Biology Climate Change Paleoceanography

Ironing Out the Details of the Last Ice Age

Dr. John Martin, a graduate of the University of Rhode Island and long standing director of the Moss Landing Marine Laboratories, California State University is well remembered for having said, “Give me a half tanker of iron, and I will give you an ice age!” at a lecture held at Woods Hole Oceanographic Institute in 1988.  Although an outrageous claim, there is validity and sound science in his aggrandizement.

Photosynthetic marine organisms, phytoplankton, rely on minerals composed of macronutrients such as nitrate and phosphate, as well as sunlight and carbon dioxide to produce organic matter and oxygen.  However, insufficient availability of micronutrients such as iron can limit productivity of phytoplankton as well.  A simple introduction of more iron into the surface ocean can ramp up productivity.  As phytoplankton begin to thrive, they use up CO2 in the surface ocean, allowing drawdown of atmospheric CO2 to the surface ocean.  The connection between atmospheric concentration of CO2 and global climate has been thoroughly investigated.  At very high levels of CO2, the climate is typically very warm, in a “greenhouse” state.  At very low levels of CO2, the climate is more typically very cool, in an “ice house” or glacial state.  Dr. Martin was suggesting in his famous quote that large introductions of iron into the surface oceans would cause extreme lowering of atmospheric CO2, hence driving the global climate towards a glacial state.

Dr. John Martin  Photo:
Dr. John Martin

The featured study investigates the Iron Hypothesis (IH) and how well it explains the observed changes in atmospheric CO2 associated with glacial / interglacial cycles.  The IH states that increased fluxes of iron to the ocean during glacial periods increase export productivity (the raining down of organic matter to deep isolated waters) and lower atmospheric CO2.  The IH had fallen out of favor when it was determined that the Antarctic Zone (south of the Antarctic Polar Front) experienced lower productivity and the Subantarctic Zone (north of the Antarctic Polar Front) experienced higher productivity during the last ice age.  Recently the IH has come back into the spotlight due to difficulty explaining the full 80 parts per million (ppm) decrease in atmospheric CO2 concentration observed during last ice age.  The Antarctic Zone experienced a slow-down of deep ocean ventilation, a process responsible for releasing deeply stored CO2 back to the atmosphere.  The reduction only accounts for ~40 ppm reduction in atmospheric CO2.  The remaining reduction of 40 ppm can be neatly explained by iron fertilization of the Subantarctic, which is fortified by sediment core proxy observations suggesting higher iron and productivity during the last ice age.

The study tests the validity of the IH as a mechanism for lowering atmospheric CO2 by evaluating iron accumulation, productivity, and surface nutrient contents in a sediment core in the South Atlantic Ocean (Figure 1).  Site 1090 is located downwind of the Patagonian dust plume and is characterized by high concentrations of unused nitrate today.  During the last ice age, drier conditions and exposed continental shelves due to lower sea-level lead to more dust (containing iron) being transported into the ocean.  If the IH is valid, productivity should only increase due to the addition of iron to the surface ocean, with no increase of macronutrients.  Furthermore, macronutrients should be more completely consumed.  To test this hypothesis, isotopic analysis of Nitrogen within the calcium carbonate shells of foraminifera (phytoplankton) was performed.  The foram shells are preserved in sediment, and for this study were sourced from the Ocean Drilling Program Site 1090 sediment core.  Previous studies have shown that the ratio of nitrogen isotopes (15N / 14N; also written δ15N) of forams changes depending on the consumption of the macronutrient nitrate (NO3).  The findings of this study show that δ15N rose during the last ice age, indicating more complete consumption of nitrate.  Coincident with the consumption of nitrate was an increase in iron flux and productivity.  This finding supports the validity of the IH, that increased iron supplied to the surface ocean increases productivity and results in lowering atmospheric CO2.  The timing of the changes in δ15N occurs concurrently with increased iron fertilization.

Location of ODP Site 1090, downwind of the Patagonian dust source.
Figure 1. Location of ODP Site 1090, downwind of the Patagonian dust source.

A somewhat puzzling observation is the behavior of δ15N at transitions from deglacial to glacial periods.  The rise in δ15N occurs before increased productivity at the onset of an ice age, and δ15N decreases well after the decline in iron fertilization following the onset of an ice age (Figure 2).  The behavior of δ15N at these transitions appear inconsistent with the IH.  The authors suggest that low sedimentation rates at Site 1090 are not ideal for analyzing the timing of ice age / warming events.

Figure 2.
Figure 2. (A.) Atmospheric CO2 reconstructed from an Antarctic ice core.  (B.) Iron flux (black line) and foram d15N (red line) at ODP Site 1090 (iron accumulation and productivity) (C.)  Alkenone flux at ODP Site 1090 (productivity). (D.) Dust flux at EDC Antarctic ice core.

Regardless, the authors describe two processes that would influence the behavior of δ15N at deglacial/glacial transitions, maintaining that the IH remains valid:

1. The movement of the boundaries of the Subantarctic zone towards the equator during ice ages and towards the South pole during warming.  The equatorward movement of the Subantarctic Zone through Site 1090 during the onset of the last ice age would increase the concentration of nitrate, yet show an apparent decrease in δ15N, suggesting lower productivity.

2. The degree of nitrate consumption in Antarctic surface waters rose during the last ice age.  The Antarctic is a source of nitrate to Subantarctic waters and thus the δ15N may be tied to the nitrate composition of Antarctic surface waters.  Changes in Antarctic δ15N may weakly influence the δ15N of Subantarctic waters, independent of local productivity.



1. Increases in dust accumulation and productivity at Site 1090 are coincident with CO2 drawdown over the last ice age.

2. Model simulations suggest iron fertilization driven drawdown of major nutrients in the Subantarctic can drive a 40 ppm decrease in atmospheric CO2 concentrations.

3. Site 1090 suggests >40ppm decrease, however, it is expected that other regions (i.e. South Pacific) had significantly less iron fertilization.

4. Recent studies discuss changes in Atlantic overturning circulation as the cause of millennial scale CO2 changes.  It is possible that Atlantic overturning circulation changes and iron fertilization work in concert and vary atmospheric CO2 concentrations over glacial-interglacial timescales.


Humans’ combustion of fossil fuels has undoubtedly contributed to rising atmospheric CO2 concentrations, which is rapidly changing the Earth’s climate system.  Iron fertilization has been proposed as a method to reduce atmospheric CO2 concentrations, to counteract climate change.  It is therefore important to understand how the natural climate system has responded to iron fertilization of the ocean’s surface waters.  Although iron fertilization is probably not a viable management strategy for reduction of atmospheric CO2, it is important that we have a robust understanding of how the Earth’s climate system responds to natural variation, which will help to direct policy to better overcome the challenges of global climate change.

One thought on “Ironing Out the Details of the Last Ice Age

Leave a Reply

Your email address will not be published.