Abstract– We detail the Kamil crater (Egypt) structure and refine the impact scenario, based on the geological and geophysical data collected during our first expedition in February 2010. Kamil Crater is a model for terrestrial small-scale hypervelocity impact craters. It is an exceptionally well-preserved, simple crater with a diameter of 45 m, depth of 10 m, and rayed pattern of bright ejecta. It occurs in a simple geological context: flat, rocky desert surface, and target rocks comprising subhorizontally layered sandstones. The high depth-to-diameter ratio of the transient crater, its concave, yet asymmetric, bottom, and the fact that Kamil Crater is not part of a crater field confirm that it formed by the impact of a single iron mass (or a tight cluster of fragments) that fragmented upon hypervelocity impact with the ground. The circular crater shape and asymmetries in ejecta and shrapnel distributions coherently indicate a direction of incidence from the NW and an impact angle of approximately 30 to 45°. Newly identified asymmetries, including the off-center bottom of the transient crater floor downrange, maximum overturning of target rocks along the impact direction, and lower crater rim elevation downrange, may be diagnostic of oblique impacts in well-preserved craters. Geomagnetic data reveal no buried individual impactor masses >100 kg and suggest that the total mass of the buried shrapnel >100 g is approximately 1050–1700 kg. Based on this mass value plus that of shrapnel >10 g identified earlier on the surface during systematic search, the new estimate of the minimum projectile mass is approximately 5 t.
Although small impact craters on Earth are very young (20 of 21 are younger than 1 Ma, and 15 of 21 are younger than 100 ka), their original features are often strongly modified or obliterated. This is primarily due to both the nature of the target rocks and climate conditions in the crater areas. In addition, some of these craters are partially or completely covered by vegetation (e.g., Whitecourt, Alberta, Canada; Morasko, Poland; Macha, Russia; Sobolev, Russia; Ilumetsä, Estonia; Sikhote Alin, Russia) or host ponds or small lakes (e.g., Kaalijärv, Estonia; Morasko; Macha).
The recently discovered 45 m diameter Kamil Crater in southern Egypt (Figs. 1 and 2) constitutes an exception, and represents a type structure for the study of small impact craters on Earth (Folco et al. 2010, 2011). The very young age of the crater (possibly <5000 yr; Folco et al. 2011), the ideal dry climate of the Sahara Desert, and the lithology of the target rocks are such that the impact structure has been preserved in a nearly pristine state. Kamil Crater preserves its original morphology, breccia lens, rayed ejecta pattern, various types of shock-metamorphosed materials and impact melt rocks, and a nearly intact assemblage of fragments of the iron meteorite impactor (D’Orazio et al. 2011; Folco et al. 2011).
During the first Italian–Egyptian expedition to Kamil Crater in February 2010 (six working days from February 19th to February 24th), we performed geological, differential global positioning system (DGPS), seismic refraction (SR), ground penetrating radar (GPR), and geomagnetic surveys, as well as the systematic collection of macroscopic meteorite specimens and meteorite microparticles. Field work was complemented by remote sensing analyses. In our discovery papers (Folco et al. 2010, 2011), we reported the circumstances of the find and the main results of our first geophysical investigation, highlighting the remarkable pristine state of preservation of the crater and addressing current models for meteorite–atmosphere interaction. We subsequently reported on the petrography and classification of the Gebel Kamil iron meteorite (namely, an ungrouped ataxite) that formed Kamil Crater, providing a detailed account of the systematic collection of macroscopic meteorite specimens. In particular, we distinguished between meteorite fragments generated upon impact with the ground, or shrapnel, versus individuals that detached from the main mass prior to impact, and showed their geographic distribution and its bearing on the impact trajectory and the estimated minimum impactor mass (D’Orazio et al. 2011).
In this article, we report the overall geology and detailed analyses of all the geophysical data collected at Kamil Crater during the February 2010 expedition. Largely expanding on the main results presented in Folco et al. (2010, 2011), we detail here (1) the geological setting and nature of the target rocks by integrating remote sensing data with the results of the geological and seismic refraction surveys; (2) the surface and subsurface crater structure and morphology by integrating information from remote sensing analyses, the geological survey, the digital terrain model (DTM) derived from the DGPS survey, and the high-frequency GPR survey; (3) the ejecta distribution pattern on the basis of remote sensing and field data; (4) the distribution of the buried impactor masses on the basis of geomagnetic data. The data set presented in this article allows us to improve the impact scenario discussed in our previous papers (Folco et al. 2010, 2011; D’Orazio et al. 2011), particularly in terms of impact trajectory, impact angle, cratering mechanism, and mass of the impactor. Furthermore, the structure of Kamil Crater defined in this work is characterized by a number of features that may be diagnostic for the identification of obliquity of the impact in terrestrial small, simple craters.
Remote Sensing Analysis
Satellite image analysis was performed on two datasets: a panchromatic QuickBird frame (15 × 12 km) acquired on 22 October 2005 with a nominal resolution of 0.6 m per pixel, and a radar COSMO-SkyMed frame (11 × 10 km) acquired on 13 March 2009 with a resolution of 1 m per pixel (Fig. S1). Data processing was carried out using the ENVI 4.8 software. The first data set was used to obtain general information on the geomorphology of the area, including details of the ejecta distribution. The second data set was used to obtain additional geomorphological data and to reconstruct a first-order DTM of the area. Owing to the low resolution of the radar frame, the fine topographic details of Kamil Crater and the surrounding area were defined through the DGPS survey described below. Both the panchromatic QuickBird frame and the radar COSMO-SkyMed frame were used to carefully inspect an area of approximately 220 km2 around Kamil Crater in the search of possible companion craters.
Differential Global Positioning System (DGPS) Survey
The main goal of the DGPS survey was the production of an accurate DTM of the Kamil Crater area so as to determine its preimpact topography and the surface morphometric parameters of the impact structure.
A suitable reference point was identified in the field, and its coordinates were recorded through a 6 h long acquisition on a three-point-net (one Trimble 4000SSI and two Trimble 5700 geodetic receivers were used). The achieved precision of the reference base station coordinates was less than 1 cm. Differential correction was transmitted at 1 Hz to the rovers (Trimble 5700) using one observation per second (real time kinematic). The rover coordinates were processed using the Trimble Geomatic Office software and referenced to the Ellipsoid using the UTM projection (Zone 35N, WGS 84). A total surface of about 0.48 km2 (800 × 600 m) centered on the crater was covered. Positions were acquired on a 1 m grid within an area about 50 m from the crater rim.
Seismic Refraction (SR) Survey
To gain information on the shallow subsurface lithology and structure of the target, an SR survey was carried out about 200 m due east of the crater in an area almost devoid of important ejecta (Figs. 1B and S2).
A 100 m long seismic profile was acquired using a 48 channel Geometrics Strataview seismograph. A shot consisting of several stacks of 15 kg sledgehammer blows was used as the seismic source. The geophone spacing was 2 m, and a total of five shots were acquired. The Seisimager software (Oyo Corp.) was used to process and model the seismic data based on delay time and ray tracing methods to determine the specific depth of each refractor beneath each geophone. The adopted method allowed investigation to a depth of about 15 m below the preimpact surface, i.e., just below the elevation of the transient crater floor (see below).
Ground Penetrating Radar (GPR) Survey
The GPR survey was carried out to obtain subsurface information on the structure and morphology of Kamil Crater. In particular, we aimed at defining morphometric parameters such as the crater floor, which is largely buried under a recent sand deposit, and the transient crater floor, which is completely buried under the breccia lens. Furthermore, we aimed at defining the mode and extent of bedrock deformation in the subsurface crater surroundings.
Unfortunately, because it was impossible to tow the GPR antennas coherently through the large blocks on the crater rim and at the top of the steep crater walls, the GPR survey was limited to within the crater and from the crater rim outward. Note that the area is off limits for military reasons and airborne surveys are strictly forbidden.
