The Availability of Timely Global Oceanographic Data from Satellites
7
anomalies in the TOPEX-POSEIDON sea surface elevation to predict anomalies in the
depth of the 20
C isotherm, as a proxy for mixed-layer depth, and so to model changes
in new production rate using a relationship established by observation (Turk et al.,
2001). This particular analysis went further by using satellite-derived zonal wind stress
to demonstrate the primacy of nonlocal dynamics of the tropical ocean in determining
local production rates.
Finally, although it is not directly of relevance to us here because we are not concerned
directly with bathyal ecosystems I cannot refrain from noting the amazing—to those
not involved—recent advances in mapping the sea floor by satellite radar altimetry that
requires precision at the millimeter scale. Even today, of course, the solid surface of much
of our planet is mapped with a horizontal resolution of only about 15 km, and about
250 m in the vertical, whereas parts of the surface of Mars are mapped to resolutions of
1 km in the horizontal and 1 m in the vertical.
The global measurement of phytoplankton chlorophyll biomass in the ocean depends
on analysis of the small fraction of incident radiation that is not scattered or absorbed
at the surface or deeper and consequently may be observed as water-leaving irradiance
by a passive spectroradiometer. Both pure seawater and chlorophyll have, of course,
wavelength-specific characteristics: seawater absorbs red-green and reflects blue light,
while chlorophyll absorbs the deeply penetrating blue light and preferentially reflects
red-green wavelengths. Thus, reflected red-green light seen from space is proportional
to chlorophyll integrated over a large fraction of the first optical attenuation depth for
the relevant wavelengths, biased toward the surface when the chlorophyll profile is not
uniform. The sensed depth is Z e /46, where Z e is the euphotic depth (Morel and Berthon,
1989). Using the attenuation analysis of Smith (1981b), we can infer that sensed depths
range from about 25 m in clear oligotrophic water (0.1 mg chlorophyll m
−3 ) to about
5 m in eutrophic ocean water (10 mg chlorophyll m
−3 ).
Multispectral radiometers have the further capability of discriminating between some
taxon-specific combinations of individual chlorophyll pigments and of distinguishing
coccolithophore blooms by their white reflectance (Brown and Yoder, 1994). Global
maps of coccolith blooms show that these occur preferentially in the open NE Atlantic
and NE Pacific and also in some large shelf regions: the North Sea, off eastern Canada,
on the Falklands plateau, and off northern Australia. Other taxonomic discriminations
are now becoming possible; one such is the ability to classify the individual pixels
comprising a SeaWiFS or MODIS image as representing either a “diatom-dominated”
or a “mixed” population of generally smaller autotrophic cells (Sathyendranath et al.,
2004). The algorithm used to achieve this is based on the fact that the absorptive
properties of diatoms differ from those of other phytoplankton taxa, and also requires the
use of a theoretical reflectance model that relates water-leaving radiance to chlorophyll
concentration. We may then prepare maps to show the probability on a pixel-by-pixel
basis that the phytoplankton represented was diatom-dominated; such maps describe
the real dimensions of episodic diatom-dominated blooms. More recently, Alvain et al.
(2005) have analyzed a large set of relationships between different phytoplankton taxa
and the spectral characteristics of their water-leaving radiance based on the spectral
signatures derived from taxon-specific inventories of seven phytoplankton pigments.
Using these relationships, Alvain et al. are able to separate four major phytoplankton
groups in sea surface chlorophyll data from satellite sensors: haptophytes, Prochlorococcus,
Synechococcus-like bacteria, and diatoms. Experimental plots of relative distributions of
the four taxonomic groups thus obtained (e.g., Color plate 1) conform to our general
understanding of how phytoplankton taxa are distributed, obtained from surveys done
at sea. Diatoms really do dominate in spring blooms and coastal upwellings, whereas
Prochlorococcus- and Synechococcus-like bacteria dominate open-ocean, low-latitude gyral
situations and also the high-pigment signal from the South Subtropical Convergence zone
7
anomalies in the TOPEX-POSEIDON sea surface elevation to predict anomalies in the
depth of the 20
C isotherm, as a proxy for mixed-layer depth, and so to model changes
in new production rate using a relationship established by observation (Turk et al.,
2001). This particular analysis went further by using satellite-derived zonal wind stress
to demonstrate the primacy of nonlocal dynamics of the tropical ocean in determining
local production rates.
