Atmospheric inversions
Atmospheric transport models using CO2 distribution measurements
indicate that there is a large sink of CO2 at mid to high northern latitudes. However, the locations
and mechanisms for this CO2 sink remain difficult to identify. In addition to a large oceanic sink,
centred in the temperate and sub-polar North Atlantic, it is recognised that there must be large
mid-latitude sinks in the terrestrial biosphere. The carbon budgets of Europe and North America
cannot however at present be reliably separated by atmospheric inversions. An important reason
contributing to this uncertainty is the fact that the North Atlantic sink must be accurately known
to partition the land sinks with precision. Better determination and monitoring of these sinks is
therefore important for the Kyoto emission reduction targets to be successfully implemented and
verified. Theme 1 of CARBOOCEAN will reduce the uncertainty in the North Atlantic sink by
providing a seasonally resolved estimates if fluxes into the North Atlantic. Theme 1 will include
atmospheric inversion modelling using the fluxes inferred from the observing system, and these
studies will be integrated with those to be done under Theme 3 (coastal ocean) CarboEurope IP.
North Atlantic Ocean carbon fluxes
During the FP5 CAVASSOO project, we tested elements of
an observing system designed to deliver accurate estimates of the flux of CO2 into the North
Atlantic. To be useful to constrain atmospheric inversions, such a system must have basin-wide
coverage and a time resolution approaching one month since it is known that the sink will be
highly variable on from one season to the next. The backbone of such a system must be the use
of voluntary observing ships (VOS) instrumented to measure surface pCO2 and ancillary
measurements, including nutrients, temperature and salinity.
During CAVASSOO, we have shown that a network of such ships can produce accurate and
complete data using automated instruments with no scientist on board, We have been able to
produce maps of CO2 in much better detail than ever before, for limited areas of the North
Atlantic. We have additionally shown during CAVASSOO that the North Atlantic sink appears
to be changing decreasing slightly, rather than increasing rapidly as ocean carbon model
studies suggest. Such studies have shown us how to design an observing system that can deliver
the required information needed to accurately specify the North Atlantic. The CARBOOCEAN
system will have a better coverage of VOS and will supplement them. It will consist of:
Increased coverage by 6 VOS lines funded from Europe, in combination with VOS lines
of our US colleagues
Time series at selected sites
High accuracy atmospheric sampling for CO2 and O2/N2 ratios from selected vessels and
sites, as an added constraint on atmospheric inversion studies.
In addition, in recognition that time series of the surface at remote locations would be very
helpful in such a system, we will study a modification to an ANIMATE-style mooring to enable
surface-layer measurements.
We include in the North Atlantic the equatorial Atlantic: We expect the tropical regions to be
variable. For example, the Amazon influences the carbon budget in the western equatorial zone,
but there are comparatively few measurements there. New VOS lines going into the tropics will
be commissioned, and we will take advantage of a collaboration to put CO2 measuring
equipment on PIRATA moorings in the equatorial region to produce time series here.
Interpolation, data assimilation
It will be necessary to interpolate the data from the network to
the entire North Atlantic. During CAVASSOO we have investigated how to do this, and have
tried both regression techniques and neural networks to attempt the interpolation in time and
space, with encouraging results. We expect that by including remotely sensed parameters such as
chlorophyll and other ocean pigments we will be able to improve on our present methods and
produce more detailed estimates of the surface pCO2. Remote sensing will also be used to derive
estimates of the gas transfer velocity, and hence CO2 fluxes.
Longer-term, a better method for interpolating the data would be to assimilate it into a model of
the CO2 fluxes for the regions under study. Several modelling groups will be investigating the
route of data assimilation using different ocean carbon models having different strengths. In
addition, we will investigate the use of very high resolution models of the North Atlantic, and
how assimilating physical data into models in the EU-MERSEA IP on operational ocean models.
Southern ocean biogeochemistry
The Southern ocean is the least well measured region for air-sea fluxes. It is the region of most
disagreement between models, such that it is not even clear whether it is a source or a sink for
atmospheric CO2. Such measurements as we do have, suggest that the sub-Antarctic, where
mode waters are formed, are a relatively strong sink region. This seems to be equally true for the
Atlantic and Indian sectors of the Southern Ocean.
Because of the its remoteness, the Southern Ocean cannot be studied using the same methods as
the Atlantic. We will use the new European CARIOCA buoys to obtain data over long periods
from this remote area. We will contribute a new set of observations for the Southern Ocean,
concentrating on the sub-antarctic in the Indian and Atlantic sectors. We will attempt to define
the net CO2 flux there, and understand the processes responsible for it. The CARIOCA data will
be supplemented by those from Antarctic supply vessels. Remote sensing data will be used to
interpolate and extend the data, in a similar manner to that described above for the North
Atlantic.
Oxygen measurements - constraints on gas exchange and the global carbon cycle
The issue of how to parameterize the gas exchange coefficient has hitherto limited the
confidence that we are able to put into our air-sea flux measurements: Historically gas exchange
has been difficult to measure at sea, and the small number of measurements means there remains
uncertainty in how to parameterize it. For the North Atlantic, we propose to obtain an
independent measure of the transfer velocity over large areas, obtained simultaneously with our
CO2 measurements. We will do this by measuring oxygen in the mixed layer simultaneously
with our pCO2 data collection - this has recently become possible using a new generation of
oxygen sensors based on optical technology. With, in addition, information on nutrient
concentrations and mixed layer depth, we can then apply Redfields classic technique for
measuring gas transfer by budgeting the oxygen in the mixed layer. This will enable us to
remove the biggest single uncertainty in calculating air-sea fluxes of CO2.
Objectives
O1.1 Implement an observing system for surface marine CO2 in the Atlantic Ocean. This will
document at seasonal time resolution the present variability of the North Atlantic and tropical
Atlantic fluxes of atmospheric CO2, and O2, with sufficient accuracy to place strong constraints
on adjacent continental sources and sinks. The South Atlantic will also be covered, but with less
resolution.
O1.2 Refine estimates of the magnitude of the Southern Ocean source/sink in the Atlantic and
Indian Ocean sectors, and document seasonal and inter-annual variability there.
O1.3 Develop methods for diagnosing and predicting Atlantic and Southern Ocean marine CO2
fluxes from remotely sensed data, constrained with in-situ measurements.
O1.4 In co-operation with national programs of other countries, integrate the above studies into a
detailed, ongoing global view of the ocean sinks and sources by region.
O1.5 Improve understanding of the processes that control seasonal to interannual variability of
air-sea CO2 fluxes (as realized by agreement between observations and ocean carbon models).
