Physical characterization of a suite of Buzzard Coulee H4 chondrite fragments


Corresponding author. E-mail:


On November 20, 2008, the Buzzard Coulee H4 chondrite fell to Earth outside of Lloydminster, Alberta, Canada. Eighteen fresh samples obtained by the National Meteorite Collection of Canada, ranging from 8.80 to 109.14 g, were investigated in this study. Physical properties of the samples were first obtained using a suite of nondestructive techniques. The bulk density (Archimedean bead method: 3.48 ± 0.04 g cm−3; 3-D laser imaging: 3.46 ± 0.03 g cm−3; micro-computed tomography: 3.44 ± 0.03 g cm−3), porosity (6.2 ± 0.1%), bulk magnetic susceptibility (log χ: 5.364 ± 0.056 × 10−9 m3 kg−1 at 825 Hz; 5.329 ± 0.052 × 10−9 m3 kg−1 at 19,000 Hz), and other derived magnetic properties (frequency dependence: 8.7 ± 6.2%; degree of anisotropy A%: 22.0 ± 2.0%; ellipsoid shape B%: −18.7 ± 8.7%) are typical of H chondrites. The coefficient of variation associated with the properties measured directly was low (0.10–1.15%), indicating that the samples are homogenous at the interfragment scale. The study then proceeded with detailed analyses at the intrafragment scale. Visual inspection of micro-computed tomographic images allowed the identification of an anomalous large clast with low metal content in a fragment. Another fragment exhibited macroscopic shock veins that warranted further examination. These fragments were cut and polished thin sections prepared for petrological analysis by optical and scanning electron microscopy. Based on mineralogical and textural similarities with several chondrules, the large clast was interpreted to be a macrochondrule. In a larger context, this study proposes a protocol for the systematic investigation of extraterrestrial material that can be exported to other new meteorite falls and finds, and specimens from sample return mission.


On the early evening of 20 November 2008, the Buzzard Coulee meteorite fell to Earth in a widely witnessed fireball south of the City of Lloydminster, near the Saskatchewan–Alberta interprovincial border, Canada. Video footage, all-sky-cameras, and eyewitness reports were initially used to estimate its trajectory (Hildebrand et al. 2009; Milley et al. 2010; Brown et al. 2011). The fireball proceeded from north to south, passing nearly directly over Lloydminster, terminating just north of the Battle River to the east of the interprovincial border. The first meteorite fragments were found seven days after the fall on a small frozen pond in Buzzard Coulee, Saskatchewan.

Infrasound records from six stations captured acoustic shockwaves from the fireball, providing an estimate of the total energy of the event that implied an initial meteoroid mass of approximately 15 tons upon atmospheric entry. Detailed examination of calibrated video and security camera records yielded an entry angle of 67° toward 168° azimuth, with the first significant fragmentation having occurred at >60 km altitude. Buzzard Coulee underwent multiple fragmentation episodes, notably at altitudes of 12 km and 2–3 km. Video records and eyewitness accounts indicate that at least three large fragments were observed as red glowing objects below approximately 12 km (Hildebrand et al. 2009; Milley et al. 2010). An initial entry velocity of approximately 18 km s−1 was determined from video records, which, coupled with the resolved trajectory, indicates that Buzzard Coulee was delivered to the Earth on an Apollo-type orbit that would place its origin amongst the evolved, near-Earth object (NEO) population (Hildebrand et al. 2009; Milley et al. 2010). Buzzard Coulee is one of only 14 meteorites for which a preatmospheric orbit is known, and is the second known direct representative of the evolved NEO population inside of Mars' orbit (Brown et al. 2011).

The Buzzard Coulee strewn field covers an area of 10 km north-to-south by 3 km wide, with a further 3 km extension eastward due to wind deflection of small fragments. More than 200 Buzzard Coulee fragments (>50 kg) were recovered by organized searchers on private land before the arrival of snow halted search efforts on December 6, 2008 (Hildebrand et al. 2009). Possibly a similar number were recovered by members of the public from public and private land during the same period (Paulson 2009). To date, at least three thousand fragments have been found with a total known mass in excess of 200 kg (Milley et al. 2010; Wilson and McCausland 2012).

