E. Lioubimtseva & J.M. Adams

Conference keynote address. rough version! Still being written.


Global biogeochemical modelling of soil and peat carbon reservoirs is essential for forecasting future CO2 levels. However, at present various key uncertainties and also basic problems with published data sources are not properly acknowledged in broad-scale models of the carbon cycle. Here we examine some of these uncertainties and suggest how they may affect our prospects of realistically modelling the global carbon cycle.

As an example of the difficulties which can come about from failing to consider 'details', standard soil carbon databases tend to concentrate on sites of agricultural potential, biasing the soil carbon content in arid regions in favour of sites with deep and well-watered soils. Global peat carbon estimates very greatly between authors, and it is important to use the most rigorous and recent studies on northern peatlands that take into account spatial heterogeneity in these bogs. Lack of attention to these difficulties has sometimes led to unrealistic figures being used in certain global reservoir estimates and carbon cycle models.

Further serious scaling difficulties remain to plague the field of carbon cycle modelling. The direct-CO2 fertilization effect on soil carbon remains an almost intractable problem, possibly varying greatly on a very small spatial scales, with little prospect of simple relationships that can be modelled on a global scale. The broad time dimension is an additional source of data which is both poorly understood and not widely acknowledged; it may well be that unknown 'lag' factors are involved in broad scale ecosystem and soil development and that these could operate during future global climate change.

If biogeochemical modelling is to make the greatest progress, it will be necessary for the community to critically examine all numerical data sources with close reference to the detailed knowledge and experience of field ecologists and pedologists who understand the spatial 'graininess' of the data. Modellers must also be prepared to accept a certain element of 'anecdotal' expertise (i.e. unquantified and not systematically gathered information) as necessary to understanding the complexity of ecosystems, and to accept that certain aspects of the carbon cycle (e.g. direct CO2 effects) simply cannot realistically be modelled at present due to the sheer spatial and temporal complexity of what is being represented.


Understanding of broad scale patterns in soil carbon storage is very important in forecasting CO2 fluxes to and from the atmosphere. A great deal of effort currently focuses on modelling changes in terrestrial carbon reservoirs in order to forecast future CO2 levels in response to land use and climate change. Here we examine some of the general problems which much be bourne in mind when considering the spatial distribution of soil carbon, and its potential for change over time.

These include; the existing distribution of soil carbon between bioclimatic regions, the amount of carbon in peats, and the nature of direct-CO2 effects on soil carbon storage. Here we briefly consider some of problems which confront understanding of the present and future soil/peat carbon reservoir.

1) The distribution of carbon between bioclimatic regions; large volumes of data do not necessarily make for reliability!

Published databases of the spatial distribution of soil organic carbon are widely used to parameterize global carbon cycle models. To ensure that the assesments of reservoirs and the models are well founded, it is necessary to point out some of the problems which may arise from the use of existing databases. The major source of data for carbon cycle modelling is still the global database compiled by Zinke et al. (1985). To make further progress, it is time to start reassessing this database in terms of detailed study of the data sources it contains using knowledge from the broad soil science and ecological community, and to bring in more recent additional sources of data, comparison and study of types of data sources. An updated source is the STATSGO database, summarized in this map of US carbon storage put together by Bill Hargrove (Fig. 1, showing map of USA soil carbon).


Some paradoxical features emerge if one examines the Zinke et al. and STATSGO databases in more detail. For example, their results suggest that there is as much carbon in temperate deserts as in temperate forests and woodlands! This is in the Zinke et al database and so it is trusted. Yet we must continually ask ourselves whether any given data source is misleading.


9.7 tC/ha in cool desert vs 10.0 tC/ha in temperate forest.

Most of the data comes from desert soils from the SW USA. From the database, Zinke et al. suggest global values for cool desert soil carbon of around 10 kg/m2, and these figures have been used in most of the long-term global carbon cycle estimates cited by Peng et al. (in press), such as (LIST THE REFERENCES). The present examination of the more detailed STATSGO database also suggests relatively high soil carbon values (around 4.8 tC/ha) for the warm western deserts of North America. Such large amounts of carbon in desert or semi-desert soils would certainly seem paradoxical, considering the low rates of input of organic matter.

Detailed assessment of the databases suggests that in fact the soil carbon values for SW USA and central Asian desert regions and are inflated due to a number of factors.

Firstly, in the face of many thousands of soil data points from around the world, some data points were accidentally mis-assigned. This has particularly been true of cool deserts, but it might well be true for some other biomes too. In the central Asian dataset (co-ordinates given below in Table 1.), all 30-odd points fall within the dry steppe and steppe-marginal zones!

Table 1. Datapoints for 'cool desert scrub' soil organic carbon in Central Asia from the database of Zinke et al. (1985). In fact, all of these fall outside the 'true' desert or semi-desert regions.

