Content uploaded by Jay Quade
Author content
All content in this area was uploaded by Jay Quade on Jan 13, 2017
Content may be subject to copyright.
3
Reviews in Mineralogy & Geochemistry
Vol. 66, pp. XXX-XXX, 2007
Copyright © Mineralogical Society of America
1529-6466/07/0066-0003$05.00 DOI: 10.2138/rmg.2007.66.3
Paleoelevation Reconstruction
using Pedogenic Carbonates
Jay Quade
Department of Geosciences
University of Arizona
Tucson, Arizona, 85721, U.S.A.
quadej@email.arizona.edu
Carmala Garzione
Department of Earth and Environmental Sciences
University of Rochester
Rochester, New York, 14627, U.S.A.
garzione@earth.rochester.edu
John Eiler
Division of Geological and Planetary Sciences
California Institute of Technology
Pasadena, California, 91125, U.S.A.
eiler@gps.caltech.edu
ABSTRACT
Paleoelevation reconstruction using stable isotopes, although a relatively new science,
is making a signi cant contribution to our understanding of the recent growth of the world’s
major orogens. In this review we examine the use of both light stable isotopes of oxygen and
the new “clumped-isotope” (Δ47) carbonate thermometer in carbonates from soils. Globally,
the oxygen isotopic composition (δ18O) of rainfall decreases on average by about 2.8 ‰/km
of elevation gain. This effect of elevation will in turn be archived in the δ18O value of soil
carbonates, and paleoelevation can be reconstructed, provided (1) temperature of formation
can be estimated, (2) the effects of evaporation are small, (3) the effects of climate change
can be accounted for, and (4) the isotopic composition of the carbonate is not diagenetically
altered. We review data from modern soils to evaluate some these issues and nd that
evaporation commonly elevates δ18O values of carbonates in deserts, an effect that would
lead to underestimates of paleoelevation. Some assessment of paleoaridity, using qualitative
indicators or carbon isotopes from soil carbonate, is therefore useful in evaluating the oxygen
isotope-based estimates of paleoelevation. Sampling from deep (> 50 cm) in paleosols helps
reduce the uncertainties arising from seasonal temperature uctuations and from evaporation.
The new “clumped-isotope” (Δ47) carbonate thermometer, expressed as Δ47, offers an in-
dependent and potentially very powerful approach to paleoelevation reconstruction. In contrast
to the use of δ18O values, nothing need be known about the isotopic composition of water from
which carbonate grew in order to estimate of temperature of carbonate formation from Δ47
values. Using assumed temperature lapse rates with elevation, paleoelevations can thereby be
reconstructed.
Case studies from the Andes and Tibet show how these methods can be used alone or in
combination to estimate paleoelevation. In both cases, the potential for diagenetic alteration
03_Quade_etal.indd 103_Quade_etal.indd 1 7/3/2007 12:41:09 AM7/3/2007 12:41:09 AM
2Quade, Garzione, Eiler
of primary carbonate values rst has to be assessed. Clear examples of both preservation and
alteration of primary isotopic values are available from deposits of varying ages and burial
histories. Δ47 values constitute a relatively straightforward test, since any temperature in
excess of reasonable surface temperatures points to diagenetic alteration. For δ18O values,
preservation of isotopic heterogeneity between different carbonate phases offers a check on
diagenesis. Results of these case studies show that one area of south-central Tibet attained
elevations comparable to today by the late Oligocene, whereas 2.7 ± 0.4 km of uplift occurred
in the Bolivian Altiplano during the late Miocene.
INTRODUCTION
There is considerable debate over the timing and causes of uplift of the earth’s great moun-
tain ranges and plateaux, such as Tibet, the Himalaya, and the Andes (
Fig. 1
Fig. 1). Estimates of
paleoelevation change provide key constraints for competing general models that link such
large-scale orogenic events with lithospheric-scale geodynamic processes (Currie et al. 2005;
Garzione et al. 2006; Rowley and Currie 2006; Kent-Corson et al. 2007). The use of the isotopic
composition of carbonates, silicates, and oxides is at the forefront of these paleoelevation recon-
structions. In this chapter we review use of one of several geologically abundant secondary car-
bonates—calcite formed in soils (or “pedogenic carbonate”)—in paleoelevation reconstruction.
The H and O isotopic composition of rainfall (expressed as δ18Omw and δDmw values in per
mil, or ‰, relative to SMOW) varies primarily as a function of the degree of rainout (removal
of water condensate) from a vapor mass. As rainout proceeds, 2H (D) and 18O are preferentially
removed from the vapor mass, decreasing δD and δ18O values of both the water vapor and
rainfall derived from it. As a vapor mass ascends over mountains, adiabatic expansion causes
cooling and condensation of water as so-called “orographic precipitation” (e.g., Roe 2005),
producing some of the largest isotopic gradients in rainfall observed on Earth. A number of
studies document large fractionations in the H and O isotopes in rainfall, snowfall, and associated
surface water with increasing elevation (e.g., Ambach et al. 1968; Siegenthaler and Oeschger
1980; Rozanski and Sonntag 1982, Ramesh and Sarin 1995; Garzione et al. 2000a; Gon antini
et al. 2001; Stern and Blisniuk 2002; Rowley and Garzione 2007; but see Dutton et al. 2005).
The δ18O-elevation relationship has been calibrated both empirically (e.g., Garzione et al.
2000a, Poage and Chamberlain 2001) and theoretically using basic thermodynamic principles
that govern isotopic fractionation of an ascending vapor mass (Rowley et al. 2001; Rowley
and Garzione 2007). Globally, the δ18Omw value of rainfall on average decreases by 2.8 ‰/km
(Poage and Chamberlain 2001), although the scatter around this average is very considerable
(Blisniuk and Stern 2005), as we discuss for the cases of Tibet and the Andes in the next section
of this review.
Studies using oxygen (and hydrogen) isotopes take advantage of this strong sensitivity
of δ18Omw values to elevation change to reconstruct paleoelevation. Stated simply, the basic
approach to paleoelevation reconstruction is this: the oxygen isotope composition of paleo-
carbonate buried in geologic sections can be used to estimate the oxygen isotope composition
of soil water, which in turn is used as an estimate of δ18Omw values; and it is this estimate of
δ18Omw values that is compared with empirical or theoretical isotopic lapse rates to estimate
paleoelevation.
Two key complications in this approach to paleoelevation reconstruction are that
carbonates, including pedogenic carbonate which forms in soils, record the oxygen isotopic
composition of soil waters (which may differ from meteoric water) and of the temperature of
carbonate formation, both of which vary with elevation. Early studies used the δ18O value of
pedogenic carbonates (or δ18Osc, expressed in ‰ relative to PDB) to reconstruct elevation by
assuming temperatures of formation that allowed calculation of the δ18O value of soil water
03_Quade_etal.indd 203_Quade_etal.indd 2 7/3/2007 12:41:24 AM7/3/2007 12:41:24 AM
Paleoelevation Reconstruction using Pedogenic Carbonates 3
75°W
75°W
70°W
70°W
65°W
65°W
25°S 25°S
20°S 20°S
15°S 15°S
0370185
Kilometers
0 m
5000 m
Elevation
Bolivia
Peru
Argentina
Chile
A
l
t
i
p
l
a
n
o
P
u
n
a
W
e
s t
e
r
n
C
o
r
d
i
l
l
e
r
a
Callapa
E
a
s
t
e
r
n
C
o
r
d
i
l
l
e
r
a
3UBANDES
!
Corque
Tambo
Tambillo
Paposa
Socompa
Calama/Paso
de Jama
Figure 1. (a) Elevations of the Central Andean plateau between 15°S and 26°S, constructed with SRTM30
dataset. White circles show our late Oligocene to late Miocene sites discussed in the text. White boxes
enclose study sites for modern soil carbonate in the Atacama Desert in Quade et al. 2007; (b) location gure
for study locations in Himalayan-Tibet orogen (Himalaya/Hindu Kush shaded).
Penbo
Dumri Fm.
China
Lhasa
Nepal
India Bangladesh
Bhutan
Siwalik Grp. Nima
90°E80°E
35°N
25°N
300
0kms
N
(a)
(b)
03_Quade_etal.indd 303_Quade_etal.indd 3 7/3/2007 12:41:25 AM7/3/2007 12:41:25 AM
4Quade, Garzione, Eiler
(or δ18Osw, expressed in ‰ relative to SMOW) (Garzione et al. 2000a,b; Rowley et al. 2001;
Blisniuk et al. 2005; Currie et al. 2005; Rowley and Currie 2006). The δ18O value of soil water
can differ from δ18Omw values due to evaporative enrichment of 18O in soil water (Cerling and
Quade 1993). Moreover, the season of carbonate precipitation may also affect the isotopic
composition of pedogenic carbonate, both in terms of seasonal variability in the temperature
of carbonate formation and the isotopic composition of rainfall and soil water (e.g., Quade et
al 1989a; Amundson et al. 1996; Liu et al. 1996; Stern et al. 1997).
These competing effects can be mitigated by the application of a new “clumped-isotope”
carbonate thermometer that allows the determination of the temperature of carbonate
formation (Ghosh et al. 2006a). This technique relies on the abundance of bonds between rare,
heavy isotopes (i.e., 13C-18O) in the carbonate mineral lattice, the relative abundance of which
increases at lower temperatures. Recent paleoelevation studies (Ghosh et al. 2006b) applied
this carbonate thermometer to better constrain δ18Osw values from which paleoelevation was
calculated using the traditional approach (Garzione et al. 2006).
We will begin this review by evaluating modern variability in δ18O values of surface
waters with elevation change across the Andean plateau (Fig. 1a) and the Himalayan-Tibet
plateau (Fig. 1b). Global patterns of the δ18Omw/elevation relationship are comprehensively
reviewed in Poage and Chamberlain (2001) and Blisniuk and Stern (2005). We will then
compare δ18Omw to δ18Osc results from the Mojave Desert, the Andes, and Tibet, among the
few regions where coupled water and carbonate isotopic data from modern soils are available,
providing an essential interpretive backdrop for the paleosol carbonate records. Finally, we
present case studies from the Andes and Himalaya-Tibet as examples of the use of δ18O and
“clumped-isotope” values from pedogenic carbonates in paleoelevation reconstruction.
DEPENDENCE OF THE ISOTOPIC COMPOSITION
OF RAINFALL ON ELEVATION
δ18Omw values decrease with elevation gain on both sides of the Andes and Tibet (
Fig. 2
Fig. 2a).
On the windward side, these changes re ect the expected effects of progressive distillation
on air masses as they rise, cool, and rain out over an orographic barrier. Interestingly, δ18Omw
values also decrease with elevation on the leeward sides of large mountain ranges, especially
in warm climates (summarized in Blisniuk and Stern 2005), including the western ank of the
central Andes, the Himalaya, and to some extent the Sierra Nevada’s. This is not predicted
by Rayleigh distillation models of cloud mass rain out (Rowley and Garzione 2007). Some
research suggests that in dry leeward settings, raindrops are evaporatively enriched in 18O
during descent, and a small fraction of the original raindrop reaches the ground (Beard and
Pruppacher 1971; Stewart 1975). Thus, evaporative gains in raindrop 18O apparently more than
offset Rayleigh-type depletions in 18O in the original cloud mass.
One striking example of this pattern is from the Andes, drawing mainly upon published
data between 17-21°S from Gon antini et al. (2001) and our unpublished water data for the
eastern side of the Andes, and from Fritz (1981) and Aravena et al. (1999) for the Paci c slope
(Fig. 2a). For the eastern slope, δ18O values decrease by ~2.6 ‰/km across an elevation gain
of 2200 m. These results match well with δ18O values of snow from Mt. Sajama (Hardy et
al. 2003), and rain falling on Lake Titicaca (Cross et al. 2001). On the Paci c slope, samples
collected from 22 stations over a four-year period show an even steeper isotopic lapse rate
of −6.2 ‰/km. We will return to these data when we look at isotopic results from modern
pedogenic carbonates in this region.
Stream-water sampling from Garzione et al. (2000a) along the Kali Gandaki up the south
face of the Himalaya at about 78°E shows a very regular decrease in δ18Omw values (Fig. 2b).
The best t to these data is a second-order polynomial, with an average decrease of around 2.5
03_Quade_etal.indd 403_Quade_etal.indd 4 7/3/2007 12:41:30 AM7/3/2007 12:41:30 AM
Paleoelevation Reconstruction using Pedogenic Carbonates 5
to 2.7 ‰/km. δ18Omw values decrease to below −20‰ for elevations above 5000 m, as sampled
near high-elevation glaciers (Zhang et al. 2002; Tian et al. 2005) and from streams owing from
the highest elevations on both sides of the Himalaya (Tian et al. 2001). The more unexpected
pattern is the increase in δ18Omw values north of the Himalayan crest, as observed by Tian et
al. (2001) and Zhang et al. (2002). The sampling stations are widely dispersed, and we can add
our own unpublished stream water samples from about 29-34°N. The combined data display an
increase of ~1.5 ‰/° latitude. Virtually all the collection stations range from 3700 m to 5000 m.
With our own stream-water samples, we show sampling elevations (Fig. 2b) unadjusted for the
more relevant mean elevation and latitude of the catchment. But since we sampled mostly small
catchments, we believe the correlation with Tian’s and Zhang’s data is not a coincidence.
The southwest Indian monsoon dominates rainfall south of the Himalayan front as well as
stations immediately north of the crest, perhaps mixed from some contribution from the East
Asian monsoon. The seasonal distribution of δ18Omw values for Lhasa, located 1.5° north of
Himalayan ridgecrest, illustrates this. Over ninety per cent of the rain falls in the summer rainy
season, and that rainfall displays much lower δ18Omw values than winter snow and rain (Araguás-
Araguás et al. 1998). The strong 18O and D depletions in summer rainfall can be modeled in
terms of gradual, Rayleigh distillation of moisture (Rowley et al. 2001) that derives water vapor
-25
-20
-15
-10
-5
0
Sajama
Titicaca
Gonfiantini
Aravena/Fritz
64
65
66
67
68
69
70
71
1
6
7
0
Titicaca/Uyuni
Pacific
Ocean
Cordillera
Occidental Cordillera
Oriental
W
E
°longitude
2
3
4
5
h = -472.5δ18O-2645
Atacama
Desert
(a)
3
6
9
0
y = 1.4x - 18.7
r2=0.50
-24
-22
-20
-18
-16
-14
-12
-10
-8
-6
-10123456789
°N of Himalayan crest
JQstreams
Tian et al.
Lhasa
Kali Gandaki
SN
Himalaya
(b)
Figure 2. Topography and δ18Omw (SMOW) values for (a) the Andes, compiled from Aravena et al.
(1999) and Fritz et al. (1981); Hardy et al. (2003) (Sajama); and Cross et al. (2001)(Titicaca); uncorrected
Gon antini et al. (2001); dashed line and equation are corrected Gon antini data combined with our
unpublished data (see also Fig. 14 and text); and (b) for Tibet, compiled from Garzione et al. 2000a (Kali
Gandaki); Tian et al. 2001; Quade unpublished data (JQ streams); and Araguás-Araguás 1998 (Lhasa).
03_Quade_etal.indd 503_Quade_etal.indd 5 7/3/2007 12:41:30 AM7/3/2007 12:41:30 AM
6Quade, Garzione, Eiler
from the equatorial Indian Ocean or East China Sea, is drawn northward and westward over
land in the summer, and rises, cools and rains out as it passes over the Himalaya and Tibet.
