Archives / 2018 / June

Remarkable Regions – The Kenya Rift

Remarkable Regions – The Kenya Rift

Every 8 weeks we turn our attention to a Remarkable Region that deserves a spot in the scientific limelight. After looking at several convergent plate boundaries, this week the focus lies on part of a nascent divergent plate boundary: the Kenya Rift. The post is by postdoctoral researcher Anne Glerum of GFZ Potsdam.

Of course an active continental rift is worthy of the title “Remarkable Region”. And naturally I consider my own research area highly interesting. But after seeing it up-close and personal on a recent 10-day trip organized by the University of Potsdam, Roma Tre and the University of Nairobi (stay tuned for the travel log, or read that of the University of Potsdam), I must say, the Kenya Rift is a truly beautiful and fascinating region.

Figure 1. Topography (Amante and Eakins 2009) and kinematic plate boundaries (Sarah D. Stamps based on Bird 2003) of the East African Rift System (EARS). Plate boundary colors schematically indicate the western and eastern branches of the EARS.

Constituting one segment of the 5000 km long East African Rift System (EARS, Fig. 1), the Kenya Rift is host to an amazing landscape, wildlife and people, all of which somehow tie back to continental rifting processes. Although the youngest rifting phase in Kenya commenced in the Miocene, the east African region as a whole has been shaped by rifting episodes since Permian times (Bosworth and Morley 1994). The present active rift system runs from the Afar region in the north all the way south to Mozambique and is split into a western and an eastern branch that run around the Archean Tanzanian Craton (Chorowitz 2005, see Fig. 1). Generally speaking, the western branch is more seismically active, but deprived of magmatism, compared to the eastern branch, of which the Kenya Rift is part (Chorowitz 2005). Three processes characterize the EARS (Burke 1996) as well as the Kenya Rift specifically: normal faulting, volcanism and uplift.


The Tanzanian Craton together with the enveloping western and eastern EARS branches constitutes the broad, uplifted area coined the East African Plateau (~1200 m elevation, Strecker 1991; Simiyu and Keller 1997, Fig. 2). The onset of uplift of this plateau can be constrained to the Early Miocene with the help of one of the longest phonolitic lava flows on Earth (> 300 km, Wichura et al. 2010; 2011) and a whale that stranded inland 17 Ma (and was only recently found again after going missing for 30 years, Wichura et al. 2015). Plume-lithosphere interaction is thought responsible for the uplift (e.g. Wichura et al. 2010), although there is disagreement about the continuity of the low seismic velocity anomalies seen in the east African upper mantle and whether they are connected to the lower mantle. For example Ebinger and Sleep (1998), Hansen et al. (2012), Sun et al. (2017) and Torres Acosta et al. (2015) advocate for one East African superplume, while Pik et al. (2006) distinguish separate lower and upper mantle plumes and Davis and Sack (2002) and Halldórsson et al. (2014) consider a lower mantle plume splitting in the upper mantle.

Figure 2. Topography (Amante and Eakins 2009) and fault traces (GEM) of the central EARS. Triangles indicate off-rift volcanoes, dotted grey lines the three segments of the Kenya Rift.

Magmatism and volcanism

The northward motion of Africa over this hot mantle anomaly has been thought the cause of a north-to-south younging trend in the age of the ensuing EARS volcanism and rifting (e.g. Ebinger and Sleep 1998; George et al. 1998; Nyblade and Brazier 2002), although more recent studies arrive at a more spatially disparate and diachronous rifting evolution (Torres Acosta et al. 2015 and references therein). In general, massive emplacement of flood-phonolites preceded the onset of rifting in Kenya around 15 Ma (Torres Acosta et al. 2015). With ongoing rifting, and localization of faulting towards the rift axis, volcanism also migrated towards the center of the rift. Since the Miocene, massive amounts of volcanics have thus been emplaced (144,000-230,000 km3, MacDonald 1994; Wichura et al. 2011). Moreover, dyking also accommodated a significant part of the extension, with 22 to 26 % of the crust in the rift valley being composed of dykes (MacDonald 2012). Not surprisingly, the highlands directly around the rift valley, the Kenya Dome (Fig. 2) formed through a combination of volcanism and uplift (Davis and Slack 2002) with elevations of up to 1900 m.

