Control of normal fault
interaction on the distribution
of major Neogene sedimentary
depocenters, Lake Tanganyika,
East African rift
Kiram E. Lezzar, Jean-Jacques Tiercelin,
Caroline Le Turdu, Andrew. S. Cohen, David J. Reynolds,
Bernard Le Gall, and Christopher A. Scholz
ABSTRACT
The Tanganyika continental rift basin is one of the most important
structural features of the East African rift system and provides an
opportunity to observe the early stages of rift basin development
unobscured by postrift deformation and erosion. The geometry of
half grabens and their zones of linkage have a great influence on rift
development and depositional environments. Topographic features
associated with zones of linkage between half grabens exert a direct
control on drainage basin evolution, sediment supply, and synrift
stratigraphy. Previous structural studies, based on widely spaced
(⬃15 km) seismic reflection profiles, focused mainly on large-scale
geometrical fault descriptions and not on the spatial and temporal
linkage of the individual border faults controlling each half graben.
In this article, using newly available basin age estimates, multichannel seismic reflection data, high-resolution single-channel sparker
seismic data, and onshore structural data (remote directioning and
microstructural field observations), we have constructed a detailed
late Miocene–Holocene kinematic model for the evolution of the
northern part of the Lake Tanganyika rift basin. A classification of
fault interaction geometry is proposed to describe the initiation and
development through time of major depocenters. Fault correlation
lessons are provided for exploration seismic interpreters in extensional settings. The development of the depocenters of northern
Lake Tanganyika is complex, and this article clearly shows that antecedent structures control subbasin initiation and development. As
the rift evolves, border faults become dominant, producing more
AUTHORS
Kiram E. Lezzar ⬃ UMR 6538 Domaines
Oceaniques, Institut Universitaire Europeen de la Mer,
Universite de Bretagne Occidentale, Place Nicolas
Copernic, 29280 Plouzane, France; current address:
Department of Earth Sciences, 204 Heroy Geology
Laboratory, Syracuse University, Syracuse, New York,
13244; klezzar@syr.edu
Kiram E. Lezzar received an M.S. degree in marine
geosciences (DEA 1991) and a Ph.D. (1997) from the
University of Western Brittany in Brest, France. His Ph.D.
topic was the tectonics and stratigraphy of the northern
Lake Tanganyika rift basin. Since 1998, he has been a
research associate in the Department of Earth Sciences at
Syracuse University, New York. His main research is
concentrated on reflection seismic analysis (structure and
stratigraphy) of the East African rift great lakes, such as
Tanganyika, Malawi, Albert, and Edward. He is also the
limnogeology mentor of the Nyanza Project (since 1999),
a National Science Foundation–funded research training
program in tropical lake studies in East Africa, open to
undergraduates, graduate students, and secondary school
teachers.
Jean-Jacques Tiercelin ⬃ UMR 6538
Domaines Oceaniques, Institut Universitaire Europeen de
la Mer, Universite de Bretagne Occidentale, Place Nicolas
Copernic, 29280 Plouzane, France; tiercelin@univ-brest.fr
Jean-Jacques Tiercelin is research director in the CNRS/
IUEM at the University of Western Brittany in Plouzane,
France. He received his Doctorat d’Etat from the AixMarseille II University in 1981. He has worked on the
sedimentology and tectonics of various lake basins in the
East African rift system from Ethiopia down to Botswana.
Caroline Le Turdu ⬃ UMR 6538 Domaines
Oceaniques, Institut Universitaire Europeen de la Mer,
Universite de Bretagne Occidentale, Place Nicolas
Copernic, 29280 Plouzane, France; current address: Elf
Petroleum Norge AS, Dusavik, P.O. Box 168N, Stavanger,
Norway; caroline.le-turdu@technoguide.com
Caroline Le Turdu is an account manager for France and
eastern Europe at Technoguide in Oslo, Norway. She
received her Ph.D. in structural geology from the
University of Western Brittany in Brest, France, in 1998.
After a two-year postdoctorate at Elf Petroleum Norge in
Stavanger, Norway, she joined Technoguide, a provider of
advanced software for 3-D modeling of oil and gas
reservoirs.
Andrew. S. Cohen ⬃ Department of
Geosciences, University of Arizona, Tucson, Arizona,
85721; acohen@geo.arizona.edu
Andy S. Cohen is a professor of geosciences and joint
professor of ecology and evolutionary biology at the
University of Arizona. He is interested in rift lake history
and paleolimnology.
David J. Reynolds ⬃ Exxon Production Research
Company, P.O. Box 2189, Houston, Texas, 77252–2189;
djreyno@upstream.xomcorp.com
Copyright 䉷2002. The American Association of Petroleum Geologists. All rights reserved.
Manuscript received May 26, 2001; revised manuscript received June 4, 2001; final acceptance January
2, 2002
AAPG Bulletin, v. 86, no. 6 (June 2002), pp. 1027–1059
1027
David J. Reynolds is a senior research specialist at
ExxonMobil Upstream Research Company. He received his
B.A. degree from Princeton University, an M.S. degree
from Duke University, and a Ph.D. from Lamont-Doherty
Earth Observatory. His research interests include rift
tectonics and structural modeling.
continuous and elongate depocenters, although the influence of
transverse structures is still evident.
Bernard Le Gall ⬃ UMR 6538 Domaines
Oceaniques, Institut Universitaire Europeen de la Mer,
Universite de Bretagne Occidentale, Place Nicolas
Copernic, 29280 Plouzane, France; blegall@univ_brest.fr
Bernard Le Gall received his Ph.D. in 1983 and his HDR
(research director) in 1995 from the University of
Western Brittany (UBO/CNRS-Brest). Since 1995, he has
been studying extensional deformation in the active rift
system of East Africa (Kenya, Tanzania), and in the Karoo
rifted zone of Botswana and South Africa.
Christopher A. Scholz ⬃ Department of
Earth Sciences, 204 Heroy Geology Laboratory, Syracuse
University, Syracuse, New York, 13244; cascholz@syr.edu
Christopher A. Scholz is an associate professor in the
Department of Earth Sciences at Syracuse University. His
research focus is on the sequence stratigraphy and basin
analysis of rift systems and large lakes.
ACKNOWLEDGEMENTS
This article represents parts of Kiram Lezzar’s and
Caroline Le Turdu’s Ph.D. dissertations at the Doctoral
School of Marine Geosciences UMR 6538 of the Universite
de Bretagne Occidentale, Brest, France, and benefited
from the assistance of several research projects: the Casimir Project of the Royal Museum of Central Africa (Tervuren, Belgium); the Tanganydro Project; and the 3D-3G
Project of the Universite de Bretagne Occidentale, Brest.
Research permits in Zaire (currently the Democratic Republic of Congo) were provided by the Ministere de
l’Energie, Commission Nationale l’Energie and in Burundi
by the Faculte des Sciences, Universite du Burundi. The
Ministere de l’Education National at Algiers, Algeria,
granted in 1991 a Ph.D. scholarship to K. E. Lezzar. This
work was funded by grants from the Casimir Project, Elf
Aquitaine Production, AAPG Grant-in-Aid 1997 to C. Le
Turdu and K. E. Lezzar, Elf Petroleum Norge AS (3D-3G
Project; grant to J.-J. Tiercelin, principal coordinator), the
Ministry of Foreign Affairs of France, INSU-CNRS (France),
and the Universite de Bretagne Occidentale (SUCRI). We
would like to thank the coordinator of this AAPG special
issue, John Underhill, for his patience and great help during a long period of revision. Thanks also to our three reviewers, Bruce Trudgill, Christopher Morley, and Patience
Cowie, for their immense and essential input; they provided extremely helpful guidance and advice to improve
the quality of the submitted manuscript. Thanks also to
Max Fernandez-Alonso, Karel Theunissen, Joel Rolet, and
Christophe Coussement for helpful discussions and suggestions. Special thanks to Elf Aquitaine Production for
providing access to the Georift Project database. Thanks
to Bernadette Coleno for her assistance in preparation of
illustrations and to Peter Cattaneo for his helpful advice
to solve computer system and software problems. We express our gratitude to the management of Elf Petroleum
Norge AS for having authorized this publication. We thank
the National Science Foundation (Grant #ATM-9619458
Nyanza Project) for financial support of this research. This
article is publication #128 of the International Decade of
East African Lakes (IDEAL) program.
1028
INTRODUCTION
Lake Tanganyika occupies a north-south–trending continental rift
basin, one of the most dominant structural features of the East African rift system (EARS). The lake has a length greater than 700
km, a mean width of 50 km, a maximum water depth of 1.5 km,
and a sedimentary thickness greater than 4 km, and up to 7 km of
throw has been identified along the major border faults (Reynolds,
1984; Rosendahl et al., 1986; Morley, 1988; Ebinger, 1989b; Morley et al., 1989; Tiercelin and Mondeguer, 1991) (Figure 1). In this
article, we have constructed a detailed late Miocene–Holocene kinematic model for the evolution of the northern part of the Lake
Tanganyika basin (between long. 3⬚20⬘E and lat. 4⬚30⬘S) (Figure 2).
Our data sources include the following: (1) selected, multichannel,
air-gun seismic reflection data from Project PROBE (Rosendahl et
al., 1986, 1988); (2) high-resolution, single-channel sparker seismic
reflection data (lines S1–S9) from the Casimir Project; (3) very high
resolution (5 kHz) echo-sounding seismic profiles from the Georift
Project (Bouroullec et al., 1992; Lezzar et al., 1996; Lezzar, 1997);
(4) lake basin age estimates (Tiercelin and Mondeguer, 1991; Cohen et al., 1993; Lezzar et al., 1996; Lezzar, 1997); and (5) onshore
structural data (remote directioning and microstructural field observations) (Chorowicz et al. 1987; Reynes et al. 1993; Coussement
et al., 1994; Coussement, 1995; Le Turdu, 1998). We investigated
many important previous studies on rift tectonics, volcanism, and
magmatism in which were developed models of rift fault geometry,
growth, and propagation and their relationships to volcanic activity
and preexisting prerift fabrics (Baker, 1986; Leeder and Gawthorpe, 1987; Steckler et al., 1988; Ebinger et al., 1989a, b, 1991;
Karson and Curtis, 1989; Kuznir and Egan, 1989; Cowie and
Scholz, 1992; Schlische, 1992, 1995; Gawthorpe and Hurst, 1993;
Ring et al., 1993; Anders and Schlische, 1994; Coblentz and Sandiford, 1994; Gawthorpe et al., 1994, 1997; Hendrie et al., 1994;
Morley, 1994, 1999a, b, c; Trudgill and Cartwright, 1994; Cartwright et al., 1995, 1996; Childs et al., 1995; Dawers and Anders,
1995; Rohrman and van der Beek, 1996; Schlische et al., 1996;
Contreas et al., 1997, 2000; Simiyu and Keller, 1997; Zhao et al.,
1997; Gupta et al., 1998, 1999; Sharp et al., 1999; Dawers and
Underhill, 2000; McLeod et al., 2000). Comparing and integrating
all these models with our data interpretation, we examined spatial
and temporal relationships between major rift structures in Lake
Tanganyika, as well as fault propagation and interaction during successive late Tertiary phases of rifting. We studied the influence of
rift fault propagation and abandonment on coarse detrital deep lacustrine fan development and distribution. Finally, we offer fault
correlation lessons for petroleum exploration geologists to avoid
wrong fault interpretation of seismic data in extensional settings.
Neogene Sedimentary Depocenters (East African Rift)
28°E
Figure 1. (A) The East African
rift system and location of the
Lake Tanganyika basin.
(B) Geological setting of the
north Tanganyika trough (study
area) related to the Precambrian basement fabrics.
30°E
2°S
Lake
Kivu
Oc
ian
I nd
n
ra
ba
Ki
ch
50 Km
an
Approximative trend
of basement fabrics
M
Br
yika
Tangan
Trough
B.
T
er n
n
We s t
ia
Trough
N)
IA
ND
BE
(U
iz
La k e
us
4°S
zi
u si
Fig. 2
R
BURUNDI
ea n
East African Rift
R
DEMOCRATIC
REPUBLIC
OF CONGO
A.
T - Lake Tanganyika
M - Lake Malawi
TANZANIA
Lacustrine sediments
and alluvium
Upper Proterozoic
sediments
Neogene volcanics
Late Proterozoic orogenic
belt
Permian–Triassic sediments
(Karoo)
Tanzanian Craton
GEOLOGICAL SETTING
Interpretations of the development of the Tanganyika
rift basin have been based upon two different models
deduced from a coarse seismic data set (Rosendahl et
al., 1986, 1988; Morley, 1988). One model assumes
pure east-west rifting (Morley, 1988, 1995, 1999b,
2002; Morley et al., 1992; Coussement et al., 1994),
whereas the other model assumes northwest-southeast
extension with a strong wrenching component (Chorowicz et al., 1983; Scott et al., 1992). If present-day
seismicity is an accurate indicator of the long-term regional extension direction, then focal mechanism solutions support east-west extension for the Lake Tanganyika area (Fairhead and Stuart, 1982; Shudofsky,
1985). The maximum east-west extension calculated
using PROBE seismic lines on major faults has been
estimated at about 12–14 km (Morley, 1989).
Previous age estimates based on the reflection
seismic–radiocarbon method (RSRM) suggest that the
Tanganyika basin began to form before 12 Ma in its
central region, at about 7–8 Ma for the northern basin
(our study area), and at about 2 Ma for the southern
basin (Cohen et al., 1993; Lezzar et al., 1996). Even if
the development of the Tanganyika basin occurred exclusively during the Neogene, however, the position
and orientation of the modern basins has originated
from movement of a series of major fault zones associated with ancient tectonic and metamorphic events
(Versfelt and Rosendahl, 1989). These include structures within cratonic areas and orogenic belts that
range from early Precambrian (3600 Ma) to late Precambrian (600 Ma) in age (McConnell, 1972).
The Lake Tanganyika basin is formed within two
mobile belts west of the rigid Tanzanian craton. The
N130–140–trending Ubendan (or Rusizian) belt is
formed by granites, gneisses, and mica schists overprinted by the Eburnian orogeny at 1950–1850 Ma
(Paleoproterozoic), whereas the N30–50–trending Kibaran belt is related to the Kibaran orogeny (Mesoproterozoic) (Daly, 1988; Theunissen et al., 1996). The
N0–20–trending north Tanganyika basin divides the
Kibaran belt into southwestern and northeastern belts
(Fernandez-Alonso and Theunissen, 1998). On the
western (Congo) side of the North Tanganyika trough,
the southwestern Kibaran belt consists mainly of highgrade metamorphic orthogneisses and high-grade
metasediments with an N30–50 fabric. On the eastern
Lezzar et al.
1029
0
10
Figure 2. Bathymetric map of
the northern Lake Tanganyika
rift basin showing the location
of Project PROBE (P) and
Casimir Project (S) seismic lines
used in this study and of magnified seismic sections displayed
in Figures 3 and 4. Location of
Figure 2 shown in Figure 1. Refer to Table 1 for all other
abbreviations.
P228
N
S9
250
10 km
S1:Sparker Reflection Seismic Survey
Project Casimir
P6A
S7-2
P16: Multichannel Reflection Seismic
Survey. PROBE Project
S7
3
S7-2:Seismic line blow up displayed
in Figure 4
S8
P200
1-South Bujumbura subbasin/South
Rusizi half graben/RSHG
2-Rumonge subbasin/North Kigoma
half graben/NKHG
3-North Bujumbura subbasin/North
Rusizi half graben/NKHG
4-Ubwari horst
S6-2 Cape
Magara
WU
F
S6
S5
S5-2
WUF: West Ubwari Fault
250
CM F
300
300
50
1
EUF: EastUbwari fault
S3-2
S2
S2-2
P16
S2
EUF
S2-1
4
2
400
0
200
300
100
10
0
200
10
S3
CMF: Cape Magara fault
S4
Cape
Banza
(Burundi) side, gneissic formations belong to the Archean basement, and the Kibaran belt is represented
by quartzites and metaquartzites, schists, and intrusive
granites with N0, N110, and N130–140 fabrics (Carte
Geologique du Burundi, 1986, 1988, 1989; Fernandez-Alonso and Theunissen, 1998).
