5. The Measurement of Primary Production in Aquatic Ecosystems
81
shallow estuary that rates of oxygen exchange at
any given wind velocity were less than predicted
from the regression in Marino and Howarth
(1993), perhaps because the shallowness of the
water column reduced the amount of surface turbulence created by the wind shear.
In situ diel techniques work best when the correction for atmospheric exchange is relatively
small, since there can be a significant error associated with this correction (Howarth et al. 1992;
Marino and Howarth 1993). The exchange of oxygen with the atmosphere is a function both of wind
and of the gradient in partial pressures; thus, this
correction is least when oxygen concentrations are
near saturation levels. Consequently, the atmospheric exchange becomes greater as either GPP
becomes much greater than respiration (leading to
degassing of oxygen from the water to the atmosphere) or as respiration becomes much greater than
GPP (leading to large fluxes from atmosphere to
water). Further, for a given amount of production
per area, changes in oxygen concentration are
greater in more shallow water columns, leading to
steeper gradients with the atmosphere. Consequently, in situ diel techniques have the greatest
precision and accuracy in deeper water columns
and in ecosystems where GPP and respiration are
roughly equal and relatively high (Marino and Howarth 1993).
When applied to nonflowing waters, in situ diel
techniques can simply measure changes in O 2 or
CO2 over 24 hours at one location (Nixon and
Oviatt 1972; Oviatt et al. 1986). The situation is
more complex in flowing waters. One approach
has been to sample at only one location, and simply assume that the waters upstream are homogenous so that the one station represents the entire
ecosystem (Odum 1956). Another approach in rivers is to compare dissolved gas concentrations at
two stations, and infer rates of metabolism from
the difference between the two stations and a
knowledge of the time of travel for water between
the two stations (Owens 1969). A third approach
used in a tidal river with a high rate of mixing but
with water advection changing direction over the
tidal cycle is to measure diel changes at many stations spaced more broadly than the range of tidal
flow over a day and simply average the station
results (Howarth et al. 1992). In large, partially
mixed estuaries with complex circulation patterns,
the use of in situ diel techniques for estimating
GPP and respiration has proven problematical
(Kemp and Boynton 1980). However, recently
Swaney et al. (1999) have demonstrated a multistation technique in which oxygen concentrations
are regressed across stations as a function of depth,
salinity, and time. The time-dependent rate of
change of oxygen as a function of depth and salinity can be used to obtain reasonably highprecision estimates of GPP and respiration despite
complex estuarine circulation.
An analogous approach has been developed for
diel changes in particles in the oceanic water column (Siegel et al. 1989; Walsh et al. 1995). In this
approach, an optical instrument called a beam
transmissometer is used as a rapid profiler to collect
data on the vertical patterns in the attenuation of
light. The instrument emits a 25-cm-long beam of
light at 665-nm wavelength. The attenuation of this
beam of light is recorded and corrected for the attenuation of pure water to estimate the attenuation
from the scattering by particles. Samples of the suspended particles are collected on filters and compared with the optical measurement to determine a
regression between optical attenuation and particulate organic carbon.
The attenuation of light in oceanic waters sometimes shows a marked diel pattern with daytime
maxima and nighttime minima. The variable fraction of the particulate matter is likely due to biological production and removal terms, although
changes in cell size, particle aggregation, and other
processes can cause diel patterns that are not related
to biomass (Durand and Olson, 1996). After correction to suspended particulate variations with the
calculated regression, this approach allows an estimate of the diurnal fluctuations in particulate organic matter. These correlate well with the production measured by 14C incubations (Siegel et al.
1989). As with the gas methods, the nocturnal decrease is a measure of particle removal or respiration. The daylight increase is the total particle production (analogous to NPP). When corrected for the
nocturnal respiration (with the same assumptions as
above), the total production is probably analogous
to GPP. This method has a variety of assumptions
that are difficult to test. It has the advantage that
it is very easy to make and it can be done automatically from moorings or electronic profiling
packages.
