Journal of African Earth Sciences 44 (2006) 314–330 www.elsevier.com/locate/jafrearsci
The Trompsburg Complex, South Africa: A preliminary three dimensional model L.P. Mare´ *, J. Cole Council for Geoscience, Private Bag X112, Pretoria 0001, South Africa Received 21 January 2005; received in revised form 10 August 2005; accepted 22 November 2005 Available online 9 February 2006
Abstract The 1915 ± 6 Ma Trompsburg Complex is a large layered mafic intrusion, measuring approximately 2500 km2 in area. Seven boreholes were drilled in the 1940s to constrain a strong gravimetric and magnetic anomaly. These approximately 2 km deep boreholes intersected up to 760 m of gabbro-troctolite–anorthosite containing up to 19 massive magnetite layers. Physical property analyses including density, magnetic susceptibility, intensity of magnetization, electrical resistivity and induced polarization were performed on specimens from these cores. Gridded aeromagnetic and Bouguer gravity data, combined with the available borehole information were used to construct a three dimensional polygonal model of the Trompsburg Complex. This was then used to calculate a forward three dimensional gravity and magnetic model using an algorithm that computes the gravity and magnetic fields for polyhedrons. Unfortunately the boreholes were confined to the northwestern part of the Complex and as a result the rest of the model is not as well constrained as in this area. The model is saucer-shaped at the top, with a large feeder to the south-east. The extremely high values of the Bouguer gravity anomaly reaching a maximum of almost 93 mGal with respect to the regional suggests a substantial thickness of more than 6000 m for the Complex reaching up to 16,000 m at the feeder. 2005 Elsevier Ltd. All rights reserved. Keywords: Trompsburg; Physical properties; Layered intrusion; 3D model
1. Introduction The 1915 ± 6 Ma (Maier et al., 2003) Trompsburg Complex is a roughly circular in shape, layered mafic intrusion, with a diameter of nearly 50 km. It forms a suboutcrop at a depth of approximately 450 m near the town of Trompsburg in the Free State Province, South Africa. Granitic rocks are present near the centre of the intrusion, but the greater part of the Complex consists of layered mafic igneous rocks ranging from coarse-grained gabbro, anorthosite and mafic pegmatoid to an abundance of titaniferous magnetitite (Ortlepp, 1959; Buchmann, 1960; Reynolds, 1979). The Complex was intruded into dolomites of probable Transvaal age and is unconformably *
Corresponding author. Fax: +27 12 841 1424. E-mail addresses:
[email protected] (L.P. Mare´), jcole@ geoscience.org.za (J. Cole). 1464-343X/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.jafrearsci.2005.11.026
overlain by Karoo sediments (Table 1, Ortlepp, 1959; Buchmann, 1960). During the 1940s seven boreholes were drilled in the north-western sector of the Complex, and the information obtained from these was used in the construction of a possible cross-section of this part of the Complex. Ortlepp (1959) and Buchmann (1960) summarised the information collected during these investigations. According to the cross-section provided by Buchmann (1960), the thickness of the layered mafic rocks could be on the order of 10,000 m, although Ortlepp considered 2000–3000 m to be a more reasonable figure. Fig. 2 shows the locations of the boreholes plotted on the regional gravity and magnetic data. Several mineralogical and geochemical studies have been conducted on samples from the seven borehole cores (Reynolds, 1979; Logan, 1979) to determine the economic value of the Complex and identify any platinum-group
L.P. Mare´, J. Cole / Journal of African Earth Sciences 44 (2006) 314–330
