Ocean Stagnation and end Permian anoxia

Abstract

Ocean stagnation has been invoked to explain the widespread occurrence of organic-carbon-rich, laminated sediments interpreted to have been deposited under anoxic bottom waters at the time of the end-Permian mass extinction. However, to a first approximation, stagnation would severely reduce the upwelling supply of nutrients to the photic zone, reducing productivity. Moreover, it is not obvious that ocean stagnation can be achieved. Numerical experiments performed with a three-dimensional global ocean model linked to a biogeochemical model of phosphate and oxygen cycling indicate that a low equator to pole temperature gradient could have produced weak oceanic circulation and widespread anoxia in the Late Permian ocean. We find that polar warming and tropical cooling of sea-surface temperatures cause anoxia throughout the deep ocean as a result of both lower dissolved oxygen in bottom source waters and increased nutrient utilization. Buildup of quantities of H2S and CO2 in the Late Permian ocean sufficient to directly cause a mass extinction, however, would have required large increases in the oceanic nutrient inventory.

Full text

q2001 Geological Society of America. For permission to copy, contact Copyright Clearance Center at www.copyright.com or (978) 750-8400.
Geology; January 2001; v. 29; no. 1; p. 7–10; 3 figures; Data Repository item 20016. 7
Ocean stagnation and end-Permian anoxia
Roberta M. Hotinski* Department of Geosciences, Pennsylvania State University, University Park, Pennsylvania 16802, USA
Karen L. Bice Department of Geology and Geophysics, Woods Hole Oceanographic Institution, Woods Hole, Massachusetts
02543, USA
Lee R. Kump Department of Geosciences, Pennsylvania State University, University Park, Pennsylvania 16802, USA
Raymond G. Najjar Department of Meteorology, Pennsylvania State University, University Park, Pennsylvania 16802, USA
Michael A. Arthur Department of Geosciences, Pennsylvania State University, University Park, Pennsylvania 16802, USA
ABSTRACT
Ocean stagnation has been invoked to explain the widespread
occurrence of organic-carbon–rich, laminated sediments interpret-
ed to have been deposited under anoxic bottom waters at the time
of the end-Permian mass extinction. However, to a first approxi-
mation, stagnation would severely reduce the upwelling supply of
nutrients to the photic zone, reducing productivity. Moreover, it is
not obvious that ocean stagnation can be achieved. Numerical ex-
periments performed with a three-dimensional global ocean model
linked to a biogeochemical model of phosphate and oxygen cycling
indicate that a low equator to pole temperature gradient could
have produced weak oceanic circulation and widespread anoxia in
the Late Permian ocean. We find that polar warming and tropical
cooling of sea-surface temperatures cause anoxia throughout the
deep ocean as a result of both lower dissolved oxygen in bottom
source waters and increased nutrient utilization. Buildup of quan-
tities of H
2
S and CO
2
in the Late Permian ocean sufficient to di-
rectly cause a mass extinction, however, would have required large
increases in the oceanic nutrient inventory.
Keywords: Permian, anoxia, ocean circulation, phosphate cycle, oxy-
gen cycle.
INTRODUCTION
A number of authors have proposed that sluggish or stagnant
ocean circulation contributed to anoxia and the creation of large chem-
ical gradients in ancient oceans (e.g., Fischer and Arthur, 1977; Bra-
lower and Thierstein, 1984; Holser and Magaritz, 1987; Malkowski et
al., 1989; Gruszczynski et al., 1992; Kajiwara et al., 1994; Isozaki,
1997). Geochemical evidence suggests that the Permian-Triassic
boundary interval was such a period. Studies of boundary interval sed-
iments reveal large negative excursions in carbon, sulfur,and strontium
isotopic compositions of surface waters, which have been interpreted
as evidence of chemically distinct deep waters upwelling into a pre-
viously isolated surface ocean (Holser and Magaritz, 1987; Gruszcyn-
ski et al., 1992; Malkowski et al., 1994; Kajiwara et al., 1994; Knoll
et al., 1996). In addition, widespread laminated, pyritic sediments in
Upper Permian and Lower Triassic sections (Wignall and Hallam,
1992, 1993; Wignall and Twitchett, 1996; Isozaki, 1997) suggest that,
unlike today, decomposition of organic matter exhausted oxygen sup-
plied by circulation throughout much of the water column. The mag-
nitude and global nature of the excursions and anoxia led these inves-
tigators to propose that the entire ocean was severely stratified prior to
the boundary, and that stagnation-induced anoxia may have played a
role in the end-Permian extinction.