Database no. ....Biome..Organic C........Grid co-ordinates....Life zone type

2050001 245.01 SB 15.3 1001. 53.0N 78.6E 200. 13 B 40

2050002 245.03 SB 6.3 0. 54.0N 78.7E 200. 13 B 40

2050003 245.04 SB 8.0 0. 53.3N 78.3E 200. 13 B 40

2050004 245.05 SB 5.5 0. 52.3N 79.1E 200. 13 B 40

2051001 248.01 SB 8.3 0. 53.5N 80.2E 200. 13 B 40

2051002 248.02 SB 6.6 0. 53.5N 80.2E 200. 13 B 40

2052001 251.01 SB 4.3 0. 53.4N 79.0E 200. 13 B 40

2052002 251.02 SB 11.3 0. 52.7N 83.0E 200. 13 B 40

2053001 254.01 SB 6.4 711. 53.8N 91.4E 375. 13 B 40

2053002 254.02 SB 3.5 0. 53.8N 91.4E 375. 13 B 40

2053003 254.03 SB 4.9 394. 52.2N 90.7E 750. 13 B 40

2053004 254.04 SB 6.7 377. 53.2N 90.5E 375. 13 B 40

2053005 254.05 SB 5.5 357.. 53.2N 90.5E 375. 13 B 40

2053006 254.06 SB 10.6 658. 52.1N 90.7E 750. 13 B 40

2053007 254.07 SB 6.3 473. 53.2N 90.5E 375. 13 B 40

2053008 254.08 SB 6.4 938. 56.1N 92.8E 0200 13 B 40

2054001 257.01 SB 17.4 948. 50.3N 87.6E 1600 13 B 40

2060001 VV14 SB 5.6 768. 50.1N 88.9E 2000 42 B 60

2060002 VV87 SB 3.3 225. 50.1N 88.3E 42 B 60

2060003 VV2 SB 8.7 869. 50.1N 88.9E 2000 42 B 60

2060004 VV9 SB 4.0 271. 50.1N 88.9E 42 B 60

2069001 SC01 RS 8.7 1180. 49.6N 57.2E 300 19 B

2069002 SC01A RS 6.0 980. 49.6N 57.2E 300 19 B

2069003 SC02 RS 7.8 1237. 49.6N 57.2E 300 19 B

(Also map showing the location of the data points).

Other problems with the Zinke et al. and STATSCO databases are likely to have arisen due to selective sampling in agricultural prospecting (which was the reason most soil profiles were first taken), and exclusion (in the STATSCO database) from the database of shallow soil profiles from the 100cm depth calculation. The separate STATSCO dataset of carbon in soil profils only to 25cm depth and to 50cm depth show a greater number of samples, each with much less carbon. Since desert soils are generally shallower than 100cm depth, and since much of the soil work in the SW deserts of the USA has been to assess sites potentially suitable for agriculture, there is a bias in favour of deep soils in favourable sites such as hollows which collect more water and sediment (Zinke et al.).

We are moving in the direction of improved understanding of spatial distribution of soil carbon, but also need to be sceptical. Figures in a data base look impressive, but they are biased toward selection of potential agricultural sites. Yet modellers have ended up using figures quite happily which suggest as much carbon in desert soils as in a temperate forest!

Another source of confusion in terms of understanding spatial heterogeneity and overall totals of soil carbon is the important and fundamental distinction between organic and inorganic carbon. Calcrete, carbonate precipitated into the soil, often as nodules or a crust near the surface, is a major reservoir of carbon in present-day soils in arid and semi-arid regions. Inorganic carbon is excluded from the Zinke et al. and STASGO databases, butthere is still possibiity of confusion from other sources. Many biogeochemists (e.g. Schlesinger 1992) discuss this soil reservoir in a purely empirical sense, pointing out its importance but not its actual behaviour as part of the carbon cycle. Yet its effect on the carbon cycle is opposite to that of organic carbon! Any decrease in calcrete, by chemical weathering, takes up CO2 from the atmosphere, at least temporarily. A decrease in organic matter in a soil should increase atmospheric CO2. This is an obvious distinction, yet it is suprising how many good scientists fail to notice this when they discuss soil carbonate reservoirs.

In summary, the extraordinarily high values for soil carbon that are found in published soils databases from cool desert bush regions of the western USA and central Asia can thus be assigned to a combination of problems;

1. Most samples in the Zinke et al database have been mis-assigned in the biome sense; they fall in steppe regions and not desert.

2. Of those samples which do fall within desert regions, many were gathered with an eye to agricultural potential and so favour locally moist pockets within the landscape, no matter how unrepresentative these might be.

3. Most semi-desert and desert soils do not reach 100cm depth and are thus excluded from the STATSCO database (this is allowed for in the Zinke et al. database, however), which favour locally deeper soil pockets in moister sites. P>4. Even those areas in the western USA which do fall within 'desert' as generally described from that region are in fact from relatively dense sagebrush semi-desert, which is capable of sustaining commericial ranching. On a global scale this would not qualify as 'true' desert, yet the high carbon storage values are extrapolated to include all cool desert zones, no matter how dry.

The Zinke et al. database is an impressive piece of work; when something on this scale is attempted it is inevitable that some errors will enter the system. It is unfortunate that carbon cycle modellers have not noticed these problems, instead (in effect) preferring to treat data from a published source as if it were infallable. Even a very cursory knowledge of soil geography and ecology would have made clear that there was something badly awry with the data. As long as modellers continue to ignore the broad 'natural history' of what they are studying, these problems will arise continually and decrease the validity of their work.

Another warning lesson in the potential problems with spatial sampling of soil carbon is evident from the very extensive STATSCO soil carbon database (Fig.1). Although we are not aware of anyone using the database for the particular purpose of modelling natural ecosystems or broadscale soil carbon coverage, we will discuss it further to point out one potential (and not necessarily evident) pitfall within this soil carbon database: the sampling in the STATSCO database is mostly based upon agricultural soil surveys of sites with an existing crop cover. Thus, it is not representative of either semi-natural ecosystems, nor is it a spatially unbiased picture of soil carbon distribution. This gives (for example) much lower average soil carbon contents for the cool temperate forest climate zone (around 70 tC/ha), than for the Zinke et al database (about 120-140 tC/ha) which selected semi-natural forested sites rather than cultivated land. As with the desert zones. This is true even in portions of the country which are predominantly forested; still, most of the STATSCO sites are from cultivated areas within those. It is always necessary to be wary of the land use heterogeneity of each region, and the origonal process by which data points were selected.

2) Peatlands and their carbon content.

Various figures are used in the literature to suggest how much carbon there is in peat bogs at present. Often in models of the carbon cycle one sees a single figure accepted as 'the' figure for the peatland carbon reservoir. Yet (as the table below shows) there is room for considerable scepticism that we are even close to the true figure for peatland carbon storage. Accepted figures for peatland carbon in the high latitudes have changed over past 30 years or so; recent attempts deal more explicitly with spatial heterogeneity in northern peatland regions suggest that there may be much more peat than previously thought. There is a need for further careful field sampling and GIS work to back up these assertions. Much of the difficulty at present centres on the varying proportions of 'hummocks' and 'pools' in peat bogs in different regions. If there are problems with adding up the amount of peat in the high latitudes, the problems are much greater for the tropical peats. We still essentially have no idea how much peatland carbon there is in the tropics; it could be hundreds of gigatonnes according to at least one geologist who has worked on tropical peats (H. Faure pers. comm.).

The problem of allowing for spatial heterogeneity within peat bogs is difficult enough. A further problem is distinguishing the area covered by peat bogs, as opposed to other forms of carbon-containing ecosystems such as tundra or closed forest. This remains a further uncertainty which is difficult to allow for in calculating broad scale carbon storage: it is important to subtract the area of land surface covered by peat bog from that covered by better-drained vegetation types, or one would end up counting the same area twice. Yet one needs a clear definition of each of them in order to do this. There is in fact no clear and consistently used definition in the literature of exactly how one should distinguish peat bogs from other ecosystems. Hence there is considerable ambiguity in many studies on global or local carbon storage as to where the category of 'peat' ends and that of 'soil' begins. Without enough care and understanding they might easily be dealt with in either an overlapping or incomplete way, so that when an estimate for global soil carbon storage is added to an estimate for global peat carbon storage, some of the areas could be counted twice or not at all. One widely accepted definition for peat is a pure organic layer at least 20 cm in thickness, and this was used by the widely cited studies by Post et al. and Zinke et al. However one can take as another example the study of Canadian peatland areas by Tarnocai (1980), which defines peatlands as having peat depths (i.e. an organic matter layer) greater than 40cm, and mineral wetlands as having an organic matter layer of less than 40cm. The recent wide-ranging study of northern peatlands by Gorham (1992) used a minimum figure of 30cm organic matter as its dividing line between peat and non-peat, so there is no sign of a true consensus emerging!

For the tropics, the actual area and depth of peatland remains largely unknown (Walter 1971, Radjagukuk 1985). For this reason it has been necessary to exclude tropical peatlands from our estimates (instead ascribing to these areas a 'normal' forest and soil cover). This probably leads to the present study underestimating the carbon that has accumulated in the tropics since the early Holocene, although the underestimate will be compensated to some extent by the thinner forest cover which tropical swamp areas tend to have (Walter 1971).

From the values in the table below, we see that peatland carbon mass is still an unknown. It may be much more than has generally been suspected. Of the estimates for global peatland, the recent one of 461 Gt compiled by Gorham (1992) is generally seen as the most robust by experts in the field whom we have spoken to (e.g. R. Clymo, C. Kreminetski). It was based on a wide range of data sources including much recently gathered data, and compiled using a GIS-based approach. However, the latest estimates currently emerging from studies on Canadian peatlands (C. Kreminetski pers. comm., May 1994) seem to suggest figures of around 200 Gt for Canada alone, and if one takes the Russian and Scandinavean peatlands as together containing about twice this amount of carbon (a conservative estimate) this would give a total of around 600 Gt. This figure also does not include any tropical peatlands, which might well contain 100 Gt or more of peat carbon (H. Faure, unpublished calculations).


860 Gt (8.) Peats, present-day Bohn (1976) (s.)
300 Gt (8.) Peats Sjors (1980) (s.)
202 Gt (8.) Peats Post et al. (1982)
377 Gt (8.) Peats Bohn (1976,82)
500 Gt (8.) Peats Houghton et al. (1985) (s.)
249 Gt (8.) Northern peatlands Arm.& Men. (1986). (s.)
210 Gt (8.) Boreal peatlands Oeschel (1989) (s.)
180-227 Gt (8.) Peats Gorham (1990) (s.)
461 Gt (9.) Subarctic and boreal peat Gorham (1992)
500 Gt (10.) Global peats Markov et al. (1988) (s.

(6.) Based on a range of IBP and other data. With 36 Gt lost from soils since the mid-1800's.

(7.) Siegenthaler & Sarmiento 1993, in a box model summary, using numbers approximating to 1990 IPCC assessment.

(8.) Values cited by Schlesinger (1985).

(9.) Of the estimates for global peatland, the recent one of 461 Gt compiled by Gorham (1992) is generally seen as the most robust by experts in the field whom I have spoken to (e.g. R. Clymo, C. Kreminetski). It was based on a wide range of data sources including much recently gathered data, and compiled using a GIS-based approach. However, the latest estimates currently emerging from studies on Canadian peatlands (C. Kreminetski pers. comm., May 1994) seem to suggest figures of around 200 Gt for Canada alone, and if one takes the Russian and Scandinavean peatlands as together containing about twice this amount of carbon (a conservative estimate) this would give a total of around 600 Gt. This figure also does not include any tropical peatlands, which might well contain 100 Gt or more of peat carbon (H. Faure, unpublished calculations).

(10.) Cited from the Russian by C. Kreminetski (Acad. of Sciences, Moscow), who regards this as a conservative estimate and probably an underestimate.