The crater floor and the transient crater floor were investigated through a number of NW–SE and NE–SW GPR profiles within the crater for a total length of 200 m using a GSSI Sir2000 unit equipped with 400 and 200 MHz monostatic antennas (Figs. 1B and S2). The extent of the deformed bedrock volume surrounding the crater was determined along ten radial profiles extending outward from the rim for a total length of approximately 300 m using the 400 MHz antenna (Figs. 1B and S2). The acquisition range (two-way-traveltime, TWT) varied from 150 ns for the 400 MHz antenna to 200–250 ns for the 200 MHz antenna, corresponding to approximately 8 m and approximately 10–12.5 m, respectively, considering an averaged electromagnetic wave speed of about 0.11 m ns−1. Due to the relatively small size (tens of meters) of the crater and the occurrence of a dry medium with good dielectric properties, the 400 and the 200 MHz antennas used in this work enabled the investigation of wide ranges with vertical interface resolutions of 0.2 and 0.4 m, respectively. All GPR profiles were normalized and filtered horizontally and vertically. The gain was adjusted and finally corrected for topography using the DGPS altitude. The electromagnetic wave speed required as an input parameter for the time to depth conversion was calculated by both hyperbola analysis and common mid point measures.
A geomagnetic survey covering an area of approximately 230 × 230 m centered on the impact crater was carried out after systematic collection of meteorite specimens larger than approximately 10 g present at the surface (see D’Orazio et al.  for details on the meteorite collection method and results). Due to the high magnetization of the Gebel Kamil iron meteorite (approximately 100 A m−1; Rochette, personal communication), the survey was expected to reveal magnetic anomalies related to buried fragments of the impactor, thereby enabling the identification of their geographic distribution and their total mass. The survey was performed using the rover unit GemSystem GSM 19 Overhauser scalar magnetometer and a GemSystem GSM19 proton precession magnetometer as base station for monitoring diurnal variations in the magnetic field. After clock synchronization between the two magnetometers, the base station and rover unit magnetic values were acquired simultaneously to remove the diurnal magnetic-field variations from rover measurements (time reduction correction). The spatial positions of magnetic stations around the crater were determined using the rover internal GPS board; in this configuration, a value of latitude and longitude is associated with each magnetic measurement during data acquisition.
Parallel magnetic profiles were acquired in a N–S direction following the “walking mode” survey procedure, in which the magnetometer unit acquires continuously at 2 Hz while the operator walks along profiles (about 1 measure every 0.5 m). Magnetic profiles were acquired along profiles spaced 1.5 m apart for a total of approximately 40 km in length and about 60,000 stations. The total surface covered around the crater was about 50,000 m2. The rover unit sensor was maintained about 0.20 m from the ground surface, whereas the base station sensor was kept about 2 m from the ground surface with a sampling rate of 1 measure every 30 s.
The survey within the crater was carried out using a “flying mag” configuration. We extended a cable across the crater from the rim crest along eight NNE–SSW profiles. The rover, running in continuous acquisition mode, was secured to the cable and moved from one side of the crater rim to the other using a pulley system. In this way, it was possible to move the unit while maintaining a constant elevation corresponding to that of the outer crater rim, thereby reducing the crater’s topographic magnetic effect. This configuration allowed us to check for the presence of magnetic signatures ascribable to a large buried impactor mass under the crater floor.
All measures were corrected by removing the International Geomagnetic Reference Field (IGRF 2010) global magnetic field, calculated for the latitude, longitude, and elevation of the site, and for the survey date. Residual data were gridded, and microleveling numerical corrections were applied to decorrugate the resulting color-scale map representing the intensities of magnetic anomalies.
To better discriminate the contribution of meteorite fragments to the magnetic signal, we first measured the magnetic susceptibility of the target rocks in several outcrops. Low values of around 10−4 SI were detected for the more ferruginous and coarse-grained sandstones, which constitute the deepest layers of the excavated target rock (see below). Their weak susceptibility produces a clear magnetic effect in such a remote and undisturbed area of the Egyptian Sahara. We subsequently measured the magnetic anomalies generated by meteorite specimens of different sizes using the same instrument configuration adopted during the survey of Kamil Crater. We found that a single meteorite fragment of 1400 g can generate a magnetic anomaly with a total amplitude in excess of 100 nT, and that only individual masses >100 g could be correctly detected.
Geological Setting and Target Rocks
The 45 m diameter Kamil Crater (22° 01′ 06′′ N, 26° 05′ 15′′ E; Figs. 1–4) is located in a rocky desert area in the East Uweinat district of southern Egypt, approximately 2 km due north of the Sudanese border and approximately 110 km due east of the Libyan border. This low-relief (<50 m) piedmont area is about 595 m above sea level and slopes gently southwestward. It lies at the base of the southeast end of some fault-aligned, NE–SW-elongated hills, and mesas as high as 750 m above sea level (Fig. 1A). Few shallow (<1 m deep) drainage lines, namely dry wadis, slope down to the southwest (Figs. 1B and 2A). The area is virtually devoid of vegetation.
Two main stratigraphic units crop out in the Kamil Crater area, as shown in the schematic block diagram of Fig. 5. The lower unit consists of a Precambrian crystalline basement belonging to the Nile Craton, which is one of the oldest African cratons. The Nile Craton is composed largely of migmatite, gneiss, amphibolite, charnockite, and granulite primarily of quartzofeldspathic composition, with intercalations of quartzite, marble, calc-silicate rocks, and amphibolites (Klerkx 1980; Schandelmeier et al. 1987, 1988). Based on U-Pb zircon ages of anorthosite samples from the Gebel Kamil (approximately 65 km due east-northeast of Kamil Crater), these rocks could be dated to 2.67 Ga, i.e., from Archean to Paleoproterozoic (Sultan et al. 1994). In the study area, the main foliation of the metamorphic rocks typically dips to the NW at high angles between 40 and 60°. The upper unit is the Gilf Kebir Formation (maximum thickness in the area: approximately 100 m), which consists of unmetamorphosed sandstones with minor siltstone intercalations and conglomerates. The Gilf Kebir Formation unconformably overlies the Precambrian Basement, with a peneplaned surface marking an enormous stratigraphic gap. A Late Jurassic to Early Cretaceous age is assigned to the Gilf Kebir Formation (Klitzsch et al. 1979). Paleozoic ages cannot be totally excluded for the lower part of the Gilf Kebir Formation, which forms outliers of dark tarnished siliceous sandstone directly overlying the basement. The undisturbed bedding of the Gilf Kebir Formation in the area ranges from subhorizontal to gently eastward-dipping. The two stratigraphic units are commonly traversed by silicified, mineralized fracture zones, which are occasionally associated with fault breccias and iron-manganese oxides. In the piedmont areas, the Precambrian and Mesozoic rocks are discontinuously covered by a m-thick layer of Quaternary colluvium locally mixed with eolian sand.
A NE–SW trending normal fault running about 100–150 m due northwest of the crater is inferred to separate a northwestern uplifted block from a southeastern downthrown block (Figs. 2B and 5). As a result, the rock outcrops northwest of the fault are dominated by crystalline basement rocks, whereas those southeast of the fault, where Kamil Crater is located, consist of layered sandstones with subhorizontal bedding belonging to the Gilf Kebir Formation. Here, the colluvium is typically less than a few decimeters thick, and no unequivocal evidence of the chronological relationship between the impact and the drainage system could be observed.
Figure 6A shows the rocks occurring at the crater walls from the rim to the crater floor at a depth of approximately 7 m below the average ground surface outside the crater structure (i.e., the preimpact surface with average elevation of approximately 594 m a.s.l.; see the Crater Morphology and Structure section and Fig. 9 for its definition), or 10 m below the average rim crest. They consist of layered, pale (top), and reddish brown (bottom), coarse to very coarse-grained, gritty, and ferruginous quartz arenites. The topmost layers are whitish, kaolinitic, and fine-grained. The base of the crater is filled by a breccia lens (see also Figs. 6B and 6C), preventing direct investigation of the lower layers of the target rocks; however, no evidence of lithologies other than sandstones was observed in the ejecta. A wind-blown sand deposit covers part of the northern crater wall and nearly two thirds of the crater floor (Fig. 6A).