Description of the Fragments Studied

The Buzzard Coulee meteorite was classified as an H4 ordinary chondrite based on the degree of chondrule-matrix integration and fayalite and ferrosilite contents of olivine and pyroxene (Hutson et al. 2009; Walton et al. 2009). It has an intermediate amount of well defined chondrules and little alteration, and has been assigned a low shock classification of S2/S3 (Hutson et al. 2009; Walton and Herd 2009). Having been recovered soon after the fall occurred, it has a weathering grade of W0 (Hutson et al. 2009).

The National Meteorite Collection of Canada obtained 18 individual fragments of the Buzzard Coulee H4 chondrite collected by members of the public within 3 weeks of the fall, before the winter of 2008–2009 (Tables 1 and 2). The fragments originate from the northern portion of the strewn field, within a few hundred meters of location 53°0.45′ N and 109°50.60′ W. They range from 8.80 to 109.14 g, covering more than an order of magnitude in mass, and exhibit shapes from rounded to angular. All have fusion crust or fresh ablation surfaces, with varying degrees of oxidation affecting them. Rusting tends to be concentrated in depressions or at the center of a given surface, and has developed in patches. Usually three generations of fusion crust development can be identified. The oldest generation is dusky brown (5YR 2/2) (Goddard et al. 1963), very smooth, and fairly rounded, giving the surfaces a convex shape with sometimes a web of shallow, polygonal-shaped cracks. The middle generation tends to be concave, with a number of pits or regmaglypts, which are usually a few millimeters to a centimeter across. The bottoms of the pits are grayish brown (5YR 3/2) in color (Goddard et al. 1963). This crust is seen to vary between fully developed fusion surfaces and ones which are barely affected. The youngest generation is a glassy black coating of melted crust, which often displays flow lines. It is found mostly on edges of fragments and on the high points of the dimpled surfaces. At times it can cover an entire surface. The older generations of crust can be seen peeking out from under it. The ablation surfaces are black to dusky brown in color, with the dark reddish brown (10R 3/4) color of minor rust (Goddard et al. 1963). The ablation surface is cracked on some of the samples and, in a few places, has chipped away showing a fresh, gray interior (Fig. 1). Some of the irregular fracture surfaces are “painted” with a thin fusion crust, likely reflecting exposure late in atmospheric entry. The exposed interior of the Buzzard Coulee fragments is fairly homogeneous. It is composed of white and gray silicate material, with lighter gray spots and dark to silvery inclusions of metal. The metal inclusions exhibit oxidation. Fragment A1 displays dark interior veins that suggest brecciation or shock (Fig. 1). The angular fragments, painted surfaces, and some of the large, fresh interior surfaces indicate the very fragmented nature of the meteorite.

Table 1. Physical properties of 18 Buzzard Coulee fragments. Outliers (defined as data falling outside two standard deviations from the mean) shown in parentheses and excluded from statistics
FragmentMass (g)Bulk density (g cm−3)Helium pycnometry dataMagnetic susceptibility data
Archimedean bead methoda3-D Laser imagingMicro-CTGrain volumeb (cm3)Grain density (g cm−3)Porosity (%)Bulk mass magnetic susceptibility log χ (10−9 m3 kg−1)Frequency dependence (%)Degree of anisotropy A%d (%)Ellipsoid shape B%d (%)
825 (Hz)c19,000 (Hz)c
  1. a

    Eight measurements were made, and the average and standard deviation are reported.

  2. b

    Five measurements were made, and the average and standard deviation are reported.

  3. c

    Six measurements were made, and the standard deviations ranged from 0.001 to 0.00002.

  4. d

    The uncertainty for each value of A% and B% was determined as the standard deviation of all 11 measurements, on the assumption that all the fragments represent random pieces from a single population (in a statistical sense) with homogeneous characteristics. For A% this value is 2.0 and for B% this value is 8.7.

