PARTICULATE ORGANIC MATTER I N SEA WATER
81
drifts into more or less circular patches five or ten metres across,
which readily identify these areas as points of convergence and confirm the fact that convection cells are indeed columnar.
The temperature is commonly 0-1-0.3°C cooler in the convergences
than in the surrounding water, and the temperature is essentially
constant with respect to depth. In the water between the convergences
a slight negative temperature gradient is likely to develop during these
calm periods. Often, particularly during the night and early morning
hours, a temperature inversion can be found just at the top of the
thermocline. This is a thin layer, generally less than half a metre thick,
in which the temperature is slightly warmer than that immediately
above it but generally cooler than the surface water. I n order to be
stable in the position in which it is found, this water also must be
slightly more saline than the water just above it.
On one occasion the writer took a series of closely spaced bathytherms as the ship drifted slowly across and beyond one of these convergences. The temperature in the downdraft was the same as in the
inversion beyond, and the profile gave the impression that the latter
was derived by spreading from the base of the convergence. There
seems to be little doubt that this water is the product of surface cooling
and evaporation.
Particulate matter in a thin layer of this sort will not have to sink
very far in order to be trapped in the thermocline, and it will be apparent that the rate of loss from the mixed layer will depend more on the
rate of convective overturn than on the actual sinking rate of the
particles.
A direct estimate of rates of loss is impossible in the present state
of our knowledge. However, the problem is essentially the same as
that of removal of phytoplankton from the mixed layer, a subject that
has been given some attention over the years.
Riley et al. (1949) developed some mathematical models of the
vertical distribution of phytoplankton, and in order to obtain a realistic
vertical distribution it was necessary to assume a sinking rate of about
3 m/day in temperate waters and 6 m/day in the subtropics. Steele
(1958) and Riley (1965) used similar assumptions in further models
that have been developed.
The assumption of a regional variation in mean sinking rates was
adopted on the grounds that variations in viscosity associated with
regional temperature differences could have this effect. However,
measured sinking rates of phytoplankton vary enormously, and most
of them are less than the values postulated here. The idea that average
sinking rates of natural populations should be so nearly constant, or
81
drifts into more or less circular patches five or ten metres across,
which readily identify these areas as points of convergence and confirm the fact that convection cells are indeed columnar.
The temperature is commonly 0-1-0.3°C cooler in the convergences
than in the surrounding water, and the temperature is essentially
constant with respect to depth. In the water between the convergences
a slight negative temperature gradient is likely to develop during these
calm periods. Often, particularly during the night and early morning
hours, a temperature inversion can be found just at the top of the
thermocline. This is a thin layer, generally less than half a metre thick,
in which the temperature is slightly warmer than that immediately
above it but generally cooler than the surface water. I n order to be
stable in the position in which it is found, this water also must be
slightly more saline than the water just above it.
On one occasion the writer took a series of closely spaced bathytherms as the ship drifted slowly across and beyond one of these convergences. The temperature in the downdraft was the same as in the
inversion beyond, and the profile gave the impression that the latter
was derived by spreading from the base of the convergence. There
seems to be little doubt that this water is the product of surface cooling
and evaporation.
Particulate matter in a thin layer of this sort will not have to sink
very far in order to be trapped in the thermocline, and it will be apparent that the rate of loss from the mixed layer will depend more on the
rate of convective overturn than on the actual sinking rate of the
particles.
A direct estimate of rates of loss is impossible in the present state
of our knowledge. However, the problem is essentially the same as
that of removal of phytoplankton from the mixed layer, a subject that
has been given some attention over the years.
Riley et al. (1949) developed some mathematical models of the
vertical distribution of phytoplankton, and in order to obtain a realistic
vertical distribution it was necessary to assume a sinking rate of about
3 m/day in temperate waters and 6 m/day in the subtropics. Steele
(1958) and Riley (1965) used similar assumptions in further models
that have been developed.
The assumption of a regional variation in mean sinking rates was
adopted on the grounds that variations in viscosity associated with
regional temperature differences could have this effect. However,
measured sinking rates of phytoplankton vary enormously, and most
of them are less than the values postulated here. The idea that average
sinking rates of natural populations should be so nearly constant, or