Le Dossier de l'environnement de l'INRA n°22
D22 : INRA faced with Sustainable Development : Landmarks for the Johannesburg Conference

Agronomic Research faced with Green House Gases

The contribution of agriculture and forests
Research carried out at INRA on the greenhouse effect and climate change
Impacts of climate change on agricultural production
To conclude

Bibliographical references

In the international context linked to emission greenhouse gases (GHG), the European Union (EU) stands halfway between the highly industrialised countries and the developing countries. Indeed, all the EU States have approximately the same gross domestic product (GDP), i.e. roughly 20,000 $/inhabitant. Average emission per inhabitant is 12 tons of CO2 (carbonic gas), as compared to 23 for the United States and 21 for Canada. In the negotiations linked to the Kyoto protocol, the EU committed itself to reducing its emissions by 8% between 2008-2012.
In the EU, France is itself in an intermediate position, with an emission of roughly 6 tons/inhabitant for a total GDP of 15,000$. Therefore, given its importance in terms of area and population, France can be considered as representative and stands between the Northern countries (Germany, Netherlands, Belgium) and the Southern countries (Spain, Greece, Portugal). Under the Kyoto protocol, the internal EU agreements simply commit them to stabilising its emissions at the level of those of 1990. The six gases listed in the protocol were being emitted at the rate of 144 million tons of carbon equivalent in 1990, and should no specific measures be implemented, these emissions would reach 175 million tons by 2010: a notable effort is therefore required to reach the objective set. This led to the implementation of a national programme in 1999. It must be added that the fight against climate change was officially declared a national priority and is based on a co-ordinating structure set up in 1992, the Interministerial Mission on the Greenhouse Effect (MIES). France is currently considering creating an observatory of climate change.

[R] The contribution of agriculture and forests

In this general context, the French situation (which is representative of the general situation in Europe) is marked by a differently distributed contribution of greenhouse gases than that known at the planetary level, as shown in the table below (Germon, 2000).

Table I. Contribution of the different greenhouse gases:
1 2
CFC, HFC et dérivés

1. radiatif forcing at the planetary level (IPCC- Intergovernmental Panel on Climate Change, 1996);
2. the potential of global warming in metropolitan France in 1998
(CITEPA, 1999).

The situation is characterised by higher levels of CO2 emission, but this may be due to a different way of computing the fluorocarbon compounds (FCs) and, moreover, to the inversion of the two other main GHGs, nitrous oxide or nitrogen dioxide (N2O) that comes quite clearly in second place, whereas methane (CH4) emissions are lower. This can be mainly attributed to the important contribution of the agricultural and forestry sector.
Agriculture alone contributes 18% of French GHG emissions, although it represents less than 3% of the GDP. The forestry sector produces net emissions comprised between 0 and 1 million tons of carbon/year, according to the way changes in land use are calculated. In the agricultural and forestry sectors, N2O is the main gas emitted, with 56% of all emissions, as compared to 33% for CH4, and 11% for CO2. But the recent evolutions in the application modalities of the Kyoto protocol have reinforced the weight of this heading, given its possibilities of carbon storage. See Fig. 1 (data provided by CITEPA, 1997).


Figure 1. Relative contribution of the agriculture and forest sectors to greenhouse gas emissions (left) and relative contribution of the different gases in the agriculture and forest sector (right)

A last piece of information to complete the context (from data published by the SCEES -Service Central des Enquêtes et Etudes Statistiques, 1998): although the economic weight of agriculture is now weaker, agriculture is still dominant as regards land use. Out of 55 million hectares, 33 are allocated to farming (i.e. roughly 60%), 3 of which are not cultivated (5.4%). Woods and forests represent an equivalent of 15 million hectares (27.5%) and non-farmland represents only 7% (12.6%). The agricultural component regularly decreases (at an average of 130,000 ha/year). The land released partly goes to increase the woodland areas (80,000 ha/year). Computations give a net absorption (under the heading "change in use of soils and forests") roughly equivalent to 10% of the emissions in other sectors.
The utilised agricultural area (UAA) represents 54.5% of the French territory (i.e. 30 million hectares), and arable land makes up over half of the French UAA (18 million hectares, i.e. 61%). This proportion regularly increases at the expense of grasslands, which still represent just over 10 million hectares (i.e. 35% of the UAA). A clear drop as compared to the 13 million hectares in 1960! As for permanent crops (vines, orchards), they have also decreased and now only represent just over 1 million hectares (i.e. less than 4% of the UAA).
To sum things up, the agricultural and forestry sector amount to only 3% of the GDP, but still represents 87% of the national territory and contributes, via agriculture alone, to the emission of 18% of all GHGs over roughly 60% of the area. Considering the recent Marrakech agreements and the perspective of deducing "sinks", the extent of the area used for agriculture and forests reinforces the weight of this sector in the context of greenhouse gases and climate change. What impacts may be forecasted? What solutions can be adopted to curb emissions and increase sequestration capacities? This paper will deal with these issues and its arguments will be based on French research work in agronomy carried out in recent years at the Institut National de la Recherche Agronomique (INRA).

