Models of Regional Nutrient Flux and Limitation
83
Southern Ocean, it seems improbable that enhanced Fe arising in this manner could be
conserved within frontal jets over long distances. It appears more likely that the vertical
motion and convergence associated with frontal dynamics will ensure a continual supply
of Fe from the deeper water masses below the ferricline. In this case, the rate of supply
will respond to relative strength of meandering of the fronts; perhaps the strengthening of
chlorophyll biomass downstream of shallow topography is due to this mechanism rather
than to supply from shoal water. It is also surely significant that meandering fronts are
rich in sloping, relatively shallow stratification that may serve to sustain blooms once
initiated. All this must be integrated with the observations that, between the annular
frontal zones, each Fe fertilization experiment (though all done in the Atlantic sector)
has succeeded in generating an ephemeral diatom bloom.
I have not dealt here with the situation in the coastal biome of the Southern Ocean
and will note in passing that irradiance is very frequently limiting here even in summer,
because of much cloudiness and low sun angle, yet near-surface stratification is widely
induced by meltwater, itself relatively high in dissolved Fe. Though not well observed by
satellite sensors, surface blooms are, not unexpectedly, relatively widespread here.
How then should we evaluate the balance between the several mechanisms of bloom
inception and limitation in the Southern Ocean? Certainly, it will be essential to examine
the conflicting possibilities in the light of William Dickinson’s evocation of scientific
myths. Perhaps the general model of Boyd (2002) is the best guide yet available: this
suggests a seasonal progression of limitation and colimitation of Fe, irradiance, and
silicate so that irradiance limits in all months except mid-October to mid-March and Fe
is limiting only in that period alone, while silicate may be limiting, in some regions and
for some cells, in October and November.
Models of Regional Nutrient Flux
and Limitation
So far, I have made little or no reference to global or regional models of production
processes that involve Fe inputs from aerosol or upward mixing, largely because no
consensus has been reached by modeling studies. They have not greatly advanced our
understanding of the problem, although their indications cannot be ignored.
Because it directly supports the concept that I believe has become mythic, it will
be useful to address at once the model of Erickson et al. (2003). These authors merge
SeaWiFS data with the output from the global model of dust-deposition of Ginoux et al.
(2001) that, incidentally, they describe as a “state-of-the-art geophysical data set.” They
obtain fields of anomaly correlation between surface chlorophyll and dust deposition at
2
× 25
resolution over the Southern Ocean, where the anomalies are the time-series
mean subtracted from each monthly value. Correlations of >06 are obtained for a
zonal swath from 40 to 55
S across the entire ocean, but principally in the Atlantic and
western Indian Ocean. The location of high correlation is commonly displaced by 2–5
of latitude from locations of chlorophyll maxima, and so Ekman transport is invoked to
explain the resultant bloom. Their analysis, they suggest, elucidates “the spatial response
of chlorophyll to iron flux” and identifies “those regions where ocean biology is possibly
tightly coupled with atmospheric Fe deposition.”
Really, of course, it does neither of these things. Statistical correlation between two
factors that have been selected from among the very many that are involved in a complex ecological process cannot, of itself, demonstrate causation: that much we learn in
elementary statistics courses. So I remain unimpressed by apparent correlation between
dust deposition and those places where phytoplankton growth and accumulation is most
83
Southern Ocean, it seems improbable that enhanced Fe arising in this manner could be
conserved within frontal jets over long distances. It appears more likely that the vertical
motion and convergence associated with frontal dynamics will ensure a continual supply
of Fe from the deeper water masses below the ferricline. In this case, the rate of supply
will respond to relative strength of meandering of the fronts; perhaps the strengthening of
chlorophyll biomass downstream of shallow topography is due to this mechanism rather
than to supply from shoal water. It is also surely significant that meandering fronts are
rich in sloping, relatively shallow stratification that may serve to sustain blooms once
initiated. All this must be integrated with the observations that, between the annular
frontal zones, each Fe fertilization experiment (though all done in the Atlantic sector)
has succeeded in generating an ephemeral diatom bloom.
I have not dealt here with the situation in the coastal biome of the Southern Ocean
and will note in passing that irradiance is very frequently limiting here even in summer,
because of much cloudiness and low sun angle, yet near-surface stratification is widely
induced by meltwater, itself relatively high in dissolved Fe. Though not well observed by
satellite sensors, surface blooms are, not unexpectedly, relatively widespread here.
How then should we evaluate the balance between the several mechanisms of bloom
inception and limitation in the Southern Ocean? Certainly, it will be essential to examine
the conflicting possibilities in the light of William Dickinson’s evocation of scientific
myths. Perhaps the general model of Boyd (2002) is the best guide yet available: this
suggests a seasonal progression of limitation and colimitation of Fe, irradiance, and
silicate so that irradiance limits in all months except mid-October to mid-March and Fe
is limiting only in that period alone, while silicate may be limiting, in some regions and
for some cells, in October and November.
Models of Regional Nutrient Flux
and Limitation
So far, I have made little or no reference to global or regional models of production
processes that involve Fe inputs from aerosol or upward mixing, largely because no
consensus has been reached by modeling studies. They have not greatly advanced our
understanding of the problem, although their indications cannot be ignored.
Because it directly supports the concept that I believe has become mythic, it will
be useful to address at once the model of Erickson et al. (2003). These authors merge
SeaWiFS data with the output from the global model of dust-deposition of Ginoux et al.
(2001) that, incidentally, they describe as a “state-of-the-art geophysical data set.” They
obtain fields of anomaly correlation between surface chlorophyll and dust deposition at
2
× 25
resolution over the Southern Ocean, where the anomalies are the time-series
mean subtracted from each monthly value. Correlations of >06 are obtained for a
zonal swath from 40 to 55
S across the entire ocean, but principally in the Atlantic and
western Indian Ocean. The location of high correlation is commonly displaced by 2–5
of latitude from locations of chlorophyll maxima, and so Ekman transport is invoked to
explain the resultant bloom. Their analysis, they suggest, elucidates “the spatial response
of chlorophyll to iron flux” and identifies “those regions where ocean biology is possibly
tightly coupled with atmospheric Fe deposition.”
Really, of course, it does neither of these things. Statistical correlation between two
factors that have been selected from among the very many that are involved in a complex ecological process cannot, of itself, demonstrate causation: that much we learn in
elementary statistics courses. So I remain unimpressed by apparent correlation between
dust deposition and those places where phytoplankton growth and accumulation is most