minerals in the intrusion itself and in the adjacent metamorphosed dolomitic country rocks. According to Reynolds (1979) the vanadium-bearing titaniferous-iron ores of the Trompsburg Complex can be compared chemically and mineralogically with similar ores from other basic complexes. The ores are locally rich in apatite, and their composition at certain levels approaches that of the olivine apatite–magnetites of the Villa Nora area of the Bushveld Complex (Reynolds, 1979). Although Logan (1979) identified no platinum-group minerals, an adequate assessment of the economic potential of the Complex was not possible due to the limited core samples. The fact that the Trompsburg intrusion shares several compositional, lithologic and stratigraphic features with the Bushveld Complex, the world’s most important source of PGE, Cr, and V deposits, had at the time raised the interest of mining companies as well as government research agencies (Reynolds, 1979). Anhaeusser (2004) reflects that although the Trompsburg intrusion could represent a significant exploration target for platinum group metals, the great depth of the intrusion will probably preclude any mineral occurrences being exploited in the foreseeable future. Economic or sub economic Cu–Ni–PGE sulphide ores are also found in some satellite intrusions of Bushveld lineage, notably the Uitkomst Complex (De Waal et al., 2001). Satellite intrusions of the Bushveld Complex (e.g. Molopo Farms,
315
Uitkomst, Roodekraal, Lindequesdrift, Rietfontein etc., Fig. 1) occur up to several hundred km in the periphery of the Bushveld (Anhaeusser, 2004; De Waal et al., 2001; Tankard et al., 1982 and Von Gruenewaldt et al., 1988) which previously led to the thinking that the Trompsburg intrusion constitutes part of the Bushveld igneous Complex (Maier et al., 2003). However, results of a SHRIMP U/Pb isotope study on zircons from two gabbroic samples of the Trompsburg intrusion (Maier et al., 2003) indicated a crystallisation age of 1915 ± 6 Ma. This is younger than the minimum U–Pb age of 2058.9 ± 0.8 Ma (Buick et al., 2001) on titanite from the Rustenburg Layered Suite of the Bushveld Complex, suggesting that the Trompsburg intrusion is not directly related to the Bushveld magmatic event. Hanson et al. (2004) reported on intraplate magmatism at 1928 to 1915 Ma along the western and northern edges of the Kaapvaal Craton. These authors suggest that the collective data from the (Fig. 1) Hartley basalts in the west (Cornell et al., 1998), the Trompsburg Complex in the south (Maier et al., 2003), the Moshaneng Dolerite to the north in Botswana and possibly even the Sebanga dykes on the Zimbabwe Craton (Hanson et al., 2004) indicate widespread Palaeoproterozoic intraplate magmatism, as also pointed out by Maier et al. (2003). This lends support to the proposed occurrence of a global 1.9 Ga superplume event (Condie et al., 2001). However, Anhaeusser (2004) does not rule
Fig. 1. Simplified map of the Kaapvaal Craton showing the location of the Bushveld Complex as well as some of the Bushveld related intrusions (numbered circles): Koringkoppies intrusion (1), Renosterhoekspruit intrusion (2), Golden Valley intrusion (3), Vogelstruisfontein intrusion (4), Kaffirskraal Complex (5), Losberg Complex (6), Roodekraal, Lindeques Drift and Rietfontein Complexes (7), Dwarsfontein intrusion (8), Uitkomst Complex (9), Moshaneng Complex (10) and Molopo Farms Complex (11). Evidence for enhanced igneous activity along the border of the Kaapvaal Craton at 1.9 Ga (Maier et al., 2003 and Hanson et al., 2004) is indicated by numbered squares: andesitic lavas from the Marydale Formation (1), Vioolsdrift pluton and associated metalavas in Namaqualand (2), Hartley Basalt (3), Moshaneng dolerite (4) and the Soutpansberg Basalts (5) (modified after Anhaeusser, 2004).
316
L.P. Mare´, J. Cole / Journal of African Earth Sciences 44 (2006) 314–330
out a closer link between the Trompsburg and Bushveld Complexes. A study by McCarthy and Cawthorn (1980) demonstrating that fractionation of Rb/Sr ratios, coupled with protracted crystallisation can produce changes in 87 Sr/86Sr ratio greater than analystical error, let Anhaeusser (2004) to suggest that the lesser Sr initial ratio (R0 = c. 0.704) calculated by Maier et al. (2003) for the Trompsburg Complex might be in error.
