238
bance, because soil disturbance influences the biological and physical processes that regulate gas
production, consumption, and transport. Potential
disturbances include inadvertent mixing of surface
soil and establishment of channels for gas flow during insertion of enclosures or enclosure bases; altering soil permeability or inducing pressure gradients by trampling areas adjacent to enclosures;
and disturbing vegetation within or adjacent to enclosures. With care, disturbance and its influence
on gas flux measurements can be minimized. We
recommend collection and analysis of a gas sample
collected at time zero (t = 0) (collected immediately after the chamber is placed) to aid in detection
of serious disturbance effects; time zero concentrations that are inordinately high relative to ambient
may indicate significant disturbance.
Enclosure design and deployment are just the
first steps in flux estimation. Air sample collection,
transport, storage, and analysis also require the development of protocols and safeguards. Contamination of gases by the storage materials, loss or
dilution due to leakage, and sorption onto walls of
the storage container all must be evaluated for each
gas of interest. Different syringe types vary in terms
of leakage rate and must be tested to evaluate potential gas loss over time. Ideally, samples should
be analyzed within several hours after collection.
Glass containers may allow longer storage if logistics demand it, but they also have the potential for
leakage and contamination (Brooks et al. 1993) and
should be tested for specific uses and sample holding times. In the case of both syringes and glass
containers, samples of known concentration should
be handled, stored, and analyzed regularly in the
same manner as samples as part of standard
practice.
The rate of gas exchange per unit surface area
can be calculated from a series of observed chamber air concentrations. Flux calculations require
trace gas concentration observations as a function
of time, enclosure basal area and volume, as well
as temperature and pressure of the enclosed air at
the time of sampling. In non-steady state systems,
both linear and nonlinear models have been used to
describe the relationship between concentration and
time; determination of the appropriate model requires a reasonably large number of gas collections
over time in each enclosure (for N 2 0 and CH 4 , we
recommend at least four collections over 20- to 30Pamela Matson and Allen Goldstein
minute sampling periods; for nitric oxide [NO] and
carbon dioxide [C0 2 ], where in situ analysis
and continuous measurement is straightforward,
much more frequent measurements are taken over
much shorter time periods). Livingston and Hutchinson (1995) present a rigorous set of guidelines for
estimating exchange rates in both steady state and
non-steady state systems, and discuss potential errors associated with estimations.
Trace gas fluxes typically have high spatial and
temporal variability; indeed, this variability is one
of the reasons that micrometeorological techniques,
which integrate over whole ecosystems and allow
for long-term continuous sampling, are the preferred approach in some situations. However, if the
sampling design is carefully developed to match the
characteristics of the site over time, and if appropriate ancillary measurements are carried out, the
variability inherent in gas fluxes can often be used
to increase understanding of the processes controlling fluxes. Indeed, enclosures are often uniquely
useful for process studies because of the small domains over which observations are made. Focusing
of sampling in areas assumed a priori to be "representative" or "typical" can lead to biased extrapolation estimates and erroneous conclusions. Sampling to explicitly include and evaluate on-site
spatial and temporal variability provides an opportunity both to develop site averages that encompass
natural variability as well as to increase understanding of the factors controlling fluxes. Spatial and
geostatistical approaches have great advantages for
extrapolation of fluxes to scales beyond which the
measurements were made (Robertson 1987, Rossi
et al. 1992); however, the intense data requirements
associated with geostatistical approaches suggest
they may be most useful in pilot studies that direct
future sampling designs.
Just as there is enormous spatial and temporal
variability within sites, cross-system variability has
confounded regional and global estimates of trace
gas exchange based on enclosure-based methods.
Ironically, because of this great variability among
ecosystems, enclosure methods are essential for
regional- and global-scale evaluations of trace gas,
as they are inexpensive, mobile, can be deployed
widely in any geographical area, and can be used
to study the factors controlling trace gas fluxes.