Figure 7 shows sampling location and optical micrographs of representative samples of the pale and reddish brown quartz arenites from the NW crater wall. They vary mainly in terms of grain size, porosity, nature of cement, and degree of kaolinitization, which is higher at the top of the sequence. Samples from the crater walls show no evidence of shock deformation, unlike some samples of the ejected debris (Figs. 7D and 7E).
Figure 8 shows a representative seismic profile acquired approximately 200 m east of the crater along a NNW–SSE direction on a flat surface with average elevation of approximately 594 m a.s.l., i.e., the elevation of the preimpact surface. This seismic profile provides information on the preimpact subsurface stratigraphy of the target down to a depth of 15 m below the ground surface. The travel time of the received seismic waves is compatible with a model with a three-layer subhorizontal stratification (Fig. 8B). The first layer from the top of the sequence is approximately 2 m thick and characterized by low P-wave velocity of approximately 0.3 km s−1. This layer is ascribed to an unconsolidated, dry top cover. The second layer shows a P-wave velocity of approximately 2.1 km s−1, compatible with the layered sandstones, which crop out in the area and on the crater walls. The third layer is characterized by higher P-wave velocity of approximately 3.2 km s−1 down to a depth of 15 m. At a depth of about 7 m, the slight gradual variation in P-wave velocity between the second and the third layers is interpreted as a transition in sandstone stiffness. Note that a P-wave velocity of approximately 3.2 km s−1 is too low for crystalline basement rocks, even if severely weathered, such as those cropping out beyond the normal faults due NW of the crater (Figs. 2B and 5). Seismic data and the surface geology thus indicate that the first 15 m of the preimpact stratigraphic sequence consists of layered sandstones with subhorizontal bedding belonging to the Gilf Kebir Formation, topped by an approximately 2 m thick unconsolidated cover, as schematically shown in Fig. 5.
Crater Morphology and Structure
The Kamil Crater morphology and structure were determined by combining surface geological observations with the DTM derived from the DGPS survey (Fig. 9) and with subsurface data derived from the GPR survey (Fig. 10). For the definition of the crater morphometric parameters used in this article, the reader is referred to fig. 3 of Collins et al. (2005). Many morphometric parameters used in impact cratering are measured from the preimpact surface level. A model preimpact surface for the Kamil Crater area (Fig. 9B) was obtained by interpolating DTM data from the crater surroundings. The model preimpact surface is located at an elevation of approximately 594 m a.s.l. and slopes very gently southwestward. As mentioned before, the chronological relationship between the impact structure and the drainage system in the area is not unequivocal; however, had the two dry wadis due east and west of the crater formed after the impact, the preimpact surface at the very site of the impact would have had essentially the same elevation and morphology. The DTM of Kamil Crater relative to the preimpact surface is shown in Fig. 9C. Figure 9D shows the same model after removing the eolian sand deposit covering part of the northern wall (whose base has been defined through GPR investigation, see below), whereas Fig. 9E shows the SW–NE and NW–SE cross-sections of this model.
No evidence of important erosional features is observed in the crater and crater rim. Figure 9 shows that Kamil Crater is essentially circular, although a somewhat polygonal outline of the crater rim crest is observed in places. The mean rim-to-rim diameter, Dfr, is 45 ± 2 m. Kamil Crater has a bowl shape with an upraised rim. A 1 m deep pit dug at the crater floor into the breccia lens shows that its topmost layer (<1 m) mainly consists of fallback material, namely laminated deposits of powdered sandstone (Fig. 6B) with abundant microscopic impact melt glass particles (Fig. 6C). About 50 cm below the surface, frequent blocks tens of centimeters in size were found embedded in the powdered sandstone. The crater depth measured from the crater floor to the average rim, dfr, is approximately 10 m. The excavation depth measured from the preimpact surface level to the crater floor is 7 m. The excavated volume, i.e., the volume of the crater between the top of the breccia lens and the preimpact surface, is approximately 3800 m3. After removing the eolian sand deposit that covers part of the northern wall of the crater, the DTM shows no significant steepening of the crater walls in any direction (Fig. 9E). The crater rim height above the model preimpact surface, hfr, is variable with a mean value of 3 ± 0.7 m. The rim of the whole SE sector of the crater is consistently depressed relative to the average rim height (Figs. 3B and 11). In this sector, we also observe an approximately 2.5 m southeastward (approximately 130°) offset of the rim relative to the best fit circle (Fig. 11).
Figure 4 shows a panoramic view of the crater walls. The target rocks are in place along most of the higher portions of the walls, and the bedding of the autochthonous rim strata can be observed. Seven sectors with contrasting bedding are separated by subvertical tear faults arranged in a radial pattern. In the S and SW sectors and the opposite N and NE sectors, the bedding is mostly upturned with a radial dip up to 40°. In the WNW sector, the bedding is dominantly overturned. The bedding in the opposite E and SE sectors is dominantly vertical and overturned, respectively. The two sectors with overturned bedding are roughly aligned along a 135° axis (Fig. 4C).
The basal portions of the crater walls that are not buried by the eolian sand deposit are covered by a <1.5 m thick layer of centripetal debris slump from the relative higher portions of the crater walls (Figs. 4 and 6A). This consists of sandstone blocks with a maximum size of 1 m. As a result, the slope of the crater walls gradually decreases toward the crater center.
Figure 10A shows a representative GPR section across Kamil Crater. It represents three combined radargrams obtained from GPR profiles carried out using a 400 MHz frequency antenna: NW of the crater (PR4-6), inside the crater (PR4-7), and on the SE side of the crater (PR4-1) in a NW–SE direction (see Fig. 1 and S2 for their locations). The inside crater profile (PR4-7) was conducted by towing the GPR antenna from wall to wall across the eolian sand deposit covering part of the northern crater wall and the crater floor. Figure 10B shows a radargram carried out at 200 MHz frequency (PR2-7) containing details of the crater structure along the PR4-7 profile. The analysis of the PR2-7 radargram (Fig. 10C) reveals the occurrence of two main discontinuity surfaces separating three layers with contrasting dielectric properties. The shallowest discontinuity (dashed black line) is a continuous concave surface, which separates a nearly homogeneous, yet finely stratified, less than 2 m thick material at the top (EM wave speed 0.15 m ns−1) from an inhomogeneous material (EM wave speed 0.11 m ns−1). The latter is dominated by wave diffraction, mostly in the central part, i.e., from −10 to 10 m from crater center, and is characterized by several faint, low-energy discontinuities (dashed gray lines). The deepest observable discontinuity (solid thick line) is a smooth, fairly continuous high-energy reflector, which slopes gently southeastward at a depth of between 5.8 and 9 m with an average inclination of approximately 12%. The absence of strong reflections or diffractions below this surface indicates the presence of a fairly homogeneous material. Coupled with field observations, GPR data are interpreted as follows: The top layer is the eolian sand deposit covering part of the northern crater wall and the crater floor. The layer below is the breccia lens. This consists of a mixture of stratified fallback debris at the top comprising about 25% of the section and displaced sandstone blocks at the bottom. The latter is most likely slump debris from the crater walls. The deepest discontinuity (thick solid line) represents the transient crater floor separating the in situ bedrock sandstones. The main discontinuities were tracked in all in-crater GPR profiles and their depths added, as surfaces, to the DTM database to build the crater structure more accurately. In particular, the 3-D data confirm that the very bottom of the transient crater floor is located off-center, in the southeast sector of the crater; that the maximum thickness of the breccia lens, tbr, is about 6 m; and that the volumetric ratio between fallback and slump debris in the breccia lens is about 1/4–1/5.
