What can we learn from the paleo record
about past changes in ocean productivity and
controls of atmospheric CO2?
QuickTime™ and a
TIFF (Uncompressed) decompressor
are needed to see this picture.
Bob Anderson, Gisela Winckler, Martin Fleisher
Lamont-Doherty Earth Observatory
Columbia University
Exploring Ocean Iron Fertilization, WHOI, September, 2007
Ice core records show tight
coupling between CO2 & Climate
From Brook, 2005
Comment on
Siegenthaler et al.,
2005
CO2
Climate
How did the ocean lower glacial
atmospheric CO2 levels?
Plausible mechanisms
1) Increased strength of the biological pump
• Increase nutrient inventory (capacity)
• Increase nutrient utilization (efficiency; today at ~50%)
2) Increase ocean ALK/DIC ratio ([CO32-])
• Continental weathering
• Shelf-basin fractionation (“Coral Reef” hypothesis)
• C-org/CaCO3 ratios (“Rain Ratio” hypothesis)
Biological pump of Sigman & Boyle, 2000
CO2
CO2
What does
“efficiency” of the
biological pump
mean?
It is the fraction of
upwelled nutrients
that are utilized and
exported to depth as
organic matter.
+ DIC !!
Regeneration
Preformed nutrients
are the master
variable to
characterize the
efficiency of the
biological pump.
Sensitivity of CO2 to preformed nutrients
pCO2 vs Preformed PO4
Princeton Ocean GCM runs with different nutrient utilization scenarios.
Constant ocean nutrient inventory
Marinov et al., Nature, 2006
Annual average Nitrate Concentration at 20 m
Levitus Nitrate - Color
Levitus Nitrate
Only about half of the upwelled nitrate is used by phytoplankton.
Efficiency of the Biological Pump today is low. Potential to alter CO2 is high.
From: iridl.ldeo.columbia.edu/SOURCES/.LEVITUS94
Martin’s “Iron Hypothesis”
Dust is inversely correlated with CO2 in Antarctic ice
core records -- is there a causal relationship?
Martin (1990) reasoned that increased dust fluxes relaxed Fe
limitation in the glacial Southern Ocean, allowing increased
efficiency of the biological pump to draw down atmospheric CO2
Fe flux (µg m-2 yr-1)
CO2 (ppm)
Antarctic Ice Core
Dust (Fe) - CO2 (anti)correlation
Fe flux from Wolff et al., Nature 2006; CO2 from Brook, Science 2005
Questions to ask of the paleo record:
1) Did dust affect Productivity in HNLC regions?
2) Did other sources of Fe have a significant impact
on productivity?
3) What caused glacial CO2 to be 80-100 ppm lower?
Questions to ask of the paleo record:
1) Did dust affect Productivity in HNLC regions? (No)
2) Did other sources of Fe have a significant impact
on productivity? (I think so)
3) What caused glacial CO2 to be 80-100 ppm lower? (?)
Equatorial Pacific
RC17-177
X
VNTR 08
Search for evidence of dust influence in regions with paired
records of dust flux and paleoproductivity.
Equatorial Pacific Dust-Climate Correlation
West
Winckler et al., submitted
Dust flux proxy
East
Global ice volume proxy
• Dust flux proxy is tightly correlated with climate
• Glacial-interglacial amplitude ~2.5X at all sites
Equatorial Pacific - Antarctica Correlation
Winckler et al., submitted
Internally-consistent change in dust flux from at least 3 sources
suggests control by global hydrological cycle
CEP - No Productivity Response
TT013-PC72 Equator, 140°W
Barite PP
Th-232 Dust
0.4
2.0
0.3
1.5
0.2
1.0
0.1
0.5
0
50
100
150
Age (ka)
200
250
Flux (µg cm-2 ky-1)
Dust
0.5
232Th
Productivity
Barite Flux (mg cm-2 ky-1)
2.5
0.0
300
232Th
flux (Dust proxy) - Winckler et al., submitted
Barite concentration - Paytan, 1995
Barite flux (PP proxy) - Anderson et al., in press
Proxy records for paleoproductivity and dust flux are
uncorrelated over the last 3 glacial cycles
EEP - No Productivity Response
ODP849-Th
PC72-Th
2.5
0.45
2.0
0.35
1.5
0.25
1.0
0.15
0.5
0.0
0.05
0
5
10
15
20
25
30
Flux (µg cm-2 ky-1)
Dust
VNTR08-Barite
232Th
Productivity
Barite Flux (mg cm-2 ky-1)
TT013-PC72 Equator, 140°W
VNTR08 & ODP 849, Eq., 110°W
Age (ka)
232Th
flux (Dust proxy) - PC72, Anderson et al., 2006; ODP849, Winckler et al., sub.