This simple picture does not explain the gradual increase in δ18Omw values north of the
Himalayan crest. Mean elevation of sampling sites north of the crest lies almost entirely within
4500 ± 500 m, and even increases slightly northward, the opposite of the expected pattern where
elevation is the main determinant of δ18Omw values. Likewise, the continentality effect—simple
rain out of air masses with distance from coastlines—should cause δ18Omw values to decrease
inland, not decrease.
The explanation provided by Tian et al. (2001) and supported by a variety of data, including
seasonality of rainfall and deuterium excess, is that other, either local or northerly derived air
masses penetrate southward during the winter months. The δ18Omw value of stream samples and
amount-weighted rainfall samples increase northwards, re ecting a decreasing proportion of
summer (monsoonal) rainfall in year-round weighted rainfall. The key point for paleoelevation
reconstruction is that δ18Omw values change quite signi cantly across the Tibetan Plateau for
climatic, not elevational reasons. The result is that lapse rates in the southern plateau (−2.6 ‰/
km) are almost twice those of the northern plateau (−1.5 ‰/km). Consequently, paleoelevation
estimates depend on distance of sampling sites from the Himalayan range crest, and how climate
change in the past, such as a strengthened or weakened Asian monsoon, might have increased or
decreased isotopic lapse rates northward on the plateau.
MODERN PEDOGENIC CARBONATES
Pedogenic carbonate formation
The general equation for weathering (to the right) and calcite precipitation (to the left) in
soils is:
CaCO3(s) + CO2 (g) + H2O = Ca++ (aq) + 2HCO3− (aq) (1)
Here we use calcite as the parent material but in its place just as easily could have used Ca-Mg
silicate minerals. Pedogenic carbonate formation (to the left) is driven by both soil water and
CO2 loss. Seasonality of carbonate formation is not known for certain, but in most settings it is
likely concentrated in the summer half year when soils are thawed or warmer, plants are active
and evapotranspiring, and evaporation is greatest.
In well-drained soil pro les, water for weathering reactions like (1) is supplied directly by
local rainfall. Pedogenic carbonate thus enjoys a key advantage over other types of carbonate
in paleoelevation reconstruction in that it forms from rainfall that fell on the site, and not
runoff from higher elevations, as in the case of many riverine and lacustrine carbonates. For
the oxygen system, molar water to rock ratios are extremely high in soils, and thus the δ18O
value of secondary weathering phases such as clays, iron oxides, and carbonates should not
be in uenced by parent material during weathering. Analogously, for the carbon system in
most soils, the molar Cplant/Cparent material is also very large, and hence plant CO2 mixed with the
atmospheric CO2, and not carbon from local parent material such as limestone, will determine
the δ13C value of pedogenic carbonate. In dry climates, soils dewater largely by a combination of
evapotranspiration (ET) and evaporation (E), with some percolation through to the local water
table. The ratio of dewatering by these processes (FE/FET) decreases with increasing soil depth
and increasing rainfall. For oxygen isotopes in soils, evaporation is a fractionating process,
enriching residual soil water in 18O. By contrast, evapotranspiration is a non-fractionating
process; water is drawn unfractionated into plant roots from the soil. Hence, in drier climates
and shallower in soils, evaporative enrichment in 18O has been widely observed (e.g., Allison
et al. 1983; Hsieh et al. 1998). After correction for temperature (discussed later), this effect
should also be expressed in secondary mineral phases, such as carbonates in soils.
03_Quade_etal.indd 603_Quade_etal.indd 6 7/3/2007 12:41:34 AM7/3/2007 12:41:34 AM
Paleoelevation Reconstruction using Pedogenic Carbonates 7
In the next section we compare δ18Osc values to local δ18Omw values from rainfall collected
sometime in the last few decades. To make this comparison, we use pedogenic carbonate
samples that formed recently, preferably over the latter part of the Holocene, on the assumption
that this is the carbonate most likely to represent isotopic equilibrium conditions between
calcite and the sampled water. In the absence of radiometric dates from our samples, we use
pedogenic carbonate morphology as a qualitative indicator of age, following the conventions
of Gile et al. (1966) (see also Machette 1985), in which thin carbonate coatings on alluvial
clasts represent Holocene-age cements, and continuous lling of the soil matrix by carbonates
represents older (104-106 yrs) Quaternary cementation.
The oxygen isotopic composition of Holocene pedogenic carbonate
Salomans et al. (1978), Talma and Netterberg (1983) and Cerling (1984) were among the
rst to show that δ18Omw values are positively correlated with the δ18Osc values, a relationship
later shown to hold true for a variety of meteoric cements (Hays and Grossman 1991). The
details of this relationship are worth exploring in light of newer data and from the perspective
of paleoelevation reconstruction. To do this we selected three data sets that encompass the
broad range of rainfall (~20 to 1000 mm/yr) in which pedogenic carbonates develop. These
include modern pedogenic carbonates from the relatively wet mid-western USA (Cerling and
Quade 1993), the Mojave Desert (Quade et al. 1989a), and from the extremely arid Atacama
Desert (Fig. 1a; Quade et al. 2007). We compare δ18Osc values from Holocene-age soils against
predicted δ18Osc based on local mean annual temperature (MAT) + 2 °C (see defense of this
formulation in the next section of the paper) and known δ18Omw values (
Fig. 3
Fig. 3). In higher rain-
fall areas, predicted and observed δ18Osc values are positively correlated, as observed by others
previously (i.e., Salomans et al 1978; Cerling 1984). The departures of observed δ18Osc values
from the 1:1 predicted line are larger where the climate is drier, consistent with evidence from
modern soil water studies (e.g., Hsieh et al. 1998). Hence, pedogenic carbonates from the driest
sites in the Mojave (MAP < 100 mm/yr) and in nearly all sites in the hyperarid Atacama Desert
-16.0
-12.0
-8.0
-4.0
0.0
4.0
-16.0 -12.0 -8.0 -4.0 0.0 4.0
δ
18
O (PDB) predicted
δ
18
O (PDB) observed
southern Nevada
Atacama Desert
moist climates
Death Valley
soils increasing evaporation
1:1
Figure 3. Observed δ18Osc (data from Quade et al. (1989a) and Cerling and Quade (1993), and Quade et al.
(2007)) versus predicted δ18Osc values based on local mean annual temperature and δ18Omw values.
03_Quade_etal.indd 703_Quade_etal.indd 7 7/3/2007 12:41:34 AM7/3/2007 12:41:34 AM
8Quade, Garzione, Eiler
(2-114 mm/yr) show the strongest
soil evaporation effects. Examina-
tion of δ18Osc values collected with
depth in soil pro les is also instruc-
tive. In dry settings like the Mojave
Desert, δ18Osc values increase to-
ward the surface (
Fig. 4
Fig. 4), whereas
from moister settings like Kansas,
δ18Osc values show little variation
with soil depth (Cerling and Quade
1993).
In the context of paleo-elevation
reconstruction, the following con-
clusions can be drawn: (1) in moist-
er climates, δ18Osc values should
provide a close approximation of
δ18Omw values after correction for
temperature, (2) in drier settings,
δ18Osc values will be in uenced by
evaporative enrichment, and hence
in this instance (3) sampling later-
ally and deeper (≥ 50 cm) in soils
should produce a better approxima-
tion of unevaporated δ18Omw values.
Because of evaporation effects,
δ18Osc values should be viewed to
provide only maximum estimates
of δ18Omw values and thus minimum
estimates of paleo-elevation.
Soil temperature considerations
One must be able to constrain
soil temperature (T) to account for
the fractionation between calcite
and water. In this regard, several
factors favor the use of pedogenic
carbonates. One is that the
fractionation factor between calcite
and water is T small (~ −0.22 to
−0.24 ‰/°C in the range 0-30 °C)
when compared, for example, to the
steep average dependence of δ18Omw
values on elevation of −2.8 ‰/km.
Thus, δ18Osc values are widely used
by geologists to constrain δ18Omw values, as in paleo-elevation reconstructions, but only used
for paleo-temperature reconstructions in a narrow set of circumstances where paleo-δ18Omw
values are relatively invariant. Having said this, the effects on δ18Osc values of decreasing
T with elevation can partially to entirely offset the in uence of decreases in δ18Omw values
with elevation. Temperature lapse rates globally are about 5-6°C/km. This temperature lapse
rate combined with a −2.8 ‰/km decrease in δ18Omw values would produce a ca. −1.5 ‰/km
change with elevation gain in δ18Osc values.
a. SM-3b
1900 masl
-140
-130
-120
-110
-100
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
-15 -13 -11
-9 -7 -5
b. SM-2b
1550 masl
-100
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
-14 -12 -10
-8 -6 -4
c. PAM-1
300 masl
-80
-70
-60
-50
-40
-30
-20
-10
0
-8 -6 -4 -2
0
δ
18
O(VPDB)
Figure 4. Depth pro les of δ18Osc (PDB) versus soil depth,
from three sites in the Mojave Desert, (a) SM-3 (1990 masl),
(b) SM-2b (1550 masl), and (c) PaM-1 (300 masl), all from
Quade et al. (1989a).
03_Quade_etal.indd 803_Quade_etal.indd 8 7/3/2007 12:41:35 AM7/3/2007 12:41:35 AM
Paleoelevation Reconstruction using Pedogenic Carbonates 9
One feature of soils that confers a signi cant advantage compared to other carbonates
(such as lacustrine carbonate) is that soil T converges toward mean annual air T at depth. In
contrast, shallow aquatic systems where many carbonates form can experience sharp diurnal
and seasonal changes. Soil T varies with time (t) and exponentially as function soil depth (z)
following (from Hillel 1982):
Tzt T A t zd e
avg
zd
(,) sin=+ −
()
⎡
⎣⎤
⎦−
oω
where Tavg is average air temperature, Ao is seasonal or diurnal maximum and minimum T
amplitude at z = 0, ω is radial frequency, z is soil depth, d is the “damping” depth (or 1/e folding
depth) characteristic of the soil:
dC
V
=
()
22
12
κω ()
where κ is thermal conductivity and CV is volumetric heat capacity. As regards surface-T uctu-
ations, the 1/e folding depth of diurnal uctuations in the order of centimeters, much shallower
than the average depth of carbonate accumulation. This is due mainly to the high diurnal radial
frequency (in which ω = 2π/86,400 seconds), which in Equation (2) is inversely related to the
damping depth. Seasonal temperature uctuations have a much lower radial frequency, hence
greater damping depth. Therefore, it is seasonal uctuations in T, and their attenuation at deep
soil depths (> −20 cm) where soil carbonate forms, that is the chief concern here.
Taking the hypothetical example of a Tibetan soil near Lhasa at 4100 m, we assume in the
following simulation 0.0007 cal/cm sec, and 0.3 cal/cm3 °C for κ and CV (from Hillel 1982),
respectively. This yields 1/e folding lengths for seasonal temperature change of plateau soils of
about 150 cm.
Average ground temperature differs by 1-3 °C from average air temperature due to excess
ground warming during the summer months (Bartlett et al. 2006). This offset is correlated with
mean radiation received at a site (1.21 K/100 W m−2), which is in turn is a function of vegetation
cover, slope aspect, and latitude. For our calculations we added a constant 2 ± 1 °C to average
air temperature to arrive at average ground temperature at depth.
Depth of pedogenic carbonate formation is typically between 50 and 150 cm on the Tibetan
plateau. At these high elevations the soil is at or below freezing half of the year. Thus we can
assume that pedogenic carbonate formation occurs during the summer half year, when soils
thaw or are warmer, evaporation is greatest, and when plants are actively transpiring soil water.
For our Tibetan soil site at 4100 m, monthly mean extremes in air temperature range between
11.6 and −6.4 °C, around a mean of about 2.6 °C. This yields a modeled seasonal T amplitude
of about 12.7 °C at 50 cm to 6.7 °C at 150 cm (
Fig. 5
Fig. 5). For the summer half-year, this produces
a range of 1.5 to 0.7‰ uncertainty in the estimation of paleo-δ18Omw from pedogenic carbonate
sampled from 50 to 150 cm soil depth, respectively (Fig. 5b). Here we use the fractionation
factor (α) between calcite and water of 1000lnαcalcite-water = (18030/T) − 32.42, where T is in
Kelvins (Kim and O’Neil 1997). We can conclude that deeper sampling is better for reducing
the error in paleo-elevation reconstruction arising from seasonal uctuations in temperature.
The contribution to the error can be no less than on the order ~0.8‰ for typical Tibetan soils; in
other words, about 300 m in paleoelevation terms.
Evaluation of aridity
Aridity is of interest in paleoelevation reconstruction for two reasons: (1) in extreme aridity
cases, the δ18Osc values are dominated by evaporation, and realistic estimates of paleoelevation
are probably not obtainable, no matter the sampling density, and (2) aridity develops in rain
shadows, and thus may provide qualitative evidence of orogenic blockage of moisture.
Evaluation of paleoaridity can (and should) be approached both qualitatively and quanti-
03_Quade_etal.indd 903_Quade_etal.indd 9 7/3/2007 12:41:36 AM7/3/2007 12:41:36 AM
10 Quade, Garzione, Eiler
tatively. The qualitative approach involves both soil morphology and carbonate leaching depth.
Rech et al (2006) provides one recent example from the central Atacama Desert of the utility
of paleosols in establishing the limits on paleo-aridity (Fig. 1a). They showed that hyperarid
(< 20 mm/yr) conditions have prevailed—based on the presence of buried nitrate and gypsic
paleosols—in the central Atacama Desert since at least 12 Ma. In this setting where evapora-
tion strongly distorts δ18Osc values, useful estimates of paleo-elevation are hard to obtain, as
we will argue in a later section of the paper. Rech et al (2006) go on to describe older calcare-
ous paleosols in the ~20 Ma range characterized by strong reddening, visible secondary clay
accumulations, and vertic fracturing of the paleosol. These are all features uncharacteristic of
calcareous paleosols of the current Atacama Desert, and a good match to modern soils now
found in moister central Chile where mean annual rainfall (MAP) is > 200 mm/yr. These kinds
of pedogenic carbonates are much better targets for paleoelevation reconstruction.
In addition to reddening and clay content (Birkeland 1984), depth of pedogenic carbonate
leaching is a useful qualitative index of paleoaridity. In modern soils, leaching depth increases
by 0.2 to 0.3 cm/cm rainfall (Jenny and Leonard 1934; Royer 1999; Wynn 2004). The scatter
around this relationship is very large, due to a host of factors such as erosion of paleosol
surfaces, texture of parent material, local drainage conditions, and so forth (Royer 1999). In our
experience, depth of leaching is useful where multiple paleosols can be observed, vertically and
along strike, in an attempt to average out some of the non-climatic in uences listed above.
-8
-6
-4
-2
0
2
4
6
8
10
12
14
0612
month
soil T
d = 0cm
d = 50cm
d = 150 cm
d = 300cm
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
0612
month
deviation from δ
δ
18
Osc value at MAT
50 cm
150 cm
Figure 5. Depth pro les of soil temperature for a site at 4100 m in southern Tibet (see text) calculated using
MAT from nearby Lhasa (3660 m) of 8°C and a local air-T lapse rate of 6 °C/km, a) showing modeled
seasonal uctuations in soil temperature at various depths, and (b) seasonal deviations from predicted
δ18Osc values at MAT.
03_Quade_etal.indd 1003_Quade_etal.indd 10 7/3/2007 12:41:36 AM7/3/2007 12:41:36 AM
Paleoelevation Reconstruction using Pedogenic Carbonates 11
For a more quantitative approach to paleoaridity, carbon isotopes from pedogenic carbonate
can provide evidence of soil respiration rates, which are closely related to plant cover at a site,
and hence aridity. Respiration rates cannot be calculated from δ13Csc values without independent
knowledge of δ13C values of local plant cover, because both C3 and C4 plants compose modern
vegetation. C4 plants expanded across low-latitude and mid-latitude ecosystems between 8
and 4 Ma, prior to which time C3 plants appear to dominate most ecosystems (Cerling et al.
