There is not yet consensus on whether nutrient deficiency affects the response of plants and ecosystems to elevated CO2, but legumes are of particular interest due to their high tissue nitrogen content and ability to fix atmospheric N2. Grazed pasture in southern Australia covers over 60 million ha and accounts for approx. 60% of total Australian biological nitrogen fixation (BNF). Furthermore, BNF provides the largest flux of N into Australian soils.
For my initial three-year contract I addressed the question of whether 'BNF and pasture legume growth responses to elevated CO2 are affected by P limitation' using a simple two-species community of Trifolium repens and the C4 grass Stenotaphrum secundatum, grown in sand filled 0.2 m2 mesocosms for 15 months under ambient (aCO2) or elevated CO2 (eCO2), with either low P / zero N, high P / zero N or high P / plus N nutrient treatments.
Production was measured by clipping to 5 cm on ten occasions. Surprisingly, aboveground standing crop at the end of the experiment was not affected by nutrient treatment and was 17-40% higher under eCO2. However, when the clipped material was included, annual aboveground production was halved by low P availability: approx. 12.5 or 25 t ha-1 for low or high P, respectively. At high P, eCO2 stimulated aboveground production by 12-24%, but at low P there was no effect of CO2. This apparent difference in growth response was due to changes in plant stature, so clipping had removed a smaller proportion of the biomass in the low P mesocosm under eCO2. T. repens became the dominant species in all the mesocosms, with S. secundatum making up 2-13% of aboveground biomass.
Table 1: Regressions of aboveground production against cumulative radiation. Slopes are g dry mass MJ-1.
Once the mesocosms were established, aboveground production was strongly seasonal in all treatments (Table 1), with the highest growth rates occurring during the highest incident radiation in summer. Effects of nutrient and CO2 treatment were not seasonal nor dependent on rate of production.
The response of BNF to the treatments was similar to that of production. In the low P treatments BNF was reduced by more than 50% and was unaffected by eCO2, whereas with adequate P there was an increase in BNF of 30% (Fig. 1). Evidently growth and BNF of pasture legumes such as Trifolium spp. are unlikely to respond to eCO2 where soil P is limiting. Further, the tissue nutrient data indicated that growth was P rather than N limited. Therefore, C flux models must include P availability and its effects as well as the N cycle, particularly in areas such as Australia where P is commonly limiting.
Fig. 1: Biological nitrogen fixation in clover dominated mesocosms grown under either ambient (white bars) or elevated CO2 (grey bars) with no N and low (-P) or high P (+P) nutrient treatment.
Effects on photosynthesis
A trial experiment using individual T. repens plants, grown in pots with high or low P & N availability, demonstrated that, in the short-term at least, nutrient stress rapidly removed stimulation of photosynthesis by eCO2. A similar response was also initially seen in the mesocosm grown T. repens, with no eCO2 enhancement of Asat observed at low P availability. However, in the second half of the experiment, despite Asat being reduced by low P, rates were higher under eCO2 in all treatments. The explanation was that in both experiments Asat of T. repens was strongly dependent on leaf N per unit area, which was only affected by eCO2 in the first half of the mesocosm experiment.
Examining the belowground responses of the mesocosms provided a more complex picture than looking at the aboveground data alone. Allocation of C, allocation of N and new soil C were all affected by interactions between eCO2 and P limitation. Root mass fraction (RMF) at the final harvest was doubled by P limitation and eCO2 increased RMF in all nutrient treatments. I did not measure root turnover directly, but the low C status of the sand used in the mesocosms allowed the detection of soil C that had accumulated in the system. Total belowground C increased with nutrient limitation under aCO2 (Fig. 2). However, under eCO2 there was no effect of soil nutrient status on C content, indicating that increases seen in soil C sequestration with eCO2 may not occur in nutrient limited soils. The fraction of N allocated belowground also increased under eCO2. Even in the P limited mesocosms, where there was no effect of CO2 on BNF, there was a 30% increase in both soil and root N.
Fig. 2: Root C (hatched bars) and total belowground C (plain bars) at the final harvest under ambient (white bars) or elevated CO2 (grey bars) with low P/zero N, high P/zero N or high P/plus N nutrient treatment.
As well as providing evidence that BNF and legume growth does not respond to eCO2 when P limited, the results of this major experiment suggest that effects of a CO2 enriched atmosphere on nutrient limited pasture will be largely belowground. Of particular interest is the possibility of an interaction between CO2 and nutrient availability on root turnover with consequent effects on soils.