New
results relevant to air-sea gas flux studies have come in from the
SSTWIND-Carbon Fugacity project, carried out by ALTRAN Brest team in 2012 in collaboration with the Laboratoire
d'Oceanographie Spatiale at IFREMER.
The flux between ocean and atmosphere, F,
of a gas such as CO2 is given by
where k is the gas transfer velocity, α is
the solubility of the gas in water and p is the partial pressure of the gas in
question.
We can substitute the fugacity of CO2
as a proxy for pCO2 and one aim of the project was to verify
empirical formulae for fugacity of atmospheric and oceanic CO2 (fCO2
ATM and fCO2 OCEAN, respectively). Once the accuracy of the
formulae was assessed, the formulae were used to generate monthly mean
climatologies of fCO2 OCEAN and fCO2 ATM and ΔfCO2 (=
fCO2 OCEAN – fCO2 ATM ) in the subtropical gyre of the
North Atlantic.
The formula tested are of the
form
where (pCO2)ATM is
calculated as a function of molar fraction of CO2, sea surface
temperature (SST), sea surface salinity and atmospheric pressure. Two separate
sets of constants (c1 to c5) are used for the periods
February-July and August-January.
The data used to recreate the fugacities of
atmospheric and ocean using these empiric functions come from different
sources. Atmospheric molar fraction was measured at a World Meteorological
Organisation (WMO) station in the Barbados. Surface air pressure was taken from
the National Center for Envionmental Prediction (NCEP) reanalyses and SST from
the satellite sensors AVHRR (Advanced Very High Resolution Radiometer) and AMSR
(Advanced Microwave Surface Radiometer). A constant salinity was used, as the
formulae are a weak function of this parameter.
To validate the formulae, data from the
Surface Ocean CO2 Atlas (SOCAT) were used. This database contains
observations from Voluntary Observing Ships across the global oceans, gridded
to a 1ºx1º monthly average value. The fields are very heterogenous in quality,
however, with numerous obervations per
month along well-travelled ship routes but very few, if any, observations in other
areas. Figure 1 shows the number of observations in the SOCAT database in the
zone of study.
 |
| Figure 1. Number of observations in the
SOCAT database over the study zone. |
When the spatial heterogeneity
is taken into account and the modelled fCO2 is compared to points
with more than 72 observations (equivalent to 6 consecutive years), the
correlation with the observed fCO2 is greater than 0.9. Equally, the mean error
is approximately 2% and the maximum error is 6% over the zone.
 |
| Figure 2. Annual trend, in μatm/year, of
fCO2 ATM (top), fCO2 OCEAN (middle) and ΔfCO2
(bottom) between 1997 and 2012. |
The trends of the fCO2 ATM and
fCO2 OCEAN in the area increase over the period 1997-2012, being
1.85 uatm/year and 1.96 uatm/year respectively. These trends follow that of SST
trends over the same period. Both show a marked annual cycle, with a phase lag
of 6 months: fCO2ATM peaks in winter, and fCO2 OCEAN in
summer. The fCO2 OCEAN drives the spatial variability of ΔfCO2,
however, as shown in figure 2.
 |
| Figure 3. Monthly
mean ΔfCO2, in μatm. |
 |
| Figure
4. Monthly mean fCO2 ATM, in μatm. |
 |
| Figure 5. Monthly mean fCO2 OCEAN,
in μatm. |
Monthly climatologies (figures 3, 4 and 5)
show that the ocean acts as a sink of CO2 from January to April and as a source
between May and December, with the most pronounced seasonal differences in the
north of the study zone.
