How does one make a patch of something in the turbulent, surface
waters of the ocean, and follow it long enough to detect a
biological signal? When the iron experiment was being planned, it
was not clear if this was even possible---in fact, Martin designed
the expedition simply to test if such a patch could be created and
followed. The ambient iron concentration at the experimental site
was 0.1 nM Fe (= 10
moles per liter), and the target
concentration for the patch was 4 nM Fe. To this end, 7,800 moles
of Fe (as FeSO
) were added to 15,600 liters of seawater with
enough concentrated HCl to bring it to pH 2 (iron is very
insoluble in seawater). This mixture was dispensed into the ship's
propeller wash as it steamed back and forth at 9 km hr
,
creating a 64 km
patch over a 24-hour period. Horizontal
mixing homogenized the surface streaks of iron within one day, and
convective overturn of the mixed layer supplied the additional
force necessary to mix the iron down to 35 m in the first 24 hours
of the experiment (Martin et al., 1994).
The patch steadily drifted in absolute position, thus a buoy
equipped with a Global Positioning System (GPS) receiver was
deployed at the center of the patch, and served as the ship's
reference point throughout the experiment (Fig. 4). In addition,
4 drifter buoys were deployed in the four corners of the patch to
keep track of the periphery (Fig. 4). Although these buoys served
as a crude reference, they could not be relied upon as indicators
of when one was ``in'' and ``out'' of the iron-enriched area,
because it was impossible to know a priori how much lateral
mixing would take place. Since iron was consumed by the
phytoplankton throughout the experiment, it could not be used as a
tracer for the patch. Thus the inert tracer, sulfur hexafluoride
(SF
), was added to the propeller wash along with the iron
(Martin et al., 1994).
The patch was followed for 9 days. During the first five days it maintained its integrity surprisingly well (Fig. 4). On the 6th day a low salinity front moved in from the east, subducting the patch to 20 to 40 m. Although it was still trackable under the surface, the experiment was compromised because the subduction subjected the phytoplankton to a light-shift, thus introducing a new variable (Martin et al., 1994).