Introduction

Terrestrial ecosystems play an important role in the global carbon (C) budget. The C amounts released by anthropogenic activities are higher than the observed increase in atmospheric C02 concentrations measured globally by almost a factor of 2 (IPCC, 1996). After accounting for a large oceanic sink, models predict a large C sink to be located in the Northern Hemisphere, particularly in the terrestrial biosphere (Tans et al, 1990; Ciais et al, 1995; Enting et al, 1995; Francey et al, 1995). However, the annual partitioning among different terrestrial carbon sinks is still under debate (budget for 1980 to 1990; IPCC, 1996; Keeling et al, 1996; Schimel et al, 2000).

Thus, ecosystem physiology, specifically the C02 gas exchange between terrestrial ecosystems and the atmosphere, is of primary interest for global change research (Walker and Steffen, 1996). Atmospheric C02 concentrations and the corresponding carbon global biogeochemical cycles in the climate system

Copyright ' 2001 by Academic' 1'rcss. All rights of reproduction in any form reserved.

isotope ratios of that CO, (<513Ctrop) fluctuate seasonally, mainly due to changes in the terrestrial C fluxes (Conway et al, 1994; Trolier et al, 1996). In addition, measurements starting in the early 1980s showed that atmospheric i5' 3C ratios decreased by — 0.025%o during the 1980s, but that this rate of change approached almost zero between 1990 and 1993 (Trolier et al, 1996). Understanding these changes in <5l3Ctrop requires detailed knowledge about the l 3C signature of different compartments and fluxes in terrestrial ecosystems as well as about the l3C fractionation taking place during the biospheric CO, exchange with the atmosphere.

The 13C02 exchange of the biosphere with the atmosphere is influenced by the interactions of the turbulence regime and ecosystem physiology (Fig. 1). The turbulence regime will influence the mixing of C02 with different isotopic compositions between the biosphere and the atmosphere. Ecosystem physiology will affect the signature of the biospheric flux and the magnitude of this flux. Strong feed back mechanisms exist such as the effect of high turbulence on ecosystem assimilation or of low ecosystem gas-exchange rates on CO, concentrations. The main ecosystem processes that alter the signature of canopy C02 are assimilation, autotrophic respiration, and heterotrophic respiration, each carrying l3C signals integrated over different time spans. The leaf carbon-isotope ratios (Sl3Cieat) reflect current carbon isotope ratios of tropospheric CO, (<513Ctrop) or canopy C02 (Sl3Calnopy) as this C02 is fixed during leaf photosynthesis. During this fixation and subsequent carboxylation, discrimination against the heavier l3CO, takes place (A|eat). In contrast, the 5I3C ratios of litter (<5l3Ciilter) and soil organic carbon (Sl3Csoc) carry isotopic signals from past times due to the long residence times of organic matter in the soil. Thus, both reflect conditions with lower tropospheric C02 concentrations ([C02]) and higher tropospheric SI3C ratios,

JER

troposphere

1

Strop

"canopy

5 leaf

"canopy

5 leaf

RS K f "litter

pedosphere ' C >

5 ^

FIGURE 1 Conceptual model of ecosystem nCO, exchange with the atmosphere. ¿¡„„o,,, = 5"C of canopy air C02, Ae = ecosystem carbon isotope discrimination, SER = 8' 'C of CO, respired by the ecosystem, <\,.lt = SI !C of foliage, A|taf = leaf carbon-isotope discrimination, 8|inel. = 5I3C of litter, <5Rs — <S'3C of soil respired CO,, Ss0(- ~ 8I3C of soil organic carbon, <5, = 5"C of tropospheric CO,.

FIGURE 1 Conceptual model of ecosystem nCO, exchange with the atmosphere. ¿¡„„o,,, = 5"C of canopy air C02, Ae = ecosystem carbon isotope discrimination, SER = 8' 'C of CO, respired by the ecosystem, <\,.lt = SI !C of foliage, A|taf = leaf carbon-isotope discrimination, 8|inel. = 5I3C of litter, <5Rs — <S'3C of soil respired CO,, Ss0(- ~ 8I3C of soil organic carbon, <5, = 5"C of tropospheric CO,.

prior to the combustion of l3C-depleted fossil fuel (isotopic disequilibrium; Enting et al, 1995). Consequently, the <5I3C ratio of soil respiration (<5l3CRs) carries this long-term "ecosystem memory," dependent on the turnover rates of soil organic matter. At the ecosystem level, two parameters integrate these various spatial and temporal scales and describe the 13CO, fluxes: the 513C of ecosystem respiration (<5I3CER; Flanagan and Ehleringer, 1998) and the ecosystem carbon discrimination (Ae; Buchmann et al, 1998).

<5i3Cer ratios and Ae estimates describe the l3C signature of ecosystem C02 fluxes and quantify the biospheric fractionation at the ecosystem level (Flanagan and Ehleringer, 1998; Buchmann et al, 1998). Both parameters integrate the l3C signature of ecosystem CO, exchange with the troposphere, weighted by both the flux rates of above and below-ground processes and the contribution of all species present (see below for necessary measurements and calculations). Furthermore, both <5I3CER and Ae values reflect land-use history, due to mixing of litter and slow turnover rates of soil carbon (Buchmann and Ehleringer, 1998). Thus, both estimates of the l 3C signature of terrestrial C02 exchange can be used to constrain general circulation models' identification and quantification of C sinks or sources. <5l3Ctrop ratios, which have been measured at selected stations within international networks since 1990 (Ciais et al, 1995; Enting et al, 1995; Fung et al, 1997), have found application in general circulation models only very recently. These models rely on scaling modeled <5I3C ratios of soil and plant components to the ecosystem level. To date no comparison between flask-derived and ecosystem-level estimates of ecosystem discrimination has been accomplished.

Bakwin et al (1998) estimated the l3C signatures of biospheric discrimination using flask data from 17 stations of the NOAA/CMDL network (National Oceanic and Atmospheric Administration/Climate Monitoring and Diagnostics Laboratory). Their global estimate of biospheric discrimination averaged 16.8 ± 0.8%o. However, the comparison of this mean with two recent model simulations showed large discrepancies (Lloyd and Far-quhar, 1994; Fung et al, 1997). The model of Fung et al (1997) predicted a much stronger latitudinal gradient, whereas the model of Lloyd and Farquhar (1994) resulted in higher discrimination values. Differences in the distribution of C5 versus Q vegetation and the smoothing effect of atmospheric transport were thought to be responsible for the disagreement of modeled versus measured data (Bakwin et al, 1998).

hi this paper, we compare modeled Ac estimates from the BIOME3.5 model and measured Ae estimates from ecosystem studies. Using flask data from canopy air collections within and above 50 different forest and agricultural sites might allow us to test the modeled Ae estimates more realistically since the smoothing effect of atmospheric transfer observed with the tropospheric air collections will be eliminated. Furthermore, this comparison will be used to identify major gaps in the spatial representation of our study sites and to test the hypothesis that Ae estimates are related to the water use efficiency of terrestrial vegetation as postulated by Buchmann et al (1998).

0 0

Post a comment