B1109.143.46 ± 0.043.47Fragment too large29.30 ± 0.043.73 ± too largeFragments too large
A167.323.52 ± 0.043.453.4518.11 ± 0.033.72 ± too largeFragments too large
B261.673.53 ± 0.043.453.4216.66 ± 0.033.70 ± too largeFragments too large
A251.303.46 ± 0.043.473.4513.97 ± 0.023.67 ± too largeFragments too large
A350.503.44 ± 0.043.503.4613.66 ± 0.013.70 ±−14.8
C48.823.55 ± 0.043.493.5313.25 ± 0.063.68 ± too largeFragment too large
A447.283.41 ± 0.043.48(3.67)12.83 ± 0.013.68 ±−24.2
A546.663.48 ± 0.043.473.4512.69 ± 0.013.68 ±−16.3
A645.223.48 ± 0.083.493.4512.35 ± 0.023.66 ± too largeFragment too large
B342.073.47 ± 0.043.463.4011.54 ± 0.043.65 ±−2.9
B438.183.49 ± 0.043.47(3.74)10.35 ± 0.093.69 ±−29.2
B537.383.48 ± 0.043.473.4310.17 ± 0.023.67 ± too largeFragment too large
A736.593.50 ± 0.043.433.429.80 ± 0.013.73 ±−21.7
A836.503.53 ± 0.043.523.479.76 ± 0.023.74 ±−29.7
B628.743.50 ± 0.043.413.447.77 ± 0.013.70 ±−24.7
B725.193.43 ± 0.043.463.436.81 ± 0.013.70 ±−6.5
A920.603.44 ± 0.043.423.425.63 ± 0.023.66 ±−25.0
A108.80Fragment too small3.443.382.41 ± 0.033.66 ±−10.9
Number of fragments 1718151818181818181111
Average 3.483.463.4412.063.696.25.3645.3298.722.0−18.7
Standard deviation
Coefficient of variation (%) 1.150.870.8747.650.7316.91.0440.1071.89.246.3
Table 2. Fragment description
FragmentMass (g)Fusion crustFeaturesNumber of surfaces generatedWeatheringShapeWidth × depth × height
B1109.14Covers 95% of surface, with small pieces missing.Radiating flow lines cover one of the surfaces.3Light rusting8-sided asymmetrical polyhedron.4.5 cm × 3 cm × 3.5 cm
A167.32Covers 70% of surface. 2 cm × 2.5 cm piece missing, along with two smaller pieces on other sides. Looks to be a piece of a larger fragment.Some dark veins crosscut surface at an angle. Has been cut parallel to veins. Polygonal cracks in the crust.3Light rustingOblong, tapering to one end.4.5 cm × 2.5 cm × 3 cm
B261.67Covers 95% of surface, with small dings and pieces missing. One side is broken and only partially crusted.One side is dimpled.3Light rustingTapered rectangular prism.4.5 cm × 2.5 cm × 2 cm
A251.30Covers 85% of surface. 1 × 1.5 cm, 1 × 1.5 cm, 2 × 0.5 cm pieces missing. Bits of metal can be seen sticking into the crust.Youngest fusion surface has flow lines. Has radiating flow lines at one end.≥3Light rustingTriangular off-centered pyramid, with a concave side.4 cm × 2.7 cm × 2 cm
A350.50Covers 60% of surface. 1 × 0.7 cm, 1 × 1.5 cm, 1 × 0.5 cm pieces missing. One surface is only weakly crusted, while others can be quite smooth. There is a surface that is dimpled.≥3Light rustingPolyhedron.4 cm × 2.5 cm × 3 cm
C48.82Crust covers 99% of surface, with small mm-sized bits missing.Large metal shard is exposed on crustal surface. Suggestion of flow lines on some edges.3Light rustingPolyhedron; truncated triangular pyramid.5 cm × 3 cm × 1.75 cm
A447.28Covers 90% of surface. 1 × 0.8 cm piece is missing. A 1.2 cm × 0.8 cm surface is partially crusted.Has been cut to reveal a macrochondrule in the fragment's interior. Can be oriented by splashes/flow lines that extend over the rounded end.≥3Light rustingBalloon shaped; one end is hemispherical, while the other forms a ridge.3.5 cm × 2.5 cm × 3 cm
A546.66Covers 99% of surface, with one or two small bits missing.Has very fresh surface features, flow lines, and polygonal fracturing.4Light rustingResembles a triangular pyramid, with rounded edges.4 cm × 3 cm × 2.2 cm
A645.22Covers 99% of surface, with minor amounts missing.Some flow lines on surface.3Light rustingDumbbell shaped. 5 cm × 2.5 cm × 2.2 cm
B342.07Covers 85% of surface. A1.5 × 1.5 fragment is missing at one end. 3Light rustingA tapered polyhedron. 4 cm × 2.5 cm × 2 cm
B438.18Covers 99% of surface. Has some small dings.Maybe possible to orient from a fusion ring formed by the youngest generation of crust.3Light rustingHalf a hemisphere; one side is very rounded with 3 flatter ones.4 cm × 3 cm × 2.5 cm
B537.38Covers 99% of surface. 3Light rustingTapered, cigar shape.5.5 cm × 1.75 cm × 1.75 cm
A736.59Covers 90% of surface. 0.7 cm × 0.7 cm and 1 × 1 cm pieces missing, and a few nicks.Can be oriented from flow lines on surface.3Moderate rustingResembles a triangular pyramid, with rounded edges and flat apex.3.75 cm x 2.75 cm x 1.75 cm
A836.50Covers 95% of surface. 3 small pieces (>0.5 cm × 0.5 cm) missing.Flow lines are present on one end.3Moderate rustingFlat; triangular prism; almost tabulate/heart shape.3.5 cm × 3.5 cm × 1.2 cm
B628.74Covers 90% of surface. Missing small pieces of crust.Has rollover/splash from melting crust. Can be oriented based on youngest generation.3Light rustingAsymmetrical 8-sided polyhedron; edges are subangular to rounded.3 cm × 2.5 cm × 1.75 cm
B725.19Covers 70% of the surface. Multiple pieces of the crust have flaked off. 2Light rustingAsymmetrical 7-sided polyhedron; edges are subangular to rounded.3 cm × 2 cm × 1.5 cm
A920.60Covers 95% of surface. Some of the surface is lightly crusted and appears glassy.Has one very rounded face, reminiscent of a sector out of a sphere. Flow lines present.4Light rustingVery asymmetrical; almost a polyhedron, but has one large, very rounded side.2.5 cm × 2 cm × 1.75 cm
A108.80Covers 95% of surface. Small dings and nicks have removed pieces of the fusion crust.Some flow lines on surface.3Light rustingAlmost pyramidal; has rounded edges. 2.2 cm × 1.75 cm × 1.2 cm
Figure 1.

