\documentclass[11pt]{article}

\usepackage[margin=1in]{geometry}
\usepackage{amsmath}
\usepackage{amssymb}
\usepackage{graphicx}
\usepackage{booktabs}
\usepackage{tikz}
\usepackage{pgfplots}
\pgfplotsset{compat=1.18}
\usepackage[colorlinks=true, linkcolor=black, citecolor=black, urlcolor=blue]{hyperref}

% --- skin palette: magma rust, gold, teal, violet (the four cluster colours) -
\definecolor{afaraccent}{HTML}{B3421F}
\definecolor{afargold}{HTML}{93701C}
\definecolor{afarteal}{HTML}{1C7A68}
\definecolor{afarviolet}{HTML}{6A4B8F}
\definecolor{afarneutral}{HTML}{6E6A5F}

\title{A Pulsing Mantle Plume, Timed by the Plate Above It}
\author{The Science Journaling Club}
\date{Field Notes --- Paper Analysis --- Geodynamics}

\begin{document}
\maketitle

\begin{abstract}
Under the Afar triple junction, where the African, Arabian and Somali plates
are simultaneously pulling apart, a mantle plume is supposed to rise the way
textbook plumes rise: one hot column, one composition gradient fading
smoothly outward in every direction. Watts et al.\ measured five radiogenic
isotope ratios and three trace-element ratios, plus geophysical constraints
such as crustal thickness and extension rate, across more than 130 new young
volcanic rock samples spanning all three rift arms (79 from across Afar, 52
from dedicated Main Ethiopian Rift fieldwork), supplemented with 93 existing
analyses from the GEOROC database, and sorted the results with $k$-means
clustering on 13 standardized variables \cite{watts2025}. No bullseye
appeared. Instead the same handful of cluster identities recur along each
rift arm at a fixed spacing that differs by rift: roughly 50--100~km in the
Main Ethiopian Rift (MER, four clusters) against roughly 150--200~km in the
Red Sea Rift (RSR, three clusters), a rift extending two to four times faster
over notably thinner crust \cite{watts2025}. This Field Note reconstructs that
comparison and then performs an arithmetic step the paper itself does not
publish: converting each rift's spacing back into an implied time between
pulses using that rift's own reported extension rate, $T = \lambda/R$. The two
rift arms' implied periods overlap almost entirely, in a window of roughly
9.6--19.0~million years, despite differing four-fold in raw kilometres --
consistent with one upwelling pulsing at a broadly steady rate and each plate
stretching the record of it at its own speed. The paper's own reading of the
mechanism inverts the usual causal arrow in rift-plume stories: the
overriding plate's thickness and speed are shown to actively shape how the
plume's signal is laid down, rather than the plume simply dictating outcomes
to a passive rift above it \cite{watts2025}.
\end{abstract}

%============================================================================
\section{Three Plates, One Corner of Ethiopia}

The Afar Depression sits at the one place on the planet where three
tectonic plates -- African, Arabian and Somali -- are actively rifting apart
from a shared hinge. Three rift arms radiate from that hinge, each at a
different stage of turning into an ocean basin: the Red Sea Rift to the
northwest, already advanced enough to be flooded by seawater in places; the
Gulf of Aden Rift to the east-southeast, the most tectonically mature of the
three at roughly 35~million years of activity; and the Main Ethiopian Rift
to the south, youngest of the three and still tearing through thick
continental crust rather than the thinner, more oceanic-style crust the
other two arms have already built \cite{watts2025}. Geologists have treated
this as a single natural laboratory for decades: the same rifting process,
caught at three different ages within one region, widely attributed to a
single underlying mantle plume feeding all three arms
\cite{ebingersleep1998}.

%============================================================================
\section{What a Plume Is Supposed to Do}

W.\ Jason Morgan proposed in 1971 that narrow columns of anomalously hot,
buoyant mantle rock rise from deep inside the Earth and remain roughly fixed
in place while tectonic plates drift over them, producing trails of extinct
volcanism behind an active hotspot \cite{morgan1971}. The rock need not melt
by heating up further; it can cross its own melting point purely by rising
into lower pressure, the same way a rock's melting temperature falls as
confining pressure falls. Watts et al.\ describe the upwellings feeding
places like Afar as anomalously hot mantle originating between roughly
1{,}000 and 2{,}800~km depth -- deep enough to implicate a genuinely
lower-mantle process \cite{watts2025}, a depth range broadly consistent with
earlier seismic imaging connecting the Ethiopia/Afar hotspot to the wider
African superplume structure \cite{hansennyblade2013}.

Ebinger and Sleep argued in 1998 that the entire chain of Cenozoic volcanic
fields from Afar down through Kenya and Tanzania could be explained by a
single plume head spreading beneath the lithosphere and erupting through
pre-existing weaknesses, rather than by several independent small plumes
\cite{ebingersleep1998}. A single plume head predicts something specific
about the geochemistry directly above it.

