Interpretation of Project MAGNET Aerornagnetic Profiles Across Africa

Geophys. J. R. astr. SOC. (1976) 44, 203-228.

Interpretation of Project MAGNET Aerornagnetic Profiles Across Africa

A. G. Green*

(Received June 10 1975)t

Summary

Long aeromagnetic profiles recorded across central Africa are analysed in terms of the tectonics and geology. The short wavelength magnetic anomalies (6 km < widths of anomalies < 60 km) are correlated with the local geology while the long wavelength anomalies (widths of anomalies > 60 km) are compared with the regional tectonics and geology. Important features observed in the magnetic data are (a) the ' magnetic quiet zone ' across the Mozambique Orogenic Belt, (b) the low magnetic signature of the East African Rift, (c) the double-peaked anomalies across the Lake Malawi-Shire Rifts, (d) the high amplitude long and short wavelength magnetic anomalies recorded across a vast region of western central Africa which includes the Kasai Shield, the Angola Basement and the Kalahari and Congo Sedimentary Basins to the south of latitude 4" S, and (e) the extraordinary Bangui and Kibalian Magnetic Anomalies.

The qualitative interpretation is supplemented with an average depth to magnetic body analysis using the Statistical Models Method (Spector & Grant; Green). This analysis reveals two ensembles of magnetic bodies for most of the lengths of profile which recorded significant magnetic anomalies (> 15 gammas). The shallow ensemble of magnetic bodies are at an average depth of 0.2 km beneath the exposed Precambrian Shield, while across the large sedimentary basins they are significantly deeper at 0-8 km. The depths to the deep ensembles of magnetic bodies covers a wide range from 4 to 30 km.

Introduction In recent years aeromagnetic data have been useful in outlining the major tectonic

provinces in North America (Zietz et af. 1966; Hall 1968; MacIaren & Charbonneau 1968; Caner 1969; Douglas 1970; Haines, Hannaford & Riddihough 1971; Riddihough Haines & Hannaford 1973), USSR (Simonenko 1969, 1972; Krutikhovskaya, Pas- hkevich & Simonenko 1973) and Western Europe (Hahn & Zitzmann 1969; Hall & Dagley 1970). These studies have yielded information about the lithology of the basement as well as the areal and depth extent of magnetic bodies. The present study uses Project MAGNET profiles in a first attempt to analyse transcontinental aeromagnetic data across Africa.

The profile flight paths shown in Fig. 1 were chosen with the following objectives: * Present address: Department of Earth Sciences, University of Manitoba, Winnipeg, Canada. '- Received in original form 1975 February 13.

203

Downloaded from https://academic.oup.com/gji/article-abstract/44/1/203/664304by gueston 17 February 2018

204 A. G. Green

loo 20" 30* 40"

FIG. 1. Project MAGNET flight lines across Africa. Eight approximately transcontinental profiles may be obtained from flights T211-522, T213-517, T225, T203, T220, T219a, T202-T201, T204T205. There are in addition four shorter

profiles, 506, 508, 511 and T208.

(i) To relate the magnetic data to the surface geology and tectonics. The eight transcontinental profiles obtained from flights T211-522, T213-517, T225, T203, T220, T219a, T202-201, T204-205 (Fig. 1 and Fig. 3) are the principal data for this.

(ii) To look at the effects, if any, of the Karroo and East African Rift Systems on the magnetic pattern of Precambrian Africa. In addition to the transcontinental profiles, four shorter profiles (250-1000 km. length) 506, 508, 51 1 and T208 are analysed.

(iii) Obtain information about the depth distribution of magnetic bodies.

As a first step the geologic and tectonic framework of Africa is reviewed. To save repetition the maps that have been referenced are given in Appendix B. The Project MAGNET profiles are presented and their relationship with the geology and tectonics is outlined. Average depth analysis using the Statistical Models Method (Spector & Grant 1970; Green 1972) is then applied to lengths of profile which cross individual tectonic units.

Tectonic and geologic framework of Central Africa The Precambrian basement rocks of central Africa (Fig. 2) may be described in

terms of four broad age divisions. The rocks of the oldest division, the Precambrian D (Tectonic Map of Africa, 1968), have ages greater than 2100My and are mostly found on the shields of Rhodesia, Tanzania (Tanganyika Shield), Zaire (Kasai Shield) and Cameroun (Cahen & Snelling 1966; Clifford 1968, 1970). Granites,


Project MAGNET aeromagnetic profiles 205

granodiorites, gneisses, migmatites, charnockites, other high grade metamorphic rocks and some basic to ultrabasic intrusives are predominant on the shields. In addition there are large areas, particularly on the Tanganyika and Rhodesia Shields, which have greenstone belts wedged between large granite bodies (Quennell, McKinlay & Aitken Macgregor 1947, 1951a, b). High grade metamorphic rocks of this age are also observed in parts of the North Congo Shield and the Uganda Basement (Cahen & Snelling 1966; Hepworth & Macdonald 1966; Spooner, Hepworth & Fairbairn 1970; Mestraud 1971).

The principal outcrops of the next oldest division, the Precambrian C (1600-2100 My), occur in three orogenic belts, the Kibali-Buganda-Toro, the Ubendian-Rusizian and the Limpopo Belts (Fig. 2). The first-named belt is a metasedimentary feature with some granite intrusives while the latter belts are characterized by mostly high grade metamorphic rocks (Cahen & Snelling 1966; Clifford 1968, 1970). Other rocks of this age are found in the crystalline basement of Angola, the Upper Mayumbian of Angola and Gabon, various low grade metasedimentary formations on the north and east of the Rhodesia Shield and two small orogenic belts on the Kasai Shield. Parts of the Zambia Shield, known as the Tumbide Orogenic Belt, may have been mobile at this time (Ackerman 1950; Ackerman & Foster 1960; Vail, Snelling & Rex 1968).

CENTRAL AFRICAN 30'

TECTON i CS

Precambrian D(>2100my) Unsure

Precambrian D with K Ar ages in Precambrian A

Precambrian C(1600 to E3 2100 m y 1 Precambrian A(unmetamorphosed a Precombrtan B(1000 to plat form sediments) 1600 m y 1 Precambrian A(metamorphosed p la t fo rm sediments)

Precambrian A ( 6 0 0 to1OOOmy) rn (Mozombique type)

D Damaran Belt GD Great Dyke I Irumtde Belt K Korroo KA Katangon Belt KB Kibaran Belt KC Kibalton Belt KA Karogwe Ankolian K B Kalahari Basin K S Kosai Shield LA Ltndtan LC Limpopo Belt L B Lik i Bembian LV Lake Victoria M Z Mozambique Bel t NCS North Congo Shield 0 Oubonguien Ot Otavt R Rusizian Bel t RS Rhodesian Shield RV Ri f t Volcanics SO Sembe Ouesso TS Tanganyika Shield U Ubendion Belt UB Ugondo Basement Um Umkondo Frontier

W West Congoltan ZS Zambian Shield ZM Zambesi Mozambique

Goirezi

Bel t

FIG. 2. Tectonic map of central Africa. Taken from the Tectonic Map of Africa (1968) with modifications based on more recent publications.


206 A. G. Green

Most of the rocks of the next division, the Precambrian B (100&1600My), belong to two near parallel orogenic belts in the south and east of the region (Fig. 2), the Irumide and the Kibaran-Karagwe-Ankolean Belts (Cahen & Snelling 1966; Cahen 1970; Vail et al. 1968; Pallister 1971). These orogenic belts consist of highly folded, but mostly low grade metasedimentary rocks. Within the Irumide Belt there are outcrops of high grade metamorphic rocks which are interpreted as 'older granulites ' reworked during the Irumide event.

The vast regions of central Africa which were active in the Precambrian A (600- 1000 My) may be split into two subdivisions based on the geology (Fig. 2). The first subdivision consists of relic basement rocks from the ancient shields interspersed with high grade metamorphic rocks of undefined origin (Sanders 1965; Vail 1966; Hepworth, Kennerly & Shackelton 1967; Hepworth & Kennerly 1970; Hepworth 1972). Regions of this type include major parts of the enormous Mozambique Orogenic Belt (including the Zambezi Belt), the Uganda Basement and possibly the North Congo Shield. The rocks in these areas are surprisingly uniform with schists, gneisses, migmatites, quartzites, charnockites and some granites. Within the Mozambique Belt large granite bodies are common only in Malawi and Zambia (see Bloomfield 1970). In fact Pallister (1971) has remarked on the low number of large granite masses in other parts of the belt and this may be important in the interpretation of the magnetic profiles.