Buzzard Coulee fragment A1. (Left) Digital photograph with a 1 cm scale bar. The broken fusion crust reveals linear features interpreted as shock veins, indicated by an arrow. (Middle) 3-D laser model. (Right) Micro-CT isosurface model.


The overarching objective of this paper is to demonstrate a methodology for rapid characterization of a suite of meteorite samples while preserving the integrity of the samples. More specifically, this study is concerned with documenting each Buzzard Coulee fragment without changing its current condition, and for that reason, a variety of nondestructive techniques have been used to characterize its physical properties. Because each fragment was collected right after the fall, but before the winter, this characterization should reflect fairly accurately the characteristics of the parent asteroid that this meteorite came from. Properties investigated are bulk density, grain density, porosity, and magnetic susceptibility, with focus on interfragment and intrafragment variability (Fry 2011) (Table 1). Results were used to strategically select a sample to be cut into a thin section for further mineralogical and textural analysis (Melanson 2012).

Methods and Instruments

Bulk density is the mass per unit of bulk volume of each fragment, including pore spaces. Each fragment was weighed upon acquisition to obtain the mass. Bulk volume was later measured using 3 different nondestructive techniques: (1) the Archimedean bead method (Consolmagno and Britt 1998; Macke et al. 2009), (2) 3-D laser imaging (Smith et al. 2006b), and (3) X-ray micro-computed tomography (micro-CT) (Ebel and Rivers 2007; McCausland et al. 2010). The Archimedean bead method involved burying each fragment under 100 μm silica beads in a container of known volume. The beads imitate the flow of a liquid and closely surround the fragment. They, however, suffer from errors due to variable compaction and are sensitive to environmental conditions. Errors increase as the fragments get smaller, making the approach impractical for fragments smaller than 5 cm3 (Macke et al. 2009). In addition, a Konica-Minolta Vivid 9i 3-D laser camera was used to take 52 images of each fragment at a resolution of 640 by 480 voxels. These images provided comprehensive coverage of a fragment as it was rotated on a turntable and flipped so that all surfaces could be captured. Images were imported into visualization software that allows an operator to assemble them into highly accurate 3-D models and to calculate volume automatically (Fig. 1[middle]). This approach has been demonstrated to be reliable for fragments as small as 0.5 cm3 (McCausland et al. 2011). Finally, the fragments were imaged using a GEHC eXplore speCZT micro-CT scanner operating at 110 kV (peak) at the Robarts Research Institute of Western University, London, Ontario, Canada (McCausland et al. 2010). Each 5 min scan consisted of 900 separate X-ray radiographs taken over one full 360° rotation about the sample. The radiographs were used to reconstruct a 3-D image of the sample's radio-density displayed in grayscale Hounsfield units (HU). Each voxel in the image is a cubic element with a 50 μm edge length. GE Microview software was used to drape an isosurface over the image based on a threshold HU level and volume was calculated automatically (Fig. 1[right]). Micro-CT images were also used to visually inspect the interior of each fragment. After detailed examination of their micro-CT images, fragments A1 and A4 were cut to prepare polished thin sections for investigation using transmitted and reflected light microscopy, scanning electron microscopy (SEM), and energy dispersive spectrometry (EDS).