[R] Research carried out at INRA on the greenhouse effect and climate change

In the 1980s, research was carried out on carbon and nitrogen cycles in INRA's Bioclimatology Department (especially at Grignon and Avignon, on the impact of microclimate and local climate change) and in the Agronomy (Clermont-Ferrand and Laon) and Soil Science Departments (Avignon, Dijon, Grignon, Orléans, etc.). This research benefited from the necessary scientific and technical potential. However, it was not until the early 1990s that a group of researchers and laboratories was mobilised on this theme. The mobilisation was based on an incentive (supported by targeted finances), in the framework of the Agrotech programme implemented at the beginning of the 1990s (see INRA report, 1995). It brought together a community of twenty researchers who studied the two following complementary points:
- evaluation of the impact of CO2 doubling and global warming on agricultural and forestry production.
- quantification of the exchanges of agricultural and forest areas for all GHGs and appreciation of the efficiency of practices to reduce emissions or increase absorptions.
These two complementary points were simultaneously studied and their general philosophy is shown in the two graphs below (Figs. 2 and 3); we will not deal with forest production here, only with that of agriculture.

"Impacts of Climatic Change" section reminding 1990-2000 Works

Figure 2. Organisation of research on the impact of climate change

Figure 3. Organisation of research on the greenhouse effect

[R] Impacts of climate change on agricultural production

Over the past decade, all studies in this field were first focussed on commodities representing a large proportion of the utilised agricultural area, that is arable crops, grasslands and forage crops.
Arable crops
Research on arable crops combined ecophysiological studies on the influence of CO2 doubling on the functioning of crops and analyses based on simulation models of these crops (CERES, then STICS) developed at INRA (Brisson, 1998), simultaneously taking into consideration the effect of CO2 and the climate scenarios for the end of the century (Fig. 4).
These studies applied essentially to wheat and maize and provided estimates of the expected effects on crop yields according to various climate situations (Delécolle et al., 1995), the main points of which are presented in Table II below (Delécolle et al., 1999).

Figure 4: The different factors taken into account in the simulation of crop functioning

Table II. Simulated effects for different species and conditions
Cultivation Place Yield Water consumption
1 - Wheat


2 - Wheat
Versailles (no irrigation)
Avignon (irrigated)
Versailles (no irrigation)
Toulouse (irrigated)
Versailles (irrigated)
- 1,7
- 5,7
- 12,4
- 16,2
- 5,8

Sources : Delécolle (data non- published)

These results reveal effects situated between +10% and -15% due to the relative weights of the positive effect of carbon fertilisation (Fig. 5) and negative effect of the shortening of the plant cycle, which can be attributed to the predicted global warming. We must also add that, in any case, decreased stomatic conductance, will lead a decrease in water consumption.

Figure 5. Effects of CO2 concentration on maize conversion efficiency (adapted from results by Ruget et al, 1996)

