P AND S BODY WAVE TOMOGRAPHY OF THE NORTHERN LAKE MALAWI RIFT BASIN AND RUNGWE VOLCANIC PROVINCE

P AND S BODY WAVE TOMOGRAPHY OF THE NORTHERN LAKE MALAWI RIFT BASIN AND RUNGWE VOLCANIC PROVINCE

ABSTRACT

The Lake Malawi Rift, situated at the southern end of the East African Rift System, is in an early stage of rifting. The estimated available tectonic forces are insufficient to initiate rupture, which suggests a weakening mechanism such as magmatism and/or dike intrusions is assisting rift development. Magmatism can thermally weaken the lithosphere and thus enable continental rupture to occur at a lower tectonic stress. To determine if there is seismic evidence for thermal perturbations in the upper mantle, I investigated the P- and S-wave velocity structure of the upper mantle beneath the northern end of the Lake Malawi Rift basin and Rungwe Volcanic Province (RVP). The upper mantle velocity structure has been tomographically imaged using P- and S-wave relative arrival time residuals from earthquakes recorded on the SEGMeNT seismic network and several previous networks in eastern Africa. Tomographic images of P- and

S-wave velocity models reveal a prominent low wave speed anomaly (LWA) with reductions in Vp of 1.5 – 2.2% and Vs of 2 – 3% beneath the RVP between depths of ~50 to 300 km. This LWA can be attributed to a 100 – 200 K thermal anomaly. The tomographic images indicate that there is no basin wide ( > 75 km) thermal anomaly extending beneath the northern part of the Lake Malawi rift from the RVP. However, the presence of localized heating and/or magmatic modification of the lithosphere beneath the Lake Malawi rift from magmatic activity such as dike swarms and small pockets of melt on a sub-basin scale cannot be ruled out. Three geodynamic models for the origin of the LWA beneath the RVP, superplume, lithospheric drip, and smallscale convection, are examined. The lithospheric drip and small-scale convection models cannot readily explain magmatism in the RVP and thus are not favored. In examining the superplume model, a direct connection is not seen in the tomographic models between the African superplume and the RVP LWA. However, the LWA could be linked to the superplume structure by a small upwelling that is too narrow to be imaged or alternatively, via a small thermal plumelet that rose from the superplume structure in the mantle transition zone.

 

TABLE OF CONTENTS

List of Figures……………………………………………………………………………………………………………….. vii

List of Appendices………………………………………………………………………………………………………… xiii

 

Acknowledgements …………………………………………………………………………..     xv

Chapter 1 Introduction ………………………………………………..……………………..       1

1.1 East African Rift System …………………………………………………     1

1.2 Background ……………………………………………………………….    4

1.3 Rungwe Volcanic Province ……………………………………………….     7

1.4 Previous Seismic Studies …………………………………………………     8

Chapter 2 Data and Methodology………………………………………………………..……     9

2.1 SEGMeNT Network ………………………………………………………    9

2.2 Selection of Data ………………………………………………………….    12

2.3 Determination of relative arrival-time residuals …………………………..   15

2.4 Model Parameterization and Inversion Method …………………………..    19

Chapter 3 Results………………………………………………………………….…….……..   23

3.1 Event Relocations …………………………………………………….……  23

3.2 Station Terms ……………………………………………………….……..   24

3.3 P-wave Model …………………………………………………….……….   26

3.4 S-wave Model ……………………………………………………………..   30

3.5 P-wave Resolution Tests …………………………………………………..   34

3.6 S-wave Resolution Tests …………………………………………………..   36

Chapter 4. Discussion………………………………………………..……………………..…    38 Chapter 5. Summary and Conclusions…………………………….………….………………    53

Bibliography……………………………………………………………..………………..…..    55

Appendix  ……….……………..……………………………………………….………….….   65

Chapter 1: Introduction

1.1: East African Rift System

An outstanding challenge in continental dynamics is understanding how strong

continental lithosphere ruptures in the early stages of rifting (Buck 2006; 2004). Continental rupture can be achieved with a modest amount of tectonic stress if a weakening mechanism is present, for example, magmatism and/or dike intrusions, but is difficult to achieve in the absence of a weakening mechanism (Buck 2006; 2004). Many studies (e.g., Bastow and Keir, 2011;

Karner et al., 2004; Buck, 2004, 2006; Kampunzu and Lubala, 1991; White and McKenzie,

1989; Morgan, 1971, 1981, 1983; Bott, 1991; Sengor and Burke 1978; Spohn and Schubert, 1982; Ebinger, 1989) have suggested magmatism can assist in the development of early-stage rifts by thermally altering the lithosphere and consequently weakening and reducing its strength.

