynes, C.V., Grolier, M.J., Issawi, Bahay, Elachi, Charles, and Blom, Ron, 1982, Subsurface valleys and geoarchaeology of Egypt and Sudan revealed by Radar, Science 218:1004-1020.

McCauley, J.F., Breed, C.S., Schaber, G.G., McHugh, W.P., Haynes, C.V., Issawi, Bahay, Grolier, M.J., and El-Kilani, A., 1986a Paleodrainages of the eastern Sahara--The radar rivers revisited, IEEE Geoscience and Remote Sensing GE-24,:624-648.

McCauley, J.F., Breed, C.S., and Schaber, G.G., 1986b, The megageomorphology of the radar rivers of the eastern Sahara, The Second Spaceborne Imaging Radar Symposium (April 28-30, 1986), JPL Pub. 86-26, Jet Propulsion Laboratory (Pasadena CA), pp. 25-35.

McCauley, J.F., Issawi, Bahay, Breed, C.S., and Schaber, G.G., 1997a, Spaceborne imaging radar (SIR) geologic results in Egypt (A review: 1982-1997), 1996-Annals of the Egyptian Geological Survey, Proceedings of the Centennial science meetings of the Egyptian Geological Survey, Cairo (Nov. 19-22, 1996), in press.

McCauley, J.F., Breed, C.S., Schaber, G.G.,1997b, The Sahara paleodrainages: SIR-C/X-SAR Flights; Final Report to NASA (Jet Propulsion Laboratory), 26p.,17 illustrations, this volume

Moreira, J., and 11 others, 1995, X-SAR interferometry: First results, IEEE Trans. on Geoscience and Remote Sensing 33:950-956.

Schaber, G.G., McCauley, J.F., Breed, C.S., and Olhoeft, G.R., 1986, Shuttle Imaging Radar: Physical controls on signal penetration and scattering in the eastern Sahara, IEEE Geoscience and Remote Sensing GE-24:603-623.

Schaber, G.G., McCauley, J.F., Breed, C.S., and Issawi, Bahay, 1995, The roles of wavelength and polarization in geologic studies of Bir Safsaf (Egypt) using SIR-C/X-SAR data, Geological Society of America Abstracts with Programs , v. 27, No. 6, p A110, (New Orleans, Nov. 6-9, 1995).

Schaber, G.G., McCauley, J.F., and Breed, C.S., 1997, The use of multifrequency and polarimetric SIR-C/X-SAR data in geologic studies of Bir Safsaf, Egypt, Remote Sen. of Environ., v. 59, no. 2- pp. 337-363.

Schandelmeier, H., Richter, A., and Franz, G., 1983, Outline of the geology of magmatic and metamorphic units between Gebil Uweinat and Bir Safsaf (SW Egypt/NW Sudan), J. of African Earth Sciences 1:275-283.

Schandelmeier, Heinz, and Darbyshire, Fiona, 1984, Metamorphic and magmatic events in the Uweinat-Bir Safsaf uplift (Western Desert/Egypt), Geologische Rundschau. 73:819-831.

Sheffield, Charles, 1985, Selecting band combinations from multispectral data, Photogramm. Eng. and Remote Sensing 51:681-687.

Ulaby, F.T., Moore, R.K., and Fung, A.K., 1982, Radar remote sensing--Active and Passive, vol. II-Radar remote sensing and surface scattering and emission theory , 608 p., Addison-Wesley Publishing Company, Reading, Mass.

Von Hippel, A.R., 1954, Dielectric and waves, 284 p., John Wiley and Sons, Inc., New York.

Figure 1- General view looking south at Bir Safsaf showing the isolated date palm trees and Safsaf grasses rising above a moderately thick (several meters) local accumulation of blow sand (Photo acquired March 1997).

Figure 2- Index map showing localities and geology of the Bir Safsaf region, Egypt (modified from Jas et al., 1988). The large bold rectangles outline the three SRL-1 SAR data takes (d.t.) 82.41 (dot-dash), d.t. 114.4 (solid), and d.t. 130.4 (dashed; see Figs. 4, 5). Elevations within the sand plain covered by the SAR images range from about 230 to 260 m above sea level. Local topographic highs reach about 290 m at the southern end of dt. 130.4.

Figure 3- General view looking south from north of Bir Safsaf across nubbins of granite and granite gneisses protruding through a thin (centimeters to a meter or two) sand sheet (photo acquired March 1997).

Figure 4- (a) Mosaic of an enhanced and color-stretched Landsat TM scene (stretches- blue: 0-0, 119-0, 160-127, 195-255, 255-255; green: 0-0, 72-0,103-127, 134-255, 255-255; red: 0-0, 119-0, 172-127, 235-255, 255-255) scaled to overlay the northwestern half of an L-band (HH polarization) SAR scene from d.t. 130.4 of the Bir Safsaf site (acquired 17 April, 1994). Landsat image (acquired 14 November, 1984; Path 177, Row 44; quads 2 and 3) was processed at the U.S. Geology Survey Flagstaff Field Center (Flagstaff, Arizona). S, location of palm trees at Bir Safsaf; (b), Repeat of the same LHH SAR scene shown in (a). A circular structural feature (11.5 km in diameter) centered at D may reflect the presence of a granitoid stock at depth (feature not covered on narrower image swath acquired during d.t. 114.4, see Fig. 6). North is toward upper left of the images coincident with the edge of the partial Landsat TM subscene shown in (a). SAR image flight direction left to right (descending orbit); illumination direction northeast. SAR incidence angle at image center is 26°; pixel spacing 12.5 m (azimuth and range). Radar image size is 52 km X 100 km. S, location of Bir Safsaf.

Figure 5- Enlarged X, C, and L-band false-color composite image of the basement crystalline complex north of Bir Safsaf produced from a portion of SRL-1 d.t. 114.4 by assigning blue to XVV; green to CHH, and red to LHH. North arrow and location of palm trees at Bir Safsaf (S) indicated; incidence angle at the center of composite is about 45°. SAR flght direction left to right (descending orbit); SAR illumination direction is northeast. Pixel spacing 12.5 m azimuth and range. Image covers 30 km X 52 km; Digital data prepared at the U. S. Geological Survey Flagstaff Field Center (Flagstaff, Arizona) and processed into the color composite at the Jet Propulsion Laboratory (Pasadena ,California).

Figure 6- Enlargements of a part of SRL-1 d.t. 114.4 showing details of Stock B and the highly fractured metamorphic terrane immediately to the north; (a) XVV, (b) CHH, (c) CHV, (d) LHH, (e) LHV. Incidence angle approximately 45°. North indicated on (a); radar illumination direction to northeast. Stock B indicated on (c). Images 23.7 km X 35.5 km.

Figure 7- Sketch maps showing undifferentiated lineaments, fractures, and other structural elements at the Bir Safsaf site; (a) XVV, (b) CHH, (c) CHV, (d) LHH. Maps were compiled from (a)-(d) shown as Figure 4 in Schaber et al. (1997). Note, especially, the markedly increased level of structural detail observable with decreasing SAR frequency, especially in the basement rocks north of Safsaf (see Figs. 2, 4). Note also the increased level of geologic detail on the CHV map compared to the CHH map. S, location of Bir Safsaf.

Figure 8- Correlation of SRL-1 d.t. 114.4 XVV backscatter coefficient values with CHH (a), CHV(b), LHH (c), and LHV (d) backscatter coefficient values for eleven diverse geologic surfaces within the Bir Safsaf site, as described in Schaber et al. (1997). Solid lines represent linear regression fits to values with R= linear correlation coefficient. Error bars (standard deviation) are indicated. [Note that the R value decreases and the scatter of points increases with decreasing SAR frequency compared with XVV, reflecting the increased geologic diversity actually observed on the longer wavelength SAR images (see Multifrequency SAR Backscatter From Selected Surfaces)] (See Table 1 for additional information on SRL-1 d.t. 114.4)

Figure 9- View of tent camp north of Bir Safsaf in March, 1997.

Figure 10-(a) Ground view of rugged, exfoliated, red-granite outcrops forming the sharp, radar-bright contact on the east side of stock B (see uppermost B in Fig. 6d for location), (b) ground view looking west across stock B taken near the location shown in Figure 10a. This surface is mantled by sand sheet and is radar-dark at C and L-bands (see Fig. 6c-e)

Figure 11- Ground view taken within stock B showing radar-bright hill of exfoliated, granodiorite outcrops in distance, and radar-dark, sand-sheet covered aureole surrounding granite outcrop in forground. (see G in Fig. 6d for location).

Figure 12- Ground view taken atop radar-bright low hill seen in distance in Figure 11 (see caption of Fig. 11).

Figure 13 -Ground view across the sand-blanketed, granitic basement complex studied near the Dungul Oasis in southern Egypt.

Figure 14- Enlarged CHH (a) and LHH (b) images of the basement complex near Dungul shown in Figure 8. Images acquired during SRL-2 d.t. 50.4. The L-band image includes some external radar interference patterns thought to arise from ground based radars located near the Aswan dam 150 km to the northeast. Circular radar-bright feature is an inselberg; Barq el Sahab.

Figure 15 -Close up of pit excavated beside granite outcrop at the Dungul basement site to examine weathered (saprolitized) granite is present, as is the case at Bir Safsaf.

Table 1- SIR-C/X-SAR data takes (d.t.'s) acquired of Bir Safsaf and vicinity during SRL-1 and SRL-2.

Table 2 - Theoretical radar imaging depth (m) and attenuation (dB/m) vs. loss tangent for X-, C-, and L-band radar signals in desert sediments with low to medium loss.