2006-07-21 Göte Bertilsson

PERSPECTIVES ON SOIL CARBON.

And a simple model for practical farming, Cpersp.

More background in Appendix.

Introduction.

Soil carbon is important. Humus - the word itself gives associations to soil fertility. And humus is soil carbon. A more recent concern is the role of soil carbon in the global carbon cycle and the relation to greenhouse warming.

We know quite a lot about factors influencing soil carbon. Manure and leys are positive, excessive tillage is negative. Numerous field experiments have been carefully analyzed and reported. There is a lot of information in our scientific litterature.

The executive manager of agricultural soil carbon is the farmer. He is in charge but often he does not think of it. The tools at his disposal are a part of the normal shortterm operational measures: which crops to grow, how handle harvest residues, use of cover crops, how and when to till the soil. The key issue is to provide him with adequate information to also include the longterm aspects of soil fertility development. That is the aim of this work.

The concept originates from three sources:

Discussions within the farm environmental project “Odling I Balans”. 19 pilot farms scattered over Sweden work with the common goal to improve the environmental performance of agriculture.
The German system “Humusbilanzierung”, launched by the counterpart to NJF: VDLUFA.
The work at the Swedish Agricultural University, especially the integrated process thinking summarized by the model ICBM (Andrén et al).

This resulted in the system “Carbon Perspectives”, Cpersp.

Why a “new” system?

To be practically workable in contact with a farmer the needed input for the work must be directly available. The farmer knows about crops and yields and farm operations but cannot quantify for instance soil moisture or climate data. Also, he works with a system, not with a crop.

Humusbilanzierung has gone a long way in this direction but concentrates on summarizing the effects of rotations and crop factors only without considering the individual soil.

Inclusion of a process concept like in ICBM led to a dynamic “model” based on common farm data, Cpersp.

It needs to be stressed: Cpersp is for giving perspectives in discussion with a farmer, not scientific predictions of future soil values..

The basic principle.

Cpersp calculates balances of “stable humus carbon”, just as Humusbilanzierung. This means that about 80% of the carbon turnover is outside the calculation.

An example:

We have a soil with 2% organic C and 3 million kg topsoil per hectare. That means 60000 kg stable humus carbon per hectare. We assume that 1.5% is mineralized annually which means an annual loss of 900 kg C.

But we have additions. We assume a cereal system with average yield of 5000 kg per hectare. Straw is plowed in and in addition there are stubble, roots and rhizodeposition. In total 7500 kg organic matter is returned to the soil. With a carbon content of 40% this means 3000 kg gross C input of which 20% ( 600 kg) in the end remains as stable soil C.

Since the loss was 900 there will be a net loss of 300 kg C per year initially, but as the years go by the soil C will go down, the net loss will be reduced and eventually an equilibrium is reached at a soil C content of 1.3% hundreds of years ahead.

The process is described both in figures and diagram. Since this is more for giving management support than for scientific calculations the time span considered should not exceed 30 years.

Quantification issues and influencing factors.

Stable soil C formation.

Converging results från German work (Humusbilanzierung) and Swedish longterm research (Persson): as a rough estimate 20% of the carbon in plant material and 30% from manure will end up as longterm soil C. It could be noted, however, that a lower value for maize residues, around 12%, are reported by Puget et al (2005).

Amount of organic matter returned from crops:

Present guidelines used in Cpersp, expressed as total organic matter input, dry matter.

Cereals: 1.5* grain yield, of which 1/3 is straw.

Oilseeds: 2 * seed yield, of which ½ is straw.

Beets and potatoes: 0.1 * root or tuber yield.

Leys: a flat figure 6000 kg per year, to which “extra” not harvested growth should be added.

For a wheat yielding 8 tons this means an input of 12 tons organic dry matter. This may seem high. It includes 4 tons of straw, maybe 2 ton of stubble etc, 2-3 tons of visible roots. This adds up to at most 9. The rest is rhizosphere deposition, the sum of fine roots, dead roots, root exudation, microorganisms feeding on the root system during the vegetation period.

