Hydrodynamics An Examination of Sediment Trap Accuracy Issues During VERTIGO John Andrews 1, Carl Lamborg *1, Steve Pike 1, Debbie Steinberg 2, Stephanie Wilson 2, Jim Valdes 1 and Ken Buesseler 1 1 Woods Hole Oceanographic Institution, MS #25, Falmouth, MA United States ( * corresponding author) 2 Virginia Institute of Marine Science, P.O. Box 1346, Gloucester Point, VA United States Paper Number OS 26A-05 Hydrodynamics. One of the largest potential sources of sediment trap collection inaccuracy is the flow of water past the collection tubes. This shear can occur with bottom moored arrays as water flows past the traps, or for surface tethered traps, as the devices are dragged through the water by wind stress on the surface float. Buesseler (1991) noted that this effect can result in both under and over collection of particles. The ultimate solution to this problem is the use of neutrally bouyant sediment traps that have no surface expression at all and therefore experience little or no shear as they drift with currents at depth. During VERTIGO, we deployed neutrally bouyant sediment traps (NBSTs) and surfaced tethered traps, drogued at depth (CLAP Traps). The CLAP Traps were deployed with an elastic “bungee” near the surface to minimize wave motion, and with current meters directly below the trap frame. These figures show the comparison between CLAP and NBST trap fluxes for total mass, particulate organic carbon (POC) and 234 Th. Comparing the 234 Th fluxes one sees CLAP traps over collect at ALOHA in a low flux environment and under collection in high flux (K2 deployment 1) and agreement in medium flux (K2 deployment 2). The same story is told in mass flux; a slight over collection at ALOHA in low flux conditions and the same under collection and agreement within the two K2 deployments. Buesseler, K.O., Do upper-ocean sediment traps provide an accurate record of particle flux? Nature 353, Clap Trap current meter data. The current meters recorded significant shear flows of water relative to the sediment traps. These show hydrodynamic effects during deployment 1 and 2 at K2. The tilt is directly related to horizontal speed as the surface buoy drags the trap. Both tilt and speed greatly affect the flow cell within a trap tube and its particle collection efficiency. In theory the NBST data would show no tilt and no speed. Shear flows were fairly similar during deployments 1 and 2, and yet there were substantial differences in the fluxes of a number of parameters from one deployment to the other. This suggests that hydrodynamics are just one part of determining trap accuracy, and that particle properties likely aggravate or alleviate this bias. Trap Poisons. It is common to add poison/preservatives of some kind to traps, held in a brine at the bottom. Due to different requirements for VERTIGO collaborators’ analyses, we used mercuric chloride in some trap tubes and formalin in others on both the CLAP Traps and NBSTs. On each array, two tubes of each type of poison/preservative were deployed and back aboard these two tubes were combined and split 8 ways. The compared results show no obvious bias for mass or carbon. Sediment traps are intended to collect sinking particles, but also collect “swimmers”, motile zooplankton that are thought to actively enter a trap and are caught in the brine/poison along with passively settling materials. Two common methods for swimmer removal are screening and picking. In VERTIGO, all trap samples were gravity fed through a 350 µm screen to separate swimmers (>350) from sinking particles, and screens were further rinsed with prefiltered seawater to remove adhering particles. Both the screens and solution containing trap particles were examined microscopically to test the accuracy of this separation. Additional “picking” was used to quantify smaller swimmers that passed through the screen (a positive flux bias for sinking material) and as well as sinking particles that were caught on the screen, either in association with swimmers and/or simply due to their larger size (a negative flux bias for sinking material). At both sites, swimmer associated carbon decreases with depth, following zooplankton abundance with depth. In 150m traps at both sites, more carbon is removed from each sample as swimmer associated material than remains in the sample as passively sinking material. This high fraction of swimmer carbon in the traps illustrates the importance of proper swimmer removal protocols in any upper ocean trap study. At ALOHA and K2, there was only a small fraction of swimmer carbon that passed through the screen (light green) relative to the sinking flux and