The Tertiary rift structures are dominated by half
grabens in the offshore study area and are defined as
1030
Neogene Sedimentary Depocenters (East African Rift)
individual major sedimentological domains (Tiercelin
and Mondeguer, 1991). The deep basin structure (Figure 3A, C, E) consists of three individual half grabens,
for each of which the internal structure is controlled
by a predominance of minor faults that are mainly synthetic to the major boundary fault. These are, from
north to south, the North Rusizi half graben (NRHG),
the South Rusizi half graben (SRHG), and the North
A.
W
E
0
0
North of
Pemba
LMF
200
Bujumbura Subbasin
Deep Basin
(c)
?
(a)
(b)
?
(f)
Î
(a)
(b)
(a)
(c)
UBFSn
600
(Channels system)
Î
(b)
(d)
(a)
(b)
(d)
(c)
(f)
(d)
LMF
RL
(c)
(d)
(c)
(c)
(c)
(f)
(d)
(d)
(f)
(f)
(f)
Horst
?
RL
North Rusizi Half Graben
~0.4/0.3 Ma
400
?
(a)
(c)
(d)
(f)
800
(a)
(b)
(b)
BBFSn
(a)
400
200
Bujumbura
Slope
Transverse Fault (NE-SW)
Foot slope
600
?
800
(ms tw-tt)
(ms tw-tt)
2 km
E
W
B.
0
~0.4/0.3 Ma
0
(f)
Ru
?
1
1
(NE)
RL
LM F
3
?
UBFS n
4
5
4
~3.5 Ma
C. K. Morley,
pers. com
2 km
6
~0.4/0.3 Ma
D.
W
E
~0.4/0.3 Ma
400
WUF
f
800
8 Km
Sec tw tt
~5 Ma
PMF
PMF
600
4 Km
CMFw
CMFw
f
3
5
(s. tw-tt) 6
C.
2
a
igom
th-K ben
Nor alf-gra
)
H
(NE
(NE)
BB
FS
n
?
2
Msec twtt
E
E.
W
Banza
Shoal
Rumonge Subbasin
200
CMFw
zi
be
Ka
(c)
Slope
~0.2 Ma
(a)
D
(c)
B
(d)
(b)
C
400
(b)
A
?
(c)
(f)
600
?
F-E
(f)
CBFe
WUF
Baraka
Furrow
Buried
Magara Shoal
D
F
(f)
Rumonge
Channel
Î
(b)
(c)
CBFw
Kaboge
Dome
BL
me
Do
(a)
EUF
?
C
(d)
~0.4 Ma
(a)
Î
CMFe
RL
C
(e)
B
?
D
E
A
(b)
(c)
KAF
(a)
(c)
F
Magara
Slope
(a)
B
(d)
(f)
Capart
Channel
Baraka
Channel
A
(b)
Magara-Banza Depression
EUF'
Bujumbura Subbasin
Banza
Slope
multiple
800
M
?
(ms tw-tt)
Ubwari Horst
South Rusizi Half Graben
Buried Magara
Shoal
~0.4 Ma
W
0
~1.1 Ma
KFZ
~7.4 Ma
(N
M
E)
F
South Rusizi
Half graben
0
(f)
RM
2 km
CBF e
4
(s. twtt)
Ma
E)
?
EUF
(N
E)
(N
? ~5
KAF
Ru
(f)
RL
E
K
(f)
Bu
1
2
Banza
Shoal
C B Fw
F.
North Kigoma Half Graben
WUF
2 km
~3.5 Ma
Ubwari
Horst
1
2
4
North Kigoma
Half graben
Figure 3. Interpretation of main seismic lines used to reconstruct the northern Lake Tanganyika rift kinematic model. (A) Singlechannel sparker line S7; (B) multichannel reflection seismic line drawing P6 (modified after Rosendahl et al., 1986); (C) multichannel
line drawing P200 (modified after Rosendahl et al., 1986); (D) single-channel sparker line S6 (modified after Rosendahl et al., 1986);
(E) single-channel sparker line S2; (F) multichannel reflection seismic line drawing P16 (modified after Rosendahl et al., 1986). Filled
circles indicate RSRM age estimation (see text for details). Key seismic sections (parts A, C, E) are magnified in Figure 4. Refer to
Figure 2 for location of lines and to Table 1 for fault and sequence abbreviations. More seismic data are also available from Project
PROBE in Rosendahl et al. (1986) and from the Casimir Project in Lezzar et al. (1996) and Lezzar (1997).
Lezzar et al.
1031
E
W
E
W
400
400
500
500
400
400
500
500
600
600
600
600
700
700
(Ms-twtt)
(Ms-twtt)
500 m
1 km
1 km
W
Bujumbura Subbasin
W
C3
E
B
A
400
400
(b)
A
(a)
C
(c)
C2
Î
(a)
(a)
(b)
400
A
B
C
B
B
(b)
(b)
C
500
A
(a)
E
Barakal
Channel
C1u
C2
(b)
(c)
500
500
C
?
500
(d)
E
D
C
(d)
600
(c)
(c)
D
600
LMF
(Ms-twtt)
500 m
(Ms-twtt)
F-E
(f)
Baraka
Furrow
?
(d)
(d)
700
600
F
Kaboge Dome(
(e)
600
F-E
700
F-E
1 km
E
W
E
W
400
400
500
500
600
600
700
400
400
500
500
600
600
700
700
700
(Ms-twtt)
LineS6-2
1 km
1 km
(Ms-twtt)
E
W
E
W
400
400
400
400
A
(a)
A
(b)
C
(c)
D
(d)
?
600
(c)
600
600
?
E
(d)
PMF
700
(Ms-twtt)
D
600
?
F-E
(f)
500
C
(c)
(c)
(d)
(d)
(d)
B
(b)
500
500
700
700
(e)
F
1 km
700
WUF
B
(c)
?
(b)
?
(a)
?
WUF
B+A
(b)
?
500
1 km
300
W
300
400
400
500
500
600
600
SE
NW
E
400
400
500
500
600
600
700
700
800
800
(Ms-twtt)
500 m
E
W
SE
NW
400
Ubwari Horst
Eastern Side
400
300
300
1 km
(Ms-twtt)
C3
C2
(c)
B
A
(a)
B
(d)
(f)
Deep Lacustrine Fans
Basin Fill Deposits
1032
(c)
(d)
C
C
Buried Banza Ridge
700
(d)
600
D
Down Slope Bars or Hanging
Fault Deposits
Sheet Drape Deposits
Neogene Sedimentary Depocenters (East African Rift)
(f)
(Ms-twtt)
800
500
1 km
(b)
C
?
(c)
D
Magara-Banza
Depression
?
Rumonge Channel
A
B
EUF
500 m
CBFe
CMFe
(Ms-twtt)
500
?
BL
(a)
?
600
F
?
(a)
?
D
(f)
(b)
(b)
?
600
A
(a)
B
400
C2
EUF'
Î
B
500
A
500
Capart Channel
CBFw
C2 C1u
400
(d) D
(f)
600
(c)
?
700
800
Precambrian Acoustic Basement
Refer to Table 1 for seismic lines location
and fault abbreviations.
Kigoma half graben (NKHG). The NKHG and SRHG
are separated by the positive structure of the Ubwari
Peninsula, defined also as a high-relief accommodation
zone (Rosendahl et al., 1988).
To the north, the Lake Tanganyika basin extends
onshore through the 100 km–long and 30 km–wide
Rusizi trough (between lat. 3⬚10⬘ and 3⬚20⬘S), which
runs N0 (Rusizi Plain), then N350, and joins the N10–
20–oriented Lake Kivu basin (Ebinger, 1989b) (Figure
1B). The north Rusizi-Kivu region and the Rungwe
volcano north of Lake Malawi are the main volcanic
zones of the western branch of the EARS. The northern volcanic basin formed during three main cycles of
interacting volcanism and faulting: (1) a first stage of
tholeiitic volcanism in the late Miocene (10–6 Ma),
suggesting that the eastern border fault of the Kivu basin formed first (Ebinger, 1989b); (2) a second stage of
alkalic volcanism, dated from 8 to 4 Ma, indicating that
formation of the West Kivu border fault started during
the latest Miocene or early Pliocene (Ebinger, 1989b);
and (3) a third stage (⬍1.9–1.6 Ma) in which basalts
erupted along the West Kivu border fault (Ebinger,
1989b). Such a chronology of events demonstrates that
border fault segments in the north Rusizi-Kivu area
developed diachronously and propagated along the
length of the rift (Bellon and Pouclet, 1980; Ebinger,
1989a, b; Pasteels et al., 1989).
LAKE TANGANYIKA RIFT BASIN
STRATIGRAPHY
Multichannel and single-channel pairs of seismic reflection data from Project PROBE, the Casimir Project,
and the Georift Project have been used to reconstruct
the tectono-sedimentary history of the northern end of
the Lake Tanganyika basin from the late Miocene to
the Holocene. Some of these seismic data sets have
been published previously (Morley, 1988; Rosendahl
et al., 1988; Bouroullec et al., 1992; Lezzar et al.,
1996; Cohen et al., 1997; Lezzar, 1997). Only key features illustrating structural and stratigraphic relationships in the vicinity of major border faults are presented in this article. Figures 3 and 4, however,
illustrate the main seismic sequences and facies deduced from Casimir high-resolution sparker seismic reflection data (K. E. Lezzar, 1997, unpublished data).
Seismic lines P16/S2, P200/S5, and P06/S7 have
been selected to illustrate the structure and stratigraphy of the southern, central, and northern parts of the
northern Lake Tanganyika basin (Figure 3). Multichannel seismic reflection data (Project PROBE; Rosendahl
et al., 1988) show that the synrift fill can be divided
into two main sequences (lower and upper) separated
by a major unconformity, named “(f)” (see table 1 in
Lezzar et al., 1996). These sequences accumulated
above a major surface defined on multichannel PROBE
profiles as the Nyanja event ([NE] surface), likely representing the surface of the prerift basement (Burgess
et al., 1988; Rosendahl et al., 1988). From highresolution sparker reflection seismic data (Casimir Project; Lezzar et al., 1996; Lezzar, 1997), the multichannel upper sequence above the (f) unconformity can be
subdivided into six seismic sequences, designated F–A
from oldest to youngest. These sequences are bounded
by five well-defined hiatal surfaces (which appear to be
mainly erosional surfaces from toplap and indications
of erosional truncations) that are named (from oldest
to youngest) (f), (d), (c), (b) and (a) (Figures 3, 4),
with (e) being a minor, locally represented surface. The
four most common seismic facies signatures seen
throughout the north Tanganyika basin can be interpreted to have formed as follows (Figures 3, 4, 5).
(1) Basin fill units form the base of each sequence and
spread over the underlying, irregularly shaped unconformity. The associated seismic facies are typically
chaotic or weakly stratified, with unit boundaries represented by medium amplitude reflectors with hyperbolic character. (2) Deep sublacustrine fans are mostly
lens-shaped units with a seismic facies that is typically
chaotic. (3) Sheet drape units form the upper parts of
individual sequences. They are expressed by parallel to
subparallel, high-amplitude reflectors that alternate
with thin chaotic to subtransparent layers. These units
also may be represented by another facies type that
consists of subtransparent or transparent layers alternating with thin discontinuous, medium-amplitude
reflectors.
Gravity-flow processes dominate the steep western and eastern borders of the northern Lake Tanganyika rift basin. In the last, upper four sequences (D, C,
B, and A) (Figures 3, 4, 5), proximal and distal sedimentary bodies have been discovered at the outlet (at
the foot) of the main rivers (Lezzar et al., 1996). The
Figure 4. Interpretation of magnified single-channel sparker seismic reflection sections above the (f) surface (location of sections
is shown in Figure 2). Mapping of the basin fill and deep lacustrine fan seismic units within the uppermost sparker sequences (from
D to A) is shown in Figure 5. For abbreviations, refer to Table 1.
Lezzar et al.
1033
?
C.
*
*
?
From Rusizi River
?
?
?
?
50
50
100
100
50
25
A.
25
LUHANGA
LUHANGA
Y
25
75
25
?
75
PEMBA
PEMBA
25-50
Y
?
50
?
25
25
25
?
r iv.Ruzibazi
50
r iv.Ruzib
?
*
Cap Magara
*
?
<25
N
100
N
?
?
10 km
<50
5
25 0
10 km
75
75
50
100
?
25
100
50
?
Cap Magara
50
50
?
?
25
25-50
B.
Y
*
?
*
25
?
?
?
50
1 00
100
25
?
r iv
.M
ure
m
50
Cap
Banza
D.
?
100
*
?
25
*
*
r iv. Bar aka
150
Cape
Banza
?
50
BARAKA
*
50
?
?
BARAKA Y
e
bw
<50
100
75
?
25
?
Y
50-75
75
?
RUMONGE
25
50
25
<25
we
25
r iv
.M
ur e
mb
50
25
25-50
RUMONGE
100
100
?
75
?
75
150
?
LUHANGA
25-50
Y
50
50
LUHANGA
*
50
?
PEMBA
PEMBA
?
Y
>50
25
*
50
?
RUZibazi
N
?
50
50
25
?
*
*
*
25
75
?
75
?
50
100
50
50
<25
*
RUMONGE
r iv
.M
ure
mb
50
25
50
25-50
?
we
RUMONGE
25
?
25
100
1034
?
*
50
<25
?
?
?
Y
*
?
?
25
25
?
?
BARAKA
*
?
<25
?
r iv. Bar aka
10 km
<25
?
?
N
?
10 km
25
?
Cap Magara
*
<25
50
75
*
r iv. Ruzib zi
Cape Magara
25
25
?
Cap
Banza
Cape
Banza
?
Deep Lacustrine Fans
Precambrian Acoustic Basement
Deep Fan Source
Basin Fill Deposits
Channels and Canyons
Sedimentary transit
Neogene Sedimentary Depocenters (East African Rift)
seismic facies is typically chaotic. Moving offshore,
however, the facies becomes roughly stratified then
very stratified in the very distal parts of the fans. The
spatial density of the high-resolution sparker data allowed a strict control of the morphology of the fans
(Figure 4). They are lens- and fan-shaped, and their
maximum dimensions are 15–30 ms two-way traveltime (TWTT) (9–20 m) thick, 10–15 km long, and
2.5–10 km wide. They lie at depths of 150–300 m,
subperpendicular to the shore or parallel with the rift
lake axis. Their position in the seismic sequence is
generally intermediate, in between basal basin-fill
units and sheet-drape units at the top (Figure 4). The
sedimentary processes are linked to the combined
lack of littoral shelf and the presence of steep slopes
and sublacustrine canyons, which allow the eroded
material to be directly transported to the foot of the
slopes and farther into the deepest parts of the basin.
In this case, the variation of the seismic facies along
the different parts of the fans is symptomatic of the
decreasing dynamics during lake level rises. The process of the corresponding deposits is probably coarse
gravity flow. The lens and fan shapes of these sedimentary bodies are due to the recurrent activity of
sublacustrine canyons in front of the main rivers that
provide intermittent detrital input, concentrating
them in a single sedimentary fan. Spatial distribution
through time (from the time of sequence D to the
present-day sequence A) of these fan-shaped coarse
detrital deposits appears to be strongly controlled by
the location and the evolution of submeridian and
transverse faults that structure the north Tanganyika
basin (Figures 4, 5). The length, width, and thickness
of deep lacustrine fans seem to be totally guided by
the direction of sedimentary transits, as well as by
localization of sedimentary traps (nascent depocenters) or sedimentary barriers (active fault hurdles).
Unfortunately, this type of interpretation cannot be
extrapolated below sequence D (sequences older
than 400 ka) because of a lack of high-resolution reflection seismic data.
The ages of these individual sequence boundaries
and the duration of individual depositional sequences
have been estimated within various structurally defined zones. Cohen et al.’s (1993) original methodol-
ogy for RSRM (reexplained in detail in Lezzar et al.
[1996]) was modified to allow for the variability in
sediment accumulation rates at each study site over
time that results from the variability in morphotectonic settings between sites (open basinal; proximal to
fault escarpment; steep channel side; platform–shallow
water) as interpreted from the seismic data (Lezzar et
al., 1996; Lezzar, 1997). These minimum ages are as
follows: about 7.4 Ma for the (NE) surface, about 1.1
Ma for (f), about 0.4 Ma for (d), about 295–262 ka for
(c), about 193–169 ka for (b), and about 40–35 ka for
the most recent surface (a) (Lezzar et al., 1996; Lezzar,
1997). These surfaces have been interpreted in terms
of responses to regional tectonic and volcanic events
and/or regional to global climatic changes (Lezzar et
al., 1996; Cohen et al., 1997; Lezzar, 1997). Other
sequence boundaries may exist in sediments between
7.4 and 1.1 Ma, but these have not yet been defined
seismically, given the available sparker single-channel
and air-gun multichannel resolution and penetration.
