Background
Particulate Organic Matter (POM) in the ocean consists of a variety of material such as live organisms, zooplankton carcasses, organic debris, zooplankton fecal pellets, transparent exopolymer particles, and minerals of both biogenic and terrestrial origin. The adhesive, coagulative nature of these particles leads to the formation of aggregates that become dense enough to sink towards the sea floor (McDonnell et al., 2011). It is well understood that the flux of marine particulate matter is a key process that links the surface with the deep ocean, and is a controlling factor in the distribution of many elements that form the building blocks of marine biota (Buesseler et al., 2007). These sinking particles are closely linked to the biological carbon pump (Ducklow et al., 2001), represent a key component in the ocean food web and can also act as a sink for inorganic minerals and pollutants that are scavenged by the sinking particles.

Sample collection
The BATS program uses a method developed by Knauer et al. (1979), which was used extensively at the VERTEX time-series site in the northeast Pacific, and is suitable for the assay of all levels of sinking flux found in the ocean. Particle flux is collected within a floating array of polycarbonate tubes (7 cm diameter by 53 cm height) containing a buffered brine solution with polycarbonate filters at the bottom. During monthly cruises to the BATS site, tubes are deployed at three depths (150, 200, and 300 meters) for a target time of 72 hours, depending on weather and conditions. The array is tracked using Argos, Iridium, RDF, and AIS localization units.

Before the array is deployed, tubes are arranged in deployment racks for the different depths and filled with a heavy brine (~85 ppt) and 2% formalin solution. Equipment used and the specific arrangement of the array is represented in Figure 1 (See Supplemental Files section). Four tubes are deployed for carbon and nitrogen determinations, three tubes are deployed for phosphorus counts, and three tubes are prepared (but not deployed) for blank counts.

Upon recovery, the ambient seawater above the interface (at approximately half height of the tubes) is siphoned off and the brine in the lower half of the tube is drained through the filter by means of opening a valve. Tubes are disassembled inside a fume hood and the pre-weighed polycarbonate filters containing the flux sample are transferred to clean petri dishes. Samples are misted with pH buffer solution once in the petri dish to keep them moist. Misting is done during transfer of samples as well as during the time they remain stored at 4ºC until analysis. Care is taken to store samples no longer than 3 to 6 months to prevent any biological processes that could modify organic carbon contents.

Sample analysis
After recovery, filters are examined under a microscope at 12x and 25x magnification to remove swimming zooplankton and other organisms which are not considered sediment flux. Manual removal is needed to prevent overestimation of sinking material that could skew the end results. The remaining flux on each filter is scraped into a bolus, dried and then placed into a silver capsule. For each depth, three samples are acidified to remove inorganic carbon prior to measurement of organic carbon and nitrogen. The non-acidified sample (fourth bolus) is measured for total carbon. Samples are analyzed with a CE440 Elemental Analyzer (Exeter Analytical), which combusts and reduces the sample, converting all C and N present into CO₂ and N₂. These gas concentrations are measured by thermal conductivity detectors and converted to electrical millivolt signals. Calibration coefficients are then used to convert millivolt outputs to carbon and nitrogen concentrations in micrograms. Acetanilide conditioners and machine blanks are also run to standardize the instrument and monitor performance and accuracy during a run. (See Data Processing Section below for additional information).

The PITS samples that are collected for determination of total particulate phosphorus (PPhos) are analyzed at Bigelow Laboratory of Ocean Sciences following the methods described in Lomas et al. (2010). The PPhos samples have swimmers removed (as detailed above) and are then analyzed using the ash-hydrolysis method of Solarzano and Sharp (1980). This procedure does not separate inorganic from organic phosphorus so these results should be considered as total particulate phosphorus.

Considerations, concerns, and challenges
Sediment traps are the only tools for directly collecting sinking particles in the ocean. They are largely un-calibrated in the field and there are significant unresolved questions on their accuracy and precision. It is assumed that the collection of particles is linearly related to the aperture area of the sediment trap and that this collection is an accurate estimate of the mass of sinking particles at that depth and the particle sinking speeds. But hydrodynamics and other factors influence the collection of material by sediment traps so the interpretation of trap data should be approached with caution (Antia, 2005; Butman et al., 1986; Gardner et al., 1980; Kahler and Bauerfeind, 2001; Lee et al., 1988). The U.S. JGOFS Planning Report #10 provides an overview of these issues (U.S. JGOFS Planning Office, 1989) and there have been significant published papers on trap accuracy since that report.

Any investigator using sediment traps should become aware of the facts and challenges in order to make informed decisions about appropriate methods for their investigation.
Any scientist looking to use sediment trap data should be familiar with the factors affecting flux estimates, the methods used for the study, and the assumptions made for the flux calculations