Grain density is the mass per unit of grain volume of each fragment and porosity is the percentage of void space within the bulk volume of a fragment, calculated as:

display math(1)

Bulk volume measured using 3-D laser imaging was used in Equation (1) as it was available for all samples and has been shown to be consistent with the Archimedean bead method (McCausland et al. 2011). Grain volume was measured using a Quantachrome Instruments helium gas multipycnometer, following the procedure of McCausland and Flemming (2006), yielding a grain density for each fragment based on a minimum of five grain volume measurements per sample. Porosity for each fragment was then calculated directly as a percentage of bulk volume using Equation (1), allowing for comparison between fragments and with other meteorites.

Magnetic susceptibility is the ratio of the induced magnetization of a material to the strength of an externally applied magnetic field. This study followed the methodology described in Smith et al. (2006a) and examined bulk magnetic susceptibility (reported as the logarithm of bulk mass magnetic susceptibility, log χ, for consistency with previously published data; Rochette et al. 2003, 2009), its frequency dependence, and the degree and shape of the anisotropy of magnetic susceptibility (AMS) (Tarling and Hrouda 1993). Magnetic susceptibility was measured using a SI-2B susceptibility meter manufactured by Sapphire Instruments. The instrument has two separate coils: a smaller (4.5 cm in diameter), internal coil, which operates at a frequency of 19,000 Hz, and a larger (5.1 cm in diameter), external coil, which operates at 825 Hz. The frequency dependence of magnetic susceptibility, χFD, could therefore be estimated and expressed as a percentage using:

display math(2)

where χ825 and χ19,000 are the bulk mass magnetic susceptibilities at a frequency of 825 Hz and 19,000 Hz, respectively. Furthermore, AMS was examined using two parameters known as the degree of anisotropy (A%) and ellipsoid shape (B%), defined as (Canon-Tapia et al. 1997):

display math(3)
display math(4)

where k1, k2, and k3 correspond to the maximum, intermediate, and minimum bulk volume susceptibility measured using the external coil of the SI-2B instrument. An A% value of zero is indicative of an isotropic sample. Parameter B%, ranging from −100% (oblate) to +100% (prolate), defines the shape of the magnetic susceptibility ellipsoid. Finally, two fragments (A9 and A10) were small enough to fit in the cylindrical sample holder (length = 2.4 cm, diameter = 2.4 cm) of a Schonstedt SSM-1D fluxgate spinner magnetometer, allowing their natural remnant magnetization (NRM) to be measured. The two fragments were held securely in the plastic holder and spun inside the magnetometer in the same six different orientations used for the anisotropy measurements. The measurements took place in a magnetically shielded room at the Geological Survey of Canada that reduced the external magnetic field to a level of less than 3000 nT.

Physical Properties

Bulk Density

The three different techniques used to measure bulk density for the 18 Buzzard Coulee fragments returned very consistent values (standard deviation = 0.03–0.04 g cm−3 after removing two outliers in the micro-CT data) (Table 1; Fig. 2). The densities determined using the Archimedean bead method and 3-D laser imaging average 3.48 g cm−3 and 3.46 g cm−3, respectively. The percentage difference between the two methods (fragment A10 excluded) is only 0.5%, half the value reported in previous studies (see table 3 in Smith et al. 2006b), attesting to the high accuracy of the laser camera used (Fry et al. 2011). Micro-CT densities are in most cases smaller, with an average of 3.44 g cm−3. Observationally, the isosurface obtained through the analysis of the micro-CT reconstructions shows less detail than that captured with the laser camera (Fig. 1); it is possible that the fragment volumes obtained from micro-CT are systematically slightly larger due to loss of surface detail in comparison with 3-D laser imaging. The main advantage of micro-CT was the speed at which it produced volumetric results. Overall, the density values obtained in this study are consistent with the findings of other researchers using the Archimedean bead method (3.26–3.45 g cm−3 for six fragments: Hildebrand et al. 2009) and 3-D laser imaging (3.5 g cm−3 for one 151.7 g fragment: Walton and Herd 2009; Walton et al. 2009), and are typical values for an H chondrite (Rochette et al. 2009). There is no evidence to support the bimodal bulk density distribution reported by Hildebrand et al. (2009).