Interpretation of seismic sequences/facies and age
estimation of seismic unconformities (deduced from
high-resolution reflection seismic data, as well as sedimentation rates and sediment facies computed from
piston core datations) show that the depositional period of the lower sequence corresponds to the RBM
initial synrift phase (7.4–1.1 Ma) (Lezzar et al., 1996;
Lezzar, 1997). From the (f) surface time at 1.1 Ma, the
five seismic-sequence depositional periods named F-E,
D, C, B, and A have been interpreted in terms of
transgressive-regressive periods characterized by different tectono-stratigraphic conditions. From observations of lacustrine sediment facies (piston core data),
each sequence starts at a low–lake stand period, inducing deposition of coarse detrital basin-fill units on the
basal erosional surfaces. High–lake stand conditions
prevailed during the deposition of the upper two thirds
of each sequence. Deposition at these times comprised
the formation of deep lacustrine fans and sheet drape,
fine-grained units (Bouroullec et al., 1992; Lezzar et
al., 1996; Cohen et al., 1997).
In addition to seismic data, we have attempted to
correlate onshore and offshore structures using Landsat
and SPOT satellite imagery and using a digital elevation model (DEM) (Reynes et al., 1993), geological
Figure 5. Distribution of basin fill and deep sublacustrine fan units within sparker seismic sequences (see also Figures 3, 4).
(A) Sequence D: transgressive phase D following low lake stand (d) at about 0.4 Ma, 350 m below present lake level (bpll). (B) Sequence
C: transgressive phase C following low lake stand (c) at about 0.3 Ma, 350 m bpll. (C) Sequence B: transgressive phase B following
low lake stand (b) at about 0.2 Ma, 250 m bpll. (D) Sequence A: transgressive phase A following low lake stand (a) at about 40 ka,
150 m bpll. Low–lake stand estimations are from Lezzar et al. (1996).
Lezzar et al.
1035
maps (Carte Geologique du Burundi, 1986, 1988,
1989), microstructural field data (Chorowicz et al.,
1988; Ebinger, 1989b; Reynes et al., 1993; Coussement et al., 1994; Coussement, 1995; Theunissen et
al., 1996), and seismicity (De Bremaecker, 1959; Fairhead and Girdler, 1971; Wohlenberg, 1975; Fairhead
and Stuart, 1982; Shudofsky, 1985; Wafula et al.,
1992; Zana et al., 1992).
NORTHERN LAKE TANGANYIKA RIFT
BASIN: MAIN FAULTS, GEOMETRIES, AND
DEVELOPMENT
The northern end of the Lake Tanganyika basin (maximum water depth 320 m) (Figure 2) displays a wide
variety of large-scale geometries, largely controlled by
major border faults, which trend N0–20 and N130–
140 to the rift axis (Rosendahl et al., 1986; Morley,
1988). The offshore fault pattern can be mapped using
multiscaled complementary seismic data (Figures 3, 6).
The RSRM age estimates in the hanging wall adjacent
to the major border faults of the basin provide the first
relative chronology for the evolution of each fault (Cohen et al., 1993; Lezzar et al., 1996; Lezzar, 1997).
Thus, as described in the following sections, the evolution of each half-graben domain is reconstructed using both chronological estimates and geometric information from seismic reflection profiles. The western
limit of the rift basin (Figures 6, 7) is formed by the
N0–20–trending Uvira border-fault system (UBFS),
with the northern segment (UBFSn) represented by
the Uvira escarpment (3300–3500 m maximum altitude) and the southern segment (UBFSs) represented
by the Biera escarpment (⬃2500–2700 m). The eastern limit is formed by the N0-trending Bujumbura
border-fault system (BBFS) that culminates at about
2000–2500 m (Mondeguer et al., 1986; Tiercelin and
Mondeguer, 1991). For all half-graben, fault, and seismic sequence/discontinuity abbreviations, refer to Table 1.
South Rusizi Half Graben
The SRHG (50 km long, 16 km wide, and 3.75 s
TWTT sedimentary thickness) is controlled to the east
by the north-south–trending West Ubwari fault
(WUF) and the associated, oppositely dipping Cape
Magara faults (west [CMFw] and east [CMFe]) (Figure 3E; line P16 in Figure 6). The second dominant
fault family that controlled the SRHG is represented
1036
Neogene Sedimentary Depocenters (East African Rift)
by the N140-trending, southwest-dipping Kaboge
fault zone (KFZ) and Kabezi fault (KAF) (Figure 3E;
lines P14 and P16 in Figure 6). These faults, labeled
transverse faults, crosscut the rift axis without any
present-day morphological expression in the lake bottom bathymetry (Figure 2).
From RSRM age estimates at the foot of the WUF
in the southern area of the NRHG and seismic facies
analysis (Lezzar et al., 1996), the WUF appears to
have acted from 7.4 Ma as a normal fault (line P16 in
Figure 6), as shown by the wedge-shaped geometry of
the first seismic sequence developed on the (NE) surface within the SRHG (Figure 3E) (see also Rosendahl
et al., 1988). Thinning against the WUF and induced
small WUF-synthetic faults are probably due to more
recent tectonic readjustment related to the relative
uplift/subsidence of the adjacent Ubwari horst. In the
northernmost part of the SRHG (near Cape Magara),
the minimum estimated RSRM age for the beginning
of sedimentation above the (NE) surface indicates
that fault movement along the WUF began at about
5 Ma (Figure 3C; line P200 in Figure 6). This northern
part of the WUF has been relayed to the east by the
CMFw between about 5 Ma (corresponding to the age
of the oldest sediments deposited on the [NE] surface) and about 1.1 Ma ([f] surface), giving an average
age of about 3 Ma for the SRHG at that location (lines
P200 and S2 in Figure 6). After about 3 Ma, the WUF
in this area was buried, and the CMFw became the
eastern border fault of the SRHG (line P200 in Figure 6).
Sequence and fault geometry along the transverse
KFZ indicates that initial fault movement (evidenced
at 2 s TWTT) (Figure 3E; line P16 in Figure 6) was
apparently dominated by normal displacements. From
the (f) surface up to the (d) surface, a dome-shaped
sequence (positive flower structures identified by Rosendahl et al. [1988] and named by Lezzar et al.
[1996] the Kaboge and Kabezi domes) possibly indicates an oblique component of displacement along
those transverse faults. Northward, the transverse
KAF seems always to have acted as an oblique strikeslip fault, as shown by the positive flower structure
and associated slight doming of the (f) surface (the
Kabezi dome) (line P14 in Figure 6). Those positive
flower structures are better represented on a largescale seismic atlas published by Project PROBE (Rosendahl et al., 1988). Because of the great amount of
seismic data required to illustrate every single fact in
this article, we deliberately have provided only detailed line drawings, and we strongly recommend con-
sulting key articles by Morley (1988), Rosendahl et al.
(1988), and Lezzar et al. (1996) for greater detail. We
chose not to overrepresent already published seismic
lines in this article but to focus on facts that illustrate
our main topics, such as rift kinematics, coarse detrital
fan distribution, and fault correlation lessons. To complete the data published by Morley (1988), Rosendahl
et al. (1988), and Lezzar et al. (1996), however, a few
unpublished high-resolution seismic reflection lines
from the Casimir Project are illustrated in Figures 3,
4, and 6.
Ubwari Horst Fault System
The N0–20–trending Ubwari horst (33 km long and
13 km wide) is delineated by the East Ubwari fault
(EUF) and the WUF (Figure 3E). The internal structure of this horst is mainly controlled by two N0–20–
trending west-dipping and east-dipping faults called
the Cape Banza fault western segment (CBFw) and
the Cape Banza fault eastern segment (CBFe). The
northern sublacustrine continuation of these faults
controls the deepening of the top of the Ubwari horst
(Figure 2), whereas their southern extent borders the
Ubwari Peninsula (Figures 3, 6). The Kabezi and Kaboge transverse faults also affect the Ubwari horst.
The KAF shows a slight normal component of movement, indicated by the (NE) surface vertical displacement, whereas the upper part of the fault is characterized by a positive flower structure (dome-shaped
zone) around the (f) surface (line P14 in Figure 6).
This type of dome morphology is also observed in the
upper part of the KFZ between the (c) and (b) surfaces (lines S2A and P16 in Figure 6). As in the
SRHG, these dome-shaped zones (flow structures)
might indicate an oblique component of movement
along the transverse KAF and KFZ. The RSRM age
estimates for the prerift (NE) surface indicate the
Ubwari horst formed about 4.9 Ma or slightly earlier
(Lezzar et al., 1996).
North Kigoma Half Graben
The NKHG (30 km long, 8–14 km wide, and 1.3 s
TWTT in sedimentary thickness) is bounded to the
west by the N0–20–trending EUF (Figure 3E). The
eastern border of the NKHG is marked by the N350
to N0–trending, eastward-dipping synthetic Rumonge
fault (RMF) (Figure 3E; line S2B in Figure 6). The
RSRM age estimates (Lezzar et al., 1996) indicate
that the NKHG formed about 3.6 Ma or slightly ear-
lier (in the hanging wall of the EUF) and was actively
subsiding until about 0.2 Ma ([b] surface), as shown
by the wedge-shaped geometry of seismic reflectors
below the (b) surface, in contrast with the parallel
geometry that characterizes the upper seismic sequences B and A in the Rumonge channel infill (Figure 3E; line S2B in Figure 6).
North Rusizi Half Graben
The architecture of the NRHG (35 km long, 24 km
wide, and 4 s TWTT in sedimentary thickness) is controlled in its northern part by the N0-trending UBFSn
(lines P6 and P228 in Figure 6). The NRHG is also
transversally controlled in its southern part by two
N140-trending faults, the north-eastward–dipping
normal Pemba-Magara and Luhanga-Magara faults
(PMF and LMF, respectively) (Figure 3A; lines P6, S7,
P5B, P200, S6, and S8 in Figure 6). Justification for
projecting those two faults from one side of the lake
to the other is their importance onshore around the
villages of Pemba and Luhanga. Indeed, northwestsoutheast escarpments are extremely steep and relatively fresh within the Precambrian basement rock
(Tanganydro Group, 1992; Tiercelin et al., 1993).
The seismic data (lines P6 and P200), however, suggest that LMF and PMF are not well expressed in their
southeastern tips, close to Cape Magara. These two
transverse faults have throws ranging from 4 to 0 s
TWTT along strike. Maximum displacements occur
along their northwestern segments, close to the
UBFSn (Figure 3B). Minimum displacements, almost
close to zero, are observed to the southeast, close to
Cape Magara (Figure 3C). These observations show
that the LMF and PMF delineate southwest-tilted
basement blocks that are progressively buried southeastward under the synrift infill. In addition to the
UBFSn, however, the LMF transverse fault also appears to act as a major fault, which accommodates
much of the subsidence of the NRHG. This is shown
by the wedge-shaped geometry of seismic reflectors
(4 s TWTT) between the (NE) and (f) surfaces developed adjacent to the LMF (lines P6 and S7 in Figure 6). The second transverse fault, PMF, offsets the
entire sedimentary section but with lesser displacement (⬃0.5 s TWTT) (Figure 3A; lines P200, S6, and
S8 in Figure 6).
An RSRM age estimate at the foot of the LMF (the
deepest part of the NRHG, having maximum sedimentary thickness of 4 s TWTT) (line P6 in Figure 6)
suggests that this transverse border fault acted mainly
Lezzar et al.
1037
1038
Neogene Sedimentary Depocenters (East African Rift)
Lezzar et al.
1039
Figure 6. Structural and stratigraphical interpretation of magnified seismic line sections in the northern Lake Tanganyika rift basin (modified after Rosendahl et al., 1986; Lezzar
et al. 1996; Lezzar, 1997). Lines P ⳱ multichannel seismic reflection lines from Project PROBE; lines S ⳱ single-channel sparker seismic reflection lines from the Casimir Project.
For abbreviations, refer to Table 1.
BBFS
n
UBFSn
A.
BBFSn
Rusizi Plain
10 km
X⬘
UBFSn
4
N
X
LM
Lake
F
Lacustrine
sediments
and alluvium
8B PMF
8A
CM
Fw
Basement
rocks
BK F
WUF
Biera escarpment
1
2
UBFSs
EUF
3
F
KF
Z
RM
CMFe
F
KA
BBFSs
Figure 7. (A) Structural interpretation of the Landsat and
SPOT images and the correlation of these features with the
major fault pattern deduced
from seismic data (see Figures
3, 6). (B) Vertically exaggerated
cross section of the northern
part of the NRHG derived from
the SPOT digital elevation
model (Reynes et al., 1993) and
seismic line P228 (modified
from Rosendahl et al., 1988).
For abbreviations, refer to
Table 1.
Northern part of
the subsiding NRHG
UBFSn
East Burundian escarpment
(tilted blocks related to
flexural margin response)
X⬘
(Line P 228)
750 masl
0
1
UBFS
BBF
Sn
3000
2500
2000
1500
1000
(M)
X
n
(NE)
Sn
3500
Uvira escarpment
(major tilted blocks)
BBF
B.
2
3
4
Sec
(twtt)
Precambrian
Basement
10 km
Vertical Exaggeration: x5
as a normal fault from about 3.5 Ma to the (f) surface
time (between ⬃1.1 and ⬃0.4 Ma, age uncertain because of the lack of sediment accumulation rate data
in this area) (Lezzar et al., 1996). During this phase,
from 3.5 Ma, subsidence in the NRHG thus mainly
was controlled to the west by the UBFSn. Prior to formation of the (f) surface, the southern end of the
1040
Neogene Sedimentary Depocenters (East African Rift)
NRHG was also controlled by the transverse LMF, as
shown by the 3 s TWTT–thick fan-shaped strata deposited against this fault. After the (f) surface time, the
LMF was sealed by a major overlapping basin that was
controlled by the upper part of the PMF associated
with the UBFSn (lines P6, S7, and S8 in Figure 6).
Since this period, the LMF has been quiescent, as dem-
Table 1. Abbreviations
Half Grabens and Horsts
SRHG
NRHG
NKHG
UBW
South Rusizi half graben
North Rusizi half graben
North Kigoma half graben
Ubwari horst
Major Border Faults
UBFSs
UBFSn
BBFSs
BBFSn
WUF
EUF
WRF
ERF
LMF
PMF
Uvira border-fault system south
Uvira border-fault system north
Bujumbura border-fault system south
Bujumbura border-fault system north
West Ubwari fault
East Ubwari fault
West Rusizi fault
East Rusizi fault
Luhanga-Magara fault
Pemba-Magara fault
Seismic Discontinuities
NE
(f) to (a)
Nyanja event/acoustic basement (Project PROBE)
Seismic unconformities (Casimir Project)
Seismic Sequences
RL
RU
F to A
Rusizi lower
Rusizi upper
Sequences F to A (Casimir Sparker Project)
Other Faults
BKF
CMFw
CMFe
CBFw
CBFe
KAF
KFZ
RMF
Baraka fault
Cape Magara fault west
Cape Magara fault east
Cape Banza fault west
Cape Banza fault east
Kabezi fault
Kaboge fault zone
Rumonge fault
onstrated by the overlap of the post–(f) sequences (line
S7 in Figure 6). Some slight readjustment along the
LMF is indicated by channeled coarse sediments deposited after the (d) unconformity along the hanging
wall (northeast side) of the LMF (Figure 3A; line S7
in Figure 6). Synchronous with the deposition of
these strata, a minor (0.3 s TWTT thick) sedimentary
package developed on the PMF/UBFSn–controlled,
perched tilted block (lines P6 and S8 in Figure 6).
ONSHORE-OFFSHORE STRUCTURAL
CORRELATION AND THE INHERITANCE
FACTOR
Basement fabrics strongly influence rift geometry, as
demonstrated, for example, by the location of the
EARS around the edge of the Archean Tanzanian craton (McConnell, 1972; Sykes, 1978; Daly et al., 1989).
