In this study, we reexamined whether oceanic
hotspots are hotter than ridges using com-
plete ridge segment (N=424)( 20 ) and oceanic
hotspot (N= 46) ( 5 , 12 ) catalogs (data S1 and
Fig. 1A). We inferred temperature directly
from seismic tomography in the upper mantle
by converting velocity to temperature, sim-
ilarly to Daltonet al.( 19 ) for ridges. Daltonet al.
( 19 ) converted velocity to temperature at 300 km
depth below ridges, but not hotspots, and
Jacksonet al.( 12 ) did not convert to temper-
atures and focused only on hotspots at 200 km
depth. Examining both hotspots and ridges
simultaneously is key to both computing ex-
cess temperatures and comparison with pre-
vious studies using petrological thermometers
( 14 ). We choose to computeTexwith respect
to the ridgeTpin keeping with previous work( 14 , 15 ) and understanding of MORB forma-
tion ( 21 ). For both ridges and hotspots, we
inferredTpin 20-km intervals from 260 to
600 km depth, below the depths with the high-
est seismic attenuation ( 21 ). Hereafter we re-
fer to potential temperature as temperature.
We converted velocity anomalies from the glob-
al shear wave tomography model SEMUCB-
WM1 ( 22 ) to temperature using HeFESTo ( 23 ),58 7 JANUARY 2022¥VOL 375 ISSUE 6576 science.orgSCIENCE
Cold Hotspots
Warm Hotspots
Hot Hotspots
Plume-fed Hotspots
Mid-Ocean RidgesA1123 45
8910101515
1919
2121252526263030313133333535
37373939444423456711111212141416161818(^23232222)
2424
2828
(^32323636)
3838
4040
4343 4646
6
7
8
9
10
11
12
1313
1717
2020
29292727
3434
4141
4242
4545
13
14
15
16
17
19 18
20
21
(^2322)
24
25
26
27
28
29
30
31
32
33
34
35
36
37 38
39
40
41
42
43
44
45
46
Search radius R Distance A
hotspot mid-ocean ridge
Depleted upper mantle
plume
B
Fig. 1. Map of global hotspots and mid-ocean ridges and schematic of
velocity extraction.(A) Red, green, and blue squares represent hot,
warm, and cold oceanic hotspots, respectively. Squares with black outlines
have resolved plumes ( 6 ). The numbers correspond to the hotspot IDs
in data S1. Orange triangles show mid-ocean ridges. The−1%dlnVScontour
at the CMB ( 22 ) represents LLSVPs in white dashed lines. (B) Sketch
illustrating our search strategy. RadiusRdecreases with depth to
account for the mantle’s decreasing volume. DistanceAdefines the
ridge-hotspot proximity threshold, used in ( 21 ). We useA=1000 km
andR=500 km ( 21 ).
BallenyArago
MarquesasMacdonaldJan Mayen
IcelandSocieties
Rarotonga
Lord Howe/Tasman East
Juan de Fuca/Cobb (Axial Smt.)
PitcairnSamoaBowie
Socorro/Revillagigedo
Galapagos
Tasmanid/Tasman Central
Hawa
ii
Comores
Tristan/Gough/Walvis Ridge
Juan Fernandez
Easter
Baja/Guadalupe
MadeiraLouisvilleCanary
Amsterdam and St. Paul Islands
ReunionCarolineSt HelenaSan FelixFernandoBermuda
Meteor/Shona
VemaBouvetAzores
Kerguelen/Heard
Marion/Prince Edward
Discovery
Martin/Martin Vas/Trindade
Cape Verde
Crozet
Cameroon
Manus Basin
Great Meteor/New England
Ascension
This Study Ridges
Previous Studies Ridges
1300
1400
1500
1600
1700
1800
Potential Temperature (°C)
A
Ref
1 σ
2 σ
3 σ
SEMUCBTex= 139 °C
Hot Hotspots
Warm Hotspots
Cold Hotspots
Ridges
Putirka Hotspots/Ridges Mean
Courtier Hotspots/Ridges
Hot WarmColdRidges
1300
1400
1500
1600
1700
1800
Potential Temperature (°C)
B
Ref
199
1 σ
2 σ
3 σ
104 -10 Tex (°C)
Hotspots
Fig. 2. Potential temperature of oceanic hotspots and ridges.Violin plots
of the distribution of potential temperature (Tp). (A) On the right side, the 1377°C
reference adiabat ( 21 ) is marked by a black tick. Three yellow ticks indicateTp+
1 s,2s, and 3sfor ridges. The width of the violin represents the probability
density of temperature estimates from all depths (260 to 600 km). In every
violin, horizontal bars, from top to bottom, are maximum, mean, and minimum
values. Black outlines are plume-fed hotspots. From left to right, hotspots
are arranged in order of decreasingTp. Red, green, and blue violins are hot,
warm, and cold hotspots, respectively, classified by cluster analysis, with ticks on
the right showing theirTex. The blue shaded region (Tex<2s) covers plumes that
cannot actively upwell, green (2s<Tex<3s) covers those that barely can.
Olivine thermometry estimates from Courtieret al.( 4 ) for hotspots and ridges in
purple [average and 1s, see ( 21 )] and Putirka’s( 14 ) in solid red circles. Ridge
estimates from this study and Daltonet al.( 19 ) (with purple outlines) shown by
yellow violins on the right. (B) Violin stacks for three clusters of hotspots
compared with ridges (yellow).Texfor each group shown on top. White circles are
the median values, black and white bars away from the center indicate the
central 68th, 95th, and 99.7th percentiles.
RESEARCH | REPORTS