A potential flaw in this hypothesis is neglect of the relationship
between ocean circulation and surface productivity. While ocean mix-
ing acts to destroy vertical gradients in oxygen and other chemical
species, it also returns nutrients to the surface that fuel the biological
activity and organic-matter decomposition required to sustain gradients.
*Present address: Atmospheric and Ocean Sciences Program, Princeton
University, Princeton, New Jersey 08544, USA. E-mail: hotinski@princeton.
edu.
Quantifying the results of this competition between circulation and
biology thus requires a model that accounts for both ocean physics and
a dynamic link between nutrients and productivity. Although Permian
ocean circulation has been studied quantitatively in the past (Kutzbach
et al., 1990), this is the first attempt to treat the Permian ocean’s bio-
geochemistry explicitly.
We simulate a stagnation scenario for the Late Permian ocean
using a three-dimensional ocean general circulation model that includes
a simple biogeochemical model of phosphate and oxygen cycling. To
assess whether ocean stagnation would have created the chemical pat-
terns postulated for this time period, we apply a low equator-to-pole
temperature gradient and examine the changes in circulation and chem-
ical structure of the ocean that result.
MODEL DESCRIPTION
The general circulation model (GCM) used to simulate Permian
ocean circulation is the Geophysical Fluid Dynamics Laboratory’s
Modular Ocean Model (MOM) (Pacanowski et al., 1993). The model
configuration used includes 48348resolution, 16 vertical levels, and
constant vertical and horizontal mixing coefficients (1 cm
2
/s and 2 3
10
7
cm
2
/s, respectively). The ocean model bathymetry is a simplified
flat-bottom case (5150 m depth) with continental boundaries that ap-
proximate the land and sea distribution in the Late Permian (Wordian)
paleogeographical reconstruction of Rees et al. (1999). Further details
of the model, including additional simulations and model limitations,
can be obtained elsewhere.
1
For this study, the MOM code was modified to include simple
ocean biogeochemistry. Export of organic matter from the euphotic
zone (100 m thick) is modeled using the relationship of Yamanaka and
Tajika (1996). The export flux is proportional to surface-water phos-
phate concentration and varies with the cosine of latitude to simulate
light limitation with increasing latitude. Phosphate, rather than nitrate,
was chosen as the driver of productivity because phosphate is thought
to exert the primary control on marine primary production on long
time scales (e.g., Tyrrell, 1999). The exported organic flux is instan-
taneously remineralized below the euphotic zone according to the pow-
er law of Martin et al. (1987). Any flux that reaches the lowest vertical
layer of the ocean is remineralized, consistent with observations that
only a small fraction of the flux reaching the sediments is buried.
We assume that organic-matter decomposition will proceed via
denitrification and sulfate reduction after oxygen is depleted, and spec-
ify that the rate of phosphate remineralization is independent of oxygen
concentration in our model. However, because we do not include nitrate
and sulfate as tracers, such remineralization is represented by negative
values of dissolved O
2
. Although the biogeochemical model greatly
simplifies the cycling of organic matter, it produces realistic distribu-
tions of phosphate and oxygen in the modern ocean (Yamanaka and
Tajika, 1996).
To examine the effects of a reduced latitudinal temperature gra-
1
GSA Data Repository item 20016, Additional model description and pa-
rameters, is available on request from Documents Secretary, GSA, P.O. Box
9140, Boulder, CO 80301, editing@geosociety.org or at www.geosociety.org/
pubs/ft2001.htm.

