5. The Measurement of Primary Production in Aquatic Ecosystems
79
cant changes in the community structure within the
bottle compared with the natural system, particularly in the bacteria. Some of these changes are due
to the growth of organisms that live on surfaces (the
wall of the bottle) and thus not representative of the
purely aqueous system. The large aggregate particles called "marine snow" become destroyed by the
collection process or settle to the bottom of the
sampler and are underrepresented in the sample
(Gardner 1977). Further, the incubation bottles
rarely exceed a liter in volume. Large grazing animals will not be included in the incubation bottles
and the change in grazing pressure may affect rates
of production (Collos et al. 1993).
In Situ Diel Approaches
In situ diel methods use changes in dissolved O2,
CO 2 or particle matter over a 24-hour period to
infer rates of GPP and respiration. Where this approach is feasible, it is often preferable to bottleincubation methods since natural levels of turbulence, grazing, nutrient availability, and light fields
are maintained and since accurate integration in
space and time is possible. The basic approach is
simple and relatively easy to apply to nonflowing
systems with relatively high rates of production,
such as ponds. In oceanic environments, similar
methods have been used where spatial heterogeneity on scales of kilometers seems smaller than the
diurnal variability.
For the dissolved gas methods, the in situ concentration of O2 or CO2 is measured over time both
during daylight and at night. The rate of change
during daylight hours is a function of GPP, respiration by all organisms, and exchange of gases with
the atmosphere. The rate of change in oxygen overnight is the net result of respiration by all organisms
and the exchange of gases with the atmosphere.
Generally O 2 decreases at night and increases during the day, while CO 2 increases at night and decreases during the day, but this need not always be
so. For instance, O 2 concentrations can decrease
during daylight hours if rates of respiration are sufficiently greater than GPP or if the supersaturation
of the surface waters by cooling or previous production causes a gas flux to the atmosphere that
exceeds net ecosystem production during the day.
By correcting for the atmospheric exchange of
O 2 (or CO 2 ) at night, the rate of whole-ecosystem
respiration can be estimated by the nighttime
changes in concentration. For instance, if O 2 is undersaturated in the water body, there is a flux of O 2
from atmosphere to the water, and this needs to be
added to the rate of decrease in O 2 during the night
to estimate respiration; if O 2 is supersaturated, there
is a net flux of O 2 from the water to the atmosphere,
and this must be subtracted from the rate of decrease in dissolved O 2 to estimate respiration. Approaches for estimating atmospheric exchanges are
discussed below. If one assumes that the rate of
respiration is the same during the day as at night,
then the daily rate of respiration can be estimated
by mUltiplying the hourly nighttime respiration by
24 hours. This assumption that respiration rates are
constant between night and day is not necessarily
true, as diel temperature changes, photorespiration
by phytoplankton, or bacterial responses to organic
carbon excretion from phytoplankton during active
photosynthesis may all tend to make respiration
rates greater during the daytime.
GPP can be estimated by determining the hourly
rate of O 2 change during daylight hours, adding the
rate of whole-ecosystem respiration determined at
night to this (since the O 2 would have increased
more rapidly during the day if some of it were not
being respired away), and correcting for atmospheric exchange of O 2 , To obtain a daily rate of
GPP, this hourly rate is multiplied by the number
of daylight hours. If changes in CO2 are used rather
than O2, then the respiration rate is subtracted from
the rate of change of CO2 during daylight. Generally, changes in O 2 rather than CO 2 are used because the measurement is analytically easier, often
changes are easier to detect (if concentrations of
dissolved O 2 are less than dissolved CO 2 , as is true
in seawater), and nonbiological processes such as
carbonate precipitation and dissolution need not be
considered. However, a detailed comparison of using both gases to measure production in estuarine
mesocosms yielded similar results (Oviatt et al.
1986).