The current study does not try to resolve the issues surrounding the age of emplacement, but is an attempt to model the Complex three dimensionally. By incorporating all available geophysical information (physical properties, aeromagnetics and gravity), a three dimensional model of the Trompsburg Complex was created that satisfies the magnetic and gravity data (see Fig. 2).
Fig. 2. Positions of the boreholes shown on the regional gravity data (A) and aeromagnetic data (B).
L.P. Mare´, J. Cole / Journal of African Earth Sciences 44 (2006) 314–330
317
Fig. 3. Density distribution along the individual Trompsburg cores. Sample locations along cores indicated as diamonds. Lithological units colour coded as indicated.
318
L.P. Mare´, J. Cole / Journal of African Earth Sciences 44 (2006) 314–330
Fig. 4. Magnetic susceptibility distribution along the individual Trompsburg cores. Sample locations along cores indicated as diamonds. Lithological units colour coded as indicated.
L.P. Mare´, J. Cole / Journal of African Earth Sciences 44 (2006) 314–330
2. Geophysical reassessment 2.1. Physical property analyses Since the 1950s several research studies on the Trompsburg Complex were conducted (Ortlepp, 1959; Buchmann, 1960; Logan, 1979; Reynolds, 1979; Maier et al., 2003). From the original approximately 4 km core intersecting the Complex itself, a mere 16 kg of core remained. A total of 212 split and quartered core pieces, ranging between 30 and 80 mm long, were stored at the University of Pretoria. The samples were often not clearly labelled or the labelling has worn off, making them unsuitable for this study. The labelled samples indicated the core number and the depth in feet, and were on average 20–30 m apart. Unfortunately none of the granite intercepted by core TGX was left. Where possible, up to three smaller specimens were cut from each sample to increase the statistical reliability of the physical property analyses. The physical properties measured on each of the specimens at the Council for Geoscience included density, magnetic susceptibility, magnetic intensity (NRM), electrical resistivity as well as electrical charge and discharge (induced polarization). The standard specimen size required for magnetic property analysis is 12.87 cm3. However, due to the small core pieces, the cut specimens were in most cases undersized and volume corrections were applied to the results.
319
Even before looking at the sample composition, a signature could be recognized from the plotted physical property results (Figs. 3 and 4). This let to the decision to model the Complex using these physical property results as constraint. Initial groups were identified using the physical property signature as guide. Thereafter, Buchmann’s (1960) borehole logs as well as Ortlepp’s (1959) identified principal rock types (Table 1), were consulted to group the samples further into different ‘lithological’ units. We use the term ‘lithology’ loosely here since the identified units were selected primarily using the physical properties. Table 1 Principal rock types intersected in the Trompsburg boreholes (after Ortlepp, 1959) Lithology
Rock type
Karoo Dolerite Suite
Dolerite (sills and dykes)
Karoo Supergroup
Shale and Sandstone (Beaufort group) Shale (Ecca group) Tillite (Dwyka group)
Intrusive Complex
Granite Titaniferous Iron Ore Troctolite Olivine Gabbro Melanocratic Gabbro Gabbro and Anorthosite Hybrid Rocks
Transvaal Supergroup
Marble
Table 2 Mean physical property values (10% trimmed) for ‘lithological’ units Zones of associated mineralogy
N (n)
Density (kg/m3)
Magnetic susceptibility (·106) S.I.