Figure 2.

Bulk density versus mass for 18 Buzzard Coulee fragments (outliers in the micro-CT data in Table 1 excluded).

Grain Density and Porosity

The 18 Buzzard Coulee fragments have grain densities determined by helium pycnometry, which range from 3.65 to 3.74 g cm−3, averaging 3.69 ± 0.03 g cm−3 (Table 1). Using the bulk volume found for each fragment using 3-D laser imaging (Equation (1)), fragment porosity ranges from 4.7 to 8.7% with an average of 6.2 ± 1.0%. The porosity in each fragment likely represents void space in the form of microcracks and grain boundary separations. The Buzzard Coulee grain density and porosity averages are typical of those reported for H chondrite falls (e.g., Wilkinson et al. 2003; Consolmagno et al. 2008).

Interestingly, fragment porosity is greater with higher grain density (Fig. 3), whereas there is no significant variation in fragment porosity with bulk density (Table 1). This relationship is consistent with a weathering process in which a high-density component (FeNi metal or sulfide) is removed and replaced with oxide-weathering products, which begin to occupy the existing pore spaces (Consolmagno et al. 2008). But this weathering effect is most unlikely for the Buzzard Coulee samples, as they lay on the Earth's surface for only a very short time. Another possibility to be pursued is that melting and ablation during atmospheric entry removed denser metallic phases and essentially replaced them with the thin variably oxidized glassy fusion crusts, more for some samples than others. How this might have affected densities is unknown, although the specimen suite contains relatively small samples for which the crust area/sample volume ratio is high.

Figure 3.

Grain density versus porosity for 18 Buzzard Coulee fragments.

Magnetic Properties

The average bulk mass magnetic susceptibility of the Buzzard Coulee samples was estimated to be log χ 5.364 ± 0.056 × 10−9 m3 kg−1 when measured at a frequency of 825 Hz and log χ = 5.329 ± 0.052 × 10−9 m3 kg−1 at a frequency of 19,000 Hz (Table 1). The Buzzard Coulee bulk magnetic susceptibility data are within the range of a typical H chondrite. A sampling of 34 H4 chondrites was found to possess a similar average bulk mass magnetic susceptibility of log χ = 5.29 ± 0.09 × 10−9 m3 kg−1 (Rochette et al. 2003). The graph of bulk mass susceptibility versus sample mass does not show any trend (Fig. 4).

Figure 4.

Logarithm of bulk magnetic susceptibility versus mass for 18 Buzzard Coulee fragments, at two different frequencies.

In all cases, the bulk mass magnetic susceptibility measured using the 19,000 Hz coil is less than or equal to that measured using the 825 Hz coil. The frequency dependence was quantified using equation (2) and χFD ranges from 0.4 to 28.1% with an average of 8.7 ± 6.2% (Table 1). These values are similar to the average χFD of 10.0% and maximum of 25.6% reported in Smith et al. (2006a) for 22 H chondrites.

AMS was measured for 11 of the 18 Buzzard Coulee fragments; the remaining 7 could not be measured because they were too large to fit into the external coil of the SI-2B susceptibility meter. The data are expressed in terms of A% (Equation (3)) and B% (Equation (4)) (Table 1). The degree of anisotropy A% is moderate and ranges from 18.0% to 26.1%, with an average of 22.0 ± 2.0%. The ellipsoid shape of susceptibility given by variable B% has a range of −2.9 to −29.7%, with an average of −18.7 ± 8.7%. These negative values for B% indicate that the ellipsoid is oblate, which implies that the magnetic mineral distribution in the fragments takes on a planar aspect, an observation consistent with previous studies on the AMS of chondrites (Smith et al. 2006a). On a graph of A% versus B% (Fig. 5), the narrow range for A% values indicates that the degree of anisotropy remains fairly constant regardless of ellipsoid shape. It has been noted that meteorites with high shock levels have high degrees of anisotropy (Gattacceca et al. 2005). The low shock classification and moderate anisotropy of the Buzzard Coulee fragments, however, rather support a nonimpact origin. Perhaps the magnetic anisotropy is linked with the overburden pressure that occurs while the parent body is accreting (Gattacceca et al. 2005). Petrological analysis (see below) does, however, indicate that some samples contain shock features indicative of a higher shock grade than previously recognized.