Thus, they form the basis for regional and global
budgets and models of N20, CH4, NOx, volatile
bance, because soil disturbance influences the biological and physical processes that regulate gas
production, consumption, and transport. Potential
disturbances include inadvertent mixing of surface
soil and establishment of channels for gas flow during insertion of enclosures or enclosure bases; altering soil permeability or inducing pressure gradients by trampling areas adjacent to enclosures;
and disturbing vegetation within or adjacent to enclosures. With care, disturbance and its influence
on gas flux measurements can be minimized. We
recommend collection and analysis of a gas sample
collected at time zero (t = 0) (collected immediately after the chamber is placed) to aid in detection
of serious disturbance effects; time zero concentrations that are inordinately high relative to ambient
may indicate significant disturbance.
Enclosure design and deployment are just the
first steps in flux estimation. Air sample collection,
transport, storage, and analysis also require the development of protocols and safeguards. Contamination of gases by the storage materials, loss or
dilution due to leakage, and sorption onto walls of
the storage container all must be evaluated for each
gas of interest. Different syringe types vary in terms
of leakage rate and must be tested to evaluate potential gas loss over time. Ideally, samples should
be analyzed within several hours after collection.
Glass containers may allow longer storage if logistics demand it, but they also have the potential for
leakage and contamination (Brooks et al. 1993) and
should be tested for specific uses and sample holding times. In the case of both syringes and glass
containers, samples of known concentration should
be handled, stored, and analyzed regularly in the
same manner as samples as part of standard
practice.
The rate of gas exchange per unit surface area
can be calculated from a series of observed chamber air concentrations. Flux calculations require
trace gas concentration observations as a function
of time, enclosure basal area and volume, as well
as temperature and pressure of the enclosed air at
the time of sampling. In non-steady state systems,
both linear and nonlinear models have been used to
describe the relationship between concentration and
time; determination of the appropriate model requires a reasonably large number of gas collections
over time in each enclosure (for N 2 0 and CH 4 , we
recommend at least four collections over 20- to 30Pamela Matson and Allen Goldstein
minute sampling periods; for nitric oxide [NO] and
carbon dioxide [C0 2 ], where in situ analysis
and continuous measurement is straightforward,
much more frequent measurements are taken over
much shorter time periods). Livingston and Hutchinson (1995) present a rigorous set of guidelines for
estimating exchange rates in both steady state and
non-steady state systems, and discuss potential errors associated with estimations.
Trace gas fluxes typically have high spatial and
temporal variability; indeed, this variability is one
of the reasons that micrometeorological techniques,
which integrate over whole ecosystems and allow
for long-term continuous sampling, are the preferred approach in some situations. However, if the
sampling design is carefully developed to match the
characteristics of the site over time, and if appropriate ancillary measurements are carried out, the
variability inherent in gas fluxes can often be used
to increase understanding of the processes controlling fluxes. Indeed, enclosures are often uniquely
useful for process studies because of the small domains over which observations are made. Focusing
of sampling in areas assumed a priori to be "representative" or "typical" can lead to biased extrapolation estimates and erroneous conclusions. Sampling to explicitly include and evaluate on-site
spatial and temporal variability provides an opportunity both to develop site averages that encompass
natural variability as well as to increase understanding of the factors controlling fluxes. Spatial and
geostatistical approaches have great advantages for
extrapolation of fluxes to scales beyond which the
measurements were made (Robertson 1987, Rossi
et al. 1992); however, the intense data requirements
associated with geostatistical approaches suggest
they may be most useful in pilot studies that direct
future sampling designs.
Just as there is enormous spatial and temporal
variability within sites, cross-system variability has
confounded regional and global estimates of trace
gas exchange based on enclosure-based methods.
Ironically, because of this great variability among
ecosystems, enclosure methods are essential for
regional- and global-scale evaluations of trace gas,
as they are inexpensive, mobile, can be deployed
widely in any geographical area, and can be used
to study the factors controlling trace gas fluxes.
Thus, they form the basis for regional and global
budgets and models of N20, CH4, NOx, volatile