81
shallow estuary that rates of oxygen exchange at
any given wind velocity were less than predicted
from the regression in Marino and Howarth
(1993), perhaps because the shallowness of the
water column reduced the amount of surface turbulence created by the wind shear.
In situ diel techniques work best when the correction for atmospheric exchange is relatively
small, since there can be a significant error associated with this correction (Howarth et al. 1992;
Marino and Howarth 1993). The exchange of oxygen with the atmosphere is a function both of wind
and of the gradient in partial pressures; thus, this
correction is least when oxygen concentrations are
near saturation levels. Consequently, the atmospheric exchange becomes greater as either GPP
becomes much greater than respiration (leading to
degassing of oxygen from the water to the atmosphere) or as respiration becomes much greater than
GPP (leading to large fluxes from atmosphere to
water). Further, for a given amount of production
per area, changes in oxygen concentration are
greater in more shallow water columns, leading to
steeper gradients with the atmosphere. Consequently, in situ diel techniques have the greatest
precision and accuracy in deeper water columns
and in ecosystems where GPP and respiration are
roughly equal and relatively high (Marino and Howarth 1993).
When applied to nonflowing waters, in situ diel
techniques can simply measure changes in O 2 or
CO2 over 24 hours at one location (Nixon and
Oviatt 1972; Oviatt et al. 1986). The situation is
more complex in flowing waters. One approach
has been to sample at only one location, and simply assume that the waters upstream are homogenous so that the one station represents the entire
ecosystem (Odum 1956). Another approach in rivers is to compare dissolved gas concentrations at
two stations, and infer rates of metabolism from
the difference between the two stations and a
knowledge of the time of travel for water between
the two stations (Owens 1969). A third approach
used in a tidal river with a high rate of mixing but
with water advection changing direction over the
tidal cycle is to measure diel changes at many stations spaced more broadly than the range of tidal
flow over a day and simply average the station
results (Howarth et al. 1992). In large, partially
mixed estuaries with complex circulation patterns,
the use of in situ diel techniques for estimating
GPP and respiration has proven problematical
(Kemp and Boynton 1980). However, recently
Swaney et al. (1999) have demonstrated a multistation technique in which oxygen concentrations
are regressed across stations as a function of depth,
salinity, and time. The time-dependent rate of
change of oxygen as a function of depth and salinity can be used to obtain reasonably highprecision estimates of GPP and respiration despite
complex estuarine circulation.
An analogous approach has been developed for
diel changes in particles in the oceanic water column (Siegel et al. 1989; Walsh et al. 1995). In this
approach, an optical instrument called a beam
transmissometer is used as a rapid profiler to collect
data on the vertical patterns in the attenuation of
light. The instrument emits a 25-cm-long beam of
light at 665-nm wavelength. The attenuation of this
beam of light is recorded and corrected for the attenuation of pure water to estimate the attenuation
from the scattering by particles. Samples of the suspended particles are collected on filters and compared with the optical measurement to determine a
regression between optical attenuation and particulate organic carbon.
The attenuation of light in oceanic waters sometimes shows a marked diel pattern with daytime
maxima and nighttime minima. The variable fraction of the particulate matter is likely due to biological production and removal terms, although
changes in cell size, particle aggregation, and other
processes can cause diel patterns that are not related
to biomass (Durand and Olson, 1996). After correction to suspended particulate variations with the
calculated regression, this approach allows an estimate of the diurnal fluctuations in particulate organic matter. These correlate well with the production measured by 14C incubations (Siegel et al.
1989). As with the gas methods, the nocturnal decrease is a measure of particle removal or respiration. The daylight increase is the total particle production (analogous to NPP). When corrected for the
nocturnal respiration (with the same assumptions as
above), the total production is probably analogous
to GPP. This method has a variety of assumptions
that are difficult to test. It has the advantage that
it is very easy to make and it can be done automatically from moorings or electronic profiling
packages.