%============================================================================
\section{The Null Model: a Bullseye}

Treat a plume as one thermally and chemically zoned conduit and the natural
prediction is radial symmetry: composition should vary smoothly with
distance from the conduit axis, and -- this is the falsifiable part -- vary
the \emph{same way} regardless of which compass bearing you walk away from
the centre. Watts et al.\ state the null model directly: a single upwelling
centre in which the variation of a given geochemical property with distance
is ``assumed common to all rifts'' \cite{watts2025}. Walk out along the Red
Sea arm or the Main Ethiopian Rift arm, and a given fractional change in
isotope ratios should, under this model, appear at comparable distances in
both directions.

%============================================================================
\section{Measuring the Signal: Isotopes and Trace Elements in 130-Plus Rocks}

Radiogenic isotope ratios are the standard tool for reading a rock's mantle
ancestry, because the ratio of a radiogenic daughter isotope to a stable
isotope of the same element depends on how much of the radioactive parent
element that piece of mantle accumulated, and for how long. Decades of work,
formalized by Zindler and Hart in 1986, have catalogued recurring mantle
``flavours'' worldwide: DMM (depleted MORB mantle, the default background),
HIMU (high time-integrated U/Pb), and the enriched components EMI and EMII
\cite{zindlerhart1986}.

Watts et al.\ measured five isotope ratios --
$^{87}\mathrm{Sr}/^{86}\mathrm{Sr}$, $^{143}\mathrm{Nd}/^{144}\mathrm{Nd}$,
and three lead ratios,
$^{206}\mathrm{Pb}/^{204}\mathrm{Pb}$, $^{207}\mathrm{Pb}/^{204}\mathrm{Pb}$
and $^{208}\mathrm{Pb}/^{204}\mathrm{Pb}$ -- plus three trace-element
ratios, La/Sm, Ce/Pb and the niobium anomaly $\Delta$Nb, across more than
130 young volcanic rocks: 79 new samples from across Afar and 52 more from
dedicated Main Ethiopian Rift fieldwork, roughly doubling the number of
high-quality analyses previously available from the area
\cite{watts2025}. They supplemented this with 93 existing analyses drawn
from the GEOROC geochemical database, bringing the total dataset feeding the
statistical analysis to 224 samples -- a sum the authors do not themselves
report as a single figure, arrived at here simply by adding their reported
counts.

\begin{table}[htbp]
\centering
\begin{tabular}{@{}lcc@{}}
\toprule
\textbf{Source} & \textbf{Samples} & \textbf{Status} \\
\midrule
New Afar samples & 79 & new \\
New MER fieldwork samples & 52 & new \\
Existing GEOROC records & 93 & previously published \\
\midrule
\textbf{Total} & \textbf{224} & (club's sum) \\
\bottomrule
\end{tabular}
\caption{Sample provenance for the statistical analysis \cite{watts2025}. The
79+52 figure matches the paper's reported ``more than 130'' new samples; the
224 total is the club's own addition.}
\end{table}

%============================================================================
\section{Thirteen Variables, One Cluster Label}

Eight geochemical ratios per rock already exceeds what a reader can compare
by eye across 224 samples, so the authors added five geophysical variables
-- crustal thickness and extension rate named specifically among them -- for
13 standardized variables total, and ran $k$-means clustering: an iterative
algorithm that assigns each data point to whichever group centre it is
closest to, then recomputes the centres, until the assignments stop
changing \cite{watts2025}. The output is a small number of discrete cluster
labels rather than a continuous composition map -- a rock is ``cluster 1''
or ``cluster 3,'' each cluster defined jointly by all 13 inputs. A cluster
label built this way behaves like a compact chemical fingerprint: sharper
than any one isotope ratio taken alone, and specific enough to look for
again somewhere else on the map.

%============================================================================
\section{No Bullseye: Repeating Clusters Instead}

The observed pattern was not radial decay. The same small set of cluster
identities recurred, disappeared and reappeared moving along a rift arm.
The Main Ethiopian Rift resolved into four clusters recurring at
50--100~km spacing; the Red Sea Rift resolved into three clusters recurring
at 150--200~km spacing \cite{watts2025}. Neither spacing matches a single
radially symmetric source, and the two rift arms do not even agree with
each other on the spacing -- which a shared plume centre obeying one
universal gradient also fails to predict. The authors' own description:
Afar is fed by ``a spatially and chemically heterogeneous upwelling''
\cite{watts2025}, more specifically ``a heterogeneous pulsing mantle
upwelling that is not symmetrical but is instead shaped by both variable
lithospheric thinning and extension rates within each rift''
\cite{watts2025}.