Some of the first uses of in situ diel methods to
estimate GPP were in small streams and rivers
(Odum 1956; Edwards and Owens 1962). In such
systems, rates of gas exchange with the atmosphere can be estimated from the partial-pressure
gradient between the water and the atmosphere
79
cant changes in the community structure within the
bottle compared with the natural system, particularly in the bacteria. Some of these changes are due
to the growth of organisms that live on surfaces (the
wall of the bottle) and thus not representative of the
purely aqueous system. The large aggregate particles called "marine snow" become destroyed by the
collection process or settle to the bottom of the
sampler and are underrepresented in the sample
(Gardner 1977). Further, the incubation bottles
rarely exceed a liter in volume. Large grazing animals will not be included in the incubation bottles
and the change in grazing pressure may affect rates
of production (Collos et al. 1993).
In Situ Diel Approaches
In situ diel methods use changes in dissolved O2,
CO 2 or particle matter over a 24-hour period to
infer rates of GPP and respiration. Where this approach is feasible, it is often preferable to bottleincubation methods since natural levels of turbulence, grazing, nutrient availability, and light fields
are maintained and since accurate integration in
space and time is possible. The basic approach is
simple and relatively easy to apply to nonflowing
systems with relatively high rates of production,
such as ponds. In oceanic environments, similar
methods have been used where spatial heterogeneity on scales of kilometers seems smaller than the
diurnal variability.
For the dissolved gas methods, the in situ concentration of O2 or CO2 is measured over time both
during daylight and at night. The rate of change
during daylight hours is a function of GPP, respiration by all organisms, and exchange of gases with
the atmosphere. The rate of change in oxygen overnight is the net result of respiration by all organisms
and the exchange of gases with the atmosphere.
Generally O 2 decreases at night and increases during the day, while CO 2 increases at night and decreases during the day, but this need not always be
so. For instance, O 2 concentrations can decrease
during daylight hours if rates of respiration are sufficiently greater than GPP or if the supersaturation
of the surface waters by cooling or previous production causes a gas flux to the atmosphere that
exceeds net ecosystem production during the day.
By correcting for the atmospheric exchange of
O 2 (or CO 2 ) at night, the rate of whole-ecosystem
respiration can be estimated by the nighttime
changes in concentration. For instance, if O 2 is undersaturated in the water body, there is a flux of O 2
from atmosphere to the water, and this needs to be
added to the rate of decrease in O 2 during the night
to estimate respiration; if O 2 is supersaturated, there
is a net flux of O 2 from the water to the atmosphere,
and this must be subtracted from the rate of decrease in dissolved O 2 to estimate respiration. Approaches for estimating atmospheric exchanges are
discussed below. If one assumes that the rate of
respiration is the same during the day as at night,
then the daily rate of respiration can be estimated
by mUltiplying the hourly nighttime respiration by
24 hours. This assumption that respiration rates are
constant between night and day is not necessarily
true, as diel temperature changes, photorespiration
by phytoplankton, or bacterial responses to organic
carbon excretion from phytoplankton during active
photosynthesis may all tend to make respiration
rates greater during the daytime.
GPP can be estimated by determining the hourly
rate of O 2 change during daylight hours, adding the
rate of whole-ecosystem respiration determined at
night to this (since the O 2 would have increased
more rapidly during the day if some of it were not
being respired away), and correcting for atmospheric exchange of O 2 , To obtain a daily rate of
GPP, this hourly rate is multiplied by the number
of daylight hours. If changes in CO2 are used rather
than O2, then the respiration rate is subtracted from
the rate of change of CO2 during daylight. Generally, changes in O 2 rather than CO 2 are used because the measurement is analytically easier, often
changes are easier to detect (if concentrations of
dissolved O 2 are less than dissolved CO 2 , as is true
in seawater), and nonbiological processes such as
carbonate precipitation and dissolution need not be
considered. However, a detailed comparison of using both gases to measure production in estuarine
mesocosms yielded similar results (Oviatt et al.
1986).
Some of the first uses of in situ diel methods to
estimate GPP were in small streams and rivers
(Odum 1956; Edwards and Owens 1962). In such
systems, rates of gas exchange with the atmosphere can be estimated from the partial-pressure
gradient between the water and the atmosphere