What is evident from this range of values is that at present there is little justification in selecting any one particular value from the literature to represent the size of the global peat carbon reservoir. Despite the impression of 'reliability' in the published values that are cited again and again from the biogeochemical literature, the subject is (if you will excuse the pun) on very shaky ground. If anything, all previous published values for global peat are likely to be serious underestimates.

3) Direct CO2 effects on soil carbon storage

There has been a fair amount of discussion during recent years, concerning how rising ambient CO2 evels might affect soil carbon storage, through affecting the plants that root into the soils and supply carbon to the soil reservoir. Various models have attempted to allow for aspects of the CO2-fertlization effect on ecosystem processes. At present it seems that those models which do include direct CO2 fertlization factors on plant growth do so inadvisably. It seems that soil effects will be even harder to model because of the spatial complexity of direct-CO2 effects.

Some clues to the way in which higher CO2 levels might affect soil processes can be gleaned from experiments which have been carried out on plant growth (e.g. Wullschleger et al. 1995). Under higher-than-present future CO2 levels, a higher photosynthetic rate of vegetation could have mean changes in the net flux of primary production reaching the soil as dead leaves, roots, branches etc. The actual extent to which CO2 fertilization of plant growth occurs in natural and semi-natural ecosystems is highly uncertain (Koerner & Arnone 1992, Allen et al. 1987, Koerner & Arnone 1992, Mooney & Koch 1995, McConnaughy et al. 1993). The rapid rise of 80-90ppm CO2 (of the same order as the glacial-interglacial increase) over the last 200 years has provided no clear evidence of a ldirect-CO2 effect on vegetation productivity or biomass (Adams & Woodward 1992, Shiel & Phillips 1995) that cannot possibly be explained by climate fluctuation and increased sulphur and nitrogen fertilization by air pollution. Basically, the direct CO2 effect must be present in natural vegetation to some extent, but how important it is in affecting the productivity of vegetation is largely a matter of guesswork.

Quite possibly, even where there was no significant change in plant growth rate above ground, the result of higher CO2 in terms of faster turnover time of roots below ground would constitute a substantial increase in the amount of organic carbon to the soil, without any corresponding increase in microbial decomposition rates. For example, some experiments on CO2 fertilisation in artificial tropical microcosms have found that raised CO2 levels above 600 ppm give faster turnover rates of fine roots than under present-day ambient CO2 levels (Mooney & Koch 1994). Many experiments on both tropical and temperate plants (Mooney & Koch 1994) also indicate that at higher CO2 levels, the root mass is increased much more than the aboveground material; this might have implications for the supply rate of organic matter directly into the soil from dead root material, but once again the subject is potentially so complex and so poorly understood that any attempt to model it on a global scale seems almost unwarrented.

Bazzaz (1990) has suggested that changes in the carbon-to-mineral ratios in the plant materials reaching the soil surface could also have far reaching effects on the levels of long-lived carbon in soils. Under higher CO2 there might for instance be a higher ratio of carbon to minerals in the soil litter (due to relative carbon starvation of the plants), slowing the rate of fungal and bacterial decomposition. This would tend to result in more carbon accumulating in the soil.

Others have suggested that the opposite could occur, with less carbon accumulating in soils despite the greater input of plant material from CO2-fertilized plants, due to a 'priming' effect in which microbial activity on the new organic matter is increased. This is because at higher growth rates each resource type becomes more steadily available (due to increased total input of materials from each plant type) and better able to maintain a specialised microbial community that will break it down quickly (Schimel et al. 1995).

In their experiment on several artificial tropical ecosystems exposed to high CO2 levels, Koerner & Arnone (1992) found a decrease in soil carbon at higher-than-present CO2 concentrations (i.e. lower CO2 = more soil carbon), which is what the 'priming' hypothesis predicts.

However, the result may well vary greatly on a spatial basis, with soil type and with overlying crop and vegetation type. A recent experiment comparing C4 sorghum with C3 soybean has shown (Torbert et al. 1998). Under sorghum, the soil carbon increased under the higher-CO2 regime, whereas under soybean it decreased. Torbert et al's experiment contrasts two very different photosynthetic metabolisms, and it is not clear how more subtle differences in metabolism and in growth characteristics might affect the soil carbon response to raised CO2. However, this degree of complexity is a bad omen for any real prospect of global-scale modelling of direct-CO2 effects on soils. In natural plant communities, as well as cropland mosaics, the very direction of change in soil carbon might vary on a scale of few centimetres depending on the individual plant species growing above.

Despite certain bold attempts to model CO2 effects on the present-day and future biosphere (Esser 1984, 1987), there are grounds for considerable scepticism that at our present state of knowledge we can even approximately quantify the effects of the present anthropogenic phase of CO2 increase on broadscale ecosystem processes (Mooney & Koch 1994, Wullshetger et al. 1995, Amthor & Koch 1996). These problems apply both to plants, and (even more severely) to the soils underneath them which are only indirectly affected by the CO2 rise. Soils are of course heterogenous, the responses of the plants themselves to vary in response to CO2 changes are very poorly understood and (see below) soil carbon may take many decades and centuries to equilibriate. These complexities of spatial and temporal patterning in soil and vegetation responses may well mean that the soil carbon response to direct-CO2 effects is unquantifiable for the forseeable future at least.

4) How rapidly will soil carbon increase with succession to natural ecosystem cover?

There is a tendancy to assume that as we leave any area of disturbed vegetation alone, with time the soil carbon will build up. So if we stop cultivating an area, soil carbon will accumulate as it goes to natural vegetation. Some general understanding of these processes has already been attained but a great deal of further consideration will be necessary before realistic understanding can be reached. However, various complexities are evident.

It is generally accepted that with succession from cultivated to natural vegetation in humid climates, soil organic carbon will build up. This generally seems to hold true in the temperate zones (Johnson 1992) but it is not always the case, at least in the tropics. For example a study in Hawaii finds that going from sugar cane fields to secondary tropical forest results in a net decrease in soil carbon (Bashkin & Binkley 1998) . This seems reasonable if one compares natural prairies with other ecosystems, but is still perhaps in general a suprise. Such a result shows once again that the direction in which the soil carbon storage moves depends on where one is, and precise details of the agricultural system. Have to consider details before we can understand generalities.

In order to gain further understanding of how soil carbon may vary in response to future climate and land use change, we may need to zoom out to look on a very different scale, looking over broad spans of time rather than spatially. This is the perspective that comes from the Quaternary sciences, which study changes over thousands of years and tens of thousands of years, rather than the years, decades and centuries that soil scientists and biogeochemists most commonly think on.

There is evidence from well-dated studies that a natural disequilibrium in soil carbon can last thousands of years following climatic changes or other more localised disturbances in the environment (Schlesinger 1990). In addition to the examples of increasing carbon storage cited by Schlesinger, another more recent example found by Schwartz (1991) shows how the 'imprint of the past' can persist in a soil's carbon reservoir for thousands of years. Schwartz found that central African savanna soils still contain some carbon at depth that bears the isotopic imprint of forest vegetation, and the deeper into the soil profile one goes, the more and more forest carbon one finds. In a large total column depth (many tropical soils are very deep) the overall amount of 'old' carbon from the previous ecosystem could be quite large. No-one has yet attempted to assess this factor systematically on a global scale.

From dating this carbon it seems that these areas were covered by forest during the early Holocene, and that this relatively small deep soil reservoir still persists thousands of years later after the forest has retreated. One should ideally take into account the possibility that the soil carbon density we observe in natural sites nowadays might differ greatly from the levels back at 8,000 years ago or 5,000 years ago, when the soil carbon might not have had as much chance to equilibrate with the vegetation conditions. A similar situation could exist in a future scenario of climate change, affecting the patterns in sol carbon storage.

There can be no doubt that disequilibrium in carbon storage is especially significant in the case of peat build-up. There is abundant evidence that this process can continue at more-or-less the same rate for many thousands of years, adding incrementally to a waterlogged column of almost pure organic matter which can reach many metres in thickness (although it may eventually reach an equilibrium point at which input is balanced by overall decay rate down through the peat column; Clymo 1984).

There is some circumstantial evidence that the slowness of soil development may have retarded vegetation colonisation of many formerly glaciated or barren areas, for as long as hundreds or even thousands of years. Possible clues to the importance of this effect include the surprisingly slow rate of recolonisation of deglaciated landscapes by local tree populations after sudden warming events in northern Europe and North America at around the beginning of the Holocene (Pennington 1977). As was mentioned above, Magri (1994) has found slow exponential rises in pollen input to an enclosed lake basin in central Italy, for relatively constant species composition, taking thousands of years during which the sites were apparently being recolonised by vegetation following climatic disturbance events. She suggests that this pattern might be due to lags in vegetation build-up resulting from slow soil maturation and nutrient limitation (Magri 1994). It is also important to bear in mind that carbon storage could have been affected by more subtle and undetectable differences in vegetation structure that might have persisted in many ecosystems that formed on previously barren surfaces in the high and low latitudes following the last glacial cold period.

Just as small-scale disturbances tend to throw back the process of carbon accumulation in ecosystems, one can imagine global-scale changes in both the past and the future having a similar effect. There is some circumstantial evidence of this from Quaternary vegetation history. Disequilibrium in species migrations and in soil development may have been important in producing some of the 'no-present-analogue' species assemblages (discussed above) that occurred during the late glacial and early Holocene (Adams & Woodward 1992), when the Earth's climates and ecology were changing fastest. Ecological disequilibrium in vegetation, particularly in forest vegetation, may have prevented maximum vegetation carbon storage from being reached for thousands of years after the climate initially became suitable for it in many areas of the world, in both temperate and tropical environments (Adams & Woodward 1992). Lags in vegetation development could themselves have fed back in terms of lags in soil carbon storage, and this in itself may have led to reductions in vegetation. These are poorly understood processes which we can't readily explain from our limited present-day perspective, yet they may be important in the future.

Despite such concerns about disequilibrium in soil carbon storage following past or future environmental change, it seems that at many sites most of the humic carbon entering such soils does so within the first millennium after formation (Schlesinger 1990, and see discussion in Adams & Woodward 1992), and very often within the first few decades. Thus the longer-term lag will probably not be especially great as a proportion of the total carbon in a soil. This general picture of a very rapid initial build-up of soil organic matter following a change in circumstances is found in a great many studies from around the world, in many different sorts of vegetation.

To take just one fairly representative example, in the classic Rothampstead experiments in England where arable land was allowed to revert to deciduous temperate woodland, soil organic carbon increased 300-400% from around 20 t/ha to 60-80 t/ha in less than a century (Jenkinson & Rayner 1977). The rapidity with which organic carbon can build up in soils is also indicated by examples of buried steppe soils formed during short-lived interstadial phases in Russia and Ukraine. Even though such warm, relatively moist phases usually lasted only a few hundred years, and started out from the skeletal loess desert/semi-desert soils of glacial conditions (with which they are inter-leaved), these buried steppe soils have all the rich organic content of a present-day chernozem soil that has had many thousands of years to build up its carbon (E. Zelikson, Russian Academy of Sciences, pers. comm., May 1994).