These results provide estimates of predictable effects on arable crops. For the time being, they have not been transposed to other major crops such as sunflower, rape or sugarbeet, although this is technically feasible. First, in the present state of affairs, the models may not entirely reflect interactions between carbon fertilisation and crop functioning, in particular as regards of hydric constraints. Then, they do not incorporate the effects of other biotic constraints (weeds, diseases, insects) that must be taken into consideration in crop systems. Besides, the climate scenarios implemented were still fairly imperfect, on the one hand, as regards rainfall and thus hydric constraints, and on the other hand, in terms of spatial resolution, with a minimum grid unit size of 200 km. Due to recent progress made by climate modellers on these two points, these studies may now be resumed and predictions sharpened to be extended to other annual crops. This research does not only provide accurate data on yield perspectives at the end of the XXth century, but should also indicate directions for research into crop genetics and explore crop systems adapted to these new conditions. However, all in all, the results given above lead us to think that climate change should not pose insurmountable problems for arable crops, as their capacity to adapt quickly (a few years) has been tested in the past, following the successive orientations of the Common Agricultural Policy (CAP) at EU level. The most problematic factor will probably be that of water, if current trends revealed by the present scenarios are confirmed (Déqué, 2000), i.e. a decrease of rainfall in seasons that are already considered to be dry.
Grassland production also has been extensively investigated. As seen above, grasslands still occupy a large area even though they are regularly decreasing (the 10.5 million hectares of natural grasslands and the 2.7 million hectares of temporary or artificial grasslands make up a total of 13 million hectares, i.e. approximately the same as arable crops).
The research carried out in Clermont-Ferrand, under the responsibility of J.-F. Soussana also associated ecophysiological studies in controlled conditions and simulations based on "end-of-XXth century" scenarios.
The first allowed the evaluation in field conditions (mini-FACE) of the vegetation dynamic of permanent grasslands subjected to a 70% increase in CO2 (600 ppm) (Teyssonneyre et al., 2002 a and b). The botanical composition was significantly modified as a result and, although grasses remained dominant, the balance between species shifted towards an increase in nitrogen-fixing legumes when these grasslands were well-managed, and an increase in non-fixing dicotyledons when they were little used. These structural changes in the grassland community have consequences on both the biodiversity and pastoral value. The decline of biodiversity generally observed in grasslands subjected to frequent cutting was avoided in the case of high CO2 concentration, and the pastoral value increased in the case of a frequent cutting. In the case of infrequent cutting, the yearly production of aerial biomass significantly increased (+30% in two years); this increase was not so high (+11%) grasslands that were cut with greater frequency.
To evaluate the simultaneous effect of a scenario with CO2 doubling and +3°C global warming, an experiment was carried out in the greenhouse in Clermont-Ferrand, on a temporary ryegrass grassland grown with two levels of N inputs (Casella et al., 1996). Water was more efficient (from 17 to 30%) in the case of double CO2 input without warming, but the combination of the two effects led to a clear increase in evapotranspiration, especially in spring; during the summer months, the soil was drier over a longer period and draining was strongly reduced in winter. As a consequence, in spite of an increase in forage production (+19% and +14% respectively, according to the nitrogen input rate) in the case of doubling, the association of doubled CO2 with a +3°C simulated warming did not induce any significant change in production: the increase in spring and autumn, along with a 3-week increase of the vegetative growth, was practically annulled by the low summer production due to the warmer summer climate. As regards the feed value, composition and digestibility of the plant cell walls were barely altered, but the balance between energy value and nitrogen value of the forage was modified due to an increase in soluble sugars and a decrease in total nitrogen matter (Casella and Soussana, 1997).
These ecophysiological studies delivered detailed information, which was completed by the simulation studies based on climate scenarios that stem from the ARPÈGE model developed by Météo-France, the French International Weather Office (Déqué, 2000). The simulations were performed on the basis of an 80 km grid size unit and applied to three sites in the Massif Central, using the grassland ecosystem simulation model, PASIM. The main results are (Soussana et al., 2001): on a yearly average, greatly positive effects in these high-rainfall areas (1080 to 1300 mm) in the case of management via cutting, with increased photosynthesis, aerial biomass production and protein content. These would respectively reach 37%, 25% and 11% (of which 25%, 18% and 5% could be attributed to CO2 doubling - according to the model -, which matches the experimental results mentioned above). This positive effect may have little relation with the management modes; indeed the different combinations of cutting frequency, nitrogen fertilisation and grazing modes tested at one particular site (Allier) only caused production to increase from 18 to 25%. In the case of grazing land, positive effects are lower (a maximum of 7% under high stocking rates) due to the constraints linked to grazing.
To get a fuller picture at the livestock system level, these studies were extended to adaptation conditions in these new situations: if the current manner of using grasslands were to induce an increase in harvested hay (from 9 to 34% according to the different management styles), there would be little effect on animal production. If on the other hand the stocking rate was increased by 20% or if the grazing season was extended by three weeks, then intake and meat yield would rise by 7 to 20% for the first and 2 to 20% for the second.
As is currently the case for arable crops, it seems difficult, however, to proceed any further for the time being, on the one hand, because of the uncertainty of the climate scenarios, especially concerning rainfall, and on the other hand, owing to the limitations of the grassland model. These limitations relate to the description of the grassland functioning (there is no within-year modulation of production and thus dry summer spells are not taken into consideration) and and second, to the grazing management modalities (constant stocking rates throughout the season, results at the field, and not the farm scale). Besides, and as is the case for annual crops, the matter of hydric constraints is still unresolved.