The composition of rift magmatism is bimodal, showing phonolites and trachytes on the one side and nephelinites and basalt on the other, predominantly resulting from fractional crystallization of a basaltic source. The low viscosity of these magmas allows the young volcanoes in the volcano-tectonic axis to reach significant heights (see Fig. 3; MacDonald 2012). The most impressive volcanoes are to be found outside of the rift however (Fig. 2), with Mnt. Elgon reaching 4321 m and Africa’s highest mountains Mnt. Kenya and Mnt. Kilimanjaro reaching up to 5200 m and 5964 m, respectively (Chorowitz 2005).

Figure 3. View on the crater rim of the 400 ky old Mnt. Longonot volcano in the tectono-magmatic rift axis, at 2560 m asl. Courtesy of Corinna Kallich, GFZ Potsdam.

Normal faulting

The Kenya rift itself is composed of 3 asymmetric segments, distinguished by sharp changes in their orientation (Chorowitz 2005, Fig. 2). The 2300-3000 m high Elgeyo, Mau and Nguruman escarpments result from the steep Miocene east-dipping border faults in the west, while the antithetic border faults on the eastern side formed later during the Pliocene (Strecker et al. 1990). The older border faults formed along preexisting foliation generated by the Mozambique Belt orogeny in the late Proterozoic (Shackleton 1993; Hetzel and Strecker 1994). A change in strike of this foliation from NNE in the northern and southern Kenya rifts to NW determined the change in orientation in the central Kenya rift (Strecker et al. 1990). Consequently, different generations of faults in the northern and southern rift segments run parallel, while in the central segment, the Pleistocene change in extension direction from ENE-WSW/E-W to the present-day WNW-ESE/NW-SE directed extension results in obliquely reactivated border faults and younger, en echelon arranged left-stepping NNE-striking fault zones along the rift axis (Strecker et al. 1990). Extension is transferred between the different zones by coeval normal and strike-slip faulting or dense sets of normal faults.

Figure 4. View of lake Magadi and the Nguruman escarpment. Lake Magadi is a saline, alkaline lake, commercially mined for trona. Courtesy of Corinna Kallich, GFZ Potsdam.

Human evolution

The uplift, volcanism and normal faulting together have set the stage for human and animal evolution. For example, the shift in hoofed mammals from eating predominantly woods to grazing species evidences that the large-scale uplift modified air circulation patterns resulting in aridification and savannah-expansion at the expense of forested areas (Sepulchre et al. 2006; Wichura et al. 2015). The rift basins enabled the formation of large lakes, which were subsequently compartmentalized by tectonic and volcanic morphological barriers (Fig. 4). On the short-term, lake coverage varied due to tectonically induced changes in catchment areas, drainage networks and outlets. Maslin et al. (2014) actually found a correlation between this ephemeral lake coverage and hominin diversity and dispersal. Lake highstands link with the emergence of new species and allowed the spread of hominins north and southward out of east Africa. Remarkable, or what!

Amante, C. and Eakins B. W., 2009. NOAA Technical Memorandum NESDIS NGDC-24. National Geophysical Data Center, NOAA.
Bosworth, W. and Morley, C.K., 1994.  Tectonophysics 236, 93–115.
Burke, K., 1996. S. Afr. J. Geol. 99 (4), 339–409.
Chorowitz, J., 2005. J. Afr. Earth Sci. 43, 379-410.
Davis, P. M. and Slack, P. D. 2002. Geophys. Res. Lett. 29 (7), 1117.
Ebinger, C.J. and Sleep, N.H., 1998. Nature 395, 788-791.
George, R. et al., 1998.  Geology 26, 923–926.
Halldórsson, S. A. et al., 2014. Geophys. Res. Lett. 41, 2304–2311,
Hansen, S. E. et al., 2012.  Earth Planet. Sc. Lett. 319-320, 23-34.
Hetzel, R., Strecker, M.R., 1994. J. Struct. Geol. 16, 189–201.
Macdonald, R. et al., 1994a. J. Volcanol. Geoth. Res. 60, 301–325.
Macdonald, R., et al., 1994b. J. Geol. Soc. London 151, 879–888.
MacDonald, R., 2012. Lithos 152, 11-22.
Maslin, M. A. et al., 2014. Quaternary Sci. Rev. 101, 1-17.
Nyblade, A. A. and Brazier, R. A., 2002. Geology 30 (8), 755-758.
Pik, R. et al., 2006. Chem. Geol. 266, 100-114.
Sepulchre, P. et al., 2006. Science, 1419-1423.
Shackleton, R.M., 1993. Geological Society, London, Special Publications 76, 345–362.
Simiyu, S.M., Keller, G.R., 1997. Tectonophysics 278, 291–313.
Strecker, M., 1991. Das zentrale und südliche Kenia-rift unter besonderer berücksichtigung der neotektonischen entwicklung, habilitation, Universität Fridericiana.
Sun, M. et al., 2017.  Geophys. Res. Lett. 44, 12,116–12,124.
Torres Acosta, V. et al., 2015. Tectonics 34, 2367–2386.
Wichura, H. et al., 2010. Geology 38 (6), 543–546.
Wichura, H. et al , 2011. The Formation and Evolution of Africa: A Synopsis of 3.8 Ga of Earth History, eds. D. J. J. Van Hinsbergen, S. J. H. Buiter, T. H. Torsvik, C. and Gaina, S. J.
Wichura, H. et al., 2015. P. Natl. Acad. Sci. USA 112 (13), 3910-3915.