In this brief review we have attempted to point out some of the difficulties which remain in assessing and forecasting the behaviour of soil carbon reservoirs.

We note that it is often important to 'zoom in' to the small spatial scale for further understanding of soil carbon processes, and checking of generally accepted figures for reservoirs and responses. It is necessary to realise that many broad extrapolations published in the literature do not subdivide the environment enough, or in a clearly defined way. Extensive soils databases do not always contain data that are reliable for the purposes one might want to use them for, because of the precise ways in which the samples were taken. Equations which predict the rate of carbon buildup in a soil over time, or the CO2-fertlization effect on total soil carbon, are unlikely to be reliable under anything more than the most spatially restricted and controlled conditions. Databases of peat carbon storage may well be unreliable if they make incorrect assumptions about small-scale spatial patterns within the peat bogs. Thus, simply standing back and looking at global soil and peat carbon on a regional and global scale - as is necessary for global modelling - can lead to very unreliable conclusions if one does not also carefully allow for the local scale complications.

Paradoxically, there are also occasions when standing back still further can give a better perspective for current understanding of soil carbon storage. This includes the extra dimension of long time scales, where ecosystems have continually retreated and returned in the face of large global climate changes. It is continually necessary to look across a range of spatial and temporal scales, if one is to attain a realistic understanding of present and future patterns and processes.

Sometimes, all one gains from this added perspective of different spatial and temporal scales is a pessimistic outlook. One may have to acknowledge that there are certain aspects of the soil carbon cycle which cannot realistically be modelled in the foreeable future, because small-scale heterogeneity in reservoirs and responses is too great. This includes the direct-CO2 effect on soil carbon, which can be affected by an array of factors, including the features of individual plants that are rooted into the soil. A position of well-founded pessimism is much better than poorly-founded optimism; in science, it is always better to know with good foundation that one may be wrong, than to believe unrealistically that one knows the answer.

Searching for simplicity and numerical rigour in biogeochemistry is in itself desirable, and it is necessary for the full progress of the field. However, it is important not to let the desire to find broad, consistent spatial patterns deteriorate into wishful thinking. For example, a published numeric database of soil carbon (such as Zinke et al., or the STASCO database) is generally regarded as infallable with 'no questions asked' about exactly how the data were gathered. CO2-fertilization parameters are confidently extrapolated from tiny amounts of shaky data, without any real attempt to acknowledge the huge bounds of the uncertainties. Published peatland estimates are taken and used without an analysis of the uncertainties and contradictions within this literature. In the current culture of global biogeochemistry, more anecdotal evidence and opinion from the community of ecologists, foresters and soil scientists is regarded as suspect and rather unworthy; yet as we hope we have shown here, it is in fact vital to the subject.

It is necessary to accept that global biogeochemistry is not currently in the position to be a 'hard' science like physics and chemistry, because ecosystem patterns and processes are too complex and still too poorly understood. This is not a reason to give up; real progress in understanding of the world's ecosystems has been made and will continue to be made. However, only when the importance of continually adding a touch of 'natural history' is acknowledged, will biogeochemical modelling be able to make the greatest progress.


Torbert H.A., Rogers H.H., Prior S.A., Schlesinger W.H. & Runions G.B. (1998).

