| Popis: |
Trap array characteristics were monitored concurrent with particle collections for surface-tethered and bottom-moored cones and cylinders (MultiPITs) at the North Atlantic OFP-JGOFS site in the Sargasso Sea. At depths ranging from 145 to 3200 m, velocities of approaching fluid and those inside the traps were recorded at 5 Hz in bursts of 3–10 min every half hour during particle collections. A thermistor, a high resolution pressure gauge and two inclinometers concurrently monitored trap movements. Burst-averaged slip velocities experienced by both shallow and deep tethered traps reached 37 cm s−1, while a bottom-moored trap recorded 10-day averaged speeds of 4 cm s−1. Independent of deployment technique, for both cones and cylinders, flow cells inside the traps led to an intense flushing of fluid and particles. None of the surface-tethered traps tilted more than 8 degrees from vertical, even under strong flow accelerations. Tether-line motions, induced by the surface waves, generated high flow acceleration peaks of trap arrays at all depths, even for bungie-cord decoupled MultiPIT arrays. The flow cells inside traps were thus agitated with the result of intense turbulence prevailing close to the collection cup in the apex of tethered cones. Moored cone arrays recorded less dynamic environments. Trap fluxes by tethered cones were up to a factor of 8 smaller than by tethered MultiPITs at the same depth and time, cones collected more material with higher approaching fluid flows (untested so far for cylinders), and for the same conical geometry tethered traps collected less material than bottom-moored traps. The in situ deployments revealed substantial flow- and geometry-related differences in collection behavior among the different trap arrays, all of which deviated from steady-state flume simulation results. The diameter of the retention cup at the trap apex rather than the trap mouth diameter may be a controlling design parameter of particle collection rates for conical traps. Efforts to link trap and in situ fluxes require that hydrodynamics of individual trap arrays at depth are monitored, including line motions. Drift velocities rarely coincided with trap-experienced approach velocities. Trap simulation studies utilizing steady-state flume flows may be accurate only under very specialized conditions. Our data provide a hydrodynamic rationale for earlier recommendations by others of cylinders with adequate length-width ratio. |