The second subdivision of Precambrian A rocks consists of highly-folded sedimentary and volcanic sequences which range from unmetamorphosed to highly metamorphosed. Orogenic belts of this type include the economically important Katangan Belt, the Damaran Belt and the West Congolian Belt (Cahen & Snelling 1966; Clifford 1967, 1968, 1970). There are also a number of folded sedimentary formations of this age around the perimeter of the Congo Basin (Fig. 2).

Central Africa was relatively quiescent from the Palaeozoic to the start of the Karroo activity (Carboniferous to Jurassic), when major troughs were initiated by rifting and subsidence (Cox 1970, 1972). After long periods of tectonic activity and massive sedimentation the end of the Karroo was marked by extremely widespread basaltic lava flows (Stormberg Volcanics). These lavas are confined to south of approximately 10" S latitude where in parts they extend almost the width of the continent.

A period of between 30 and 60My separated the final phase of the Stormberg lavas and the beginning of a more minor, but none the less important phase of rifting and volcanicity in eastern Africa (Vail 1968; Woolley & Garson 1970; Flores 1970; Pallister 1971). This activity occurred mainly in Malawi (e.g. the Lake Malawi- Shire-Urema rifts, the Chilwa and Lupata alkaline igneous rocks) and Mozambique where volcanic rocks of this age (Lower Cretaceous) are believed to underlie the Tertiary and later sediments of the ' Mozambique Bulge '.

Tertiary and Quaternary tectonic and magmatic activity have been almost entirely confined to the East African Rift and related structures, while other regions were the sites of continental deposition (e.g. the Kalahari and Congo Basins). The history of the East African Rift is relatively well documented and is known to have evolved by major uplifting, rifting (classical graben structures and splay faulting) and extensive alkaline and basaltic volcanicity (Baker et al. 1971; Baker, Mohr & Williams 1972, King 1970; King & Chapman 1972; King, Le Bas & Sutherland 1972; Fairhead, Mitchell & Williams (1972).

Project MAGNET data and its reduction All the magnetic intensity profiles used in this study (Figs 1 and 3) were obtained

by the United States Naval Oceanographic Office as part of their ' Airborne Geomag- netic Survey Project MAGNET ' programme. The primary purpose of this very large



I

CB 8

I 0 KB ' ' RS Lc K

E

l 25.E

FIG. 3. Project MAGNET profile data for Africa. The tectonic units sampled by the profiles are outlined beneath each profile. The key to these units is given in Fig. 2. The lengths of profile defined by the outward pointing arrows cross volcanic rocks of the East African Rift, while the inward pointing arrows represent the position of Karroo volcanic rocks. G.D. shows the position of the Great Dyke of Rhodesia.

scale survey was to chart more accurately the magnetic field of the Earth. It is fortunate that the data shows sufficient detail to be useful for regional studies. The measurements of the total magnetic intensity were made with a Naval Ordnance Laboratory Vector Airborne Magnetometer (NOL report 1187, 1957) and were recorded on continuous strip charts. Microfilm copies of these data were used in this study. Navigational fixes are given every 5 min of flight. The probable maximum error of the absolute total magnetic intensity values and the navigational fixes have been estimated to be 2 25 gammas and rf: 9 km, respectively (G. R. Lorentzen, private communication).

From the navigational fixes the distances travelled per 5 min of flight were computed and using a linear interpolation scheme the data were digitized approximately every 3 km of track. This interval (20-40 s of flight) was sufficient to outline practically every anomaly encountered on the profiles. Anomalies of width less than 3 km are


208 A. G. Green

mostly filtered due to the high flight altitudes ( N 3 km). The low wavenumber geomagnetic field (Alldredge, Van Voorhis & Davies 1963; Bullard 1967) was removed from the profiles by subtracting the appropriate value of the International Geomagnetic Reference Field (IAGA 1969) from each of the total intensity values. The long flight time of some profiles required that there be consideration of the effect of the magnetic diurnal variation (Chapman 8c Bartels 1945). If the amplitude of the diurnal variation exceeded 15 gammas over the flight time of any profile then the effect was removed by using the magnetic intensity records obtained from the Addis Ababa, Niarobi and Lwiro observatories.

After removing the geomagnetic field and the magnetic diurnal variation a back- ground trend or constant value was observed on some of the profiles. It was therefore necessary to remove an additional linear gradient term from these profiles before the wavenumber domain analysis (see section on depth estimates) was performed.

Observations

The Project MAGNET profiles and the major tectonic units crossed are show in Fig. 3, A comparison of this figure with the African Tectonic Map (Fig. 2) demon- strates that most of the tectonic units have been sampled by at least one profile. In a more detailed analysis (Green 1973) the relationships of the topography and the finer detailed geology with the magnetic data were studied. From this previous work it may be stated:

(i) Except for a section of profile T205, there are no obvious correlations of the magnetic data with the topography.

(ii) Depending on the profile, between 30 and 70 per cent of the high amplitude short wavelength magnetic anomalies may be correlated with ' expected magnetic rocks '. Failure to obtain a higher correlation may be due to (a) inaccurate navigation, (b) insufficient geological mapping, and (c) subsurface magnetic bodies.

In the following an attempt is made to summarize the important magnetic characteristics of each tectonic unit. Most of the units are only crossed by one or two profiles so this summary is necessarily tentative. However, similar studies across North America proved to be essentially correct when more data were available.

For convenience the profiles in Fig. 3 will be referenced by the profile number (e.g. 517).

Precambrian D The magnetic anomaly patterns of the four oldest shields are broadly similar

(Figs 2-3). They each contain large regions of low magnetic relief and, except for possibly the Tanganyika Shield, large regions of high magnetic relief. Magnetic lows are observed across the gneissic and granitic rocks of all shields and across the greenstone belts of Tanzania and Rhodesia. The magnetic relief remains low even in the Rhodesian greenstone belts extensively intruded by Precambrian ultrabasic dykes. In particular, note the relatively low amplitude anomalies across the Great Dyke (GD on 517, see Worst 1960; Bichan 1970 for details of this enormous dyke).

A wide variety of rocks are associated with the high amplitude magnetic anomalies. On and around the Rhodesian Shield (517, 522) the widespread Karroo volcanics and Precambrian dyke swarms (away from the greenstones) are the principle magnetic rocks. Within the Kasai Shield (T220, T203) the major sources of the anomalies are extensive basic to ultrabasic intrusives (dolerite, serpentinite, pyroxenite) and some massive iron ore deposits (haematite-magnetite bodies with 40-60 per cent Fe content, Cahen & Lepersonne 1967). Evrard, who has conducted a ground magnetic survey



over parts of the Kasai Shield, reports very high amplitude anomalies (up to 16 000 gammas) across the ore deposits (private communication, 1972). The bodies causing the negative anomalies on the Cameroun Shield (T202) are not clearly defined by the surface geology. The pyroxene gneisses shown in this locality could be highly magnetic. Alternatively, a number of small dolerite bodies near to the profile may indicate a subsurface igneous origin to the anomalies. On the Tanganyika Shield (T208, T219a, 505) the only sizable anomaly (T219a) is directly related to a large granodiorite body.

Prrcantbrian C The lengths of profile which cross Precambrian C tectonic units are dominated by

high amplitude short and long wavelength magnetic anomalies. In the Limpopo Belt (522) these anomalies are a continuation of the Rhodesian Shield anomalies and are related to similar Precambrian dyke swarms. Across the northwestern parts of the Ubendian-Rusizian Belt (west of ‘ h ’ on T220) the shorter wavelength magnetic relief may be confidently correlated (a 1 : 1 correlation is possible) with large granite masses, basic intrusives and a large outcrop of ‘ ancient acid lavas ’. In the vicinity of ‘ h ’ (T220) the Precambrian basement is interrupted by Pleistocene to Recent lavas. The large negative anomaly at ‘ h ’ marks the Rungwe volcanic cone. Further east, where the Ubendian terrain is partly altered due to later orogenies, the amplitudes of anomalies are again continuously high.