Berglund et al (2002) have for Swedish conditions given harvest residue and root data for different crops. A further calculation based on these data give the following figures for total input of organic matter in relation to yields: winter wheat factor 2, winter rape 3, barley 1.6. The topsoil deposition will be somewhat lower.

Wilts et al (2004), quote from abstract, SC= total source C input): “The SC input from unharvestable sources was 1.8 times more than SC from aboveground stover when N was added …”

Johnson et al (2006) give a comprehensive review. The importance of the rhizodeposition is stressed. Based on their extensive litterature review they state “ ..estimates are that 2.5 to 6 times the amount of C incorpored into root biomass may be represented as rhizodeposition”. Based on their calculations of crop C turnover, the figures expressed as “input of organic matter to field/ grain yield” will be for wheat 2.6 and for barley around 3. These should for comparisons with topsoil soil C be reduced for the C deposited below topsoil. Nevertheless, the figures given above for Cpersp seem low and conservative in this perspective.

Franko (1997) has given yield-based parameters for estimation of total carbon input from crops. His values are somewhat lower than the Cpersp preliminary standard given above.

Still, they give some support for a fairly high value, but the root/rhizospere contribution relative to straw is too small to explain the results in straw removal experiments.

For leys the connection between harvested vegetation and organic matter input is less clear. ”. For instance, Skinner et al (2006) found no relation between soil C and canopy photosynthesis. The use of a constant value of 6000 kg organic matter (d.m.) per ley year fits well with experimental data, but there is of course a close interrelation with the factor “soil resting months”.

The carbon content of organic material is assumed to be 40% for the Cpersp calculations.

Soil management.

The “normal” value for mineralization is based on autumn tillage which disturbes the soil and promotes mineralization. If we have a ley or overwintering spring sown cover crop this disturbance in the autumn does not occur and the loss of soil C is reduced. Somehow this must be considered. It is also an important practical management factor in the agriculture of today. A concept “soil resting months” is introduced as a preliminary try to include and quantify these aspects.

Present system in Cpersp: since it has a Scandinavian origin 3 winter months are supposed to give near zero mineralization and consequently there are 9 active months. The background for the “normal” mineralization figure is autumn tillage after harvest. If the soil is not disturbed we have a “soil resting month”, which reduces the annual mineralization by 1/9. The soil of overwintering leys is not disturbed at all, which should mean 9 resting months and no mineralization at all. However, when comparing with experimental results it was evident that this is too extreme, and the present guideline is 6 resting months for overwintering leys.

The guidelines given should be interpreted with common sense and local knowledge. In practice this factor is very important. The positive effects of leys cannot be explained by their input of organic matter only.

A problem is that longterm experiments where this factor can be put to test only can be found for leys. Uhlen (1991) describes rotations with different ley combinations: 6 years with 0, 2 or 4 years of ley. Good agreement with soil C development is achieved with 6 soil resting months.

Cover crops.

Cover crops work in two ways: they contribute with organic matter and they may mean that the soil is left longer undisturbed.

Guidelines for cover crops are given in “Humusbilanzierung”. The factors to use in Cpersp are organic matter input and “soil resting months”.

Cover crops is a fairly recent development. Experimental quantification of their carbon effect cannot yet be found. Some circumstantial results from longterm experiments with strong undergrowth of weeds exist from Rothamsted.

Tillage effects.

Tillage effects on soil carbon are well documented but there are some conflicting results concerning the total C sequestration.

West et al (2002) in a review of experimental data, calculate net carbon differences between tillage systems. Both soil fluxes and emissions from fuels etc are considered. In average they found that conversion to minimum tillage could save 370 kg C per hectare and year.

