Page 24 - i1052-5173-27-2

DISCUSSION
The most prominent trend in our record
is the 4‰ decrease in d18Oc that occurs in
the late Neogene. This trend is opposite to
nearly all other d18Oc records in Central
Asia that lie downwind of the Zaysan
Basin and either increase or remain approxi- changes observed in Central Asia (Fig. 3)
mately constant over the Neogene (Fig. 6)
(Caves et al., 2015). Notably, the only
other records that decrease during the
Neogene are those that also lie on the
windward flank of the Tian Shan (Issyk
Kul) and/or within the modern-day spring
and fall precipitation regime (Junggar
Basin) (gray stars; Fig. 2B).
What mechanisms might explain a
simultaneous decrease in d18O on the
windward flanks of the Tian Shan and
Altai, while permitting constant or
increasing d18O in their lee?
One potential mechanism is that
increased orographic precipitation due to
uplift might shift windward d18O to lower
values, particularly as high-elevation pre-
cipitation is increasingly captured (Mulch,
2016). This hypothesis has been used to
explain the decrease in d18O at both Issyk
Kul and the Junggar Basin (Charreau et
al., 2012; Macaulay et al., 2016). However,
increased orographic precipitation appears
unlikely to explain the full decrease in the the interaction between cyclones routed
Zaysan Basin. The long-term increase in
d13C in the Zaysan Basin—at a location
well north of where abundant C4 vegeta-
tion is typically found (supplemental data
Fig. S3)—indicates that soil respiration
rates dropped, suggesting a decline in
productivity and, hence, precipitation
(Breecker et al., 2009; Caves et al., 2016). both. The mean position of the jet is
Further, increased orographic precipita-
tion should also result in coupled d18O
decreases in the lee of these ranges, which
is not observed, though d18Oc in lee basins
is complicated by evaporative effects
(Mulch, 2016). An additional mechanism
could be an increase in the formation tem-
perature of pedogenic carbonate, given
that the fractionation of 18O between water
and calcite decreases as temperature
increases (~-0.2‰/°C) (Kim and O’Neil,
1997). However, an increase in tempera-
ture is unlikely because global climate
cooled in the late Neogene following the
Miocene Climatic Optimum (Zachos et
al., 2001). Further, if changes in global
temperature were the dominant factor
driving changes in d18Oc, one would
expect similar decreases across much of
24 GSA Today | February 2017
northern Central Asia, which are not
observed (Caves et al., 2015).
In contrast to the above two mechanisms,
shifting precipitation seasonality in
Kazakhstan might have a substantial effect
on d18Oc given the large seasonal d18Op
(Araguás-Araguás et al., 1998). Pedogenic
carbonates are seasonally biased recorders
of d18Op, typically recording wet-season
d18Op (Breecker et al., 2009). Thus, a shift
toward dominantly spring and fall precipi-
tation might change the timing of pedogenic
carbonate formation and decrease d18Oc,
without necessarily increasing total precipi-
tation. Such a shift may have also lowered
the temperature of carbonate formation,
raising d18Oc, and partially offsetting the true Issyk Kul, which lies 5° south, decline
change in precipitation-weighted d18Op;
this implies that the change in precipitation
seasonality may have been substantial.
Importantly, this change in precipitation
seasonality would decouple windward d18Oc
records (Zaysan, Issyk Kul, and Junggar)
from leeward d18Oc records in interior China
and Mongolia, which receive dominantly
JJA precipitation.
IMPLICATIONS
The modern spring and fall precipita-
tion regime in Kazakhstan is a result of
along the mid-latitude westerly jet as it
migrates seasonally north and south and
the high topography of the Tian Shan and
Altai (Schiemann et al., 2009). Thus, a
change in precipitation seasonality would
imply either a shift of the mid-latitude jet,
the development of high topography, or
thought to have been farther northward
during the late Miocene given warmer
Arctic temperatures (Micheels et al., 2011),
suggesting that cooling into the late
Miocene may have contributed to a south-
ward shift of the jet.
Further, northward growth of high
topography in Asia would result in
increased interaction between cyclones
routed along the jet and this topography,
resulting in orographic precipitation dur-
ing the spring and fall (Baldwin and
Vecchi, 2016). There is substantial evi-
dence that deformation from the collision
of India-Asia has propagated northward
during the Cenozoic, resulting in—most
recently in the latest Miocene—uplift of
the Altai range (De Grave et al., 2007).
The Altai rain shadow likely formed by
the latest Miocene, as indicated by a sub-
stantial increase in pedogenic carbonate
d13C in the lee of the Altai (Caves et al.,
2014) (Fig. 6). In the northern Tian Shan,
Charreau et al. (2009) found increased
sedimentation rates in the Junggar Basin
at ca. 11 Ma, which they proposed resulted
from an acceleration of uplift. Though the
individual uplift histories of the local
Saur-Manrak ranges are unknown, but are
likely linked to uplift in the Altai or north-
ern Tian Shan (Campbell et al., 2013), we
propose that the timing of this accelerated
uplift is broadly synchronous with our
observed decrease in d18Oc. Northward-
propagating deformation might also
explain why oxygen isotopic records from
starting in the early Miocene, reflecting
earlier growth of high topography to the
south (Macaulay et al., 2016).
Our new oxygen isotope record from
the windward side of the Tian Shan and
Altai ranges indicates that these ranges
were sufficiently elevated by the late
Miocene to impact climate in Central
Asia. Most notably, this interaction
resulted in a substantial reorganization of
Central Asia climate, creating a stark pre-
cipitation seasonality boundary between
eastern and western Central Asia. Such
high topography would have further
blocked moisture from reaching down-
wind Central Asia, contributing to the
increasingly arid conditions and formation
of the Taklamakan and Gobi deserts in the
late Miocene (Caves et al., 2016; Sun et al.,
2009). The Altai are the largest source of
lee cyclones in Asia (Chen et al., 1991),
and the combined interaction of high
topography with the jet’s oscillations
as it moves northward in the spring
(Schiemann et al., 2009) creates the dust
storms that blanket the Loess Plateau
(Shao and Dong, 2006; Roe, 2009). Thus,
acceleration of loess deposition in the late
Miocene (Zhang et al., 2014; Rea et al.,
1998) may be a consequence of increasing
cyclogenesis as a result of upwind topo-
graphic growth (Caves et al., 2014; Shi et
al., 2015) (Fig. 6). The subsequent
Pliocene decline in eolian accumulation
rates may reflect a northward shifted jet
during the mid-Pliocene Warm Period,
which precluded interaction with the Altai
until Quaternary cooling shifted the jet
southward again.