Very long wavelength anomalies are the rule for the Kibali-Buganda-Tor0 Belt (T201, T204, 508). Some of the low amplitude shorter wavelength anomalies in the Buganda-Tor0 section may be correlated with outcrops of granite, but in general the high amplitude long wavelength anomalies of this belt cannot be caused by the surface metasediments. The steep sided, plateau like anomalies of the Buganda-Tor0 Belt are continuous from its western margin across Lake Victoria almost to its eastern shore (see flight paths of T204 and 508), suggesting a Buganda-Tor0 type basement for that part of the Lake. The long wavelength anomaly across the Kibalian (Kibalian Anomaly) is discussed later with the Bangui Anomaly.

On the Angola basement, the undifferentiated granodiorite-granite rocks could be sufficiently magnetic to produce most of the high amplitude anomalies recorded there (T203, T225). If so, then the magnetic data could be useful for differentiating the more basic rocks. An alternative source of deeper igneous bodies is indicated as a possibility by the frequent occurrence of ‘ Post Permian Dolerites ’ in the region of the profiles; De Sousa Machado in Haughton (1963) correlates these dolerites with the Karroo magmatism.

Precambrian B For reasons apparent later the Irumide Belt is discussed with the Katangan Belt

under the Precambrian A heading. Only the two profiles which cross the Kibaran- Karagwe-Ankolean Belt are examined here (T219a7 T220). Both these lengths of profile reveal a very low magnetic relief, which is somewhat surprising as there are a wide variety of igneous rocks within this belt, including the Lake Kivu lava field (rift volcanics on T219a). The detailed geology is still relatively unknown beneath large parts of the profiles so the rocks flown over could prove to be mostly sediments and and low grade metasediments.

Precambrian A There is no single magnetic pattern associated with the Precambrian A tectonic

units, but there are some peculiar magnetic characteristics within the individual units. One of the most important of these is the ‘ magnetic quiet zone ’ across the Mozambique Orogenic Belt. Except for an area of Malawi and Zambia outlined in Fig. 4 the low magnetic relief (amplitudes mostly less than 100 gammas) is continuous

N


c! 0

Ban

gui

Ano

mal

y K

ibal

ian

Ano

mal

y

T208

---J---

-T20

3

Moz

ambi

que

Oro

gent

c B

elt

rMo

zam

biq

ue

Oro

geni

c B

elt

522

;+(in

clud

ing

Mol

awi

and

Zam

bia)

3,

: : D

oubl

e pe

aked

ano

mal

y as

soci

ated

with

:'

:

tect

onic

act

ivity

nea

r th

e La

ke M

alaw

i R

ift

25"E

t 500

Y

1

20

0 K

m

a- b

Kal

ohar

i/Con

go

Bas

ins

Mag

netic

Hig

h

?

0

FIG. 4. P

roje

ct M

AG

NE

T p

rofil

es w

ith th

e im

port

ant m

agne

tic fe

atur

es o

utlin

ed.


Project MAGNET aeromagnetic profiles 21 1

across the belt from latitude 14"s. (T225) to latitude 3"N (T205). This is quite extraordinary when all the profiles are viewed together, as there are few regions in Africa where such low magnetic relief is recorded. The most common rocks in the ' quiet zone ' are schists, amphibolites, quartzites, gneisses, migmatites, charnockites and some granulites. From the magnetic data it would seem that these rocks are the low magnetic, perhaps less basic varieties.

In the exceptional areas of Malawi and Zambia to the east of Lake Malawi, the igneous rocks of Precambrian A age and younger (Karroo and Cretaceous) may be the sources of the high amplitude anomalies observed there. It is not generally possible to correlate specific outcrops with the anomalies, but the large plutonic masses in these areas are certainly atypical of the Mozambique Belt.

The length of profile which crosses the North Congo Shield and adjacent sedimentary formations contains the most spectacular magnetic anomaly recorded anywhere in the world (Bangui Anomaly). Although anomalies of similar amplitude have often been recorded (e.g. on profile T220) none have had the lateral extent of the Bangui Anomaly. The 750 km wide anomaly (between a and b on T204) has an amplitude of 1200 gammas (peak to trough) at an altitude of 2-4 km and 900 gammas at the upward continued height of 22.4 km (Fig. 5). Even at satellite altitudes ( N 550 km) the anomaly is still prominent; it was chosen from the Pogo and Cosmos 49 satellite magnetic data for special investigation (Regan, Davis & Cain 1973, see also Benkova, Dolginov & Simonenko 1973). The origin of this anomaly is examined later, it is only pertinent to state that it is unrelated to the surface geology of sediments and metamorphic rocks.

To the west of the Bangui Anomaly the magnetic relief is mostly low as profile T204 crosses high grade metamorphic rocks. The single negative anomaly does not correlate with any obvious change in the surface geology and may therefore be caused by a subsurface body. A similar explanation may be required for the high amplitude negative anomaly recorded by profile T208 across the gneisses and migmatites of the Uganda Basement (see depth estimate in Fig. 7(b)). The two other profiles which cross the Uganda Basement are affected by the structural discontinuities in the region of the Lake Albert-Semliki Rift (margin of the Uganda Basement with the Kibalian Orogenic Belt on T204 and T201).

Although the African Tectonic Map (Fig. 2) delineates a boundary between the Irumide and Katangan Orogenic Belts, the detailed geology and magnetic data (T203, T225) suggest that no positive demarcation is possible in the regions sampled by the two profiles. The geology crossed by the profiles appears to be similar, with Katangan /Irumide metasediments (including ironstones) and metavolcanics inter- mingled with the higher grade metamorphic rocks of the Zambia Shield. Strangely the magnetic patterns of the two profiles are completely different. On profile T225 the magnetic relief is very low, except for the high amplitude negative anomaly at the tectonic boundary probably caused by a large granite mass near to the flight path. In contrast profile T203 records high amplitude long and short wavelength anomalies. An increase in the occurrence of Katangan ironstones and /or igneous bodies beneath the latter profile would explain the data.

The profiles which cross the Damaran Orogenic Belt sample the lower grade metamorphic rocks of the Otavi facies (T2I 1, T213). Both reveal a uniform pattern of low magnetic relief which is interrupted only when the profiles cross Karroo volcanic rocks.

Late Palaeozoic to Recent There is considerable magnetic relief associated with the Karroo basaltic lavas.

Whenever these lavas outcrop beneath a profile, high amplitude anomalies are recorded (T203, T211, T213, 517, 522). High magnetic relief is also found across the


212 A. G. Green

T204 ------- T200 i I 4 0 5 1-

I T201

I 51 1 I

I b I

I I

I

- 1 - - 506 I - - 508 T219a- I - T219a

- T220-i -T220

T 2 0 3 \ - T203

T225-' T225

I I------- 517 T213- ---- I

I 3 O o Y T211 -- 522 200Km

25'E I E W

FIG. 5. Project MAGNET profiles upward continued by 20 km in the manner described by Green (1972). a-b describes the Kalahari Congo Basin magnetic high

(see Fig. 4).

Kalahari Basin and the Congo Basin south of approximately 4" S (T203, T211, T213, T219a, T220, T225, see also the vertical component anomaly map of Evrard et al. 1959). North of 4" S the magnetic relief across the Congo Basin is very low (T202, T219a, Evrard, Jones & Mathieu 1959). This magnetic data may be useful for de- termining the type of basement beneath large parts of the sedimentary basins (see discussion later).

In contrast to the Karroo igneous complexes, the East African Rift volcanic rocks have very little magnetic expression (T201, T205, T219a, T220, 506, 508, 511). Sizable magnetic anomalies are observed only when a profile crosses a volcanic cone such as Ngorongoro (508) and Rungwe (T220).

Depth estimates It is readily observed that each of the profiles in Fig. 3 are composed of over-

lapping long and short wavelength magnetic anomalies. From these profiles, the rare isolated anomaly may be interpreted by the usual modelling technique with the geology and magnetic moment vector of the underlying rocks as constraints. This method is practically useless for analysing most of the profiles. The stochastic nature of the data demands a statistical approach. The technique used here is the profile version of the Statistical Models Method (Green 1972; Spector & Grant 1970).

The basic assumptions of this method are (i) The majority of magnetic bodies causing the anomalies along a given length

of profile belong to one or more ensembles of 'two-dimensional prisms' which have strikes perpendicular to the flight path.


