STYLE>
Reply to Comment by Roberta M. Hotinski, Lee R. Kump, and
Karen L. Bice on Could the Late Permian deep ocean have
been anoxic?
R. Zhang,
1
M. J. Follows, and J. Marshall
Department of Earth, Atmospheric and Planetary Sciences, Massachusetts Institute of Technology, Cambridge,
Massachusetts, USA
Received 21 September 2002; revised 27 May 2003; accepted 1 August 2003; published 18 December 2003.
I
NDEX
T
ERMS
: 4805 Oceanography: Biological and Chemical: Biogeochemical cycles (1615); 4267 Oceanography: General:
Paleoceanography; 4255 Oceanography: General: Numerical modeling; 4802 Oceanography: Biological and Chemical: Anoxic
environments; 4845 Oceanography: Biological and Chemical: Nutrients and nutrient cycling; K
EYWORDS
: Permian, anoxia, phosphate,
extinction, stagnation, ocean modeling
Citation:
Zhang, R., M. J. Follows, and J. Marshall, Reply to Comment by Roberta M. Hotinski, Lee R. Kump, and Karen L. Bice on
Could the Late Permian deep ocean have been anoxic?, Paleoceanography, 18(4), 1095, doi:10.1029/2002PA000851, 2003.
[
1
] Two recent studies [Hotinski et al., 2001; Zhang et
al., 2001] have modeled and discussed Late Permian ocean
circulation and oxygenation. These studies have reached
significantly different conclusions. Zhang et al. [2001] find
that a thermal mode ocean circulation driven by cooling
in polar latitudes is unlikely to support deep-sea anoxia,
but a haline mode ocean circulation, a shallow over-
turning cell driven by enhanced evaporation from the
subtropics, perhaps could lead to global-scale deep-sea
anoxia. Hotinski et al. [2001], using a different model,
find that a thermal mode ocean circulation with a high
equator-to-pole gradient gives well-oxygenated Permian
deep water, while a thermal mode ocean circulation
with a very low equator-to-pole gradient leads to global
deep-sea anoxia. This discrepancy prompted a comment by
Hotinski et al. [2002].
[
2
] In response, we would like to begin by emphasizing
the common underlying feature of the two studies. In
both models, deep ocean oxygen concentrations are sig-
nificantly reduced if the overturning circulation becomes
weaker with warmer, less oxygenated waters ventilating
the deep ocean. In some circumstances this can lead to a
dominance of respiration of organic matter and a very
significant depletion of deep ocean oxygen. However, as
noted above, the two studies differ greatly in the nature of
the ocean circulation in which such a scenario can be
achieved.
[
3
] Hotinski et al. [2002] argue that the difference be-
tween the two model results are in large part the result of
dissimilar definitions of weak or reduced meridional
temperature gradients. We agree that the two studies apply
different pole-equator temperature gradients at the surface.
The high-latitudes surface temperature used to force the
model in our study [Zhang et al., 2001, Figure 5b] is much
warmer than that of modern climate and is based on an
atmospheric circulation model of the Late Permian
[Kutzbach and Gallimore, 1989], corresponding to a warm
climate with enhanced greenhouse effect (5
 CO
2
) and
reduced solar luminosity (1% decrease). The high-latitudes
surface temperature is more than 10 C (15 C) warmer than
modern climate at 70 N (70 S). The warmer high-latitudes
surface temperature is also similar to that of another
modeling study of the Late Permian climate [Kutzbach
and Ziegler, 1993], and is consistent with the Late Permian
paleobotanical data interpreted by Rees et al. [1999]. The
tropical surface temperature is about 3 C warmer than
present, resulting in pole-equator temperature gradient that
is weaker than that of modern climate. On the other hand,
our modeled Late Permian polar sea surface temperature
(SST) is cool relative to of the low-gradient scenario of
Hotinski et al. [2001] forced by the estimated Paleocene-
Eocene thermal maximum (PETM) SST [Bice and Marotzke,
2001].
[
4
] If we use much higher polar surface temperature to
force the model to give the estimated PETM polar SST as
in the work by Hotinski et al. [2001], the thermal mode
ocean circulation would be weakened. However, our recent
theoretical study [Zhang et al., 2002] showed that when the
pole-equator surface temperature gradient is reduced to a
certain threshold, the weakened thermal mode ocean
circulation would switch to the haline mode circulation,
and the critical freshwater flux required for the ocean to
switch to the haline mode circulation is much smaller with
the reduced pole-equator surface temperature gradient. In
that case, the deep ocean oxygen distribution would be
similar to what we found with the haline mode circulation.