Figure 5.

Degree of anisotropy (A%) versus ellipsoid shape (B%) for 11 Buzzard Coulee fragments.

The two NRM measurements yielded angles of 43° (fragment A10) and 80° (fragment A9) between the NRM direction and the k1k2 plane of the dominantly oblate AMS fabric. A much larger measurement population would be needed to test for any consistency between these two directions which, if observed, would imply a common origin for NRM and AMS.

X-ray Micro-Computed Tomography and Preliminary Petrology

By manipulating the 3 orthogonal cross sections in the GE Microview software, the micro-CT images of each fragment were visually inspected systematically (Fig. 6). High atomic number features, such as metallic grains, coincide with bright voxels on micro-CT images, whereas the silicate matrix appears as a darker gray medium.

Figure 6.

Three orthogonal cross sections oriented to show a macrochondrule (diameter = 8.0 mm) in the micro-CT image of fragment A4 (left), and the corresponding planar projections (right).

None of the fragments exhibited a preferential distribution of metallic grains, which is consistent with the observation that they were found to only have a moderate degree of magnetic anisotropy (A%). In almost fully crusted sample A4, an essentially metal-free interior region, invisible from the exterior, was revealed. The large clast region was dimensioned using the GE Microview software (Fig. 6). A plane containing the clast was dimensioned and outlined on the surface of fragment A4. The fragment was cut along the transcribed plane using a diamond-wire saw and a polished thin section was prepared. Similarly, another plane was outlined near the broken edge of fragment A1 to further investigate the dark veins (Fig. 1).

A selection of 24 chondrules from the polished thin sections were described (Melanson 2012). In general, these chondrules are composed of typical chondritic minerals, dominated by forsteritic olivine and pyroxenes, and display a wide array of textures (classes PO, PP, BO, RP, POP from the Gooding and Keil 1981 chondrule classification scheme are represented). Of the observed chondrules, most are intact. There is a variable, but consistently observable, proportion of mesostasis material between grains in the chondrules, often including most of the chondrule's metals. Many EDS analyses of the mesostasis indicated a plagioclase–pyroxene composition, often with a signature of metallic minerals (including troilite and kamacite) or trace elements (including titanium and chromium).

A mosaic of petrographic microscope images in a mix of plane-polarized transmitted and reflected light was compiled, centered on the large clast identified in fragment A4 to reveal textural details and an overall lack of metals (Fig. 7). The petrological analysis found that approximately 45% of the clast consists of equant subangular olivine macrocrysts with lesser elongate or subrounded macrocrysts and microcrysts, as well as minor equant subangular pyroxene macrocrysts. The clast is rather homogeneous, contains a small number of troilite grains, and has an intact glassy-metallic rounded envelope. Two olivine macrocrysts analyzed with EDS were found to have forsterite numbers of 82.16 and 82.05. The mesostasis, which accounts for approximately 15% of the clast, was found to have a mixed pyroxene–plagioclase composition and low metal content. These mineralogical-textural findings are similar to other chondrules observed in the thin section, and led to this clast being designated a macrochondrule based on its size (Melanson 2012).

Figure 7.

Mixed plane-polarized transmitted and reflected light photo mosaic of the macrochondrule identified in Fig. 6. Metallic grains appear in tan-yellow in reflected light. Low metal content in the macrochondrule is consistent with the absence of bright speckles in the micro-CT image in Fig. 6.

Following the same approach, a polished thin section from fragment A1 was examined using optical and SEM, focused on macroscopically identified dark veins (Figs. 1 and 8). The SEM image displays potential mobilization of metal within the vein zone (Fig. 8 [middle]). Optical investigation showed significant glass, variable thickness, a fractured/displaced olivine grain, and a higher proportion of metals. Rounded olivine grains on either side of the vein had compositions of Fo82 and Fo83 and pyroxenes had En80 and En83, all indicating high magnesium composition. A metallic analysis in the vein zone returned a mix of troilite and the nickel-rich alloy taenite. Imaging these metals showed a fine-grained intergrowth texture of phases not observed in other areas. All these observations are consistent with shock veins observed in Buzzard Coulee (Walton and Herd 2009; Walton et al. 2009) and other chondrites (Friedrich et al. 2012), and support their initial interpretation as shock veins.