O’Donnell et al. (2016) calculated yield strength envelopes to estimate the strength of the lithosphere in non-magmatic regions of the East African Rift System (EARS) using a suite of reasonable crust and mantle compositions with various grain sizes, Byerlee’s law (Byerlee,

1968), and a strain rate of 1 x 10^-16 s^-1. From the strength envelope calculations, O’Donnell et al

(2016) determined that extensional stresses in excess of 280 MPa would be needed to rift the

East African lithosphere. This contrasts with a range of estimated extensional tectonic stresses in East Africa (10-22 MPa, Stamps et al. 2014; 35 MPa, O’Donnell et al. 2016; 125 MPa, Craig et al. 2011). It is evident that the available tectonic stresses are insufficient to initiate rupture of the lithosphere, and hence thermal weakening of the lithosphere is often invoked to explain the onset of rifting in East Africa. Thermal perturbation in the lithosphere from magmatism is prevalent in several parts of the EARS that coincide with volcanic centres (e.g., Ebinger and Casey, 2001; Ebinger, 2005; Kendall et al., 2005). However, it is unknown to what extent magmatism plays a role, if any, in accommodating extension in parts of the rift system where there is no surface expression of magmatism (Chorowicz, 2005).

The Lake Malawi (Nyassa) rift, situated near the southern end of the EARS, is a weakly extended continental rift (< 15%) that has developed in Proterozoic mobile belts, and there is no evidence for volcanism within the rift (Furman, 2007). The Cenozoic Rungwe Volcanic Province (RVP), however, lies at the northern end of the rift, and therefore weakening of the lithosphere under the rift may have occurred if the RVP magmatism at depth extends to the south under the

rift.

To illuminate crustal and upper mantle structure for investigating the process of earlystage rifting, the SEGMeNT (Study of Extension and maGmatism in Malawi aNd Tanzania) experiment (Shillington et al. 2016) obtained a suite of datasets, including active- and passivesource seismic, GPS, magnetotelluric, and geochemical in and around the northern part of the Lake Malawi rift (Fig. 1). In this thesis, I utilize teleseismic data from the SEGMeNT experiment, as well as teleseismic data from previous networks in East Africa to investigate upper mantle structure beneath the northern part of the Lake Malawi rift and RVP to determine if there is any seismic evidence for magmatic and/or thermal alteration of the lithosphere beneath the northern part of the rift. This study provides the first detailed seismic tomographic images of upper mantle P- and S-wave speed variations beneath the RVP and the northern Lake Malawi region.

 

1.2: Background

The Lake Malawi (Nyassa) rift lies near the southern end of the Western Branch of the East African Rift System (EARS), an active continental rift system in an early stage of breakup (Fig. 1).

Figure 1: Topographic map of the East African Rift System. Regional structures shown include the Tanzania craton and Cenozoic rift faults. The red box in the inset shows the region in the map, and the black bold box encompasses the study area where the SEGMeNT project is located. Blue arrows show extension rate, 1 mm/yr.

The EARS is comprised of two branches, the Western Branch and the Eastern Branch, that developed in Proterozoic mobile belts surrounding the cold and thick lithosphere of the Archean Tanzania craton (Figs. 1 and 2) (Cahen el al., 1984; Shackleton, 1986).

Figure 2: Geologic map of the region surrounding Lake Malawi showing the locations of the SEGMeNT seismic stations. The North, Central, and South Basins in Lake Malawi are labeled as 1, 2, 3, respectively. The geology is modified from Le Gall et al (2008), (Pinna et al (2004), and Mulibo and Nyblade (2013).