The detail elevating this from noise to structure: the three-cluster
sequence spanning the whole Red Sea Rift matches the initial roughly
200~km of the clustering sequence observed across the Main Ethiopian Rift
\cite{watts2025} -- the same signature, read out on two rift arms moving at
two different speeds, not three unrelated local phenomena. The abstract's
own phrase for the underlying source is a ``single, asymmetric upwelling''
\cite{watts2025}.

\begin{table}[htbp]
\centering
\begin{tabular}{@{}lcccc@{}}
\toprule
\textbf{Rift arm} & \textbf{Extension rate} & \textbf{Crustal thickness} & \textbf{Cluster spacing} & \textbf{Clusters} \\
\midrule
Main Ethiopian Rift & $\sim$5.2~mm/yr & 25--33~km & 50--100~km & 4 \\
Red Sea Rift & 10.5--19.5~mm/yr & 16--25~km & 150--200~km & 3 \\
\bottomrule
\end{tabular}
\caption{Reported values for the two rift arms compared directly in the
clustering \cite{watts2025}.}
\end{table}

%============================================================================
\section{The Club's Own Arithmetic: From Kilometres to Megayears}

Every calculation in this section is ours, built only from numbers the paper
reports, and is not itself a claim the authors make. The rock recording a
given cluster erupted at some point and has since been carried away from the
rift axis by the same plate motion that stretches the rift, so a present-day
spatial spacing $\lambda$ can, in principle, be converted back into an
implied time between pulses $T$ using the local extension rate $R$:
\begin{equation}
T \;=\; \frac{\lambda}{R}.
\end{equation}
The units collapse conveniently: a spacing in kilometres divided by a rate
in millimetres per year returns a period directly in millions of years,
because one million years of one-millimetre-per-year motion is exactly one
kilometre.

Substituting the Main Ethiopian Rift's own numbers ($\lambda =
50$--$100$~km, $R \approx 5.2$~mm/yr) gives an implied period of
$50/5.2 \approx 9.6$ to $100/5.2 \approx 19.2$~million years. Substituting
the Red Sea Rift's numbers ($\lambda = 150$--$200$~km, $R =
10.5$--$19.5$~mm/yr) gives an implied period ranging from
$150/19.5 \approx 7.7$ to $200/10.5 \approx 19.0$~million years, depending
which end of the reported extension-rate range is used. The two ranges
overlap almost completely, in a window of roughly 9.6--19.0~million years,
despite the raw spacings differing by a factor of two to four. That overlap
is what a single upwelling pulsing at a broadly consistent underlying rate
would produce if each rift arm simply stretches the same pulse train at its
own local speed: fast, thin-crusted rifts smear the record into a long
wavelength; slower, thicker-crusted rifts compress it into a short one.

We want to be precise about the status of this calculation: the paper
reports the spatial cluster spacings and the extension rates; it does not,
as far as we found, publish an explicit period in years between pulses. The
overlap above is our own consistency check on the paper's spatial numbers,
not a result stated in the paper itself, and it assumes the modern
extension rate has held roughly steady for the tens of millions of years
implied -- an assumption real, evolving rifts are not obliged to honour.

\begin{figure}[htbp]
\centering
\begin{tikzpicture}
\begin{axis}[
  width=12cm, height=5.5cm,
  xlabel={implied time between pulses (Myr) -- club's arithmetic},
  xmin=0, xmax=22,
  ytick={1,2}, yticklabels={RSR, MER},
  y dir=reverse,
  ymin=0.5, ymax=2.5,
  axis lines=left,
  xmajorgrids=true,
  grid style={afarneutral!25},
]
\addplot[afargold, fill=afargold, fill opacity=0.85, bar width=14pt, xbar interval, mark=none]
  coordinates {(9.6,1.75) (19.2,2.25)};
\addplot[afarteal, fill=afarteal, fill opacity=0.85, bar width=14pt, xbar interval, mark=none]
  coordinates {(7.7,0.75) (19.0,1.25)};
\draw[afaraccent, dashed, line width=1pt] (axis cs:9.6,0.5) -- (axis cs:9.6,2.5);
\draw[afaraccent, dashed, line width=1pt] (axis cs:19.0,0.5) -- (axis cs:19.0,2.5);
\end{axis}
\end{tikzpicture}
\caption{The club's conversion of each rift's cluster spacing and extension
rate into an implied period, $T=\lambda/R$. MER (gold): 9.6--19.2~Myr. RSR
(teal): 7.7--19.0~Myr. Dashed lines mark the shared window, roughly
9.6--19.0~Myr. Not a figure from the paper.}
\end{figure}

%============================================================================
\section{The Plate Decides How the Signal Gets Laid Down}

Why would the same pulsing source produce a tighter spatial pattern under
one rift arm than another? The paper's proposed mechanism is mechanical: the
Main Ethiopian Rift retains thicker continental crust, 25--33~km against the
Red Sea Rift's 16--25~km, and extends more slowly, $\sim$5.2~mm/yr against
10.5--19.5~mm/yr \cite{watts2025}. A thicker, slower-moving lid gives
upwelling mantle less room and less time to flow sideways before freezing
into the plate or erupting, compressing the same underlying heterogeneity
into a shorter spatial footprint. In the authors' words, progressive
lithospheric thickening away from the upwelling centre ``should reduce the
volume capacity for melt, impeding mantle flow,'' so that ``the
heterogeneous nature of the pulsed upwelling would exhibit a more condensed
spatial pattern within the [Main Ethiopian Rift] compared with [the Red Sea
Rift]'' \cite{watts2025}.

