PROJECT GOALS
Our study is a concerted modeling and measurement program focused on carbon and oxygen isotope exchange by terrestrial ecosystems.  The experiments are being conducted primarily at the WLEF tall tower in Park Falls, Wisconsin, and will leverage off of the comprehensive suite of meteorological, ecophysiological, ecological and remote sensing studies also being conducted at this site.  The project goals are to:

¥     Provide a model that can be used to define basis functions for the global inversion of atmospheric measurements of CO₂, d¹³C of CO₂ , and d¹⁸O of CO₂.

¥     Test the model against observations at WLEF and other sites.

¥     Develop a robust experimental strategy to expand the scope of these measurements to other continental sites around the world.

¥     Develop a strategy to estimate tropospheric values of CO₂, and d¹³C and d¹⁸O of CO₂ from near surface measurements in continental areas (unlike marine sites, continental sites are strongly influenced by local fluxes). 

The modeling work is being conducted by Scott Denning of Colorado State University.  Joe Berry has set up a system to obtain and analyze the concentration and isotopic composition of CO₂ in air from at least two levels on the WLEF tower and in the canopy of adjacent forests. Jim Ehleringer of the University of Utah is conducting ecosystem measurements to sample the isotopic composition of key vegetation components of the ecosystems surrounding the tower to provide the data needed to interpret isotopic composition of atmospheric CO₂.  These measurements will be analyzed and integrated with modeling studies at a hierarchy of scales from individual elements (the canopy and soil), complete ecosystems, to regional and continental simulations.

The detailed measurements and modeling at the WLEF site will provide a good basis for designing simpler  yet powerful approaches that could be applied in other ecosystems around the world to calibrate and test the ecosystem isotope fractionation model.  Such studies will increase confidence in the basis functions produced by the model and provide important new insights into physiological and hydrological processes.

RESULTS AND ACCOMPLISHMENTS
We conducted an intensive sampling program at the WLEF tower and in surrounding ecosystems during the time of the COBRA flights at WLEF by Steve WofsyÕs group (late summer 2000).   In this campaign, we collected 40 flask samples at regular intervals over four days from  the 11 and 244 m levels of the WLEF tower.  These samples (Figure 1 & Figure 2) characterize the co-variation of CO₂ concentration and the d¹³C and d¹⁸O of CO₂ over the surrounding region.  These values agree well with 10 samples taken in 2 liter flasks and analyzed by CMDL.  During the day we observe a small but consistent gradient in CO₂ concentration and isotopic composition between the top and bottom levels.   Preliminary calculations indicate that the apparent discrimination against ¹³C, d¹³C, was 17 ä ø a plausible  result.  However many more samples would be needed to obtain an accurate determination of the photosynthetic discrimination.

Two sampling sorties with light aircraft were conducted in collaboration with John Birks and Mike Jensen of NCAR .  Consistent with the COBRA flights, the samples collected from above (7000Õ) and below (1500Õ) the top of the pbl showed a substantial midday gradients with the CO2 concentration in the pbl 4-8 ppm lower than that of the troposphere.   These measurements demonstrate the power of  aircraft sampling to enhance the value of measurements taken from the tower.  We plan to use these measurements to construct pbl budget calculations of the surface fluxes of heat and CO₂.

Keeling plots analysis of the samples collected  from levels 11, 30, 76,  244 m.  Figure 3 shows that a very consistent relationship was obtained for co-variation of d¹³C with CO₂ concentration indicating that ecosystem respiration was producing CO₂ of d¹³C ø26.3 ä.  This did not appear to vary with time of day or over the sampling period, hence all samples were pooled in the same plot.  In contrast, Figure 4, the d¹⁸O of respired CO₂ was quite variable from time to time.  Each line on the plot is a separate regression analysis on a set of samples from the 4 levels taken at the indicated times.  There is a clear trend for the d¹⁸O values to become more negative as the night progresses.  We are not sure what is causing this large variation and it will be an interesting challenge to model this result. 

Samples for Keeling plot analysis were also collected from canopy access scaffolding located in representative ecosystems in an area just north of the WLEF tower.  As at the main tower, samples were taken simultaneously from four levels on several occasions during nights.  The data (Figures 5, 6, 7, 8, 9, & 10) show the d¹³C of respired CO₂ was quite consistent from measurement to measurement and these data have been pooled for analysis.  The bog and aspen ecosystems had d¹³C values heavier than the WLEF tower.  The maple/basswood site was slightly more negative. However, at this time there was little variation in carbon isotope discrimination between sites, and these data would not offer strong constraint on estimates of the ecosystem contribution to the regional exchange monitored at the WLEF tower.   The d¹⁸O analysis shows progressively lower values from each set of samples taken over the course of the night. The d¹⁸O of respired CO₂ from these ecosystems is considerably more negative (-10 to 14 ä) than that observed at the WLEF tower.  These ecosystems are major components of the land cover surrounding and the result from the tower seems to indicate a substantial influence from the clearing that surrounds the tower.  More work is required to establish this.