Now the issue is when the pole-equator surface temperature
gradient is larger than the threshold of mode switching,
whether the weakened thermal mode circulation (with a
maximum of
$25 Sv as shown by Hotinski et al. [2001])
originating from warm polar surface water can drive global
deep-sea anoxia. We note that the weakened thermal mode
PALEOCEANOGRAPHY, VOL. 18, NO. 4, 1095, doi:10.1029/2002PA000851, 2003
1
Now at Geophysical Fluid Dynamics Laboratory/Atmospheric and
Oceanic Sciences Program, Princeton University, Princeton, New Jersey,
USA.
Copyright 2003 by the American Geophysical Union.
0883-8305/03/2002PA000851$12.00
19 -
1 circulation of Hotinski et al. [2001, Figure 2d] is of
comparable strength to that in the modern North Atlantic
Ocean. Why then, is the modern North Atlantic deep ocean
at depths of 3000 m to 3500 m (which is mainly dominated
by the North Atlantic deep water) so well oxygenated
(
$246 m mol L
À1
), whereas wide-spread, deep-sea anoxia
is sustained in the late Permian scenario of Hotinski et al.
[2001] in a circulation of similar vigor? The difference in
oxygen solubility due to changes in the temperature at the
deep water source region partially accounts for this. The
warm polar sea surface temperature (12 C) of Hotinski et
al. [2001] results in lower oxygen concentration (
$250 m
mol L
À1
) at the deep water source region. However,
comparing the modern North Atlantic value deep water
source region of
$336 m mol L
À1
, the oxygen solubility
effect can only account for
$86 m mol L
À1
reduction in
oxygen concentration at the deep water source region, far
from sufficient to induce deep-sea anoxia.
[
5
] Zhang et al. [2001] suggested that use of negative
oxygen by Hotinski et al. [2001] leads to the impression
of widespread anoxia in their well-ventilated, slow,
Permian Ocean circulation. Hotinski et al. [2002] argue
that allowing negative oxygen concentrations represents
the transport and oxidation of H
2
S, a product of the
oxidation of organic matter by sulfate. The negative
oxygen approach enhances the potential for anoxia, rela-
tive to the scheme used by Zhang et al. [2001], where we
chose to limit oxygen to nonnegative concentrations. In a
simulation of the modern ocean using our model in which
negative oxygen values are permitted, we find very exten-
sive regions where the dissolved oxygen is negative in the
Pacific and Indian oceans. However, these regions are not
anoxic in the present ocean. Hence the interpretation of
similar regions as anoxic condition in a Permian simula-
tion with the negative oxygen approach must be viewed
with extreme caution. In addition, in the absence of
oxygen, the primary oxidant of particulate organic material
is nitrate and significant denitrification of the ocean could
occur. This could lead to extensive nitrogen limitation of
new and export production [e.g., Falkowski, 1997] before
sulfate becomes the primary oxidant. Besides, oxidation of
H
2
S must not be necessarily by use of elemental O
2
.
Because of these complications, Zhang et al. [2001] chose
not to allow oxygen concentrations to become negative,
leading, we feel, to a clearer interpretation. However, we
recognize that the complex interactions under anoxic
condition must ultimately be understood and modeled in
more detail. With different parameterizations of the anoxic
condition in the biogeochemical models of the two studies,
it is difficult to compare the results directly. Further
investigations on the processes under the anoxic condition
in the real ocean and their parameterizations in the
biogeochemical models are deserved.
[
6
] Clearly the choice of physical and biogeochemical
parameterizations and boundary conditions in both models
strongly impacts upon the results and inferences. Our
studies illustrate some of the range of possibilities, though
by no means all. We have been stimulated by these
contrasting inferences and look forward to continued prog-
ress and future dialogue.
References
Bice, K. L., and J. Marotzke, Numerical evi-
dence against reversed thermohaline circula-
tion in the warm Paleocene/Eocene ocean,
J. Geophys. Res., 106, 11,529 11,542, 2001.
Falkowski, P., Evolution of the nitrogen cycle
and its influence on the biological sequestra-
tion of CO
2
in the ocean, Nature, 387, 282
285, 1997.
Hotinski, R., K. L. Bice, L. R. Kump, R. G.
Najjar, and M. A. Arthur, Ocean stagnation and
end-Permian anoxia, Geology, 29, 7 10, 2001.
Hotinski, R. M., L. R. Kump, and K. L. Bice,
Comment on Could the Late Permian deep
ocean have been anoxic? by R. Zhang et al.,
Paleoceanography, 17(4), 1052, doi:10.1029/
2001PA000680, 2002.
Kutzbach, J. E., and R. G. Gallimore, Pangean
climates: Megamonsoons of the megaconti-
nent, J. Geophys. Res., 94, 3341 3357,
1989.
Kutzbach, J. E., and A. M. Ziegler, Simulation of
Late Permian climate and biomes with an
atmosphere/ocean model: Comparisons with
observations, Philos. Tra