On the basis of morphological data so far gathered, we can estimate or constrain some parameters of the transient crater. The outline of the transient crater is similar to that of the crater rim (Fig. 9). The transient crater depth (i.e., the depth of the transient crater floor below the preimpact surface), dtc, is 13 m. The transient crater diameter, Dtc, is constrained to less than approximately 35 ± 2 m (measured at 12° intervals along the intersection between the crater and the preimpact surface), and the transient crater volume is constrained to less than approximately 5300 m3. As small-scale impact craters do not undergo considerable modification after transient crater excavation (i.e., there is some wall slumping, but no major transient crater collapse during the modification stage; e.g., Grieve 2001) as also observed at Kamil Crater, the above maximum values for Dtc and transient crater volume should be close to the actual ones.
The radargram SE of the crater (PR4-1) reported in Fig. 10A shows the lateral extent of shock deformation within the target. The radargram starts about 50 m from the crater center and ends after approximately 24 m, close to the SE crater rim crest. At distances >35 m from the crater center, the radargram shows a sharp horizontal contact between two different electromagnetic layers at a depth of approximately 2 m. The lower layer shows electromagnetic properties of a homogeneous medium similar to that observed below the transient crater floor. The upper layer shows similar electromagnetic properties typical of a more heterogeneous medium. This electromagnetic stratigraphy is coherent with the seismic refraction stratigraphy obtained about 100 m due E of the crater rim (Fig. 8). We therefore interpret the top layer as an unconsolidated cover and the bottom layer as a relatively undisturbed bedded sandstone layer. Moving toward the crater, at distances <35 m from the crater center, this contact disappears and a wedge-shaped zone of strong diffraction hyperbolas, which thickens and deepens toward the crater rim, is observed. This wedge-shaped zone is interpreted as evidence of increasing brecciation and displacement of the layered sandstones toward the crater walls. The radargram NW of the crater essentially shows a reflection of the features observed on the SE side of the crater. The brecciated zone was observed in all eleven radial GPR profiles (Fig. S2) acquired outside the crater. Its termination was always detected at a maximum depth of about 2 m and up to 12 m from the crater rim, thereby defining a circular aureole of brecciation around the crater probably associated with the upturning of the sandstone layers during the formation of the crater rim. Note that the starting point of brecciation almost coincides with the surface break in slope in most of the radial profiles (Fig. 10A). This relationship provides further evidence of the freshness of the impact structure, considering that eolian sedimentation tends to smooth out surface morphologies.
Ejecta Deposits and Their Distribution
The bulk of the bedrock debris ejected from the impact crater and deposited outside the crater (Fig. 2) consists of sandstone fragments ranging from large boulders, with a mass of up to 4 t, to dust (Fig. 12). Ejecta distributed within one crater diameter comprise debris from both the pale and reddish quartz arenites, whereas that thrown further away comprise debris composed almost exclusively of pale quartz arenites, i.e., debris from the topmost sandstone layers observed along the crater walls (Figs. 6A and 7A). As a result, the dominant color of the ejecta is pale and contrasts with the darker color of the weathered original ground surface, allowing it to be quite easily detected both in the field and in satellite images (Figs. 2A and 3). The petrographic analyses of sectioned samples from three ejected blocks show shock deformation features related to different degrees of shock metamorphism, from negligible to high. The latter is documented by the association of planar features (PF), mosaicism, and planar deformation features (PDF) in quartz grains (Figs. 7D and 7E). According to Langenhorst and Deutsch (1994), PDF develop at shock pressures ranging from 10–15 to 35 GPa.
The satellite image featured in Fig. 2A shows the radial, locally discontinuous and patchy distribution of ejecta. The image shows that water runoff has not substantially modified the general pattern of ejecta distribution, as the overall pattern is continuous across the local drainage system. Figure 2B shows a sketch map of the ejecta field obtained from satellite image analyses and ground checks. The edge of the ejecta field is somewhat arbitrary due to the discontinuity of the ejecta deposit, and limits only the areas covered by conspicuous debris that appear bright in the satellite image (i.e., strewn fields of scattered blocks that may occur beyond the mapped edge are not included). Alignments of boulders define the crests of the ejecta rays (Fig. 12A). Secondary pits generated by the impact of relatively large blocks ejected from the crater are also shown (Figs. 2B and 12C).
The radial pattern of ejecta distribution is evidenced by the occurrence of three nearly straight major ejecta rays trending to the N (approximately 355°), SE (approximately 130°), and SW (approximately 210°). The first extends for 170 m from the crater rim, the other two are approximately 300 m long. The ray trending to the southeast is also the largest one in terms of the volume of ejected material and gives rise to a positive topography close to the crater rim (Figs. 1B, 9D, and 12A).
Field observations reveal that a nearly continuous veneer of ejecta, namely the ejecta blanket, consisting of a chaotic mixture of boulders, blocks, and fine-grained material, extends radially for approximately 50 m from the crater rim in a northward, eastward, and southward direction. On the western side of the crater, however, this veneer is more discontinuous and mainly consists of large boulders and blocks. Due to the widespread occurrence of patches of eolian sand, it is somewhat difficult to define the actual limit of this continuous veneer; however, at distances greater than approximately 50 m from the crater rim, the amount of fine-grained ejecta material decreases rapidly in all directions, with trails of blocks dominating the bulk of the ejecta deposit toward the edge of the ejecta field.
The distribution of ejecta to the NE of the crater is very patchy (Fig. 2A). Here, particularly within the 45–75° sector, chains and clusters of shallow (typically less than 20 cm deep) secondary pits up to some meters in diameter and a few decimeters deep (Figs. 2B and 12C) occur up to 250 m from the crater rim. They usually show streaks of secondary ejecta downrange, i.e., in NE direction. Fragments of the parent impactor, typically less than 10 cm across, are usually found along the rim of the pit with some scattered on the floor of the pit. Although less abundant, secondary pits are also frequently observed along the southeastern major ray (Fig. 2B).
Lastly, note that the bulk of the ejecta material is preferentially concentrated within the sector from 355 to 210°, i.e., between the main northern and the southwestern ejecta rays. According to the map in Fig. 2B, the ejecta field in this sector extends for approximately 40,000 m2, corresponding to approximately 88% of the extension of the entire ejecta field. The western sector of the ejecta field is virtually devoid of conspicuous ejecta rays (i.e., those extending more than approximately 50 m from the crater rim), although discontinuous alignments of scattered ejected stones are observed up to a distance of 200 m from the crater rim in an approximately 280° direction. The latter is nearly parallel to the bisector of the angle formed by the major northern and southwestern rays and opposite to the direction of the largest, southeastern ejecta ray.
Geomagnetic Anomaly Distribution Map
Figures 13A and 13B show the magnetic anomaly map of the impact crater area and its interpretation, respectively. Note that the geomagnetic survey was conducted after the systematic collection of meteorite fragments typically >10 g in mass (D’Orazio et al. 2011) to gain information on the distribution and mass of the buried impactor fragments in and around the crater.
Magnetic anomalies in Fig. 13A can be divided into three main groups; (1) dipolar anomalies (mainly) in the crater surroundings that are related to tectonic and geomorphological structures; (2) strong, grouped dipolar anomalies outside the crater that are related to the crater structure; and (3) the magnetic anomaly in the crater.
Interpretation of the first group of anomalies (i.e., “a” in Fig. 13B) is well constrained because they correlate with structural features occurring in the crater surroundings. Due to the close distance between the magnetometer sensor and the ground surface, the topographic effect of the weakly magnetized ground is discriminated. For example, the magnetic lineaments due NW of the crater are related to the mineralized fractures and faults mentioned above.