At a smaller scale, the extent of rift faulting outside the
Lake Tanganyika region across the broad uplifted
flanks of the rift is poorly understood.
The onshore fault pattern, established from Landsat and SPOT imagery and field geology, indicates two
predominant trends, N0–20 and N130–140 (Figures 7,
8). The N0–20–trending faults are dominant and
clearly observed on both sides of the lake. An east-west
cross section in the northern part of the basin has been
constructed using DEM (Reynes et al., 1993), topographic maps, and the interpretation of the P228 seismic line (line P228 in Figure 6; Figure 7A, B). This
cross section shows the characteristic asymmetric geometry of the two margins of the basin. The western
margin is formed (Figures 7, 8, 9) by large, elevated
(up to 3000 m above sea level [masl]), tilted basement
blocks related to the N0-trending UBFS. The UBFSn
is represented on land by the Uvira escarpment (3400
masl altitude), which constitutes the Lake Tanganyika
shoreline in the northern part of the study area. Toward the north, this system extends onto land by 3–4
northeast–striking normal faults, each with throws of
several kilometers. These faults collectively form the
East Rusizi and West Rusizi faults (ERF and WRF, respectively), which bound on its western side the Rusizi
trough (Ebinger, 1989a, b).
To the south, the UBFSs corresponds to the major
N0-trending Biera escarpment (2500–2700 masl). This
escarpment lies onshore between lat. 4⬚ and 4⬚40⬘S, 20
km to the west of the parallel Baraka fault (BKF),
which in turn forms a linear, weakly eroded escarpment culminating at about 2000 masl a few kilometers
west of the lake shoreline (Figure 9). The eastern margin of the northern end of the Tanganyika basin consists of several, N0–20–trending, antithetic normal
faults that form the BBFS (⬃2500 masl altitude) (Figure 9). The northern segment (BBFSn) is represented
by N0–20–trending faults that resulted from the reactivation of structural elements belonging to the
Lezzar et al.
1041
N
A.
LM
F
on
5 Km
lan
ion
?
Lake
Tanganyika
F
nd
n
io
ns
UBFSn
Uvira escarpment
te
ex
?
LANDSAT
lineaments
UBFSn
la
on
PM
F?
115
-
F
LM
Normal
faults
80
100
SPOT
lineaments
Subvertical
foliation
trend
Neogene Sedimentary Depocenters (East African Rift)
120
Pemba
hydrothermal
area
N90-110
direction
extensio of
n
N
100
Luhanga
Kibaran belt (Klerkx et al., 1998; Carte Geologique du
Burundi, 1986, 1989). The southern segment (BBFSs)
is characterized by north-south–, north-northwest–,
and north-northeast–trending faults, superimposed on
mylonitic zones of Kibaran age as a result of Cenozoic
extension (Carte Geologique du Burundi, 1988) (Figure 7).
Transverse trends can be observed clearly on the
western shore of the lake, where microstructural field
work has been conducted, associated with Landsat and
SPOT imagery interpretations. Closely spaced N115–
120–trending faults or lineaments locally interact with
en echelon N0 normal faults belonging to the UBFSn
(Figure 8). This pattern of transverse faults coincides
with the general trend of the northwest end of the Tan1042
B.
UBFSn ?
ns
xte
de
Lake
Tanganyika
PM
Figure 8. (A) Structural interpretation of the compilation of
Landsat and SPOT images on
the western shore of Lake
Tanganyika and correlations
with field data (foliation measurements) and offshore faults.
Interpretation of transverse
lineaments and Z-shaped faults
shows a distribution characteristic of one major transverse dextral slip zone of deformation.
This zone is correlated with the
LMF and PMF offshore fault
areas. Foliation measurements
(Coussement, 1995) in the delineated area (B) trend parallel
to transverse faults. Foliation
planes and faults may have
been reactivated to create the
present-day observed transverse offshore faults and onshore lineaments. (B) Sketch illustrating the normal and
transverse fault intersection at
the Pemba hydrothermal site
(Coussement et al., 1994). The
N115–120–trending fault shows
a normal displacement with a
small dextral component, compatible with an extension direction close to N90–110. This
transverse fault trends parallel
to onshore basement foliation
planes and is inferred to be inherited. For abbreviations, refer
to Table 1.
PM
B
115
-
F
1 Km
1 Km
120
Pemba
area
ganyika-Rukwa-Malawi (TRM) fault zone (Chorowicz
et al., 1983; Tiercelin et al., 1988), which is superimposed on the Paleoproterozoic Ubendian belt (Theunissen et al., 1996). No relative chronology can be obtained using imagery or field data. At Luhanga (Figure
8A), one normal fault with a trend of N0–20 shows an
S-shaped trace between two transverse lineaments and
may result from a dextral component of movement
along the transverse trend. This interpretation is consistent with a previous study carried out at the Pemba
hydrothermal site (Tanganydro Group, 1992; Tiercelin
et al., 1993; Coussement et al., 1994) (Figure 8B),
which demonstrated that transverse faults are characterized by predominantly normal displacements with
a small dextral component (Figure 8B). These dextral
Major and minor faults
NORTH RUSIZI
HALF GRABEN
Bujumbura
RSRM dates
774 m
UBFSn
7.4 Ma
er
Rusizian Belt
Riv
ER F
Kibarian Belt
WRF
Rusizi delta
plain
Main initial depocenter
3.5/5 Ma
BBFSn
3000 m
(asl)
2760 m
Figure 9. Schematic structural
map of the northern Lake
Tanganyika rift basin showing
location of the three main half
grabens and their associated
major border faults and initial
depocenters. See also Figure 3
to locate the RSRM-dated seismic sequence boundaries in
each half graben. For abbreviations, refer to Table 1.
N
Luhanga
F
F
LM
PM
10 km
Pemba
Cap Magara
5 Ma
CM
Fe
BBFSs
KA
F
Ubwari
CM Fw
BKF
7.4 Ma
CBFw
Baraka
Rumonge
2500 m
EUF
Ubwari
peninsula
WUF
UBFSs
SOUTH RUSIZI
HALF GRABEN
4.9 Ma
CBFe
Biera escarpment
F
RM
Z
KF
3.6 Ma
oblique-slip movements on N130–160 transverse
faults are compatible with a regional direction of extension close to N90–110 (Morley, 1988; Boccaletti et
al., 1994; Coussement et al., 1994; Coussement,
1995).
These onshore structural interpretations correlate
well with the offshore fault pattern. The WUFcontrolled SRHG is delineated on its eastern margin by
the onshore N0-trending BKF (Figure 9), which can be
interpreted as a normal fault antithetic to the WUF,
resulting from the SRHG flexural margin response.
Along with the major Biera escarpment, the BKF and
NORTH KIGOMA
HALF GRABEN
WUF appear to relate to the general north-south trend
of the Mesoproterozoic southwestern Kibaran belt
(Figures 1, 6, 7, 8). In the central part of the SRHG,
the N140-trending KFZ and KAF may correlate with
onshore N150–170, southwest-dipping faults that belong to the fault trend of the Paleoproterozoic Ubendian belt (Figures 1, 5A). To the east, the EUFcontrolled NKHG is bounded on its eastern shoaling
margin by antithetic faults that belong to the southern
segment of the BBFS, resulting from the reactivation
of structures belonging to the northeastern Kibaran
belt (Figures 1, 7A, 9).
Lezzar et al.
1043
Within the NRHG, the offshore UBFSn, identified on seismic lines (lines P6 and P228 in Figure 6),
is clearly associated with the N0-trending system of
tilted blocks forming the Uvira escarpment (Figures
7B, 8A). On the eastern shoaling margin of the half
graben, minor N0-trending faults crosscutting the sediment pile also belong to the northern BBFS (line
P228 in Figure 6). At the south end of the NRHG,
the LMF, which appears from seismic data to be a
major transverse border fault (Figures 3A, 9), seems
to be aligned with the well-expressed transverse normal fault segments observed onshore (Figure 8A). The
offshore PMF, which does not exert a major control
on the NRHG depocenter, as compared with the LMF
(Figures 3A, 9), seems also to be less expressed onshore (Figure 8A).
In northern Lake Tanganyika, earthquake activity
is important for understanding the recent evolution of
this complex fault system. The present-day seismicity
shows that transverse faults are still active. Within the
overall north-south seismic trend of the rift, the epicentral distribution of the magnitude 3 or greater
earthquakes in the Luhanga-Pemba area suggests an
alignment trending N145, which correlates well with
the offshore-onshore transverse pattern of the southern
end of the NRHG (De Bremaecker, 1959; Coussement
et al., 1994). These observations are reinforced by the
bathymetric map (Figure 2), which shows a transverse
corridor acting as a slight barrier by controlling sublacustrine channel location (central deep basin on line
S7 in Figure 3A). This control on recent sedimentation
may be due to recent slight normal reactivation of the
LMF and PMF. For large, normal-faulting earthquakes
(magnitude 5 or greater), fault segments of tens of kilometers may be activated with vertical coseismic deformation on the order of tens of centimeters (King et
al., 1988). The swarm of seismic events recorded along
the transverse trends in the Luhanga-Pemba area may
have continuously reactivated the offshore segments of
these faults, producing significant vertical offsets,
thereby controlling the channeled systems described
previously.
LATE MIOCENE–HOLOCENE KINEMATIC
MODEL
Previous articles on rift basins have concentrated principally on their geometric characteristics (Patterson,
1983; Reynolds, 1984; Ebinger et al., 1987; Leeder and
Gawthorpe, 1987; Rosendahl, 1987; Milani and Da1044
Neogene Sedimentary Depocenters (East African Rift)
vidson, 1988; Rosendahl et al., 1988; Ebinger, 1989a,
b; Morley et al., 1990; Stock and Hodges, 1990; Peacock and Sanderson, 1991; Nelson et al., 1992; Gawthorpe and Hurst, 1993; Karner and Driscoll, 1993;
Childs et al., 1995; Mack and Seager, 1995). Recent
articles by Morley (1999) and Gupta et al. (1999) provide insights regarding fault evolution in the East African rift and the Suez rift. We propose to follow a
comparable strategy to model the structural evolution
of the kinematics of the northern Lake Tanganyika rift
faults. Using new RSRM age estimates (Figures 3A, 9)
and integrating all the observations described in the
preceding sections, we have constructed a kinematic
model for the structural and stratigraphic evolution of
the northern end of the Lake Tanganyika basin since
the late Miocene (Figure 10). In the following sections,
we discuss the initiation and development of rifting in
detail to understand how fault configurations and interactions influence the initiation and/or cessation of
subsidence within individual half grabens.
From before 12 to 7.4 Ma
The oldest age for the Lake Tanganyika basin, estimated by Cohen et al. (1993) using the RSRM in
the central part of the basin, was slightly before 12
Ma (Figure 10A). Other RSRM ages calculated in
various parts of the basin indicated a northward and
southward migration of extension from the central
basin. From tectono-volcanic studies in the KivuRusizi basin north of Tanganyika (Ebinger, 1989a, b;
Pasteels et al., 1989), this area appears to have been
affected by continuous, intense volcanic activity between 10 and 4 Ma. As a consequence, the presence
of magmatic heat may have resulted in increasing
ductility and a reduced brittle thickness of the upper
crust in this area (Weissel and Karner, 1989; Van
Wyk de Vries and Merle, 1996). During this period,
magmatic pressure also may have induced updoming
of the area, as suggested by Saggerson and Baker
(1965) for the updoming of central Kenya. This resulted in the formation of an uplifted, about 100 km
in diameter dome, referred to as the Kivu-Rusizi local dome (Figures 1, 10B). This dome is interpreted
by Coussement (1995) as a local uplift due to magmatic underplating, or it may be a possible magmatic
inflation of the crust, as suggested by Morley (C. K.
Morley, 1998, personal communication). In both
cases, the North Tanganyika–Kivu area is one of the
most elevated areas of the East African rift western
branch, which is a key element in the control of
A.
N
sc
ar
pm
en
t
E-W
Bi
er
aE
Ss
BF
U
?
Regional
extension
Fault
propagation
Z
Local
extension
?
KF
25 Km
D.
N
LM
U
BF
Sn
F
PM
F
E-W
Local
extension
Regional
extension
KA
BK
Kivu
-Rusizi Local
F
F
Z
KF
N
Ss
F
BF
EUF
U
WU
B.
Dom
e
25 Km
E-W
KA
Initial
depocenter
Extension offset
Local
extension
Regional
Extension
25 Km
South Rusizi
Half Graben
SRHG
E.
Kiv
NW-SE
n
FS
B
U
?
Regional
extension
Local Do
me
u-Rusizi
BB
FSn
Z
WU
F
F
KF
Ss
BF
U
N
?
SRHG
F
LM
F
PM
South Rusizi
Half Graben
SRHG
Ss
25 Km
Transfer of extension
Local
extension
Regional
extension
EUF
Z
KF
F
BK
F
KA
Ss
BF
U
WU
F
?
Local
extension
NKHG
North Rusizi
Half Graben
NRHG
E-W
UBH
B BF
C.
?
F
K
B
EUF
s
FS
B
U
WU
F
N
North Kigoma
Half Graben
NKHG
25 Km
Figure 10. Late Miocene–Holocene kinematic model of the northern Lake Tanganyika rift basin, deduced from complementary data
in the lake basin and on the rift shoulders and showing chronology of fault interaction and depocenter initiation. For abbreviations
of names of faults and half grabens, refer to Table 1. (A) From about 12 to about 7.4 Ma; (B) from about 7.4 Ma; (C) from about 5
Ma; (D) from about 3.5 Ma; (E) from about 1.1 to about 0.4 Ma. Rift basin stratigraphy above the (d) surface (⬃0.4 Ma) is shown
in Figures 3, 4, and 5. Refer to Table 1 for all abbreviations.
sedimentary transits, as shown by Lezzar et al.
(1996) and Cohen et al. (1997). During this active
volcanic period, based on our stratigraphic and kinematic models (in this article and in Lezzar et al.
[1996]), conditions for brittle tectonics appear to be
unfavorable in this area. General concepts of rock
physics in this case state that brittle tectonics still
exist for the upper crust, even with a high heat flow.
Those concepts also state that brittle crust gets thinner but still fractures and behaves in a brittle way.
Because the brittle layer is the main load-bearing part
of the crust, if it becomes thinner because of heating,
the stress per unit area increases, and, hence, the
chance for brittle failure increases. In our case, faulting is not totally absent during volcanic crises (magmatic periods, high heat flows) but is less dominant
than at other periods. We think that faulting exists
during volcanic periods, as mentioned by Ebinger
(1989b), but our study shows that faulting is less expressed than in periods in which brittle crust is cold
and thicker. The complexity of interaction between
forcing factors on rift processes like magmatic heat,
rift extension direction change, and preexisting basement faults during a relatively short period of time,
such as 12 m.y., allows us to state that basic and
general rock physics concepts cannot be applied
blindly and literally in our case study. We have an
example, based on strong stratigraphic and fault kinematics evidence, that shows that in a rift basin like
the north Lake Tanganyika trough, fault propagation
seems to be more favorable and faster in between
periods of active magmatic heat flows and related
volcanic crisis.
In addition, assuming that passive rifting and necking are the basic mechanisms of continental rupture,
thermomechanical modeling shows that within a cold
and, therefore, strong lithosphere undergoing extension, such as those of the cratonic area of north Tanganyika, strain tends to concentrate over a narrow zone
(Bassi, 1991; Buck, 1991). Also demonstrated are variations of the strength profile of the lithosphere during
extension, which can occur as a result of partial melting
at depth. Here the rheology and mechanical behavior
of the stretched lithosphere changes with time, leading
to a decrease of the brittle:ductile ratio in the thickness
of the crust. The resulting weakening and softening of
the crust are known to be favorable conditions for widening the rift zone, as well as accelerating extension
(Bassi et al., 1993). Morley (1995) demonstrated these
effects in the Kenyan rift valley and indicated that rifting during intense volcanic activity induces numerous
1046
Neogene Sedimentary Depocenters (East African Rift)
smaller, shallower, and lower angle border faults in rift
basins than during nonvolcanic periods.