larger swimmer fractions. This is generally a small fraction of the net C flux (<10% except K2, 150m). At K2, in the midst of a large diatom bloom, we also find that some of the material caught on the screen was in fact not swimmer material, sinking particles made up of larger diatoms and other phytoplankton debris and aggregates, i.e. we removed both swimmers and some sinking material using screening alone. Poisoned Particulate Solubilization Experiments were conducted to investigate losses of particle-phase elements as a result of in-trap solubilization. Only a few of the elements measured showed significant losses, and of these, only P and Sr approached a consistent behavior…with maximal correction factors of 30 and 60% during 5 day deployments (at 150 m), respectively. Suggests that for short deployment studies, solubilization is not a large artifact. During longer deployments, and even in poisoned traps, solubilization could be significant (e.g., Antia, 2005). Antia, A.N. (2005) Particle-associated dissolved elemental fluxes: revising the stoichiometry of mixed layer export. Biogeosciences Discussions, 2: Trap Poisons/Preservatives Zooplankton “Swimmers” Ideally, swimmers should be kept from collecting in the traps altogether. Peterson et al. (1993)* introduced the indented rotating sphere (IRS) valve as a mechanism to exclude swimmers from traps (photo is of our version of this design). The IRS ball rotates, shuttling particles that collect on the top surface into a lower collection area from which swimmers are excluded. We compared the number and type of swimmers caught at 150m during a 3 day deployment between a traditional open tube trap and the IRS trap. The IRS mechanism (15 min rotation cycle) succeeded in reducing by a factor of 5 to 10 the number of swimmers that are found in the trap sample (note large variability in swimmer number found in 3-6 separate trap samples). As a swimmer excluding device, the IRS holds considerable promise, however there is some concern that during the deployment, while the sinking material remains on the IRS ball surface for minutes (depending upon rotation cycle), it can be decomposed and/or eaten by zooplankton. In fact we see a slight decrease in the amount of material measured in two separate IRS traps relative to other samples from the same trap array at 150m. *Peterson, M. L., D. S. Thoreson, J. I. Hedges, C. Lee, and S. G. Wakeham Field evaluation of a valved sediment trap. Limnol. Oceanogr. 38: NBST vs. CLAP – Newer Ways to Catch the Rain Abstract Sediment traps have proven to be a valuable tool in the study of marine particle fluxes. However, a number of issues remain which may impact the quantity and/or quality of material collected in traps, and these issues are particularly important in applications of sediment traps in the upper m of the ocean. During VERTIGO (VERtical Transport In the Global Ocean), we have compared a variety of trapping systems including: surface tethered traps (similar to PITS traps used in many prior studies), a newly developed free vehicle trap, the neutrally buoyant sediment traps (NBSTs), and a tethered trap with an active swimmer avoidance mechanism, the IRS (indented rotating sphere) trap. These were deployed with duplication, at three depths in the mesopelagic "twilight zone" under contrasting flux conditions off Hawaii and in the NW Pacific. Here we present the results of a number of measurements designed to examine the impact of several potential influences on trap accuracy, including: hydrodynamic design of the trap; poison/preservative type and the influence of in-trap solubilization of collected material; and the magnitude and correction for zooplankton "swimmers" in different trap designs. CLAP Trap NBST Conclusions and Future Work: The NBST has removed significant hydrodynamic issues from the collection of sinking particles in the upper ocean. However, solubilization and swimmers remain a serious concern and are difficult to quantify. Combining a NBST with an IRS or other swimmer exclusion device is the next step in quantifying sinking particles. Such a device would resolve the trio of trap problems identified in many studies. By design, hydrodynamic biases and swimmers could be avoided. In addition, with a sealed top, one can study solubilization effects in situ, by comparing solution concentrations for the elements measured in the particle flux, unbiased by addition of particles and dissolved components derived from zooplankton, and with a sealed top, not impacted by variable dilution of brine/poison with ambient seawater during trap deployment and recovery that limits this approach in open tubes.