The strong magnetic anomalies of the second group (i.e., “b” in Fig. 13B) are dipolar anomalies distributed over an area of 11,996 m2 around most of the crater up to a distance of approximately 100 m from the rim. There is an obvious gap in the distribution of these anomalies W of the crater between 240 and 290°. Arrays of strong dipolar anomalies are sometimes arranged in a radial pattern. A major magnetic ray extends southeastward from the crater rim (MR1, Fig. 13B) in an approximately 130° direction parallel to the main ejecta ray (Fig. 2). There are two minor rays in the southern crater rim area (MR2 and MR3, azimuth approximately 180 and 220°, respectively). The area north of the crater shows only diffuse dipolar anomalies without an obvious radial pattern. Another area of strong magnetic anomalies due north of the crater, identified by a question mark in Fig. 13B, is possibly correlated with the northward-oriented ejecta ray (Fig. 2); as it extends beyond the edge of the present survey, we cannot give a definitive interpretation. On the whole, the radial distribution of the strong dipolar anomalies parallels the radial distribution of the ejecta (Fig. 2). The intensity of these strong dipolar anomalies and their distribution indicate that their most probable source is the highly magnetized meteorite debris generated during the impact and buried within the ejecta, namely fragments >100 g. Other alignments of strong magnetic anomalies are observed along topographic features such as little depressions and/or reliefs associated with local geological and geomorphological structures. We suggest that topographic magnetic effects are here amplified by the trapping and accumulation of shrapnel during their ejection and subsequent deposition. Note that the extension of the strong dipolar magnetic anomalies due N, E, and S, associated with meteorite fragments >100 g buried within the proximal ejecta, documents the extension of a consistent ejecta blanket. The third group of magnetic anomalies occurs inside the crater. Figure 13A shows that craters formed in weakly magnetized rocks generate a topographic annular dipolar magnetic anomaly (i.e., “c” in Fig. 13B). This develops on the inner side of the crater rim, with a minimum on the southern side and a maximum on the northern side. A magnetic minimum is located in the SE sector of the crater wall (i.e., “d” in Fig. 13B), close to the start of the main magnetic ray MR1. This magnetic minimum suggests the absence of an important buried portion of the impactor: the latter would produce a different magnetic dipolar anomaly, with positive values being dominant. The observed magnetic minimum is most likely the result of the combination of the topographic effect of the transient crater floor geometry, which slopes southeastwards (based on GPR data; Fig. 10), and the magnetic effect of the high density of highly magnetized material (namely buried shrapnel) within the main MR1 magnetic ray.
To estimate the total mass of meteorite debris that generated the geomagnetic signal described above (i.e., the meteorite debris consisting of fragments >100 g still present in the field after systematic collection), we followed the Gerovskaa et al. (2004) method for the localization of buried steel objects, namely a generalization of the unexploded ordnance (UXO) detection method. Our estimate is based on the field calibration for the generation of the magnetic signal by meteorite fragments of variable masses described in the Methods section and takes into consideration the possible spatial undersampling of the magnetic anomalies due to the distance between profiles.
The main magnetic anomaly source positions were identified, and the dipole magnetic moment of each source was calculated as a function of the distance between magnetometer sensor and meteorite fragments. Once the meteorite volume magnetization was determined (see the Methods section), it was possible to convert the cumulative dipole magnetic moment into cumulative masses of meteorite fragments. Figure 14 shows the calculated total mass of fragments as a function of the mean distance between the magnetometer sensor and fragments. The calculation was made for mean sensor–source distances of up to 2 m. To estimate the best mean magnetic sensor–source distance for the present survey, the Kolmogorov–Smirnov (K-S) nonparametric statistical test was applied to determine which calculated fragment mass distribution best approximates that of the collected fragments during systematic search (D’Orazio et al. 2011). Following the hypothesis that fragments generating magnetic anomalies were produced by the same physical process (fragmentation upon impact) of the collected ones, the test evaluates the distance between the empirical distribution of two different data sets, considering the statistical null hypothesis that the two sets come from the same distribution. The test was performed on the calculated mass distribution, considering a sensor-to-fragments distance of 0.3 to 1.0 m and a significance level of 5%. The K-S test failed for mean distances of 0.3, 0.4, 0.5, 0.8, 0.9, and 1.0 m, suggesting that the null hypothesis must be rejected because the calculated mass distribution of fragments differed considerably from the observed one. The test is positive for mean distances of 0.6 and 0.7 m, with a calculated probability of about 70% and 50%, respectively. Given that the mean distance between the magnetic sensor and the ground surface was about 0.2 m, these results suggest that the mean thickness of the ejecta blanket where magnetic fragments are embedded is less than 0.5 m. This value is comparable to the observed thickness of the ejecta blanket around the 36 m diameter Whitecourt meteorite crater in Canada (Kofman et al. 2010), which shares many geometric characteristics with Kamil Crater. Our best-fit model suggests that the total mass of meteorite fragments >100 g buried in the ejecta is between 1050 and 1700 kg assuming mean sensor–source distances of 0.6 and 0.7 m, respectively.
Remote Sensing Survey: Seeking Companion Craters
Remote sensing analyses of both the panchromatic QuickBird frame and the radar COSMO-SkyMed frame revealed that no further meteorite impact craters occur in an area of approximately 220 km2 around Kamil Crater. An extended Google Earth survey likewise confirmed that Kamil Crater is a single impact crater. The numerous nearly circular structures detected in the area (e.g., 22° 00′ 41′′ N, 26° 13′ 01′′ E) do not show any of the fresh structural features of Kamil Crater described above. For instance, many are eroded, positive structures rather than pristine bowl-shaped depressions. Some are elliptical. Others have incomplete rims. All of them lack bright ejecta. These structures are most likely due to endogenic geological processes, i.e., hydrothermal vents associated with an extended geothermal field activated after the deposition of a sedimentary cover in the Gilf Kebir–Jebel Uweinat region (Orti et al. 2008).
A Model Crater Structure for Small-Scale Hypervelocity Impacts on Earth
The 45 m diameter Kamil Crater generated by the impact of the Gebel Kamil iron meteorite occurs in a simple geological and geomorphological context. The surface of the target is nearly flat and devoid of vegetation (Figs. 1–3). The target rocks are relatively homogeneous with a simple structure consisting of layered sandstones with subhorizontal bedding, topped by an approximately 2 m thick unconsolidated layer comprising altered sandstones, colluvium, and eolian sand (Fig. 5). With the exception of an eolian sand deposit, which covers part of the northern crater wall and about two thirds of the crater floor, the crater structure is very fresh with no evidence of significant alteration due to geomorphic agents and human activity. This is exemplified—for instance—by the lack of erosion features on the crater rim (Fig. 4), and the essentially pristine distribution of bright ejecta (e.g., Fig. 2) and meteorite fragments (D’Orazio et al. 2011). Furthermore, the occurrence of shock metamorphism in the target rocks (e.g., PDFs in shocked quartz; Fig. 7E), shock melting of both the projectile and the target (Fig. 6C), and catastrophic disruption of the impactor into thousands (or many more) of fragments (D’Orazio et al. 2011; Folco et al. 2011) document that Kamil Crater is a crater produced by a hypervelocity impact rather than a low-velocity penetration crater. Therefore, Kamil Crater represents a model structure for small-scale hypervelocity impact craters on Earth.
A schematic NW–SE cross-section of Kamil Crater featuring its main structural features according to our geophysical survey is shown in Fig. 15. Beside general information, the main morphometric parameters and structural features that describe Kamil Crater are listed in Table 1. Overall, Kamil Crater has a bowl shape and raised rim typical of simple craters (Dence 1965). The classic example for a simple crater in terrestrial environments is the 1.2 km diameter Meteor or Barringer Crater, Arizona, USA (Shoemaker 1963). This crater type is characteristic of small impact craters below about 2–4 km in diameter on Earth, according to the target lithology (i.e., sedimentary or crystalline, respectively; e.g., Melosh 1989).
Table 1. Summary table listing the main features of Kamil Crater.