In their review Johnson et al (2006) summarize “ ..minimum source C inputs to maintain soil organic carbon ….averaged 2.5 +-1 Mg/ha based om mouldboard plow sites (n=13) and 1.8+- 0.44 Mg/ha (n=5) based on no-till and chisel plow sites.” This means, indirectly, that in conservation tillage the soil C losses where about 70% of those in mouldboard plow systems. Note that this is expressed as input of organic residue C, whereas normally in this report the resulting “stable soil C” is used.

“Conservation tillage” is a composite term for several practices. Some of them:

No-till

Shallow tillage (harrow or disc)

Chisel plow

Shallow mouldboard plow.

Reicosky (1999) measured carbon losses as carbon dioxide after use of different tillage implements. These direct losses were considerable, amounting to hundreds of kg C per hectare during a few weeks. Compared to mouldboard plow, the losses from chisel plow and disc implement were around 30% and from undisturbed soil 7%. It is concluded that we should minimize the volume of soil disturbed and leave residues on the soil surface.

One way to handle the tillage question in the calculations is to reduce the mineralization in relation to the soil volume left undisturbed. For a preliminary test this relation was included in Cpersp, and below the outcome is compared to research reports.

Borresen et al (1993) reported from a Norwegian experiments: 13 years, spring cereals, 25 cm mouldboard plow compared to 10 cm rotavator. Soil C data are given only for 0-5 cm (2.3 and 2.8 %C respectively) and 10-20 cm (2.3 and 2,3%). Weighed together for 0-25 cm this will be 2.3 and 2.4 % C respectively. At start 1976 soil C was 2.3%. Cpersp gives 2.25 and 2.42% C respectively in year 13 at an assumed mineralization of 1% when the cultivation depths 25 and 10 respectively were used.

Another report by Borresen et al (1994) gives data from an experiment started 1939 with different plowing depths, 12,18 and 24 cm. The soil C is very high, at start about 9% (recalculated from a diagram). Soil C data from 1975, 1981 and 1987 are given per layer. Based on these the soil C in the layer 0-24 are recalculated and used for comparison with Cpersp. After 42 years the soil C according to measurements are 6.9% and 7.5% respectively for 24 and 12 cm plowing depth. Cpersp gives the values 6.7 and 7.6. 1.2% mineralization and 6 resting months för leys were used. Obviously, at these high soil C contents leys also loose soil C. Cpersp overestimated the difference somewhat. But if the real disturbance depth was 15 instead of 12 there had been good accordance.

Puget and Lal (2005) performed a metaanalysis of published results om MT versus conventional. 56 pairs could be included and the average gain of soil C for MT was 330 kg per hectare compared to conventional plowing.

Jarecki and Lal (2005) describe two experiment in Ohio with different tillage, a silt loam soil (Charleston, 41 year experiment) and a clay soil (Hoytville. 16 years). At Charleston there was an MT effect of about 180 kg C annually but “the clay soil did not respond to tillage treatments”. However, the measurements at Hoytville are compatible with a mineralization of 1.5% and an advantage for MT of 300- 400 kg C annually. The C stock is high and the time short.

Wright et al found a soil C increase for NT of about 250 kg C annually in a 20 year experiment. All increase was in the top 5 cm, deeper layers where not affected.. Aggregate stability was higher for NT. No starting C content is given so the dynamics cannot be explored.

Slepetiene et al (2005) have studied a 40 year experiment in Lithuania: conventional plowing, shallow plowing and minimum tillage. Additional treatment and yield information

has been given (pers. comm). Shallow plowing gave insignificant differences in layer 0-30 but higher in 0-20. MT increased soil C in 0-30. Overall, at start the soil C was low 1.2%), and soil C increased in all treatments. Cpersp could describe the development well, with the parameters mineralization 1.7% and MT disturbance 5 cm.

In addition to these effects on soil C gross content, MT improved the humus quality.

Dersch et al (2001) summarized experiments in Austria. They conclude about minimum and reduced tillage that in 10 years these practices resulted in an increase of 4.7 and 3,2 tons respectively of soil C compared to conventional plowing.