8 GEOLOGY, January 2001
Figure 1. Model meridional (A) sea-surface temperature gradients and (B) surface density gradients for modern (A only), Permian high-
gradient, and Permian reduced-gradient scenarios.
dient on the Permian ocean, the model was run with two sets of cli-
matic forcings. A high-gradient forcing case used zonally averaged,
mean annual temperature, salinity, and wind forcings predicted by a
GENESIS version 2 simulation for the Late Permian (Wordian) that
included Late Permian paleogeography, 2760 ppm CO
2
(about eight
times the present atmospheric concentration), and 97.9% of the modern
solar luminosity (Rees et al., 1999). The temperature gradient predicted
by this simulation (Fig. 1A) is similar to the modern zonal average
gradient. The ocean model was run for 2700 yr with the high-gradient
forcing until a vigorous circulation was established.
To simulate climatic warming, the ocean model run wascontinued
with the same wind-stress and salinity forcing boundary conditions as
in the high-gradient case, but a new lower temperature gradient was
applied. Sea-surface temperatures (SSTs) for this simulation were fixed
at an equator-to-pole gradient of 12–28 8C (Fig. 1A), modeled after
that of the Paleocene-Eocene boundary interval (Zachos et al., 1994;
Bice et al., 2000). This gradient causes a substantial reduction in the
ocean’s pole to equator surface density gradient (Fig. 1B). Paleocli-
matological evidence (Taylor et al., 1992) suggests that the Late Perm-
ian southern polar climate was comparable to Paleocene-Eocene Arctic
climate, temperatures being .10 8C for at least a third of the year
(Ziegler, 1990; Yemane, 1993). In addition, oxygen isotope and paleo-
sol data indicate warming from the latest Permian into the earliest
Triassic (Holser et al., 1991; Retallack, 1999).
RESULTS
Circulation
The steady-state meridional overturning (zonally integrated mass
transport stream function in the meridional vertical plane) for the high-
gradient scenario is shown in Figure 2A. The pattern is asymmetrical;
there is a strong cell in the Southern Hemisphere and a weaker cell in
the Northern Hemisphere. The maximum Southern Hemisphere trans-
port value is .80 Sverdrups (1 Sv 51310
6
m
3
s
2
1
) and the circu-
lation can be characterized as vigorous.
This vigorous circulation is disrupted when the reduced temper-
ature gradient is applied (Fig. 2). At 100 yr, the ocean is poorly mixed
and is dominated below 1000 m by Northern Hemisphere sinking. By
1200 yr, weak mixing between the surface and deep ocean is reestab-
lished. After 10 k.y., circulation has reached a new steady state with
substantially reduced overturning relative to the high-gradient case,
consistent with the reduction in upper ocean density between the sim-
ulations. The steady state thermohaline circulation, driven by a weaker,
but largely symmetrical, density contrast (Fig. 1B), exhibits much less
interhemispheric asymmetry than the high-gradient circulation (Fig.
1A). Although circulation is reduced, bottom water is still formed at
high latitudes in both hemispheres and there is no low-latitude deep-
water formation. Northern Hemisphere deep waters are ventilated at a
rate nearly comparable to that of modern North Atlantic Deep Water,
estimated to be ;20 Sv (Broecker, 1991).
Dissolved Oxygen
The high-gradient case is characterized by high values of dis-
solved O
2
in the high latitudes and in deep water (Fig. 3). The ocean
is oxic everywhere except in intermediate waters off the western coast
of Pangea (Fig. 3A), where negative values of O
2
indicate oxidation
of organic matter with electron acceptors other than oxygen (i.e., nitrate
and sulfate). The deep ocean is particularly well oxygenated (Fig. 3B);
minimum O
2
values are .150 mmol/L. In comparison, modern North
Pacific deep waters average ;130 mmol/L.
When the ocean reaches a steady state after application of the
reduced temperature gradient, oxygen concentrations below the top 100
m are reduced by an average of 264 mmol/L. Deep ocean oxygen levels
are dysoxic (,45 mmol/L) to anoxic over a majority of the deep-sea
floor (Fig. 3D), and intermediate water oxygen concentrations are neg-
ative; values are as low as 2300 mmol/L off the western coast of the
supercontinent (Fig. 3C).
