An important question, on which estimates disagree, and which was
first addressed quantitatively by Smith and Prezelin
[Ray Smith Prezelin macintyre 1992],
is what is the loss, if any, of primary production in the Southern Ocean
due to ozone-related enhanced UV-B and would this loss have any feedback
effect on atmospheric 
?
Differences in the spectral properties of field incubators used by various
groups to measure UV effects complicates direct comparison of results since,
if results are not first normalized with common weighting functions,
relatively minor spectral differences in filters can cause quantitative
differences in reported results.
Prezelin and co-workers
[Prezelin Boucher Schofield 1994]
have normalized and summarized the work of several austral spring data
bases of UV-B inhibition of primary production in the Southern Ocean and
show that results, once normalized, are in relatively close agreement.
They show that in the absence of an ozone hole (>320 DU), the ambient UV-B
radiation suppresses daily primary production by at least 4% to 7% in
surface waters (2 m) and by about 12% in near surface waters (5-10 m).
This is the background influence of UV-B against which ozone-related effects
of enhanced UV-B must be compared.
Extrapolation of this information in order to estimate effects to the whole
Southern Ocean remains controversial.
Holm-Hansen, Helbling and co-workers [Helbling Villafane Ferrario Holm-Hansen 1992, Helbling Villafane Holm-Hansen 1994] estimate the loss of primary productivity for all the ice-free waters south of the Polar front. As a simplification, they treat all oceanic areas (open ocean, marginal ice zone, shelf and coastal waters) as equivalent, although the various water types in the Antarctic are significantly different in terms of hydrography and productivity [Walker Smith sakshaug 1990, holm-hansen franceschini cuhel 1977] and likely in sensitivity to UV-B. With this simplification and taking into consideration the magnitude of ozone depletion and its space/time variations, they estimate losses in primary productivity due to ozone-related enhanced UV-B to be less than 0.15% for the entire year.
Smith, Prezelin and co-workers
[Ray Smith Prezelin macintyre 1992, Prezelin Boucher Smith 1994,
Prezelin Boucher Schofield 1994]
first focused their attention on the marginal ice zone (MIZ) within the
Southern Ocean.
For pelagic waters surrounding the MIZ of the Southern Ocean, current
estimates of annual primary productivity are about
[walker Smith sakshaug 1990, elsayed turner 1977,
holm-hansen franceschini cuhel 1977].
Productivity estimates for the MIZ were derived by W.O. Smith and Nelson
[nelson 1986 importance]
[1986] using a simple model of ice-edge bloom genesis to be about
or about a 40% of the total MIZ plus pelagic
production south of the Antarctic Convergence.
Thus, the MIZ potentially plays a major role in the ecological and
biochemical cycles of the Southern Ocean.
Smith & Prezelin make an estimate of the impact of reduced ozone
on primary production for the MIZ of the Bellingshausen
Sea (Fig. 2) based upon a determination of phytoplankton productivity
data averaged for inside and outside the 
hole.
Again, it is important to note, that this simple comparison of production
inside vs. outside the 
hole avoids complicating assumptions
and focuses on the consequences to the phytoplankton community to the
increased UV-B inside the hole.
Based on these in situ data, a yearly estimate of production loss for
the MIZ of the Southern Ocean can be made by assuming that the loss they
measure is representative of the MIZ and integrating production over the
MIZ area and over the 3-month duration of the 
hole during
Antarctic spring.
They estimate (using a 6% loss of water column productivity and
conservatively assuming a given location is outside the 
hole
one-half of the time) that this productivity loss to the MIZ is
,
corresponding to about 2% of the estimated yearly production of the MIZ.
Their assumptions are such that this is a minimum loss estimate and values
could be at least two times higher depending upon the specific space-time
extent of the 
hole.
They note that they used short-term 
C studies
to assess changes in natural communities of phytoplankton caused by variations
in the ozone hole which occurred on time scales from hours to weeks.
Thus, because the water column was not actively mixing and fixed-depth
incubations of several hours were appropriate, the time scale of their
experimental protocol matched that of the processes observed.
However, caution must be used when inferring longer-term
ecological consequences from short-term observations
[Ray Smith baker science 1980].
Likewise, in environments where vertical mixing imposes variations of UV
on time scales much less than the incubation time, the possibility of
artifactual overestimation of UV effects should be considered
[Cullen Neale 1993].
The interannual variability of primary production of the MIZ has been
estimated to be substantial
[keene comiso 1988],
and such that the maximum productivity is 50% greater than the minimum.
This variability associated with the annual advance and retreat of pack ice
is thought to be a major physical determinant of space-time changes in the
structure and function of polar biota
[ainley fraser sullivan torres 1986, fraser ainley bioscience 1986,
sharon Smith vidal 1986, nelson 1986 importance,
walsh mcroy 1986, garrison 1987,
ainley fraser daly 1988, walker Smith polar oceanography 1990 book].
In particular, this interannual variability is likely to have a significant
effect on total annual primary production, although to date these natural
changes have not been accurately quantified.
Thus, Smith and Prezelin note that their estimate of (2 to 4%) loss to MIZ
productivity should be viewed in the context of a presumed natural
variability of 
25%.
Concern has been expressed
[voytek 1990]
that 
-induced phytoplankton loss may trigger a positive
feedback with respect to atmospheric
that would exacerbate the greenhouse effect.
The estimated loss of
is about 3 orders of magnitude smaller than estimates of global phytoplankton
production and thus is not likely to be significant in this context.
Furthermore, Peng [Peng 1992] using a global circulation model,
found negligible global effect (with respect to 
) of turning off
all phytoplankton production in the Southern Ocean.
On the other hand, the finding that the 
-induced loss to a natural
community of phytoplankton in the MIZ is measurable, leaves the
ecological consequences of the magnitude and timing of this early spring
loss as something to be determined.