The morphometric parameters of Kamil Crater (Table 1) defined through our geophysical survey are in reasonable agreement with the expected values for simple craters, thereby documenting that Kamil Crater has a standard simple crater morphology. Fresh simple craters on all planetary bodies have similar morphologies with dfr/Dfr typically between 0.2 and 0.33 (e.g., Melosh 1989), and the dfr/Dfr approximately 0.22 for Kamil Crater falls into this range. Similarly, the Dtc/Dfr ≤ 0.78, the dtc/Dtc ≥ 0.37, the tbr/dfr approximately 0.6 and the dfr/Dfr approximately 0.22 for Kamil Crater are consistent within error with the Dtc/Dfr = 0.84, the dtc/Dtc approximately 0.37, tbr/dfr approximately 0.5 and the dfr/Dfr = 0.2 predicted by volume conservation models (Melosh 1989). The excavated depth is approximately 40% higher than the canonical value of approximately 1/3dtc (e.g., Melosh 1989, 2011) indicating a considerable proportion of ejected material. The hfr = 3 ± 0.7 m for Kamil Crater is higher than the 0.04Dfr value predicted by the model and observed in fresh simple craters less that 15 km in diameter on the Moon (Pike 1977). Note, however, that hfr at Kamil Crater shows variable azimuthal values, possibly connected with physical and structural properties of the target, as well as with the impact angle, as discussed below. Furthermore, the extent of the ejecta blanket around the crater in the northern, eastern, and southern sectors inferred from the geomagnetic survey (Fig. 13) is about two times the crater radius as observed in lunar craters with diameters between 1.3 and 436 km (Moore et al. 1974).
Notably, however, Kamil Crater shows some interesting features that are not observed in similar terrestrial craters. The transient crater floor gently slopes southeastward with an inclination of ca 12%. The crater breccia lens comprises chaotic superposition of slump debris layers at the bottom and laminated fallback material at the top (Fig. 6B) with an overall volumetric ratio of about 4/1. The brecciated bedrock zone under the transient crater floor extends laterally under the crater rim. The shape of this pattern is qualitatively similar to the shape of the zone of the maximum total plastic strain of the bedrock in numerical simulations (e.g., Collins et al. 2012), to the reconstructed transient crater structure of the Brent Crater (Ontario, Canada, 3.8 km in diameter; Dence 1965), as well as to the fracture distribution, particularly that of the near surface fractures, in experimentally produced craters (e.g., Polanskey and Ahrens 1990).
Complete and internally consistent sets of morphometric parameters are available for a limited number of terrestrial simple craters including Barringer, Brent (Ontario, Canada), Lonar (India), West Hawk (Manitoba, Canada), Aouelloul (Mauritania), Tenoumer (Mauritania), and Wolfe Creek (Australia), ranging in diameter from 390 m (Aouelloul) to 2.44 km (West Hawk). These data sets allow definition of a number of empirical relationships between morphometric parameters like diameter, depth, rim-height, thickness of breccia lens, etc. that are of fundamental importance for constraining impact process and models (e.g., Grieve and Therriault 2004). For small, namely decametric-scale, craters like Kamil Crater, data are available only for the two recently studied Whitecourt (Kofman et al. 2010) and Carancas craters (14.2 m in diameter; Perù; Tancredi et al. 2009; Kenkmann et al. 2009). However, the former is a rather eroded crater and its geometries are complicated by the fact that impact occurred onto a slope, whereas for the latter, it is not conclusively demonstrated that the crater was produced by hypervelocity (Tancredi et al. 2009) or low-velocity (Kenkmann et al. 2011) impact. Therefore, the complete data set of morphometric parameters of Kamil Crater (Fig. 15 and Table 1) is invaluable for future refinements of the empirical relationships between morphometric parameters for simple hypervelocity impact craters in the terrestrial environment as well as for impact models, particularly at the small-diameter end.
The structural data of Kamil Crater presented above allow us to further constrain the impact scenario discussed in our previous papers (Folco et al. 2010, 2011; D’Orazio et al. 2011), particularly in terms of projectile–atmosphere interaction, impact trajectory, cratering mechanism, and mass of the impactor.
One of the reasons for the low number of small impact craters on Earth is that small asteroids may undergo catastrophic disruption during impact with the Earth’s atmosphere (e.g., Passey and Melosh 1980). Current atmosphere–bolide interaction models (e.g., Bland and Artemieva 2006) for iron and stony objects (approximately 4.5% and 95% of the modern flux) with initial diameters up to approximately 1 km predict that Earth’s atmosphere plays a crucial role in filtering out small impactors that strike Earth’s surface with a considerable fraction of their initial cosmic velocities. In particular, weaker stony impactors with entry masses up to 108 kg are catastrophically disrupted and fragments are significantly decelerated almost to terminal velocity producing strewn fields, whereas stronger iron impactors up to 106 kg undergo fragmentation, separation, and further ablation, yet, some fragments still retain part of their cosmic velocity at the time of impact with Earth’s surface, thereby producing hypervelocity impact craters up to 300–400 m in diameter (e.g., Bland and Artemieva 2006). This general model is supported by the fact that the majority of the few small craters (<300 m in diameter) on Earth are multiple craters produced by iron meteorites. Violation of the model can be ascribed to significant deviations from the input parameters, most notably, entry angle and velocity as in the case of the recently witnessed Carancas impact cratering event produced by a small stony meteoroid (Kenkmann et al. 2009).
Folco et al. (2010, 2011) and D’Orazio et al. (2011) argued that Gebel Kamil, the iron meteorite that formed Kamil Crater, did not undergo important aerodynamic break up and fragment dispersion in the atmosphere, and that it exploded upon hypervelocity impact with the ground. This argument is based on the number and mass of specimen types recovered (or identified) in the field during systematic searches, namely one single regmaglypted specimen of 83 kg with no associated crater or penetration pit, and thousands of shrapnel pieces devoid of signatures of ablative flight totaling 1700 kg (for details see D’Orazio et al. 2011).
That Gebel Kamil escaped considerable fragmentation during impact with the atmosphere and lateral dispersion of the fragments prior to impact with the ground is also supported by our remote sensing analysis. The QuickBird and radar COSMO-SkyMed satellite images (plus an extended Google Earth survey) studied in this work cover an area >220 km2 around Kamil Crater and reveal that Kamil Crater is a single crater and not part of a crater field. However, single craters can also be produced by clustered impactors.
Laboratory-scale craters produced by clustered impactors (impact angles from vertical to 45°) show that, although characterized by the general crater morphologies with raised rims and exterior ejecta blankets, the craters are anomalously shallow with nearly flat floors (Schultz and Gault 1985). The 1.2 km diameter Barringer Crater could be a natural example of single crater produced by the hypervelocity impact of a cluster of iron meteorite fragments according to numerical modeling by Artemieva and Pierazzo (2009, 2011). Based on the morphometric parameters reported by Roddy et al. (1975) and Roddy (1978) on one side, and Grieve and Garvin (1984) on the other, Barringer Crater has a dtc/Dtc of 0.28 and 0.32, respectively. Kamil Crater is characterized by a higher dtc/Dtc ≥ 0.37 (Table 1) and a concave, yet asymmetric, transient crater floor. As such, morphometric parameters obtained in this work suggest that Kamil Crater was formed by the impact of a single mass or a very tight cluster of impactor fragments. Note also that first results from the MEMIN laboratory experiments on sandstone targets (Kenkmann et al. 2011) show that the crater morphologies differ substantially from wet (namely, pore space of sandstones filled with water) to dry targets, i.e., craters in wet targets are larger, but shallower. The high dtc/Dtc ≥ 0.37 measured at Kamil Crater could at least be partly due to the dry condition of the target at the time of impact.
In summary, Gebel Kamil iron meteorite deviated at least in part from the hitherto existing general model (Bland and Artemieva 2006) for fragmentation, lateral separation, and deceleration of small iron impactors. Therefore, modeling the Kamil Crater impact event requires nonaverage atmospheric entry conditions, and/or nonaverage impactor strength input parameters as suggested by Folco et al. (2011).