%============================================================================
\section{Flipping the Arrow}

Most rift-plume narratives run one direction: a plume rises, thins the
lithosphere from below, and the rift follows where the plume points, with
the crust above cast as a largely passive lid \cite{ebingersleep1998}. The
Afar result adds a second arrow running the other way: once a rift geometry
exists, its own thickness and spreading rate feed back and reshape how the
plume beneath it is permitted to flow. Watts et al.\ state the general claim
directly: ``the evolution of mantle upwellings is influenced by the
dynamics of overriding plates'' \cite{watts2025}. The Main Ethiopian Rift's
own documented history of episodic, segment-by-segment magmatic and
structural growth on the way to becoming today's triple junction
\cite{wolfenden2004} is consistent with a plume whose surface expression was
never going to be smooth to begin with.

%============================================================================
\section{What Would Change Our Mind}

The cross-rift match of the 1--2--3 cluster sequence is the load-bearing
evidence for one pulsing plume rather than three independently noisy rifts,
and it depends on a clustering algorithm's judgment call about how many
groups to look for; the paper reports four clusters for MER and three for
RSR, and we would want to see the match-up tested against a range of other
cluster counts before treating ``same sequence, different scale'' as
settled. Our own kilometre-to-megayear conversion is more exposed still,
resting on an assumption of roughly constant extension rate over tens of
millions of years that the paper itself does not make. A direct
geochronological date on rock from neighbouring clusters, rather than a
rate-based inference, is the test that would actually confirm or refute
``9.6 to 19.0 million years between pulses.''

%============================================================================
\section{Summary for a Reader in a Hurry}

``Pulsing plume'' here is inferred from a repeating chemical pattern frozen
into rock erupted over a long span, read out in space rather than watched
in time. The surprise is not that Afar's mantle is heterogeneous -- mixed
mantle sources are unremarkable -- but that the heterogeneity is a repeating
sequence rather than noise, and that the same sequence appears on two rift
arms at two different physical scales. And the reframing that earns this
paper a field note: the rift is not merely a stage the plume performs on.
Its thickness and its speed are, in a now chemically documented sense, part
of the plume's own instrument.

%============================================================================
\begin{thebibliography}{9}

\bibitem{watts2025}
Watts, E.~J., Rees, R., Jonathan, P., Keir, D., Taylor, R.~N., Siegburg, M.,
Chambers, E.~L., Pagli, C., Cooper, M.~J., Michalik, A., Milton, J.~A.,
Hincks, T.~K., Gebru, E.~F., Ayele, A., Abebe, B. \& Gernon, T.~M. (2025).
Mantle upwelling at Afar triple junction shaped by overriding plate
dynamics. \textit{Nature Geoscience} 18, 661--669.
doi:10.1038/s41561-025-01717-0

\bibitem{morgan1971}
Morgan, W.~J. (1971). Convection plumes in the lower mantle.
\textit{Nature} 230, 42--43. doi:10.1038/230042a0

\bibitem{ebingersleep1998}
Ebinger, C.~J. \& Sleep, N.~H. (1998). Cenozoic magmatism throughout east
Africa resulting from impact of a single plume. \textit{Nature} 395,
788--791. doi:10.1038/27417

\bibitem{zindlerhart1986}
Zindler, A. \& Hart, S. (1986). Chemical geodynamics.
\textit{Annual Review of Earth and Planetary Sciences} 14, 493--571.
doi:10.1146/annurev.ea.14.050186.002425

\bibitem{hansennyblade2013}
Hansen, S.~E. \& Nyblade, A.~A. (2013). The deep seismic structure of the
Ethiopia/Afar hotspot and the African superplume. \textit{Geophysical
Journal International} 194(1), 118--124. doi:10.1093/gji/ggt116

\bibitem{wolfenden2004}
Wolfenden, E., Ebinger, C., Yirgu, G., Deino, A. \& Ayalew, D. (2004).
Evolution of the northern Main Ethiopian rift: birth of a triple junction.
\textit{Earth and Planetary Science Letters} 224, 213--228.
doi:10.1016/j.epsl.2004.04.022

\end{thebibliography}

\end{document}