Project MAGNET aeromagnetic pro6les 213

Cii) Within each ensemble, the magnitudes of the magnetizations, the declinations and inclinations of the magnetic moment and geomagnetic field vectors, the widths, the thicknesses and the depths of the magnetic bodies are independent of each other and rectangularly distributed about mean values. The extent to which these parameters are allowed to vary from the mean values is discussed by Green (1972).

With these assumptions it may be shown that the logarithmic ensemble average of the energy spectrum is given by:

In(E(v)) = 1n(4n2 K 2 ) +ln(R,) +In(RK) +ln(S(v)) +ln(C(v)) +In(P(v)) (1)

where u is the angular wavenumber equivalent to the space variable along the

length of profile,

( ) denotes ensemble average,

(4x2 K 2 ) contains the magnetic moment/unit volume information,

(RK) average inclination and declination of the magnetic moment vector term,

( RT) average inclination and declination of the geomagnetic field vector term,

<S(v)) average width term,

(C(v)) average thickness term,

(P(v)> average depth term.

Only the last three terms affect the gradient of the logarithmic energy spectrum versus wavenumber curve. If an estimate of the average width of the bodies can be obtained (Green 1972), the final gradient (width effect removed) will only depend on the average depth and thickness of the bodies. For sufficient lengths of prose the depth and thickness terms combine to produce a peak in the low wavenumber part of the spectrum (Spector & Grant 1970; Green 1972). Away from the peak towards the higher wavenumbers, the average depth term

ln(P(v)> N -2Ev (2)

where !i is the average depth, dominates the logarithmic spectrum. Therefore, the gradient of the linear part of the spectrum (least-squares fit of a straight line is adequate for this) together with equation (2) is sufficient to yield an estimate of the average depth to the ensemble of magnetic bodies. The practical procedure for computing the logarithmic energy spectrum and other details are outlined by Green (1972), while some relevant examples of this technique applied to simulated data are given in Appendix A.

The tectonic units shown beneath the profiles (Fig. 3) d e h e the separate lengths of data which have been depth analysed. By analysing the magnetic pattern of individual tectonic units, there is a greater probability that the magnetic sources will belong to only one or two ensembles. In a few cases where the magnetic boundary is not well defined or does not coincide with the tectonic boundary, two or more units have been analysed together. Lengths of profile with large variations of the magnetic pattern within a single tectonic unit were further sectioned; average depth estimates from highly non-stationary sets of data will have little meaning.


214 A. G. Green

2 - -

0 -

2 -

0 - (d 1 ( C )

0.0Km ( 2 0 Krn)

ko (el

1 . . 5 * ' , ' I I " ,

0 .02 .04 06 .OO .I0 .12 14

4 -

0 2

( I to j )

overage depths 19.6 Km (20.7 K i n ) 0 5 K r n ( l . 6 K r n )

iO, t

L . . , , I I " ~ ' ' ' ' ' '

0 .02 .04 .06 .08 .I0 .l2 34 16

FIG. 6. Energy spectra of profiles T204 and T220. The letters in brackets refer to the respective lengths of profile in Fig. 3. The crosses indicate the lengths of spectra used to determine the average depth of the deep ensemble of magnetic bodies, the circles indicate the lengths used for the shallow ensembles. All spectra have had the effect of width removed; the gap in the spectra after the second cross is due to the different width factors removed from the ' deep ' and ' shallow ' parts of the curves. The average depths are relative to the average ground level; the depths in

the brackets are relative to the flight height.

Examples of energy spectra from profiles T220 and T204 are shown in Fig. 6. These are typical of the spectra computed from the Project MAGNET data. Note the similarity of these with the spectra of the simulated data in Appendix A. The dominant, linearly decaying depth term is clearly seen in all the figures. Except for the Bangui Anomaly (Fig. 6(d)) the examples reveal two ensemble of magnetic bodies, a deep and a shallow ensemble. Two distinct sets of magnetic bodies were common in the African



data. In a few cases where the spectra revealed only a deep ensemble of magnetic bodies, the energy from the shallow ensemble (if it existed) was probably below the noise level of the data.

The distribution of average depths obtained from the high wavenumber parts of the spectra are shown in Fig. 7(a). This part of the curve always yields shallow average depths. All the depths appear to be reasonable; the few negative estimates are well within the 0.3km probable error of the aircraft’s average terrain clearance. As

zoo 40° 1 I I

FIG. 7. (a) Distribution of average depths to the shallow ensemble of magnetic bodies. These depths are obtained from the high wavenumber part of the spectrum. (b) Distribution of average depths to the ‘ deep ’ ensemble of magnetic bodies. These depths are obtained from the low wavenumber part of the spectrum. Note the similarity of the depths from near neighbouring profiles which cross the same

tectonic unit.


216 A. G. Green

expected, the mean depth to the magnetic bodies beneath the thick sediments of the Kalahari /Congo Basins (0.8 km with a standard deviation of 0.3 km) is significantly higher than the mean depth obtained from all other regions (0.2 km with a standard deviation of 0.2km).

The distribution of average depths obtained from the low wavenumber part of the spectrum is shown in Fig. 7(b). This part of the curve generally yields the average depth to the deeper ensemble of magnetic bodies. Average depths less than 4 km (6 out of 32 estimates) were obtained from spectra which yielded only a single depth (except for the section of profile T220 which crosses near to the Kasai ironstone formations, Fig. 6(a)) and are considered to represent either the average depth of bodies extending from near to the surface to greater depths or the average depths to near surface ' broad ' magnetic bodies. The following discussion is principally con- cerned with average depths greater than 4km.

The similarity of the average depth estimates from near neighbouring proiles gives some confidence in both the spectral method of depth determination and the consistency of the data (see Fig. 7(a) and (b)). Only the depths from the Mozambique Orogenic Belt and the Kalahari /Congo Basins show any indication of a long distant trend or pattern (Fig. 7(b)). All the other tectonic units are insufficiently sampled for any conclusions about trends or patterns within the given units. That there is little or no correlation of the average depth estimates between the tectonic units may be an important clue to the geological evolution of the African continent.

It seems worthwhile to compare these values to those obtained from other parts of the world (Table 1, Appendix B). The average depths in Fig. 7(b) are likely to be maximum depths so the values obtained from the African continent are not significantly different from other regions. Some of the authors referenced in Table 1 (Vacquier & Affleck 1941; McClure 1963; Bhattacharyya & Morley 1965) postulate that the base of their models represents the Curie point isotherm. Although undula- tions in the Curie point isotherm may well affect the energy spectra, it is not possible to give an estimate of its depth from the present study. It is only reasonable to state that magnetic bodies may exist to a depth of 23 km and possibly to a depth of 30 km (Fig. 7(b)).

Discussion

Project MAGNET data. The following is an attempt to explain some of the more important features of the

The Mozambique Belt ' magnetic quiet zone ' Some points relevant to the interpretation of the ' magnetic quiet zone ' are:

(a) The geology of the complete length of the Mozambique Belt (except Malawi and Zambia) is remarkably uniform. High grade, low ferromagnesian metamorphic rocks predominate at the surface. A particularly notable feature is the low number of large granite bodies outside of Malawi and Zambia.

(b) Gravity data reveal large anomalies across the Tanganyika Shield which are abruptly terminated at the boundary with the Mozambique Belt (Darracott 1972). Within the Belt the gravity field is practically featureless except for the negative regional term described by Girdler & Sowerbutts (1970).

(c) The East African Rift has affected large portions of the Belt and the high temperatures in the Rift (Searle 1970) will certainly have caused a major upwelling of the Curie point isotherm in those areas.

(d) Geomagnetic depth sounding has shown a well-defined conductivity anomaly beneath the eastern rift (Banks & Ottey 1974).



(e) The depths to the magnetic sources in the Mozambique Belt generally increase from north to south (Fig. 7(b)) which corresponds to the decreasing effects of the Rift.

It is worthwhile to use the three proposed explanations of the ' magnetic quiet zone ' within the North American Cordillera as a basis for discussion here (Pakiser & Zietz 1965; Zietz & King 1966; Zietz et al. 1966; Caner 1969; Haines et al. 1971; Berry et al. 1971). These are (after Caner 1969):

(i) The basement rocks are more silicic (less ferromagnesian minerals) in the ' quiet zone ' than other regions.

(ii) The surface to deeper structures in the ' quiet zone ' have been broken up by orogeny.