There are more critical reports. Dolan et al carefully investigated a tillage experiment in Rosemount, Minnesota. There were clear increases of soil carbon for reduced tillage in the layer 0-20 cm, but “the summation of soil organic carbon over depth to 50 cm did not vary among tillage treatments”. They stress the need for correct sampling protocol with bulk density determinations down to sufficient depth. It is suggested that in conventional plowing both rhizodeposition and downward transport of breakdown substances can increase soil C below 30 cm. To some extent this work casts a shadow of doubt over optimistic reports on carbon sequestration. However, on closer look at the results and probing with Cpersp it appears that the variation in the experiment does not allow conclusions about effects smaller than 800 kg C per year. So – a question mark remains.

Other profile studies show consistent increases in total C stock for reduced tillage (Puget, Lal and Izzuralde et al, 2005), in average an advantage of 280 kg C per hectare and year for no till..

Koch and Stockfisch (2006) introduce another question mark: what happens if a field with conservation tillage occasionally is plowed? They conclude that the accumulated gain in soil C rapidly disappears. A 9 year period of reduced tillage was followed by plowing and the development of soil C after the plowing operation was followed. Within 3 years a decline of 4400 kg C/ha was measured. It seems that this should cancel the possible gains during the 9 year period of reduced tillage, although this comparison cannot be made. Also in this experiment important changes occur in the layer below the plowing depth 30 cm, a layer which was fairly high in soil C in this loess soil.

In the few longterm tillage experiments with information enough to use Cpersp, the tillage depth factor gave good agreement with measurements. However, this input should be used with care.

Soil mineralization.

The breakdown of humus carbon is certainly more complex than can be taken care of by a simple one-step calculation as in Cpersp. Different fractions of soil C will have different decay rates, and the “breakdown activity” may be influenced by external factors not considered in the Cpersp system. Cpersp is not intended for longterm modeling There are scientific models for these issues.. When more detailed knowledge is available, for instance on how different crops and tillage influence mineralization, it can be included in Cpersp.

Mineralization is normally assumed to be about 1-2 % a year (Persson 2003) but the actual site specific value is not known. When Cpersp is used, different mineralization figures can be used to frame in the perspective.

Calibration and validation.

The best based parameter of those mentioned above is the stable soil C formation. The others need to be developed and tested against experimental results. It should be noted that these parameters to a great extent are interrelated and should be seen as a package within the system. For instance: A too high estimate of organic matter input may be compensated by a too low estimate of resting months or a too high mineralization. Two ways were used to reduce this problem:

as far as possible independent research data of for instance crop C input were used
a wide array of different experiments with different treatments and soils were used.

One base for the first stage of the calibration was the data on crop effects etc supplied by Humusbilanzierung, The experimental base is mostly Scandinavian, although one interesting work from Lithuania deserves to be mentioned (Slepetiene et al 2005).

Extensions have been made to longterm Central European, British and American experiments, but this is not included in the background presented here. Cpersp gives results which well describe the C development in the less extreme treatments (except on very erodible sites), but those very old basic experiments have some drawbacks for the validation work of Cpersp:

the initial soil C from before 1900 or so is not available and no common starting value for treatments can be found. The total input of organic material (residues, weeds etc) is often uncertain.
The depletion of soil C may have gone so far that the mineralization is considerably reduced. Cpersp is not intended to cover more than 30-50 years in “normal” agricultural soils. For longer periods scientific models should be used.

Cpersp is intended as a tool for the agriculture of today, with agricultural soils in full production and in a certain equilibrium state. Preferably, its base should not differ too much from this target group. The contradiction and the problem is that quantification of soil C trends requires several decades.

Figure 1.

Example of the diagram output of Cpersp.

Dots: soil analyses, Lines: Cpersp calculations-

For calibration/validation normally two treatments in the same experiment is compared, which means that it is presupposed that the treatments do not affect the mineralization figure. Logaritmic regression is calculated for the soil analysis values available and the mineralization figure is adapted so that the calculations gives best fit at the year 30 considering both treatments. The standard values for other parameters are in accordance with Cpersp guidelines. In this way the values for mineralization in the following list have been obtained.