Copies of data from this project are available in Excel spread sheet format at ftp://ecophys.biology.utah.edu/WLEF.

FUTURE WORK
We have constructed an automatic sampling system that will be installed in the instrument trailer at the WLEF tower in early May 2001.  At that time we will begin weekly sampling of the diurnal variation of the isotopic composition of CO₂ in air in the pbl.  Another intensive sampling campaign is planned for June 2001.   A meeting of the project scientists will be held on May 8 2001 to coordinate activities. We will review our progress and adjust our plans for future sampling campaigns.

Modeling
These data are particularly valuable because the existing algorithms are based on theoretical considerations and are largely unconstrained by macroscopic data. The isotopic model is already coupled to both CSU RAMS and the CSU GCM, and allows us to predict stable isotopic exchange and resulting d¹³C and d¹⁸O ratios in atmospheric CO₂  at multiple spatial scales (Figure 11).  Simulations can be performed as "case studies" by prescribing observed weather and wind fields at a station or across a region; or as a fully coupled "forecast" in which the biophysics, isotope geochemistry, weather, and atmospheric transport interact and influence one another.

The improved model will be tested against atmospheric data at each field site by comparing simulated and observed time series of d¹³C and d¹⁸O, forcing the model with observed weather. At the WLEF tower site, a richer suite of observations is available (multiple observing heights up to the mid-CBL; component fluxes in upland hardwood and lowland spruce forest, soil and snow physical data, etc). Here we will simulate the spatial and temporal variations of the isotopic tracers for intensive observing periods using the coupled SiB2-RAMS system, with a set of nested grids. The outer grid will be about 1000 km across, with a horizontal grid spacing of 64 km, with weather at lateral boundaries forced from operational products and tracer boundary conditions forced from the output of a global simulation with the CSU GCM. Two-way interactive nested grids will be run with spacing of 16, 4, 1, and 0.25 km within this outer grid. The innermost grid will resolve vegetation heterogeneity in the tall tower footprint and the vertical variations measured on the tower.  Biophysical properties of the vegetation will be specified from remotely sensed data products (NDVI and  subpixel mixtures of woody and herbaceous vegetation derived from 1 km AVHRR by Defries et al. (1999) for the outer grids, and 30 m LandSat Thematic Mapper imagery for the innermost grid).  These simulations will be compared against tower data and also against flasks collected further aloft during the August 2000 intensive observing campaign by John BirksÕ team using balloon and kite soundings, Ron DobosyÕs team using the LongEZ aircraft, and Steve WofsyÕs COBRA team up to 10 km. These campaigns will allow us to evaluate the fully coupled modeling system at multiple spatial scales.

The results of the high-resolution coupled RAMS-SiB2 simulations will include realistic fully populated grids of CO₂, d¹³C, and d¹⁸O over the upper Midwest, which we will use to test methods for regional extrapolation from surface measurements over the land surface. These tests will include an evaluation of the feasibility of estimating mid-CBL isotopic ratios using the "virtual tall tower" concept pioneered by Ken Davis, in which AmeriFlux data may be used to correct for local offsets to concentrations due to strong surface fluxes and turbulence to obtain regionally-representative "background" values appropriate for use in inversions. In addition, we will test the feasibility of routine aircraft soundings for stable isotopic composition of CO₂ by subsampling our simulated atmosphere with "virtual aircraft." This will include a careful assessment of spatial and temporal variability of the isotopic tracers in the model and the representativeness of these ratios as measured by the various sampling systems.

After conducting the intensive regional simulations and comparing to the data collected in the intensive observing period, we will extend the simulations to the scale of the entire continental USA using the mesoscale model in "climate" mode (ClimRAMS) on a 50 km grid. Weather on the lateral boundaries will be specified from NCEP reanalysis products. We will prescribe trace gas boundary conditions from a global simulation with the CSU GCM (Denning et al., 1996, 1999). The global fluxes that control these trace gas boundary conditions will be optimized through synthesis inversion to ensure consistency with the NOAA flask network. The ClimRAMS simulation will then produce fully populated 3-dimensional fields of the trace gases on the 50 km grid which we will archive hourly for one year.  These "pseudo-data" will be consistent with the field measurements of isotopic exchange processes, with the actual weather, with the state of the vegetation as measured from space, and with the NOAA flask measurements at the global scale.

