Abstract

Global data sets of the stable carbon isotope composition of plant leaves, of CO2 in canopy air, and of CO2 in the background atmosphere were compiled and compared to results of a global vegetation model (BIOME4) that simulated, at these three scales, the magnitude, direction, and timing of fluxes of CO2 and C-13 between the biosphere and the atmosphere. Carbon isotope data on leaves were classified into 12 Plant Functional Types (PFTs), and measurements from canopy flasks were assigned to 16 biomes, for direct comparison to model results. BIOME4 simulated the observed leaf delta(13)C values to within 1 standard deviation of the measured mean for most PFTs. Modeled delta(13)C for C-3 grasses, tundra shrubs, and herbaceous plants of cold climates deviated only slightly more from measurements, perhaps as a result of the wide geographic range and a limited set of measurements of these PFTs. Modeled ecosystem isotopic discrimination against C-13 (Delta(e)) averaged 18.6 globally when simulating potential natural vegetation and 18.1 when an agricultural crop mask was superimposed. The difference was mainly due to the influence of C-4 agriculture in areas that are naturally dominated by C-3 vegetation. Model results show a gradient in Delta(e) among C-3-dominated biomes as a result of stomatal responses to aridity; this model result is supported by canopy air measurements. At the troposphere scale, BIOME4 was coupled to a matrix representation of an atmospheric tracer transport model to simulate seasonally varying concentrations of CO2 and C-13 at remote Northern Hemisphere measuring stations. Ocean CO2 and C-13 flux fields were included, using the HAMOCC3 ocean biogeochemistry model [Six and Maier-Reimer, 1996]. Model results and observations show similar seasonal cycles, and the model reproduces the inferred latitudinal trend toward smaller isotopic discrimination by the biosphere at lower latitudes. These results indicate that biologically mediated variations in C-13 discrimination by terrestrial ecosystems may be significant for atmospheric inverse modeling of carbon sources and sinks, and that such variations can be simulated using a process-based model.

Details

Actions