Knoll et al. (1996) discussed a box model of Permian anoxia and
suggested that utilization of ;470 mmol/L of sulfate in the Late Perm-
ian caused buildup of lethal CO
2
levels in the deep ocean and contrib-
uted to the end-Permian extinction. Because sulfate has twice the re-
ducing power of oxygen, this amount of sulfate reduction would
correspond to values of ;21000 mmol/L oxygen in the model. Such
levels are not even approached in the reduced-gradient simulation.
Therefore, some factor in addition to reduced overturning seems nec-
essary to create such high levels of CO
2
in the Late Permian deep
ocean.
DISCUSSION AND CONCLUSIONS
Our results suggest that deep-ocean anoxia is consistent with re-
duced thermohaline circulation driven by a low meridional density con-
trast. However, almost half the difference between the simulated high-
gradient and reduced-gradient deep water oxygen levels is due to the
difference in solubility of oxygen in the warmed high-latitude regions
where deep waters are formed (;250 mmol/L, vs. 370 mmol/L in the

GEOLOGY, January 2001 9
Figure 2. Simulated me-
ridional overturning (zon-
ally integrated volume
transport stream func-
tion in meridional vertical
plane) for Permian ocean
in Sverdrups (Sv; 10
6
m
3
/
s). Positive values indi-
cate clockwise flow, and
negative values indicate
counterclockwise flow.
Modular ocean model
(MOM) (A) equilibrated
with high-gradient sea-
surface temperature forc-
ing, and (B) 100 yr, (C)
1200 yr, and (D) 10 k.y. af-
ter imposing reduced-
gradient forcing.
Figure 3. Maps of steady-
state oxygen concentrations
(mmol/L) for (A) intermediate
(850 m) and (B) deep ocean
(4650 m) in high-gradient
scenario; (C) upper interme-
diate and (D) deep ocean in
reduced-gradient scenario.

10 GEOLOGY, January 2001
high-gradient scenario). This result is consistent with box modeling
results for the Cretaceous (Herbert and Sarmiento, 1991). Biological
oxygen demand explains the remainder of the oxygen decline (;140
mmol/L), but the magnitude of this oxygen demand alone would not
drive the reduced-gradient ocean completely anoxic without the afore-
mentioned temperature effect. Chemical stratification is limited because
decreased upwelling rates of nutrients in the reduced-gradient case sup-
port lower export of organic matter from surface waters than in the
high-gradient case (14.3 vs. 19.8 Gt C/yr). The effects of slowed ven-
tilation and dwindling productivity do not exactly balance because the
surface biota exploit a reserve source of nutrients, i.e., high-latitude
phosphate.
Because of light limitation at high latitudes, the high-gradient sce-
nario with vigorous upwelling exhibits phosphate concentrations .2
mmol/L in southern high latitudes and 1.5 mmol/L in the northern areas
of convection. In the reduced-gradient scenario, with slower upwelling
and longer high-latitude surface-water residence times, phosphate val-
ues in surface waters of both the northern and southern high latitudes
decline by ;0.4 and 0.8 mmol/L, respectively (not shown). As a result,
3.62 310
20
mol of phosphate are lost from surface waters in the
reduced-gradient case, fueling export of organic matter that drives al-
ready low deep-water oxygen concentrations to anoxic levels. This sen-
sitivity of deep-water oxygenation to high-latitude productivity is con-
sistent with the box-modeling results of Sarmiento et al. (1988) and
Hotinski et al. (2000). Thus, although the reduced-gradient scenario
exhibits widespread anoxia, relatively high rates of ventilation and re-
duced productivity keep the deep ocean near the oxic-anoxic boundary,
rather than firmly in the regime of sulfate reduction. Increasing the
marine phosphate inventory by 50% increases chemical stratification
in the model (results not shown), but five times more phosphate would
be needed to drive the system to the values suggested by Knoll et al.
(1996). Because the phosphate content of the Permian ocean is not
well determined, such a high value cannot be ruled out (Martin, 1995).
