Global correlations of seismic elastic and anelastic tomography and
mantle temperatures
Experimental data on temperature dependence of Q, Vp, and Vs.
  • Purple dots and purple shading [J]: Q in dunite at P=300 MPa, f=0.01-0.33 Hz (Jackson et al.
    1992);
  • blue crosses and shading [S]: Q  in dry peridotite at P=0.20, 0.48, and 0.73 GPa measured at
    ultrasonic frequencies (Sato et al. 1989);  
  • green triangles and shading [BK, KB]: Q in dunite at f=0.1 Hz and 10 Hz (Berckhemer et al.
    1982; Kampfmann & Berckhemer 1985);
  • light blue rhombs [T]: Q in olivine aggregate at f=0.09 Hz (Tan et al. 1997).
  • Red symbols: seismic velocities  in peridotite.
  • Stars [K]: Vp  at P=1 kbar and 6 kbar (Kern 1978);
  • rectangles [MK]: Vp  at P=5 kbar and 10 kbar (Murase & Kushiro 1979);
  • crosses [MF]: Vs  at P=10 kbar (Murase & Fukuyama 1980).
  • Pink shaded area: the region where volume fraction of melt in peridotite increases from 0% to
    16% at P=10 kbar (Murase & Fukuyama 1980).
At high temperatures, the effect of  T on Qs and
Vs is different: a sharp increase in Qs occurs at
lower temperature than a drop in Vs associated
with partial melting
.
Qualitative comparison of Vs, Qs , and T anomalies in the upper mantle
Quantitative comparison of Vs, Qs , and T anomalies in the upper mantle
Correlation coefficients between seismic velocity,
attenuation, temperature, and surface heat flow at
different depths in the continental upper mantle
Parameters   
z=50 km
z=100 km
z=150 km
Qs vs T  (*)
-0.22
-0.33
-0.47
ln Qs vs 1/T  (*)
0.27
0.31
0.36
Vs vs T  (*)  
-0.31
.56
-0.56
Vs vs Qs  (*)  
0.38
0.66
0.41
Qs vs surface heat flow (*)
-0.19
-0.18
-0.04
Vs vs surface heat flow (*)
-0.26
-0.39
-0.36
       
Qs vs T  ("black & white")   
-0.24
-0.24
-0.25
Vs vs T  ("black & white")  
-0.20
-0.36
-0.26
Vs vs Qs  ("black & white")   
0.23
0.48
0.28
       
Qs vs T  (all data)    
-0.19
-0.23
-0.34
Vs vs T  (all data)   
-0.19
-0.38
-0.42
Vs vs Qs  (all data)
0.38
0.41
0.41
(*) At each depth, values within the band around the average value (zero
anomaly) approximately corresponding to white areas in Figs 2-4, are
excluded from the analysis. The width of the band is 10% of the total range
of parameter variations. The correlations for Vs–Qs  and Qs–T at z=50 and
150 km and for Vs–T at z=50 km should be treated with caution due to low
resolution of Vs and Qs models at these depths.

For a comparison, we calculated the correlation between the signs of the
anomalies only (
"black & white" models) and excluding the 10%-wide
band around the average value from the statistical analysis. Lower
correlations indicate that the amplitudes of the anomalies are important in
global correlations.

Correlation coefficients calculated when
all data were included into the
analysis are presented for a comparison as well.
Last modified December, 2004; irina@swave.wr.usgs.gov
Comparison of Qs, Vs, and T at a depths of 50,
100, and 150 km along the latitude 50o N in
North America and the latitude 60o N in Eurasia.
The values of Vs, Qs, and T are averaged over a
10o band, centered correspondingly on 50o N or
on 60o N. Large tics of the y-axis correspond to
4% anomaly for Vs, 40% anomaly for Qs, and
-20% anomaly for T. Qs and Vs anomalies are
shown as variation with respect to PREM. T
anomaly is shown with respect to a typical
continental geotherm of 55-60 mW/m2 (i.e.,
~750o C, ~1000o C, and ~1250o C at three
depths, correspondingly). An accuracy in the
amplitudes of T model is <10%, so that the signs
of the anomalies are the robust features. The
lateral resolution of thermal and seismic models
is different (ca. 20o for Qs and Vs anomalies and
ca. 10o for T anomalies), such that their peaks
may not necessarily coincide. The resolution of
seismic data is the lowest at 50 km depth due to
crustal effects. On the whole, there is a good
general agreement between Vs and T anomalies.
However, at 100 km depth where the resolution
of the attenuation model is the highest, Qs and T
anomalies in Eurasia also show an excellent  
agreement
Correlations between shear-wave velocity Vs (as variation in %
relative to PREM), inverse attenuation Qs (as variation in %
relative to PREM), and temperature T (in deg C) for continental
lithosphere at a depth of 100 km. Vs and Qs values are for the sites
of thermal modeling.

Red circles – all data for the continents; blue circles – data for
Archean – early Proterozoic cratons only.

The calculated correlation coefficients are less than 0.42 for any
pair of the parameters. The correlation is especially bad for the
Archean-early Proterozoic cratons, suggesting that a significant part
of lateral Vs and Qs variations in cratonic lithosphere has a non-
thermal origin.

Artemieva I.M., Billien M., Lévêque J.-J., and Mooney W.
Geophysical Journal International, 2004, v.157, 607-628.
that for the Tibetan Plateau and the Andes, where the crustal thickness exceeds 60-70 km, the anomalies refer
to the crustal depth), 100±25 km, and 150±25 km.

We visually compare maps of seismic and temperature anomalies in the upper mantle to examine if there is a
qualitative regional correlation between the parameters. Since at 50 km depth the crustal structure has the
largest influence on the Vs and Qs values, we place greater emphasis on the maps for 100 and 150 km depth,
with particular focus on the results for 100 km depth, where the resolution of both of the seismic models is
the highest.

A comparison of the three maps reveals the existence of a strong overall qualitative correlation between the
signs of Vs, Qs, and T anomalies for most of the continental lithosphere at depths of 50-150 km. There are
two types of regions where all three anomalies correlate: “cold” regions, which include most of the
Precambrian cratons, and “hot” regions such as continental rifts, Cenozoic orogens and, unexpectedly, the
Sino-Korean craton. The signs of all three anomalies correlate within both of the groups throughout the entire
lithosphere, which suggests a primarily thermal origin of most of the seismic anomalies.
    THE CONTINENTAL LITHOSPHERE
Lithosphere research: Current Projects
Thermal model,2001
    THE CONTINENTAL LITHOSPHERE