Fire management entails changing burning regimes to alter carbon pools in the landscape.
Use and Potential
Fire often is an essential tool in pastoral lands for controlling woody weeds, removing dead biomass, stimulating regrowth, hunting, controlling pests, and clearing land. In many areas, fire regimes are strongly influenced by human actions such as controlled burning, back-burning, firebreaks, and rapid response. Reduced fire frequency or fire prevention tends to increase mean soil, biomass, and litter carbon levels (Jones et al., 1990). In particular, fire management increases the density of woody species in many landscapes (Archer, 1994; Archer et al., 1995; Scholes and Archer, 1997).
The magnitude of carbon storage associated with woody growth can be large. In the Orinoco Llanos, for example, protection from fires for 25 years increased the total carbon pool by a mean of 1.4 t C ha-1 yr-1; carbon stocks increased by 5.6 and 29 t C ha-1 in the vegetation and soil pools, respectively. If open forest is allowed to form, up to 5.69 Gt C may accumulate over 51 years, if extrapolated over the full area of the Orinoco Plains (28 Mha) (San Jose et al., 1998). Similarly, in 60 Mha of savanna lands in northeastern Australia, Burrows et al. (1998, 2000) report that management and environmental changes (predominantly decreased fire frequency) are increasing carbon pools by 30 Mt yr-1 in aboveground woody biomass and 10 Mt C yr-1 in below-ground woody biomass; aboveground biomass pools could increase 2-5 t C ha-1 (open grassland) to 15-75 t C ha-1 (closed woodland), and similar changes in below-ground biomass stocks can be expected (Burrows et al., 1998). Similar changes have been found in Africa (northern Guinea savanna): Protection from burning for 26 years resulted in large increases in tree density and basal area compared with burned plots, as well as increases in soil carbon concentrations (Brookman-Amissah et al., 1980). Scholes and Hall (1996) suggest that increased tree cover in savannas could be contributing a worldwide sink of 2 Gt C yr-1. This potential could be limited in parts of Africa and South America, however, because of population pressures on land use and in other regions because of changes toward more sustainable stocking rates that will increase burning opportunities for woody vegetation management (Hall et al., 1998).
Optimizing fire timing may increase biomass in some systems while increasing productivity. Conventional spring and autumn burning in some perennial grasslands, for example, have a long-term negative effect on live biomass and standing crop; midsummer burns facilitate more effective recovery of the grasses (Cox and Morton, 1986). In some ecosystems, burning has little effect on aboveground or below-ground biomass (e.g., Senthilkumar et al., 1998).
Charcoal generated by fires can constitute 8 g C kg-1 soil and represent up to 30 percent of the soil carbon content of some Australian soils (Skjemstad et al., 1996); it probably constitutes a significant part of the inert or passive soil carbon pool. Reduction or removal of fire will result over long periods (several centuries) in reduction of this pool.
Direct, repeated measurements of basal area increase in woody species can be made cost-effectively to high levels of accuracy (e.g., Back et al., 1997; Brown, 1997; Vine et al., 1999); these measurements can be combined with allometric equations for the species involved (e.g., Burrows et al., 2000) to calculate aboveground and below-ground biomass carbon change. Associated soil carbon sampling (see Chapter 2) can be carried out to calculate total system carbon change. The results can be scaled up to regional levels by using statistical sampling methods (Austin and Heyligers, 1989) or via remote sensing (Danaher et al., 1998). For areas where woody components are not a large part of the carbon fluxes, sampling regimes described in Fact Sheet 4.6 can be used.
Sampling for charcoal pools is a unique feature related to this activity. The slow rate of change of these relatively inert pools create uncertainty about whether including this pool is appropriate to the short time frame of the Kyoto Protocol. Furthermore, analysis is likely to be expensive. The size of the pool and the relatively poorly known dynamics, however, suggest that research is needed to determine the significance of this pool under reduced-fire regimes.
