 |
| Titel |
Anthropogenic carbon estimates in the North Atlantic in CMIP5 GCMs using reconstructed transient tracers (CFC-11, CFC-12, SF6). |
| VerfasserIn |
Tobia Tudino, Marie-Jose Messias, Ute Schuster, Andrew J. Watson |
| Konferenz |
EGU General Assembly 2015
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 17 (2015) |
| Datensatznummer |
250102210
|
| Publikation (Nr.) |
 EGU/EGU2015-1518.pdf |
|
|
|
|
|
| Zusammenfassung |
| Since the beginning of the industrial revolution, atmospheric concentration of carbon dioxide
has raised from 280 ppm to nearly 400 ppm. The Global Ocean has limited the consequent
greenhouse effect on the climate system by taking up around 26% (2.3 ± 0.7 PgC/yr) of the
carbon produced by human activities (anthropogenic carbon, Cant), mostly in key areas, such
as the North Atlantic.
Estimating anthropogenic carbon in the ocean is challenging because Cant cannot
be measured directly, it is difficult to separate from the biggest unknown natural
fraction (DIC), it is influenced by the biogeochemical cycles and the circulation in the
ocean. Therefore, Cant is estimated with different methods (δC*, ΦCTO, TTD)
based on carbon parameters (DIC, alkalinity, silicate, phosphate) or transient tracers
observations.
Here, we used CFC-11, CFC-12 and sulphur hexafluoride (SF6) measured along the
subtropical North Atlantic (26°N) in 2010 as proxies for the Cantestimates. However, these
transient tracer observations are sparse and few GCMs (e.g.: OCMIP project) have reported
them in the outputs. Therefore, oxygen and potential temperature observations registered on
1992, 1998, 2004 and 2010 along the 26°N transect have been used, with a combined
CFC-12/SF6 water masses ages trend and the CMIP5 GCM (MPI-ESM, IPSL-CM5) outputs
to reconstruct transient tracers concentrations using regression models. Overall, the
2010 residual analyses confirm the validity of the statistical approach and derived
values show a consistency with the tracer observations (averaged R2 > 0.9). These
regression models have been used to derive Cant from 1992 to 2014, in CMIP5 GCMs
(MPI-ESM, IPSL-CM5), using the reconstructed transient tracers fields with the
transit-time distribution (TTD) method. Results are, finally, compared to the Cant
distributions and budgets calculated, in the same area, with other methods (δC*,
ΦCTO) and the Cant directly estimated from the GCMs as the difference between
the historical and the control DIC distributions (Cant=DIChistorical-DICcontrol). |
|
|
|
|
|
|