Most ecosystems, under constant conditions, eventually approach a steady-state C stock that is dictated by management, climate, and soil properties. But changes imposed on the ecosystem can alter the balance of C inputs and losses, shifting the ecosystem, eventually, to a new steady state (Paustian et al., 1997c). For example, after conversion of forests or grasslands to arable agriculture, losses of C often exceed inputs temporarily, resulting in a net loss of C to the atmosphere until a new, lower equilibrium level is reached (Balesdent et al., 1998; Huggins et al., 1998; Solomon et al., 2000). At least a portion of C lost, however, can often be recovered by adopting management practices that again favour higher C stocks (Cole et al., 1997). The accumulation of C in soil can continue until a new steady state is reached, often after several or more decades. Most of the additional C is stored in the soil as organic matter. Apart from agroforests, agricultural lands store very little carbon in plant biomass (Table 4.1).
There are two general ways of increasing C stocks in agricultural lands: by changing management within a given land use (e.g., cropland, rice land, grazing land, or agroforests) or by changing from one land use to another (e.g., cropland to grassland or cropland to forest) (Sampson et al., 2000). In this section, we review briefly the possible ways of increasing C stocks in agricultural lands, first within a land use and then by a change in land use. We then review recent estimates of the potential for increasing C stocks in agricultural lands globally. A more detailed assessment of management practices and corresponding rates of C accrual is reported in the IPCC Special Report on LULUCF (IPCC, 2000a).
Croplands, as referred to here, are lands devoted, at least periodically, to the production of arable crops (wetland rice, because of its unique features, is discussed separately). Soil C in these lands can often be preserved or enhanced by using farming systems with reduced tillage intensity, thus slowing the rate at which soil organic matter decomposes (Bajracharya et al., 1997; Feller and Beare, 1997; Rasmussen and Albrecht, 1997; Dick et al., 1998). Another way to promote higher soil C is to increase crop yields. This can be done by applying organic amendments, by effective use of fertilizers, by using improved crop varieties, or by irrigating. These practices help replenish soil organic matter by increasing the amount of crop residues returned to the soil (Raun et al., 1998; Huggins et al., 1998; Paustian et al., 1997b; Lal et al., 1998; Smith et al., 1997; Fernandes et al. 1997; Izac 1997). Further, soil C can often be increased by using practices that extend the duration of C fixation by photosynthesis; for example, cover crops, perennial forages in rotation, and avoiding bare fallow tend to increase organic C returns to soil (Lal et al., 1997; Singh et al., 1997a; Smith et al., 1997; Carter et al. 1998; Tiessen et al., 1998; Tian et al., 1999; Paustian et al., 1997a, 2000). Farming techniques that reduce erosion (e.g., terracing, windbreaks, and residue management) maintain productivity and also prevent loss of C from agricultural soils. The net effect of soil erosion on atmospheric CO2 is still uncertain, however, because the C removed may be deposited elsewhere and at least partially stabilized (van Noordwijk et al., 1997; Lal et al., 1998; Stallard, 1998).
Rice land, as the term is used here, refers to areas that are at least periodically flooded for wetland rice production. Carbon stocks in these systems can be preserved or enhanced by the addition of organic amendments (Singh et al., 1997b; Kumar et al., 1999) and nutrient management (Yadav et al., 1998). Rice lands, however, are an important source of methane and, from the standpoint of overall radiative forcing, management effects on methane emissions may be more important than effects on C storage (Greenland, 1995; Sampson et al., 2000). Methane emissions can be suppressed to some extent by soil amendments, altered tillage practices, water management, crop rotation, and cultivar selection (Minami, 1995; Kern et al., 1997; Neue, 1997; Yagi et al., 1997; Van der Gon, 2000). For more information on CH4 and N2O emissions from land use, see Section 3.6.
Grazing lands refer to natural grasslands, intensively managed pastures, savannas, and shrublands used, at least periodically, to graze livestock. One way to increase C stocks in these lands is to introduce new plant species. For example, the introduction of N-fixing legumes increases productivity, thereby favouring C storage (Fisher et al., 1997; Conant et al., 2001). Large increases in soil C have been also reported from the introduction of deep-rooted grasses in South American savannas (e.g., Fisher et al., 1994), though the area over which these findings apply is still uncertain (Davidson et al., 1995). Other management practices that can affect C storage include: changing grazing intensity and frequency (Manley et al., 1995; Ash et al., 1996; Burke et al., 1997, 1998); adding nutrients, especially phosphorus (Barrett and Gifford, 1999); controlling fire (Burke et al., 1997; Kauffman et al., 1998); and irrigation (Conant et al., 2001).
Agroforests include trees on farms as part of the agricultural landscape (Sampson et al., 2000). Unlike most other agricultural systems, agroforests store C in the above and below ground vegetation as well as in soil organic matter (Fernandes et al., 1997; Woomer et al., 1997). Examples of practices that can enhance C stocks include: integrated pest management, optimum tree densities, superior tree or crop cultivars, and better nutrient management (Sampson et al., 2000).
