78
Chapter 5: Nutrient Limitation: The Example of Iron
new phytoplankton production of about 114 M kg
−1 of nitrate. Further, it seems likely
that the observed range of Fe td concentrations (034–13 nM kg
−1 ) would allow uptake
of nitrate within the range of 60–195 M kg
−1 .
Allowing a modest input (modest, that is, over short time scale) from dry deposition,
it is reasonable to infer from this computation that the variable quantity of Fe relative to
N transported in vertical flux is likely to result in both iron-limited (high−S) and nitratelimited (low-S) regions in the surface waters of the oceans. This is, of course, just what we
observe and I venture to suggest that the variable NO 3 /Fe ratio just below the euphotic
zone must be regarded as the simplest possible explanation of the observations. Even
where a different macronutrient limits phytoplankton growth, the same argument may be
made; the preferential uptake of silicate by diatoms in the Subantarctic Zone of the Southern Ocean causes the water mass that is formed there in winter to be highly silica-deficient.
In fact, the relative flux of Fe td and other nutrients to the euphotic zone, across
either the nutricline or the sea surface, is rarely in the ratio required by plant cells, with
consequences that are both physiological and ecological. The study of the ecology of highS regions has greatly advanced understanding of the consequences for phytoplankton of
a nutrient supply in which Fe is relatively lacking. I shall argue that, while relative Fe
shortage in high-S regions does induce some characteristic features in their phytoplankton
ecology, it does not modify the regional distribution of chlorophyll accumulation: this is
forced principally by physics and by herbivory, both in high-S regions and elsewhere.
Regional Anomalies in Nutrient Limitation
Let us now examine the consequences of the distribution of iron and nitrate in some
characteristic regions. Of course, each of these regions will be discussed in more detail
in Chapters 9–12, but it will be useful here to deal specifically with the consequences of
differing subaerial flux of Fe in, especially, the high-S regions. The most accessible source
for defining the high-S regions is any convenient nutrient atlas; for the mean distribution
the 20 M kg
−1 contour for NO 3 at (say) 10 m may be taken as defining such areas.
The equatorial Pacific is a good entry into the characterization of high-S conditions.
However, although this region is often characterized as comprising only those areas
directly influenced by the equatorial upwelling itself, the area where NO 3 is >20 M kg
−1
at the surface is actually larger than that extending from 15
N to 15
S in the east and tapering westward to mid-ocean, so that significant parts of the subtropical gyres exhibit some
aspects of Fe limitation (Behrenfeld and Kolber, 1999). Most biological oceanographers
would now agree that the familiar pattern of productivity and chlorophyll accumulation
in the eastern tropical Pacific very largely represents the response of phytoplankton to
physical processes: however, in the early years of investigations of high-S regions this was
not the case, and it was supposed that the gyre field downstream from the Galapagos
was the one region in the HNLC waters of the eastern Pacific where phytoplankton
do bloom. This opinion was based on the analysis of Barber and Chavez (1991), who
suggested that rarefaction of the aerosol flux of Fe from Asiatic sources eastward across
the tropical Pacific at low latitudes had consequences for regional productivity. Barber
and Chavez suggested that productivity is a simple function of 60 m nitrate beyond
115
W, but is independent of nitrate in the eastern equatorial Pacific, and thus limited
to ∼36 mg Cmg chl
−1 d
−1 . Downstream of the Galapagos they postulated that Fe input
from slope sediments might maintain higher productivity. The only pattern of haze and
aerosol deposition over Pacific low latitudes then available (Duce and Tindale, 1991)
for this region rather poorly matches the pattern obtained from modern satellite AVHRR
data (e.g., Husar et al., 1997), or delivered by the GOCART model (see Fig. 5.2) of Ginoux
Chapter 5: Nutrient Limitation: The Example of Iron
new phytoplankton production of about 114 M kg
−1 of nitrate. Further, it seems likely
that the observed range of Fe td concentrations (034–13 nM kg
−1 ) would allow uptake
of nitrate within the range of 60–195 M kg
−1 .
Allowing a modest input (modest, that is, over short time scale) from dry deposition,
it is reasonable to infer from this computation that the variable quantity of Fe relative to
N transported in vertical flux is likely to result in both iron-limited (high−S) and nitratelimited (low-S) regions in the surface waters of the oceans. This is, of course, just what we
observe and I venture to suggest that the variable NO 3 /Fe ratio just below the euphotic
zone must be regarded as the simplest possible explanation of the observations. Even
where a different macronutrient limits phytoplankton growth, the same argument may be
made; the preferential uptake of silicate by diatoms in the Subantarctic Zone of the Southern Ocean causes the water mass that is formed there in winter to be highly silica-deficient.
In fact, the relative flux of Fe td and other nutrients to the euphotic zone, across
either the nutricline or the sea surface, is rarely in the ratio required by plant cells, with
consequences that are both physiological and ecological. The study of the ecology of highS regions has greatly advanced understanding of the consequences for phytoplankton of
a nutrient supply in which Fe is relatively lacking. I shall argue that, while relative Fe
shortage in high-S regions does induce some characteristic features in their phytoplankton
ecology, it does not modify the regional distribution of chlorophyll accumulation: this is
forced principally by physics and by herbivory, both in high-S regions and elsewhere.
Regional Anomalies in Nutrient Limitation
Let us now examine the consequences of the distribution of iron and nitrate in some
characteristic regions. Of course, each of these regions will be discussed in more detail
in Chapters 9–12, but it will be useful here to deal specifically with the consequences of
differing subaerial flux of Fe in, especially, the high-S regions. The most accessible source
for defining the high-S regions is any convenient nutrient atlas; for the mean distribution
the 20 M kg
−1 contour for NO 3 at (say) 10 m may be taken as defining such areas.
The equatorial Pacific is a good entry into the characterization of high-S conditions.
However, although this region is often characterized as comprising only those areas
directly influenced by the equatorial upwelling itself, the area where NO 3 is >20 M kg
−1
at the surface is actually larger than that extending from 15
N to 15
S in the east and tapering westward to mid-ocean, so that significant parts of the subtropical gyres exhibit some
aspects of Fe limitation (Behrenfeld and Kolber, 1999). Most biological oceanographers
would now agree that the familiar pattern of productivity and chlorophyll accumulation
in the eastern tropical Pacific very largely represents the response of phytoplankton to
physical processes: however, in the early years of investigations of high-S regions this was
not the case, and it was supposed that the gyre field downstream from the Galapagos
was the one region in the HNLC waters of the eastern Pacific where phytoplankton
do bloom. This opinion was based on the analysis of Barber and Chavez (1991), who
suggested that rarefaction of the aerosol flux of Fe from Asiatic sources eastward across
the tropical Pacific at low latitudes had consequences for regional productivity. Barber
and Chavez suggested that productivity is a simple function of 60 m nitrate beyond
115
W, but is independent of nitrate in the eastern equatorial Pacific, and thus limited
to ∼36 mg Cmg chl
−1 d
−1 . Downstream of the Galapagos they postulated that Fe input
from slope sediments might maintain higher productivity. The only pattern of haze and
aerosol deposition over Pacific low latitudes then available (Duce and Tindale, 1991)
for this region rather poorly matches the pattern obtained from modern satellite AVHRR
data (e.g., Husar et al., 1997), or delivered by the GOCART model (see Fig. 5.2) of Ginoux