Johnson D.W. 1992. Effects of forest management on soil carbon storage. Water, Air and Soil Pollution v.64 p.83-120. Adams J.M. (1995). The role of terrestrial ecosystems in glacial to interglacial changes in the global carbon cycle. PhD thesis. Université D'Aix-Marseille II, France Adams J.M., Faure H., Faure-Denard L., McGlade J.M. & Woodward F.I. (1990). Increases in terrestrial carbon storage from the Last Glacial Maximum to the present. Nature v.348 p.711-714. Adams J.M. & Faure H. Palaeovegetation maps of the Earth during the Last Glacial Maximum, and the early and mid Holocene: an aid to archaeologists. In press. Journal of Archaeological Science. Alley R.B., Meese R.B.& Shuman A.J. (1993). A new Greenland ice-core record from GISP2.Nature v.366 p.443-445. van Andel T.H. & Tzedakis P.C. (1996). Palaeolithic landscapes of Europe and environs, 150,000-25,000 years ago: an overview. Quaternary Science Reviews v.15 p.481-500. Barnola J.M., Raynaud D., Koroketvich Y.S. & Lorius C. (1989). Vostok ice core: a 160,000 year record of atmospheric CO2. Nature v.329 p.408-414. Broecker W.S. & Peng T-H. (1993). What caused the glacial-to-interglacial CO2 change? In; The Global Carbon Cycle. NATO ASI Series, v.15 p.95-115. M. Heinmann ed. Brown S. & Lugo A.E. (1992). The storage and production of organic matter in tropical forests and their role in the global carbon cycle. Biotropica v.14 p.161-187. Cannell M.G.R. (1982) World Forest Biomass and Primary Production Data. Academic Press. London. Chappellaz J.M., Blunier T., Ratnaud D., Barnola J.M., Schwander J. & Stauffer B. (1993). Synchronous changes in atmospheric CH4 and Greenland climate between 40 and 8 kyr BP. Nature v.366 p.443-445. Clapperton C.M. (1993). Quaternary Geology and Geomorphology of South America. Elsevier, London. CLIMAP (1976). The surface of the ice age earth. Science v.191 p.1131-1136. Clymo R.S. (1984). The limits to peat bog growth. Phil. Trans. R. Soc. Lond. B303. p.605-654. Colinvaux P.A. (1987). Environmental history of the Amazon Basin. Quaternary of South America and Antarctic Peninsula. v.5. A.A. Balkema, Rotterdam, Brookfield. Crowley T.J. (1994). Pleistocene temperature changes. Nature v.371 p.664. Crowley T.J. (1995a) A new reconstruction of ice age terrestrial carbon changes. p.56, Abstracts, 14th INQUA Congress, Berlin. Crowley T.J. (1995b) Ice age terrestrial carbon changes revisited. Global Biogeochem. Cycles. in press. Crowley T.J. & North G.R. (1991). Palaeoclimatology. Oxford University Press, Oxford. Cullen J.L. (1981). Microfossil evidence for changing salinity patterns in the Bay of Bengal over the last 20,000 years. Palaegeogr., Palaeoclim., Palaeoecol. v.35 p.315-356. Esser G. (1984). The significance of biospheric carbon pools and fluxes for atmospheric CO2: a proposed model structure. Prog. Biometeorol. v.3 p.253-294. Esser G. (1987). Sensitivity of global carbon pools and fluxes to human and potential climatic impacts. Tellus v.39B p.245-260. Faure H. (1990). Changes in the continental reservoir of carbon. Palaeogeog. Palaeoclim.. Palaeoecol. v.82 p.47-52. Franzen L. (1994). Are wetlands the key to the ice-age cycle enigma? Ambio v.23 p.300-308. Giresse P., Malounguila-Nganga N. & Moguedel G. (1990) La succession des palaeoenvironnements Quaternaires des plateformes contentales du sud du Gabon, du Congo, du Cabina et du Zaire p.71-77 in; Lanfranchi R. & Schwartz D. (ed.s). Paysages Quaternaires de L'Afrique Centrale Atlantique. Editions de l'Orstom. Paris 1990. Goreham E. (1991). The role of subarctic and boreal peatlands in the global carbon cycle, and their probable responses to 'greenhouse' climatic warming. Ecological Applications v.1 p.182-195. Grichuk V.P. (1992) Main types of vegetation (ecosystems) for the maximum cooling of the last glaciation. Frenzel B., Pecsi B & Velichko A.A.; Atlas of Palaeoclimates & Palaeoenvironments of the Northern Hemisphere. INQUA/Hungarian Academy of Sciences. Budapest. Haffer J. (1969). Speciation in Amazon forest birds. Science v.165 p.131-137. Harden J.W., Sundquist E.T., Stallard R.F. & Mark R.K. (1992). Dynamics of soil carbon during deglaciation of the Laurentide Ice Sheet. Science v.258 p.1921-1923. Harmon M.E. & Hua C. (1992) Coarse woody debris dynamics in two old-growth ecosystems. Bioscience v.41 p.604-610. Huston M.A. (1994). Biological Diversity: the coexistance of species on changing landscapes. Cambridge University Press, Cambridge, UK. van der Kaars W.A. (1990) Late Quaternary Vegetaton and Climate of Australia as reflected by the palynology of eatern Indonesian deep sea piston-cores. PhD thesis, University of Amsterdam. Klinger L.F. (1991). Peatland formation and ice ages: a possible Gaian mechanism related to community succession. Ch 28. p. 247- 255 In; Scientists on Gaia. ed.s S.H. Schneider & P.J. Boston. MIT Press, London. Klinger L.F. (1996).Potential role of peatland dynamics in Ice-Age initiation. Quaternary Research v.45 p.89-92. Koerner C. & Arnone J.A. (1992). Responses to elevated carbon dioxide in artificial tropical ecosystems. Science v.257. p.1672-1674. Lottes A.L. & Ziegler A.M. (1994). World peat occurrence and the seasonality of climate and vegetation. Palaeogeogr., Palaeoclim., Palaeoecol. v.106 p.23-37. Maley J. (1987). Fragmentation de la foret dense humide africaine et extension des biotopes montagnards et chronologiques. Implications palaeoclimatiques et biogeographiques. Palaeoecology of Africa v.18 p.307-334. Maslin M., Adams J.M., Thomas E. & Faure H. (1994) The carbon transfer between oceans, atmosphere and the terrestrial biosphere since the Last Glacial Maximum: a review of two contrasting methods of estimation. Terra Nova. McConnaughay K.D.M., Bernston K.D.M. & Bazzaz F.A. (1993). Plant responses to carbon dioxide. Nature v.361 p.24. Meadows M.E. (1988). Late Quaternary peat accumulation in Southern Africa. Catena v.15 p.459-472. Mooney H.A. & Koch G.W. (1994). Impact of rising CO2 concentrations on the terrestrial biosphere. Ambio v.23 p.74-76. Morley R.J. (1981) Development and vegetation dynamics of a lowland ombrogenous peat swamp in Kalimantan Tengah, Indonesia. Journal of Biogeography v.8 p.383-404. Neftel A., Oeschger H., Staffelbach T. & Stauffer B. (1988). CO2 record in the Byrd ice core 50,000-5,000 years B.P. Nature v.331 p.609-611. Neustadt M.I. (1982). Holocene peatland development. in; Quaternary Environments in the Soviet Union, ed; A.A. Velichko p.201-206. Nichols H. (1969). Chronology of peat growth in Canada. Palaeogeog, Palaeoclim., Palaeoecol. v.6 p.61-65. Olson J.S., Watts J.A. & Allinson L.J. (1983) Carbon in Live Vegetation in Major World Ecosystems. Environmental Sciences Division Publication No. 1997. Oak Ridge National Laboratory, Tennessee. Peng C. (1994). Reconstruction du stock de carbone terrestre du passe a partir de donnees polliniques et de modelels biospheriques depuis le dernier maximum glaciaire. Abstract, PhD thesis. Universite d'Aix-Marseille II. Peng C.H., Guiot J. & van Campo E. (1994). Reconstruction of past terrestrial carbon storage in the Northern Hemisphere from the Osnabroeck Biosphere Model and palaeodata. In press, Clim. Res. Post W.M., Emanuel W.R., Zinke P.J. & Stangenberger A.G. (1982) Soil Carbon Pools & World Life Zones. Nature v.298 p.156-159. QEN (1995). The Quaternary Environments Network Atlas and Review of Palaeovegetation during the last 20,000 years. World Wide Web address; Radjagukuk B. (1985). Prospects of peat exploitation in Indonesia. p.99-112. Tropical Peat Resources - prospects and potential. International Peat Society. Proceedings of Symposium at Kingston, Jamaica, 1985. IPS, Helsinki. Ritchie J.C. & Cwynar L.C. (1982). The late Quaternary vegetation of the North Yukon, in; Palaeoecology of Beringia. ed.s D.M. Hopkins, J.V. Matthews, C.E. Schweger & S.B. Young. Academic press, London p.113-126. Robinson J.M. (1994). Atmospheric CO2 and plants. Nature v.368 p.105. Schlesinger W.H. (1982). Carbon storeage in the caliche of arid soils: a case study from Arizona. Soil Science v.133 p.247-255. Schlesinger W.H. (1990). Evidence from chronosequence studies for a low carbon storage potential of soils. Nature v.348 p.232-234. Shier C.W. (1985). Tropical peat resources - an overview. p.29-46. Tropical Peat Resources - prospects and potential. International Peat Society. Proceedings of Symposium at Kingston, Jamaica, 1985. IPS, Helsinki. Spasskaya I.I. (1992). Dominant geomorphic processes during the maximum cooling of the last glaciation. Frenzel B., Pecsi B & Velichko A.A.; Atlas of Palaeoclimates & Palaeoenvironments of the Northern Hemisphere. INQUA/Hungarian Academy of Sciences. Budapest. Tallis J. (1990). Plant Communities. Academic Press, London. Thompson R.S. & Mead J.I. (1982). Late Quaternary environments and biogeography in the Great Basin. Quaternary Research v.17 P.39-55. Times/Harper Collins (1992). Times Atlas of the World: Comprehensive Edition. Harper Collins, London. Van Campo E., Guiot J. & Peng C. (1993). A data-based re-appraisal of the terrestrial carbon budjet at the last glacial maximum. Global and Planetary Change v.8 p.189-201. Velichko A.A. & Spasskaya I.I. (Ed.s) 1991. Legend of regions of the land during maximum stage of late Valdai glaciation. Geogeodezia SSSR. (also includes accompanying maps by a large number of authors). Walter H. (1971). Ecology of Tropical and Subtropical Vegetation. Oliver & Boyd, Edinburgh. Warner R. (1993). Peat accumulation over time. The Holocene v.3 p.110-123. Watts W.A. & Stuiver M. 1980, Late Wisconsin Climate of Northern Florida & the Origin of Species-Rich Deciduous Forest. Science v. 210 p.325-327. Webb R. et al. (1995). Bibliography and inventory of sites with information for constructing digital global maps of vegetation distribution 18,000 yr. B.P., NOAA Palaeoclimatol. Publ. Ser. Rep. Nat. Geophys. Data Cent., Boulder Colo., in preparation. West. R. (1978). Quaternary Geology and Biology, Cambridge University Press, Cambridge UK. Williams D.L., Dunkerley D.L., De Dekker P., Kershaw A.P. & Stokes T. (1993). Quaternary Environments. Edward Arnold, London. Williams M.A.J. & Clarke M.F. (1984). Late Quaternary environments in north-central India. Nature v.308. p.633-635. Zinke P.J., Stangenburger A.G., Post W.M., Emmanuel W.R. & Olson J.S. (1984). Worldwide organic soil carbon and nitrogen data. Environmental Sciences Division, publication no.2212. Oak Ridge National lab/ US Department of Energy. Zoltai S.C. (1991). Estimating age of peat samples from their weight: a study from west-central Canada. The Holocene v.1 p.68-73.