Unfortunately, the lack of deep geophysical data
prevents confirmation of localized and anomalously
hot material at depth beneath the Kivu volcanic rift
and northern Lake Tanganyika areas. Based on our
stratigraphic and kinematic models in north Tanganyika (in this article and in Lezzar et al., 1996; Cohen
et al., 1997), it appears that, during periods of no volcanism, preexisting border faults continued to break in
a colder upper crust. The extension rate probably
slowed during intervolcanic phases, but, as those faults
corresponded to weakened zones, even a limited extension induced subsidence on those faults and, thus,
was favorable for creating large and deep depocenters,
as observed today in the northern Lake Tanganyika rift
basin.
From before 12 to about 7.4 Ma, northward migration of extension from the central basin appears to
have been associated with the formation of a series of
opposite polarity half grabens, controlled by the major
border-fault systems defined by Rosendahl (1987) and
Rosendahl et al. (1988). Immediately west of the
present-day northern end of the Lake Tanganyika basin, the Biera escarpment (UBFSs in Figures 9, 10A)
represents one of these middle–late Miocene major
border-fault systems, possibly bounding a west-dipping half graben (Figure 9). The flat-bottomed,
present-day morphology of the Biera escarpment
hanging wall, occupied by river networks and extensive marshes, may be analogous to the final phase of
subsidence activity of this ancient (Miocene?) sedimentary basin (J. J. Tiercelin and A. Mondeguer,
1991, personal communication). Thus, the Biera escarpment/UBFSs can be interpreted as younger than
about 12 Ma (the oldest RSRM age estimate of the
Lake Tanganyika basin) and older than about 7.4 Ma.
As a consequence of a northward migration of extension (Cohen et al., 1993; Cartwright et al., 1995;
Klerkx et al., 1998), the UBFSs propagated in the
same direction but became locked at its intersection
with the onshore, basement-related, northwest extension of the southwest-dipping KFZ (Figure 10A).
These fault interference mechanisms are discussed in
further detail at the end of this article.
From about 7.4 Ma
The RBM initial synrift phase of Lezzar et al. (1996),
identified by interpreting the PROBE multichannel
seismic lines, began about 7.4 Ma (Figure 10B) in the
north Lake Tanganyika basin with the initiation of the
N0-trending WUF and the associated SRHG (Figure
3E). Subsequent to the locking of the UBFSs, the
maximum local extension was offset eastward, resulting in the formation of the N0–20–trending WUF by
reactivation of preexisting Kibaran fabrics. Contemporaneously, the extensional strain was also applied to
Kaboge and Kabezi transverse faults, interpreted as
the trace of Rusizian fabrics, as foliations or fault
planes (Figure 9). The initial interaction between the
convergently dipping WUF and KFZ created subsidence of a fault-controlled, triangular-shaped block,
resulting in the initial depocenter of the SRHG (Figures 3E, 10B).
A minimum RSRM age of about 5 Ma can be
estimated for the beginning of sedimentation above
the prerift (NE) surface north of the Kaboge and Kabezi transverse faults. This suggests northward propagation of the WUF across the KFZ and KAF, which
rendered the SRHG an elongate depocenter. The
SRHG retained the WUF as its major border fault,
whereas the transverse faults no longer exerted an influence on depocenter geometry. This structural interpretation concerning SRHG development from
about 7.4 to about 5 Ma suggests that this area was
out of the thermal influence of the Kivu-Rusizi local
dome during this volcanically active period and, thus,
susceptible to brittle extension (Figure 10B). This
confirms the northward migration of brittle extension
hypothesis suggested by Cohen et al. (1993) and
Klerkx et al. (1998). Morley (1995, 1999) showed in
the Kenyan rift that volcanism forced changes in
structural style with time and stopped large boundary
fault development.
From about 5 Ma
From about 5 Ma, as a result of the gradual cessation
of volcanic activity in the Kivu-Rusizi area (Ebinger,
1989a, b; Pasteels et al., 1989), extension gradually
migrated northward into an area that was becoming
favorable for brittle deformation (Figure 10C). The
northward migration of extension between the northern end of the SRHG and the NRHG area started at
about 5 Ma with an early reactivation of the N0trending Kibaran and N140-trending Ubende fabrics.
From about 3.5 Ma
Rift development of the phase from about 5 Ma resulted in the development of the observed UBFSn,
LMF, and PMF from about 3.5 Ma. At about 3.5 Ma
(RSRM age) the NRHG formed, as a consequence of
fault interaction between the convergently dipping
UBFSn and LMF, which act as major, normal, longitudinal and transverse border faults, respectively
(Figure 10D). Slightly to the south, the depocenter
located between the transverse PMF and LMF became an abandoned, perched, depocenter (line S8 in
Figure 6). Thus, the LMF northeast-dipping transverse border fault acted as a transfer fault zone,
which allowed rifting to propagate northwestward
(Figure 10D) and not northward, as in the case of
the WUF between about 7.4 and about 5 Ma. This
type of transfer fault mechanism is discussed in terms
of correlation between fault intersection angles and
dips in further detail in a subsequent section of this
article.
Geographically independent of but synchronous
with the northward propagation of faulting, an eastward migration of subsidence occurred in the SRHG
with the development of the EUF (Figure 10D). This
is suggested by the youngest RSRM age estimates
found on the western side of the Ubwari horst (⬃4.9
Ma; [NE] surface) (Figures 3, 9) and on the upper
part of the NKHG (⬃5–3.5 Ma; [NE] surface) (Lezzar et al., 1996).
From about 1.1 to about 0.4 Ma
The second major rift phase (the F-E synrift phase
from ⬃1.1 to ⬃0.4 Ma) represents the region’s most
recent major tectonic episode (Lezzar et al., 1996;
Cohen et al., 1997). This phase started with the
same structural and volcanic context as described for
the beginning of the previous RBM phase (between
about 7.4 and about 1.1 Ma). At before 1.9–1.6 Ma,
renewed volcanic activity in the south Kivu area, representing the last volcanic phase in the Kivu-Rusizi
region (Bellon and Pouclet, 1980; Ebinger, 1989a, b;
Pasteels et al., 1989), possibly produced topographic
doming across the whole NRHG. Evidence for this
uplift is suggested by the presence of the important
erosional surface dated at about 1.1–0.4 Ma (the
KMSB surface of Rosendahl et al. [1988] and the [f]
surface of Lezzar et al. [1996]), during which time
the area was probably elevated above lake level (Figures 3, 10E). The lack of extension-induced subsidence in the north is similar to that observed between about 7.4 and about 5 Ma and is again
attributable to the thermal uplift and corollary crustal
state in the region. Thus, normal faulting appears to
Lezzar et al.
1047
be especially concentrated in the SRHG, particularly
along the major WUF, where the oldest RSRM age
of the (f) surface (⬃1.1 Ma) is calculated (Lezzar et
al., 1996) (Figure 10E).
After the cessation of this last volcanic period and
related doming (phases F-E–A of Lezzar et al.
[1996]), significant vertical movement occurred primarily along the main normal rift border faults and
some normal transverse faults: the WUF, EUF, and
UBFSn. Half-graben development along the WUF induced a flexural margin response, resulting in the development of the onshore, north-south–trending, eastward-dipping BKF (Figures 3, 9, 10E). This feature
corresponds to the present-day Baraka escarpment.
Likewise, half-graben development and deepening
along the UBFSn and the EUF initiated by flexural
margin response (uplifted margin consequent to basin
subsidence) the north-south–trending BBFSn and
BBFSs, which represent the east Burundian rift escarpment (Figures 3, 9, 10E). This major tectonic activity (between ⬃1.1 and ⬃0.4 Ma) along the WUF
and EUF induced the clear separation of the SRHG
and NKHG by the development of the present-day
Ubwari horst (Figures 3E, 10E).
We have emphasized in preceding sections that,
during the RBM phase (from ⬃3.5 to ⬃1.1 Ma), the
southern border of the NRHG was represented by the
N140-trending reactivated LMF. In the SRHG, the
Kaboge and Kabezi transverse faults, in contrast to the
LMF and PMF, appear to have accommodated important intrabasinal strike-slip motion at around 1.1
Ma without any influence on the deepening of the
SRHG, as shown by the dome-shaped seismic sequences (around the [f] surface and dated ⬃1.1 Ma)
(Figure 3E) that form positive flower structure geometries (the Kaboge and Kabezi domes) (Figures 3C,
E; 9).
To the north, in the NRHG, another domeshaped structure, called the Magara dome, affects all
of the sedimentary pile above the (f) surface (between
⬃1.1 and ⬃0.4 Ma) in the area of Cape Magara (the
southeast end of the NRHG) (lines P200 and P6 in
Figure 6). This structure is located at the intersection
of the N140-trending LMF with the N0–20–trending
WUF and can be interpreted as the result of slight
strike-slip movements at the southeast tip of the LMF.
Similar movement also has been described on land in
the Pemba-Luhanga region, at the northwest tip of
the offshore LMF, in the form of strike slip (dextral)
affecting Precambrian rocks (Coussement et al., 1994)
(Figure 8A, B).
1048
Neogene Sedimentary Depocenters (East African Rift)
The existence of strike-slip motion at the northwest and southeast tips of the LMF suggests a major
change in faulting mechanisms that resulted in the
cessation of the LMF border-fault-like activity after
the (f) surface time. This is demonstrated by southward overlap of sediment observed on seismic lines
(Figure 3A; line S7 in Figure 6), which also indicates
a decrease in subsidence in the NRHG. Nevertheless,
fault-controlled channels identified on sparker seismic
line S7 east of the LMF suggest persistent, slight normal movements (Figure 3A), possibly expressed by
the present-day local earthquake activity. From about
0.4 Ma, subsidence in the SRHG appears to have
stopped, as indicated by the parallel geometry of seismic reflectors dated as post–(d) surface (Figures 3E,
4; line S2 in Figure 6). From about 0.2 Ma, subsidence
in the NKHG also stopped, as indicated by post–(b)
surface reflector geometry in the Rumonge channel.
To the north, slight normal components of displacement are observed along the transverse PMF and the
UBFSn, which are seen to control the present-day lake
bottom morphology (Figure 3A).
FORCING FACTORS CONTROLLING RIFT
BASIN EVOLUTION
The northern basin of Lake Tanganyika provided a
unique opportunity to observe the early stages of rift
basin development. We have demonstrated that elementary half-graben structures are delineated by two
main types of rift faults. The first type consists of longitudinal faults that trend N0–20 and form the
present-day asymmetric rift shoulders. Two subtypes
are distinguished based on their chronology and development mechanisms: major normal border-fault
segments that control half-graben subsidence and minor normal border-fault segments that result from
flexural processes in response to subsidence along the
major border faults. The second type consists of transverse faults that trend N130–140. Two subtypes are
identified: pure normal border faults that also control
half-graben subsidence and strike-slip faults that have
a slight normal component along which half-graben
subsidence progressively decreases toward zero.
According to the kinematic model proposed in
this article, these two types of faults may have evolved
through time from one subtype to the other. This
model also permits us to investigate in detail the complex relationships that exist spatially and temporally
between the classically defined forcing factors con-
trolling a rift evolution, that is, (1) the direction of
regional extension, (2) the reactivation of basement
fabrics, and (3) the occurrence and cyclicity of volcanic activity. Such knowledge helps provide a better
definition for the general concept of rift propagation.
Direction of Regional Extension
Our interpretations of complementary seismic data
indicate that deep (2.5–4 s TWTT) half grabens are
preferentially developed against major or minor, N0–
20–trending normal border faults, suggesting a regional purely orthogonal extension (N90–110), in
agreement with Morley’s (1988) interpretation of
PROBE multichannel seismic data. Such a direction
of extension also may have induced the development
of half grabens along transverse normal faults with a
reduced subsidence rate (between 0.5 and 3 s TWTT
sedimentary thickness), less than in the case of major
submeridional longitudinal normal border faults. Positive flower structures, seen along transverse faults on
multichannel and single-channel seismic lines, indicate
strike-slip movements along these faults after about
1.1 Ma. This suggests a reorientation of the regional
extension from pure orthogonal extension to oblique
extension, parallel to the transverse faults. This rotation resulted in half-graben subsidence decreasing
along the N0–20–trending major border faults and basin deepening against transverse faults. Such results
are in agreement with Morley’s work (1995), in which
he defined, in the case of pure orthogonal extension,
a 100% subsidence rate along major longitudinal border faults and an 80% subsidence rate along transverse
faults. In contrast, in the case of an oblique extensional regime, subsidence rates change to 80% against
longitudinal faults, and only strike-slip motion is observed along transverse faults.
At the scale of the EARS (Figure 1A), fault kinematic studies, for example, near Lake Malawi, suggest that between about 2 and about 0.2 Ma, a reorientation of the regional extension direction from
east-west to northwest-southeast reactivated some
originally normal extensional faults with a strong
oblique-slip component (Ring et al., 1993; Ring and
Betzler, 1995). Strecker et al. (1990) illustrated such
a change for the central Kenyan rift at about 0.4 Ma,
and Delvaux et al. (1992) identified a similar event in
the Rukwa–north Malawi region between 0.55 and
0.42 Ma. Such coincident timing in the two branches
of the EARS suggests that changes in stress regime
could be the result of a major plate boundaries pro-
cess. In the north Tanganyika area, a combination of
fault movements identified at the recent Pemba and
Cape Banza hydrothermal sites (Coussement et al.,
1994) and focal mechanism solutions in the same area
(Fairhead and Girdler, 1971; Fairhead and Stuart,
1982; Shudofsky, 1985) indicates a N90–110 local extension direction. This may suggest a counterclockwise rotation of extension back to initial pure orthogonal extension for the most recent period.
Reactivation of Basement Fabrics
Preexisting fabrics clearly can control the geometry
and location of rifts at various levels by initiating, diverting, or inhibiting fracture propagation. Within the
EARS, the western branch is commonly cited as a result of the multiple reactivation of the N140-trending
Paleoproterozoic Ubende belt, described as a “longlived fundamental zone of structural weakness” (Sutton and Watson, 1986) or a “perennial taphrogenic
structure” (McConnell, 1972). In the case of the
northern end of the Tanganyika basin, longitudinal
major and minor border faults are N0–20–trending,
whereas transverse faults essentially correspond to the
N130–140 trend. These two rift tectonic trends are
clearly related to basement fabrics developed in central Africa during the successive Ubendian and Kibaran orogenies.
As a consequence of their location and importance, basement fabrics of Kibaran origin essentially
control the initiation, and in some cases the propagation, of major and minor longitudinal border faults.
In terms of local migration of extension and related
rift fault propagation, it previously has been demonstrated that the regional extensional component resulting in a proto-central Tanganyika rift developed at
about 12 Ma, then migrated asynchronously northward and southward (Cohen et al., 1993). In this article, we clearly demonstrate that northward rift propagation existed for the northern end of Lake
Tanganyika from 7.4 Ma up to the present day. This
confirms a similar mechanism developing from the
central basin toward the south end of the studied
area (Biera escarpment and WUF) between about 12
and slightly before 7.4 Ma. Oldest RSRM ages and
the distribution of major longitudinal border faults
suggest that northward rift propagation from the central basin preferentially followed a submeridional
western corridor delineated by the Biera escarpment
and the WUF (Figures 9; 10A, B), developed upon
the southwestern Kibaran belt fabrics. In contrast, the
Lezzar et al.
1049
N130–140–trending Ubende basement fabrics at various scales appear to exert a major influence on propagation of longitudinal major and minor border faults,
according to their particular geometry in terms of dip
and intersection angles when interacting with transverse basement fabrics (see the models developed at
the end of this article).
Influence of Volcanic Activity
Unlike the influence of prerift fabrics on rift development, volcanism largely has been ignored as a factor
in regulating rift mechanics in the Tanganyika region
(Morley, 1995). In the Kivu-Rusizi basin, about 100
km north of the Tanganyika basin, cyclic volcanic activity began at about 10 Ma and continues up to the
present (Ebinger, 1989a, b; Pasteels et al., 1989). In
this article, we propose that the presence of magmatic
heat in relation to these volcanic events resulted in
cyclic changes in the thermal state of the upper crust
of the north Tanganyika–Kivu-Rusizi region. Such
thermal variations induced increasing ductility and reduced brittle thickness of the upper crust, which became unfavorable for strong and deep brittle tectonics
to induce large-scale faults. Under similar extensional
regimes, rift fault propagation was considerably
slowed down in the north Lake Tanganyika basin, and
formed faults were much shallower and shorter than
what could have been formed with a colder, thicker
continental upper crust . When the crust returned to
an abnormal thermal state during periods of volcanic
quiescence, rift faults propagated normally. In consequence, cyclic volcanic activity clearly acted as a delay
factor on rift fault propagation. In the absence of such
a volcanic influence, the duration of rifting processes
in north Tanganyika would have been noticeably reduced, resulting in an earlier development of major
half grabens and depocenters. Caution should be
taken regarding the limits of our model, however, because Morgan et al. (1999) showed that heat flow and
thermal conductivity are quite slow processes in rocks
and cannot just be switched on and off. Morgan et al.