The art of the 15-minute talk

The art of the 15-minute talk

We’ve all attended conferences with those dreaded 15-minute talks and we have no problem picking out which talks were amazing and which talks were abysmal. However, when it comes to our own talks, it’s hard to judge them, find out how they can be improved or break away from long-established habits (such as our layout or talking pace). This week, Matthew Herman, postdoc at the Tectonophysics research group at Utrecht University in the Netherlands, guides you towards your best 15-minute talk yet!

At some point in your career as an Earth Scientist, you will hopefully have a chance to give a 15-minute talk at a meeting, a colloquium series, or simply in your lab group. This provides a great opportunity to advertise your hard work to your colleagues in an amount of time that is well within a human attention span. Ultimately, your goal in this talk is to effectively communicate your discovery to your audience. In the process, you get to explain the importance of your field, pose a crucial research question in that field, demonstrate cutting-edge analyses and applications, and, finally, provide an answer to that initial research question, sometimes for the very first time.

Despite all the latent potential for a 15-minute talk to captivate and teach the audience, many of these presentations end up being uninformative. I do not intend this as a judgment regarding the significance or quality of the science. I have seen incomprehensible talks from people whose research is crucial to our understanding of the Earth system. Alternatively, I have seen talks presenting incremental scientific advancements that were truly enlightening. But from all the diverse presentations I have seen, there are common elements that either dramatically improved or reduced my understanding of the subject matter. My aim here is to provide what I think are some of these key characteristics that make up a really excellent talk, so that next time you have the opportunity to present, you will inspire your audience.

I think there are two general things to keep in mind for your 15-minute talk: (a) you have limited time with your audience, and (b) the expertise of your audience can vary a lot. This means that you should design a presentation that fits your extensive understanding into a brief window and tailor the details for the particular audience that will be attending. If this makes it seem like it will take time and effort to construct an effective talk, that is because it is true! Even if you have a well-received publication, simply transferring figures, analyses, and interpretations from the paper into your talk is almost guaranteed to lead to an ineffective presentation – it will probably be too long, too technical, and too difficult to see from the back of the room. If you really want your audience to concentrate on your work for the full 15 minutes, take the time (potentially up to a few weeks) to craft a great talk. And one more thing: you really should practice your talk ahead of time. Actually, I cannot emphasize this point enough: PRACTICE.

Note: If you are short on time right now, I have included a checklist at the end to summarize the main points.

How long?

Imagine: you are in the audience and the end of the talk is not in sight. You shift in your seat uncomfortably as you glance at your watch. The speaker does not appear to notice the amount of time since they started, but you definitely do: 14:30… 15:00… 15:30. Finally, two full minutes after the end of the scheduled time slot, the speaker asks if there are any questions, but of course there is no time for that. Many otherwise good talks have been ruined for me by the presenters going into overtime. All I can now remember about them is by how much they exceeded the final bell. As a speaker, you have 15 minutes – choose a topic and present it in the allotted time frame. In fact, target your talk for 12-13 minutes so your audience can ask questions at the end.

This, and that, and these…

The detailed structure of the talk is flexible, but should probably contain the following items: background/motivation (Why should we, the audience, care?); a research question or hypothesis (What is being tested?); observations, models, and analysis (How is the research question being answered?); and interpretations and conclusions (GIVE US THE ANSWERS!).

i. Background
Try to avoid dwelling on the background for too long. I know many of us (myself included) enjoy pedantically explaining the rich history of our field leading up to the present day. But you do not have the time in a 15-minute talk. As you are constructing your presentation, you should budget no more than 2-3 minutes at the start to establish the context for your research problem. At that point, your audience should be oriented and ready to be amazed by your results.