Table 1.

Experimental base for calibration and validation. (References: separate list in “References”).

“Mineralization” is the value found by Cpersp giving best fit to the experimental background. Sum of deviations from the regression lines at year 30 always <0.3% C mostly <0,1%. Normal guidelines for residues and soil resting months have been followed.

Sources of variation and error.

Soil bulk density, topsoil depth and sampling depth affect the values.

Organic matter from crops depend on species and varieties. Hard to specify.

Input from weeds etc normally not specified.

Erosion losses or additions disturb the picture.

Varying tillage depth causes unstable values in soil C content.

The soil sampling and analysis is sometimes subject to large irregular variations. Especially, when few analyses are available there is an important random effect on the regressions.

In spite of this a fairly coherent list is obtained. It contains all background worked on so far. Not one experiment with sufficient background data is discarded.

Mineralization.

All values are in the range 0.9 – 2.0 %. Of 29 sites, 5 are below 1.1 and 1 is above 1.8. A range of 1.5 +- 0.3 will cover more than 80% of the cases. There are four sites with long-term black fallow. These are of special interest, because the calculation there is unaffected by crop C input and resting months. Their mineralization values are 1.0, 1.2, 1.3 and 1.5% respectively.

No systematic influence by carbon content, clay content or geographical location can be seen.

Work with Cpersp.

Two examples.

1. Cereals 3 years, sugar beet 1 year. High yielding clayey soil, 7-9 tons of grain, 55 beets.

Fairly low organic C, 1.7%. All straw plowed in. We start with an assumed mineralization of 1.5%.

Because of the high production the carbon balance is almost even and the situation stable. But the demand for straw as biofuel is increasing.

If the straw is removed the carbon balance will shift to minus 215 kg C per year. This is negative on this kind of soil (below 2% org C), and a negative trend in productivity development is introduced.

A counteracting measure: introduce a cover crop (ryegrass) between oats and barley, let it remain over winter and delay tillage till Spring. This almost restores the carbon balance to even (-36).

Now – we assumed a mineralization of 1.5%. What would happen if it is 1.8%? In that case there is negative carbon balance from the start. Removal of the straw is still more critical and the amount should be limited to one or two crops in the rotation..

If the mineralization should be 1,2%: There is a positive carbon balance in the original system (158 kg C) but removing the straw will shift it to negative. A cover crop will almost restore the situation.

Summarizing conclusion: straw can be removed but compensatory measures are needed.

2. Sugar beets, barley, potatoes, rye, peas, rye. Sandy soil, 5 tons of rye, about 50 of beets and potatoes.. High organic C, 3,9%. Straw plowed in, 3 cover crops.

Carbon balance minus 650 kg C. Note the high figure (this is an experimental site and analyses verify the trend). This high carbon loss is no problem for the farm within several decades, but it could be noted in relation to the climate discussion of today.

What can be done? One possibility: Cut away the sugar beets, replace rye with spring sown cereals, add one more cover crop. The result is a carbon balance of minus 65, not too far from stability. And the cost? There is a gain for the farmer of SEK 1000 ($ 130) per hectare and year.

Why bother about soil carbon?

Every textbook in agriculture emphazises the importance of humus, of soil carbon. Soil fertility, water relations, soil structure and soil life are mentioned. Can this be translated into hard figures for management planning valid for the agriculture of today?

The data needed are in fact how the yield potential is influenced by different levels of soil carbon. Johnston (2005) has given a summary of British work. Capriel (undated) relates results from some German experiments. There are conclusions to be drawn from Swedish experiments. In the following figure data from these sources are put together.

Figure 2.

Some aspects in addition:

High yields are more demanding on soil conditions and are more rewarding for the positive effects of soil C.

Soil C increases soil structural stability, improves water characteristics and reduces erosion. This can be of special importance for loss of phosphorus to waters.

Soil C is positive for the retention of soluble herbicides.