(iii) The curie point isotherm in the ' quiet zone is near to the surface due to the heating effects of recent tectonic activity.

From points (a)-(e) it would seem that all three explanations have some relevance to the Mozambique Belt ' magnetic quiet zone '. The presumed low ferromagnesian content of the surface rocks (point (a)) is adequate to explain the low amplitildes of the short wavelength magnetic relief. A natural depth extension of these surface rocks would also account for the low amplitude of the long wavelength components. Point (b) hints that the Mozambique Orogeny may have broken up the original structure to such an extent that the geophysical fields have been ' homogenized '. Finally (c)-(e) indicate that the uprising of the Curie point isotherm has affected at least those regions of the Belt near to the Rift.

East African Rvt and related structures One of the surprising results of this study has been the very low magnetic relief

recorded across the East African Rift volcanic rocks (Fig. 3). In some of the profiles an increase in the amplitude of anomalies may be significant, but except for the Rungwe and Ngorongoro volcanoes. the evidence is not good. Ground magnetic surveys (Whitham & Hoge 1961; Searle 1969, 1970; Sowerbutts 1972) and a low altitude aeromagnetic survey (Wohlenberg & Bhatt 1972) have also revealed a generally low magnetic relief across the Rift volcanics. In the ground magnetic data there are some high amplitude short wavelength anomalies but these are filtered out at the altitudes of the Project MAGNET profiles.

From the lack of ' magnetic lineations ' it seems reasonable to postulate that the majority of igneous rocks have been extruded from non-linear dykes, small cones and large central volcanoes; Lippard (1973, private communication) reports that more central volcanoes are being discovered in northern Kenya with improved geological mapping. The absence of a magnetic anomaly across the high density intrusive body, postulated from the evidence of a detailed gravity survey of the Gregory Rift (Searle 1969, 1970), is best explained by a combination of the low dip angle of the geomagnetic field and the high temperatures in the rifted regions. A Curie temperature as high as 200 "C would allow the top of the intrusive body to be within 7 km of the surface, assuming the temperature depth profile of Roy, Blackwell & Birch (196% and still be non-magnetic; see O'Reilly & Readman (1971) for a Curie temperature plot superimposed on the FeO-Ti0,-Fe,O, ternary diagram and Creer (1971) for rep- resentive Curie temperatures of basalts.

In contrast to the main East African Rift, there are some large magnetic anomalies across its postulated extension along the Lake Malawi/Shire Rifts. In the vicinity of this southern rift zone, three profiles record a double-peaked magnetic anomaly (T203, T220, T225, Fig. 4). Profiles T220 and T225 show high magnetic relief to the west of Malawi, so it is possible that these double-peaked anomalies merely indicate the continuation of this basement into the rifted regions. If not, then the anomalies


218 A. G. Green

- 0 °

-1 oo

0 -2 0

-3 0'

I /

i PROJECT MAGNET

l1

1 0" 2 oo 3 0' 4 0"

FIQ. 8. Project MAGNET profiles plotted along each track projected onto a map of Africa.

may mark the position of a large intrusive body which attempted to break the surface in Jurassic to Cretaceous times, causing the major rifts and volcanics now observed in Malawi and Mozambique. From the energy spectra the depth to the top of the body is estimated to be near to 5 km (depth values of 4.5 and 5.6 km in Fig. 7(b)). Support for the intrusive body postulate comes from an impressive gravity anomaly (55 km wide with an amplitude of 750 g.u.) observed 50-km to the north of Lake Malawi at the junction of the eastern and western rifts (Sowerbutts 1972). Sowerbutts (1972) has qualitatively related this anomaly to a 13 km wide ' Precambrian ultrabasic intrusive '.

There are two reasons why a large magnetic anomaly would be expected across the southern rift zone and not across the main East African Rift. First, the magnetic dip angle is greater across the Lake Malawi /Shire Rifts (44" to 52") than across the East African Rift over which ground or aero-magnetic surveys have been made (15" to 35"). Second, the temperature of a Pre-Tertiary plutonic body is likely to be much lower than the Tertiary to Recent intrusives of the East African Rift and is therefore more likely to be below the Curie point temperature. A more detailed survey of long gravity and magnetic traverses would help to confirm or refute the intrusive body postulate for the Lake Malawi /Shire Rifts.



The Kalahari and Congo Basins Profiles which sample the Kalahari Basin (T203, T211, T213, T225, 517, 522)

show a consistent pattern of high magnetic anomalies in both the long and short wavelengths (see original and upward continued profiles in Fig. 4 and Fig. 5). Exten- sive Karroo volcanic rocks beneath the sediments are considered to be the principal source of the short to medium wavelength anomalies, with granite and granodiorite masses of the Angola Basement contributing to the anomaly pattern in the more northern profiles. The occurrence of these rocks in small outcrops and deep boreholes beneath the profiles is the basis for this interpretation. If the interpretation proves to be correct, the areal estimates of Karroo volcanics (Cox 1970) would require revision.

Across the Congo Basin, the Project MAGNET profiles (T202, T219a, T220) and a vertical component magnetic survey (Evrard et al. 1959) reveal two distinct magnetic regimes. To the south of approximately 4" S there are many high amplitude anomalies, while to the north of this latitude the magnetic relief is very low (see flight paths in Fig. 1 and profile data in Fig. 3). The source of the high amplitude short wavelength magnetic relief is postulated to be a subsurface extension of the large igneous and /or iron ore complex observed on the Kasai Shield. Parts of the low magnetic basement to the north are known, from the evidence of deep boreholes and a seismic refraction survey (Evrard 1960) to be similar to the Precambrian A platform sediments which outcrop around the Congo Basin (Fig. 2). The Project MAGNET data (see also Evrard et al. 1959) would suggest that this type of basement is extensive in the northern part of the Congo Basin.

The long wavelength magnetic structure of the Kalahari Basin and the southern part of the Congo Basin shows a certain continuity of pattern. A series of large amplitude long wavelength anomalies may be traced from latitude 20" S (T211) to 4"s (T220). These features may be seen in the original curves (Fig. 4) but are best observed in the upward continued curves (Fig. 5). Average depths to these long wavelength structures generally decrease from 13 km in the south to a minimum of 2 km in the north. From the continuity of this pattern it seems reasonable to suggest that a single deep structural unit underlies a major part of the Kalahari and Congo Basins; the continuity of the magnetic pattern to the east and west on profiles T203 and T225 may indicate that the same deep structure lies beneath the Angola Basement and the Kasai Shield. The very different pattern observed to the north of approximately 4" S suggests a change of structure in this region (T219a, T202).

The Bangui and Kibalian magnetic anomalies It is not surprising that the Bangui and Kibalian Anomalies (a to b, and c to d

respectively on T204) are unrelated to the surface geology of metamorphic and sedimentary rocks. Apart from the crude depth estimates (30 km for the Bangui Anomaly and 8 - 4 km for the Kibalian Anomaly, Fig. 7(b)) there are two clues which indicate deep origins. First, the anomalies are dominated by high amplitude very long wavelength components. A shallow body could produce this type of anomaly if the magnetization varied along its length. However, a systematic change of the magnetization over the great lengths of these anomalies seems highly unlikely. The second clue is the absence of significant short wavelength magnetic relief. With the high magnetization needed to produce the long wavelength components, small irregularities of the bodies would produce corresponding high amplitude magnetic anomalies. Tf the bodies are deep the short wavelength magnetic relief is naturally filtered.

With regard to the magnetization directions of the causative bodies, the information from profile T204 (Fig. 3) and the assumption that the anomalies are two- dimensional perpendicular to the flight path (Green 1972) would lead to the erronous conclusion that the bodies are reversely magnetized. In fact, they are normally


220 A. G. Green

magnetized. At this low latitude a ' three-dimensional prism ' produces a large amplitude negative lobe with smaller positive lobes to the north and south. The satellite data show that the Bangui Anomaly is indeed three-dimensional and also reveal the important positive lobe to the south. Profile T201 (Fig. 3) shows the southern positive lobe of the Kibalian Anomaly.

