Particulates in the water were concentrated onto 25mm glass fibre filters.
Light transmission and reflection through the filters was measured using a spectrophotometer to yield spectral absorption coefficients.
Water samples were taken from Niskin bottles mounted on the CTD rosette. Two or three depths were selected at each station, using the CTD fluorometer profile to identify the depth of maximum fluorescence and below the fluorescence maximum. One sample was always taken at 10m, provided water was available, as a reference depth for comparisons with satellite data (remote sensing international standard). Water sampling was carried out after other groups, leading to a considerable time delay of between half an hour and 3 hours, during which particulates are likely to have sedimented within the Niskin bottle, and algae photoadapted to the dark. In order to minimise problems of sedimentation, as large a sample as practical was taken. Often so little water remained in the Niskin bottle that the entire remnant was taken. Where less than one litre remained, leftover sample water was taken from the HPLC group. Water samples were filtered through 25mm diameter GF/F filters under a low vacuum (less than 5mmHg), in the dark. Filters were stored in tissue capsules in liquid nitrogen and transported to the lab for analysis after the cruise. Three water samples were filtered through GF/F filters under gravity, with 2 30ml pre-rinses to remove organic substances from the filter, and brought to the laboratory for further filtration through 0.2micron membrane filters.
Filters were analysed in batches of 3 to 7, with all depths at each station being analysed within the same batch to ensure comparability. Filters were removed one batch at a time and place on ice in the dark. Once defrosted, the filters were placed upon a drop of filtered seawater in a clean petri dish and returned to cold, dark conditions. One by one, the filters were placed on a clean glass plate and scanned from 200 to 900nm in a spectrophotometer equipped with an integrating sphere. A fresh baseline was taken with each new batch using 2 blank filters from the same batch as the sample filters, soaked in filtered seawater. After scanning, the filters were placed on a filtration manifold, soaked in methanol for between 1 and 2 hours to extract pigments, and rinsed with filtered seawater. They were then scanned again against blanks soaked in methanol and rinsed in filtered seawater.
The initial scan of total particulate matter, ap, and the second scan of non-pigmented particles, anp, were corrected for baseline wandering by setting the near-infrared absorption to zero.
This technique requires correction for enhanced scattering within the filter, which has been reported to vary with species. One dilution series was carried out at station 118 to allow calculation of the correction (beta-factor). Since it is debatable whether this factor will be applicable to all samples, no correction has been applied to the dataset. Potential users should contact JSchwarz for advice on this matter when using the data quantitatively.
Not yet complete:
Comparison of the beta-factor calculated for station 118 with the literature values.
Comparison of phytoplankton populations from station 118 with those found at other stations to evaluate the applicability of the beta-factor.
Two files: phyto_absorp_brokew.txt and phyto_absorp_brokew_2.txt: covering stations 4 to 90 and 91 to 118, respectively. Note that not every station was sampled.
File format: Matlab-readable ascii text with 3 'header' lines:
Row 1: col.1=-999, col.2 to end = ctd number
Row 2: col.1=-999, col.2 to end = sample depth in metres
Row 3: col.1=-999, col.2 to end = 1 for total absorption by particulates, 2 for absorption by non-pigmented particles
Row 4 to end: col.1=wavelength in nanometres, col.2 to end = absorption coefficient corresponding to station, depth and type given in rows 1 to 3 of the same column.
This work was completed as part of ASAC projects 2655 and 2679 (ASAC_2655, ASAC_2679).