Impact Direction and Angle
On planetary bodies, the majority of the impact craters are produced by oblique impacts. Statistics predict that approximately 77% of impacts occur at angles of incidence between 20 and 70° from the horizontal, highly oblique impacts with angles <15° have a probability of occurring of approximately 7%, whereas the probability of vertical and grazing impacts is negligible (see also Pierazzo and Melosh 2000a). Observations, experiments, and modeling show that crater shape and structure, ejecta distribution, and impactor debris distribution provide clues on the impact direction and angle.
In the case of small impact craters like Kamil Crater, part of the projectile usually survives the impact with the ground, in the form of shrapnel and impact melt spherules. Models for high-velocity oblique impacts with impact angles up to at least 45° predict that the projectile deforms into a downrange plume composed of shrapnel and impact melt spherules (O’Keefe and Ahrens 1985; Pierazzo and Melosh 2000b). This scenario is also predicted by recent numerical models for Barringer Crater adopting incident angles of 45° (Artemieva and Pierazzo, 2011). Asymmetric distribution of the impactor debris around the crater is thus an indicator of the incident direction of the projectile in oblique impacts.
Folco et al. (2010) and D’Orazio et al. (2011) showed that the distribution of the Gebel Kamil shrapnel is asymmetric around Kamil Crater. The shrapnel identified in the field during systematic visual search, namely the surface specimens, is concentrated due SE with a maximum in a sector between the 125 and 160°. On the basis of this evidence, they suggested that Kamil Crater was produced by an oblique impact from the NW. The magnetic anomaly map discussed in detail in this article (Fig. 13) is consistent with the above argument. Buried shrapnel >100 g is concentrated around the crater rim with an obvious gap in the western sector, and the largest magnetic ray, i.e., ray #R1 in Fig. 13B, extends for about 100 m from the rim crest in an approximately 125° direction.
The ejecta distribution at Kamil Crater is characterized by a preferential concentration (88% of total extension of the ejecta field) of debris between the northern (355°) and the southwestern (210°) ejecta rays (Fig. 2). Following experimental work by Gault and Wedekind (1978), such asymmetric distribution suggests an oblique impact from the NW, namely an incident trajectory of approximately 290° and an impact angle of approximately 45° or lower. The radial distribution of the three main ejecta rays (bearing: 355, 125 and 210°) deviates slightly from a symmetric three-fold axial distribution. The largest angular gap of 145° occurs between the southwestern (210°) and northern (355°) ray. Thus, one may speculate that this radial sector with little bright ejecta may represent an incipient and ill-defined uprange forbidden zone. This interpretation would imply an impact direction of 290° and impact angle close to 30° according to Gault and Wedekind (1978). Note that the southeastern ejecta ray (125°) is the largest in terms of extension (approximately 300 m from the crater rim) and volume, and gives rise to a positive morphology close to the crater rim (Figs. 2 and 9D). Experiments (e.g., Schultz 1992; Anderson and Schultz 2006) related to the ejecta distribution generated by oblique impacts show a focusing of high-velocity ejecta in the downrange direction for a 30° impact, thereby suggesting a 305° direction of incidence for the impactor that generated Kamil Crater. Note also that hypervelocity laboratory experiments are performed using spherical projectiles and homogeneous targets. We speculate that the patchy distribution of the ejecta at Kamil Crater and the lack of an obvious bilateral symmetry could be in part due to some heterogeneity of the target and/or irregular shape of the projectile. Finally, recent numerical modeling (e.g., Shuvalov 2011) has shown that, at the same impact angle, the asymmetry in the ejecta distribution in oblique impacts is more pronounced in real craters on planets than in laboratory experiments, and around larger craters (i.e., 10 km in diameter) than around smaller craters (i.e., 100 m in diameter). Therefore, the impact angle for Kamil Crater could be somewhat higher than that inferred through comparison with laboratory experiments.
The experimental work by Gault and Wedekind (1978) also documents that crater shape is informative of the impact direction and angle only for impact angles <30°, with impact velocity and properties of the target and projectile playing a role in shaping the crater. For impact angles >30°, impacts produce circular craters without obvious asymmetries in the crater wall–crater rim profiles (Gault and Wedekind 1978). This is explained by the loss of asymmetries produced in the early stages of the crater excavation through crater growth (e.g., Collins et al. 2012). For impact angles <30°, craters first tend to be elongated along the impact direction, with steeper crater walls on the uprange side of the crater, then at angles <10° craters become markedly elliptical (Gault and Wedekind 1978).
We have shown that Kamil Crater is essentially circular (Figs. 2 and 9). Furthermore, there is no evidence of significant steepening of the crater walls in any direction within the error in the DGPS data discussed above (Figs. 9 and 15). Following experimental work by Gault and Wedekind (1978), this suggests impact angles >30° for Kamil Crater.
In summary, crater shape, ejecta distribution, and shrapnel distribution provide coherent information that Kamil Crater was generated by a moderately oblique impact, from a northwestern direction. The most likely direction of incidence lies between 305 and 340°, and is given by the shrapnel distribution around the crater (see also D’Orazio et al. 2011; Folco et al. 2011). The impact angle was between approximately 30° and approximately 45°, as deduced from the overall circularity of the crater and the downrange concentration of the ejecta.
Finally, Kamil Crater indeed shows some minor morphological asymmetries. They include the southeastward off-center position of the very bottom of the transient crater floor (Figs. 10 and 15), the southeastward offset, and lower elevation of the southeastern crater rim crest (Fig. 11), and the preferential occurrence of overturned rim strata along the NW and SE sectors of the crater rim (Fig. 4). All these features are coherent in terms of orientation with the incident trajectory inferred from meteorite and ejecta distributions discussed above. As such, we suggest that they are three effects of a pronounced downrange excavation flow component generated during the oblique impact from the NW. Perhaps these field data could be useful to test peak shock pressure distribution and flow-field migration models for oblique impacts (e.g., Pierazzo and Melosh 2000b; Anderson and Schultz 2006); in particular, the fact that overturning of bedrock layers is preferentially observed not only downrange but also uprange (Fig. 4) needs explanation. Although the present, relatively fine-scale observations require an assessment of the possible structural control of the target rocks (see below) as well as of the shape of the projectile, we have listed these three asymmetric morphological features in Table 2 as potential indicators of obliquity for small-scale (<300 m in diameter) terrestrial impact craters. The fact that these features are obvious only at Kamil Crater and are not reported for other terrestrial small-scale impact craters could be in part due to the exceptional pristine state of preservation of Kamil Crater structure. Note, however, that Kofman et al. (2010) do report that the bottom of the transient crater in the Whitecourt impact crater is slightly off-center downrange.
Table 2. Diagnostic features of oblique impact trajectories (i.e., neither vertical, nor highly oblique <10–15°) for terrestrial small-scale simple craters inferred from the study of the 45 m diameter Kamil Crater impact structure.
Downrange displacement of the very bottom of transient crater floor
Maximum overturning of target rocks along impact direction
Downrange lower crater rim crest elevation
Downrange concentration of impactor debris (shrapnel)
As shown in Fig. 9 Kamil Crater is essentially circular. In places, however, it shows a somewhat polygonal outline. Some segments of the polygonal outline appear to be limited by tear faults along which differential uplift of the crater rim strata is observed (Fig. 6). In spite of the limited structural data available, we speculate that the excavation of Kamil Crater and some details of its final shape were at least in part controlled by pre-existing structures in the target. It is known that planes of weakness in the target rocks like joints, fractures, and faults play a role in affecting the excavation and final crater morphology. One of the most striking and classical examples of such structural control is given by the Barringer Crater. Its outline is more square than circular due to the occurrence of two orthogonal sets of vertical joints in the targets rocks cutting almost diagonally across the square crater (Shoemaker 1963; Poelchau et al. 2009). The excavation flow exploited these planes of weakness and peeled back the surface rocks like petals of a flower, tearing along the joint directions. Most likely, the excavation of the smaller Kamil Crater occurred in a similar fashion, thereby evidencing a general mechanism at variable scales. Future investigations should also check how “petals” relate to the ejecta patterns and thus verify the relationship between pre-existing structures in the target rocks and ejecta blanket and ray distribution.
