454
J.D. Tenhunen et al.
of stem respiration, the monitoring of gas exchange of whole moss cushions,
observations of whole-canopy and moss-cushion water loss, and further
measurements of community carbon dioxide exchange in the light and dark.
Extremely detailed studies of both lichen community structure, microclimate,
and CO2 gas exchange behavior in four ridgetop heath sites (Hahn 1991)
provide the opportunity to extend the model to such vegetation units of
the watershed. To our knowledge, studies of ecosystem carbon dioxide
exchange at the large plot scale and examinations of temporal changes
simultaneously in aboveground and belowground gas exchange processes
have not been carried out in heath vegetation. These studies are made
difficult by the rapid changes in water content of poikilohydric plants
that occur in the open vegetation and correlated change in physiological
response. Recent improvements in measuring techniques allow us to develop
a realistic picture of the dynamic changes occurring and the resulting data
help establish criteria for improvement of the model.
The driving variables for predicting ecosystem net carbon dioxide exchange, incident light, soil temperature, water table, and aboveground
phenological development, are all variables for which spatial patterns in
the landscape may be defined as a function of time. Radiation input is
predictable from physical models. Our field studies provide reasonable information on soil temperature shifts along topographic and vegetation
gradients within the Imnavait Creek watershed and on the relationship
of soil temperature to water table depth (which is related to watershed
hydration and water movement). Water table depth is a critical predictor of
carbon balance which apparently may be described spatially by further
development of a terrain-based hydrological model (Ostendorf and Reynolds
1993). Patterns in phenological development have been observed but may
also be definable from time-dependent changes in vegetation index sensed
remotely. Thus, it may be possible to link the aggregate version of GASFLUX to routines providing the driving variables and mapping carbon
balance results for the entire watershed in a dynamic fashion on a daily
time step. This would result in an interesting opportunity to view carbon
exchange properties both with fine scale resolution (10 x 10m pixels) and
at the integrated landscape level. The approach to an aggregate model
description described above is a pixel-based approach. Whether such an
approach is adequate to link simple descriptions of carbon uptake to simulations via other models of growth and storage in plant biomass will require
further study.
While there are many pitfalls in both scaling up and in extrapolation
of results, tundra ecosystems, due to their unique structure, offer us the
possibility to verify models at several scales. Tundra is one of the few
natural ecosystems, perhaps the most important one to date, for which
ecosystem response to long-term elevated CO2 exposure has been examined
(Oechel and Billings 1992). The GAS-FLUX model is a simulator based on
mechanistic descriptions, such that essential and basic characteristics of the
response to elevated CO 2 are included. While ecosystem feedback effects in
J.D. Tenhunen et al.
of stem respiration, the monitoring of gas exchange of whole moss cushions,
observations of whole-canopy and moss-cushion water loss, and further
measurements of community carbon dioxide exchange in the light and dark.
Extremely detailed studies of both lichen community structure, microclimate,
and CO2 gas exchange behavior in four ridgetop heath sites (Hahn 1991)
provide the opportunity to extend the model to such vegetation units of
the watershed. To our knowledge, studies of ecosystem carbon dioxide
exchange at the large plot scale and examinations of temporal changes
simultaneously in aboveground and belowground gas exchange processes
have not been carried out in heath vegetation. These studies are made
difficult by the rapid changes in water content of poikilohydric plants
that occur in the open vegetation and correlated change in physiological
response. Recent improvements in measuring techniques allow us to develop
a realistic picture of the dynamic changes occurring and the resulting data
help establish criteria for improvement of the model.
The driving variables for predicting ecosystem net carbon dioxide exchange, incident light, soil temperature, water table, and aboveground
phenological development, are all variables for which spatial patterns in
the landscape may be defined as a function of time. Radiation input is
predictable from physical models. Our field studies provide reasonable information on soil temperature shifts along topographic and vegetation
gradients within the Imnavait Creek watershed and on the relationship
of soil temperature to water table depth (which is related to watershed
hydration and water movement). Water table depth is a critical predictor of
carbon balance which apparently may be described spatially by further
development of a terrain-based hydrological model (Ostendorf and Reynolds
1993). Patterns in phenological development have been observed but may
also be definable from time-dependent changes in vegetation index sensed
remotely. Thus, it may be possible to link the aggregate version of GASFLUX to routines providing the driving variables and mapping carbon
balance results for the entire watershed in a dynamic fashion on a daily
time step. This would result in an interesting opportunity to view carbon
exchange properties both with fine scale resolution (10 x 10m pixels) and
at the integrated landscape level. The approach to an aggregate model
description described above is a pixel-based approach. Whether such an
approach is adequate to link simple descriptions of carbon uptake to simulations via other models of growth and storage in plant biomass will require
further study.
While there are many pitfalls in both scaling up and in extrapolation
of results, tundra ecosystems, due to their unique structure, offer us the
possibility to verify models at several scales. Tundra is one of the few
natural ecosystems, perhaps the most important one to date, for which
ecosystem response to long-term elevated CO2 exposure has been examined
(Oechel and Billings 1992). The GAS-FLUX model is a simulator based on
mechanistic descriptions, such that essential and basic characteristics of the
response to elevated CO 2 are included. While ecosystem feedback effects in