402
and the stoichiometry of planktonic food webs.
Physicallimnologists and oceanographers have the
advantage of being able to generate hydrodynamic
data at high spatial and temporal resolution, while
most biological data are only available on coarse
scales due to methodological limitations. The situation is the same for stoichiometric data, as analytical methods for elemental composition are slow
and labor-intensive. One methodology that holds
promise for reducing the mismatch between the intensities of hydrodynamic and stoichiometric observations is near-infrared reflectance spectroscopy
(NIRS). Developed largely for food product analysis, this approach analyzes the infrared signatures
produced by covalent bond structures of organic
materials and, when properly calibrated, provides
information on elemental composition (C, N, P) as
well as on composition of major biomolecules (protein, lipid, carbohydrate, nucleic acid). Furthermore, the analysis is rapid «5 minutes per sample)
and nondestructive so that the sample remains for
archiving or for analysis of other constituents. The
approach has recently been applied in a limnological setting by Malley et al. (1993), who found good
agreement between particulate samples analyzed by
NIRS and by conventional elemental analysis.
They also successfully analyzed sediment samples
and Malley indicates that the approach should be
applicable to mixed or species-specific zooplankton
samples. Thus, NIRS may allow real-time, highintensity observations of stoichiometric properties
in limnological and oceanographic settings, reducing sample analysis costs at the same time.
Improved methodological and analytical capabilities may greatly facilitate analysis of pelagic
ecosystems in stoichiometric terms. However, technical issues are not the primary constraint on progress in this realm. More important will be the rate
at which pelagic scientists begin to apply the conceptual and theoretical tools described in this chapter (and elsewhere) to incorporate the role of stoichiometric constraints into their view of nutrient
cycling and trophic dynamics.
Applications
How might improved understanding of pelagic ecosystems obtained via stoichiometric analysis be
usefully applied in addressing problems of human
James J. Elser
concern? We have already seen how the model of
Andersen (1997) has direct application to problems
of eutrophication involving increased P loading to
lakes. Indeed, stoichiometric approaches are already in direct application in a major project that
seeks to analyze and model eutrophication processes in the Mediterranean Sea, the Baltic Sea, and
the Northeast Atlantic Ocean (COMWEB project).
Stoichiometric thinking may be particularly important in understanding and making accurate predictions concerning the impacts of perturbations that,
as human influences often do, alter more than one
key factor at a time. For example, consider the effects of global climate change on water column
structure, hydrodynamics, and nutrient dynamics in
pelagic ecosystems. Stoichiometric analysis suggests that climate change effects on hydrodynamics
may push many pelagic ecosystems from the
"quantity" world to the "quality" world, with potentially serious consequences for food web performance. First, consider marine systems where it
has been suggested that global warming may create
warmer and shallower (and thus brighter) surface
mixed layers with more sluggish circulation and
thus reduced nutrient availability (Broecker 1982;
Mullin 1993; Rowe and Baldauf 1995; Karl et al.
1995). For example, EI Nino conditions in the
North Pacific SUbtropical gyre were associated with
a shallower, less turbulent mixed layer, a shift from
N- to P-limitation of production, and increased
seston CIP (Karl et al. 1995, 1997). This response
is consistent with the data for the lake-size series
analyzed by Sterner et al. (1997) indicating that reduced turbulence in pelagic ecosystems may accentuate autotroph C/nutrient imbalance, conceivably
impairing secondary production. This possibility is
particularly intriguing given recently documented
long-term declines in marine zooplankton biomass
(Roemmich and McGowan 1995). Next, consider
lakes. Climate change may affect lakes in the same
way as described above for oceanic water columns:
creating shallower, brighter mixed layers with reduced nutrient supplies, elevating C/nutrient ratios
of primary production and imposing stoichiometric
constraints on food web performance. A potential
example of this can be seen in the study of Schindler et al. (1990) of a boreal forest lake during a 20year period of climatic warming. During this period, watershed runoff decreased, resulting in lower
inputs of colored dissolved organic carbon (DOC)
and the stoichiometry of planktonic food webs.