Some general remarks and observations.

High yields are positive for soil C development. This means that nitrogen fertilizers in general are positive as long as they increase yields.

When estimating soil C development, there may be uncertainties about the absolute levels. However, there is much less uncertainty about the effect of one certain measure relative to another. Differences induced by management practices can be determined with some confidence.

Soil carbon changes are slow and longterm. Still, they need to be taken into account.

References:

Andrén, O and Kätterer, T 1997. ICBM: the Introductory Carbon Balance Model for exploration of soil carbon balances. Ecological Applications 7(4), 1226-1236.

Berglund, Kerstin, Örjan Berglund och Anna Gustafson Bjuréus 2002. Markstrukturindex - ett sätt att bedöma jordarnas fysikaliska status och odlingssystemets inverkan på markstrukturen. SLU, Institutionen för Markvetenskap, Avdelningen för lantbrukets hydroteknik, Avdelningsmeddelande 02:04.

Borresen, T. and Njös, A. 1993. Plowing and rotary cultivation for cereal production in a long-term experiment on a clay soil in southeastern Norway. 1. Soil properties. Soil and Tillage Research 28,97-108.

Börresen, T. And Njös, A., 1994. The effect of ploughing depth and seedbed preparation on crop yields, weed infestation and soil properties from 1940 to 1990 on an loam soil in eastern Norway. Soil and Tillage Research 32, 21-39.

Capriel, P, 2007 (Internet) Humusversorgung der Böden. LfL, Inst. für Agrarökologie, Freising. http://www.lfl.bayern.de/iab/bodenbearbeitung/13479/linkurl_0_3.pdf

Carlgren, K och Mattsson, L 2001. Swedish Soil Fertility Experiments. Acta Agr. Scand., Sect. B, 2001, 51, 49-78.

Dersch, G and Böhm, K 2001. Effects of agronomic practices on the soil carbon storage potential in Austria. Nutrient Cycling in Agroecosystems, 60, 1-3, 49-55.

Dolan, M S, Clapp, C E, Allmaras, R R, Baker, J M and Molina, J A E 2006. Soil Organic Carbon and Nitrogen on a Minnesota soil as related to tillage, residue and nitrogen management. Soil & Tillage Research 89, 221-231.

Franko, U. (1997). Modellierung des Umsatzes der organischen Bodensubstanz. Archiv für

Acker- und Pflanzenbau und Bodenkunde 41, 527-547.

Jarecki, M. K. and Lal, R., 2005. Soil organic carbon sequestration rates in two long-term no-till experiments in Ohio. Soil Sci. 170,4,289-291.

Johnson, J. M-F., Allmaras, R.R. and Reicosky, D. C., 2006. Estimating source carbon from Crop Residues, Roots and Rhizodepositits Using the National Grain-Yield Database. Agron. J. 89:622-636..

Johnston, A E and Poulton, P R 2005. Soil Organic Matter: its importance in sustainable agricultural systems. Proceedings No 565, The International Fertiliser Society, York, UK.

Koch, H-J and Stockfisch, N 2006. Loss of soil organic matter upon ploughing under a loess soil after several years of conservation tillage. Soil & Tillage Research 86, 1, 73-83.

Mattsson, Lennart och Larsson, Hans 2005. Att föra bort eller bruka ner halmen påverkar mullhalt, daggmaskar och skadedjur. Inst. för Markvetenskap, Avd. för Växtnäringslära, Rapport 210.

Mattsson, Lennart 1991. Nettomineralisering och rotproduktion vid odling av några vanliga lantbruksgrödor. Inst. för Markvetenskap. Avd. för Växtnäringslära, Rapport 182.

Persson, Jan 2003. Nitrogen losses and N management. Possible improvements in agriculture. Swed. Univ. of Agric.Sciences, Dept of Soil Sciences, Div of Soil Fertility, Rapport 207. Swedish, Engl. Summary).