Figure 6. Effects of global warming on first cutting date at mid-elevation in the Alps (June, 2001)

In this respect, even if hydric constraints remained moderate in our study area in the Massif Central, their effect would be far greater in the Mediterranean region. And if global warming is to be reflected in much earlier cutting dates, in the mid-elevation alpine areas for example (see Fig. 6), as simulated in the climate scenarios (those developed at the LMD - Laboratoire de Météorologie Dynamique - by Y. Polcher, see Crossley et al., 2000), it is clear that the effects on fodder production in these regions would first depend on the rainfall level. Advances in rainfall prediction based on climate models will there-fore be determinant.
Perennial crops (orchards and vineyards)
Although they only make up a small part of the agricultural area these crops are nevertheless economically important (especially vineyards). Moreover, they raise the question of impacts, not only in terms of quantity (what influence do impacts have on the yield of the crop?), but also in terms of quality (once again, especially in the case of vineyards).
Research on these crops has only recently started. Contrary to the above cases, they have not been the subject of ecophysiological studies or of simulation models, although a STICS-vigne version is currently being developed and studied using climate scenarios.
Phenology was first taken into consideration in this approach, as it plays a major role in these crops and is directly linked to air temperature, and is thus immediately affected by global warming. The first results clearly show significant earlierness of the various phenological stages, as can be seen in the following figures (Figs. 7 and 8) for orchards and vineyards that are fairly representative of the different cases observed.

Figure 7. Évolution of aplle tree flowering dates in south-eastern France (Balandran) over the 1973-2000 period
a : recorded dates; b : dates simulated using a phenological model.
(Domergue, 2001).

Figure 8. Évolution of dates of grape harvest start at Chateauneuf-du-Pape
(Ganichot, 2002)

In these two examples, there appears a clear change in phenological earliness at the different stages (approximately two weeks over a period of 30 years for flowering in apple trees, to three or four weeks over the last 50 years for the grape harvesting date). These are preliminary results, as the respective roles of changes in cultural techniques and climate still need to be worked out, at least in the case of the grape harvesting date. In any case, the results are certain for the flowering of apple trees, as the evolution observed has been confirmed with the help of a conventional phenological module (Fig. 7b).
It can therefore be said that the temperature rise in the 1990-2000 decade, which has been confirmed by climate data (not given here), has already had an impact on the phenological stages: these have been noticeably brought forward. This may be considered to reflect quite clearly the phenomenon of global warming. The results are of course only partial and, for the time being, restricted to south-eastern France. Studies are currently being carried out to gather further phenological data across the French territory and at INRA's observation sites, as well as in the different experimental stations managed by the fruit producers and winegrowers. Thus the consequences of different warming scenarios could be simulated just by considering the plant phenology (Figs. 9 and 10).

Figure 9. Estimates of flowering dates for three fruit trees (apple, apricot and peach) at two sites (South and North of the Rhone valley) based on weather data for the 1965-1989 years, 1990-2000 and two global warming scenarios 2 x CO2 (conventional) and 2 x CO2 (with variability) (Domergue, 2001).

Figure 10. Projected evolution of vine phenology (Syrah variety) in the Montpellier region
(Lebon, 2002)