(1999) calculated that the thermal effects of the asthenospheric mantle at the base of the crust 10 m.y.
have yet to be manifest as surface heat-flow anomalies
on the rift flanks. Our study in a much younger rift
system than the Atlantic Rift could indicate that in
the East African rift, at least in the western branch
around the Kivu–north Tanganyika area, heat-flow
anomalies reached the surface several times in less
than 10 m.y.
1050
Neogene Sedimentary Depocenters (East African Rift)
An increase of magmatic underplating during the
successive eruptive phases also may have induced updoming of the area, resulting in the cyclic development of the Kivu-Rusizi local dome (Lezzar et al.,
1996; Cohen et al., 1997). Resulting slope variations
within the upper drainage basin of north Tanganyika
induced strong variations of sedimentary fluxes between the main, axial sediment source in the KivuRusizi area and the distal depocenters, by creating
sediment bypass zones and resulting in the preferential infilling of the most distal depocenters.
FAULT INTERACTION CLASSIFICATION:
INFLUENCE ON DEPOCENTER
DISTRIBUTION
The rift fault propagation model proposed for the
northern Lake Tanganyika basin clearly shows how
complex the relationships are between the three forcing factors, preexisting (basement) fabrics, kinematics
of rift in terms of variations of extension direction,
and influence of local and cyclic magmatic activity. In
this section, we examine various types of fault interactions, as well as their consequences for half-graben/
depocenter initiation, distribution, and infill. The proposed late Miocene–Holocene kinematic model demonstrates that depocenter initiation and development
are functions of (1) the intersection angle between a
major or minor border fault and transverse fault
trends and (2) the convergent or divergent dip between intersecting fault planes. In addition to the angle and dip parameters, the strike-slip component related to extensional strain is also a controlling factor
in the interaction areas.
We propose a classification of these different
types of fault intersections, and we discuss them in
terms of initiation, development, and inhibition of
major depocenters through time. This classification
scheme is based on the northern Lake Tanganyika rift
basin, although it is also likely to be applicable to
other basins, because many of the fault geometries observed in Lake Tanganyika are similar to those observed in other continental rift basins (e.g., Reynolds,
1984; Rosendahl, 1987). By unraveling the chronological development of these fault families and their
associated depocenters, we are able to draw lessons
about the evolutionary models of accommodation or
transfer zones. This leads to an important distinction
between the structural conditions that exist early in
rift basin development and the final geometry of the
a compressional wedge. Such a fault arrangement is
thus unfavorable to the formation of an initial depocenter and also acts to stop the northward fault propagation. This restrains border-fault propagation and,
hence, forces the maximum local extensional strain to
be laterally offset to a more favorable geometrical configuration (Figures 10A, B; 11). This eastward extension shift occurs in a zone that may be compared to
an overlapping convergent transfer zone, as described
by Morley et al. (1990) and Nelson et al. (1992), or
to an interference accommodation zone (Rosendahl et
al., 1988). We think this previous terminology is inappropriate for the initial stage of rifting, especially
where border faults have formed diachronously. For
this situation, the notions of “overlapping convergent”
and “interference” imply processes that were not necessarily operative; it is more accurate to use “extensional offset zone” for the initial phases of rifting. In
the case of more recent major synrift phases, we completely agree with use of the terms “transfer zones” or
rift basin. Thus, we caution against the use of transfer
zone terminology based solely on observed geometry,
because these labels commonly imply a process that
was not necessarily operative.
Fault Interaction Type 1: Lock of Fault Propagation
Inducing a Lateral Offset of the Extensional Strain
10 km
N
LM
F
UBFSn
Fault interaction type (FIT) 1 is characterized by the
following geometrical conditions (Figure 11): (1) the
angle between a longitudinal major or minor normal
border fault and a transverse normal or strike-slip fault
is acute; (2) dips between the longitudinal border and
transverse faults are divergent; and (3) the area of the
fault intersection is under extension due to the
dextral component of movement along the transverse
fault.
Using a model of orthogonal extension applied to
the northern Tanganyika basin, the dextral component of strike-slip in the area of the acute angle creates
shoreline
Fault propagation
BKF
FIT 3
N
FIT 3A
WUF
Border fault propagation
N1
40
Local Extension
ult
f fa
no
tio tion
rec ga
D i ropa
p
UBFSs
ult
f fa
no n
tio tio
rec ga
Di ropa
p
Extensional
offset zone
Regional
FIT 2
extension
K
A
F
?
25 Km
t
FIT 2
K
FZ
Local Extension
l
fau
of
ion ion
ect gat
Dirpropa
FIT 3B
N
FIT 1
N
N1
40
n
io ne
ns o
te r Z
Ex sfe
an
Tr
PM
F
FIT 1:
Fault Interaction
Type 1
FIT 1
Present-day
N
14
0
Local Extension
Lake Tanganyika
Figure 11. Fault interaction type (FIT) classification, indicating favorable or unfavorable geometry to initiate a major depocenter
and chronological evolution from initial up to final subsidence stages. FIT 1 ⳱ angle of interference is acute, dips of interacting faults
are opposed, and lateral component of displacement is locally under compression; consequences ⳱ cessation of fault propagation
against the transverse fault because of unfavorable geometry to initiate and develop a depocenter. FIT 2 ⳱ angle of interference is
obtuse, dips of interacting faults are opposed, and lateral component of movement is extensional; consequences ⳱ transverse fault
is used to transfer the extensional strain. FIT 3 ⳱ angle of interference is obtuse, dips of interacting faults are similar, and lateral
component of movement is extensional; consequences ⳱ subsidence of a triangular block and initiation of a major depocenter; FIT
3 is subdivided in two subtypes, FIT 3A and FIT 3B, based on faulting and basin initiation chronology (see Figure 12 for further
details). For other abbreviations, refer to Table 1.
Lezzar et al.
1051
“accommodation zones,” because they are the product
of several synrift phases.
Fault Interaction Type 2: Transverse Fault(s) Used to
Transfer Extensional Strain
Fault interaction type 2 is characterized by the following geometrical conditions: (1) the angle between the
longitudinal major or minor normal border and
oblique fault trends is obtuse; (2) the area of the fault
intersection is under extension due to the dextral
component of movement along the transverse faults
(Figure 11); and (3) dips between the border and
transverse faults are divergent. This configuration results in a transfer of extension along the transverse
faults. As for FIT 1, we think that the area between
two overlapping opposite polarity longitudinal normal
border faults (WUF and UBFSn) evolves through
time. In the earlier stages of rifting, the transverse
faults act principally as a typical extension transfer
fault zone, transferring extensional strain from one
border fault to another (Figure 10C, D). At a later
stage, continued subsidence adjacent to the border
faults results in well-developed half grabens of opposing polarity. At this time, the terms “convergent transfer zone,” “overlapping transfer zone,” or “accommodation zone,” already used in previous articles, are
more appropriate.
Fault Interaction Type 3: Major Depocenter Initiation and
Abandonment or Development
Fault interaction type 3 is characterized by the following geometrical conditions: (1) the angle between the
longitudinal major or minor normal border fault and
the transverse normal or strike-slip fault is obtuse;
(2) the normal fault and transverse/strike-slip faults
display favorable convergent dips; and (3) the area of
the obtuse angle is under extension due to the dextral
component of strike-slip motion. Fault interaction
type 3 represents the most favorable interaction of
faults for developing a major depocenter. Considering
the chronological development and the direction of
propagation of the faults (refer to the kinematic
model text section and Figure 10), however, two subtypes can be distinguished within FIT 3 (Figures 11,
12).
In FIT 3A, the longitudinal minor normal border
fault first accommodates the initial extension and
then, while propagating, interacts with the first transverse fault, inducing its reactivation and, subse1052
Neogene Sedimentary Depocenters (East African Rift)
quently, the formation of a triangular or spoon-shaped
subsiding block (Figure 12A). This results in the initiation of an outermost depocenter in the earliest
stages (Figures 10A, 12A). In the final stage, the transverse fault is abandoned subsequently by border fault
propagation. Increased subsidence produces the formation of an elongate hanging-wall basin, which continues to evolve independently of the early transverse
faults.
Subtype FIT 3B is geometrically similar to FIT
3A, but the kinematics of rifting and its consequences
in terms of depocenter final shape are completely different. Considering the development of longitudinal
and transverse faults (direction of fault propagation in
Figure 12B) for FIT 3B (following a FIT 2 case), two
transverse faults are reactivated before the development of the north-south longitudinal border fault.
This produces two restricted triangular-shaped depocenters delineated by two elongated basement tilted
blocks (lines P6, S7, and S8 in Figure 6; Figure 12B).
As the border fault continues to propagate northward,
local subsidence becomes greatest on the northernmost transverse fault, producing a major outermost
triangular (spoon-shaped) depocenter, deeper than
the elongated one induced by a FIT 3A. The innermost (southern) depocenter may subside at a slower
rate or be abandoned as a perched depocenter. At a
later stage, both basins are filled in with sediments.
The inner transverse fault is buried, and the southern
outer transverse fault becomes the transverse border
fault of a large depocenter (final stage in Figure 12B).
BASIN SUBSIDENCE VARIATION AND
IMPLICATIONS FOR STRUCTURE AND
INFILL
Figure 13 summarizes the major events inducing the
development and distribution of each depocenter of
the northern Lake Tanganyika rift basin through a
synthesis of the final-stage fault configuration of each
key zone in terms of related hypothetical subsidence.
Figure 13 shows clearly the migration through time of
each depocenter, related to fault abandonment, propagation, and interference. These diagrams show that
strong variations in sedimentary facies, related to fault
initiation (coarse bodies) or to development of an
elongated quiet depocenter (laminated lacustrine facies), may occur in such basins both within the vertical sedimentary pile and along strike following the
propagation of major faults.
A.
KAF
KFZ
40
UF
W
KFZ
N1
ult
f fa n
n o tio
tio aga
r e c p r op
UF
Di
W
N
KAF
ult
f fa
n o on
tio ati
rec pag
D i pro
FIT 3A
N
Triangular
outermost r
depocente
Early stage
Final stage
Initiation of the first depocenter by reactivation
Abandonment of the KFZ and propagation
of the KFZ in conjunction with the formation of the WUF border fault that controls the major
of the WUF border fault.
depocenter, without any reactivation of the KAF.
Figure 12. Subdivision of FIT
3 into to two subtypes, based
on faulting and basin initiation
chronology. (A) FIT 3A ⳱ major depocenter initiation and
abandonment; consequence ⳱
border fault propagation and
elongated depocenter. (B) FIT
3B ⳱ major depocenter initiation and development; consequence ⳱ triangular or spoonshaped deep depocenter. For
abbreviations, refer to Table 1.
B.
FIT 3B
40
PMF
Local Extension N
MF
Early stage
E
ost
Outerm
major
nter
ce
depo
PMF
Initiation of two depocenters using
the reactivation of the two parallel
N
transverse faults (LMF and PMF)
associated with the normal fault UBFSn.
The southernmost triangular block is perched,
N 14
0
LMF
Cha nn
e ls
system
Intermediate stage
whereas the external one becomes the major
depocenter of the area controlled by the LMF.
ult
f fa
n o on
tio ati
n
rec ag
FS
D i pro p
UB
N
U
Perched
basin L
ult
f fa
no n
tio atio
rec ag
Sn
D i pro p
BF
F
UB
LMF
N1
ult
f fa
no n
tio atio
rec ag
D i prop
Sn
PMF
ppin
Overla
basin
g
N
Final stage
25 Km
Abandonment of the LMF and southward extension
of an elongated basin, controlled by the PMF and UBFSn.
FAULT CORRELATION CLASSIFICATION:
LESSONS FOR SEISMIC INTERPRETERS
EXPLORING IN EXTENSIONAL SETTINGS
The degree of correlative to noncorrelative rift faults
can be classified into specific fault correlation types
(FCTs) based on the kinematic model and fault interaction classification proposed in this article and shown
in Figures 10, 11, and 12. These correlations have been
investigated at the scale of the entire rift basin by integrating satellite images, microstructural onshore observations, and offshore multichannel and singlechannel seismic reflection data (Figures 7, 14).
Fault Correlation Type 1: Noncorrelative Onshore-Onshore
Faults
Fault correlation type 1 is illustrated by the eastdipping UBFSn and BKF, which previously have been
interpreted as a single continuous fault forming the
western escarpment of the north Tanganyika basin sets
(Figures 7, 14). Our kinematic reconstruction indicates
that this escarpment is broken into two distinct fault
segments (UBFS and BKF), separated by the transverse
KFZ/KAF system. In addition, the UBFSn is interpreted as a major N0-striking normal border fault controlling the NRHG since 3.5 Ma, whereas the BKF is
interpreted as a flexural response to the major West
Ubwari border fault since 1.1 Ma.
Fault Correlation Type 2: Poorly Correlated OffshoreOnshore Faults
Fault correlation type 2 is illustrated by the southwestdipping offshore transverse faults of the KFZ/KAF.
These faults behaved as normal faults for only a brief
time during the early history of the rift sets (Figures 7,
14); they have been transverse strike-slip faults during
Lezzar et al.
1053
PMF
FIT 3B
U
L MF
N 14
0
Local Extension
12 11 10
Approximate depth
on
a ti
ig r n
n m atio
sio ag
ten rop
Ex ult p
Sn
Fa
BF
BA
C
9
7
8
6
5
4
3
2
1
Initiation of two triangular
depocenters controlled by
the PMF and LMF associated with
the UBFSn
Age (Ma)
C
Development of the outermost
major depocenter controlled
by the LMF transverse fault
B
The PMF-controlled initial
triangular depocenter acts
as a perched basin
A
25 km
N
C
12 11 10
B
N
A
N1
40
PM
F
SRHG
on
a ti
igr n
n m atio
sio ag
ten rop
Ex ult p
Fa
C
B
A
KFZ
UF
W
N
SRHG
UBFS s
B
C A
on
a ti
igr n
n m atio
sio ag
ten rop
Ex ult p
Fa
FIT 1
Approximate depth
K AF
UF
W
FIT 3A
N
K FZ
N1
40
7
8
5
6
3
4
1
2
Northward transfer of the
extension along the PMF
transverse fault and initiation of
the NRHG northern half graben
9
7
8
6
5
4
3
2
1
Abandonment of the initial
depocenter and northward
migration of the WUF
controlling the SRHG elongated
basin
9
8
7
6
5
Age (Ma)
C
B
A
Initial triangular
depocenter controlled by
the WUF and KFZ
12 11 10
Age (Ma)
B
C
A
Northward migration of
rifting (fault propagation) and progression of
the SRHG elongated depocenter
12 11 10
on
a ti
igr n
n m atio
sio ag
ten rop
Ex ult p
Fa
Approximate depth
FIT 2
Approximate depth
NRHG
9
4
3
2
Reactivation of the
N140 inherited fault
KFZ
Northward migration of
rifting and formation of
the western border fault
UBFSs
1
Age (Ma)
C
B
A
Initiation of the first depocenter to
the east and abandonment of the
UBFSs and the KFZ
Figure 13. Basin geometries and hypothetical subsidence curves showing synthesis of the final-stage fault configuration of each
key zone of the northern Lake Tanganyika rift basin. See definitions of fault interaction types in Figures 11 and 12; for other
abbreviations, refer to Table 1.
1054
Neogene Sedimentary Depocenters (East African Rift)
shore on the eastern rift shoulder from 1.1 Ma as a
flexural response fault to the major normal UBFSn that
controls the subsidence of the NRHG.
Fault Correlation Type 4: Well-Correlated OnshoreOnshore Faults
Fault correlation type 4 is illustrated by the BBFSn
and BBFSs sets (Figures 7, 14) that, during the 1980s
and the 1990s, were interpreted by Chorowich and
Mukonk (1980) to form the single, continuous eastern
Burundian escarpment of the northern Lake Tanganyika rift basin. From our data, however, this escarpment appears to correspond to two distinct fault segments, the BBFSn and BBFSs, both active from around
1.1 Ma as flexural response faults to the deepening of
the NRHG and NKHG, respectively.