In summary the causative bodies are (i) unrelated to the surface geology of metamorphic and sedimentary rocks, (ii) deep seated, (iii) highly magnetic (at least for the Bangui Anomaly), (iv) normally and probably inductively magnetized, and (v) extensive. The origin of these bodies is one problem which needs to be resolved. Could the bodies be massive iron meteorites? On a similar line, perhaps some late arriving primordial Earth material failed to sink and differentiate into core-mantle constituents resulting in an iron rich lower crust. Alternatively there could have been a major upwelling of ultrabasic magma which cooled in an exceptionally favour- able environment for the production of highly magnetic minerals. At the present time there is insufficient information for a complete diagnosis. Large scale gravity, magnetic and seismic surveys are required in the important areas of Cameroun, Central African Republic, and Zaire.

0 10 2 0 3 0 4 0

0

10

2 0

30

. B A N G U I K I B A L I A N

Y

z

W E S T C E N T R A L

A F R I C A H I G H

M A G N E T I C

R E L I E F L A K E M A L A W I - S H I R E R I F T Y

10 2 0 3 0 4 0 0

FIG. 9. Summary diagram showing the important magnetic features observed in the Project MAGNET data of Fig. 8; a and b are as in Figs 4 and 5



Conclusions

From this study of the Project MAGNET profiles, which cross central Africa, the following may be concluded:

(1) A characteristic magnetic pattern is observed across many of the tectonic / geologic units. Of particular note is (a) the Mozambique Belt ' Magnetic Quiet Zone ', (b) the near absence of significant magnetic anomalies across the East African Rift volcanic rocks, (c) the double-peaked magnetic anomalies associated with the Lake Malawi /Shire Rifts, (d) the high amplitude magnetic relief recorded across western central Africa which includes the Kasai Shield, the Angola Basement and the Kalahari and Congo Sedimentary Basins to the south of latitude 4" S, and (e) the extraordinary Bangui and Kibalian Magnetic anomalies observed on the North Congo Shield and the Kibalian Orogenic Belt respectively. In Fig. 8 the Project MAGNET profiles are shown plotted along each track projected onto a map of Africa. The geographic locations of the important features are then outlined in the summary diagram of Fig. 9. It is emphasized that the boundaries of the magnetic units shown in the figure are at best tentative; many more profiles need to be obtained before any conclusive maps can be drawn.

(2) Application of the Statistical Models Method (Spector & Grant 1970; Green 1972) to the Project MAGNET data generally reveals two ensembles of magnetic sources (Figs 6 and 7). The average depth to the shallow ensembles of magnetic bodies is significantly higher for lengths of profile which cross the Kalahari and Congo Sedimentary Basins (0.8 km) than for the lengths of profile which sample the exposed Precambrian Shield (0.2 km). The surface rocks and their deeper equivalents are interpreted as the principal members of the shallow ensembles of magnetic bodies. The depths to the deep ensembles of magnetic bodies covers a wide range from 4 to 30 km.

Within a given tectonic unit (i.e. the Mozambique Orogogenic Belt or the Kalahari and Congo Basins) the depth estimates to the deep ensembles seem to show a simple pattern or trend, but there is no obvious relationship between the depth estimates from neighbouring units. It is suggested that further geophysical investigation, particularly long seismic refraction profiles, is required before the nature of the deep ensemble of magnetic bodies can be clearly discerned.

Acknowledgments

I would like to express my sincere thanks to Dr R. W. Girdler, Dr R. L. Coles and Dr H. Pollock for their critical reviews of the first draft of this manuscript. I also thank Dr D. H. Tarling for useful discussion on the Bangui Anomaly. I thank Dr H. Stockard of the US Naval Oceanographic Office for making the magnetic data available and members of the Addis Ababa, Lwiro and Nairobi observatories for supplying the observatory magnetic records. The work was financed by a grant from the Natural Environment Research Council.

School of Physics, University of Newca.de,

Newcastle upon Tyne

References

Ackerman, E., 1950. Ein neuer Faltengurtel in Nordrhodesian und seine tectonishe Stellung in Afrikanishen Grundgeburg, Geol. Rund. 38, 24.


222 A. G. Green

Ackerman, E. & Foster, A., 1960. Grundzuge der Stratigraphie und Struktur des Irumide-orogens. C.R. XXI p. 182.

Alldredge, L. R., Van Voorhis, D. G. & Davies, T. M., 1963 Magnetic profile round the world., J. geophys. Res., 68, 3679.

Baker, B. H., Williams, L. A. J., Miller, J. A. & Fitch, F. J., 1971. Sequence and geochronology of the Kenya Rift volcanics, Tectonophysics, 11, 191.

Baker, B. H., Mohr, P. & Williams, L. A. J., 1972. Geology of the Eastern Rift System of Africa, Geol. SOC. Am. Spec. paper 136.

Banks, R. J. & Ottey, P., 1974. Geomagnetic deep sounding in and around the Kenya Rift Valley, Geophys. R . astr. Soc., 36, 321.

Benkova, N. P., Dolginov, Sh. Sh. & Simonenko, T. N., 1973. Residual geomagnetic field from the satellite Cosmos, 49, J. geophys. Res., 78, 798.

Berry, M. J., Jacoby, W. R., Niblett, E. R. & Stacey, R. A., 1971. A review of geophysical studies in the Canadian Cordillera, Can. J. earth Sci., 8,787.

Bhattarcharyya, B. K. & Morley, L. W., 1965. The delineation of deep crustal magnetic bodies from total-field aeromagnetic anomalies, J . Geomang. Geoelect. 17, 237.

Bichan, R. 1970. The evolution and structural setting of the Great Dyke, Rhodesia, in African magnetism and tectonics, eds. Clifford and Gass, Oliver and Boyd.

Bloomfield, K., 1970. Orogenic and Post Orogenic Plutonism in Malawi, in African magnetism and tectonics, eds. Clifford and Gass, Oliver and Boyd.

Bloomfield, K. & Hapgood, F., 1966. Explanatory notes on the structural map of Malawi, Geological Survey of Malawi Report.

Bulina, L. V., 1961. The use of airborne magnetic prospecting data in deep-seated structure of the Earth’s crust within the Siberian platform, Sovetskaya Geologiya, 5 , 134.

Bullard, E. C., 1967. The removal of trend from magnetic surveys, Earth Planet. Sci. Lett. 2,293.

Cahen, L., 1970. Igneous activity and mineralisation episodes of the Kibaride and Katangan Orogenic Belts of Central Africa, in African magmatism and tectonics, eds Clifford and Gass, Oliver and Boyd Ltd, Edinburgh.

Cahen, L. & Lepersonne, J., 1967. The Precambrian of the Congo, Rwanda and Burundi, in The Precambrian, Vol. 3, ed. K. Rankama.

Cahen, L. & Snelling, N. J., 1966. Geochronology of equatorial Africa, North-Holland. Caner, B., 1969. Long aeromagnetic profiles and crustal structure in western Canada,

Chapman, S. & Bartels, 1945. Geomagnetism, Clarendon Press, Oxford. Clifford, T. N., 1967. The Damaran episode in the Upper Proterozoic-Lower Pal-

aeozoic structural history of Southern Africa, Geol. SOC. Am. Spec. Paper 92. Clifford, T. N., 1968. Radiometric dating and the Pre-Silurian geology of Africa,

in Radiometric dating for geologists, eds E. I. Hamilton and R. M. Farquhar, Interscience Publishers.

Clifford, T. N., 1970. The structural framework of Africa, in Africaa magmatism and tectonics, eds Clifford and Gass, Oliver and Boyd Ltd, Edinburgh.

Cox, K. G., 1970. Tectonics and vulcanism of the Karroo Period and their bearing on the postulated fragmentation of Gondwanaland, in African magmatism and tectonics, eds Clifford and Gass, Oliver and Boyd Ltd, Edinburgh.

Earth Planet. Sci. Lett., 7, 3.

Cox, K. G., 1972. The Karroo volcanic cycle. J. Geol. Soc., 128, 31 1. Creer, K. M., 1971. Geophysical interpretation of remanent magnetisation in oxidized

Basalts, 2. Geophys., 37, 383. Darracott, B. W., 1972. Gravity and magnetic survey of the Speke Gulf, Serengeti and

Mara Regions and their relevance to rifting in East Afiica, PhD Thesis, University of Newcastle.

Douglas, R. J. W., 1970. Geology and economic minerals of Canada, Geological Survey of Canada, Economic Report no. 1.


Project MAGNET aeromagnetic proBles 223

Evrard, P., 1960. Resultats Scientifiques des Missions du Syndicat pour 1’6tude Geologique et Mini&re de la Cuvette et Travaux Connexes; Scismique. Ann. de Musee Royal du Congo Belge, 33.

