- Top of page
- Materials and Methods
Fossil fuel consumption and the clearing of forests by humans have increased the concentration of CO2 in the atmosphere by 30% above the preindustrial concentration of 280 µmol mol−1. The observed annual increase in atmospheric CO2 concentration is less than the estimated annual anthropogenic emissions due to uptake by the oceans and terrestrial biosphere. The Intergovernmental Panel on Climate Change (IPCC) Third Assessment Report (Houghton et al., 2001) estimated that terrestrial and oceanic uptake were both 1.9 Pg C yr−1. Given that the carbon and nitrogen cycles are tightly coupled, it is evident that continued C sequestration would require additional input of N (Hungate et al., 2003). At the ecosystem level, if one assumes a C : N ratio of 25–50, then the current annual terrestrial C sink requires an additional 38–76 Tg N yr−1. Gifford (1992, 1994) has hypothesized that this can be provided by biological nitrogen fixation (BNF) and atmospheric deposition. In contrast, Hungate et al. (2003) argued that models used by the IPCC to calculate C sequestration during this century failed to deal adequately with potential N limitation, and concluded that increases in BNF would be inadequate to realize the projected C sink.
An increase in the amount of N fixed in response to elevated CO2 (eCO2), either per plant or per unit area of soil surface, has been observed across a range of legume species and legume-containing ecosystems, including Acacia spp. in polycarbonate greenhouses (Schortemeyer et al., 2002); Alnus glutinosa (Vogel et al., 1997); the tropical tree Gliricida sepium (Thomas et al., 1991); Lupinus perennis (Lee et al., 2003); field-grown lucerne (Lüscher et al., 2000); Trifolium repens (Zanetti et al., 1996); Vigna radiata (Srivastava et al., 2002); and Galactia elliottii in a naturally occurring stand of scrub oak (Hungate et al., 1999). Typically, more N is fixed because of an increase in plant size, resulting in a greater nodule mass per plant, rather than an increase in N fixation per unit nodule mass. However, the increase in fixed N is not a universal response, with no effect of eCO2 on the legume N pool reported in alpine (Schäppi & Körner, 1997; Arnone, 1999) and calcareous grasslands (Niklaus et al., 1998).
Legumes as a group are not expected to respond more than nonlegumes to eCO2, in terms of biomass (Poorter & Navas, 2003; Nowak et al., 2004). However, members of the Fabaceae typically have a high N content irrespective of environment (McKey, 1994). This produces litter with a high N content, even under eCO2, and can negate the reduction in litter N concentrations seen in nonlegumes (Hartwig et al., 2000). Potentially this could maintain N availability under eCO2 in grasslands containing legumes, and lead to greater long-term C storage. An analysis by Soussana & Hartwig (1996) concluded that exposing legume-containing grasslands to eCO2 increased not only BNF, but also the contribution of BNF to total grassland N acquisition.
Trifolium spp. are important forage legumes in many pastures, and T. repens in particular is the most important legume in temperate zones, including Europe, North America and New Zealand (Frame & Newbould, 1986). BNF by T. repens has been found to increase under eCO2 in controlled environments (Ryle et al., 1992), open-top chambers (Manderscheid et al., 1997) and in the field (Zanetti et al., 1996). However, these studies were all conducted with a nonlimiting nutrient supply. Stöcklin et al. (1998) have suggested that the lack of a BNF response to eCO2 seen in some natural ecosystems is caused by phosphorus limitation, but experiments where P is added to model or actual native communities that are P-deficient have not reported BNF estimates to date (Stöcklin & Körner, 1999; Grünzweig & Körner, 2003). Niklaus et al. (1998) do not provide BNF data for multispecies microcosms grown with P fertilization and eCO2, but do describe an increase in total legume N. A controlled environment experiment using T. repens plants grown under P deficiency saw no effect of eCO2 on BNF, but also saw no effect of eCO2 when P was supplied in excess (Almeida et al., 2000).
Many Australian soils are lacking in P, and the application of phosphate fertilizers is used to improve biological N fixation (Peoples et al., 1998). Furthermore, BNF (6.5 million t yr−1) accounts for 90% of the N input into Australian soils, with grazed pasture in south-eastern and south-western Australia accounting for approx. 60% of this (Raupach et al., 2001). Consequently the N and P cycles in Australian pasture are linked, and interactions between CO2 and P on BNF may have a substantial role in determining continental C sequestration under a future, CO2-rich atmosphere.
Our study used white clover (T. repens) grown in a 0.2 m2 two-species system with buffalo grass (Stenotaphrum secundatum), a stoloniferous C4 grass native to the east coast of Australia, for a duration of 15 months to examine (i) the response of BNF to eCO2; (ii) whether any such response was dependent on P availability; and (iii) whether there was any effect on the N content of a co-occurring, nonfixing species.