1997; but see also Fox and Koch 2004). This ecologic simpli cation makes estimation of paleo-
respiration rates from δ13Csc values possible prior to about 8 Ma, since the only two end members
contributing to soil CO2 are C3 plants and the atmosphere (
Fig. 6
Fig. 6). Prior to the mid-Oligocene,
pCO2 appears to have been much higher (Pagani et al. 1999; Pearson and Palmer 2000; Pagani
et al. 2005), once again complicating estimation of paleo-respiration rates. But for the period
30-8 Ma, paleo-respiration rates can be calculated from δ13Csc values alone, allowing us to trace
the development of rain shadows in the lee of both the emerging Himalaya and Andes.
The basic principle, rst modeled in Cerling (1984) and modi ed in Davidson (1995) and
Quade et al. (2007), is that the δ13C value of soil CO2, and the pedogenic carbonate from which
it forms, is determined by the local mixture of C3 plant CO2 and atmospheric CO2. In general,
the deeper in the soil pro le or the higher the soil respiration rate, the greater the in uence of
plant-derived CO2. Thus, for a given soil depth, δ13Csc values are determined by soil respiration
rate. Temperature has a very modest effect on carbon isotope fractionation (~ −0.11 ‰/°C). As
an example, δ13Csc values will increase from −4.6 to −9.4‰ as respiration rates increase from
0.5 to 8 mmoles/m2/hr at 9 °C (Fig. 6), where > 0.5 mmoles/m2/hr roughly translates into MAP
> 200 mm/yr in the Mojave Desert. Later in the paper we will use this framework to interpret
δ13Csc values from Tertiary-age carbonates in Himalaya/Tibet and the Andes.
-100
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
-10 -8 -6 -4 -2 0 2 4 6
δ
13C(PDB)
0.1 mmoles/m2/hr
0.3
0.4
8410.5
Figure 6. Changes in the δ13C (PDB) value of soil carbonate at various respiration rates (indicated) in a
pure C3 world and at low pCO2 (400 ppmV) from 30 to 8 Ma, as predicted by the one-dimensional diffusion
model of Cerling (1984) and Quade et al. (2007), assuming δ13C (PDB) of plants = −25‰, δ13C (PDB)
of the atmosphere = −6.5‰, an exponential form to CO2 production with depth, and characteristic CO2
production depth (k) of 32 cm. Black bar denotes range of sampled carbonate soil depths and average δ13Csc
(PDB) values from late Oligocene paleosol carbonate samples from Nima, southern Tibet; horizontal error
bar re ects uncertainties in soil porosity = 0.5 ± 0.1 (see Hillel 1982), T = 9±3°C, and pCO2 = 400 ± 100
ppmV (see Pagani et al. 2005).
03_Quade_etal.indd 1103_Quade_etal.indd 11 7/3/2007 12:41:37 AM7/3/2007 12:41:37 AM
12 Quade, Garzione, Eiler
Elevation variation in the δ18Osc value of pedogenic carbonate
It follows from our previous discussion of modern pedogenic carbonates that there should
be a strong correlation between elevation of sites and local δ18Omw, and hence δ18Osc values. For
all the interest in paleoelevation reconstruction, there is surprisingly little ground-truthing of
the elevation-δ18Osc relationship by direct sampling of modern carbonates. Most studies simply
calculate theoretical δ18Osc values using some assumed δ18Omw/elevation relationship, and local
MAT. Here we explore the validity of these assumptions by looking at actual, not theoretical,
δ18Osc values from modern soils from different elevations in southern Nevada, Tibet, and the
Paci c slope of the Andes. Our conclusion is that δ18Osc values from the warm and arid Mojave
and hyperarid Atacama are severely affected by evaporation, whereas evaporation effects are
much reduced in carbonates from moist India/Nepal and dry but very cold Tibet.
δ18O values of modern (= Holocene) pedogenic carbonate collected from 50 cm depth in
Death Valley and up the east faces of the Spring and Grapevine Mountains of southern Nevada
decreases steeply with elevation gain by about −5.5 ‰/km (r2 = 0.65) (Quade et al. 1989a)
(
Fig. 7
Fig. 7). By comparison, the δ18Omw values compiled from 32 stations located across the region
decreases with elevation by a modest 1.3 ‰/km (r2 = 0.4) (Friedman et al. 1992). If these δ18Omw
values are representative of rainfall falling on soil sites, the δ18Osc values predicted to form in
equilibrium with them (adjusting for temperatures at each site) would all be around −13‰
(Fig. 7), and virtually invariant with elevation. We only observe these very low values deep
in soils at the wetter sites (MAP > 200 mm/yr). The residual between observed and predicted
δ18Osc values is strongly correlated (
Fig. 8
Fig. 8; r2 =0.76) with mean annual rainfall. This shows that
almost all of the steep gradient in δ18Osc values that we observe over a 3000 m elevation range
0
500
1000
1500
2000
2500
3000
-14 -12 -10 -8 -6 -4 -2 0
δ
18O (VPDB)
elevation (masl)
limestone
non-limestone
profile SM-2b
profile PaM-1
profile SM-3b
predicted
Figure 7. δ18Osc (PDB) of soil carbonate collected from ~50 cm soil depth versus elevation (masl) from
Holocene-age soils in the Mojave Desert (from Quade et al. 1989a). Soil carbonate collected on both
limestone and non-limestone parent materials show the same approximate decrease with elevation of −5.5
‰/km. Soil pro le values are the same as in Figure 4. The line denotes the δ18Osc (PDB) predicted from
local δ18Omw values from 33 rainfall collection sites in the Mojave Desert (δ18Omw (SMOW) = −0.0013
× elevation (m) − 10.68; r2 = 0.39; converted from δD values using δD = 7.2δ18O + 7.8) compiled from
Friedman et al. (1992), using MAT + 2 °C at each site (T (°C) = −0.0079 × elevation (m) − 23.07; r2 = 0.93;
8 climate stations).
03_Quade_etal.indd 1203_Quade_etal.indd 12 7/3/2007 12:41:38 AM7/3/2007 12:41:38 AM
Paleoelevation Reconstruction using Pedogenic Carbonates 13
is produced by evaporation, and changes in δ18Omw values have little in uence. At rainfall rates
below 200 mm/yr, δ18Osc values from the same pedogenic carbonates are highly enriched in 18O
by evaporation (Fig. 8), and can yield large (> 2 km) underestimates of actual elevation.
We recently completed a transect up the Paci c slope of the Andes at 24°S (Fig. 1a;
Fig. 9
Fig. 9).
The transect spans the central sector of the Atacama Desert, one of the driest regions in the
world. δ18Osc values for samples > 50 cm depth from this transect vary between −5.9 and
+5.4‰. Scatter of data within one pro le or across several pro les in a narrow elevation range
is very large, often > 5‰, even from samples deeper than 50 cm. In spite of this scatter, the
most negative δ18Osc values decrease modestly with elevation, from as low as −2‰ on the coast
to −5.9‰ at 3400 m, with an excursion toward higher values in the driest part of the transect
between about 1000 and 2500 m, a region so dry that it is virtually plantless.
The δ18Omw values from Fritz (1981) and Aravena et al. (1999) can be used to predict
δ18Osc values assuming MAT + 2 °C at soil depth, with an error of ≤ 1.5‰ at ≥ 50 cm, as
calculated in a previous section. In the coastal Atacama, fog drip outstrips rainfall from the
rare Paci c storm (Larrain et al. 2002), and hence we compare δ18Omw values of fog (Aravena
et al. 1989) and of local rainfall as assumed parent waters in the coastal zone. We nd that
observed δ18Osc values always exceed predicted δ18Omw values (Fig. 9). Predicted δ18Osc values
along the transect between sea level and 4000 m are between about −3 and −15‰, whereas the
most negative observed δ18Osc values range between −3 and −5.9‰. Clearly, the prognosis for
meaningful paleoelevation reconstruction in such arid conditions is poor. For example, even
with our very large sample size of >60 from between 3000 and 4000 m, we would underestimate
paleoelevation using only the most negative values by at least 1 km. This is why it is vital to
make some assessment of paleoaridity independent of the conventional δ18Osc measurements.
The second data set that we brie y consider comes from selected locations along a north-
south transect across the Himalaya and Tibet (
Fig. 10
Fig. 10). This is work in progress, but enough
information is already in place to make some useful observations. As with other data sets, we
focused on sampling deeper (> 50 cm) carbonate from younger pro les, although in the case of
y = -0.02x + 9.86
R2 = 0.76
-2
0
2
4
6
8
10
12
14
0 50 100 150 200 250 300 350 400 450 500
mean annual rainfall (mm/yr)
residual (observed-predicted)
δ18O observed = δ18O predicted
Figure 8. The residuals calculated from Figure 7 plotted against
mean annual rainfall for each Mojave Desert site.
03_Quade_etal.indd 1303_Quade_etal.indd 13 7/3/2007 12:41:39 AM7/3/2007 12:41:39 AM
14 Quade, Garzione, Eiler
0
500
1000
1500
2000
2500
3000
3500
4000
4500
-20 -15 -10 -5 0 5 10
δ
δ
18O (VPDB)
elevation (m)
1.0 0.5 0.2
coastal
fog
Figure 9. δ18Osc (PDB) values of Holocene soils versus elevation in the Atacama Desert, between 22.5
and 25°S. The thin lines (solid, dashed, dotted) denote modeled fraction (f) of soil water remaining after
evaporation, using local site MAT and δ18Omw (SMOW) values for rainwater (from Aravena et al. 1999 and
Fritz 1981); and the thick line for f = 1.0 for coastal fog from Aravena et al. (1989).
-25
-20
-15
-10
-5
0
27 28 29 30 31 32 33
°N latitude
δ
δ
18
O (VPDB) carbonate
Quaternary-age soils
predicted at 2.5°C
predicted at 8.5°C
Nima paleosols
evaporation
p
redicted
carbonate values
Figure 10. δ18Osc (PDB) values from Quaternary-age soils across the Himalayan-Tibetan orogen, and
“corrected” values (see text) from late Oligocene paleosols at Nima. Lines denote δ18Osc (PDB) values
predicted at a range of soil temperatures (indicated), and using the regression equation of unpublished
water data across the Tibetan plateau shown in Figure 2b.
03_Quade_etal.indd 1403_Quade_etal.indd 14 7/3/2007 12:41:39 AM7/3/2007 12:41:39 AM
Paleoelevation Reconstruction using Pedogenic Carbonates 15
Tibet we also obtained carbonates from older geomorphic surfaces, in order to develop some
perspective on the isotopic effects of climate variability during the Quaternary.
In general, δ18Osc values follow the expected pattern based on changes in modern δ18Omw
values across the Himalayan-Tibetan orogen (Fig. 10). To make this comparison we drew on
our published and unpublished data on Holocene-age carbonates from three areas: modern
Pakistan, southern Tibet, and a few scattered sites in NE Tibet. In Pakistan, twelve analyses
of modern pedogenic carbonates from three pro les yield an average value of −6.3‰ (Quade
and Cerling 1995). This value is very similar to the long-term average for the Quaternary-age
paleosols from Pakistan (−6.4‰) and from central Nepal (−7.0) (Quade et al. 1995b). The most
negative modern δ18Osc values of −7.7‰ in Pakistan, and −8.5‰ for the late Neogene Siwalik
paleosols are also very similar, and are close to that predicted (−8.7‰) to form from weighted
modern δ18Omw values from New Delhi (Araguás-Araguás et al. 1998). So, based on this admit-
tedly restricted sample of carbonates and rainfall, the system appears to be well-behaved: we
can accurately reconstruct paleo-elevation for these sites during the late Neogene, which have
always been at low elevation (< 500 m), using the most negative δ18Osc values, and the scatter of
δ18Osc values is relatively low (< 3‰), indicating only modest evaporative modi cation.
Turning brie y to southern Tibet, we have preliminary data from sixteen surface soils
collected between 29 and 32°N, 79 to 92°E, and 3800-5400 m. Pedogenic carbonate from these
surface soils displays a range of morphologies from thin gravel coatings to pervasive matrix
cements, suggestive in some cases of ages >Holocene, unlike our Mojave and Atacama soils,
which are all very young (Holocene). The Tibetan samples yield a wide range of results, from
−18 to +3‰ (Fig. 10). For comparison, we can predict δ18Osc values using δ18Omw values from
Figure 2b and assuming local the range of MAT + 2 °C encompassing these sites. As with the
Mojave and Atacama Desert samples, many observed δ18Osc values exceed predicted values,
again because of evaporation. Interestingly, in many cases the observed δ18Osc values are 1 to
5‰ more negative than expected δ18Osc values. Detailed consideration of this interesting but
incomplete data set is beyond the scope of this review, but we believe that it may illustrate the
effects of perhaps the largest uncertainty in all paleo-elevation analysis using δ18Osc values,
that of changing climate. We suggest that modern meteoric waters probably have higher δ18O
values than average longer-term waters as archived by Quaternary-age pedogenic carbonates.
This may be due to cyclical changes during the Quaternary in the average penetration northward
of Asian monsoon rainfall into southern Tibet. If true, this would lead to overestimates of
paleoelevation using modern δ18Omw-elevation relationships. Our analyses illustrate the key
advantage of sampling Quaternary pedogenic carbonates as opposed to just modern meteoric
waters in calibration studies. Carbonates provide more time-depth and will tend to average out
the swings in Quaternary climate.
A few samples of pedogenic carbonate from 3000-3500 m in northeast Tibet fall between
−11 and −4‰ in δ18Osc (PDB), more positive than values in southern Tibet at comparable
elevations, and consistent with overall regional trends of increasing δ18Omw values with
latitude. δ18O values of surface waters from this area and approximate elevation (Garzione et
al. 2004) range from −11 to −8‰, consistent with the higher observed δ18Osc values.
PALEO-ELEVATION ESTIMATES USING
CARBONATE “CLUMPED ISOTOPE” PALEOTHERMOMETRY
Up to this point our review has focused on reconstruction of paleoelevation using δ18Osc
values. This approach has the advantage that suitable samples are common constituents of
paleosols, the necessary analytical methods are widely available and straightforward, and the
various complicating factors, such as soil-water evaporation and post-depositional diagenesis,
have been studied in samples with relatively well-understood histories. Nevertheless, the
03_Quade_etal.indd 1503_Quade_etal.indd 15 7/3/2007 12:41:41 AM7/3/2007 12:41:41 AM
16 Quade, Garzione, Eiler
validity of this “conventional” method ultimately rests on the ability to: (1) account for the
effect of temperature on the isotopic fractionation between soil water and pedogenic carbonate;
and (2) distinguish orographic effects on the isotopic composition of meteoric water from other
possible in uences, such as seasonality, and climate and latitude change.
Ghosh et al. (2006b) present an innovation to paleoelevation reconstruction using pedo-
genic carbonate that potentially addresses these issues. The Ghosh et al. approach uses carbon-
ate “clumped isotope” thermometry (Ghosh et al. 2006a; Schauble et al. 2006) to impose an
independent constraint on the temperatures of pedogenic carbonate growth. This information
permits a more robust estimate of the δ18Omw from which carbonate grew and provides an in-
dependent constraint on paleoelevation by comparison with the altitudinal gradient in surface
temperature.