Figure 14. Fault correlation type (FCT) classification and lessons for seismic interpreters exploring in extensional settings.
FCT 1 ⳱ noncorrelative onshore-onshore faults; FCT 2 ⳱
poorly correlated offshore-onshore faults; FCT 3 ⳱ noncorrelative offshore-onshore faults; FCT 4 ⳱ well-correlated onshoreonshore faults; FCT 5 ⳱ well-correlated offshore-onshore faults.
For other abbreviations, refer to Table 1.
most of their history. Although geometrically similar
to the northernmost transverse faults (LMF and PMF),
the KFZ and KAF have a weaker onshore expression
across the western rift shoulder. Although easy correlation can be made between the offshore and onshore
segments of the KFZ and KAF, this relationship is not
as obvious as for the northernmost transverse fault pair
(LMF and PMF) and probably is due to the strike-slip
character of these faults.
Fault Correlation Type 3: Noncorrelative Offshore-Onshore
Faults
Fault correlation type 3 is illustrated by the offshore
N0-striking and west-dipping WUF and onshore
BBFSn, both of which have been interpreted in previous articles (Mondeguer et al., 1986; Coussement,
1995) as being a single, continuous fault set (Figures 7,
14). Our interpretation indicates two distinct, asynchronous, N0-striking fault segments, with the WUF
acting as a major normal border fault controlling the
SRHG from 7.4 to 0.4 Ma. The BBFSn is active on-
Fault Correlation Type 5: Well-Correlated OffshoreOnshore Faults
Fault correlation type 5 is illustrated by the LMF/PMF
northwest-southeast–trending fault sets (Figures 7,
14). The northeast-dipping LMF/PMF offshore faults
have been described by Lezzar et al. (1996) as active
since the late Miocene (pure normal faulting from 7.4
up to 1.1 Ma, controlling subsidence of the NRHG).
Deformation on this system changed recently, with an
active strike-slip component from 1.1 Ma up to the
present day. These two transverse faults also have a
very clear onshore extension throughout the western
rift escarpment (Rolet et al., 1991; Coussement et al.,
1994).
CONCLUSIONS AND IMPLICATIONS FOR
RIFT BASIN EXPLORATION
The kinematic model established for the northern Lake
Tanganyika basin from the late Miocene to the Holocene provides a chronological framework suitable for
examining the interaction between longitudinal normal border faults and transverse normal or strike-slip
faults. The latter interact to create a suite of depocenters, each with unique subsidence histories and sedimentary fill. Recognition of these fault interactions reinforced by a chronological constraint furthers our
understanding of rift architecture, specifically how
transfer zones evolve. We believe that the descriptive
terminology currently used for existing geometries
does not adequately address the various and commonly
Lezzar et al.
1055
diachronous processes responsible for finite geometrical configuration.
The effect of cyclic volcanic activity (uplifted local
volcanic dome) close to the north Tanganyika rift basin
and its associated thermal effects probably induced a
selective effect on extension and fault propagation. A
depocenter proximal to this domed-up area is much
more disturbed by volcanic cycles and subsequent rift
fault propagation. The occurrence of such fluctuating
topographic uplifts leads to important sedimentary
fluxes all around the dome, which rapidly fill in the
depocenters. This model demonstrates that the interpretation of thick basin infills in deep, actively subsiding half grabens is not at all indicative of the age of
such systems. This is an important consideration when
exploring in extensional basins.
Maps provided in Figure 5 are essential for seismic
stratigraphers who interpret and correlate faults in extensional settings. These coarse detrital deep lacustrine
fans, developed at the front of major river systems,
probably have a high reservoir potential. Deep-thicklarge sublacustrine fans (DTL type) deposited in half
grabens (35–50 km long, 20–30 km wide, 1.5–3.5 km
deep) show dimensions of 30 ms TWTT (20 m) thickness, 10–15 km length, and 5–10 km width. Shallowthin-elongated sublacustrine fans (STE type) deposited
on a horst/high-relief accommodation zone (40 km
long, 15 km wide, 750 m deep) show dimensions of
15 ms TWTT (9 m) thickness, 10–15 km length, and
2.5–4 km width.
The example of the Lake Tanganyika basin may
also help in the interpretation and understanding of
petroleum-rich rift systems. Note, however, that very
specific characteristics, such as extension direction,
heritage (basement fabrics), and thermal state of the
upper crust (cyclic volcanic activity), can be absent or
different from other rifts. These forcing characteristics
during rifting propagation are fundamental factors in
determining the rift location, the geometry of the initial basins, and their later evolution.
Interaction of active north-south or northwestsoutheast normal fault segments has greater impact on
basin rift geometry because they control the thickness
and size of those basins. Establishing realistic and coherent fault correlation during seismic data interpretation in complex tectonic settings like extensional basins is important. For example, we show why two
geometrically similar axial and transverse faults acted
differently during the rifting phase. These faults may
be major factors in rift basin formation, controlling
basin thickness, length, and width. The proposed clas1056
Neogene Sedimentary Depocenters (East African Rift)
sification of rift fault interaction based on faulting geometry, mechanism, and chronology can help petroleum geologists/geophysicists avoid potential errors
when mapping faults. We presented a tool that can aid
exploration seismic interpreters in assessing basin age
in frontier areas that have no age or well control.
This article also shows that a relationship exists
between fault intersection angle and dip and that this
governs the initiation and development of depocenters
in the northern Lake Tanganyika basin and probably in
other continental rift systems. Recognition of these
fault geometries can be used to help predict early synrift facies distribution, basin subsidence histories, and
problems related to fault growth and scaling.
REFERENCES CITED
Anders, M. H., and R. W. Schlische, 1994, Overlapping faults, intrabasin highs, and the growth of normal faults: Journal of Geology, v. 102, p. 165–179.
Baker, B. H., 1986, Tectonics and volcanism of the southern Kenya
rift valley and its influence on rift sedimentation, in L. E. Frostick, R. W. Renaut, I. Reid, and J. J. Tiercelin, eds., Sedimentation in the African rifts: Geological Society Special Publication 25, p. 45–57.
Bassi, G., 1991, Factors controlling the style of continental rifting:
insights from numerical modeling: Earth and Planetary Science
Letters, v. 105, p. 430–452.
Bassi, G., C. Keen, and P. Potter, 1993, Contrasting styles of rifting:
models and examples from the Eastern Canadian margin: Tectonics, v. 12, no. 3, p. 639–655.
Bellon, H., and A. Pouclet, 1980, Datations K-Ar de quelques laves
du Rift-ouest de l’Afrique Centrale: implications sur l’evolution
magmatique et structurale: Geologische Rundschau, v. 69,
p. 49–62.
Boccaletti, M., T. Mammo, M. Bonini, and B. Abebe, 1994, Seismotectonics of East African rift system: evidence of active
oblique rifting: Annales Tectonicae, v. 2, p. 87–99.
Bouroullec, J. L., C. Thouin, J. J. Tiercelin, J. Rolet, J. P. Rehault,
and A. Mondeguer, 1992, Sequences sismiques haute resolution
du fosse nord-Tanganyika, Rift Est-africain: implications climatiques, tectoniques et hydrothermales: Comptes Rendus de
l’Academie des Sciences, v. 315, p. 601–608.
Buck, W. R., 1991, Modes of continental lithospheric extension:
Journal of Geophysical Research, v. 96, p. 20161–20178.
Burgess, C. F., B. R. Rosendahl, S. Sander, C. A. Burgess, J. Lambiase, S. Derksen, and N. Meader, 1988, The structural and
stratigraphic evolution of Lake Tanganyika: a case study of continental rifting, in W. Manspeizer, ed., Rifting and the opening
of the Atlantic Ocean: Amsterdam, Elsevier, p. 861–881.
Carte Geologique du Burundi, 1986, Rumonge sheet: scale
1:100.000, 1 sheet.
Carte Geologique du Burundi, 1988, Makamba sheet: scale
1:100.000, 1 sheet.
Carte Geologique du Burundi, 1989, Bujumbura sheet: scale
1:100.000, 1 sheet.
Cartwright, J. A., B. D. Trudgill, and C. S. Mansfield, 1995, Fault
growth by segment linkage: an explanation for scatter in maximum displacement and trace length data from the Canyon-
lands grabens of SE Utah: Journal of Structural Geology, v. 17,
p. 1319–1326.
Cartwright, J. A., C. S. Mansfield, and B. D. Trudgill, 1996, The
growth of normal faults by segment linkage, in P. G. Buchanan
and D. A. Nieuwland, eds., Modern developments in structural
interpretation, validation and modeling: Geological Society
Special Publication 99, pp. 163–177.
Childs, C., J. Watterson, and J. J. Walsh, 1995, Fault overlap zones
within developing normal fault systems: Journal of the Geological Society, v. 152, p. 535–549.
Chorowicz, J., and N. B. Mukonk, 1980, Lineaments anciens, zones
transformantes recentes et geotectonique des fosses de l’Est Africain, d’apses la teledetection et la microtectonique: Museum
Royal de l’ Afrique centrale, Département de Géologie et Minéralogie, Tervuren, rapp. ann., p. 143–167.
Chorowicz, J., J. Le Fournier, C. Le Mut, J. P. Richert, F. L. SpyAnderson, and J. J. Tiercelin, 1983, Observation par teledetection et au sol de mouvements dÈcrochants NW-SE dextres dans
le secteur transformant Tanganyika-Rukwa-Malawi du rift estafricain: Comptes Rendus de l’Academie des Sciences, v. 296,
p. 997–1002.
Chorowicz, J., J. Le Fournier, and G. Vidal, 1987, A model for rift
development in eastern Africa: Geological Journal, Thematic
Issue 22, p. 495–513.
Chorowicz, J., T. Nkanira, and G. Tamain, 1988, L’accident nordsud du Burundi: une faille inverse kibarienne visible par satellite: Son role dans la formation du fosse cenozoique nordTanganyika: Comptes Rendus de l’Academie des Sciences,
v. 307, p. 1663–1668.
Coblentz, D. D., and M. Sandiford, 1994, Tectonic stresses in the
African plate: constraints on the ambient lithospheric stress
state: Geology, v. 22, p. 831–834.
Cohen, A. S., M. J. Soreghan, and C. A. Scholz, 1993, Estimating
the age of formation of lakes: an example from Lake Tanganyika, East African rift system: Geology, v. 21, p. 511–514.
Cohen, A. S., K. E. Lezzar, J. J. Tiercelin, and M. Soreghan, 1997,
New palaeogeographic and lake level reconstructions of northern Lake Tanganyika: implications for tectonic, climatic and
biologic evolution in a rift lake: Basin Research, v. 9, p. 107–
132.
Contreas, J., C. H. Scholz, and G. C. P. King, 1997, A model of rift
basin evolution constrained by first-order stratigraphic observations: Journal of Geophysical Research, v. 102, p. 7673–
7690.
Contreas, J., M. H. Anders, and C. H. Scholz, 2000, Growth of a
normal fault system: observations from the Lake Malawi basin
of the East African rift: Journal of Structural Geology, v. 22,
p. 159–168.
Coussement, C., 1995, Structures transverses et extension intracontinentale: Le role des zones de failles d’Assoua et TanganyikaRukwa-Malawi dans la cinematique neogene du systeme de Rift
Est-Africain: These de Doctorat Nouveau Regime (Ph.D. dissertation), Universite de Bretagne Occidentale, Brest, France,
222 p.
Coussement, C., P. Gente, J. Rolet, J. J. Tiercelin, M. Wafula, and
S. Buku, 1994, The North Tanganyika hydrothermal fields, East
African rift system: their tectonic control and relationship to
volcanism and rift segmentation: Tectonophysics, v. 237,
p. 155–173.
Cowie, P. A., and C. H. Scholz, 1992, Displacement-length scaling
relationships for faults: data synthesis and discussion: Journal of
Structural Geology, v. 14, p. 1149–1156.
Daly, M. C., 1988, Crustal shear zones in Central Africa: a kinematic
approach to Proterozoic tectonics: Episodes, v. 11, p. 5–11.
Daly, M. C., J. Chorowicz, and J. D. Fairhead, 1989, Rift basin evolution in Africa: the influence of reactivated steep basement
shear zones, in M. A. Cooper and G. D. Williams, eds., Inversion tectonics: Geological Society Special Publication 44,
p. 309–334.
Dawers, N. H., and M. H. Anders, 1995, Displacement-length scaling and fault linkage: Journal of Structural Geology, v. 17,
p. 607–614.
Dawers, N. H., and J. R. Underhill, 2000, The role of fault interaction and linkage in controlling syn-rift stratigraphic sequences:
Statfjord East area, northern North Sea: AAPG Bulletin, v. 84,
p. 45–64.
De Bremaecker, J. Cl., 1959, Seismicity of the west African rift valley: Journal of Geophysical Research, v. 64, p. 1961–1966.
Delvaux, D., K. Levi, R. Kajara, and J. Sarota, 1992, Cenozoic
paleostress and kinematic evolution of the Rukwa–North Malawi rift valley (East African rift system): Bulletin Centre Recherche Exploration Production Elf-Aquitaine, v. 16, p. 383–
406.
Ebinger, C. J., 1989a, Tectonic development of the western branch
of the East African rift system: Geological Society of America
Bulletin, v. 101, p. 952–967.
Ebinger, C. J., 1989b, Geometric and kinematic development of border faults and accommodation zones, Kivu-Rusizi rift, Africa:
Tectonics, v. 8, p. 117–133.
Ebinger, C. J., B. R. Rosendahl, and D. J. Reynolds, 1987, Tectonic
model of the Malawi rift, Africa: Tectonophysics, v. 141,
p. 215–235.
Ebinger, C. J., T. D. Bechtel, D. H. Forsyth, and C. O. Bowin, 1989a,
Effective elastic plate thickness beneath the East African and
Afar plateaux and dynamic compensation of the uplifts: Journal
of Geophysical Research, v. 94, p. 2893–2901.
Ebinger, C. J., A. L. Deino, R. E. Drake, and A. L. Tesha, 1989b,
Chronology of volcanism and rift basin propagation: Rungwe
volcanic province, East Africa: Journal of Geophysical Research, v. 94, p. 15785–15803.
Ebinger, C. J., G. D. Karner, and J. K. Weissel, 1991, Mechanical
strength of extended continental lithosphere: constraints from
the western rift system, East Africa: Tectonics, v. 10, p. 1239–
1256.
Fairhead, J. D., and R. W. Girdler, 1971, The seismicity of the East
African rift system: Geophysical Journal of the Royal Astronomical Society, v. 24, p. 271–301.
Fairhead, J. D., and G. W. Stuart, 1982, The seismicity of the East
African rift system and comparison with other continental rifts,
in G. Palmason, ed., Continental and oceanic rifts: American
Geophysical Union Geodynamics Series 8, p. 41–61.
Fernandez-Alonso, M., and K. Theunissen, 1998, Airborne geophysics and geochemistry provide new insights in the intracontinental evolution of the Mesoproterozoic Kibaran belt (Central Africa): Geological Magazine, v. 135, p. 203–216.
Gawthorpe, R. L., and J. M. Hurst, 1993, Transfer zones in extensional basins: their structural style and influence on drainage
development and stratigraphy: Journal of the Geological Society, v. 150, p. 1137–1152.
Gawthorpe, R. L., A. J. Fraser, and R. E. Ll. Collier, 1994, Sequence
stratigraphy in active extensional basins: implications for the
interpretation of ancient basin fills: Marine and Petroleum Geology, v. 11, p. 642–658.
Gawthorpe, R. L., I. R. Sharp, J. R. Underhill, and S. Gupta, 1997,
Linked sequence stratigraphic and structural evolution of propagating normal faults: Geology, v. 25, p. 795–798.
Gupta, S., P. A. Cowie, N. H. Dawers, and J. R. Underhill, 1998, A
mechanism to explain rift basin subsidence and stratigraphic
patterns through fault array evolution: Geology, v. 26, p. 595–
598.
Gupta, S., J. R. Underhill, I. R. Sharp, and L. Gawthorpe, 1999,
Role of fault interactions in controlling synrift sediment
Lezzar et al.
