Geology Reference
In-Depth Information
Fig. 5.11 The mechanics of the EI correction technique.
The blue curve is the TK03 GAD fi eld model used in the
correction. It predicts north-south elongated directional
distributions as a function of inclination. The red curve is
the directional distribution elongation of a particular set of
directions that are being corrected. The directions have an
east-west elongation at f = 1 (uncorrected) and an
inclination of about 35° in this example. As the directions
are 'unshallowed', the elongation decreases in an east-west
sense according to tan I c = f tan I 0 (dashed curve), reaches a
minimum (nearly circular) at about 42° and f = 0.6 in this
example, and then becomes more elongated in a north-
south sense (solid curve) until at a particular elongation-
inclination pair it crosses the TK03 GAD fi eld model curve
(green circle). The intersection is the corrected inclination.
In the application of the technique many thousands of
corrections are made with bootstrapped data based on the
initial directional distribution of directions. This creates
many crossing points with the fi eld model. The crossing
points give a quantitative measure of the mean crossing
point and the 95% confi dence interval around the mean.
(See Colour Plate 9)
Fig. 5.10 LIP data in support of the TK03 fi eld model used
for the EI inclination correction. Red line is the TK03 model,
dashed lines are from Constable & Parker's (1988) fi eld
model while the dotted line is Quidelleur & Courtillot's
(1996) fi eld model. The Parana LIP is Cretaceous (about
120 Ma) and was considered problematic by Tauxe et al .
(2008). Keweenawan is 1.1 Ga and is from the Northshore
basalts of Lake Superior (Tauxe & Kodama 2009), E is the
30 Ma Ethiopian traps of Rochette et al . (1988) . Plot
modifi ed from Tauxe et al . (2008). Reprinted from Physics of
the Earth and Planetary Interiors , 169, L Tauxe, KP Kodama
and DV Kent, Testing corrections for paleomagnetic
inclination error in sedimentary rocks: A comparative
approach, 152-165, copyright 2008, with permission from
Elsevier. (See Colour Plate 8)
fi eld model, the corrected inclination and fl attening
factor are determined by the intersection of the two
curves (Fig. 5.11). Tauxe & Kent (2004) also use boot-
strap statistics to determine confi dence limits for the
corrected inclination.
In order to obtain a robust correction, many more
site means are needed than are typically collected in a
paleomagnetic study (> 100 - 150; Tauxe et al . 2008 ),
but standard paleomagnetic measurement and demag-
netization techniques are all that are required to reduce
the data. For this reason, the best EI corrections come
from analysis of magnetostratigraphic studies that
typically collect many sites. The EI technique was dem-
onstrated, in Tauxe's (2005) introduction of the EI
technique, with a correction of central Asian Oligo-
Miocene red beds collected near Subei, China that were
sampled by Gilder et al . (2001) for a magnetostrati-
graphic study. These rocks showed anomalously low
inclinations. The study had 222 sites and the inclina-
tion was corrected from 43.7° to 63° ( f = 0.4). Tauxe
(2005) also tested the technique with 105 site means
from the Siwalik Group of Pakistan (Tauxe & Opdyke
1982) and observed a fl attening factor of f = 0.77 for
the correction ( I 0 = 33.7, I c = 41; Table 5.2 ).
One important limitation of the EI technique that is
not often recognized is that it is assumed that each site
is an independent 'spot' measure in time of geomag-
netic fi eld secular variation. Because each sedimentary
rock sample will integrate its recording of the fi eld over
the time of its deposition, paleomagnetic samples of
fi nite thickness will average some secular variation.
Furthermore, unless all the samples from a paleomag-
netic site are collected from exactly the same horizon,
Search WWH ::




Custom Search