 |
| Titel |
Carbon storage in the South Atlantic Ocean. Assessing time trends in Cant with a novel, multiple-cruises variant of the MLR-technique: Time Series Residuals (TSR) |
| VerfasserIn |
Steven van Heuven, Harro A. J. Meijer, Mario Hoppema, Hein J. W. de Baar |
| Konferenz |
EGU General Assembly 2011
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 13 (2011) |
| Datensatznummer |
250053979
|
|
|
|
|
|
| Zusammenfassung |
| We present and apply the Time Series Residuals (TSR) approach, a novel variation of the
MLR technique (Wallace, 1995). The new approach allows the evaluation of time trends in
CT in datasets comprising multiple cruises in a region of interest.
The TSR methods evaluates, on a per-water mass basis, the time trends of the residuals of
a single multivariate linear regression performed on the data from all cruises in the dataset.
This technique exploits the premise that the use of regression coefficients that were
derived using all data from a given time interval, will, for the oldest cruises in that
interval yield predictions for CT that are too high, and for the youngest cruises are too
low. The temporal distribution of the residuals of the regression reflects the rate
of increase of CT in a water mass, which is attributed to uptake of Cant from the
atmosphere.
Due to the increasingly rapid buildup of Cant in the atmosphere, the rate of buildup of
Cant in the ocean will not be constant through time, and thus neither throughout
a water mass. Rather, this rate will be higher at the formation region and lower
further ‘downstream’ (this ‘older’ water was ventilated at times of lower atmospheric
Cant). In order to retain this characteristic upon application of our method, the
dependence of the evolution of CT on the distance from the formation region of the
water mass is quantified as a function of Apparent Oxygen Utilization (AOU), the
value of which can be considered to be an indication of water mass ventilation
age.
A notable benefit of this method is that the rate of change of CT is parameterized on the
basis of only AOU and water mass fraction (as derived from Optimum Multiparameter
Analysis, OMP). Hence, the results obtained from time-stamped cruise datasets such as
GLODAP (Key et al., 2004) and CARINA (Key et al., 2010) are easily projected onto
high-resolution climatologies such as the World Ocean Atlas 2005 (Levitus, 2006), as is done
in the present study.
Application to data of 68 cruises performed in the South Atlantic Central Water between
1972 and 2009 yields trends in CT ranging from +0.1±0.2 μmol kg-1 year-1 for the interior
part of the water mass to +0.7±0.1 μmol kg-1 year-1 close to the formation region of the
same water mass. The spatially integrated rate of storage is calculated to be 103±12 TgC
year-1. Results for the other water masses in the South Atlantic Ocean, obtained in similar
fashion, add up to a total rate of uptake of Cant of 0.4±0.1 PgC year-1 for the region south
of the Equator between the longitudinal boundaries of 65ºW (Drake Passage) and 30ºE. |
|
|
|
|
|
|