Current Knowledge and Scientific Uncertainties
There is copious documentation of the increase in woody density in savannas and other woodlands (e.g., Archer, 1994; Archer et al., 1995), supported by analyses such as soil C13/C12 profiles (e.g., Boutton et al., 1998; Burrows et al., 1998). Management appears to be a more significant factor in these changes than environmental factors such as CO2, climate, and increased nitrogen deposition (Archer et al., 1995). There is difficulty in definitively attributing the proportion of change from each factor, which will vary by location. There is significant variation in rates of accumulation of woody biomass by location (e.g., 0.25-2.5 t ha-1 yr-1 for South African savannas) (Scholes and van der Merwe, 1996), and differences in potential pool sizes are likely. This situation requires some spatial disaggregation to meet uncertainty levels specified for reporting.
Monitoring, Verifiability, and Transparency
Monitoring to detect change can be carried out by repeat sampling procedures (Back et al., 1997; Critchley and Poulton, 1998). Detailed guidelines for establishing a monitoring network and ensuring its representativeness are given by Brown (1997), MacDicken (1997), and Vine et al. (1999). At the individual plot level, the combination of allometry plus measured stem growth increment (including the use of dendrometers) is a powerful and accurate indicator of aboveground biomass flux. A verification and auditing team could evaluate satellite imagery to confirm that the integrity of registered sites was maintained, and auditing could be undertaken of a subset of these sites.
Biomass carbon increases are likely to continue for at least 50 years, though at reduced rates with time (Scholes and van der Merwe, 1996; Burrows et al., 1998). In some systems, the period of accumulation may be 100 years or longer (Archer, 1989).
Burning transfers a large proportion (up to 90 percent) of the aboveground carbon and nitrogen pools in some grasslands, and 3 percent of the total nitrogen pool to 10 cm in the soil (e.g., Kauffman et al., 1994), into the atmosphere as CO2, CO, CH4, N2O, NOX, and particulates. The mix of the gases depends on the material burned and the conditions of burning. Burning results in greater soil temperature, which increases soil CO2 fluxes (Knapp et al., 1998). Thus, burning results in a short-term loss of carbon from ecosystems. Replacement of this carbon, however, generally occurs within 1 to 3 years for grasslands (somewhat longer for woody plants). Where demands for fuel wood or agricultural products dictate or where population density is high (e.g., West Africa), increase in woody plant density will most likely be kept in check by ongoing management.
GHG emissions directly from burning (CO2, CO, CH4, N2O, NOX) and associated grazing activities (CH4, N2O) cause most productive savannas to be net sources of greenhouse emissions (0.06-0.2 t CO2-eq ha-1 yr-1 for semi-arid grasslands in Australia; Moore et al., 1997). Elimination of burning along with subsequent increases in woody plants and reductions in grazing livestock numbers can cause these systems to become net sinks for GHGs (about 1 t CO2-eq ha-1 yr-1; Moore et al., 1997).
In the absence of human intervention, catastrophic fire is the major threat to carbon storage in savannas. This situation may be impossible to prevent in the long term (50-100 years) (Scholes and van der Merwe, 1996), but because this vegetation type evolved under a regular burning regime, recovery after fire to the pre-fire structure is usually rapid. Extended droughts can also cause mortality (Fensham, 1998). Where mature woody plants die, they are likely to be replaced, provided that fire frequency remains low. Dead trees may remain standing, undergoing slow decomposition and providing a continuing but decreasing carbon store for periods of up to decades. Tree clearing and thinning may occur on dense stands to improve agricultural productivity.
For much of the world's broad-leaved savannas, as woody plant density increases, potential livestock carrying capacity declines (e.g., Scanla, 1992)-as does production for other purposes (e.g., Brookman-Amissah et al., 1980). In many ecosystems, fauna and flora species are fire-dependent, so removing or reducing fires may result in localized extinction or decline (e.g., Edroma, 1986). Increasing woody biomass may also reduce environmental services such as catchment water yield. Reducing burning will reduce atmospheric loads of particulates and various other forms of pollution.
Relationship to IPCC Guidelines
The Guidelines deal with savanna burning only in terms of non-CO2 gases; it is assumed that there are no net losses of carbon, and the system is assumed to be in balance on average.
Other reports in this collection