Dataset: Trap Flux
Deployment: OC399-03

Detailed methods for all data collected as part of this study can be found in one of the three
publications arising from this study (references given below). This contains information on
analytical machines and certified standards where applicable.

Sample QA/QC procedures followed those of the Bermuda Atlantic Time-series Study (BATS).
At the point of collection, any leaking niskin bottles were noted on the master cast sheets
and samples were taken from a different niskin fired at the same depth as the leaking bottle.
No data are reported for leaking Niskin bottles. During sample analysis standard curves and/or
certified standards were carefully examined to ensure that they were consistent with expectations
and accurate. Next, data were plotted as depth profiles and compared to a quality control window
for the February/March period at BATS. The QC window consisted of the upper and lower 95% confidence
limits based upon all data collected at BATS during February/March from 1989-2005. If our data
fell well within these bounds it was considered 'acceptable'. For those data that fell near or
outside the QC window, we went back to the original data run to ensure there was no miscalculation
or other error. If nothing was found, then we examined other data from that niskin to see if other
samples are in question. If no obvious error or problem was found, the data were considered OK and
in the range of extremes that this study hoped to observe.

Sample accuracy and precision. Sample accuracy was assessed by using certified standards, for those
measurements where standards are available (dissolved oxygen, nutrients, salinity, dissolved inorganic
and organic carbon). Certified standards were run with each analytical run and compared to long term
control charts for respective analyses. Samples were not run until certified standards were shown to
be accurate for that analytical run. Sample precision was determined by analyzing replicate samples
(not replicate analyses on the same sample) and therefore is higher than analytical precision due to
the inclusion of sampling error. At the concentrations observed during this study, sample precision
was <5% for stock measurements and <10-15% for rate measurements. Some analyses, namely dissolved
oxygen and salinity, were much better and often had a sample precisions <1%. These precision estimates
are consistent with the long term QA/QC seen with the BATS program.

The provided data files are complete matrices and therefore not every sample (columns) will be taken
from every Niskin fired (rows). Data that were either not collected, or were associated with leaking
Niskins, or were found to be in error for other reasons are denoted by "-nd" in the spreadsheets.

References:
Detailed information on phytoplankton analysis.
Lomas, M.W., Roberts, N., Lipschultz, F., Krause, J.W., Nelson, D.M., and Bates, N.R. 2009.
Biogeochemical responses to late-winter storms in the Sargasso Sea. IV. Rapid succession of
major phytoplankton groups.
Deep Sea Research I, 56: 892-909.
doi:10.1016/j.dsr.2009.03.004

Detailed information on all silica cycle measurements.
Krause, J.W., Nelson, D.M., and Lomas, M.W. 2009. Biogeochemical responses to late-winter
storms in the Sargasso Sea. 2009. II. Increased rates of biogenic silica production and export.
Deep Sea Research I, 56: 861-875.
doi:10.1016/j.dsr.2009.01.002

Maiti, K., Benitez-Nelson, C.R., Lomas, M.W., and Krause, J. W. 2009. Biogeochemical responses
to late-winter storms in the Sargasso Sea. IV. Comparison of Export Production by 234Th and
Sediment Traps.
Deep Sea Research I, 56: 875-892.
doi:10.1016/j.dsr.2009.01.008

Detailed information on general biogeochemical measurements.
Lomas, M.W., Lipschultz, F., Nelson, D.M., and Bates, N.R. 2009. Biogeochemical responses
to late-winter storms in the Sargasso Sea. I. Pulses of new and primary production.
Deep Sea Research I, 56: 843-861.
doi:10.1016/j.dsr.2008.09.002