VNTR08 Barite Flux (PP proxy) - Barite conc. Paytan, 1995
Sediment flux - Pichat et al., 2004
Productivity shows no response to a 2-fold drop in dust
flux over the last deglaciation
Equatorial Pacific:
Increased glacial dust fluxes had no detectable effect on
export production.
What about the Southern Ocean?
Levitus Nitrate - Color
Levitus Nitrate
Here, increased nutrient utilization south of the Antarctic Polar Front has the
greatest potential to affect atmospheric CO2.
From: iridl.ldeo.columbia.edu/SOURCES/.LEVITUS94
LGM minus Modern Export Production
(synthesis of published data; all proxies)
High glacial productivity is restricted to the Subantarctic zone
Iron fertilization was not pervasive throughout the Southern Ocean
Kohfeld, LeQuéré, Harrison and Anderson, Science, 2005
Sites around the Southern Ocean with detailed
records showing glacial productivity < interglacial
Nutrient utilization
south of the APF
has the greatest
potential to impact
global inventory of
preformed nutrients.
Marinov et al., (2006)
SW Pacific - Two Cores & Three Proxies
Consistently show glacial productivity < Holocene
Anderson et al., 2002
300
250
2
200
150
1
100
50
0
0
20
40
60
80
100
120
140
4
0
160
300
250
3
200
2
150
100
1
50
0
0
20
40
60
80
100
120
140
0
160
EDC Fe Flux (µg m-2 y-1)
3
EDC Fe Flux (µg m-2 y-1)
Opal Flux (g cm-2 ky-1)
S Atlantic core RC13-259 - 54°S, 5°W
xsBa Flux (mg cm-2 ky-1)
Export Production Proxy
Diatom Productivity Proxy
S Atlantic Productivity anti-correlated with dust
Age (ka)
Opal & Ba fluxes: Anderson et al., 2002 - EPICA Dome C Fe flux: Wolff et al., 2006
Site is downwind of the Patagonian dust source
If dust-borne Fe
stimulated nutrient
utilization in the
glacial Southern
Ocean, then it
should have been
evident here.
Southern Ocean (South of APF):
Any iron fertilization by increased glacial dust fluxes was
more than offset by other factors that reduced export
production.
Did Fe have any impact on glacial
productivity in the Southern Ocean?
LGM minus Modern Export Production
(synthesis of published data; all proxies)
“Hot Spots” - Subantarctic Sites Experienced High Productivity
Kohfeld, LeQuéré, Harrison and Anderson, Science, 2005
Examples from Subantarctic “Hot Spot”
Higher Subantarctic Productivity in LGM supported
by order of magnitude greater C-org burial
C-org Flux (mmol m-2 yr-1)
0
Patterns
reproduced in
two cores
10
20
0
20
40
Glacial
Changes
were BIG!
Age (ky)
60
80
100
Interglacial
120
140
RC13-254
RC15-93
160
Anderson et al., 1998, 2002
Why such different behavior among cores
downwind of Patagonia?
Blue = Lower glacial productivity; Red = Higher glacial productivity
Contours = Summer Nitrate µM; ample nutrients N of APF
Why such different behavior among cores?
Is the APF (convergence) a barrier to supply of essential factor?
Patagonian ice
sheet during
glacial times
delivered Ice-
Rafted Debris
(IRD) to the
Southern Ocean
Modern
ALACE float
tracks show
that currents
would have
carried
Patagonian
IRD into the
S Atlantic
Courtesy of
S. Gille, SIO
Icebergs as a
source of Fe location matters!
QuickTime™ and a
TIFF (Uncompressed) decompressor
are needed to see this picture.
Subpolar
Antarctic
Alaskan photos from John Crusius
QuickTime™ and a
TIFF (Uncompressed) decompressor
are needed to see this picture.
APF would have been a barrier to icebergs, IRD,
and any Fe released from IRD
QuickTime™ and a
TIFF (Uncompressed) decompressor
are needed to see this picture.
Evidence for Patagonian Fe fertilization?
1) YES - Isotopic and mineralogical data; Diekmann, Walter,
Kuhn, & others at AWI;
2) Nd isotopes in Cape Basin (highlighted star)
TN57-21-
TN57-21 eps Nd & productivity
1500
-7.0
-7.5
1000
-8.0
500
-8.5
-9.0
0
0
200
400
600
800
1000
1200
Alkenone Flux
(ng/cm2/ka)
Alkenone Flux
 Nd
S America
-6.5
2000
eNd
230Th-n
-6.0
1400
Depth (cm)
MIS 3
MIS 4
-6.0
2.0
-6.5
1.5
1.0
 Nd
S Africa
-7.0
-7.5
0.5
-8.0
0.0
eNd
-8.5
U auth
-9.0
-0.5
0
200
400
600
800
Depth (cm)
1000
1200
1400
Authigenic U (dpm/g)
MIS 2
Cape Basin:
Nd isotopes
correlate with
productivity
proxies…
May reflect Fe
supply.
Alkenone Flux
Sachs & Anderson, 2003
Uranium
Sachs & Anderson, 2005
Nd
Piotrowski et al., 2005
LGM minus Modern Export Production
(synthesis of published data; all proxies)
Ongoing
Hypothesis- “Hot Spots” reflect Fe from Patagonia & Kerguelan
Current work on S Pacific shows no hot spots; supports local Fe fertilization
Kohfeld, LeQuéré, Harrison and Anderson, Science, 2005
Summary:
No evidence for Fe fertilization of HNLC regions (EqPac &
So. Ocean) by increased glacial dust fluxes.
Subantarctic: Hot spots of high productivity may have been
fertilized by local sources of Fe; not dust, maybe
icebergs.
Impact of Subantarctic on CO2 minor because disconnected
from main inventory of preformed nutrients.
Increased ocean stratification, with feedbacks from CaCO3
compensation, lowered glacial atm. CO2
(Marchitto et al., Science, 2007)
What caused lower glacial CO2?
Increased ocean stratification was a primary factor.
Marchitto et al,
Science, 2007
Indirect evidence from 14C of benthic forams at 700m in N Pacific.
Accelerated overturning of deep waters brought CO2 to the atm.,
and 14C-depleted DIC, both to intermediate depths and to the atm.
More direct evidence: Deglacial increase in So Ocean upwelling
coincided with rise in CO2 and drop in 14C of Atm. CO2
F(opal)
6
Hughen06
5
400
IntCal04
4
300
3
200
2
ACR
100
1
14C of Atm. CO2 (‰)
Opal MAR (g cm-2 kyr-1)
500
0
7 0
5
10
15
20
25
F(opal)
30 300
CO2
6
280
5
260
4
240
3
220
2
ACR
1
200
0
180
0
5
10
15
Age (ka)
20
25
30
Dome C CO2 (ppm)
CO2 from
Monnin et al.,
EPSL, 2004
53.2°S, 5.1°E
0
Opal MAR (g cm-2 kyr-1)
Hughen Atm. 14C
from Cariaco
Basin. Hughen et
al., 2006
Increased Upwelling
IntCal Atm. 14C
from Reimer et al.,
2004
Increased Upwelling
Opal flux from TN057-13PC
7
Deglacial increase in upwelling is evident at sites
all around the Southern Ocean
Red star = TN057-13