The first example shows that the flowering date of different species did not occur much earlier at the most southern sites (owing to the antagonist play of cold and heat requirements), whereas it responded to warming at the northern site. Besides, contrary to what could be expected from the rise in night temperatures, frost damage has notably increased in some cases, as a result of the earlier flowering date. Beyond these preliminary results which need confirming, the conclusion is that global warming has probably already had a noticeable effect over the last 10 years and will involve complex mechanisms (alternate cold and heat requirements) that must be taken into consideration in plant variety selection. It may also lead to reconsider the geographic distribution of production areas. Besides, frost risks will never disappear, quite the opposite. To overcome the problem, climate models, which up to now have only considered average values, will need to be more efficient regarding threshold values.
As regards vineyards, the research carried out in the South of France by Lebon (2002) shows that warming has a marked effect on production (the effect of cold requirements is negligible), although its consequences are still difficult to assess. If, for instance, the ripening were to be considerably advanced, the maturation stage would occur during the hottest period of the year, with especially warm nights (minima >18°C). Even during hot periods, nights are at present temperate (between 14 and 18°C) and these cooler temperatures are a quality-building factor in the Mediterranean area. Obviously, this is only one aspect and more detailed analyses will be needed to produce a reliable diagnosis. It is also obvious that the effects of climate change will be quite different in northern vineyards. However, we can already consider that these changes will be significant for the geographic distribution of types of production (and even for the choice of varieties) and that traditional terroirs will have to adapt to this probable new climate situation.
Current research avenues
These examples of perennial crops which act as phenological "indicators", lead us to think that global warming is already taking place, or at least that the climate of the past decade comprises certain particularities which will persist or even increase. Beyond the studies carried out on the "end-of-XXth-century" scenarios, this induces us to prioritise a detailed analysis of the recent past and present, so as to deliver answers by 2020-2030 (and not only by 2070-2100).
Although future research should not neglect arable crops and grasslands (especially regarding plant breeding, weeds and pests, that have been little dealt with up to now) their capacity for rapid adaptation should make us focus in priority on perennial crops, the adaptation time which have a longer adaptation time of between 10 and 20 years.
In this area (but also, at a lesser level, in that of other commodities), we need to obtain from the climate modellers scenarios that are more reliable than presently existing ones regarding factors other than temperature (rainfall, total radiation), more detailed regional information (for 50 km-sized grid units) and more detailed indications than average values (variability, threshold-values).
Finally, it will be necessary to deal with both quantity and quality aspects and include environmental effects in the impacts to be taken into consideration, as these two components take prominent place in any research on production in contemporary European agriculture. It must be added that the environmental effects of these crops also play a role in the greenhouse effect, via the soil or exchanges with the atmosphere whose retroaction must be taken into consideration: climate change/greenhouse effect. leads us to the studies undertaken on the greenhouse effect.
Agriculture and greenhouse gases: a summary
The teams working for INRA already had in-depth knowledge of, and were able to control, in terms of experimentation and modelling, CO2 exchanges between the soil or a plant cover and the atmosphere (bioclimatology). They were also familiar with the denitrification processes occurring in the soil, mainly applied at the time to the problem of nitrate rejection in the soil or the evolution of organic matter in soil physics. These programmes had to be redirected to a large extent to address the issue of the contribution of agriculture to the greenhouse effect.
The Agrotech programme paved the way for another programme, AGRIGES, which lasted from 1994 and 1998. Its role was to stimulate and co-ordinate these studies (INRA, 1995; Académie d'Agriculture, 1999). The programme was funded by several institutions and achieved the results described hereafter (work on carbon fluxes and storage in forests is not dealt with in this paper).
Carbon storage in soils
Earlier studies by Arrouays and Pélissier (1994) produced a detailed description of carbon stock evolution over a 30-year period in a transition situation from forest to crop (maize) on the loamy soils of the Pyrenean foothills (Fig. 11).

Figure 11. Evolution of carbon stocks in a transition situation from forest to agriculture
(Arrouays et Pélissier, 1994).

One aim of the AGRIGES programme was to work on a wider scale (that of the French territory) by calculating stocks at a depth of 0-30 cm. To do so, the characteristics of the different soil types and land use types were crossed for each of these variables by using the data available in the various soil maps of the French territory (Fig. 12).

Figure 12. Carbon stocks and land use
(Arrouays et al., 1999).
Land use and soil C stocks (t/ha sur 0-30 cm)

The resulting map (Fig. 13) both makes it possible to assess the geographical distribution of these stocks and to propose a first estimate (3 billion tons of carbon for total France for a depth of 0-30 cm).

Figure 13. Geographical distribution of organic C in French soils (Arrouays, 1999).

Compared with world-scale evaluations (1500 billion tons), the French metropolitan territory represents 1/500 of all world stocks, but one should add that these stocks are estimated across a one meter depth and that, on this basis, the French situation is closer to 1/400 to 1/300. From another standpoint, this stock can be associated with the French carbon emissions, estimated to be 100 million tons, i.e. approximately 3% of the stock.

Figure 14. Estimates of carbon storage in French soils since 1850 (with enveloppes corresponding to extreme assumptions: de stock à l'équilibre sous forêts ou surfaces toujours en herbe)
(Balesdent et Arrouays, 1999).

To highlight the evolution of the C stock in relation to land use, Arrouays and Balesdent's studies (1999) combined historical data with an estimation of storage for each type of land use with a two-compartment model of carbon dynamic. Figure 14 sums up the main points of this evolution by showing a continuous storage since 1850 due to a decrease in cultivated areas. Although the latter have expanded since 1970 at the expense of grasslands, the balance remains positive as both grasslands and forests planted earlier are now moving towards maturity and continue to accumulate carbon. The introduction of bare set-aside, imposed by the CAP around 1990, led to low storage at the end of the period, although the current storage potential is thought to lie around 2 to 4% of French emissions (for a scenario maximum involving the conversion of 3 million hectares of crops and 0.8 million hectares of current set-aside land into grasslands and forests). This estimation should, however, be modulated by taking into consideration the temperature increase. The rise in temperature is thought to have resulted in an 80 million ton "destocking" since 1975, due to its effect on organic matter mineralisation as considered in this model, and could alone annihilate all or part of the stored carbon! It must also be added that this study does not take into consideration the productivity increases in farming and forestry systems, for lack of available data.
Much remains to be done to evaluate these scenarios accurately. Once developed, the MORGANE model (Arrouays et al., 1999) will be an essential tool to refine present estimates, and complement on a finer scale available data on evaluation of the effects of practices (Fig. 15).

Figure 15. Évaluation of the storage potential of different practices
(Balesdent, 1995).

It must also be noted that research on the use of crops as biofuel may also induce a potential storage of carbon. This is being currently investigated.
Emissions of nitrous oxide and methane
These two gases have also been investigated by several INRA teams, but I will deal with them more briefly than carbon storage which has taken on increasing importance since the recent Marrakech agreements and is the object of reflection on the actions to be implemented at the political level.
As regards N2O, these studies have combined detailed modelling of denitrification and nitrification processes that result from the soil microbial activity (Germon et al., 1999) and on-site measures via micrometeorological methods or gas exchange chambers (Laville et al., 1999). They enabled the re-evaluation of regional-level emissions estimated by the models (Renault, 1999) or by the GIEC/IPCC, based on emission factors (Cellier et Laville, 1999). Although these evaluation methods deliver generally reliable results at the scale of the territory (with an approximate value of 1 kg/ha of nitrogen + (1.25 + 1%) fertiliser inputs), the results show a strong soil effect and an influence of mineral nitrogen availability that could be more than proportional. To further work out means of action (other than cutting down fertiliser inputs, the only measure considered by the emission control programme, in connection with measures to reduce nitrate pollution of water), the specificities of the natural environment (soil, climate) and the impact of cultural techniques at the regional level, will need to be taken into consideration. Such is the orientation of research in this field. It is based on the evaluation of emission factors adjusted to local characteristics and more generally on the use of the mechanistic models currently being developed (Cellier and Laville, 1999). In terms of agronomic practices locally and beyond the recommendation to strictly adjust fertiliser inputs to the needs of crops, suggestions can already be made about limiting mineral nitrogen availability, on the one hand, and maintaining a soil structural state favourable to aeration in order to limit anoxy periods and emissions attributable to denitrification, on the other hand (Germon et al., 1999).
As regards methane, a gross estimate of the methane balance in French soils proposed by Roger et al. (1999), showed that although cultivated soils, and soils in general, constitute sinks (at the respective levels of 274 and 236 tons/day), they are of an order of magnitude inferior to emissions from livestock farming (4500 t/day) and public tips (2250 t/day). This justifies the fact that INRA research is mainly focussed on emissions linked to livestock farming. Beyond estimations at the national level that could be transposed on the regional scale, these studies revealed major differences (from 1 to 60) depending on the cattle population (Sauvant et al., 1999). The studies produced a detailed analysis of the methanogenesis mechanisms and their nutritional importance. This enabled researchers to analyse the variation factors among and within species, as well as the type of diet.
Suggestions may be made to reduce emissions at the individual level (in terms of diet and food supplements), yet the margin of manoeuvre seems relatively narrow, especially given that more detailed research into the impact of these measures would be needed (excluding, of course, all use of animal flour and antibiotics). Reducing animal performance is the only way to significantly decrease emissions. But on the other hand the production of methane, expressed per kg of milk, is notably decreased in high-producing cows! There is thus no clearly marked course in this field, except by diminishing the animal population, which would of course raise other problems!
To complete this overview, we should mention ongoing research at the Rennes Centre of INRA. Its objective is to reduce emissions from livestock farming buildings, especially by using litter. However, it is still too early to give any results. These will ultimately complement CEMAGREF (Centre d'Etude du Machinisme Agricole, des Eaux et des Forêts) results on the treatment and spreading of animal excrement (Martinez et al., 1999), given that this sector will probably represent an equivalent of 0.9 M tons of carbon/year in 2010.