High steady-state phosphate concentrations in pore waters of modern
marine sediments, which commonly approach 10 mmol/L, indicate that
high levels of phosphate are sustainable (Emerson et al., 1980; Van
Cappellen and Berner, 1988). Thus, it is possible that the Permian
ocean’s phosphate concentration was significantly higher than today’s
value of 2.15 mmol/L.
ACKNOWLEDGMENTS
Financial support was provided by the NASA Astrobiology Institute Cooperative
Agreement (NCC2-1057), the National Science Foundation (grant EAR-98-05139), the
Shell Oil Company Foundation, and the Pennsylvania Space Grant Consortium. We thank
Tim Bralower for a thoughtful review of the manuscript.
REFERENCES CITED
Bice, K.L., Sloan, L.C., and Barron, E.J., 2000, Comparison of early Eocene isotopic
paleotemperatures and the three-dimensional OGCM temperature field: The potential
for use of model-derived surface water d
18
O, in Huber, B.T., et al., eds., Warm
climates in Earth history: Cambridge, UK, Cambridge University Press, p. 79–131.
Bralower, T.J., and Thierstein, H.R., 1984, Low productivity and slow deep-water circu-
lation in mid-Cretaceous oceans: Geology, v. 12, p. 614–618.
Broecker, W.S., 1991, The great ocean conveyor: Oceanography, v. 4, p. 79–89.
Emerson, S., Jahnke, R., Bender, M., Froelich, P., Klinkhammer, G., Bowser, C., and Set-
lock, S., 1980, Early diagenesis in sediments from the eastern Equatorial Pacific, I.
Pore water nutrient and carbonate results: Earth and Planetary Science Letters, v. 49,
p. 57–80.
Fischer, A.G., and Arthur, M.A., 1977, Secular variations in the pelagic realm, in Cook,
H.E., and Enos, P., eds., Deep-water carbonate environments: Society of Economic
Paleontologists and Mineralogists Special Publication 25, p. 19–50.
Gruszczynski, M., Hoffman, A., Malkowski, K., and Veizer, J., 1992, Seawater strontium
perturbation at the Permian-Triassic boundary, West Spitsbergen, and its implications
for the interpretation of strontium isotopic data: Geology, v. 20, p. 779–782.
Herbert, G.D., and Sarmiento, J.L., 1991, Ocean nutrient distribution and oxygenation:
Limits on the formation of warm saline bottom waters over the past 91 m.y.: Ge-
ology, v. 19, p. 702–705.
Holser, W.T., and Magaritz, M., 1987, Events near the Permian-Triassic boundary: Modern
Geology, v. 11, p. 155–180.
Holser, W.T., Scho¨nlaub, H.P., Boeckelman, K., and Magaritz, M., 1991, The Permian-
Triassic of the Gartnerkofel-1 core (Carnic Alps, Austria):Synthesis and conclusions,
in Holser, W.T., and Scho¨nlaub, H.P., eds., The Permian-Triassic boundary in the
Carnic Alps of Austria (Gartnerkofel region): Abhandlungen Geologische Bunde-
santstalt Austriche, v. 45, p. 213–232.
Hotinski, R.M., Kump, L.R., and Najjar, R.G., 2000, Opening Pandora’s Box: The impact
of open-system modeling on interpretations of anoxia: Paleoceanography, v. 15,
p. 267–279.
Isozaki, Y., 1997, Permo-Triassic boundary superanoxia and stratified superocean:Records
from lost deep-sea: Science, v. 276, p. 235–238.
Kajiwara, Y., Yamakita, S., Ishida, K., Ishiga, H., and Imai, A., 1994, Development of a
largely anoxic stratified ocean and its temporary massive mixing at the Permian/
Triassic boundary supported by the sulfur isotopic record: Palaeogeography,Palaeo-
climatology, Palaeoecology, v. 111, p. 367–379.
Knoll, A.H., Bambach, R.K., Canfield, D.E., and Grotzinger, J.P., 1996, Comparative Earth
history and Late Permian mass extinction: Science, v. 273, p. 452–457.
