South Atlantic Climatology

Thurnherr, A. M. and K. G. Speer, 2004: Representativeness of meridional hydrographic sections in the western South Atlantic. Journal of Marine Research, 62: 37-65.

[Take me straight to the atlas.]

Introduction

One of the goals of the WOCE AIMS Deep Basin Experiment is an investigation of the large-scale circulation in the deep Brazil Basin in the South Atlantic. When plotting isopycnal tracer fields it was soon realized that data sets from different cruises had to be bias-corrected in order to remove obvious artifacts. The O2 corrections calculated by Gouretski and Jancke, (Progr. Oceanogr., 2001) proved unsatisfactory in that their data set lacks an important section in our region of interest (the 1983 Oceanus section at 24S). Furthermore it appears that there are still consistent biases in the O2 data corrected using their data; this may be because of typos in the paper, however. It was therefore decided to do our own bias correction.

Methods

Data Sets

• high-quality sections of known provenance

• primarily sections with O2, SiO2 data; augmented with sections to increase number of repeat-sampling regions (RFZ2, M413) and along-crest salinity section (JR65); NB: B2 has O2 but not SiO2

Bias/Scale Corrections

• repeat-sampling regions with 150km max. distance between stations; extending to depths below 1500m; (station plot)

• corrections derived from mean values on 10 potential temperature surfaces for each data set in 3x3 degree regions centered around nearest-neighbor station

• properties subsampled to nearest 50dbar average (CTD data) or linearly interpolated (bottle data)

• potential temperature is a function of salinity adjustment was repeated until solution converged

• temperature range of low variability for each repeat-sampling region and property manually selected; some repeat-sampling regions discarded for each property

• bias/scale corrections using weighted least-squares as in Johnson et al. (JAOT, 2001); additive for salinity and O2, multiplicative for O2 and SiO2; comparison of bias- and scale-corrected O2 data reveals little differences we use bias corrected O2 values
• a priori weights calculated from observed (10 temperature levels) and expected (0.002 psu, 2% for O2 and SiO2) uncertainties

• reported a posteriori uncertainties are observed ones (from 10 temperature levels)

Presentation

• data averaged in 0.5x0.5 degree regions

• data smoothed with a 1 degree Gaussian filter

• data interpolated and extrapolated by constructing harmonic surfaces no extrema away from control-data points

• manually selected contour levels, different for different depths

• below the nominal depth of 1500m (at _n27.89), the property surfaces are constructed separately in the Brazil, Argentine, Angola, and Cape Basins; this leads to some contouring artifacts at the boundaries because contouring is done across the entire domain

• extrapolated data are not to be trusted; this is mainly a problem in the Argentine Basin (SW corner) near 2500m where the properties are extrapolated from profiles lying south of our region of interest

An Atlas

• properties on selected isoneutral surfaces

• _n levels taken from a 50dbar average Brazil-Basin profile at 19S subsampled every 250m and rounded to 0.01

Acknowledgements

This material is based upon work supported by the National Science Foundation under Grant No. 0220407. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation.