Abstract

Models of how forests will respond to climate change usually assume that carbon dioxide output from decomposing organic matter will increase with global warming. That assumption may be wrong.

Each year, the world's vegetation absorbs about 60 billion tonnes of carbon by photosynthesis and releases a similar — but not exactly equivalent — amount by respiration¹. These fluxes of carbon are large, dwarfing the 6.5 billion tonnes emitted by the burning of fossil fuel. Most of the terrestrial photosynthesis and respiration is carried out by ecosystems that produce woody material — forests and savannahs ( Fig. 1). The respiratory flux is partly from the plants themselves, but about 50% is from microbial decomposition of the organic matter that the plants have produced. There is a large stock of this organic matter in the soil, much of it in the form of very old and only slowly degradable residues of lignin (the main constituent of the cell walls of woody plants).

Figure 1: Carbon production by terrestrial ecosystems.

(Adapted from ref. 14.)

High resolution image and legend (8K)

From the global patterns of CO₂ concentration, it seems that terrestrial photosynthesis and respiration are not in balance^{2, 3} — photosynthesis appears to exceed respiration by 2 billion tonnes of carbon per year³. It is now possible to measure these fluxes at specific forest sites⁴. The measurements show quite clearly that old and undisturbed forests, as well as middle-aged forests, are net absorbers of CO₂ (refs 5–7). The reason may be increased CO₂ fertilization (CO₂ stimulates photosynthesis) together with increased deposition of anthropogenic nitrogen (which also acts as a fertilizer)⁸. This is good news: it means that forests are serving as a carbon sink, providing a global environmental service by removing CO₂ from the atmosphere and thus reducing the rate of CO₂-induced warming.

However, the forest sink may not persist. In all forests studied so far, the net gain of carbon is the small difference between the two huge numbers representing photosynthetic gains and respiratory losses. In the future 'greenhouse' world, photosynthesis is likely to increase with rising CO₂ levels and nitrogen deposition, enhancing the sink strength. However, as every physiologist knows, respiration rates increase sharply with temperature⁹. So it is generally believed that respiration — both of the vegetation itself and of microbial decomposition of organic matter — will increase with global warming. The commonly held view, then, now enshrined in models of global change, is that the carbon sink provided by forests will weaken, and that in the long term the world's forests may eventually become a source of carbon to the atmosphere.

Two papers on pages 858and 861 of this issue^{10, 11} make us think otherwise. Giardina and Ryan's results¹⁰ corroborate and extend findings from Finnish soils¹² suggesting that, over long periods of time (decades), decomposition of organic matter is not very sensitive to temperature. It seems that short-term warming experiments showing a rise in respiration do not capture the long-term characteristics of its response to rising temperatures.

But the results reported in the second paper, which come from a network of CO₂ flux-measurement stations set across Europe's forests, are even more surprising. Valentini and colleagues¹¹ show that respiration is a more important component of the carbon balance in northerly latitudes despite the low temperatures there, and that it is really respiration, not photosynthesis, that varies over the latitudinal band from Iceland to Italy.

Are the results of Valentini et al. likely to be general? Despite the great difficulty of sampling forests that are truly representative over the whole of Europe, we suspect that they do indeed reveal a real trend with latitude. We await results from a similar network of stations in the United States, which may confirm this trend. Meanwhile, we know that carbon fluxes in the tropics are larger than those in temperate and northern forests^{5, 6}. But there is not enough information yet to comment on the long-term effects of temperature on the carbon balance, and further data are needed from rainforests and savannahs to complete the global picture.

Why should soil respiration be higher in a colder climate? Perhaps it is because the soil in the north is wetter for longer, and so microbes that are adapted to work at low temperature are active for most of the year. In more southerly latitudes, by contrast, the microbes become inactive for much of the year when the soil is dry. Or perhaps the more northerly ecosystems contain much carbon as organic matter which has accumulated in the soil over previous, colder periods, and is only now decomposing as the soil warms in response to climate change.

