This paper is about the effects of global warming on the decomposition of soil organic matter—often known as humus. Carbon dioxide from the combustion of fossil fuels is one of the key drivers behind global warming. But global warming will also warm terrestrial soils—and there is about twice as much carbon in the top meter of soil than there is in the atmosphere as CO₂. If global warming accelerates the decomposition of this soil carbon, more CO₂ will be released to the atmosphere, further enhancing the warming trend. Positive feedbacks like this are always worrying in a dynamic system, since they destabilise the system still more. Towards the end of the 1980s, several people realised the importance of feedback from soil warming and produced various estimates of its size, some extremely disturbing.

Some years earlier, James Rayner and I had developed a model for the turnover of organic matter in soil that operated over the decades-to-centuries time scale (RothC)^1,2. This model, originally designed for use on agricultural soils, was based on changes in soil carbon in the Rothamsted long-term field experiments, which have been running for more than 150 years. It contained modules that relate decomposition to temperature and soil water content. Alan Wild (then Head of the Department of Soil Science at the University of Reading) and I decided to see if we could use RothC to calculate how much extra carbon would be released from the global soil stock by a specified increase in temperature. We obtained a grant from the Leverhulme Trust to look at feedback from soil warming, appointed Denise Phillips to help with the computing (a much more formidable task then than now) and started work.

We first tested RothC to see how well it worked, using data from experiments on the decay of ¹⁴C-labeled plant material in a range of climates and soils. Next, we split the world's soils into 16 climate zones, based on the aggregation by Post et al.³ of Holdridge's "life zones"⁴. Using the model, we then calculated how much dead plant material would have to be returned to each of the zones each year to maintain the present stock of soil organic carbon in the present climate. As the results looked plausible on the world scale, we went on to calculate how much CO₂ would be released from the world stock of soil organic carbon if world temperatures rose uniformly by 0.3 ^oC per decade. This was the Intergovernmental Panel on Climate Change's "best estimate" for global warming at that time. RothC predicted that by 2050 the additional release of CO₂ from soil organic matter would be a fifth of that released by combustion of fossil fuel—assuming that the rate of use of fossil fuel in 1990 continued unabated. These calculations suggested that increased decomposition of soil organic carbon would indeed make an important contribution to the greenhouse effect, but that the contribution was unlikely to be a runaway effect, as had been feared.

Why we decided to publish in Nature.

Publishing in a soil science journal was an option, but in the end we chose Nature because we wanted to communicate to a wider audience—scientists working on global warming. A few years later, some Rothamsted colleagues and I wrote another paper that might also have been of interest to the global warming community, showing that yields of herbage in a long-term grassland experiment had not changed measurably over the last hundred years. It was published in a good agricultural journal⁵ but sank almost without trace.

The paper has been criticised for two main reasons. First, it only deals with heterotrophic respiration—i.e. the decomposition of soil organic carbon from the moment the leaf falls, exudate leaves the root, or the root dies. It takes no account of autotrophic respiration by live roots, which can make up 10-90% of total soil respiration, depending on the time of year. This criticism is of course correct—but it misses the point of our paper: we wanted to know how global warming could influence the decomposition of soil organic matter, not how it affected root respiration, which marches to the beat of a different drum.

A more serious criticism is that by Giardina and Ryan, who argue that the decomposition of soil organic matter is not temperature sensitive⁶. We strongly disagree with this view—indeed our paper contains good evidence to the contrary. This is an important issue—there will be no feedback to the atmosphere from soil warming if Giardina and Ryan are correct. Time will show who is right.

The global stock of soil carbon (about 1,500 billion tonnes) is calculated for a soil depth of one meter. Very roughly half this carbon is held in the topsoil—say 0-25cm, and the rest in the 25-100 cm layer. There are now several good models for the turnover of organic matter in topsoils (for inter-model comparisons see Smith et al.⁷⁾ but none for the deeper layers. In our Nature paper we assumed that organic matter in topsoils and subsoils behaved similarly. We are now testing this assumption, using the pulse of radiocarbon from the thermonuclear bomb tests of the early 1960s as a tracer for the turnover of organic carbon in subsoils. Good models for topsoils and subsoils are needed if we are to understand the role of soil organic carbon in the global carbon cycle

Both Alan Wild and I began as chemists, but we soon went to ground and have stayed in soil science ever since. The work in this paper was started shortly after I retired in February 1988. Retirement at Rothamsted is compulsory at 60 and I was keen to disprove Thomas Huxley's contention that "scientists over 60 do more harm than good." I was also keen to supplement my pension. Alan also retired in 1982: since then he has written two books. Denise Adams (nee Phillips) had twin boys and, sadly, left science.

David Jenkinson
IACR-Rothamsted
Hertfordshire, United Kingdom