1057
dispersal patterns: Miocene, Abu Alaqa Group, Suez rift, Sinai,
Egypt: Basin Research, v. 11, p. 167–189.
Hendrie, D. B., N. J. Kusznir, C. K. Morley, and C. J. Ebinger, 1994,
Cenozoic extension in northern Kenya: a quantitative model of
rift basin development in the Turkana region: Tectonophysics,
v. 236, p. 409–438.
Karner, G. D., and N. W. Driscoll, 1993, Rift flank topography and
extensional basin architecture: formation of Broken Ridge,
southeast Indian Ocean: Anais da Academia Brasileira de Ciencias, v. 65, p. 263–294.
Karson, J. A., and P. C. Curtis, 1989, Tectonic and magmatic processes in the eastern branch of the East African rift and implications for magmatically active continental rifts: Journal of African Earth Sciences, v. 8, p. 431–453.
King, G. C. P., R. S. Stein, and J. B. Rundle, 1988, The growth of
geological structures by repeated earthquakes: 1. conceptual
framework: Journal of Geophysical Research, v. 93, p. 13307–
13318.
Klerkx, J., K. Theunissen, and D. Delvaux, 1998, Persistent fault
controlled basin formation since the Proterozoic along the western branch of the East African rift: Journal of African Earth
Sciences, v. 26, p. 347–361.
Kuznir, N. J., and S. S. Egan, 1989, Simple-shear and pure-shear
models of extensional sedimentary basin formation: application
to the Jeanne D’Arc basin, Grand Banks of Newfoundland, in
A. J. Tankard and H. R. Balkwill, eds., Extensional tectonics
and stratigraphy of the North Atlantic margins: AAPG Memoir
46, p. 305–322.
Leeder, M. R., and R. L. Gawthorpe, 1987, Sedimentary models for
extensional tilt-block/half-graben basins, in M. P. Coward, J. F.
Dewey, and P. L. Hancock, eds., Continental extensional tectonics: Geological Society Special Publication 28, p. 139–
152.
Le Turdu, C., 1998, Modeles tectono-sedimentaires 3D des bassins
en extension. Exemples du Rift du Kenya (Baringo-Bogoria et
Magadi), du Rift Ethiopien (Ziway-Shala) et du Fosse NordTanganyika: These de Doctorat Nouveau Regime (Ph.D. dissertation), Universite de Bretagne Occidentale, Brest, France,
416 p.
Lezzar, K. E., 1997, Evolution du Bassin Nord du Lac Tanganyika,
Rift Est-Africain, depuis le Miocene Superieur. Analyse de la
cinematique du rifting et des processus sedimentaires a partir
de donnees de sismique reflexion, imagerie satellitaire et carottages: These de Doctorat Nouveau Regime (Ph.D. dissertation), Universite de Bretagne Occidentale, Brest, France, 300 p.
Lezzar, K. E., J. J. Tiercelin, M. De Batist, A. S. Cohen, T. Bandora,
P. Van Rensbergen, C. Le Turdu, M. Wafula, and J. Klerkx,
1996, New seismic stratigraphy and late Tertiary history of the
North Tanganyika basin, East African rift system, deduced from
multifold reflection and high resolution seismic data and piston
core evidence: Basin Research, v. 8, p. 1–28.
Mack, G. H., and W. R. Seager, 1995, Transfer zones in the southern
Rio Grande rift: Journal of the Geological Society, v. 152,
p. 551–560.
McConnell, R. B., 1972, Geological development of the rift system
of Eastern Africa: Geological Society of America Bulletin, v. 83,
p. 2549–2572.
McLeod, A. E., N. H. Dawers, and J. R. Underhill, 2000, The propagation and linkage of normal faults: insights from the Strathspey–Brent–Statfjord fault array, northern North Sea, in S.
Gupta and P. A. Cowie, eds., Thematic set on processes and
controls in the stratigraphic development of extensional basins:
Basin Research, v. 12, p. 263–284.
Milani, E. J., and I. Davidson, 1988, Basement control and transfer
tectonics in the Reconcavo-Tucano-Jatoba rift, northeast Brazil:
Tectonophysics, v. 154, p. 41–70.
1058
Neogene Sedimentary Depocenters (East African Rift)
Mondeguer, A., J. J. Tiercelin, M. Hoffert, P. Larque, J. Le Fournier,
and P. Tucholka, 1986, Sedimentation actuelle et recente dans
un petit bassin en contexte extensif et decrochant: la baie de
Burton, fosse nord-Tanganyika, Rift Est-Africain: Bulletin des
Centres de Recherches Exploration-Production Elf Aquitaine,
v. 10, p. 229–247.
Morgan, W. J., M. A. Crooks, and M. J. Phipps, 1999, A thin spherical shell model of global asthenosphere flow: European Union
of Geosciences Conference Abstracts, EUG 10, Journal of Conference Abstracts, v. 4, no. 1, p. 347.
Morley, C. K., 1988, Variable extension in Lake Tanganyika: Tectonics, v. 7, p. 785–801.
Morley, C. K., 1989, Extension, detachments, and sedimentation in
continental rifts (with particular reference to East Africa): Tectonics, v. 8, p. 1175–1192.
Morley, C. K., 1994, Interaction of deep and shallow processes in
the evolution of the Kenya rift: Tectonophysics, v. 236, p. 81–
91.
Morley, C. K., 1995, Developments in the structural geology of rifts
over the last decade and their impact on hydrocarbon exploration, in J. J. Lambiase, ed., Hydrocarbon habitat in rift basins:
Geological Society Special Publication 80, p. 1–32.
Morley, C. K., 1999a, How successful are analogue models in addressing the influence of pre-existing fabrics on rift structure?:
Journal of Structural Geology, v. 21, p. 1267–1274.
Morley, C. K., 1999b, Patterns of displacement along large normal
faults: implications for basin evolution and fault propagation,
based on examples from East Africa: AAPG Bulletin, v. 83,
p. 613–634.
Morley, C. K., ed., 1999c, Geoscience of rift systems—evolution of
East Africa: AAPG Studies in Geology 44, 242 p.
Morley, C. K., 2002, Evolution of large normal faults: evidence from
seismic reflection data: AAPG Bulletin, v. 86, no. 6, p. 961–
978.
Morley, C. K., W. A. Wescott, S. M. Cunningham, and R. M.
Harper, 1989, Recent exploration in the Lake Rukwa area, East
African rift: 28th International Geological Congress, v. 2,
p. 462–463.
Morley, C. K., R. A. Nelson, T. L. Patton, and S. G. Munn, 1990,
Transfer zones in the East African rift system and their relevance to hydrocarbon exploration in rifts: AAPG Bulletin,
v. 74, p. 1234–1253.
Morley, C. K., W. A. Wescott, D. M. Stone, R. M. Harper, S. T.
Wigger, and F. M. Karanja, 1992, Tectonic evolution of the
northern Kenyan rift: Journal of the Geological Society, v. 149,
p. 333–348.
Nelson, R. A., T. L. Patton, and C. K. Morley, 1992, Rift-segment
interaction and its relation to hydrocarbon exploration in continental rift systems: AAPG Bulletin, v. 76, p. 1153–1169.
Pasteels, P., M. Villeneuve, P. De Paepe, and J. Klerkx, 1989, Timing
of the volcanism of the southern Kivu province: implications
for the evolution of the western branch of the East African rift
system: Earth and Planetary Science Letters, v. 94, p. 353–363.
Patterson, M. B., 1983, Structure and acoustic stratigraphy of the
Lake Tanganyika rift valley: a single-channel seismic survey of
the lake, north of Kalemie, Zaire: Master’s dissertation, Duke
University, Durham, North Carolina, 89 p.
Peacock, D. C. P., and D. J. Sanderson, 1991, Displacements, segment linkage and relay ramps in normal fault zones: Journal of
Structural Geology, v. 13, p. 721–733.
Reynes, P., J. Rolet, J. P. Richert, P. Gruneisen, J. M. Palengat, and
D. Coquelet, 1993, Apports des techniques 3D de la teledetection dans la recherche des blocs bascules du fosse nordTanganyika, Rift Est-Africain, Zaire: Bulletin des Centres de
Recherches Exploration-Production Elf Aquitaine, v. 17, p. 1–
17.
Reynolds, D. J., 1984, Structural and dimensional repetition in con-
tinental rifting: Master’s dissertation, Duke University, Durham, North Carolina, 175 p.
Ring, U., and C. Betzler, 1995, Geology of the Malawi rift: kinematic
and tectonosedimentary background of the Chiwondo Beds,
northern Malawi: Journal of Human Evolution, v. 28, p. 7–21.
Ring, U., C. Betzler, and D. Delvaux, 1993, Normal vs. strike-slip
faulting during rift development in East Africa: the Malawi rift:
Geology, v. 20, p. 1015–1018.
Rohrman, M., and P. van der Beek, 1996, Cenozoic postrift domal
uplift of North Atlantic margins: an asthenospheric diapirism
model: Geology, v. 24, p. 901–904.
Rolet, J., A. Mondeguer, J. L. Bouroullec, T. Bandora, C. Coussement, J. P. Rehault, and J. J. Tiercelin, 1991, Structure and
different kinematic development faults along the Lake Tanganyika rift valley (East African rift system): Bulletin des Centres
de Recherches Exploration-Production Elf Aquitaine, v. 15,
p. 327–342.
Rosendahl, B. R., 1987, Architecture of continental rifts with special
reference to East Africa: Annual Review of Earth and Planetary
Sciences, v. 15, p. 445–503.
Rosendahl, B. R., D. J. Reynolds, P. M. Lorber, C. F. Burgess, J.
McGill, D. Scott, J. J. Lambiase, and S. J. Derksen, 1986, Structural expressions of rifting: lessons from Lake Tanganyika, Africa, in L. E. Frostick, R. W. Renaut, I. Reid, and J. J. Tiercelin,
eds., Sedimentation in the African rifts: Geological Society Special Publication 25, p. 127–139.
Rosendahl, B. R., J. W. Versfelt, C. A. Scholz, J. E. Buck, and L. D.
Woods, 1988, Seismic atlas of Lake Tanganyika, East Africa:
Durham, North Carolina, Duke University, Project PROBE
Geophysical Atlas Series, unpaginated.
Saggerson, E. P., and B. H. Baker, 1965, Post-Jurassic erosion surfaces in eastern Kenya and their deformation in relation to rift
structure: Quarterly Journal of the Geological Society, v. 121,
p. 51–72.
Schlische, R. W., 1992, Structural and stratigraphic development of
the Newark extensional basin, eastern North America: implications for the growth of the basin and its bounding structures:
Geological Society of America Bulletin, v. 104, p. 1246–1263.
Schlische, R. W., 1995, Geometry and origin of fault-related folds
in extensional settings: AAPG Bulletin, v. 79, p. 1661–1678.
Schlische, R. W., S. S. Young, R. V. Ackerman, and A. Gupta, 1996,
Geometry and scaling relations of a population of very small
rift-related normal faults: Geology, v. 24, p. 683–686.
Scholz, C. A., B. R. Rosendahl, J. W. Versfelt, K. J. Kaczmarick, and
L. D. Woods, 1989, Seismic atlas of Lake Malawi (Nyasa), East
Africa: Durham, North Carolina, Duke University, Project
PROBE Geophysical Atlas series, unpaginated.
Scott, D. L., M. A. Etheridge, and B. R. Rosendahl, 1992, Obliqueslip deformation in extensional terrains: a case study of the lakes
Tanganyika and Malawi rift zones: Tectonics, v. 11, p. 998–
1009.
Sharp, I. R., R. L. Gawthorpe, J. R. Underhill, and S. Gupta, 1999,
Fault propagation folding in extensional settings: examples of
structural style and synrift sedimentary response from the Suez
rift, Sinai, Egypt: Geological Society of America Bulletin,
v. 112, p. 1877–1899.
Shudofsky, G. N., 1985, Source mechanisms and focal depths of East
African earthquakes using Rayleigh-wave inversion and bodywave modelling: Geophysical Journal of the Royal Astronomical Society, v. 83, p. 563–614.
Simiyu, S. M., and G. R. Keller, 1997, An integrated analysis of
lithospheric structure across the East African plateau based on
gravity anomalies and recent seismic studies: Tectonophysics,
v. 279, p. 291–313.
Steckler, M. S., F. Berthelot, N. Lyberis, and X. Lepichon, 1988,
Subsidence in the Gulf of Suez: implications for rifting and
plate kinematics: Tectonophysics, v. 153, p. 249–270.
Stock, J. M., and K. V. Hodges, 1990, Miocene to Recent structural
development of an extensional accommodation zone, northeastern Baja California, Mexico: Journal of Structural Geology,
v. 12, p. 315–328.
Strecker, M. R., P. M. Blisniuk, and G. H. Eisbacher, 1990, Rotation
of extension direction in the Central Kenya rift: Geology, v. 18,
p. 299–302.
Sutton, J., and J. V. Watson, 1986, Architecture of the continental
lithosphere, in H. G. Reading, J. Watterson, and S. H. White,
eds., Major crustal lineaments and their influence on the geological history of the continental lithosphere: Philosophical
Transactions of the Royal Society of London, Series A: Mathematical and Physical Sciences, v. 317, no. 1539, p. 5–12.
Sykes, L. R., 1978, Intraplate seismicity, reactivation of preexisting
zones of weakness, alkaline magmatism, and other tectonism
postdating continental fragmentation: Review of Geophysics
and Space Physics, v. 16, no. 1, p. 55–81.
Tanganydro Group, 1992, Sublacustrine hydrothermal seeps in
northern Lake Tanganyika, East African rift: 1991 Tanganhydro
expedition: Bulletin des Centres de Recherches ExplorationProduction Elf Aquitaine, v. 16, p. 229–247.
Theunissen, K., J. Klerkx, A. Melnikov, and A. Mruma, 1996, Mechanisms of inheritance of rift faulting in the western branch of
the East African rift, Tanzania: Tectonics, v. 15, p. 776–
790.
Tiercelin, J. J., and A. Mondeguer, 1991, The geology of the Tanganyika trough, East African rift, in G. W. Coulter, ed., Lake
Tanganyika and its life: Oxford, British Museum Natural History Publications and Oxford University Press, p. 7–48.
Tiercelin, J. J., J. Chorowicz, H. Bellon, J. P. Richert, J. T. Mwanbene, and F. Walgenwitz, 1988, East African rift system: offset, age and tectonic significance of the Tanganyika-RukwaMalawi transcurrent fault zone: Tectonophysics, v. 148,
p. 241–252.
Tiercelin, J. J., et al., 1993, Hydrothermal vents in Lake Tanganyika,
East African rift system: Geology, v. 21, p. 499–502.
Trudgill, B. D., and J. A. Cartwright, 1994, Relay ramp forms and
normal fault linkages—Canyonlands National Park, Utah: Geological Society of America Bulletin, v. 106, p. 1143–1157.
Van Wyk de Vries, B., and O. Merle, 1996, The effect of volcanic
constructs on rift fault patterns: Geology, v. 24, no. 7, p. 643–
646.
Versfelt, J., and B. R. Rosendahl, 1989, Relationships between prerift structure and rift architecture in lakes Tanganyika and Malawi, East Africa: Nature, v. 337, p. 354–357.
Wafula, M., N. Zana, T. Mavonga, and A. M. Sassa, 1992, Recent
seismicity of the Virunga volcanic zone, western rift, Zaire:
Tectonophysics, v. 209, p. 259–260.
Weissel, J. K., and G. D. Karner, 1989, Flexural uplift of rift flanks
due to mechanical unloading of the lithosphere during extension: Journal of Geophysical Research, v. 94, p. 13919–13950.
Wohlenberg, J., 1975, Geophysikalische Aspekte der Ostrafrikanishen Gragenzonen: Geologisches Jahrbuch, Reihe E, Geophysik, v. 4, p. 1–82.
Zana, N., K. Tanaka, and M. Kasahara, 1992, Main geophysical features related to the Virunga zone, Western rift, and their volcanological implications: Tectonophysics, v. 209, p. 255–
257.
Zhao, M., C. A. Langston, A. A. Nyblade, and T. J. Owens, 1997,
Lower-crustal rifting in the Rukwa graben: East Africa: Geophysical Journal International, v. 129, p. 412–420.
Lezzar et al.
1059