Magnetic intensity (·103) A/m
Electrical resistivity (Xm)
Induced polarization (%)
Shale
2(5)
2617 [37] (2591–2643)
5 [5] (2–8)
1(1) 3(5)
2958 2643 [17] (2623–2653)
Anorthosite
25(46)
2833 [65] (2737–3034)
Gabbro
66(113)
2905 [105] (2701–3308)
Olivine Gabbro
22(40)
3114 [64] (3005–3267)
Trochtolite
11(20)
3034 [113] (2889–3287)
Mineralized Gabbro
26(41)
3209 [160] (2809–3859)
Magnetite Pegmatoid
2(2)
3796 [42] (3766–3826)
Magnetitite
11(20)
4322 [176] (3884–4557)
Hybrid Rock
7(13)
3251 [198] (2826–3414)
3492 1392 [2313] (34–4062) 1210 [3661] (2–18,509) 27,986 [28,611] (4–234,886) 51,295 [53,671] (1628–192,194) 76,412 [136,290] (2029–453,830) 49,463 [53,330] (2020–493,988) 864,097 [94,596] (797,207–930,987) 311,376 [328,646] (49,787–1,115,830) 2814 [53,330] (3–17,500)
1098 [263] (912–1284) 6569 6750 [1751] (4787–8151) 10,536 [6935] (698–29,483) 7 329 [2287] (893–13,782) 6448 [4034] (272–17,615) 4490 [4133] (88–12,000) 3222 [2820] (61–12,328) 182 [216] (29–335)
17 [8] (11–23)
Dolerite Granite
Marble
2(5)
2565 [26] (2546–2583)
347 [90] (283–410) 13,646 10,589 [9830] (422–20,043) 2210 [5437] (95–24,116) 31,570 [37,133] (164–197,427) 135,419 [56,908] (14,031–488,631) 104,909 [89,391] (10,005–273,163) 192,548 [181,485] (6995–1,122,712) 296,051 [25,729] (277,858–314,245) 1,200,929 [162,974] (953,070–1,386,517) 4700 [181,485] (217–24,646) 824 [350] (577–1072)
Dolomite
1(2)
2704
87,690
46,301
1897 [277] (1701–2093)
12 8 [3] (5–10) 11 [5] (5–26) 14 [4] (7–29) 16 [5] (7–24) 17 [5] (8–23) 18 [7] (8–33) 12 [7] (7–17)
61 [17] (36–94)
5 [3] (2–10)
9031 [2820] (1094–24,608) 1471 [816] (894–2048) 4478
10 [7] (8–14) 19 [8] (13–24) 15
The standard deviation is given in square brackets and the ranges in curved brackets. N represents the number of samples, while n indicates the number of specimens.
L.P. Mare´, J. Cole / Journal of African Earth Sciences 44 (2006) 314–330
tion of the whole Complex. Taking this into account, the three dimensional model presented in this paper is only one possible model for the Complex. Improved data coverage, by drilling more boreholes into the Complex, is needed to completely constrain the model. 2.2. Gravity data Regional gravity data have been collected over the whole of South Africa by the Council for Geoscience (previously Geological Survey of South Africa). A total of 1370 gravity points have been extracted from the database over the study area, indicating a measurement roughly every 9 km2. The gravity stations were mainly located along roads and the localities are indicated in Fig. 5. Gravity measurements were collected with La Coste and Romberg
25°15'E
25°30'E
25°45'E
26°E
26°15'E
Edenburg
Fauresmith
29°45'S
29°45'S
Reddersburg
30°S
Jagersfontein
30°S
Philippolis
30°15'S
Trompsburg
Springfontein
30°30'S
The physical property results displayed good contrast between these ‘lithological’ units (Table 2). No attempt was made to classify the samples in more detail. This allows for a simplified geological model for the purpose of potential field modelling. Table 2 is a summary of the means for these ‘lithological’ units trimmed by 10% to exclude anomalous values. This trimming was allowed to accommodate for the fact that the selected ‘lithological’ units might include thin layers of unrelated mineralogy not otherwise accommodated for. Figs. 3 and 4 display the density and magnetic susceptibility values along each core (TG1– TG6) with above mentioned ‘lithological’ units indicated in different colours. On petrographic evidence mainly from borehole TG6, Logan (1979) subdivided the intrusion into a Basal, Intermediate and Upper Zone. However, considering the limited data available, the additional fact that no individual unit could be traced from one borehole core to the next (Logan, 1979) as well as the suggestion by Reynolds (1979) that the Trompsburg Complex may well be a multiple intrusion, no attempt was made to use these zones in our model of the Complex. Table 3 is a summary of available physical property results for the Bushveld Complex as well as related results from the Karoo and Transvaal Supergroups from the South African Geophysical Atlas, Volume IV (Mare´ and Tabane, 2004). Since the Trompsburg data was only loosely