The following methodology section is from McKay (in prep), with minor modifications.

Sample collection
During the Multitracers Project (September 1987 to July 1991) bottom-moored sediments traps were deployed at three stations (Nearshore, Midway, and Gyre) in the Northeast Pacific Ocean. This dataset provides major (Al), minor (Ti, Ba, Mn), and trace element (Cd, Ag, Mo) data for the Nearshore and Gyre stations. The data are for the first 2 years of the project, where "year" refers to the time a trap mooring was deployed until it was recovered and not a calendar year (e.g., Year 1 is September 1987 to September 1988 and Year 2 is September 1988 to September 1989). In addition, there are data for the Nearshore site in Year 3 (September 1989 and March 1990). Sediment trap samples were obtained from the Oregon State University Marine and Geology Repository.

Each mooring had sediment traps deployed at 500 m (Year 1 only), within the OMZ at 1000 m, just below the OMZ at 1500 and/or 1750 m, and 500 m above the bottom (near-bottom traps; ~2300 m at the Nearshore site and ~3200 m at the Gyre site). In Years 1 and 2 most traps, with the exception of the 1500 m trap, had 6 sample cups (C1 to C6) that were open for 30 to 33 days (C1), 60 days (C2 to C5), and 87 to 94 days (C6). The 1500 m trap at both sites had 13 cups that were open for ~30 days each. In Year 3 all Nearshore traps had 6 cups, which were open for 30 days (C1 to C5) and 23 days (C6, near-bottom trap only). No traps were deployed at the Gyre site in Year 3.

In Year 1 the 500 m trap at the Nearshore station failed after the third cup, presumably due to clogging during a high flux event. There is also evidence that the 500 m trap at the Gyre site under-collected material (i.e., lower fluxes at 500 m compared to 1000 m). For these reasons, the 500 m traps were not deployed at either site in subsequent years. The 1500 m traps also failed at both sites in Year 1. Fortunately, traps were also deployed at 1750 m and they functioned correctly.

The trap sample cups were filled with unfiltered seawater collected from the near-bottom trap depth at each site and poisoned with 15g/L of sodium azide. To compare the effects of different "preservatives" a pair of traps were deployed at the Nearshore site (1750 m) in Year 3; one set of cups was poisoned with sodium azide and the others preserved with formalin. No special trace metal clean sampling procedures were utilized during the Mulitracers Project; however, the traps were constructed out of plastic and fiberglass thus reducing possible trace metal contamination.

When the sediment traps were recovered the sample cups were removed, sealed, and not opened until they were ready to be processed in the laboratory. Once back at the lab the samples were allowed to sit undisturbed until the particulate material had settled. Once opened the supernatant was poured off and then used to rinse the particulate fraction through a 2 mm sieve. The ≥2 mm size-fraction was transferred to a bottle containing formalin and refrigerated for future use. The <2 mm size-fraction, which was used in this study, was split into 10 aliquots. Between 1 and 4 aliquots were frozen for future use. The other 6 to 9 aliquots were centrifuged and the supernatant was discarded. These samples were then rinsed with buffered distilled water and centrifuged; this step was carried out twice. Finally, the samples were freeze-dried and then homogenized.

Analytical methods
Sample preparation for major (Al), minor (Mn, Ba), and trace (Cd, Ag, Mo) metal analysis involved placing ~50 mg of sample into teflon reaction vessels and pretreating overnight at room temperature with a mixture of concentrated, quartz-distilled HNO3 (1 ml) and HCl (1 ml). The following day concentrated, quartz-distilled HNO3 (2 ml), BDH Aristar Ultra HF (2 ml), and Milli-Q water (2 ml) were added and the samples were digested using a CEM MARS microwave digestion system. This was followed by 3 cycles of acid evaporation using the same system. After the final evaporation, samples were diluted with 10 ml of 5% quartz-distilled HNO3 and heated for 24 hr at 60°C to remove fluorides complexes (Muratli et al., 2012) The concentrations of Ag, Cd, and Mo in these digests were analyzed by Inductively Coupled Plasma Mass Spectrometry (ICP-MS) using a Thermo Scientific X-Series II instrument located in the Keck Collaboratory at Oregon State University. No column chemistry was carried out prior to ICP-MS analysis. Instead, a collision cell was utilized to reduce the formation of oxides that interfere with the analysis of Ag-109 (Nb oxides and Zr oxides and oxyhydroxides) and Cd-114 (Sn oxides). However, it was still necessary to measure the oxide formation, which was typically <5 %, and apply a small oxide correction to the data. The same digests were used to analyze the major and minor element concentrations by ICP-OES (Prodigy, Teledyne Instruments) in the Keck Collaboratory. Metal concentrations were calculated using an external standard calibration curve.

Precision and accuracy were determined using the National Research Council of Canada sediment standards MESS-3 and PACS-2, which were put through the same digestion and analytical procedures as the samples. In general, precision was better than 8% and accuracy was within error for all metals except Al (see the "Table 1" Supplemental File (PDF)). The Al values for both sediment standards are slightly (~4 to 5 %) lower than their certified values.

The non-lithogenic (i.e., excess) concentrations of the metals were estimated using Eq. (1), where xsM is the excess concentration of the metal, Msmpl is the metal concentration in the sample, Alsmpl is the aluminum concentration in the sample, and (M/Al)lith is the metal/Al ratio in lithogenic material.

[xsM] = [Msmpl] – [Alsmpl] * (M/Al)lith

This method of calculating excess concentrations relies on measuring or assuming a reasonable lithogenic value for each metal. In this study, the average shale values reported in Brumsack (2006) were used (i.e., 8.83 wt% Al; 580 µg/g Ba; 850 µg/g Mn; 0.13 µg/g Cd; 0.07 µg/g Ag; 1 µg/g Mo). In addition, because variations in trace metal concentrations can be driven by changes in the concentrations of major components (e.g., dilution by lithogenic material in late fall and winter) all metal data have been converted from concentrations to fluxes using the total particulate flux data provided by Dr. Collier.

Known problems/issues
(1) Year 1 data for the 500 m sediment traps at both sites (i.e., Nearshore samples MT31584 to MT31586 and Gyre samples MT31607 to MT31612) are problematic because evidence (e.g., lower total mass fluxes compared to the 1000 m trap) suggest these shallow traps under-collected settling particulate material.

(2) The 1500 m trap at the Nearshore site in Year 1 (samples MT31396 to MT31401) should have had samples for 13 cups; however, only 6 samples were collected. It is unclear whether the trap failed after cup 6, meaning this cup was open for 237 days, or the trap was programmed improperly and thus each cup was actually open for twice as long. Flux data suggest the latter.

(3) At times fluxes were low and not enough sample was available for an analysis. This was most common at the Gyre site.

(4) The midway (MS) sediment trap samples were not analyzed as a part of this study. Also, there was no Gyre sediment trap samples collected in Year 3.