To illustrate the importance of these papers to our understanding of the future of the terrestrial carbon sink, we set parameters for a simple ecosystem model of conifer forest using routines and procedures described elsewhere¹³. For case 1 we adopted the commonly held view and assumed that the ecosystem respiration will rise in line with long-term increases in temperature. For case 2 we modified the model and assumed that ecosystem respiration would continue to respond to daily and seasonal variations in temperature, but would be insensitive to longer-term temperature changes. In both cases we ignored any age-related changes in the forests, as this would complicate the picture needlessly.

The result (Fig. 2) clearly shows that the assumption made about the temperature sensitivity of ecosystem respiration has a profound effect on the long-term future of the forest carbon sink in coniferous forest. If case 1 is correct, the sink diminishes, and the forest becomes less effective at removing CO₂ from the atmosphere; if case 2 is correct, the effect of an increasing rate of photosynthesis is not masked by an increasing respiration rate, and the forest becomes increasingly more effective as a sink for atmospheric CO₂.

Figure 2: A long-term prediction of the net ecosystem exchange of carbon in a generic northern European forest.

Data plotted are ten-year means of model output, where respiration is (case 1) or is not (case 2) sensitive to long-term changes in temperature. The possibility of case 2 applying stems from research reported in refs 10–12. In case 1 the sink diminishes; in case 2 the sink strengthens to over 10 tonnes of carbon per hectare per year. A more negative number means an increasing uptake of carbon by the forest. (Model run with climate data output from a general circulation model, using scenario IS92a of the International Panel on Climate Change¹⁵.)

High resolution image and legend (7K)

In any case, the results from these two papers^{10, 11} should send a powerful message to those working with models of global vegetation change. Setting the parameters for soil respiration models using only the results of short-term experiments may be quite misleading. When respiration models are eventually fully coupled to models of climate change, the resulting positive feedback between respiration and temperature that magnifies global warming may proceed only for a limited time — until the easily decomposed soil organic matter is depleted. Does this mean that the doomsday view of runaway global warming now seems unlikely? We hope so.

References

1. Schimel, D. S. Glob. Change Biol. 1, 77–91 (1995). | ISI |
2. Francey, R. J. et al. Nature 373, 326–330 (1995). | Article | ISI | ChemPort |
3. Rayner, P. J. et al. Tellus B 51, 213–232 (1999). | Article | ISI |
4. Moncrieff, J. B. et al. J. Hydrol. 189, 589– 611 (1997). | Article | ISI |
5. Grace, J. et al. Science 270, 778–780 (1995). | ISI | ChemPort |
6. Malhi, Y. et al. J. Geophys. Res. 103, 31593– 31631 (1998). | Article | ISI | ChemPort |
7. Lindroth, A. et al. Glob. Change Biol. 4, 443– 450 (1998). | Article | ISI |
8. Lloyd, J. & Farquhar, G. D. Funct. Ecol. 10, 4–32 (1996). | ISI |
9. Lloyd, J. & Taylor, J. A. Funct. Ecol. 8, 315–323 (1994). | ISI |
10. Giardina, C. P. & Ryan, M. G. Nature 404, 858–861 (2000).  | Article | PubMed | ISI | ChemPort |
11. Valentini, R. et al. Nature 404, 861–865 (2000). | Article | PubMed | ISI | ChemPort |
12. Liski, J. et al. Ambio 28, 171–174 (1999). | ISI |
13. Lloyd, J. et al. Plant Cell Environ. 18, 1129– 1145 (1995). | ISI |
14. Ajtay, G. et al. in The Global Carbon Cycle (eds Bolin, B. et al.) 144 (Wiley, Chichester, 1979).
15. Houghton, J. T. et al. (eds) Climate Change 1994 (Cambridge Univ. Press, 1995).

1. John Grace and Mark Rayment are at the Institute of Ecology and Resource Management, University of Edinburgh, Edinburgh EH9 3JU, UK.

Correspondence to: John Grace¹ e-mail: Email: e-mail: jgrace@ed.ac.uk