Evrard, P., Jones, L. & Mathieu, P. L., 1959. Syndicat pour I’itude Geologique et MiniZre de la Cuvette Congolaise, Turvuen.

Fairhead, J. D., Mitchell, J. G. &Williams, L. A. J., 1972. New K/Ar determinations on Rift volcanics of S . Kenya and their bearing on age of rift faulting, Nature, 238, 66.

Flores, G., 1970. Suggested origin of the Mozambique Channel, Trans. geol. SOC. South Africa 73,l.

Fotidadi, E. E. & Karatayev, G. I., 1963. Structure of the Earth‘s crust of Siberia and the Far East according to data of regional geophysical investigations, Geology and Geophysics, Siberian Section, Academy of Sciences of the USSR, NO. 10.

Girdler, R. W. & Sowerbutts, W. T. C., 1970. Some recent geophysical studies of the Rift System in East Africa, J. Geomagn. Geoelect. 22, 153.

Green, A. G., 1972. Magnetic profile analysis, Geophys. J. R. astr. SOC., 30, 393. Green, A. G., 1973. Interpretation of Project MAGNET Data (1959 to 1966) for

Africa and the Mozambique Channel, Part 1 of PhD. Thesis, Newcastle Upon Tyne University.

Haines, G. V., Hannaford, W. & Riddihough, J. 1971. Magnetic anomalies over British Columbia and adjacent Pacific Ocean, Can. J. earth. Sci. 8, 387.

Hall, D. H., 1968. Regional magnetic anomalies, magnetic units and structure in the Kenora district of Ontario, Can. J. earth Sci., 5, 1277.

Hall, D. H. & Dagley, P., 1970. Regional magnetic anomalies, an analysis of the smoothed aeromagnetic map of Great Britain and Northern Ireland, NERC Report 70 110.

Hanh, A. & Zitzmann, A., 1969. The relation of magnetic anomalies to topography and geologic features in Europe in the the Earth’s crust and upper mantle, Geophysical Monograph 13, ed. Pembroke J. Hart.

Haughton, S . H., 1963. The stratigraphic history of Africa, south of the Snhara, Oliver and Boyd.

Hepworth, J. V., 1972. The Mozambique Orogenic belt and it’s foreland in North- east Tanzania, J. geol. Sac. (Lond.) 128,461.

Hepworth, J. V. & Kennerly, J. B., 1970. Photogeology and structure of the Mozambi- que Orogenic front near Kolo, Northeast Tanzania. Qf. J. geol. SOC. (Lond.), 125,447.

Hepworth, J. V., Kennerly, J. B. & Shackleton, R. M., 1967. Photogeologic investigation of the Mozambique Front in Tanzania. Nature, 216, 146.

Hepworth, J. V. & Macdonald, R., 1966. Orogenic belts of the northern Uganda basement, Nature, 210,726.

King, B. C., 1970. Vulcanicity and rift tectonics in East Africa, in African magmatism and tectonics, eds Clifford and Gass, Oliver and Boyd Ltd, Edinburgh.

King, B. C. & Chapman, G. R., 1972. Vulcanism of the Kenya rift valley, Phil. Trans. R. SOC. (Lond.) A271, 185.

King, B. C., Le Bas, J. M. & Sutherland, D. S., 1972. The history of the alkaline volcanoes and intrusive complexes of Eastern Uganda and Western Kenya. J. Geol. SOC., 128,173.

Krutikhovskaya, Z . A., Pashkevich, I. K. & Simonenko, T. N., 1973. Magnetic anomalies of Precambrian shieIds and some problems of their geoloical interpretation, Can. J. earth Sci. 10, 629.

Macgregor, A. M., 1947. An outline of the geological history of Southern Rhodesia, Geol. Surv. S. Rhod. Bull. 38.

Macgreaor, A. M., 1951a. Some milestones in the Precambrian of Southern Rhodesia, Trans. Geol. SOC. of South Africa, 54, 27.


224 A. G. Green

Macgregor, A. M., 1951b. A comparison of the geology of Northern Rhodesia (Zambia) and Southern Rhodesia and adjoining territories, Int. Geol. Cong. 18th.

Maclaren, A. S . & Charbonneau, B. W., 1968. Characteristics of magnetic data over major subdivisions of the Canadian shield, Proc. geol. Assoc. Can., 19, 57.

McGrath, P. & Hall, D. H., 1969. Crustal structure in North Western Ontario: regional magnetic anomalies, Can. J . earth Sci., 6, 101.

McClure, D., 1963. Interpretation of long aeromagnetic profiles, MSc Thesis, Univer- sity of Alberta.

Mestraud, J. L., 1971. The tectonics of Central Africa, in Tectonics of Africa, UNESCO, Earth Sciences 6, eds G. Choubert and A. Faure-Muret.

Mundt, W., 1965. On the statistics of geomagnetic anomalies in the territory of the German Democratic Republic, Geophysik, 74, 106.

O’Reilly, W. & Readman, P. W., 1971. The preparation and unmixing of cation deficient titanomagnetites, 2. Geophys., 37, 321.

Pallister, J. W., 1971. The tectonics of East Africa, in Tectonics of Africa, UNESCO, Earth Sciences 6, eds. G. Choubert, and A. Faur-Muret.

Pakiser, L. C. & Zietz, J., 1965. Transcontinental crustal and upper mantle structure, Rev. Geophys. 3, 505.

Puchkova, L. I. & Ladynin, A. V., 1966. On the relationship of magnetic anomalies to crustal structure in the south eastern part of Central Kazakhstan, Acad. Nauk SSSR.-U.S. Geol. Surv. Geophys. Abs. 238.

Quennel1,yA. M., McKinlay, A. C. M. & Aitken, W. G., 1956. Summary of the Geology of Tmganyika pt. I , Introduction and Stratigraphy. Geol. Surv. Tan- kanyika Mem. 1.

Regan, R. D., Davis, W. M. & Cain, J. C., 1973. The Bangui magnetic anomaly, EOS 64,236.

Riddihough, R. P., Haines, G. V. & Hannaford, W., 1973. Regional magnetic anomalies of the Canadian Arctic, Can. J . earth Sci., 10, 157.

Roy, R. F., Blackwell, D. D. & Birch, F., 1968. Heat generation of Plutonic Rocks and continental heat flow provinces, Earth. Planet. Sci. Lett. 5, 1.

Sanders, L. D., 1965. Geology of the contact between the Nyanza Shield and the Mozambique Belt in Western Kenya. Geol. Surv. Kenya Bull. 7.

Searle, R. C., 1969. Barometric hypsometry and a geophysical study of part of the Gregory R v t Valley in Kenya, PhD Thesis, University of Newcastle upon Tyne.

Searle, R. C., 1970. Evidence from gravity anomalies from thinning of the lithosphere beneath the Rift Valley in Kenya. Geophys. J. R. astr. SOC., 21, 13.

Serson, P. M. & Hannaford, W. L. W., 1957. A statistical analysis of magnetic profiles, J. geophys. Res., 62, 1.

Simonenko, T., 1969. Relation of magnetic anomalies to topography and geology in USSR the Earth’s Crust and Upper Mantle, Geophysical Monograph 13, ed. Pembroke J. Hart.

Simonenko, T., 1972. The magnetic field of regions of the continental crust, Izvestiya, 9, 599.

Sowerbutts, 1972. Long gravity and magnetic profiles over the East African Rift system and their interpretation, PhD Thesis, University of Newcastle upon Tyne.

Spector, A. & Grant, F. S., 1970. Statistical models for interpretating aeromagnetic data, Geophys., 35,293.

Spooner, C. M., Hepworth, J. V. & Fairbairn, H. W., 1970. Whole rock Rb-Sr isotopic investigation of some East African granulites, Geol. Mag., 107,511.

Talwani, M. & Heirtzler, J. R., 1964. Computation of magnetic anomalies caused by two-dimensional structures of any shape, in Computers in the mineral industries, ed. G. A. Parks, Stanford University.

Tukey, J. W., 1961. Discussion emphasizing the connection between the analysis of variance and spectrum analysis, Technometrics, 3, 191.



Vacquier, V. & Affleck, J., 1941. Computation of the depth to the bottom of the Earth's magnetic crust, Trans. Am. geophys. Un. 446.