Principles, methods and instrumentation
Carbonate clumped isotope thermometry is based on the temperature dependence of the
abundances of bonds between 13C and 18O in carbonate minerals. This temperature sensitivity
stems from the fact that isotope exchange reactions such as:
Ca13C16O3 + Ca12C18O16O2 = Ca13C18O16O2 + Ca12C16O3 (3)
are driven to the right with decreasing temperature. That is, ordering, or “clumping,” of heavy
isotopes into bonds with each other is favored at low temperatures.
We are aware of no analytical method for directly measuring the abundances of reactant and
product species in equation (3) with precision suf cient for useful geothermometry. Therefore,
the carbonate clumped isotope thermometer uses the abundances of analogous species in CO2
produced by phosphoric acid digestion of carbonate (i.e., 13C16O2, 12C18O16O, 13C18O16O and
12C16O2). These measurements are made using a gas-source isotope ratio mass spectrometer that
has had its collection array modi ed to permit simultaneous collection of all cardinal masses 44
through 49 corresponding to CO2 isotopologues.
Ion corrections, standardization and nomenclature for analyses of 13C16O2, 12C18O16O, and
12C16O2 follow established conventions (Santrock et al. 1985; Allison et al. 1995; Gon antini et
al. 1995). 13C18O16O is a special case in several respects (Eiler and Schauble 2004; Wang et al.
2004; Affek and Eiler 2006): most of the variations in its abundance in natural CO2 (including
that produced by acid digestion of naturally occurring carbonates) arise from variations in bulk
isotopic composition. In other words, if a population of CO2 molecules contains an unusually
large amount of 13C and/or 18O, then it will contain an unusually large number of 13C18O16O mol-
ecules simply because the probability of a molecule incorporating both rare isotopes is relatively
high. This cause of variations in 13C18O16O abundance is relatively uninteresting because it tells
nothing beyond what is known from measuring δ13C and δ18O. Therefore, the nomenclature for
reporting analyses of 13C18O16O considers its enrichment or depletion relative to the amount that
would be present in the analyzed population of molecules if that population had a “stochastic
distribution” of isotopes among all possible isotopologues; that is, if the probability that a mole-
cule contains a given isotope in a given location within the molecule were entirely random. This
enrichment or depletion is reported using values of “Δ47,” which are de ned as the difference, in
per mil, of the ratio of mass 47 to mass 44 isotopologues in a sample from that ratio expected for
a stochastic distribution of isotopes in that sample. Note that the population of mass 47 isotopo-
logues includes contributions from 12C17O18O and 13C17O2, which cannot be mass discriminated
from 13C18O16O using existing gas source isotope ratio mass spectrometers. These species must
be accounted for when calculating the expected stochastic distribution and when interpreting
observed values of Δ47, but are not a signi cant complicating factor for most materials of inter-
est to the present discussion. Standardization of Δ47 values requires comparison of samples with
materials that are known to have a stochastic distribution. This is generally accomplished by
comparison with gases that have been internally equilibrated by heating to 1000 °C.
03_Quade_etal.indd 1603_Quade_etal.indd 16 7/3/2007 12:41:41 AM7/3/2007 12:41:41 AM
Paleoelevation Reconstruction using Pedogenic Carbonates 17
The relationship between growth temperature of a carbonate mineral and the Δ47 value of
CO2 produced by acid digestion of that carbonate is known for a variety of natural and synthetic
materials (including recent pedogenic carbonate) (
Fig. 11
Fig. 11). The temperature dependence
of −0.005‰ per °C is clear but subtle. Moreover, 13C18O16O is an exceedingly rare isotopic
species, only ~45 ppm of most natural CO2. These factors give rise to the two principal technical
limitations to carbonate clumped isotope thermometry: (1) exceptionally long analyses of ~one
hour and large samples of ~5 mg are required to generate measurements of Δ47 having precision
good enough for many problems in paleothermometry; and, (2) samples must be thoroughly
cleaned by cryogenic, gas chromatographic and other procedures prior to analysis (Eiler
and Schauble 2004; Affek and Eiler 2006; Ghosh et al. 2006a). The primary effect of these
limitations is to slow the rate at which usefully precise measurements can be made. Previous
experience suggests that the highest precision that can be consistently achieved is ~ 0.005 to
0.010 ‰ (Came et al. in review), and involves such a high level of replicate measurements and
standardization that only two or three unknowns per day can be analyzed.
Advantages and disadvantages
Carbonate clumped-isotope thermometry brings several signi cant capabilities to the
problem of paleoelevation reconstruction. Most importantly, the thermometer is based on
the thermodynamics of a homogeneous reaction, involving a reaction among components
of a single phase. Therefore, the temperature of carbonate growth is de ned rigorously by
analysis of the isotopic constituents of carbonate alone, or, more precisely, CO2 produced from
that carbonate. Nothing need be known about the isotopic composition of water from which
carbonate grew to determine the temperature of carbonate growth.
Figure 11. Data documenting the calibration of the carbonate clumped-isotope thermometer for inorganic
calcite grown in the laboratory ( lled circles) and aragonitic corals grown in nature at known temperatures
(an example of one of several biogenic materials we have also calibrated; un lled symbols). The large, gray
circle shows the result of analyses of a modern soil carbonate collected from the Bolivian Altiplano plateau.
The horizontal position of this data point is based on the mean annual surface temperature near the site of
collection between 2004 and the present.
03_Quade_etal.indd 1703_Quade_etal.indd 17 7/3/2007 12:41:41 AM7/3/2007 12:41:41 AM
18 Quade, Garzione, Eiler
Secondly, the material analyzed to determine growth temperature is the same as that used
to de ne the δ18Osc value (in fact, the two measurements are made simultaneously on the
same aliquot of CO2). Because one knows the growth temperature from Δ47 measurements and
δ18Osc values through conventional measurements, one can calculate the δ18Osw value from
which it grew based on the known temperature dependence of the carbonate-water oxygen
isotope fractionation (e.g., Kim and O’Neil 1997). Thus, clumped isotope thermometry,
combined with conventional stable isotope geochemistry, provides two complementary bases
for estimating paleoelevation: comparison of the record of carbonate growth temperatures with
inferred altitudinal gradients in surface temperature, and comparison of calculated values of
the δ18Osw with inferred altitudinal gradients in the δ18Omw values.
Finally, the correlation of calculated δ18Osw values with carbonate growth temperature
imposes one more potentially useful constraint: the slope of this trend, ∂δ18Owater/∂T, for an
altitudinal transect is, in many cases, a known or predictable quantity that differs from slopes
resulting from variations in season, latitude or climate (Ghosh et al. 2006b; see also Figure 15
of this chapter). Therefore, knowledge of this slope for a suite of related pedogenic carbonates
can potentially distinguish between variations in surface temperature and/or the δ18Omw value
caused by changes in elevation versus variations caused by other factors. In practice, this
constraint will be more useful in some cases than in others because of natural variations in
∂δ18Owater/∂T slopes for processes of interest, and because some suites of pedogenic carbonates
will be in uenced by more than one factor. Nevertheless, it provides another basis for improving
the con dence and precision of paleoelevation estimates.
Approaches to paleoelevation reconstruction based on carbonate clumped isotope
thermometry also carry with them several disadvantages, most importantly, that the analytical
techniques involved in a clumped isotope paleotemperature estimate (Eiler and Schauble 2004;
Affek and Eiler 2006; Ghosh et al. 2006b) are slow, technically dif cult, and presently made
in only one laboratory. It seems likely that these measurements will become more routine
as they are adopted in other laboratories and improved through various possible innovations
(such as automation). Nevertheless, the supply of data is a limitation at present. Secondly, the
precision of data is poor for some problems. This too appears likely to improve through time
(e.g., Came et al. in review) report carbonate clumped isotope measurements with precisions
corresponding to errors of only ± 0.9 °C), but will remain a signi cant issue for the foreseeable
future due to the subtle temperature sensitivity of isotopic clumping in Reaction (1). Finally,
because carbonate clumped isotope thermometry is a relatively new technique, the potential
exists that there are unrecognized systematic errors in its application to pedogenic carbonates,
such as unexpected fractionations during pedogenic carbonate growth, or poor preservation.
Diagenesis
Any attempt to reconstruct elevation using geochemical proxies must consider the
possibility that diagenesis and burial metamorphism have overprinted the carbonate record.
The δ18O values of all carbonates are quite susceptible to diagenetic alteration, as witnessed by
the extraordinary dif culty in reconstructing δ18O values of sea-water through deep geologic
time using carbonates, as contrasted with, for example, δ13C values and 87Sr/86Sr ratios (e.g.,
Veizer et al. 1999). Carbonates buried to depths of several km or less can be overprinted
through dissolution and re-precipitation reactions with ground and formation waters. The best
approach to evaluating the potential effects of diagenesis depends on the material under study.
For example, studies that use non-marine shells can take advantage of the fact that unaltered
shell retains its primary structure (Veizer et al. 1999) and has not been converted to calcite from
aragonite. In the case of soil nodules, recrystallization to spar is a clear warning sign, since the
primary texture of pedogenic carbonate is mostly micritic to microsparitic (Chadwick et al.
1988; Deutz et al. 2002; Garzione et al. 2004). These replacement textures can be identi ed
with the naked eye, optical microscope or cathodoluminiscope, and avoided by sampling
03_Quade_etal.indd 1803_Quade_etal.indd 18 7/3/2007 12:41:42 AM7/3/2007 12:41:42 AM
Paleoelevation Reconstruction using Pedogenic Carbonates 19
with a slow-speed drill or similar tool, as in both our Andean and Tibet studies discussed
below. This strategy appears to be suf cient for recovering records of primary variation in
δ18Osc and δ13Csc values (e.g., Koch et al. 1995). However, there is no means of knowing, a
priori, whether this approach will avoid overprinting of the 13C-18O ordering that governs the
carbonate clumped isotope thermometer. Hence, in the next section on the Andes we discuss
other ways to evaluate the delity of the clumped isotope thermometer.
Another approach to diagenesis discussed in the nal section of this paper is to look for
heterogeneity in δ18Osc values of adjacent calcitic phases such as interbedded marine, detrital
(Quade and Cerling 1995), lacustrine (e.g., Cyr et al. 2005; DeCelles et al. 2007a), and paleosol
carbonates. The concept here is that diagenesis should tend to homogenize δ18O values that in
a primary state may have had very different values.
PALEOSOL RECORDS OF PALEOELEVATION CHANGE: CASE STUDIES
We now turn to published and unpublished results of case studies from the Andes and Asia
as illustrations of how δ18Osc and Δ47 measurements from paleosol carbonates can be used in
paleoelevation reconstruction. The key points that we address include: (1) accounting for the
role of climate change in reconstructing paleoelevation; (2) testing for diagenetic alteration
of primary δ18Occ values due to burial in deep sedimentary basins; (3) evaluating paleoaridity,
and therefore extent to which evaporation has increased the δ18Omw values of paleowaters,
producing underestimates of paleoelevation, (4) comparing pedogenic carbonate records to other
sedimentary carbonates to determine the delity of various elevation proxies, and (5) the power
of combining conventional δ18Omw and Δ47 measurements in paleoelevation reconstruction.
Paleoelevation reconstruction of the Bolivian Altiplano
We can combine both δ18O and Δ47 measurements of pedogenic carbonate from the
Miocene deposits of Bolivia to trace uplift of the Bolivian Altiplano through time. The northern
Altiplano is an attractive target for both approaches to paleoelevation reconstruction for several
reasons. It preserves carbonate-bearing sediments deposited over much of its uplift history; the
modern altitudinal gradients in temperature and the δ18Omw values are known (Gon antini
et al. 2001), providing a locally calibrated basis for reconstructing past elevation; modern
rainfall rates are > 250 mm/yr, providing adequately wet conditions to reconstruct meteoric
water compositions; previous studies of the mountain belt history and geomorphology of the
Altiplano provide observations that can be used to test and elaborate on one’s results; and,
the amplitude of elevation change is large (several km) and thus the analytical precision of
clumped isotope measurements is not a severe limitation.
The Altiplano basin is a broad internally drained basin that occupies the central Andean
plateau between 15°S and 25°S. At an average elevation of ~3800 m, the Altiplano is situated
between the Eastern and Western cordilleras that reach peak elevations in excess of 6 km
(Figs. 1a, 2a). The Western Cordillera is the modern magmatic arc of the Andes, and the
Eastern Cordillera is a fold-thrust belt made up of deformed Paleozoic metasedimentary rocks.
The Altiplano basin has been internally drained since at least late Oligocene time, evidenced
by both westward paleo ow and derivation of sedimentary sources from the Eastern Cordillera
(Horton et al. 2002; DeCelles and Horton 2003). The late Miocene stratigraphic sections near
Callapa are ~3500 m thick and are exposed in the eastern limb of the Corque synclinorium.
Three measured sections include uvial and oodplain deposits in the lower 1200 m and the
upper 800 m and widespread lacustrine deposits (laterally continuous for more than 100 km
along strike) in the middle part of the section (Garzione et al. 2006). Age constraints on these
rocks come from 40Ar/39Ar dates on tuffs within our measured section (Marshall et al. 1992)
and magnetostratigraphy (Roperch et al. 1999; Garzione et al. 2006).
03_Quade_etal.indd 1903_Quade_etal.indd 19 7/3/2007 12:41:42 AM7/3/2007 12:41:42 AM
20 Quade, Garzione, Eiler
Fluvial- oodplain and lacustrine sediments contain authigenic carbonates for which δ18Osc
values and δ13Csc values were obtained (Garzione et al. 2006). The oodplain deposits contain
both paleosol carbonate nodules and palustrine carbonates. Paleosols are massive and red to
reddish brown. Discrete carbonate (Bk) horizons include rare carbonate rhizoliths and occur
below the upper part of the B horizon that has been leached of carbonate. Paleosol carbonate
nodules, 0.5 to 3 cm in diameter, were sampled between ~20 and 80 cm below the top of the
paleosol where their depth within the soil pro le could be measured. In many instances, it was
not possible to determine the top of the soil pro le because, in general, oodplain lithofacies
show extensive oxidation and pedogenesis that makes it dif cult to identify the top of individual
soil pro les. Palustrine carbonates represent marsh or shallow pond deposits in the oodplain
adjacent to uvial channels. These laminated, mud-rich micrites presumably precipitated
seasonally when evaporation rates and productivity were higher. Within the lacustrine interval
in the middle part of the section, carbonates are rare and include laminated, very thinly bedded,
micritic limestone that contains vertical worm burrows and laminated, thinly to thickly bedded,
calcareous mudstone. In thin section, paleosol, palustrine, and lacustrine carbonates lack sparry
calcite, suggesting that they have not undergone extensive, late-stage diagenesis. This inference
is supported by Δ47 paleothermometry data that indicate that pedogenic carbonates do not show
a systematic increase in formation temperature with burial depth (Eiler et al. 2006;
Fig. 12
Fig. 12).
Garzione et al. (2006) excluded lacustrine carbonates from their paleoelevation analysis
while including palustrine and pedogenic carbonates. Both modern and ancient lake studies
show that closed-basin lakes, such as the paleolake that occupied the northern Altiplano in
late Miocene time, experienced moderate to extreme evaporative enrichments in the 18O of
T(˚C)
Burial depth (m)
Estimated burial geotherm
Salla 28.5-25.5 Ma
Corque 23.7-23.6 Ma
Tambo Tambillo 25.0-23.0 Ma (lacustrine)
Callapa 11.4-4.6 Ma
Modern Callapa
Figure 12. Apparent growth temperatures for various Altiplano carbonates based on clumped isotope
thermometry, plotted as a function of estimated maximum burial depth. Symbols discriminate among
soil carbonates from sections near Callapa, Corque and Salla and lacustrine carbonates from near Tambo
Tambillo, as indicated by the legend. The heavy solid line indicates an estimated burial geotherm, assuming
a surface temperature of 20 °C and a gradient of 30 °C per km. The dashed lines de ne a ±10° offset from
this trend, which we consider a reasonable estimate of its uncertainty. Carbonates deposited within the last
28.5 Ma and buried to 5000 meters or less exhibit no systematic relationship between apparent temperature
and burial depth, and show no evidence for pervasive resetting of deeply buried samples. Error bars are ±1σ
(when not visible, these are approximately the size of the plotted symbol).