Estimate of the Recovered Projectile Mass
The exceptional status of preservation of the Kamil Crater structure and the pristine distribution of the meteorite specimens we found at the time of our first geophysical investigation (see D’Orazio et al.  for details) represent a unique opportunity for constraining directly the mass of the projectile that formed Kamil Crater, and thus for establishing an empirical relationship between crater structure and impactor mass.
Systematic visual search for meteorite specimens in and around the crater led to a minimum estimate of approximately 3400 kg of shrapnel >10 g lying on the surface (D’Orazio et al. 2011; Table 1). The geomagnetic survey conducted to identify buried impactor masses after completion of a visual systematic search in a 50 × 50 m gridded area about 450 × 450 m presented in this work reveals that no masses >100 kg are buried in the Kamil Crater area. Furthermore, the strongest dipolar magnetic anomalies arranged in a radial pattern around the crater up to a distance of 100 m from the crater rim (Fig. 13) are most likely generated by buried shrapnel >100 g for a total mass of approximately 1050–1700 kg (Table 1).
Therefore, the minimum mass of the projectile that formed Kamil Crater based on shrapnel count amounts to approximately 4450–5100 kg. Future investigations aimed at the determination of the total mass of surface and buried shrapnel less than 10 g and less than 100 g, respectively, as well as of the total mass of the projectile-derived impact melt particles preliminarily described in Folco et al. (2011) and D’Orazio et al. (2011) will provide a more complete and accurate estimate of the projectile mass. Note that scaling relations for impact craters forming in competent rocks (e.g., Collins et al. 2005; Melosh 2011) suggest that the projectile mass that generated Kamil Crater is of the order of approximately 10–20 t, as already mentioned in Folco et al. (2011). In particular, scaling relations predict that a transient crater approximately 35 m in diameter can be generated in a sedimentary target by iron projectiles ranging from approximately 1.2 to 1.7 m in diameter, assuming meteoroid entry velocities ranging from 12 to 18 km s−1 and entry angles ranging from 30 to 45°.
This is a comprehensive report on the geological and geophysical data collected at Kamil Crater, Egypt, during the first Italian-Egyptian expedition in February 2010. Data allow us to draw the following conclusions in terms of crater structure, impact scenario and, more in general, oblique impact processes:
1 Kamil Crater represents a model structure for small-scale hypervelocity impact craters on Earth. It occurs in a simple geological and geomorphological context; type of surface: rocky desert; topographic preimpact surface: flat; target rocks: layered sandstones with subhorizontal bedding. It is exceptionally well preserved with no evidence of important erosion or burial of the crater structure. Its ejecta ray pattern and its shrapnel distribution are essentially pristine (see also D’Orazio et al.  for details). The crater structure described above and summarized in Table 1 and Fig. 15 constitutes a solid set of input parameters for future modeling of the impact event that produced Kamil Crater and, in general, for calibrating analog and numerical models of small-scale impacts (<300 m in diameter) on Earth.
2 Kamil Crater was formed by the impact of a single mass, or a very tight cluster of fragments, that fragmented into thousands of shrapnel pieces upon hypervelocity impact. This conclusion is based on the fact that Kamil Crater is not part of a crater field, that it is characterized by a high transient crater depth-to-diameter ratio (dtc/Dtc ≥ 0.37; Table 1) and by a concave, yet asymmetric transient crater floor with an off-center bottom in downrange direction (Figs. 10 and 15). The Gebel Kamil iron meteorite that formed Kamil Crater (ungrouped ataxite) underwent only minor fragmentation and separation prior to impact with the ground. Gebel Kamil thus deviated at least in part from the current general model for projectile–atmosphere interaction (Bland and Artemieva 2006) predicting that fragmentation, lateral separation, and deceleration are the norm for small (<3 × 106 kg) iron impactors. Models for the Kamil Crater impact event must therefore consider nonaverage atmospheric entry conditions and/or different impactor strength input parameters, as suggested by Folco et al. (2011).
3 Comparison of the crater shape of the Kamil Crater (Fig. 15), its ejecta pattern (Fig. 2), and buried shrapnel distribution (Fig. 13) with laboratory experiments and models from the literature indicates that Kamil Crater formed through a moderately oblique impact from a northwestern direction. This is consistent with previous conclusion of an azimuth of the incoming impactor between 305 and 340° (D’Orazio et al. 2011), based on shrapnel distribution identified on the surface in and around the crater. Crater shape and ejecta pattern also suggest an impact angle most likely comprised between approximately 30 and 45°. In general, however, the patchy and not exactly bilateral symmetry of the ejecta distribution at Kamil Crater deviates somewhat from the ejecta distribution observed in hypervelocity laboratory experiments. This confirms that anisotropic properties of the target (e.g., structure and lithology) and projectile (e.g. shape) play a role in ejecta distribution.
4 Although analog models (Gault and Wedekind 1978) show that the crater shape is not very sensitive to impact angle for impact angles >30° (e.g., see also Pierazzo and Melosh 2000; and Collins et al.  for recent reviews), Kamil Crater is characterized by some minor but significant asymmetric morphological and structural characteristics that can help establish the obliquity of impacts in well-preserved craters. They include (Table 2) the off-center position of the bottom of the transient crater floor downrange (Figs. 10 and 15), the maximum overturning of target rocks along the impact direction (Fig. 4), and the lower crater rim crest elevation downrange (Figs. 3 and 11).
5 The somewhat polygonal outline observed in places at Kamil Crater (Fig. 9) points to a structural control on the cratering excavation flow that should be investigated further. Some segments of the polygonal outline appear to be limited by tear faults along which differential uplift of the crater rim strata is observed (Fig. 4). The excavation flow at Kamil Crater probably exploited pre-existing planes of weakness and rotated back sectors of the surface rocks like petals of a flower. This has been previously suggested in the case of the 1.2 km diameter Barringer Crater (Shoemaker 1963).
6 Our geomagnetic survey rules out the occurrence of buried masses >100 kg at Kamil Crater. A model of the detected magnetic anomalies indicates that the total mass of the buried shrapnel >100 g is between approximately 1050 and 1700 kg. By adding this mass value to that of the shrapnel >10 g identified on the surface during systematic visual search (D’Orazio et al. 2011), we can provide here an improved estimate of the minimum projectile mass, which is close to 5 t, namely comprised between approximately 4450 and 5100 kg (Table 1).
Acknowledgments— This work was carried out within the framework of the “2009 Italian-Egyptian Year of Science and Technology.” We thank Prof. M. Alsherbiny (President of the Egyptian National Academy for Scientific Research and Technology) and Prof. F. Porcelli (Scientific Attaché, Italian Embassy, Egypt) for diplomatic and institutional support; the Egyptian Army for logistical support; Pierre Rochette (CEREGE, France) for the measurements of the magnetic properties of the Gebel Kamil meteorite; Mario Di Martino (INAF - Pino Torinese, Italy) and Romano Serra (Bologna University, Italy), who took part in the geophysical expedition, for providing some pictures of Kamil Crater; the European Commission (through the Marie Curie Actions–RTNs ORIGINS project), the Italian Ministry of Education, University and Research (MIUR-PRIN 2008 project, code 008222KBS_005), the Fondazione Cassa di Risparmio di Torino, the Fondazione Banca Monte dei Paschi di Siena, the Banca Monte dei Paschi di Siena S.p.A., and Telespazio S.p.A. (in the framework of the ESA SSA program) for financial support. Vincenzo de Michele, former curator of the Civico Museo di Storia Naturale in Milan, discovered the crater in June 2008 through a Google Earth survey. We thank Dr. Natalia Artemieva, an anonymous reviewer, and associate editor W. U. Reimold for constructive review and editorial handling.