The air–sea CO 2 -flux map (Plate 16.1) also shows this biological carbon transferinto the sea. Much influx to the ocean interior occurs in regions without deep-water
formation: the subarctic North Atlantic, subarctic Pacific and those subantarctic areas
just downwind of land. In those regions, the substantial influx of CO 2 to the ocean
interior is by the “biological pump” (a term from Longhurst & Harrison 1989). In
these areas, surface CO 2 is reduced because photosynthesis removes it from the water,
increasing the atmosphere-to-water gradient and the flux into the ocean. Respiration
does not balance phytoplankton uptake of CO 2 in surface layers across these regions,
because sinking particulates (from tiny fecal pellets to dead whales) and vertical
migrations move carbon out of the surface to accumulate at depth. Below the
permanent pycnoclines typical of mid-latitudes or the haloclines in subarctic and
subantarctic regions, these particles eventually oxidize, adding to the supersaturation
of mid-waters with CO 2 . Over intervals longer than just the industrial era, changes in
atmospheric CO 2 , as will be seen for glacial–interglacial cycling, certainly involve
changes in operation of both the physical and biological carbon pumps.
(^) The data upon which the flux map is based make possible areal summations of input
and output (Gruber et al. 2009; Plate 16.2). There are four ways to calculate these
sums, represented by the bars of different colors in the plate. While the areas to which
the bars apply are not explicit, it is clear that input to the ocean is greater than output,
a very large proportion of which is located in the tropical Pacific. Within their errors,
the methods agree with the O 2 /N 2 method that the net uptake in the period of the
surveys (1995) was ∼2 GtC yr−1. Approximate chemical calculations allow separation