03_Quade_etal.indd 2003_Quade_etal.indd 20 7/3/2007 12:41:42 AM7/3/2007 12:41:42 AM
Paleoelevation Reconstruction using Pedogenic Carbonates 21
water and carbonate that precipitated from it (e.g., Talbot 1990). Studies in the Altiplano region
support the inference that closed-lake waters in the Altiplano and Eastern Cordillera have
higher values of δ18O than local rainfall (Wolfe et al. 2001). In addition, Altiplano lacustrine
deposits contain abundant gypsum suggestive of extreme evaporation that would systematically
bias carbonates toward lower estimates of paleoelevation. Comparing lacustrine deposits
to paleosol carbonates above and below the lacustrine interval shows that δ18Omw values of
lacustrine carbonates are on average 2.8‰ to 3.8‰ higher than pedogenic carbonates (
Table 1
Table 1),
which reinforces the notion that they record the effects of evaporative enrichment in 18O.
Palustrine carbonates form in settings comparable to open-basin lakes through which water
ows and where evaporative enrichment is much less than in closed basins. These depositional
environments are in communication with river waters, lack interbedded gypsum, and therefore
should not be as evaporitic. Despite lack of evidence for extreme evaporation, these carbonates
generally record δ18O values that are ~1‰ higher than δ18Osc values of similar age (Table 1).
The consistently lower δ18Osc values indicate that paleosol carbonates are probably the most
faithful recorders of the isotopic composition of meteoric water and therefore provide the best
estimates of paleoelevation.
In modern soils globally, δ13Csc and δ18Osc values covary, with higher δ values re ecting
more arid conditions and lower respiration rates (Cerling 1984; Quade et al. 1989a; Deutz et
al 2002). In contrast, within the Altiplano basin, δ13Csc values increase at 7.6 Ma, whereas
δ18Osc values decrease (
Figs. 13
Figs. 13a and b). A plot of δ18O values versus δ13C values (Fig. 13c)
(excluding one outlier) show an inverse relationship between δ18O and δ13C values (R2 = 0.37).
The timing of the increase in δ13C values could be interpreted to re ect lower soil respiration
rates (e.g., Cerling and Quade 1993) or an increasing proportion of C4 grasses in the Altiplano
at this time, coincident with the global expansion of C4 grasses (Cerling et al. 1997; Latorre et
al. 1997). Analysis of both modern and fossil tooth enamel of large grazing mammals in the
Altiplano shows no evidence for signi cant C4 presence in the region today or at any time in the
Table 1. Oxygen and carbon isotope data from stratigraphic sections sam-
pled in the eastern limb of the Corque syncline (from Garzione et al. 2006).
Sample
type δ18O(VPDB)
(mean ‰ ± 1σ)δ13C(VPDB)
(mean ‰ ± 1σ)# of
samples
11.5 to 10.3 Ma
paleosols −11.8 ± 0.9 −9.0 ± 1.0 14
palustrine −10.9 ± 1.5 −8.2 ± 0.8 9
10.1 to 9.1 Ma
lacustrine −9.0 ± 1.3 −8.8 ± 3.2 21
7.6 to 6.8 Ma
paleosol −12.8 ± 0.9 −4.3 ± 2.1 2
palustrine −11.3 ± 1.4 −8.4 ± 1.6 12
6.8 to 5.8 Ma
paleosol −14.7 ± 0.7 −6.1 ± 1.0 4
s.s. cement −14.8 ± 0.4 −8.6 ± 1.2 5
Standard deviation (1σ) of the mean of each sample set is reported. VPDB - Vienna Peedee
belemnite. Two pedogenic carbonate data points have been excluded, one that shows much
higher temperatures relative to other pedogenic nodules of the same age (based on Δ47
measurement) and one that has a much higher δ18O relative to other pedogenic nodules of
the same age.
03_Quade_etal.indd 2103_Quade_etal.indd 21 7/3/2007 12:41:44 AM7/3/2007 12:41:44 AM
22 Quade, Garzione, Eiler
past 8 Myr (Bershaw et al. 2006). Therefore, an increase in the proportion of C4 grasses is an
unlikely cause for the observed increase in δ13Csc values. We suggest that this increase may be
the result of lower respiration rates due to increasingly arid conditions in the Altiplano.
Increasing aridity in Altiplano basin is supported by several lines of evidence from the
sedimentary record. First, both the thickness of uvial channel deposits and their lateral
continuity decrease up-section. In the oldest part of the section, channel sandstone bodies up
to 15 m thick continue laterally for 100s of meters. Internally, these sand bodies consist of
amalgamated lenticular beds that show low-angle cross strati cation, interpreted to re ected
lateral accretion on migrating longitudinal bars. Individual bar deposits are up to 2 m thick,
re ecting the maximum channel depth during high discharge events. Above the lacustrine
interval, channel sandstone bodies are up to 5 m thick and continue laterally for tens of meters.
Individual beds often contain trough cross-strati cation and are <1 m thick, indicating shallower
channel depths compared to older deposits. Both shallower channel depths and a decrease in
the lateral extent and overall thickness of sandstone channel deposits suggest lower discharge
in the uvial systems above the lacustrine interval. The second line of evidence comes from the
depth of pedogenic carbonate formation. Below the lacustrine interval, pedogenic carbonates
were observed to have formed below 30 cm in the paleosol pro le, whereas pedogenic
4
5
6
7
8
9
10
11
12
-12 -10 -8 -6 -4 -2
4
5
6
7
8
9
10
11
12
-16
-15
-14
-13
-12
-11
-10
-9
-8
-14 -12 -10 -8 -6 -4 -2 0
-16 -14 -12 -10 -8
Age (Ma)
Age (Ma)
δ
18
O (VPDB)
δ
13
C (VPDB)
δ
18
O(VPDB)
δ
13
C (VPDB)
AB
C
Figure 13. (a) δ18O (VPDB) and
(b) δ13C(VPDB) versus age for pa-
leosol carbonates near Callapa. (c)
δ13C(VPDB) versus δ18O(VPDB)
for paleosol carbonates near Cal-
lapa. The linear regression shows
the inverse relationship between
δ13C and δ18O, excluding one data
point (open diamond), discussed
in the text.
03_Quade_etal.indd 2203_Quade_etal.indd 22 7/3/2007 12:41:44 AM7/3/2007 12:41:44 AM
Paleoelevation Reconstruction using Pedogenic Carbonates 23
carbonates formed at depths as shallow as 16 cm in paleosols above the lacustrine interval.
The observation in Holocene soils that depth to the Bk horizon increases with increasing mean
annual precipitation (Royer 1999) suggests that the youngest part of the soil record re ects
the most arid conditions. The combined observations of increased aridity in late Miocene time
point to lower soil respiration rates as the cause of the increase in δ13C values up section.
We revise and update paleoelevation estimates for the northern Altiplano here and in the
following section on “clumped isotope thermometry.” Realizing that Gon antini et al. (2001)
mislabeled data in their Table 5 as weighted means, when they are in fact unweighted means,
we estimate elevation as in Garzione et al. (2006) using the weighted mean values for 3 years
of rainfall data reported in Table 6 of Gon antini et al. (2001). Seven sites for which rainfall
amount was also reported can be used for the regression. The linear regression to these data is:
h = − 472.5δ18Omw − 2645 (4)
where h = elevation in meters, and with an r2 = 0.95 (
Fig
Fig. 2a,
14
14). We choose a linear regression
because there are not enough data points in the weighted mean data set to evaluate whether a
polynomial provides a better t to the data. Surface water data collected over 2004 and 2005
years from small tributaries along the Coroico River, where rainfall samples were collected,
agree with the weighted mean values observed in the sparse rainfall data set and show a
similar isotopic gradient to Equation (4), which indicates that the isotopic gradient observed in
meteoric water is re ected in surface water. Use of Equation (4) produces elevation estimates
of: 400 to 2200 m in carbonates deposited before 10.3 Ma, 2000 to 3800 m in carbonates
deposited between 7.4 Ma and 6.8 Ma, and 4000-4700 m in carbonates deposited after 6.8
Ma. This reanalysis of the data based on weighted mean precipitation suggests essentially the
same amount of surface uplift (2.5 to 3.6 km) as our original estimates of 2.5 to 3.5 km. It also
brings the calculated elevations for the oldest part of the section into a more reasonable range,
Elevation (m)
D O(vsmow)
18
0
1000
2000
3000
4000
5000
6000
-18 -16 -14 -12 -10 -8 -6 -4 -2
2005 tribuatry data
2004 tributary data
weighted mean
annual precipitation
Figure 14. δ18O (VSMOW) of rainfall and surface waters across the Eastern Cordillera. Rainfall data
represent the weighted mean isotopic composition (1983-1985) from Gon antini et al (2001). Tributaries
to the Coroico river were sampled in late May 2004 and early May 2005 and are plotted relative to the
sampling elevation. A linear regression to the rainfall data (gray triangles) de nes Equation (4) in the text.
03_Quade_etal.indd 2303_Quade_etal.indd 23 7/3/2007 12:41:45 AM7/3/2007 12:41:45 AM
24 Quade, Garzione, Eiler
with all values above 400 m, as opposed to the negative elevations calculated for the oldest
part of the record in Garzione et al. (2006). Given that paleoelevation estimates are prone
to systematic biases that tend to underestimate elevations, we suggest using the difference
between the highest elevation estimates prior to 10.3 Ma and after 6.8 Ma of 2.5 km as a
conservative estimate for the amount of late Miocene surface uplift, with an uncertainty of ±1
km (see Rowley and Garzione 2007 for discussion of uncertainties)
Ideally, to apply stable isotopes to paleoelevation reconstruction, low-elevation records of
regional climate are required to constrain variations in climate that may in uence the isotopic
record. For example, the Siwalik foreland basin deposits south of the Himalaya provide
low-elevation constraints on climate and the isotopic composition of meteoric water used to
estimate elevations in southern Tibet (Garzione et al. 2000a; Rowley at al. 2001; Currie et al.
2005). The accuracy of the current estimates of paleoelevation of the Altiplano is limited by
lack of data from simultaneous low elevation records in the Subandean foreland (Rowley and
Garzione 2007). Several potentially confounding variables that are currently unconstrained
are the proximity and isotopic composition of sources of water vapor, secular changes in
terrestrial temperature associated with late Cenozoic cooling, and the effects of climate change
on the isotopic lapse rate. Because of extensive evapotranspiration over the Amazon basin,
modern water vapor delivered to the eastern ank of the Andes is isotopically similar to water
vapor at the Atlantic coast (Vuille et al. 2003). However, the continental isotopic gradient
may have changed through time as a result of changes in plant cover and/or changes in the
extent and nature of inland water bodies in the Amazon basin. Recent studies document a
late Miocene marine incursion in the northern Subandean foreland. Although these marine
rocks lack precise age constraints, Hernández et al. (2005) show that at least one marine unit
was slightly younger than 7.7±0.3 Ma. Hernández et al. (2005), Uba et al. (2006), and Hoorn
(2006) show that marine conditions existed brie y, while lacustrine conditions prevailed for
longer time periods. A lacustrine origin is supported by stable isotopic evidence that indicates
a predominantly Andean freshwater and cratonic freshwater origin for the water isotopic
composition recorded in microfossils deposited between 18°S and 19°S (Hulka 2005). The
changing paleogeography of the Andean foreland points to the need to better establish foreland
records of both climate and paleogeography to more accurately constrain paleoelevation of the
Altiplano in late Miocene time.
Results from clumped isotope thermometry
Burial effects. Eiler et al. (2006) examined the effects of post-depositional alteration on
the carbonate clumped isotope thermometer in soil and lacustrine carbonates from the northern
Altiplano, spanning a range in age from 0 to 28.5 Ma and maximum burial depths from ~0
to ~5 km. Age and burial depth estimates for these samples are based on data presented in
MacFadden et al. (1985), Kennan et al. (1995), Kay et al. (1998), Garzione et al. (2006),
and Ghosh et al. (2006b) and on unpublished magnetostratigraphic, U-Pb geochronologic,
and stratigraphic data summarized in Eiler et al. (2006). The results (Fig. 12) demonstrate
that the temperatures recorded by carbonate clumped isotope thermometry are uncorrelated
with burial depth, generally (i.e., with one exception) indistinguishable from earth-surface
temperatures, and preserve temperatures as low as ca. 20 °C in pedogenic carbonates as old as
23.6-23.7 Ma and buried as deep as 5 km. No systematic differences in apparent temperature
were observed between soil and lacustrine carbonates, suggesting that the thermometer is
insensitive to variations in textural characteristics such as grain size and porosity that might
in uence susceptibility to overprinting. These results suggest that burial diagenesis has no
systematic effect on the temperatures recorded by the carbonate clumped isotope thermometer
in micritic portions of pedogenic carbonates over timescales of tens of millions of years and up
to burial temperatures approaching 200 °C. Nevertheless, there are reasons to remain cautious
when applying this approach to new suites, because one of the samples originally examined by
03_Quade_etal.indd 2403_Quade_etal.indd 24 7/3/2007 12:41:46 AM7/3/2007 12:41:46 AM
Paleoelevation Reconstruction using Pedogenic Carbonates 25
Ghosh et al. (2006b) recorded a temperature of 50 °C—higher than any plausible depositional
temperature. This result is not representative of the record as a whole, but clearly indicates the
possibility for overprinting during burial. Moreover, the samples in Figure 12 collected near
Salla are shallowly buried but record temperatures that could lie on a burial geotherm. Eiler et
al. (2006) speculate that hydrothermal activity associated with nearby ~15-25 Ma magmatism
might have partially or completely overprinted these samples.
Paleoelevations of pedogenic carbonate from the Corque syncline
Only the 11.4 to 4.6 Ma paleosols exposed in the Callapa section (Figs. 1a, 12) of the
Corque syncline have been studied in enough detail to justify a detailed reconstruction of
elevation history. Ghosh et al. (2006b) present such a reconstruction, which we now review.
Our discussion of these data differs in some respects from that presented in Ghosh et al. (2006b)
because we have re-evaluated the constraints on long-term climate change (which in uences
the low-elevation “base line” of a paleoalititude estimate) and re-calculated the elevation trend
to δ18Omw values across the central Andes today. Figure 12 includes data for Callapa pedogenic
carbonates from two sources; those from Ghosh et al. (2006b) were analyzed with a relatively
high degree of replication and have uncertainties in Δ47 equivalent to ±2 to 3 °C, whereas data
from Eiler et al. (2006) were not as thoroughly replicated and have uncertainties in apparent
temperature averaging ±5 °C. This latter data set is suf ciently precise to test for burial
metamorphic overprinting (Fig. 12), but provides inferior constraints on paleoelevation. We
therefore focus our discussion on the data from Ghosh et al. (2006b).
Ghosh et al. (2006b) found that pedogenic carbonates from Corque syncline paleosols
deposited between 11.4 and 10.3 Ma record carbonate precipitation temperatures of 28.4 ± 2.6
°C (standard error, or s.e., of ±0.9 °C); those grown between 7.6 and 7.3 Ma record temperatures
of 17.7 ± 3.1 °C (s.e. of ±2.2 °C); and those grown between 6.7 and 5.8 Ma record temperatures
of 12.6 ± 5.6 °C (s.e. of ±2.8 °C). This range in apparent temperature through time is broadly
similar to the modern temperature variation with elevation between the east anks of the
Andes and the Altiplano, suggesting that the Altiplano may have risen from low elevation (ca.
