Geoscience Reference
In-Depth Information
87
Sr/
86
Sr in seawater with
mine the variation of
Table 16.1
Statistics of helium in MORB
[
statistical tests helium reser-
voirs
]
time.
The helium isotopic ratio in MORB is usually
quoted as 8
±
1
R
A
but this represents strongly
Mean
Standard deviation
n
filtered
data,
filtered
to
'remove
any
plume
influence'.
3.6
R
A
(Table 16.2). Spreading ridges average the isotopic
ratios from large volumes of the mantle and it
is fairly straightforward to compile meaningful
statistics. Unfortunately, the helium data on and
near spreading ridges is highly selected and
fil-
tered
prior to statistical analysis. Table 16.2 gives
both filtered and unfiltered estimates. Ocean
island basalt data is harder to analyze because
the samples are not collected in systematic or
random ways and anomalous regions are over-
sampled (the reverse of the situation for MORB
and ridges). Nevertheless, one can compile aver-
ages of all samples for each island and average
these to get global estimates of OIB statistics.
The
statistics and the distribution
of helium
in mantle magmas show that there
is no statistical difference in
3
He/
4
He (
R
) ratios
between the available data for midoceanic ridge
basalts (MORBs) and ocean island basalts (OIBs).
It is usually assumed, however, that these two
classes of basalts require the existence of at
least two distinct mantle reservoirs that have
been preserved over long periods of the Earth's
history. The distribution in ocean ridge basalts
(Figure 16.1) is more Gaussian than OIB. The
variance is relatively low but, importantly, is a
large fraction of the mean. OIBs have a non-
Gaussian distribution, with a very large variance.
The largest
R
values in the OIB population are
often considered to be diagnostic of the OIB
reservoir and large values in MORB are consid-
ered to be plume contamination. There is no
obvious cut-off between MORB values and OIB
values. Model calculations suggest that the sepa-
ration between the low
R
ratio components and
the high ratios found in some OIB-type compo-
nents is relatively recent (Anderson, 1998a, b;
Seta
et al.
, 2001). The variance differences, and
differences in extreme values, between various
basalt populations is a consequence of the
aver-
aging during the sampling process
. The extreme
3
He/
4
He ratios are averaged out
upon melt-
ing and averaging (SUMA)
, as at the global
Unfiltered
data
gives
9.1
±
All ridge
9.14
3.59
503
Atlantic
9.58
2.94
236
Pacific
8.13
0.98
245
Indian
8.49
1.62
177
OIB
7.67
3.68
23
flow and
4
He flux from the mantle implies that
He is trapped in the upper mantle; it is not escap-
ing as fast as it is produced. Apparently,
4
He is
accumulating in the mantle. Helium is more sol-
uble in magma than the heavier noble gases and
may be trapped in residual melts and cumulates.
Helium trapped in residual melts or fertile man-
tlewillevolvetolow
R
because of the presence
of U and Th. Helium trapped in olivine, olivine-
rich cumulates or depleted restites will main-
tain nearly its original isotopic ratio. Thus, high
3
He/
4
He samples may retain an ancient
frozen-in
ratio rather than a current
primordial
ratio. Such
material is probably intrinsic to the upper mantle
and upper mantle processes, such as degassing of
ascending magmas.
The
3
He
reference
isotope
for
helium
is
4
He
∗
) is usually writ-
ten
3
He/
4
He or
R
and this in turn is refer-
enced to the current atmospheric ratio
R
a (or
R
A
), and written
R
/
R
a (or
R
/
R
A
). The atmospheric
value involves degassing from the mantle and
crust --- which have quite different values and
time scales --- a contribution from interplane-
tary dust particles (IDP), and the rate of escape
from the atmosphere. These all change with time
and the atmospheric ratio is also expected to
change with time. There is no guarantee that
ancient magmas, corrected for radioactive decay,
had the same relation to the atmospheric values
atthetimeastheyhavetopresent-dayatmo-
spheric ratios. It is always assumed in mantle
geochemistry calculations that the atmospheric
ratio is invariant with time and that the atmo-
sphere is a well-stirred homogenous reservoir.
It would be useful to be able to measure the
time
3
He/(
4
He
and
the
ratio
+
variability,
just
as
it
is
useful
to
deter-