 

The Lake Malawi rift stretches over 800 km from the RVP in the north to the Urema

Graben (Mozambique) in the south (Fig. 1) (Delvaux, 1991). Lake Malawi, a 550 km long, 50 – 80 km wide and 700 m deep lake, occupies the main part of the rift (Delvaux, 1991). The lake is surrounded by Malawi to the west, Tanzania to the northeast, and Mozambique to the southeast. The geometry of the rift is defined by side stepping en echelon faults, with a high angle border fault on one side which accommodates most of the extension (Ebinger et al., 1989). Three basins

(North, Central, and South) partition Lake Malawi and are linked via transfer faults (Fig. 2). Only a modest amount of crustal stretching is thought to have occurred, with estimates from flexural models of the basin and flank morphology, as well as fault reconstructions, suggesting < 15% (Ebinger et al., 1987). GPS  data indicate a plate-opening velocity of ca. 1 mm/yr (Saria et al., 2014).

The Lake Malawi rift developed along the boundary of three Proterozoic mobile belts, the

Paleoproterozoic Ubendian Belt, the Mesoproterozoic Irumide Belt, and the Neoproterozoic Mozambique Belt (Fig. 2). The Ubendian Belt consists of granulite and amphibolite facies gneisses and meta-sedimentary rocks that formed between 2.03 and 1.86 Ga (Cahen et al., 1984;

Lenoir et al., 1994; Schlüter and Hampton, 1997). The Irumide belt, situated south of the

Ubendian Belt, is comprised of 2.05–1.85 Ga volcano-plutonic complexes and gneisses, a 1.85 Ga quartzite–metapelite succession with minor carbonates, and granitoids emplaced between

1.65 and 1.55 Ga (Begg et al., 2009). The Mozambique Belt, lying east and southeast of the Archean Tanzania craton and Paleoproterozoic Usagaran Belt, consists of granulites and gneisses that formed between 1200 and 450 Ma (Cahen et al., 1984; Shackleton, 1986).

 

1.3: Rungwe Volcanic Province

The RVP, situated within a complex accommodation zone between the Rukwa, Malawi and Usanga rifts, is the only surface expression of magmatism within the southern part of the Western Branch (Ebinger et al. 1989) (Fig. 2). Three NW-SE trending, alkaline volcanoes within the province (from the NW to SE: Ngozi, Rungwe, Kyejo) have erupted in the past 10 Ka (Fontijn et al. 2010a, b; 2012). Basalts, trachytes, and phonolites are the most common lava types found within the RVP (Fontijn et al., 2010a, b). The youngest lavas erupted from Rungwe and

Ngozi (< 1 Ka) have evolved to being mostly silica-saturated compared to the youngest lavas of

Kyejo (< 500 Ka), which are comprised of more nephelinitic compositions (Fontijn et al., 2010b; Furman, 1995).

 

The relative timing of the initiation of rift development within the Rukwa-Malawi rift zones is debated. Roberts et al. (2012) suggest that the onset of rifting initiated as early as 25 Ma, which is earlier than many of the ages reported for the RVP volcanism (7 – 9.2 Ma) (Ebinger et al., 1989, 1993). A recent geochronology study using argon dating on mafic lavas found to phonolite domes in the Usangu Basin to be 17.6 – 18.5 Ma old (Mesko et al., 2014).

 

7

1.4: Previous Seismic Studies

 Previous seismic studies of crustal and upper mantle beneath the Lake Malawi region have been restricted to a few basin-scale (shallow stratigraphy of the Lake Malawi rift basins) and regional-scale studies (Lyons et al. 2011; Rosendahl, 1987; Adams et al. 2012; Mulibo and Nyblade 2013; O’Donnell et al. 2013, 2016). The regional-scale tomography studies of Adams et al. (2012), Mulibo and Nyblade (2013), and O’Donnell et al. (2013, 2016) imaged a ~100 km wide, circular low velocity anomaly centered under the RVP that extends to a depth of ~140 km. O’Donnell et al. (2013) attribute the anomaly to a 100 – 200 K thermal perturbation of the uppermost mantle, and also possibly to the presence of partial melt in the mantle lithosphere. Limited resolution in these regional-scale studies make it difficult to resolve upper mantle structure beneath the northern Lake Malawi region, leaving open the question of whether or not the thermally perturbed upper mantle beneath the RVP extends to the south beneath Lake Malawi.

P AND S BODY WAVE TOMOGRAPHY OF THE NORTHERN LAKE MALAWI RIFT BASIN AND RUNGWE VOLCANIC PROVINCE

Sharing is caring!