Figure 8.

(Top) Mixed plane-polarized transmitted and reflected light photo mosaic of the shock veins identified in Fig. 1. (Middle) scanning electron microscopy (SEM) image of the boxed area in the top figure, showing mobilized metallic minerals. (Bottom) SEM image of the boxed area in the middle figure, showing metallic intergrowth texture.

Discussion and Concluding Remarks

This study has contributed several new observations on the physical properties of 18 fragments from a single meteorite fall, the Buzzard Coulee H4 chondrite. The new data fall consistently within the normal range for fresh H chondrite falls (Rochette et al. 2003, 2009; Smith et al. 2006a; Consolmagno et al. 2008). The fragments are good representatives of the original meteoroid because of their limited exposure to the effects of weathering.

With regards to interfragment variability, the coefficient of variation attached to the basic physical properties measured directly—bulk density, grain density, magnetic susceptibility—falls within a small range between 0.10 and 1.15% (Table 1). The distribution of measured values appears to be drawn from a single population, suggesting that the Buzzard Coulee meteoroid, as represented by these fragments, was homogenous down to the 10 to 100 g scale. Similar homogeneity has been noted elsewhere for multiple fragments from the Holbrook L6 chondrite fall (Wilkinson et al. 2003) and the Gao-Guenie H5 chondrite fall (Beech et al. 2009). The fragments considered in this study, however, were recovered from a restricted area in the northern end of the Buzzard Coulee strewn field; the suite of measurements presented here could in the future be performed on fragments in other strewn field locations to check if the physical properties are indeed uniform at the scale of the strewn field and representative of the whole meteoroid.

Micro-CT imaging was instrumental in the reconnaissance investigation of heterogeneity between Buzzard Coulee fragments, enabling follow-up targeted petrological analyses. Overall, the chemical composition and mineralogical assemblage of the fragments were relatively uniform, although their associated chondrules exhibited a rich variety of textures. Anomalous features included a large macrochondrule and shock veins. These findings for this suite may offer support for continued study of the physical and mineralogical-textural properties of Buzzard Coulee to better constrain variabilities, which have been shown to exist in this and other studies (Walton et al. 2009; Melanson et al. 2012; Ruzicka et al. 2012).

Shock veins observed to be present on the broken surface of fragment A1 were spatially defined using micro-CT imaging, and then confirmed petrographically with observed shock-related textures in polished thin sections, such as occasional parallel-fractured grains. The effects of minor static-recrystallization or annealing processes (Winter 2009) have perhaps been observed, possibly explaining the abundant rounded and embayed (e.g., resorbed) grains, the remarkably consistent olivine and enstatite compositions, and the lack of zoned crystals. Features like shock veins and parallel fractures are typically caused by late deformation. With these observations in mind, it is possible that this Buzzard Coulee suite may have experienced more shock than its initial S2 classification suggests, at which shock veins and parallel fractures are not supposed to have developed. Shock veins are defined as being present in S3 or greater (Stöffler et al. 1991); in this case, the limited observation of shock veins across all micro-CT imaged Buzzard Coulee fragments may mean that the minimal S3 level is appropriate for the meteorite, or perhaps that the H4 breccia may consist of clasts with varying shock states.

Importantly, this suite of Buzzard Coulee fragments afforded an opportunity for systematic and multifaceted study, to help define a protocol and methodologies for the nondestructive examination of new meteorite falls and finds, or for specimens from sample return missions. The approach presented here preserves the integrity of the samples before they are cut or reduced to powder for further analysis.


We are grateful to the Department of Canadian Heritage Movable Cultural Property program for assistance in purchasing most of the specimens, and to the National Meteorite Collection of Canada for the loan of samples. We thank Ian Nicklin and Kim Tait from the Royal Ontario Museum, who helped with sample preparation and cutting; Pat Hunt from the Geological Survey of Canada, Natural Resources Canada, for her assistance with the SEM and EDS analyses; Kenneth Buchan, also from Natural Resources Canada, for performing the magnetic remanence measurements; and Fred Gaidies of the Department of Earth Sciences at Carleton University for his insights into micro-CT. In Summer 2011, Dave Melanson worked on this project with the support of the Cox endowment in mineralogy from the Department of Earth Sciences at Carleton University.

Editorial Handling

Dr. Edward Scott