Methodology:
\n
Samples were collected during two deployment cycles (termed deployment 1 and deployment 2) occupied during the RRS Discovery cruise DY077 to the Porcupine Abyssal Plain Sustained Observatory (PAP-SO) Site in April 2017 (Figure 1). In each of the cycles, we conducted particle flux sampling method intercomparisons between two types of neutrally buoyant sediment traps (NBST and PELAGRA), a surface tethered array of sediment traps (STT), and fluxes derived from upper water column deficits of 234Th vs. its parent isotope, 238U. DY077 samples analyzed in US (WHOI and Skidmore College) are archived here; DY077 samples analyzed in the UK (NOC) are archived in the British Oceanographic Data Centre.
\n
Neutrally Buoyant Sediment Trap (NBST)
\nNBSTs consist of four cylindrical sediment trap tubes (collection area 0.0113 m2) and a 0.25-m pathlength transmissometer (C-Rover 2000, WETLabs, Inc.) arranged around a central SOLO profiling float. The traps are programmed to sink to a predetermined depth, drift while collecting sedimenting particles, close the trap lids, and then rise to the surface at a programmed time for recovery. Recovery aids consist of GPS/Iridium and a flashing strobe light. The transmissometer operates as an optical sediment trap and measures attenuance flux as a function of time, which is a proxy for sinking particulate carbon flux. For the deployments conducted on DY077, trap tubes were set up as follows: Three tubes were prepared with a layer of 500 mL of 70 ppt brine poisoned with 0.1% formaldehyde and borate buffered to pH 8.5. This brine layer was overlain with 1-\u03bcm filtered seawater from 350 m. In the fourth trap tube, a jar containing approximately 50 mL of polyacrylamide gel replaced the brine layer and allowed preservation of collected particles for microscopic imaging after recovery. During deployment 1, the NBST sampled at a depth of 200 m; during deployment 2, it sampled at a depth of 350 m.
\n
PELAGRA neutral sediment traps
\nThe Particle Export LAGRAngian (PELAGRA) trap was designed at the National Oceanography Centre, Southampton, UK (Lampitt et al. 2008) and consists of an arrangement of four conical traps (collection area 0.5 m2) around an APEX float (Teledyne-Webb Research, Inc.), with mechanically opening and closing collection cups. For this cruise, samples were collected from three PELAGRA traps: P4, P7, and P9. P4 and P7 each carried two conventional sediment funnels, two non-funnelled collectors for gel sampling and a camera/flash system for capturing time-lapse images of sinking particles. For these traps, the two cups situated beneath the conventional funnels were filled with the same brine solution used in the bottom of the NBST traps. Under the non-funnel collectors, jars containing polyacrylamide gel (described above) and commercially available cryogel were attached. P9 carried four conventional sediment funnels, each with a brine cup installed.
\n
Surface Tethered Trap (STT) arrays
\nAlongside the neutral traps was deployed a drifting mooring carrying cylindrical sediment trap tubes set up identically to those on the NBSTs as well as tubes of a different design provided by collaborator C. Lamborg (and henceforth termed modified PIT tubes). Modified PIT tubes are identical to standard PIT tubes (collection area 0.00385 m2) but with detachable bottoms. During deployment 1 two arrays were deployed, one at 200 m and one at 350 m. Both arrays contained two NBST-style brine-filled tubes, one NBST-style gel tube, and two modified PIT tubes which collected samples into 125-mL bottles filled with the same poisoned brine used in the other tubes. A programmable burnwire controller was set up to close the NBST-style tube lids at the same time as on the NBST traps. The burnwire controller at 200 m operated as planned but the controller at 350 m did not, due to a hardware failure. During the second deployment, a single trap array at 350 m was deployed using the fully-functioning burnwire controller. Two NBST-style tubes were set up to close at the same time as the NBSTs, two more were set up to remain open, and a third pair of modified PIT tubes (without lids) were included. During both deployments a Nortek current meter was deployed looking downwards approximately 2 m below the bottom of the 350-m trap array
\n
Upon platform retrieval, trap brine samples were processed as follows.
\nNBST samples and NBST-style tubes on STT: After a period of 1-3 hours to allow particles to finish settling in trap tubes, overlying filtered seawater was removed via peristaltic pump. The bottom brine layer was screened through 350-\u03bcm nylon mesh to aid in swimmer removal. The replicate brine tubes were drained through a single screen and combined. The screen was picked under 12x magnification to remove obvious swimmers while leaving behind passively sinking particles. Material remaining on the screen was rinsed back into the main sample while swimmers were filtered onto a QMA filter for later carbon and thorium-234 analysis. Combined trap samples were split eight ways using a custom rotary splitter. Splits were filtered onto QMA filters for C/N, PIC, and 234Th analysis or polycarbonate filters for biogenic Si and mass analysis on shore. Splits were also kept aside to return to collaborators labs at NOC.
\n
Modified PIT tubes on STT: Overlying seawater was siphoned off as above, then the 125-ml sample collection bottles were removed and combined into an extra NBST tube used as a dispenser. The sample was processed from this point identically to the NBST tubes.
\nPELAGRA trap samples: Brine cups were removed and either kept by the NOC lab for parallel processing (generally cup 1 on P4 and P7 and cups 3 and 4 on P9) or treated as described for modified PIT tubes above, minus the siphoning step.
\n
QMA filters were dried at 45C, mounted, and immediately counted for low-level \u03b2 emission onboard the ship. At WHOI, a subset of samples was re-counted within one month on shore. Final background counts to measure non-234Th related \u03b2 emissions were conducted several months later. At this point, QMA filters were unmounted, re-dried, and gravimetrically subdivided into four sections. One half of the filter was analyzed for total carbon and nitrogen after high-temperature combustion on a Thermo Electron FlashEA 1112 C/N analyzer. Coulometric analysis for PIC after sample acidification was performed on a quarter of the filter (Johnson et al, 1985; Honjo et al, 2000). The remainder of the filter was archived. At Skidmore College, polycarbonate filters for mass and bSi determination were dried and weighed repeatedly on a microbalance until stable weights with a precision better than 0.01 mg were achieved. Filter tare weights were subtracted and net mass accumulation was calculated. Then the filters were digested to release bSi using a weak alkaline digest (0.2 N NaOH for 2 hours at 95C) and analyzed following standard spectrophotometric methods (Strickland and Parsons, 1972).
\n
A replicate set of each type of trap collector was prepared as described above, held in the shipboard laboratory during each deployment, and then analyzed in parallel to provide a process blank determination. The blanks from the two deployments were averaged to determine the mean process blank for the cruise (Table 1).
For each analyte, the mean content of the process blank was subtracted from the corresponding sample value (Table 1), and blank-corrected values were normalized by the total number of splits, the trap collection area, and the deployment length to yield flux. The fluxes of each replicate sample at a given depth were averaged to yield the mean flux. Particulate organic carbon (POC) flux was computed as the difference between the mean TC flux and the mean PIC flux.
\n
BCO-DMO Data Manager Processing Notes:
\n* added a conventional header with dataset name, PI name, version date
\n* Added column "ISO_DateTime_start" from separate date|time columns
\n* changed format of date columns to ISO 8601 dates.
\n* In data version 3, trailing zeros were preserved for the appropriate level of precision.