grouped into ‘lithological’ units, any correlation between the two data sets (Tables 2 and 3) should be made with caution. For a general correlation we compared the magnetite-gabbro of the Upper Zone of the Rustenburg Layered Suite (RLS) with the mineralized gabbro in Table 2, the gabbro-norite of the Main Zone (RSL) with the gabbro in Table 2, and the norite–anothosite of the Critical Zone (RSL) with the anorthosite in Table 2. All the values from the Rustenburg Layered Suite fall within the associated data ranges in Table 2. The available samples from the Trompsburg Complex were all highly mineralized, but since they originate only from the north-western side of the Complex where the anomalously high magnetic signature (Fig. 2) prevails, these results cannot be a true reflec-
30°15'S
320
Bethulie
25°15'E
25°30'E
25°45'E
Gravity point Town
26°E
26°15'E
National route Main road Secondary road
Fig. 5. Distribution of gravity points over study area.
Table 3 Mean physical property values for the Bushveld Complex as well as related lithologies of the Karoo and Transvaal Supergroups calculated from the South African Geophysical Atlas, Volume IV (Mare´ and Tabane, 2004) Lithology
Rock type
Density (kg/m3)
Magnetic susceptibility (·106) SI
Magnetic intensity (·103) A/m
Induced polarization (%)
Karoo Dolerite Suite
Dolerite
2889
9190
16,608
12
Karoo Supergroup
Ecca Group
Shale
2606
197
7.3
5289
9
Bushveld Complex
Lebowa Granite Suite Rustenburg Layered Suite Upper Zone Main Zone Critical Zone
Granite
2625
930
487
148,767
9
Magnetite–Gabbro Gabbro–Norite Norite–Anorthosite
3115 2970 2852
71,515 1865 470
8578 2417 10
11,703 12,648 10,248
17 12 12
Malmani Subgroup
Dolomite/Marble
2785
190
3676
20,656
6
Transvaal Supergroup
28,848
Electrical resistivity (Xm)
L.P. Mare´, J. Cole / Journal of African Earth Sciences 44 (2006) 314–330
gravimeters and elevations were determined using microbarometers. Bouguer anomaly values were calculated assuming a mean crustal density of 2670 kg/m3. The measurements were tied to the International Gravity Standardisation Net values (Morelli et al., 1974) and were referred to the gravity formula based on the 1967 geodetic reference system (Moritz, 1968). A maximum error in the regional Bouguer anomaly value is calculated by combining an inaccuracy of 2 m (4 gravity units) in the barometrically determined elevation with a maximum error in the observed gravity of 2 g.u. and a positional error of 150 m in a north south direction (1 g.u.). This worst case error amounts to about 7 g.u., or roughly 1 mGal. More detailed information on the data acquisition and processing can be found in Stettler et al. (2001a,b,c,d). The maximum variation in topography in the study area is on the order of 560 m. Terrain corrections using an elevation grid obtained from the Surveyor General’s offices were applied to the data. The terrain corrections were calculated by dividing the area surrounding the observation station into grid segments bounded by concentric rings, and the influence of each grid segment at the observation station is calculated and added together (Stettler et al., 2001a,b,c,d). For grid segments further away than 1 km
321
from the observation point the expression for the vertical component of gravitational attraction is approximated by (Telford et al., 1990) dgT ðr; hÞ ¼ cqhfðro ri Þ þ ðr2i þ Dz2 Þ1=2 ðr2o þ Dz2 Þ1=2 g where c is the universal gravitational constant (6.672 · 108 dyne cm2/g2), q is the density contrast with respect to the laterally equivalent material (g/cm3), h is the sector angle (radians) Dz = jzs zaj with zs the station elevation (m) and za the grid segment elevation (m), ro is the outer boundary of the grid segment (m) and ri the inner boundary of the grid segment (m). The terrain correction is the sum of the contribution of all sectors XX DgT ¼ dgT ðr; hÞ r
h
For segments less than 1 km from the observation point the gravity formula for a rectangular parallelepiped as calculated by Stettler (1979) was used. The gravity data were gridded with a minimum curvature gridding algorithm (Briggs, 1974) using a grid interval of 1000 m. Fig. 2(A) shows the terrain corrected regional gravity data over the Trompsburg Complex.
