Our main aims are to:

Chlorofluorocarbons

Ocean distributions of anthropogenic transient tracers, such as chlorofluorocarbons (CFCs) and radiocarbon (14C) reveal pathways, rates and mechanisms of ocean circulation and mixing. CFC-11 is entirely man-made, and its concentration in the atmosphere increased rapidly from the 1930s to the 1990s. It is soluble and has been invading the oceans, marking waters which have recently been near the surface. Here you can see an animation of the invasion of CFC-11 into the deep North Atlantic in a one-degree resolution configuration of the MIT ocean circulation model.

The Carbon cycle

Understanding the circulation and CO₂ biogeochemistry of the oceans is key to our ability to understand past climates and assess the future evolution of climate. The ocean is important in the regulation of heat and moisture fluxes, and oceanic physical and biogeochemical processes are major regulators of natural atmospheric carbon dioxide (as well as being an important sink of fossil fuel CO₂). The physics, chemistry and biology of the ocean are coupled since the oceanic CO₂ cycle is controlled by the ocean circulation, the supply of sunlight, and major and trace nutrients delivered by the oceanic and atmospheric circulations.

Interannual Variability of Air-Sea Carbon and Oxygen Fluxes

How do changes in ocean circulation and mixing on interannual timescales affect the air-sea exchange of trace gases, such as carbon dioxide and oxygen? We use numerical models of ocean circulation and biogeochemistry to examine the possible effect of interannual changes in meteorological forcing and upper ocean circulation on the fluxes of carbon dioxide and oxygen between the ocean and atmosphere.

With former graduate student Galen McKinley, using global ocean circulation estimates from the ECCO consortium we estimated the interannual variability in the global air-sea exchange of carbon dioxide.

The time-mean air-sea flux of carbon dioxide from a global, interannually varying simulation made by Galen McKinley. The ocean model absorbs carbon from the atmosphere in regions of sea surface heat loss and biological uptake of carbon (e.g. North Atlantic). The ocean model is losing carbon dioxide from the ocean in regions where carbon rich deep waters upwell to the surface (e.g. Tropical Pacific).

Interannual variability in the global flux of carbon dioxide across the model's sea surface is dominated by the role of changes in the Tropical Pacific associated with El Nino. In that region the model's predicted fluxes show general agreement with inferences from observed data and atmospheric inverse models. The ocean model suggests little role for the subtropical and subpolar oceans in modulating atmospheric CO₂ on inter-annual timescales in contrast with inferences from atmospheric inverse models. This apparent disagreement is an outstanding scientific question which remains to be resolved. What is the role of the extra-tropical oceans in modulating atmospheric CO₂ on interannual timescales?

Interannual Variability of the North Atlantic Spring Bloom

In collaboration with Dr. Stephanie Dutkiewicz we have examined the relationships between regional and interannual variations in meteorological forcing and the variability of the spring bloom of biological productivity in the North Atlantic ocean using remotely sensed ocean color data from SeaWiFs mission.

An animation, showing the seasonal and interannual variability of the North Atlantic spring bloom, as observed from space by SeaWiFS, can be seen here (1.9Mb animated gif, images produced by Stephanie Dutkiewicz). Monthly images are sequenced, for the two years 1998 and 1999.

We have also demonstrated the relationship between meteorological forcing and the bloom response in a numerical model of the time dependent ocean circulation with a highly simplified representation of the ecosystem of the North Atlantic (a collaboration with Dr. Watson Gregg of NASA/Goddard Space Flight Center).

Modeling the Ocean Iron Cycle

With former graduate student Payal Parekh and Prof. Ed Boyle at MIT, we have developed an explicit parameterization of the iron cycle in the global ocean. Iron is supplied to the remote ocean by Aeolian (wind-borne) dust deposition. Biological processes in the upper ocean utilize iron and other nutrients, including nitrate and phosphate, creating organic matter some of which eventually sinks into the deep ocean where it is remineralized. In the ocean interior iron it is scavenged onto sinking particles and can become bound to an organic ligand, or complexed though neither of these processes happens to nitrogen or phosphorus. Hence iron is decoupled from the macro-nutrients. In many upwelling regions there is insufficient iron available to utilize all of the available macro-nutrients. Hence, it is thought that iron limits the efficiency of the biological storage of carbon in the oceans.

In the framework of a coarse resolution ocean model we explicitly account for the iron limitation in remote ocean regions. The figure illustrates the modeled dissolved iron distribution (nM) at the surface and 1000m. The distributions are generally consistent with the available, but very sparse, observations. The low surface iron concentrations in the Southern Ocean limit biological production and export.

In collaboration with Stephanie Dutkiewicz, we have coupled the iron model to an ecosystem model which resolves two functional groups of phytoplankton and explicitly represents the cycles of the limiting nutrients phosphorus, silica, iron and light. This is again in a coarse resolution framework. Here we compare the mean surface chlorophyll as observed by the SeaWiFS remote instrument and from the numerical model.

Ocean Carbon-Cycle Model Intercomparison Project (OCMIP)

Comparing model simulations of ocean tracers with observed data can inform us about the tracer transport circulation of our models. We have taken participated in the Ocean Carbon Model Intercomparison Project, OCMIP in which a number of numerical global ocean circulation models are used to simulate a suite of ocean tracers. Comparisons show significant differences between models and data, and between the different models.

For the OCMIP study we used a coarse-resolution configuration of the MIT ocean circulation model (2.8 degree horizontal resolution, 15 vertical levels).