Puget, P. and Lal, R., 2005. Soil organic carbon and nitrogen in a Mollisol in central Ohis as affected by tillage and land use. Soil and Tillage Research, 80, Issues 1-2. 201-213.

Puget, P, Lal, R, Izzuralde, C, Post, M and Owens, L 2005. Stock and distribution of total and corn-derived organic carbon in aggregate and primary particle fractions for different land use and soil management practices. Soil Sci. 170, 4, 256-279.

Reicosky, D. C., 1999. Effect on Conservation Tillage on Soil Organic Carbon Dynamics: Field experiments in the U. S. Corn Belt. In: (editors) Stott, D. E., Mohtar, R. H. and Steinhardt, G. C., 2001. Sustaining the global farm. Selected papers from 10^th International Soil Conservation Organization Meeting, May 1999 at Purdue University.

Sauerbeck, D., R., 2001. CO2 emissions and C sequestration by agriculture – perspectives and limitations. Nutrient Cycling in Agroecosystems, 60, 253-266.

Skinner, R. H., Sanderson, M. A., Tracy, B. F. and Dell, C. J., 2006. Above- and belowground productivity and soil carbon dynamics of pasture mixes. Agr. J., 98,320-326.

Swedish Agricultural University, Dept of Soil Science. 2007. Experimental results available at Internet. Manager: Lennart Mattsson.

Slepetiene, A. and Slepetys, J., 2005. Status of humus in soil under various long-term tillage systems. Geoderma, 127, issues 3-4, 207-215.

Uhlen, G., 1991. Long-term effects of fertilizers, manure, straw and crop rotation on total N

and total C in soil. Acta Agric. Scand. 41,119-127.

West, T., O. and Marland, G., 2002. Net carbon flux from agricultural ecosystems: methodology for full carbon cycle analyses. Environmental Pollution 116, 439-444.

Wilts, A. R., Reicosky, D. C., Allmaras, R. R. and Clapp, C. E., 2004. Long Term Residue Effects. Soil Sci. Soc. of Am. J. 68:1342-1351.

Wright, A. L. and Hons, F. M., 2005. Tillage impacts on soil aggregation and carbon and nitrogen sequestration under wheat cropping systems. Soil and Tillage Research, 84, 67-75.

VDLUFA 2004. Humusbilanzierung - Standpunkt. www.vdlufa.de

References for the experimental background, Table 1 and figure 2.

Christensen, B, Petersen, J and Trentemuller, U, 2006. The Askov Longterm Experiments on Animal Manure and Mineral Fertilizers. I. The Lermarken site. DIAS report 121. Danish Institute of Agricultural Sciences.

Cultivation Systems. Hushållningssällskapet Kristianstad. Report 2007, under preparation.

www.odlingssystem.se

Cuvardic, M, Tveitnes, S, Krogstad, T and Lombnaes, P, 2004. Longterm effects of crop rotation and different fertilization systems on soil fertility and productivity. Acta Agriculturae Scandinavica, Section B Plant Soil Science, 54,4, 193-201.

Dersch, G and Böhm, K 2001. Effects of agronomic practices on the soil carbon storage potential in Austria. Nutrient Cycling in Agroecosystems, 60, 1-3, 49-55.

Mattsson, Lennart och Larsson, Hans 2005. Att föra bort eller bruka ner halmen påverkar mullhalt, daggmaskar och skadedjur. Inst. för Markvetenskap, Avd. för Växtnäringslära, Rapport 210.

Petersen, B, Berntsen, J, Hansen, S and Jensen, L, 2005. CN-SIM - a model for the turnover of soil organic matter. I. Long-term carbon and radiocarbon development. Soil Biology and

Biochemistry 37, 359-354. (Source of data for some experiments).

Slepetiene, A. and Slepetys, J., 2005. Status of humus in soil under various long-term tillage systems. Geoderma, 127, issues 3-4, 207-215.

Soil Fertility (SF) and SLU. Internet Database, Swedish Agricultural University. Manager: Lennart Mattsson.