Kutzbach, J.E., Guetter, P.J., and Washington, W.M., 1990, Simulated circulation of an
idealized ocean for Pangaean time: Paleoceanography, v. 5, p. 299–317.
Malkowski, K., Gruszczynski, M., Hoffman, A., and Halas, S., 1989, Oceanic stable iso-
topic composition and a scenario for the Permo-Triassic crisis: Historical Biology,
v. 2, p. 289–309.
Martin, J.H., Knauer, G.A., Karl, D.M., and Broenkow, W.W., 1987, VERTEX: Carbon
cycling in the northeast Pacific: Deep Sea Research, v. 32, p. 267–286.
Martin, R.E., 1995, Cyclic and secular variation in microfossil mineralization, clues to the
biogeochemical evolution of Phanerozoic oceans: Global and Planetary Change,
v. 11, p. 1–23.
Pacanowski, R., Dixon, K., and Rosati, A., 1993, The Geophysical Fluid Dynamics Lab-
oratory modular ocean users guide: Geophysical Fluid Dynamics Laboratory Ocean
Group Technical Report 2.
Rees, P.M., Gibbs, M.T., Ziegler, A.M., Kutzbach, J.E., and Behling, P.J., 1999, Permian
climates; evaluating model predictions using global paleobotanical data: Geology,
v. 27, p. 891–894.
Retallack, G., 1999, Postapocalyptic greenhouse paleoclimate revealed by earliest Triassic
paleosols in the Sydney Basin, Australia: Geological Society of America Bulletin,
v. 111, p. 52–70.
Sarmiento, J.L., Herbert, T.D., and Toggweiler, J.R., 1988, Causes of anoxia in the world
ocean: Global Biogeochemical Cycles, v. 7, p. 417–450.
Taylor, E.L., Taylor, T.N., and Cu´ neo, N.R., 1992, The present is not the key to the past:
A polar forest from the Permian of Antarctica: Science, v. 257, p. 1675–1677.
Tyrrell, T., 1999, The relative influences of nitrogen and phosphorus on oceanic primary
production: Nature, v. 400, p. 525–531.
Van Cappellen, P., and Berner, R.A., 1988, A mathematical model for the early diagenesis
of phosphorus and fluorine in marine sediments: Apatite precipitation: American
Journal of Science, v. 288, p. 289–333.
Wignall, P.B., and Hallam, A., 1992, Anoxia as a cause of the Permian/Triassic mass
extinction, facies evidence from northern Italy and the western United States: Pa-
laeogeography, Palaeoclimatology, Palaeoecology, v. 93, p. 21–46.
Wignall, P.B., and Hallam, A., 1993, Griesbachian (Early Triassic) palaeoenvironmental
changes in the Salt Range, Pakistan and southeast China and their bearing on the
Permo-Triassic extinction: Palaeogeogragraphy, Palaeoclimatology, Palaeoecology,
v. 102, p. 215–237.
Wignall, P.B., and Twitchett, R.J., 1996, Oceanic anoxia and the end Permian mass ex-
tinction: Science, v. 272, p. 1155–1158.
Yamanaka, Y., and Tajika, E., 1996, The role of the vertical fluxes of particulate organic
matter and calcite in the oceanic carbon cycle; studies using an ocean biogeochemical
general circulation model: Global Biogeochemical Cycles, v. 10, p. 361–382.
Yemane, K., 1993, Contribution of Late Permian palaeogeography in maintaining a tem-
perate climate in Gondwana: Nature, v. 361, p. 51–54.
Zachos, J.C., Stott, L., and Lohmann, K.C., 1994, Evolution of early Cenozoic marine
temperatures: Paleoceanography, v. 9, p. 353–387.
Ziegler, A.M., 1990, Phytogeographic patterns and continental configurations during the
Permian Period, in McKerrow, W.S., and Scotese, C.R., eds., Palaeozoic palaeo-
geography and biogeography and climatology: Geological Society [London]Memoir
12, p. 363–379.
Manuscript received May 1, 2000
Revised manuscript received September 21, 2000
Manuscript accepted October 4, 2000
Printed in USA