Physicallimnologists and oceanographers have the
advantage of being able to generate hydrodynamic
data at high spatial and temporal resolution, while
most biological data are only available on coarse
scales due to methodological limitations. The situation is the same for stoichiometric data, as analytical methods for elemental composition are slow
and labor-intensive. One methodology that holds
promise for reducing the mismatch between the intensities of hydrodynamic and stoichiometric observations is near-infrared reflectance spectroscopy
(NIRS). Developed largely for food product analysis, this approach analyzes the infrared signatures
produced by covalent bond structures of organic
materials and, when properly calibrated, provides
information on elemental composition (C, N, P) as
well as on composition of major biomolecules (protein, lipid, carbohydrate, nucleic acid). Furthermore, the analysis is rapid «5 minutes per sample)
and nondestructive so that the sample remains for
archiving or for analysis of other constituents. The
approach has recently been applied in a limnological setting by Malley et al. (1993), who found good
agreement between particulate samples analyzed by
NIRS and by conventional elemental analysis.
They also successfully analyzed sediment samples
and Malley indicates that the approach should be
applicable to mixed or species-specific zooplankton
samples. Thus, NIRS may allow real-time, highintensity observations of stoichiometric properties
in limnological and oceanographic settings, reducing sample analysis costs at the same time.
Improved methodological and analytical capabilities may greatly facilitate analysis of pelagic
ecosystems in stoichiometric terms. However, technical issues are not the primary constraint on progress in this realm. More important will be the rate
at which pelagic scientists begin to apply the conceptual and theoretical tools described in this chapter (and elsewhere) to incorporate the role of stoichiometric constraints into their view of nutrient
cycling and trophic dynamics.
Applications
How might improved understanding of pelagic ecosystems obtained via stoichiometric analysis be
usefully applied in addressing problems of human
James J. Elser
concern? We have already seen how the model of
Andersen (1997) has direct application to problems
of eutrophication involving increased P loading to
lakes. Indeed, stoichiometric approaches are already in direct application in a major project that
seeks to analyze and model eutrophication processes in the Mediterranean Sea, the Baltic Sea, and
the Northeast Atlantic Ocean (COMWEB project).
Stoichiometric thinking may be particularly important in understanding and making accurate predictions concerning the impacts of perturbations that,
as human influences often do, alter more than one
key factor at a time. For example, consider the effects of global climate change on water column
structure, hydrodynamics, and nutrient dynamics in
pelagic ecosystems. Stoichiometric analysis suggests that climate change effects on hydrodynamics
may push many pelagic ecosystems from the
"quantity" world to the "quality" world, with potentially serious consequences for food web performance. First, consider marine systems where it
has been suggested that global warming may create
warmer and shallower (and thus brighter) surface
mixed layers with more sluggish circulation and
thus reduced nutrient availability (Broecker 1982;
Mullin 1993; Rowe and Baldauf 1995; Karl et al.
1995). For example, EI Nino conditions in the
North Pacific SUbtropical gyre were associated with
a shallower, less turbulent mixed layer, a shift from
N- to P-limitation of production, and increased
seston CIP (Karl et al. 1995, 1997). This response
is consistent with the data for the lake-size series
analyzed by Sterner et al. (1997) indicating that reduced turbulence in pelagic ecosystems may accentuate autotroph C/nutrient imbalance, conceivably
impairing secondary production. This possibility is
particularly intriguing given recently documented
long-term declines in marine zooplankton biomass
(Roemmich and McGowan 1995). Next, consider
lakes. Climate change may affect lakes in the same
way as described above for oceanic water columns:
creating shallower, brighter mixed layers with reduced nutrient supplies, elevating C/nutrient ratios
of primary production and imposing stoichiometric
constraints on food web performance. A potential
example of this can be seen in the study of Schindler et al. (1990) of a boreal forest lake during a 20year period of climatic warming. During this period, watershed runoff decreased, resulting in lower
inputs of colored dissolved organic carbon (DOC)