[R] To conclude

Having completed this overview, we can now go back over the stakes evoked in the introduction, first regarding the greenhouse effect:
At the level of France, a recent official document published in March 2002 provides an overall view of the situation (Troisième communication nationale à la Convention-Cadre des Nations Unies sur les changements climatiques, on-line on the MIES Internet site www@effet-de-serre.gouv.fr). After briefly recalling the danger of a decrease in soil water reserves for agricultural and forest production and the results mentioned above for grasslands, the document states that France has not yet defined a specific agenda on adaptation to climate change, except for the laws on regional development and protection of the environment. As for GHGs, after mentioning ongoing studies on electricity production via the biomass and ongoing experiments on an industrial scale for the production and distribution of biofuel (liable to decrease CO2 emission by 1Mt/year), the text indicates that the 1977 national plan to fight climate change (PNLCC) contained only a few voluntary actions to control emissions in the agricultural sector, but described the impact of certain CAP evolutions on these emissions. At present, beyond specific measures in the forestry sector, the text underlines the need for action in the field of animal excrement (liable to produce approximately 3 M tons CO2 per year, in 2010), as well as the need to limit nitrogen inputs by taxing them in the framework of the law on water being currently discussed, and to improve our knowledge with the help of research (on biomass, enteric fermentation of ruminants, N2O emissions from soils, as well as carbon storage in the same soils). Finally, agriculture is directly concerned by the objective of converting 30,000 hectares/year (now 10,000 hectares/year due to the 1999 storm) of farming land into forests.
At the European level, there is no common system and the elaboration of the new CAP in 2003, already mentioned for France, will noticeably condition the relationships between agriculture and the environment.

The same remark can beThe same remark can be made about the impacts of climate change. However, although it seems necessary to immediately take these impacts into consideration for forests (for one, because of their long response time, but also because the forestry studies carried out by INRA show a notable productivity increase (by approximately 30%) and we do not really know the respective contribution of carbon dioxide increase, global warming and nitrogen fertilisers in rain), more detailed knowledge is needed on this point, on the one hand concerning the climate scenarios, and on the other, the analysis of recent past and possible evolutions. In this respect, the priority given to climate change is clearly declared in the orientations of the 6th PCRD (Common Plan for Research and Development) being currently launched, a good guarantee of progress at the scale of Europe.

Bernard Seguin is currently responsible for the "Climate and Greenhouse Effect Mission" in INRA.
This article is taken from the "Courrier de l'environnement de l'INRA, n°46", by B. Seguin.
Translated from French by Nicole Scott.