Fig. 6. Survey blocks of the regional aeromagnetic coverage of South Africa. The position of the study area is indicated.
322
L.P. Mare´, J. Cole / Journal of African Earth Sciences 44 (2006) 314–330
Fig. 7. Flight lines along which aeromagnetic data was collected. The graph summarises the clearance that was maintained above the Earth’s surface during the surveys.
2.3. Airborne magnetic data Regional aeromagnetic data coverage exists for the whole of South Africa. The data were collected between the 1960s and 1980s under the auspices of the Geological Survey of South Africa in the blocks depicted in Fig. 6. The study area around the Trompsburg Complex falls in an area where four survey blocks meet. Data were collected along north–south directed flight lines spaced 1 km apart with tie lines at 10.5 km intervals (Fig. 7). A mean flying height of 114 m above the surface was maintained. The histogram in Fig. 7 and the statistical information of the flying height listed in Table 4 show that the heights varied quite substantially in some cases. The data were levelled by minimising the discrepancies between readings at the points where the flight- and tie-lines intersected. The magnetome-
Table 4 Statistical summary of the flying height maintained above the surface during the aeromagnetic survey Number of lines Minimum clearance (m) Maximum clearance (m) Mean clearance (m) Standard deviation
463 62 546 114.3629 63.0841
ter cycling time was 1 s and at an aircraft velocity of 250 km/h this translates into a magnetometer reading every 63 m. The magnetometer was towed in a bird below the aircraft. It is therefore clear that these types of surveys provide strictly regional coverage. The magnetic data were gridded using the same minimum curvature algorithm applied to the gravity data. A
L.P. Mare´, J. Cole / Journal of African Earth Sciences 44 (2006) 314–330
gridding interval of 250 m was used for this data set. Minimum curvature algorithms are well suited to potential field data (Briggs, 1974). The data were corrected for the 1 July 1975 magnetic epoch, closest to when they were collected. Fig. 2(B) shows a sun shaded image of the total field aeromagnetic data.
323
3. Integrated interpretation and 3D model The Trompsburg anomaly is situated near the southern boundary of the Kaapvaal Craton. Ortlepp (1959) suggested that the Complex intruded into dolomites of the Transvaal Supergroup, and borehole TG5 confirmed that the sur-
Fig. 8. Regional gravity (A) and aeromagnetic data (B) for the area around the Trompsburg Complex showing the approximate location of the ColesbergMafikeng lineament.
324
L.P. Mare´, J. Cole / Journal of African Earth Sciences 44 (2006) 314–330
rounding material consisted of dolomite. The density of the dolomite is 2704 kg/m3 and is responsible for the gravity lows surrounding the Trompsburg Complex. The anomaly associated with the Complex is located on a NNE-SSW trending gravity high that is in the order of 15 mGal higher than the surrounding data. This anomaly runs parallel to the Colesberg-Mafikeng lineament located roughly 70 km to the West. The Colesberg-Mafikeng lineament is defined by high
gravity and magnetic values and has been described by De Wit and Tinker (2004) as the boundary between two sub domains of the Kaapvaal Craton, namely the younger (