We will investigate the feasibility of performing mesoscale inversions on these pseudo-data, by assuming various configurations of atmospheric sampling networks that might be deployed in the future using aircraft, surface measurements, tall towers, or micrometeorological extrapolation ("virtual tall towers"). We will subsample the large pseudo-data volume at this hypothetical network and try to recover the regional fluxes that produced it in the model. Unlike inversions of real data, we will know the flux distribution exactly, so we will be able to rigorously evaluate these inversions and quantify the error in the results depending on the configuration of the hypothetical observing network. These studies will be essential for the design of continental observing systems in the future, and we will explore the ways in which stable isotopic tracers might add value to such a network.

Finally, we will use the results of the local, regional, and continental experiments described above to develop isotopic basis functions for global inverse models which include realistic variations of coupled CO₂, d¹³C, and d¹⁸O exchanges in both space and time. These basis functions should allow inverse calculations to distinguish much more reliably between ocean and terrestrial uptake at similar latitudes (using d¹³C). In addition, the use of spatially and temporally varying distributions of d¹⁸O exchanges will allow global inversions that recover distributions of gross photosynthesis and respiration, rather than simply the net flux (Peylin, 1999). This will allow inverse modeling to move beyond simple diagnosis of flux maps toward a better understanding of the processes responsible for the sources and sinks, which is necessary if we are to predict changes in the carbon cycle in the future.

Literature Cited

Craig, H.  1953. The geochemistry of stable carbon isotopes.  Geochimica Cosmochimica Acta, 3, 53-92.

DeFries, R., J.R.G. Townshend, and M. Hansen.  1999.  Continuous fields of vegetation characteristics at the global scale at 1 km resolution.  Journal of Geophysical Research 104:16911-16925.

Denning, A.S., D.A. Randall, G.J. Collatz, and P.J. Sellers.  1996.  Simulations of terrestrial carbon metabolism and atmospheric CO₂ in a general circulation model. Part 2: Spatial and temporal variations of atmospheric CO₂. Tellus 48B:543-567.

Denning, A.S., M. Holzer, K.R. Gurney, M. Heimann, R.M. Law, P.J. Rayner, I.Y. Fung, S.M. Fan, S. Taguchi, P. Friedlingstein, Y. Balkanski, J. Taylor, M. Maiss, and I. Levin.  1999.  Three-dimensional transport and concentration of SF6: A model intercomparison study (TransCom 2). Tellus 51B:266-297.

Farquhar, G.D. 19983.  On the nature of carbon isotope discrimination in C₄ species. Australian Journal of Plant Physiology, 10, 205-226

Farquhar, G.D., J. Lloyd, J.A. Taylor, L.B. Flanagan, J.P. Syversten, K.T. Hubick, S.C. Wong, and J.R. Ehleringer.  1993.  Vegetation effects on the isotopic composition of oxygen in atmospheric CO₂.  Nature 363:439-443.

Flanagan, L.B., J.P. Comstock, and J.R. Ehleringer.  1991.  Comparison of modelled and observed environmental influences on the stable oxygen and hydrogen isotope composition of leaf water in Phaseolus vulgaris L.  Plant Physiology 96:588-596.

Flanagan, L.B., J.R. Brooks, G.T. Varney, S.C. Berry, and J.R. Ehleringer.  1996.  Carbon isotope discrimination during photosynthesis and the isotope ratio of respired CO₂ in boreal ecosystems.  Global Biogeochemical Cycles 10:629-640.

Mook, W.G., Bommerson, J.G., and W.H. Staverman.  1974.   Carbon isotope fractionation between dissolved bicarbonate and gaseous carbon dioxide.  Earth and Planetary Science Letters, 22, 169-176.

OÕLeary, M.H.  1984.  Measurement of the isotopic fractionation associated with diffusion of carbon dioxide in aqueous solution.  Journal of Physical Chemistry, 88, 823-825.

Peylin, P.  1999.  ¹⁸O isotopic composition in atmospheric CO₂: A new tracer to estimate photosynthesis at global scale. PhD Dissertation, Paris IV University, 221 pp.

CONTACTS

Joseph Berry
Department of Plant Biology
Carnegie Institution of Washington 
290 Panama Street
Stanford, California 94301
Tel.  650-325-1521
FAX 650-325-6857 
e-mail: joeberry@biosphere.stanford.edu

A. Scott Denning 
Department of Atmospheric Science
Colorado State University
Fort Collins, Colorado 80523
Tel. 970-491-6936
FAX  970-491-8449
e-mail:  denning@atmos.colostate.edu

James Ehleringer
Department of Biology
University of Utah
257 South 1400 East
Salt Lake City, Utah 84112
Tel. 801-581-7623
FAX 801-581-4665
e-mail: ehleringer@biology.utah.edu

Return to GCC Abstracts