[R] Bibliographical references

Academie d'agriculture. 1999. Compte-rendu du colloque " Bilan et gestion des gaz à effet de serre dans l'espace rural ". Paris. 18 et 19 mai 1999. C.R. Acad. Agric. FR. No spécial 85, no 6 : 1-392.
Arrouays D. et P. Pelissier. 1994. Changes in carbon storage in temperate humic loamy soils after forest clearing and continuous corn cropping in France. Plant and Soil. 160, 215-223.
Arrouays, D. et autres. 1999. Stocks de carbone dans les sols de France : quelles estimations ? C.R. Acad. Agric. Fr. 85, 278-292.
Balesdent, J. 1995. Stockage du carbone dans les sols en fonction de leur utilisation.In A. Perrier: Agriculture et gaz à effet de serre.dossiers de l'environnement, no 10, INRA éditeur. 39-48.
Brisson, N. et al. 1998. SICS : a generic model for the simulation of crops and their water and nitrogen balance. I . Theory and parametrization applied to wheat and corn, Agronomie. 5-6, 311-346.
CITEPA. 1997 et 1999. Site Web : www.citepa.org.
Casella, E, J.F. Soussana et P. Loiseau. 1996. Long-term effect of CO2 enrichment and temperature increase on a temperate grassland. I. Productivity and water use, Plant and Soil. 182, 83-99.
Casella, E. et J.F. Soussana. 1997. Long-term effect of CO enrichment and temperature increase on the carbon balance of a temperate grasssward, Journal of Experimental Botany, 48, 1309-1321.
Cellier, P. et P. Laville. Les émissions de N2O par les sols agricoles en France : quelle évaluation possible ? C.R. Acad. Agric. Fr. (déjà cité). 177-192.
Crossley, J.F. et al. 2000. Uncertainties linked to land processes in climate change simulations. Climate Dynamics. 16, 949-961.
Delecolle, R. et al. 1995. Possible effects of climate change on wheat and maize crops in France. Climate change and agiculture : analysis of potential international impacts. ASA special publication, no 59. Madison Wisconsin, ASA éditeur. 241-258.
Delecolle, R., J.F. Soussana et J.P. Legros. 1999. Impacts attendus des changements climatiques sur l'agriculture française. C.R. Acad. Agric. Fr. 85, 45-51.
Delecolle, R., P.A. Jayetet J.F. Soussana. 2000. Agriculture française et effet de serre :quelques éléments de réflexion. Impacts potentiels du changement climatique en France au XXIe siècle, Paris, MIES éditeur. 74-80.
Deque, M. 2000. Modélisation numérique des changements climatiques. Impacts potentiels duchangement climatique en France au XXIe siècle (déjà cité). 22-45.
Domergue, M. 2001. Impact du réchauffement climatique sur le parcours phénologique d'espèces/variétés dans la vallée du Rhône. Mémoire ESITPA, travail effectué à l'unité CSE de l'INRA d'Avignon. 72 pages.
Ganichot, B. 2002. Evolution de la date des vendanges dans les Côtes-du-Rhône méridionales C.R. des 6es Rencontres Rhodaniennes, Orange. Institut Rhodanien éditeur. 38-41.
Germon, J.C. et al. 1999. Mécanismes de production, régulation et possibilités de limitation des émissions de N2O à l'échelle agronomique. C.R. Acad. Agric.Fr.(déjà cité). 148-162.
Germon, J.C. et al. 1998. Nitrous oxide emission under different soils and land management conditions. Biol. Fert.Soils. 26, 199-207.
Henault, C. et al 1998. Influence of different agricultural practices (type of soils,form of N-fertilizer) on soil nitrous oxide emissions. Biol. Fert. Soils. 27, 299-306.
Henault, C. et J.C. Germon. 1995. Quantification de la dénitrification et des émissions de protoxyde d'azote par les sols. Agronomie. 15, 321-355.
INRA. 1995. Ecosystèmes et changements globaux. Dossiers de l'environnement, no 8, 296 pages.
IPCC. Climate change 1995, the science of climate change. Contribution of Working group 1 Cambridge University Press, Cambridge UK. 572 pages.
JUIN, S. 2001. Influence du réchauffement climatique sur la répartition géographique et le calendrier de production de trois systèmes fourragers. Mémoire de 3e année de l'ENSA Montpellier, travail effectué à l'unité CES de l'INRA d'Avignon. 38 pages plus annexes.
Laville, P. et al 1999. Measurement of nitrous oxide fluxes in a fertilised maize crop by micrometeorological and chamber methods. Agric. For. Meteorol. 96 ,19-38.
Lebon, E. 2002. Changements climatiques : quelles conséquences pour la viticulture. C.R des 6e Rencontres Rhodaniennes, déjà cité. 31-36.
Martinez , J. et al1999. Emissions de méthane et de protoxydes d'azote au cours du stockage, du traitement et de l'épandage des déjections animales. C.R. Acad. Agric. Fr. 87-101.
Renault, P. 1999. Les modèles opérationnels d'émission de N2O par les sols aux échelles régionales. C.R. Acad. Agric. Fr. 163-176.
Roger, P. et al. 1999. L'émission et la consommation de méthane par les sols : mécanismes, bilan, contrôle. C.R. Acad. Agric. Fr. 193-210.
Ruget, F., O. Bethenod et L. Combe. 1996. Repercussions of increased atmospheric CO2 on maize morphogenesis and growth for various temperature and radiation levels. Maydica. 41, 181-191.
Sauvant, D. et al. 1999. Production de CH4 par les ruminants : analyse des processus,quantification et modélisation, spatialisation et bilans, possibilités de réduction des émissions. C.R. Acad. Agr. Fr. 70-86.
SCEES. 1998. L'agriculture, la forêt et les industries agro-alimentaires. Collection Graph. Agri., Ministère de l'Agriculture et de la Pêche éditeur, Paris.156 pages.
Soussana, J.F. 2001. Changement climatique .Impacts potentiels sur l'agriculture et adaptations possibles. DEMETER, Armand Colin éditeur, Paris. 195-222.
Soussana, J.F. et al. Impacts régionaux de changements climatiques sur les prairies et sur la filière viande dans les régions d'élevage peu intensif. Rapport final du programme REGCLIM, Ministère de l'Agriculture.
Teyssonneyre, F. et al. 2002. Effects of elevated CO2 and cutting frequency on plant community structure. New Phytologist, (sous presse).