Finally, although it is not directly of relevance to us here because we are not concerned
directly with bathyal ecosystems I cannot refrain from noting the amazing—to those
not involved—recent advances in mapping the sea floor by satellite radar altimetry that
requires precision at the millimeter scale. Even today, of course, the solid surface of much
of our planet is mapped with a horizontal resolution of only about 15 km, and about
250 m in the vertical, whereas parts of the surface of Mars are mapped to resolutions of
1 km in the horizontal and 1 m in the vertical.
The global measurement of phytoplankton chlorophyll biomass in the ocean depends
on analysis of the small fraction of incident radiation that is not scattered or absorbed
at the surface or deeper and consequently may be observed as water-leaving irradiance
by a passive spectroradiometer. Both pure seawater and chlorophyll have, of course,
wavelength-specific characteristics: seawater absorbs red-green and reflects blue light,
while chlorophyll absorbs the deeply penetrating blue light and preferentially reflects
red-green wavelengths. Thus, reflected red-green light seen from space is proportional
to chlorophyll integrated over a large fraction of the first optical attenuation depth for
the relevant wavelengths, biased toward the surface when the chlorophyll profile is not
uniform. The sensed depth is Z e /46, where Z e is the euphotic depth (Morel and Berthon,
1989). Using the attenuation analysis of Smith (1981b), we can infer that sensed depths
range from about 25 m in clear oligotrophic water (0.1 mg chlorophyll m
−3 ) to about
5 m in eutrophic ocean water (10 mg chlorophyll m
−3 ).
Multispectral radiometers have the further capability of discriminating between some
taxon-specific combinations of individual chlorophyll pigments and of distinguishing
coccolithophore blooms by their white reflectance (Brown and Yoder, 1994). Global
maps of coccolith blooms show that these occur preferentially in the open NE Atlantic
and NE Pacific and also in some large shelf regions: the North Sea, off eastern Canada,
on the Falklands plateau, and off northern Australia. Other taxonomic discriminations
are now becoming possible; one such is the ability to classify the individual pixels
comprising a SeaWiFS or MODIS image as representing either a “diatom-dominated”
or a “mixed” population of generally smaller autotrophic cells (Sathyendranath et al.,
2004). The algorithm used to achieve this is based on the fact that the absorptive
properties of diatoms differ from those of other phytoplankton taxa, and also requires the
use of a theoretical reflectance model that relates water-leaving radiance to chlorophyll
concentration. We may then prepare maps to show the probability on a pixel-by-pixel
basis that the phytoplankton represented was diatom-dominated; such maps describe
the real dimensions of episodic diatom-dominated blooms. More recently, Alvain et al.
(2005) have analyzed a large set of relationships between different phytoplankton taxa
and the spectral characteristics of their water-leaving radiance based on the spectral
signatures derived from taxon-specific inventories of seven phytoplankton pigments.
Using these relationships, Alvain et al. are able to separate four major phytoplankton
groups in sea surface chlorophyll data from satellite sensors: haptophytes, Prochlorococcus,
Synechococcus-like bacteria, and diatoms. Experimental plots of relative distributions of
the four taxonomic groups thus obtained (e.g., Color plate 1) conform to our general
understanding of how phytoplankton taxa are distributed, obtained from surveys done
at sea. Diatoms really do dominate in spring blooms and coastal upwellings, whereas
Prochlorococcus- and Synechococcus-like bacteria dominate open-ocean, low-latitude gyral
situations and also the high-pigment signal from the South Subtropical Convergence zone