1 km or less) to its modern elevation (ca. 4 km) between 10.3 and 6.7 Ma. Note that apparent
temperatures for pedogenic carbonates grown between 11.4 and 10.3 Ma are at the high end of
modern temperature variations at low elevations on the anks of the Andes. This could re ect
a warmer Miocene climate and/or preferential growth of paleosol carbonates during the austral
summer. If we neglect this complicating factor, these data suggest cooling of 15.7 ± 2.9 °C (±1
se). Given the modern gradient in temperature with elevation in the Andes today (4.66 °C/km;
Gon antini et al. 2001 and data compiled at www.climate-zone.com), this change is consistent
with uplift of 3.4 ± 0.6 m, for an average uplift rate of 0.94 ± 0.17 mm/yr between 10.3 and
6.7 Ma. Differences in climate and latitude can account for anywhere between 1.3 °C (Savin et
al. 1975; Smith et al. 1981, Zachos et al. 2001) and 2.3 °C (Berner and Kothavala 2001) of the
observed temperature change. If we correct for this systematic error, then the amount of cooling
due to elevation change could be a minimum of 13.4 ± 2.9 °C (assuming a 2.3 °C change in
low-elevation mean annual temperature, all between 10.3 and 6.7 Ma), implying elevation gain
of 2.9 ± 0.6 m and an uplift rate of 0.80 ± 0.17 mm/yr.
The contrast in δ18O (SMOW) values of waters in equilibrium with 11.4 to 10.3 Ma
pedogenic carbonates (average −8.6 ± 0.4‰, ± 1 s.e.) versus those in equilibrium with 6.7 and
5.8 Ma pedogenic carbonates (average −14.6 ± 0.6‰, ± 1 s.e.) is 6.0 ± 0.7‰, (± 1 s.e. in the
difference of the means). The modern gradient in the annual weighted average δ18Omw values on
the slopes of the Andes (equation 4) implies an elevation gain of 2.8 ± 0.3 km (from 1.4 ± 0.2 to
4.2 ± 0.3 km) between 10.3 and 6.7 Ma, or an uplift rate of 0.78 ± 0.08 mm/yr.
Figure 15
Figure 15 plots the growth temperatures of Corque syncline pedogenic carbonates versus the
δ18O values of waters from which they grew, and compares the trend in those data to the modern
trends in surface temperatures and meteoric water δ18O values across a range of elevations in the
03_Quade_etal.indd 2503_Quade_etal.indd 25 7/3/2007 12:41:46 AM7/3/2007 12:41:46 AM
26 Quade, Garzione, Eiler
central Andes. These modern trends are shown both for the annual weighted mean (solid grey
line) and Jan/Feb seasonal extreme (dotted gray line), and have been corrected for differences
between the Miocene and today in climate, latitude of the Altiplano and the δ18O value of the
ocean (see the caption to Figure 15 for details). Fine, dashed black contour lines of constant
elevation link the mean annual trend with the Jan/Feb extreme for the inferred Miocene trends.
This gure is useful in two respects: (1) the fact that the trend de ned by data for Miocene
pedogenic carbonates parallels the trend for an elevation transect re-enforces the interpretation
that variations in pedogenic carbonate growth temperature in the northern Altiplano and δ18O
primarily re ect elevation change rather than other factors (e.g., climate change; latitude
change; changing seasonality of pedogenic carbonate growth); and (2) we can use a sample’s
position in relation to the iso-elevation contours in Figure 15 to estimate its paleoelevation,
thereby circumventing systematic errors that could occur if pedogenic carbonates had any
variations in the season of growth. This is important because the plot of points in Figure 15
suggests that the season of formation for the pedogenic carbonates examined in this study are
biased toward warm, low-δ18O conditions that prevail in the Bolivian austral summer. Our
results suggest that the Altiplano rose 2.7 ± 0.4 km (from 0.6 ± 0.2 km to 3.3 ±0.4 km) between
10.3 Ma and 6.7, implying an average uplift rate of 0.73 ± 0.12 mm/yr.
Paleoatimetry of southern Tibet and the Himalaya
Burial effects. We have been involved in several studies in the Himalaya and Tibet that
illustrate the need to carefully evaluate the potential effects of burial metamorphism on δ18Osc
values before using them to reconstruct paleoelevation. The rst comes from Penbo in southern
Tibet (Fig. 1b) where Leier (2005) analyzed paleosol carbonates in close stratigraphic proximity
11.4-10.3 Ma
7.6-7.3 Ma
6.7-5.8 Ma
Temperature (˚C)
18OSMOW
of water
-8
-10
-12
-14
-16
-18
515
25
1km
3km
4km
5km
2km
2.4 km
3.3 km
0km
0.6 km
Figure 15. Plot of the δ18OSMOW value of water in equilibrium with soil carbonate nodules vs. the growth
temperatures of those nodules. Circles are averages for the 11.4-10.3 Ma, 7.6-7.3 Ma and 6.7-5.8 Ma age
groups, as indicated in the legend. Error bars are ±1 standard error of the population. Gray lines show the
mean annual trend (solid) and trend of Jan/Feb extreme (dotted) for the modern relationships between
surface temperature and δ18OSMOW of meteoric water. These trends are contoured for altitude in 1-km
increments. The similar heavy and dotted black trends plot the expected location of the mean annual and
Jan/Feb extremes in the mid-Miocene, based on inferred changes in the latitude of Bolivia, low-latitude
climate change and secular variation in the δ18O of sea water (Savin et al. 1975; Smith et al. 1981). Fine
dashed lines connecting the mid-Miocene mean annual trend and Jan/Feb extreme trend show the slopes
of seasonal variations in T and δ18O of water at a xed altitude. Paleoaltitudes of age-group averages are
estimated by their intersections with this set of altitude contours, as indicated by the italicized text.
03_Quade_etal.indd 2603_Quade_etal.indd 26 7/3/2007 12:41:46 AM7/3/2007 12:41:46 AM
Paleoelevation Reconstruction using Pedogenic Carbonates 27
(< 200 m) to marine carbonates from the late Cretaceous Takena Formation. This association
shows that the Takena paleosols formed near sea-level. The nodules from the paleosols are a
mix of micrite with local sparry zones, both of which we analyzed. Both textures yielded δ18Osc
values −12 to −14‰ (
Fig. 16
Fig. 16). These are implausibly low for a low-latitude coastal site and
re ect isotopic exchange, possibly at elevated temperatures, with meteoric waters with very low
δ18Omw values, such as those that occur in the region today. To test this, Leier (2005) analyzed
stratigraphically adjacent marine limestone, including spars, micrite, and fossils fragments.
Again all phases—from matrix micrite to recrystallized fossils—yielded much lower δ18O
values (−12 to −14‰; Leier 2005; Fig. 16) than expected (0 to −4‰) for late Cretaceous
marine carbonates (Veizer et al. 1999). This example clearly illustrates the effects of isotopic
homogenization due to burial metamorphism of originally heterogeneous calcite phases.
The second example is unpublished data from paleosol carbonates of the Oligocene-
age Dumri Formation in central Nepal (Fig. 1b). The Dumri Formation underlies the better
known Siwalik Group, and represents the early stages (Oligo-Miocene) of Himalayan foreland
deposition. Non-paleosol carbonates are unavailable from our study sections to check for isotopic
homogenization expected from diagenesis. Petrographic evidence shows local but not pervasive
recrystallization of paleosol nodules. Most δ18Osc values, however, are < −17‰, implausibly low
given that the paleosols must have formed near sea-level along the ancestral Ganges River and
the foreland reaches of its paleo-tributaries. The Dumri Formation at this location experienced
deep (> 8 km) tectonic burial under a thick thrust wedge of Greater Himalayan nappe rocks
(DeCelles et al. 1998), so resetting of δ18Osc values at elevated temperatures is unsurprising.
A nal example comes from the Nima basin located ~450 km northwest of Lhasa at
about 4500 m, in the southern part of the Bangong suture zone, which separates the Qiangtang
and Lhasa terranes in central Tibet (Fig. 1b). The southern Nima basin contains more than 4
km of Tertiary alluvial, uvial, lacustrine, and lacustrine fan-delta deposits that accumulated
next to growing thrust-faulted ranges (DeCelles et al. 2007b). Lacustrine marl beds and well-
developed calcareous paleosols are common in the informally designated Nima Redbed unit
-20
-18
-16
-14
-12
-10
-8
-6
-4
-2
0
-12 -10
-8
-6
-4
-2
024
6
δ
13C(PDB)carbonate
Penbo
limestone
Penbo
paleosol
expected
marine values
Figure 16. δ18Osc (PDB) versus δ13Csc (PDB) values from marine limestone and paleosol carbonate in the
late Cretaceous Takena Formation at the Penbo locality, southern Tibet (Leier 2005). Range of expected
marine isotopic values for the late Cretaceous from Veizer et al. (1999).
03_Quade_etal.indd 2703_Quade_etal.indd 27 7/3/2007 12:41:47 AM7/3/2007 12:41:47 AM
28 Quade, Garzione, Eiler
which is exposed in a roughly 50 km-long outcrop belt. For age control, 40Ar/39Ar dates from
six reworked tuffs in the Nima Redbed unit securely place the age of the analyzed deposits
between 25 and 26 Ma (late Oligocene) (DeCelles et al. 2007b).
We devised an “isotopic conglomerate test” in order to assess the potential for diagenetic
resetting of the isotopic paleoelevation signal (DeCelles et al. 2007a). The Nima Redbeds
contain numerous conglomerates interbedded with the marls and paleosols that we sampled for
paleoelevation reconstruction. Among other rock types, these conglomerates contain abundant
marine limestone clasts derived from the Lower Cretaceous Langshan Formation cropping out
around the basin. Where paleochannels carrying these gravels scour into contemporaneous
marls and paleosols, reworked paleosol carbonate nodules are also present in channel lags. The
limestone clasts are composed of both micrite and sparite and commonly contain obviously
recrystallized marine fossils. In contrast, the Tertiary lacustrine marls and nodular paleosol
carbonates are dense, well-indurated micrite containing well-preserved gastropod, ostracod,
and Chara fossils.
Our hypothesis is that if resetting by diagenesis has occurred in the Nima basin, carbonates
of all types should exhibit uniformly very low δ18Osc values (< −6‰), owing to diagenetic
interaction at higher temperatures with 18O-depleted meteoric water that is characteristic of
recharge in the region today. Critical here is that clasts of marine limestone should not retain
their primary marine δ18Osc values of > −5‰ (Veizer et al. 1999), as observed in the example
from the upper Cretaceous Takena Formation.
Analysis of the reworked Cretaceous marine limestone pebbles and cobbles yielded
several δ18Osc values > −3‰ (
Fig. 17
Fig. 17), in the range expected for unaltered Cretaceous marine
carbonates. This indicates that some of the limestone pebbles are not diagenetically reset despite
a long history of uplift, erosion, and burial over a time period of approximately 80 million
years. Other Cretaceous-age limestone pebbles yielded δ18Osc values ranging between −5‰
and −15‰, indicating that they have been diagenetically reset by interaction with meteoric
waters (Fig. 17). In sharp contrast, paleosol carbonate nodules—both reworked and in situ—in
the Nima basin yielded tightly clustered δ13Csc values of −3.5 ± 0.6‰ (range −2.3 to −4.6‰;
n = 31) and δ18Osc values of −17.0 ± 0.3‰ (range −16.3 to −17.5‰). The large spread in δ18O
values of the marine limestone clasts, as contrasted with the narrow range of values from in situ
and reworked Nima paleosol carbonate (Fig. 17), indicates that resetting of the limestone must
have occurred prior to deposition in the Nima basin, during previous burial of the limestone.
Thus, we interpret paleosol carbonate in the Nima basin as preserving the original isotopic
composition of meteoric water during the late Oligocene.
Results from marl and fossil Chara and ostracodes in section with the paleosols strongly
reinforce the view that the paleosol carbonates are preserving primary values. Marls range in
δ18O values between −16.5 to −8.9‰, whereas those from fossil Chara and ostracodes fall
between −18.1 to −4.7‰. The most negative values are very similar to those from pedogenic
carbonates. The large spread in δ18O values is consistent with periodic strong evaporation in
the Nima paleolake. The most positive δ18O values from fossils and marl are very similar to
carbonate forming in modern Tibetan lakes (Quade, unpublished data).
Paleoaridity. As already discussed, we can use δ13Csc values from paleosols as potential
archives of paleoaridity. Unlike δ18Osc values, δ13Csc values in carbonates are far less prone to
diagenetic resetting because of the very low carbon (versus high oxygen) content of natural
waters. Results from the upper Cretaceous Takena Formation cannot be used to reconstruct
paleoaridity, because of much higher pCO2 at that time (Ekart et al. 1999). However, results
from the late Oligocene to early Miocene Dumri Formation and Nima redbeds are admissible.
δ13Csc values from the Dumri Formation average −10.3 ± 0.6‰. These very low values
are similar to those obtained from the overlying Siwalik Formation in paleosols > 8 Ma. They
03_Quade_etal.indd 2803_Quade_etal.indd 28 7/3/2007 12:41:48 AM7/3/2007 12:41:48 AM
Paleoelevation Reconstruction using Pedogenic Carbonates 29
imply high respiration rates of 4-8 mmoles/m2/hr, assuming δ13C values of local plants between
−24 and −25‰. This in turn indicates moist climatic conditions, consistent with paleobotanic
evidence for the period (Lakhanpal 1970).
By contrast, δ13Csc values from Nima paleosols in central Tibet (Fig. 1b) formed at about
the same time as those in the Dumri Formation, but on the other side of the Himalaya, and
are much higher (−3.5 ± 0.6‰). These results are very close to the lowest δ13Csc values from
modern pedogenic carbonates from immediately around Nima, which range from −3.4 to
+7.7‰. The relatively high δ13Csc values for modern carbonates indicate mixing of atmospheric
CO2 with plant-derived CO2 owing to low soil respiration rates, a result of arid local climate of
~200-250 mm/yr. Similarly, the high δ13Csc values from paleosol carbonates indicate very low
soil respiration rates at 25-26 Ma (Fig. 6). Ancient soil respiration rates of 0.15 to 0.7 mmoles/
m2/hr can be calculated from the one-dimensional soil diffusion model of Cerling (1984) and
modi ed in (Quade et al. 2007), assuming model conditions given in the caption to Figure 6.
These respiration rates are very low and typical of sparsely vegetated, low-elevation settings in
the Mojave (Quade et al. 1989a). Aridity in modern Tibet is produced by orographic blockage
of moisture arriving from outside the plateau, as well as by orographically induced stationary
waves that tend to x dry, descending air over the plateau during most of the year (Broccoli
and Manabe 1992). The high δ13Ccc values from Nima basin paleosols indicate that orographic
barriers to moisture also existed during the late Oligocene, most likely the Himalaya, which was
actively growing during Oligocene time and earlier.
Paleoelevation. The δ18Osc values of modern pedogenic carbonate from around Nima (12
analyses, 3 pro les) ranges from −13.8‰ to −6.1‰. δ18Omw values of meteoric water samples
from springs and small creeks in the Nima area at modern elevations of 4500-5000 m range from
−12.6‰ to −16.2‰. Using local mean annual temperature of 3°C, δ18O values of pedogenic
carbonate forming in isotopic equilibrium with these waters should be −11.7 to −13.3‰. Thus,
Figure 17. δ18Osc (PDB) versus δ13Csc (PDB) values from marine limestone cobbles (variable ages) and
paleosol carbonate in the late Oligocene-age Nima Redbeds, southern Tibet (DeCelles et al. 2007a). Range
of expected marine isotopic values for the late Cretaceous from Veizer et al. (1999).