Land-use conversion involves transferring a given land area from one use to another. Where the shift is to a land use with higher potential C storage, the conversion can result in increased C stocks. For example, conversion of cropland to grassland often increases soil C (e.g., Paustian et al., 1997b; Reeder et al., 1998; Potter et al., 1999; Post and Kwon, 2000). Carbon stocks may also be enhanced by conversion of cropland to forests (reforestation, afforestation) or to agroforests (e.g., Fernandes et al., 1997; Woomer et al., 1997; Falloon et al., 1998; Post and Kwon, 2000). In some cases, cultivated lands can be restored as wetlands (Paustian et al., 1998; Lal et al., 1999), resulting in carbon gains, though this practice may also result in higher net CH4 emissions (Willison et al., 1998; Batjes, 1999; Sampson et al., 2000).
Another form of land-use conversion is the rehabilitation of severely degraded lands. Severely degraded lands are those where previous management has caused a drastic decline or disruption of productivity. Large areas of degraded lands occur on lands previously used for agriculture; lands abandoned after excessive erosion, over-grazing, desertification, or salinization (Oldeman, 1994; Lal and Bruce, 1999). Often the degradation was caused by social and economic pressures, and land rehabilitation may depend on the amelioration of the underlying causes of degradation. Specific rehabilitation practices include: introduction of new species (e.g., reforestation), addition of nutrients, and organic amendments (e.g., Lal and Bruce, 1999; Lal et al., 1998; Izaurralde et al., 1997).
Various attempts have been made to estimate potential C storage by improved management of agricultural lands. In the IPCC Second Assessment Report, Cole et al. (1996) estimated the potential for C storage in agricultural soils from improved management of existing croplands, restoration of degraded lands, and conversion to grass or forestlands. By assuming that one-half to two-thirds of the estimated historic C loss from cultivated soils could be recovered in 50 years, they proposed potential soil C increases of about 0.4 to 0.6GtC/yr from better management of existing agricultural soils. According to their estimates, additional C could be stored by set-aside of surplus upland soils (0.015 to 0.03GtC/yr), restoration of wetlands (0.006 to 0.012GtC/yr), and restoration of degraded lands (0.024 to 0.24GtC/yr), yielding a combined potential of about 0.44 to 0.88GtC/yr over a 50-year period. Later studies have provided similar estimates. Lal and Bruce (1999), using rates of soil C gain from the literature, estimated global C storage potentials of 0.43 to 0.57GtC/yr in the next 20-50 years, from erosion control, soil restoration, conservation tillage and residue management, and improved cropping practices. Batjes (1999), based partly on C gains estimated by Bruce et al. (1999), proposed that an additional 14GtC (±7) could be stored in agricultural soils over the next 25 years by improved management of degraded and stable agricultural lands. Including extensive grasslands and regrowth forests increased the estimate to 20GtC (±10), corresponding to an average rate of 0.58 to 0.80GtC/yr.
Sampson et al. (2000) recently completed a comprehensive assessment of potential net C storage from land management as part of the IPCC Special Report on LULUCF (IPCC, 2000a). According to their estimate, improved management within a land use could result in global rates of C gain, in 2010, of 0.125GtC/yr for cropland, <0.008GtC/yr for rice paddies, 0.026GtC/yr for agroforestry, and 0.237GtC/yr for grazing land. Potential rates of C gain in 2010 for land use conversion were 0.391 GtC/yr for conversion of unproductive cropland and grasslands to agroforests, <0.004GtC/yr for restoring severely degraded land, 0.038GtC/yr for conversion of cropland to grassland, and 0.004GtC/yr for conversion of drained land back to wetland. Corresponding rates of potential C gains for 2040 were consistently higher than those for 2010, often by a factor of about 2, though confidence in these values was lower. Sampson et al. (2000) cautioned that their estimates are approximations, based on interpretation of available data and that, for some estimates of potential carbon storage, the uncertainty may be as high as ±50%.
Most of these estimates assume widespread, concerted adoption of C-conserving practices, and all have high uncertainty, stemming in part from the difficulty of predicting adoption of C-conserving practices. The various estimates, furthermore, cannot always be compared directly because of differences in practices, scope, time-frame, and underlying assumptions. Most of the more recent estimates, however, are within the same order of magnitude as those presented in the SAR (Cole et al., 1996).
Increases in soil carbon content in response to improved practices cannot continue indefinitely. Eventually, soil C storage will approach a new equilibrium where C gains equal C losses (Paustian et al., 2000). This new equilibrium will depend on the management practices adopted, as well as on soil type and climatic conditions. Consequently, rates of C gain will diminish with time, and estimates for a given year cannot be extrapolated far into the future.
Once soils reach a new equilibrium, there is little further accumulation of C. And if the C-conserving practice is discontinued (e.g., reversion from no-tillage to intensive tillage), much of the previously gained C may be lost back to the atmosphere as CO2 (Dick et al., 1998; Stockfisch et al., 1999). Consequently, the C stocks stored in soils are not necessarily permanent and irreversible.
Other reports in this collection