Amongst the soil values, the mis-assignment of high carbon storage to some 'desert' soils in the Zinke et al. database has led previous studies to assign implausibly high carbon storage to the desert zones of the LGM (P. Zinke pers. comm. 1997), in which temperate desert soil carbon content rivals that of forest ecosystems. Relatively small numbers of samples used. Sampled in agricultural areas/agric surveys very often (Zinke pers. comm.)

62 tC/ha (1.) cool/cold semidesert/shrubland Zinke et al. (1985) 97 tC/ha (2.) cool temperate desert Zinke et al. (1985) 14 tC/ha (3.) warm temperate desert Zinke et al. (1985) 99 tC/ha (4.) cool temperate desert bush Zinke et al. (1985)

(1.) Based on only 5 samples, 4 from North America. SD = 4.6. (2.) Based on only 4 samples, all from North America. SD = 0.7. Hence, this is in fact treated as semi-desert here. (3.) Based on 9 samples, all from North America. Hence treated as semi-desert here. SD = 0.7. Cool-temperate on Holdridge diagram is defined as having mean annual biotemperate < 12 degrees C (this would seem to include semi-desert north of about 45 degrees latitude in present-day climates). (4.) based on 129 samples. SD = 6. Most of these samples are from North America and are hence suggested as being semi-desert. The other 24 values from Asia show a rather lower mean of about 74 tC/ha (SD=3.4).

Figure for deserts; as much C as a temperate forest. (Schlesinger says not).