03_Quade_etal.indd 2903_Quade_etal.indd 29 7/3/2007 12:41:49 AM7/3/2007 12:41:49 AM
30 Quade, Garzione, Eiler
observed (−13.8 to −6.1‰) and predicted (−11.7 to −13.3‰) δ18Osc values overlap, but modern
carbonates exhibit much more positive values (Fig. 10). These high values are consistent with
the arid setting of Nima basin (annual rainfall is ~200 mm/yr).
The δ18Osc values of −17.0 ± 0.3‰ from paleosol nodules are much lower than δ18Osc
values of modern pedogenic carbonate from around Nima of −13.8‰ to −6.1‰ (Fig. 10). On
the face of it this would appear to imply higher paleoelevations in the past compared to the
present elevation of Nima at ~4500 m. However, this simple comparison is unrealistic because
we know that corrections for decreased ice volume, a warmer Earth, lower latitude at 26 Ma,
and likely changes in Asian monsoon intensity must be considered. Each of these factors can
be considered separately. Or, as in the case of several previous studies (Garzione et al. 2000a;
Rowley et al. 2001; Mulch et al. 2004; Currie et al. 2006; Rowley and Currie 2006), changes in
δ18Osc values from a nearby low-elevation site can be used to make the correction. The second
approach has great advantages, since it both “anchors” the temperature-δ18Osc relationship of the
paleoclimate system delivering the moisture to Tibet at some known elevation, in this case near
sea level, and it subsumes the effects on δ18Osc values of changes latitude, temperature, global
ice volume, and regional climate through time. Critical here is that—today at least— northern
India and southern Tibet share a common climate dominated by the SW Indian monsoon.
As with previous studies (Garzione et al. 2000a; Rowley et al. 2001), we can use the data
from the Siwalik Group (Quade and Cerling 1995) in the Himalayan foreland of Nepal and
Pakistan as our low-elevation reference site. These data only extend back to 17 Ma, since our
attempts to extend the lowland isotopic record to the Oligocene failed, as already described.
These records show increases in δ18Osc values in the late Miocene of about 2.5‰, comparing
the most negative values (not the means) of the pre-late Miocene pedogenic carbonates (Quade
and Cerling 1995). We have interpreted this late Miocene increase in δ18Osc values as related
to a weakening of the Asian monsoon (Dettman et al 2001), the opposite of earlier-held views
(Quade et al. 1989b; Kroon et al. 1991; Ruddiman et al. 1997). Causes aside, this ~2.5‰
“correction” can be added to δ18Osc values from paleosol carbonate (−17‰) at Nima to yield
about −14.5‰. This value is slightly lower than the lowest (least evaporated) modern pedogenic
carbonate value from Nima of around −13.8‰. On this basis we conclude that δ18Osc values of
paleosol carbonate—after correction—look very similar to modern values. This would imply
little change in elevation of the site since the late Oligocene.
Central to our reconstruction is that the δ18Osc record from Pakistan back to 17 Ma can be
applied to late Oligocene Nima, and the shift in the y-intercept of the system (Fig. 2b) of ~2.5‰
encapsulates the changes in global temperature, latitude, sea-water δ18O values, and monsoon
strength over the past 26 Ma. An additional (and related) assumption is that the slope of the
modern δ18Osc value-elevation relationship for southern Tibet has not appreciably changed back
through time.
The last assumption is an important one, because just viewing climatic patterns in Tibet
today, the slope of the δ18Osc value-elevation relationship in Tibet is highly dependent on
latitude, and should vary according to the strength of the Asian monsoon. For example, a
weaker monsoon at 26 Ma should reduce the slope of the δ18Osc value-elevation relationship
in southern Tibet—that is, northerly or local air masses would have had a stronger in uence
further south in the late Oligocene. This would have the effect of increasing our estimates
of paleoelevation compared to today. Alternatively, we might argue that the Asian monsoon
was stronger at 26 Ma, resulting in a locally steeper isotopic lapse rates with elevation and
overestimated paleoelevation. However, this reasoning is contradictory by invoking a stronger
Asian monsoon at 26 Ma in the presence of a lower plateau.
This brings us to our current “best guess” about the Nima record, that paleoelevation in the
Oligocene at Nima was at least as high as today, and no higher than today only if the modern
Asian monsoon—or something that coincidentally looks a lot like it in oxygen isotope terms
03_Quade_etal.indd 3003_Quade_etal.indd 30 7/3/2007 12:41:50 AM7/3/2007 12:41:50 AM
Paleoelevation Reconstruction using Pedogenic Carbonates 31
—was already in place at 26 Ma. We would stress, however, that Nima is only one of a hand-
full of well-dated Oligo-Miocene sites in Tibet (others examples are discussed in Garzione et
al. 2000b, Currie et al. 2005, and Rowley and Currie 2006), distributed over a very broad and
geologically diverse region. The challenge ahead is to see if our results from Nima and other
sites in southern Tibet can be replicated, and to eventually reconstruct the uplift history of the
rest of the plateau.
CONCLUDING REMARKS
Paleoelevation reconstruction is a relatively new science. Of the several approaches under
development, the use of oxygen isotopes in pedogenic carbonates enjoy several advantages.
As to soils themselves, pedogenic carbonate is relatively common in basin deposits of many
orogens, especially on the drier leeside. For another, soils form in the presence of local rainfall,
and hence are not in uenced by the δ18O value of run-off from higher elevations in the way
that, for example, riverine and lacustrine carbonates could be.
The δ18O analysis of pedogenic carbonate has the advantage of being routine but the major
disadvantage of being a function of both temperature and of the δ18O value of rainfall variably
modi ed by evaporation. Paleoelevation can also be seriously underestimated due to the effects
of evaporation. This effect can only be minimized by sampling deep in paleosol pro les and is
probably an insurmountable problem in very dry settings (< 200 mm/yr) like the Atacama and
low-elevation Mojave Desert. For this reason, Δ47 analyses hold much promise because they
are only temperature dependent. Here, a nal important advantage of soils comes into play:
soil T collapses to slightly above (+1 to +3 °C) mean annual T at soil depth. Other common
settings for carbonate formation such as oceans, lakes, and rivers can experience very large
temperature swings diurnally or seasonally, and do not converge on mean annual temperature
with depth. The combination of δ18O and Δ47 analysis is especially powerful since it provides
two largely independent estimates of paleoelevation. Δ47 measurements of pedogenic carbonate
allow mean annual temperature to be reconstructed, and paleoelevations to be estimated using
temperature lapse rates. These same paleotemperature estimates allow δ18Omw values to be
calculated from δ18Occ values, providing a second estimate of paleoelevation, assuming a local
isotopic lapse rate.
Diagenetic resetting poses a serious problem for future paleoelevation reconstructions
using carbonate, especially as we delve into the older and deeper geologic record where
resetting of δ18O values is well documented in many situations. Our Tibetan and Bolivian
studies offer two approaches to the problem, the Bolivian case where δ18O and Δ47 values
provide independent checks of each other, and the Tibetan case by analysis of co-occurring
marine limestones.
In the longer view, the δ18O and Δ47 combination entails the fewest assumptions and holds
the greatest promise for paleoelevation reconstruction. However, this potential will only be
fully realized when Δ47 analysis becomes more routine and can be calibrated and tested in very
young pedogenic carbonates formed in soils with well-constrained temperature histories.
ACKNOWLEDGMENTS
JQ thanks Paul Kapp and Pete DeCelles for involving him in their Tibetan research, and
discussions and data from Dave Dettman and Andrew Leier, and acknowledges support from
NSF-EAR-Tectonics 0438115. CG thanks David Rowley and Terry Jordan for insightful
discussions, and acknowledges support from NSF-EAR-Tectonics 0230232.
03_Quade_etal.indd 3103_Quade_etal.indd 31 7/3/2007 12:41:50 AM7/3/2007 12:41:50 AM
32 Quade, Garzione, Eiler
REFERENCES
Affek HP, Eiler JM (2006) Abundance of mass-47 CO2 in urban air, car exhaust and human breath. Geochim
Cosmochim Acta 70:1-12
Allison GB, Barnes CJ, Hughes MW (1983) The distribution of deuterium and 18O in dry soils 2. Experimental.
J Hydrol 64:377-397
Allison CE, Francey RJ, Meijer HAJ (1995) Recommendations for the reporting of stable isotope measurements
of carbon and oxygen in CO2 gas. In: IAEA-TECDOC-825, Reference and intercomparison materials for
stable isotopes of light elements. IAEA, Vienna. p 155-162
Ambach W, Dansgaard W, Eisner H, Mollner J (1968) The altitude effect on the isotopic composition of
precipitation and glacier ice in the Alps. Tellus 20:595-600
Amundson R, Chadwick O, Kendall C, Wang Y, DeNiro M (1996) Isotopic evidence for shifts in atmospheric
circulation patterns during the late Quaternary in mid-North America. Geology 24:23-26
Araguás-Araguás L, Froelich K, Rozanski K (1998) Stable isotope composition of precipitation over southeast
Asia. J Geophys Res 103:28,721-28,742
Aravena R, Suzuki O, Pollastri A (1989) Coastal fog and its relation to groundwater in IV region of northern
Chile. Chem Geol (Isotope Geos Sec) 79: 83-91
Aravena R, Suzuki O, Pena H, Pollastri A, Fuenzalida H, Grilli A (1999) Isotopic composition and origin of
precipitation in northern Chile. Appl Geochem 14:411-422
Bartlett MG, Chapman DS, Harris RN (2006) A decade of ground-air temperature tracking at Emigrant Pass
Observatory, Utah. J Clim 19:3722-3731
Beard KV, Pruppacher HR (1971) A wind tunnel investigation of the rate of evaporation of small water drops
falling at terminal velocity in air. J Atmos Sci 28:1455-1464
Berner RA, Kothavala Z (2001) Geocarb III: a revised model of atmospheric CO2 over Phanerozoic time. Am
J Sci 301:182-204
Bershaw J, Garzione CN, Higgins P, MacFadden BJ, Anaya F, Alveringa H (2006) The isotopic composition of
mammal teeth across South America: A proxy for paleoclimate and paleoelevation of the Altiplano: EOS,
Transactions, American Geophysical Union 87:Abstract #T33C-0519
Birkeland PW (1984) Soils and Geomorphology. Oxford University Press, New York
Blisniuk PM, Stern LA (2005) Stable isotope paleoaltimetry---a critical review. Am J Sci 305:1033-1074
Blisniuk PM, Stern LA, Chamberlain CP, Idleman B, Zeitler PK (2005) Climatic and ecologic changes during
Miocene surface uplift in the southern Patagonian Andes. Earth Planet Sci Lett 230:125-142
Broccoli AJ, Manabe S (1992) The effects of orography on midlatitude Northern Hemisphere dry climates. J
Climate 5:1181-1201
Came R, Veizer J, Eiler JM (in review) Surface temperatures of the Paleozoic ocean based on clumped isotope
thermometry. Nature: in review (UPDATE?)
Cerling TE (1984) The stable isotopic composition of modern soil carbonate and its relation to climate. Earth
Planet Sci Lett 71:229-240
Cerling TE, Quade J (1993) Stable carbon and oxygen isotopes in soil carbonates. In: Continental Indicators of
Climate. Swart P, McKenzie JA, Lohman KC (eds) Proceedings of Chapman Conference, Jackson Hole,
Wyoming, American Geophysical Union Monograph 78:217-231
Cerling TE, Harris JM, MacFadden BJ, Ehleringer JR, Leakey MG, Quade J, Eisenman V (1997) Global
vegetation change through the Miocene/Pliocene boundary. Nature 389:153-157
Chadwick OA, Sowers JM, Amundson RA (1988) In uence of climate on the size and shape of pedogenic
crystals. Soil Sci Soc Am J 53:211-219
Cross SL, Seltzer GO, Fritz SC, Dunbar RB (2001) Late Quaternary climate change and hydrology of tropical
South America inferred from an isotopic and chemical model of Lake Titicaca, Bolivia and Peru. Quat
Res 56:1-9
Currie BS, Rowley DB, Tabor NJ (2005) Middle Miocene paleoaltimetry of southern Tibet: implications for the
role of mantle thickening and delamination in the Himalayan orogen. Geology 33(3):181-184
Cyr AJ, Currie BS, Rowley DB (2005) Geochemical evaluation of Fenghuoshan Group Lacustrine carbonates,
north-central Tibet: implications for paleoaltimetry of the Eocene Tibetan Plateau. J Geol 113:517-533
Davidson GR (1995) The stable isotopic composition and measurement of carbon in soil CO2. Geochim
Cosmochim Acta 59(12):2485-2489
DeCelles PG, Horton BK (2003) Early to middle Tertiary foreland basin development and the history of Andean
crustal shortening in Bolivia. Geol Soc Am Bull 115(1):58-77
DeCelles PG, Gehrels G, Quade J, Ojha TP (1998) Eocene-early Miocene foreland basin development and the
history of Himalayan thrusting, western and central Nepal. Tectonics 17(5):741-765
DeCelles PG, Quade J, Kapp P, Fan M, Dettman DL, Ding L (2007a) High and dry in central Tibet during the
late Oligocene. Ear Planet Sci Lett 253:389-401
03_Quade_etal.indd 3203_Quade_etal.indd 32 7/3/2007 12:41:51 AM7/3/2007 12:41:51 AM
Paleoelevation Reconstruction using Pedogenic Carbonates 33
DeCelles PG, Kapp P, Ding L, Gehrels GL (2007b) Late Cretaceous to middle Tertiary basin evolution in the
central Tibetan Plateau: Changing environments in response to tectonic partitioning, aridi cation, and
regional elevation gain. Geol Soc Am Bull 119(5-6):654-680
Dettman DL, Kohn MJ, Quade J, Ryerson FJ, Ojha TP, Hamidullah S (2001) Seasonal stable isotope evidence
for a strong Asian monsoon throughout the past 10.7 Ma. Geology 29:31-34
Deutz P, Montanez IP, Moger HC (2002) Morphology and stable and radiogenic isotope composition of
pedogenic carbonates in late Quaternary relict soils, New Mexico, USA: an integrated record of pedogenic
overprinting. J Sediment Res 72:809-822
Dutton A, Wilkinson BH, Welker JM, Bowen G, Lohmann KC (2005) Spatial distribution and seasonal variation
in 18O/16O of modern precipitation and river water across the coterminous USA. Hydrological Processes
19:4121-4146
Eiler JM, Schauble E (2004) 18O13C16O in Earth’s atmosphere. Geochim Cosmochim Acta 68:4767-4777
Eiler JM, Garzione C, Ghosh P (2006) Response to comment on “Rapid uplift of the altiplano revealed through
13C-18O bonds in paleosol carbonates. Science 314:5800
Ekart DD, Cerling TE, Montanez IP, Tabor NJ (1999) A 400 million year carbon isotope record of pedogenic
carbonate: implications for paleoatmospheric carbon dioxide. Am J Sci 299: 805-827
Fox D, Koch PL (2004) Carbon and oxygen isotopic variability in Neogene paleosol carbonates: constraints on
the evolution of C4 grasslands. Palaeogr Palaeocl Palaeoecol 207:305-329
Friedman I, Smith GI, Gleason JD, Warden A, Harris JM (1992) Stable isotope composition of waters in
southeastern California 1. modern precipitation. J Geophys Res 97(D5):5795-5812
Fritz P, Suzuki O, Silva C, Salati E (1981) Isotope hydrology of groundwater in the Pampa del Tamarugal, Chile.
