versions with a carousel of cups changed by a stepping motor at suitable intervals (8.5
days, 21 days, etc.). After a collecting interval, the whole mooring array, often several
of these traps at different depths, is released from its anchors by a signal to an
acoustically activated hook and comes up in a tangle. That is sorted out, the cups
removed, and the samples divided into aliquots while wet. Those splits are subjected
to various chemical and biological analyses.
(^) There have been several attempts to summarize the results (Lampitt & Antia 1997;
Berelson 2001; Lutz et al. 2002; Honjo et al. 2008). It is not possible to get good
estimates of organic matter flux from traps just above the seafloor, because currents
near the bottom resuspend sediments and mix them upward for hundreds of meters.
However, the rate of resuspension slows sufficiently above those layers that good
estimates of the final seafloor input can be obtained from traps at about 3000 meters.
Almost all of the results (Fig. 13.24a) fall in the range from 0.1 to 2 mol C m−2 d−1
(1.2 to 24 mg C m−2 d−1). This is a small fraction of the primary production at the
surface. Berelson (2001) compared all of the PARFLUX-type trapping studies from
the process studies cruises of JGOFS program to the simultaneous production rate
estimates (Fig. 13.24b & Fig. 13.25) and found that ∼0.5 to 1% of the surface
photosynthate generated in the euphotic zone reaches 3000 meters or the bottom. That
result is typical for all studies before and since. Consumption in the water column
follows a vertical sequence (Fig. 13.25) reasonably well expressed by a function
suggested by Martin et al. (1987), the “Martin curve”:
(^) where b is a fitted parameter. The flux at 100 or 200 m, often termed the “export
flux”, is measured by a trap there. All of the biases of trapping are extreme above 100
m (or a bit deeper), so that the comparison is made to this somewhat arbitrary
standard for the upper level trapping rate, rather than to shallower rates or directly to
primary production. As seen in Berelson’s figure (Fig. 13.25) the best value of b
varies from ∼0.6 to twice that, and very often the decrease below 2000 m is slower
than the fitted equation. The slowing, whether the function fits exactly or not, is due
to (i) decrease in the populations of scavengers at depth reprocessing sinking marine
snow and fecal matter, and (ii) rather sharp downward acceleration of particles.
Sinking rates of typical larger particles (those carrying most of the flux) are 100 to
200 m per day, accelerating by ∼50% below that. Mostly, that is attributable to
agglomeration of particles, and larger particles sink faster (Stokes 1851). It is widely
accepted that sinking of small organic particles is largely due to several forms of
agglomeration: (i) in “marine snow” that is at least partly a matrix of polymers
secreted by algae and animals, and (ii) fecal pellets of zooplankton (small but dense
from copepods, large from salps, etc.) and nekton. The ocean is frequently full of
flocculent, filmy stuff, first termed marine snow by William Beebe who observed it