Geoscience Reference
In-Depth Information
primary carbonates within pedogenic calcretes can be
used in the reconstruction of palaeoclimates, past veg-
etation and CO
2
concentrations (Cerling, 1999; Alonso-
Zarza, 2003). Both
Golding, 1998), although the pH range under which sili-
cification occurs remains disputed (Summerfield, 1979,
1983a; Thiry and Milnes, 1991; Terry and Evans, 1994).
Groundwater, drainage-line and pan/lacustrine silcrete
formation is strongly influenced by palaeotopography,
which controls both the watertable position and ground-
water flow; silicification under both high pH (Summer-
field, 1982; Meyer and Pena dos Reis, 1985; Nash and
Shaw, 1998) and low pH (Wopfner, 1978, 1983; Meyer
and Pena dos Reis, 1985; Taylor and Ruxton, 1987; Thiry,
Ayrault and Grisoni, 1988; Thiry and Milnes, 1991) con-
ditions have been suggested. Recent analyses of an Aus-
tralian groundwater silcrete by Alexandre
et al.
(2004)
also suggest that silicification may have occurred under
cooler and wetter conditions than at present.
Siliceous duricrusts are difficult to date, although the
development of techniques for dating diagenetic events
(McNaughton, Rasmussen and Fletcher, 1999) and K-Mn
oxides in weathered profiles (Vasconcelos, 1999; Vas-
concelos and Conroy, 2003) offers scope for optimism.
Radtke and Br uckner (1991) used electron spin resonance
to estimate the ages of Australian silcretes; however, their
use of bulk samples rendered the derived dates almost
meaningless. Most studies use relative dating approaches,
where the ages of silcretes are determined from their strati-
graphic position. This concept has been employed widely
in Australia where the presence of plant fossils and the re-
lationship between silcretes and basalts of known age have
been used to age-constrain outcrops (Webb and Golding,
1998; Taylor and Eggleton, 2001). Most pedogenic sil-
cretes date from the Palaeogene or Early Neogene (Thiry,
1999), although some have been attributed to the Meso-
zoic (Langford-Smith, 1978; Wopfner, 1978; Ballesteros,
Taleg on and Hernandez, 1997), Late Neogene and Qua-
ternary (Dubroeucq and Thiry, 1994; Curlık and Forgac,
1996). Australian groundwater silcretes have mostly been
ascribed a mid-Tertiary age (Stephens, 1971; Alley, 1973;
Young, 1978; Ambrose and Flint, 1981).
Rock varnish may coat surfaces within a few decades
(Engel and Sharp, 1958; Dorn and Meek, 1995), but is
more likely to take hundreds or thousands of years to
form (Elvidge, 1982; Dorn and DeNiro, 1985). Hence,
some varnishes may provide extremely long palaeoenvi-
ronmental records (see also Chapter 3). A variety of var-
nish properties has been used to infer past environments,
including analyses of micromorphological changes (Dorn,
1986), stable carbon isotope ratios (Dorn and DeNiro,
1985), organic carbon ratios (Dorn
et al.
, 2000), trace ele-
ment geochemistry (Fleisher
et al.
, 1999),
17
O excesses in
sulfates (Bao, Thiemens and Heine, 2001) and interlayer-
ing with other rock coatings (Dragovich, 1986). Contrary
18
O levels are dependent
upon the depth beneath the soil surface at which the sam-
ples are obtained, decreasing rapidly to become almost
constant at 10-50 cm depth (Quade, Cerling and Bow-
man, 1989). The oxygen isotope composition of a calcrete
is directly related to that of the meteoric water from which
it formed. In arid zones (with
13
C and
δ
δ
<
250 mm annual rainfall),
18
O lower than
values of
do not occur and ar-
eas receiving less than 350 mm have
δ
−
5
18
O values greater
δ
13
Cval-
ues of soil carbonates at depths below 30 cm depend on
the isotopic composition of the soil CO
2
, itself controlled
by the amount of atmospheric CO
2
that penetrates the
air-soil interface, the density of vegetation cover and the
proportions of plants present that use the C
3
,C
4
and CAM
photosynthetic pathways (Cerling, 1999). Low
than -2
(Talma and Netterberg, 1983). The
δ
13
Cval-
ues are normally taken to indicate the dominance of C
3
plants while heavier values suggest greater proportions of
C
4
and CAM plant communities.
The reliable dating of calcretes is essential if they are
to be used as an effective palaeoenvironmental indicator.
The earliest age estimates for calcrete were based on rela-
tive dating. However, the possibility of radiocarbon dating
calcretes (Williams and Polach, 1971; Magaritz, Kaufman
and Yaalon, 1981) led to their widespread use in palaeoen-
vironmental studies. Great care is needed to ensure that
only one generation of carbonate is represented in the
dated sample (Rust, Schmidt and Dietz, 1984) and the
appropriate correction factors are applied to take into ac-
count natural variations in stable isotope levels (Salomons
and Mook, 1976; Salomons, Goudie and Mook, 1978).
More recent studies have used U/Th dating to establish
the age of carbonates. This technique is prone to the same
sampling issues as radiocarbon dating (Ku
et al.
, 1979;
Schlesinger, 1985), but many problems can be overcome
through microsampling of carbonate phases. For exam-
ple, a suite of Middle Pleistocene to Holocene pedogenic
calcretes in the Sorbas basin, southeast Spain, have been
dated using this approach (Kelly, Black and Rowan, 2000;
Candy, Black and Sellwood, 2004a, 2004b, 2005).
The use of silcrete as a palaeoenvironmental indicator
has been the subject of considerable debate (see Nash
and Ullyott, 2007). However, recent advances in our un-
derstanding of silcrete genesis provide a starting point
for reevaluating palaeoenvironmental signals. Pedogenic
silcrete is now thought to form mainly under climates
with alternating wet/humid and dry/arid seasons or pe-
riods (e.g. Thiry, 1981; Callen, 1983; Meyer and Pena
δ