J Hydrol 53:161-184
Garzione CN, Quade J, DeCelles PG, English NB (2000a) Predicting paleoelevation of Tibet and the Himalaya
from δ18O versus altitude gradients in meteoric water across the Nepal Himalaya. Earth Planet Sci Lett
183:215-229
Garzione CN, Dettman DL, Quade J, DeCelles PG, Butler RF (2000b) High times on the Tibetan Plateau:
paleoelevation of the Thakkola graben, Nepal. Geology 28:339-342
Garzione CN, Dettman DL, Horton BK (2004) Carbonate oxygen isotope paleoaltimetry: evaluating the effect of
diagenesis on paleoelevation estimates for the Tibetan plateau. Palaeogeog Palaeoclim Palaeoec 212:119-
140
Garzione CN, Molnar P, Libarkin JC, MacFadden BJ (2006) Rapid late Miocene rise of the Bolivian Altiplano:
Evidence for removal of mantle lithosphere. Earth Planet Sci Lett 241:543-556
Ghosh P, Adkins J, Affek H, Balta B, Guo W, Schauble EA, Schrag D, Eiler JM (2006a) 13C-18O bonds in
carbonate minerals: A new kind of paleothermometer. Geochim Cosmochim Acta 70:1439-1456
Ghosh P, Eiler JM, Garzione C (2006b) Rapid uplift of the Altiplano revealed in abundances of 13C–18O bonds
in paleosol carbonate. Science 311:511-515
Gile LH, Peterson FF, Grossman RB (1966) Morphological and genetic sequences of carbonate accumulation
in desert soils. Soil Sci 101:347-360
Gon antini R, Stichler W, Rozanski K (1995) (ARTICLE TITLE?) In: IAEA-TECDOC-825, Reference and
intercomparison materials for stable isotopes of light elements. IAEA, Vienna. p. 13-30
Gon antini R, Roche M-A, Olivry J-C, Fontes J-C, Zupi GM (2001) The altitude effect on the isotopic
composition of tropical rains. Chem Geol 181:147-167
Guo W, Eiler JM (2007) Evidence for methane generation during the aqueous alteration of CM chondrites.
Meteoritics and Planetary Sciences in press (UPDATE?)
Hardy DR, Vuille M, Bradley RS (2003) Variability of snow accumulation and isotopic composition on Nevado
Sajama, Bolivia. J Geophys Res 108 (D22):4693
Harris N (2006) The elevation history of the Tibetan Plateau and its implications for the Asian monsoon.
Palaeogeogr Palaeoclim Palaeoecol 24:4-21
Hays PD, Grossman EL (1991) Oxygen isotopes in meteoric calcite cements as indicators of continental
paleoclimate. Geology 19:441-444
Hernández R, Jordan T, Dalenz Farjat A, Echavarría L, Idleman B, Reynolds J (2005) Age, distribution, tectonics
and eustatic controls of the Paranense and Caribbean marine transgressions in southern Bolivia and
Argentina. J South Am Earth Sci 19:495-512
Hillel D (1982) Introduction to Soil Physics. Academic Press, New York
Hoorn C (2006) The birth of the mighty Amazon. Scienti c American 294:52-59
Horton BK, Hampton BA, Lareau BN, Baldellon E (2002) Tertiary provenance history of the northern and
central Altiplano (Central Andes, Bolivia); a detrital record of plateau-margin tectonics, J Sedimentary
Res 72:711-726
Hsieh JCC, Chadwick O, Kelly E, Savin SM (1998) Oxyen isotopic composition of soil water: quantifying
evaporation and transpiration. Geoderma 82:269-293
Hulka C (2005) Sedimentary and tectonic evolution of the Cenozoic Chaco foreland basin, southern Bolivia (Ph.
D. dissertation), Freien Universität, Berlin
03_Quade_etal.indd 3303_Quade_etal.indd 33 7/3/2007 12:41:51 AM7/3/2007 12:41:51 AM
34 Quade, Garzione, Eiler
Jenny H, Leonard C (1934) Functional relationships between soil properties and rainfall. Soil Sci 38:363-381
Kay RF, MacFadden BJ, Madden RH, Sandeman H, Anaya F (1998) Revised age of the Salla beds, Bolivia, and
its bearing on the age of the Deseadan South American Land Mammal “Age.” J Vert Paleont 18:189-199
Kennan L, Lamb S, Rundle C (1995) K-Ar dates from the Altiplano and Cordillera Oriental of Bolivia -
implications for Cenozoic stratigraphy and tectonics. J South Am Earth Sci 8:163-186
Kent-Corson ML, Sherman LS, Mulch A, Chamberlain CP (2007) Cenozoic Topographic and climatic response
to changing tectonic boundary conditions in western North America. Earth Planet Sci Lett: in press
(UPDATE?)
Kim S-T, O’Neil JR (1997) Equilibrium and non-equilibrium oxygen isotope effects in synthetic carbonates.
Geochim Cosmochim Acta 61:3461-3475
Koch PL, Zachos JC, Dettman DL (1995) Stable-isotope stratigraphy and paleoclimatology of the paleogene
bighorn basin (Wyoming, USA). Palaeogeogr Palaeoclim Palaeoecol 115:61-89
Kroon D, Steens T, Troelstra SR (1991) Onset of monsoonal related upwelling in the western Arabian Sea as
revealed by planktonic foraminifers. In: Proceedings of the Ocean Drilling Project, Scienti c Results. Prell
WL, Niitsuma N (eds) 117:257-263
Lakhanpal RN (1970) Tertiary oras of India and their bearing on historical geology of the region. Taxon
19(5):675-694
Larrain H, Velazquez F, Cereceda P, Espejo R, Pinot R, Osses P, Schemenauer RS (2002) Fog measurements
at the site “Falde Verde” north of Chanaral compared with other fog stations of Chile. Atmos Res 64:273-
284
Latorre C, Quade J, McIntosh WC (1997) The expansion of C4 grasses and global change in the late Miocene:
stable isotope evidence from the Americas. Earth Plan Sci Lett 146:83-96
Leier AL (2005) The Cretaceous evolution of the Lhasa terrane, southern Tibet, Ph.D. Dissertation, Univ. of
Arizona
Liu B, Phillips FM, Campbell AR (1996) Stable carbon and oxygen isotopes of pedogenic carbonates, Ajo
Mountains, southern Arizona: implications for paleoenvironmental change. Palaeogr Palaeocl Palaeoecol
124:233-246
Machette MA (1985) Calcic soils of the southwestern United States. In: Soils and Quaternary Geology of the
Southwestern United States. Weide DL, Faber ML (eds) Geol Soc Am Spec Pap 203:1-21
MacFadden BJ, Campbell KE Jr, Cifelli RL, Siles O, Johnson NM, Naeser CW, Zeitler PK (1985) Magnetic
polarity stratigraphy and mammalian fauna of the Deseadan (late Oligocene early Miocene) Salla beds of
northern Bolivia. J Geol 93:223-250
Marshall LG, Swisher CC III, Lavenu A, Hoffstetter R, Curtis GH (1992) Geochronology of the mammal-
bearing late Cenozoic on the northern Altiplano, Bolivia. J South Am Earth Sci 5(1):1-19
Mulch A, Teyssier C, Cosca MA, Vanderhaeghe O, Vennemann V (2004) Reconstructing paleoelevation in
eroded orogens. Geology 32:525-528
Pagani M, Freeman KH, Arthur MA (1999) Late Miocene atmospheric CO2 concentrations and expansion of C4
grasses. Science 285:876-879
Pagani M, Zachos J, Freeman KH, Bohaty S, Tipple B (2005) Marked change in atmospheric carbon dioxide
concentrations during the Oligocene. Science 309:600-603
Pearson PN, Palmer MR (2000) Atmospheric carbon dioxide concentrations over the past 60 million years.
Nature 406:695-699
Poage MA, Chamberlain CP (2001) Empirical relationships between elevation and the stable isotope composition
of precipitation and surface waters: consideration for studies of paleoelevation change. Am J Sci 301:1-
15
Quade J, Cerling TE (1995) Expansion of C4 grasses in the late Miocene of northern Pakistan: evidence from
stable isotopes in paleosols. Palaeogr Palaeocl Palaeoecol 115:91-116
Quade J, Cerling TE, Bowman JR (1989a) Systematic variation in the carbon and oxygen isotopic composition
of Holocene soil carbonate along elevation transects in the southern Great Basin, USA. Geol Soc Am Bull
101:464-475
Quade J, Cerling TE, Bowman JR (1989b) Development of the Asian Monsoon revealed by marked ecological
shift in the latest Miocene in northern Pakistan. Nature 342:163-166
Quade J, Cater JML, Ojha TP, Adam J, Harrison TM (1995) Dramatic carbon and oxygen isotopic shift in
paleosols from Nepal and late Miocene environmental change across the northern Indian sub-continent.
Geol Soc Am Bull 107:1381-1397
Quade J, Rech J, Latorre C, Betancourt J, Gleason E, Kalin-Arroyo M (2007) Soils at the hyperarid margin:
the isotopic composition of soil carbonate from the Atacama Desert. Geochim Cosmochim Acta, in press
(UPDATE?)
Ramesh R, Sarin MM (1995) Stable isotope study of the Ganga (Ganges) river system. J Hydrol 139:49-62
Rech JA, Currie BS, Michalski G, Cowan AM (2006) Neogene climate change and uplift in the Atacama Desert,
Chile. Geology 34(9):761-764
03_Quade_etal.indd 3403_Quade_etal.indd 34 7/3/2007 12:41:51 AM7/3/2007 12:41:51 AM
Paleoelevation Reconstruction using Pedogenic Carbonates 35
Roe GH (2005) Orographic precipitation. Ann Rev Earth Planet Sci 33:645-671
Roperch P, Herail G, Fornari M (1999) Magnetostratigraphy of the Miocene Corque Basin, Bolivia; implications
for the geodynamic evolution of the Altiplano during the late Tertiary. J Geophys Res 104(9):20415-
20429
Rowley DB, Pierrehumbert RT, Currie BS (2001) A new approach to stable isotope-based paleoaltimetry:
implications for paleoaltimetry and paleohypsometry of the High Himalaya since the Late Miocene. Earth
Planet Sci Lett 188:253-268
Rowley DB, Currie BS (2006) Palaeo-altimetry of the late Eocene to Miocene Lunpola basin, central Tibet.
Nature 439:677-681
Rowley DB, Garzione CN (2007) Stable isotope-based Paleoaltimetry. Annu Rev Earth Planet Sci (in press)
(UPDATE?)
Royer D (1999) Depth to pedogenic carbonate horizon as a paleoprecipitation indicator? Geology 27:1123-
1126
Rozanski K, Sonntag C (1982) Vertical distribution of deuterium in atmospheric water vapour. Tellus 34:135-
141
Ruddiman WF, Raymo ME, Prell WL, Kutzbach JE (1997) The climate uplift-connection. In: Tectonic Uplift
and Climate Change. Ruddman WL (ed) Plenum Press. p 3-15
Salomans W, Goudie A, Mook WG (1978) Isotopic composition of calcrete deposits from Europe, Africa and
India. Earth Surf Processes 3:43-57
Santrock J, Studley SA, Hayes JM (1985) Isotopic analysis based on the mass-spectrum of carbon-dioxide. Anal
Chem 57:1444-1448
Savin SM, Douglas RG, Stehli FG (1975) Tertiary marine paleotemperatures. Geol Soc Am Bull 86:1499-1510
Schauble EA, Ghosh P, Eiler JM (2006) Preferential formation of 13C-18O bonds in carbonate minerals, estimated
using rst-principles lattice dynamics. Geochim Cosmochim Acta 70:2510-2529
Siegenthaler U, Oeschger, H (1980) Correlation of 18O in precipitation with temperature and altitude. Nature
285:314-17
Smith AG, Hurley AM, Briden JC (1981) Phanerozoic paleocontinental world maps, Cambridge, UK, Cambridge
University Press, 102 p
Stern, LA, Chamberlain, CP, Reynolds, RC, Johnson GD (1997) Oxygen isotope evidence of climate changes
from pedogenic clay minerals in the Himalayan molasses. Geochim Cosmochim Acta 61:731-744
Stern LA, Blisniuk PM (2002) Stable isotope composition of precipitation across the southern Patagonian
Andes. J Geophys Res 107(D23): doi:10.1029/ 2002JD002509
Stewart MK (1975) Stable isotope fractionation due to evaporation and isotopic exchange of falling waterdrops—
applications to atmospheric processes and evaporation of lakes. J Geophys Res 80:1133-1146
Talbot MR (1990) A review of the paleohydrological interpretation of carbon and oxygen isotopic ratios in
primary lacustrine carbonates. Chem Geol 80:261-279
Talma AS, Netterberg F (1983) Stable isotope abundances in calcretes. In: Residual Deposits: Surface Related
Weathering Processes and Materials. Wilson, RCL (ed) Oxford: Blackwell Scienti c Publ., p 221-233
Tian L, Masson-Delmotte V, Stievenard M, Tao T, Jouzel J (2001) Tibetan Plateau summer monsoon northward
extent revealed by measurements of water stable isotopes. J Geophys Res 106:28,081-28,088
Tian L, Yao T, White JWC, Yu W, Wang N (2005) Westerly moisture transport to the middle of Himalayas
revealed from the high deuterium excess. Chinese Sci Bull 50:1026-130
Uba CE, Heubeck C, Hulka C (2006) Evolution of the late Cenozoic Chaco foreland basin, southern Bolivia.
Basin Res 18:145-170
Veizer J, Ala D, Azmy K, Bruckschen P, Buhla D, Bruhn F, Carden GAF, Diener A, Ebneth S, Godderis Y, Jasper
T, Korte C, Pawellek F, Podlaha OG, Strauss H (1999) 87Sr/86Sr, δ13C and δ18O evolution of Phanerozoic
seawater. Chem Geol 161:59-88
Vuille M, Bradley RS, Werner M, Healy R, Keimig F (2003) Modeling δ18O in precipitation over the
tropical Americas: 1. Interannual variability and climatic controls. J Geophys Res 108: doi:10.1029/
2001JD002038
Wang ZG, Schauble EA, Eiler JM (2004) Equilibrium thermodynamics of multiply substituted isotopologues of
molecular gases. Geochim Cosmochim Acta 68:4779-4797
Wolfe BB, Aravena R, Abbott MB, Seltzer GO, Gibson JJ (2001) Reconstruction of paleohydrology and
paleohumidity from oxygen isotope records in the Bolivian Andes. Palaeogeogr Palaeoclimatol Palaeoecol
176(1-4):177-192
Wynn JG (2004) In uence of Plio-Pleistocene aridi cation on human evolution: evidence from paleosols of the
Turkana Basin. Am J Phys Anthro 123:106-118
Zachos J, Pagani M, Sloan L, Thomas E, Billups K (2001) Trends, rhythms, and aberrations in global climate
65 Ma to present. Science 292:686-693
Zhang X, Nakawo M, Yao T, Han J, Xie Z (2002) Variations of stable isotopic compositions in precipitation on
the Tibetan Plateau and its adjacent regions. Science in China (Series D) 25(6):481-493
03_Quade_etal.indd 3503_Quade_etal.indd 35 7/3/2007 12:41:52 AM7/3/2007 12:41:52 AM
36 Quade, Garzione, Eiler
03_Quade_etal.indd 3603_Quade_etal.indd 36 7/3/2007 12:41:52 AM7/3/2007 12:41:52 AM