Vail, J. R., 1966. Zones of progressive regional metamorphism across the western margin of the Mozambique Belt in Rhodesia and Mozambique, Geol. Mag. 103,231.

Vail, J. R., 1968. The southern extension of the East African Rift system and related igneous activity. Geol. Rdsch., 57, 601.

Vail, 3. R., Snelling, N. J. & Rex, D. C., 1968. Pre-Katangan geochronology of Zambia and adjacent parts of Central Africa, Can. J. earth Sci., 5, 621.

Welsh, P. D., 1967. The use of the FFT for estimation of spectra-A method based on time averaging over short modified periodograms, IEEE, Trans. Audio Electro- acoust., Au-15, 70.

Whitham, K. & Hoge, E., 1961. Geomagnetic investigations in British East Africa during 1959, Dominion astrophys. Obs. Report.

Wohlenberg, J. & Bhatt, N. V., 1972. Report on an aeromagnetic survey of two areas in the Kenya Rift Valley, Tectonophysics 15, 143.

Woolley, A. R. & Garson, M. S . , 1970. Petrochemical and tectonic relationship of the Malawi carbonatite-alkaline province and the Lupata-Lebombo volcanics, in African magmatism and tectonics, eds Clifford and Gass, Oliver and Boyd Ltd, Edinburgh. in African magmatism and tectonics, eds. Clifford and Gass, Oliver and Boyd.

Worst, B. G., 1960. The Great Dyke of Southern Rhodesia, Geol. Surv. S. Rhodesia Bdl. 47.

Zietz, I. & King, E. R., 1966. Crustal study along the Transcontinental Great Circle from Washington, D.C., to San Francisco, Calif., in Continental margins and island arcs, Int. Upper Mantle Comm. Symp. Ottawa, Can. Geol. Surv. paper 66, 445.

Zietz, I., King, E. R., Geddes, W. & Lidiak, E. G., 1966. Crustal study of a continental strip from the Atlantic Ocean to the Rocky Mountains, Geol. SOC. of Am. Bull., 77, 1427.

Appendix A

Examples of the statistical models method applied to simulated data

To test the average depth technique, data were simulated using a two-dimensional magnetic anomaly computer program (Talwani & Heirtzler 1964). An example of two ensembles of magnetic bodies together with the resultant profile and logarithmic energy spectrum is shown in Fig. A1. The effects of the average widths of the bodies have been removed from the spectra in the manner described by Green (1972). The logarithmic spectrum NMOP was obtained from the simulated profile using only the broad deep magnetic bodies. The peak in the low wavenumber part of the spectrum due to the interaction of the depth and thickness terms is clearly seen, as is the linear decay of the curve away from the peak due to the dominant depth term. The average depth obtained from the section of logarithmic spectrum MOP was estimated to be 5.4 km which is to be compared to the true average depth of 5.6 km below the flight line.

The logarithmic spectrum NMOQ was obtained from the profile using both the deep and shallow bodies in the simulation (profile A in Fig. Al). The curve marked NMO is not affected by the introduction of the second ensemble (on the scale shown at least) and therefore this section may be used to obtain an average depth estimate for the deep ensemble of bodies. The linear section of curve MO yields an average depth of 6.2 km (true value of 5.6 km) for the deep ensemble, while the linear trend OQ yields an average depth of 0.54 km which compares well with the true average depth of 0.6 km for the shallow ensemble of bodies.

0


226 A. G. Green

(&I cycles/ krn

FIG. Al. Two sets of magnetic bodies from separate and distinct ensembles. F. L. denotes flight line, A is the computed anomaly curve and B is the spectra. The curve NMOP is the significant part of the energy spectrum obtained from the deep ensemble of bodies only and the curve NMOQ is the energy spectrum for both

ensembles of bodies.

Two further tests relevant to this paper, were applied to profiles obtained from sets of bodies not included in the theory presented to date. First a single body, its calculated anomaly curve and energy spectrum are shown in Fig. A2. Exactly the same procedure was followed as for an ensemble of bodies to give a depth of 5.4 km (true value of 5.3 km). Finally a set of magnetic bodies similar to the deep bodies of Fig. A1 was taken, but now two of the bodies (marked R in Fig. A3) are reversely magnetized with respect to the others. Surprisingly the average depth computed from the logarithmic spectrum was 5 - 1 km which compares extremely well with the true average depth of 5.3 km. This result is best explained by considering the wavenumber content of individual anomalies together with the ' short periodogram method ' of energy spectra determination (Tukey 1961; Welsh 1967). Ideally, the shape of the spectrum due to a ' two-dimensional prism ' is independent of the magnetic moment vector (Green 1972). Therefore, the linear decay of the logarithmic spectrum of a reversely magnetized body is the same as for a normally magnetized body. Thus, we could (i) section the profile in Fig. A3 into individual anomalies, (ii) transform each anomaly into the wavenumber domain, (iii) square and add the wavenumber co- efficients to give the periodogram of each anomaly, and (iv) compute the average periodogram to give an estimate of the spectrum. At stages (iii) and (iv) the information about the direction of the magnetic moment vector is lost. It would seem, that provided there is little interference of anomalies, this method is useful for analysing lengths of profiles which have contributions from magnetic bodies with widely varying directions of magnetization.


Project MAGNET aeromagnetic profiles

12

8

4 A

w V

- v >

K - 0

-4 EL.

20km

'x €3

X u

.1 .2 .3 ( v / 2 ~ ) cycledkm

Fro. A2. Magnetic profile and energy spectrum for a single magnetic body. All the letters are as for Fig. Al.

( L ) c y c l e s / K r n 2+

om 0 J 5 K m

20 Km

F I ~ . A3. Magnetic profile and energy spectrum for an ensemble of magnetic bodies which includes reversely magnetized bodies. The blocks marked ' R ' indicate the bodies which are reversely magnetized with respect to the remaining bodies. All

the ktters are as for Fig. Al.

227


228 A. G. Green

Appendix B

List of geological maps used in the magnetic profile-geology interpretation

Map Esboco Geologico de Angola Provisional Geological Map of Botswana Carte Geologique du Cameroun Carte Geologique du Congo Belge et du Ruanda-Urundi Carte Geologique de L'Afrique Equatorial Francaise Kenya Geological Map Geological Map of Malawi Geological Map of Mozambique Geological Map of Rhodesia Geological Map of South West Africa Geological Map of Tanzania Geological Map of Tanganyika Geological Map of Uganda Geological Map of Uganda Geological Map of Northern Rhodesia (Zambia) Geological Map of Africa Tectonic Map of Africa

Scale 1 : 2X1O6 1 : 2x106 1 : 1 x 106

1 : 2 x 106

1 : 2x106 1 : 3x106 1 : 1 x lo6 1 : 2x106 1 : 1x106 1 : 1 x 106 1 : 3x106 1 : 2x106 1 : 1.25X1O6 1 : 1.50x1O6

1 : i X 1 O 6 1 : 5x106 1 : 5x106

Year of Publication

1954 1971 1956

1951

1958 1962 1966 1968 1971 1963 1967 1959 1961 1966

1961 1963 1968

Table 1

Average depths or range of depths to the top and bottom of' ' '

from various studies. aeep . magnetic sources

Average depth or range of depths to the top of

Average depth or range of depths to the bottom

Region the bodies (km) of the bodies (km)

North America 20.0

4.0 to 30.0

4.8 to 8.0

4.0 to 03

14.4 to 25.6

40.0 to 48,O 13.0 to 20-0 4.0 to f3.0

Russia

10 km

Germany

Conrad Discontinuity (- 15 km) Conrad to ' M ' Discontinuity (15-35 km) Above the Conrad Discontinutiy 30 km

Above the Conrad Discontinuity

Gt Britain 8.0 to 15.0 15.0 to 26.0

Reference

Vacquier and Affleck (1941) Serson & Hannaford (1957) Bhattacharyya & Morley (1965) McClure (1963) Hall (1968) and McGrath & Hall (1969)

Bulina (1961)

Fotiadi & Karatayev (1963)

Puchkova & Ladynin (1966) Krutikhovskaya e ta / . (1973)

Mundt (1965)

Hall & Dagley (1 970)

This study Africa 4.4 to 29.9


Documents

Interpretation of Project MAGNET Aerornagnetic Profiles Across Africa