Example of an introduction/background slide

ii. Research Question
Do not assume that your research question or hypothesis is obvious to everyone. People come to talks for a lot of different reasons; sometimes they are experts in the field, but other times they saw a keyword in your title or abstract, or maybe there were no other interesting sessions. In any case, it is likely that a good percentage of your audience does not know what specifically you are testing if you do not tell them. After setting up the background, verbally or on the screen state your research question or hypothesis.

iii. Observations, Models, and Analysis
This will be the bulk of your presentation. Tailoring your 15-minute talk for your specific audience means you will want to use just the right amount of technical terminology. You should assume some foundational level of knowledge because there is no way to define every term and present the complete theory for your research. But for the most part, I think you should try to minimize technical jargon (particularly uncommon acronyms) in talks. If and when you need to use a term repeatedly, then take 15-30 of your precious seconds to concisely explain the concept, ideally without patronizing or condescending. [Did I mention this was a difficult balance?] Incidentally, explaining a concept has the added benefit of forcing you to understand the concept sufficiently that you can distill its definition into a compact form for your listeners.

The precise minimum level of knowledge you assume for your audience depends on the setting. In the large lecture hall of an international meeting like the EGU General Assembly, the audience may be weighted towards less experience in your field, whereas a special meeting focused on your subject area will likely have a higher percentage of experts.

A related point is that you should avoid all but the most straightforward equations. The reality is that any audience member who does not already understand the equation is not going to understand it from your talk. There is not enough time, and the medium is not amenable to higher level math. Simple equations with a couple variables are okay, but anything with multiple terms, powers, derivatives, etc. are a waste of time.

iv. Interpretations and Conclusions
Honestly, most people are pretty good at this part. This is the most fun and exciting aspect of the talk, plus it means the end is near. A couple minor pieces of advice: (a) make sure you have drawn a clear path from the background through the analyses and into the interpretations, with the common thread being answering your research question; and (b) I think it is best to limit the number of conclusions to 3-4 (consider this in the preliminary design stage of the talk as well!).

Example of a results slide

Good looks matter

I try to follow the advice of the great Jim Henson when it comes to designing the look for my talks: “Simple is good.” I will not harp on making figures, because many other people have discussed how to design good ones. In a nutshell: make them big, use good color schemes and large fonts, and keep them uncluttered. Resist the urge to copy figures straight from papers to your talk. You will probably need to simplify a figure from the published version in order to make it optimal for your talk. Sometimes you just need to design and produce a totally new figure. In fact, making figures is where I spend at least 65% of my time when I am preparing a talk.

In terms of slide layout, use the whole slide. Borders, icons, and backgrounds can be pretty flourishes, but they take up valuable real estate. Every centimeter you use for a border is a centimeter you can no longer use for a making a figure nice and big. And remember there will be people, some with poor eyesight, in the back of the room. As on figures, limit the amount of text. When you do need labels or bullet points, use a classic, simple font (I will scream if I see Comic Sans one more time…) in a large size – I typically use no smaller than 24-point font Helvetica.

Closing remarks

Many of my suggestions are more like guidelines than hard rules. I enjoy seeing creative and innovative presentations. As long as you give yourself enough time to craft an excellent presentation, then take time to practice it in front of friends, it will turn out well. Hopefully we will all see a large collection of great talks in the next few meetings. See you there!

Remember: the goal of the talk is for your audience to understand your science!

• Take time (up to several weeks) to construct your presentation
• Practice before the date of your talk, if possible in front of a test audience

• Target talk length for 12-13 minutes (do not go over 14!)
• Limit background or introductory information to 2-3 minutes
• Explicitly state research question
• Link background, analysis, and interpretations to research question
• Limit conclusions to ≤ 4

Scientific Content
• Choose technical jargon at level appropriate for audience
• Define critical terminology in 30 seconds or less
• Limit acronyms
• Avoid complex equations
• Avoid tables

Visual Content
• Fill space on slide, especially with figures
• Make thin frames to not waste precious room
• Choose large font sizes (≥ 24 pt) in a standard font
• Adjust figures from published version
• Check figure color contrasts (avoid blue/black, yellow/white)
• Use perceptually linear color palettes (no rainbow!)
• Avoid cartoons, animations, and sounds

General Life Advice
• Use common sense (e.g., do not include pictures from the bar in your talk)