3.2.2.1 Background

Higher plants acquire CO₂ by diffusion through tiny pores (stomata) into leaves and thus to the sites of photosynthesis. The total amount of CO₂ that dissolves in leaf water amounts to about 270 PgC/yr, i.e., more than one-third of all the CO₂ in the atmosphere (Farquhar et al., 1993; Ciais et al., 1997). This quantity is measurable because this CO₂ has time to exchange oxygen atoms with the leaf water and is imprinted with the corresponding ¹⁸O “signature” (Francey and Tans, 1987; Farquhar et al., 1993). Most of this CO₂ diffuses out again without participating in photosynthesis. The amount that is “fixed” from the atmosphere, i.e., converted from CO₂ to carbohydrate during photosynthesis, is known as gross primary production (GPP). Terrestrial GPP has been estimated as about 120 PgC/yr based on ¹⁸O measurements of atmospheric CO₂ (Ciais et al., 1997). This is also the approximate value necessary to support observed plant growth, assuming that about half of GPP is incorporated into new plant tissues such as leaves, roots and wood, and the other half is converted back to atmospheric CO₂ by autotrophic respiration (respiration by plant tissues) (Lloyd and Farquhar, 1996; Waring et al., 1998).

Annual plant growth is the difference between photo-synthesis and autotrophic respiration, and is referred to as net primary production (NPP). NPP has been measured in all major ecosystem types by sequential harvesting or by measuring plant biomass (Hall et al., 1993). Global terrestrial NPP has been estimated at about 60 PgC/yr through integration of field measurements (Table 3.2) (Atjay et al., 1979; Saugier and Roy, 2001). Estimates from remote sensing and atmospheric CO₂ data (Ruimy et al., 1994; Knorr and Heimann, 1995) concur with this value, although there are large uncertainties in all methods. Eventually, virtually all of the carbon fixed in NPP is returned to the atmospheric CO₂ pool through two processes: heterotrophic respiration (Rh) by decomposers (bacteria and fungi feeding on dead tissue and exudates) and herbivores; and combustion in natural or human-set fires (Figure 3.1d).

Table 3.2: Estimates of terrestrial carbon stocks and NPP (global aggregated values by biome).
Biome	Area (109 ha)		Global Carbon Stocks (PgC)f						Carbon density (MgC/ha)				NPP (PgC/yr)
	WBGUa	MRSb	WBGUa			MRSb	IGBPc		WBGUa		MRSb	IGBPc	Atjaya	MRSb
			Plants	Soil	Total	Plants	Soil	Total	Plants	Soil	Plants	Soil
Tropical forests	1.76	1.75	212	216	428	340	213	553	120	123	194	122	13.7	21.9
Temperate forests	1.04	1.04	59	100	159	139e	153	292	57	96	134	147	6.5	8.1
Boreal forests	1.37	1.37	88d	471	559	57	338	395	64	344	42	247	3.2	2.6
Tropical savannas & grasslands	2.25	2.76	66	264	330	79	247	326	29	117	29	90	17.7	14.9
Temperate grasslands & shrublands	1.25	1.78	9	295	304	23	176	199	7	236	13	99	5.3	7.0
Deserts and semi deserts	4.55h	2.77	8	191	199	10	159	169	2	42	4	57	1.4	3.5
Tundra	0.95	0.56	6	121	127	2	115	117	6	127	4	206	1.0	0.5
Croplands	1.60	1.35	3	128	131	4	165	169	2	80	3	122	6.8	4.1
Wetlandsg	0.35	-	15	225	240	-	-	-	43	643	-	-	4.3	-
Total	15.12	14.93h	466	2011	2477	654	1567	2221					59.9	62.6
a WBGU (1988): forest data from Dixon et al. (1994); other data from Atjay et al. (1979). b MRS: Mooney, Roy and Saugier (MRS) (2001). Temperate grassland and Mediterranean shrubland categories combined. c IGBP-DIS (International Geosphere-Biosphere Programme – Data Information Service) soil carbon layer (Carter and Scholes, 2000) overlaid with De Fries et al. (1999) current vegetation map to give average ecosystem soil carbon. d WBGU boreal forest vegetation estimate is likely to be to high, due to high Russian forest density estimates including standing dead biomass. e MRS temperate forest estimate is likely to be too high, being based on mature stand density. f Soil carbon values are for the top 1 m, although stores are also high below this depth in peatlands and tropical forests. g Variations in classification of ecosystems can lead to inconsistencies. In particular, wetlands are not recognised in the MRS classification. h Total land area of 14.93x109 ha in MRS includes 1.55x109 ha ice cover not listed in this table. InWBGU, ice is included in deserts and semi-deserts category.

Most dead biomass enters the detritus and soil organic matter pools where it is respired at a rate that depends on the chemical composition of the dead tissues and on environmental conditions (for example, low temperatures, dry conditions and flooding slow down decomposition). Conceptually, several soil carbon pools are distinguished. Detritus and microbial biomass have a short turnover time (<10 yr). Modified soil organic carbon has decadal to centennial turnover time. Inert (stable or recalcitrant) soil organic carbon is composed of molecules more or less resistant to further decomposition. A very small fraction of soil organic matter, and a small fraction of burnt biomass, are converted into inert forms (Schlesinger, 1990; Kuhlbusch et al., 1996). Natural processes and management regimes may reduce or increase the amount of carbon stored in pools with turnover times on the order of tens to hundreds of years (living wood, wood products and modified soil organic matter) and thus influence the time evolution of atmospheric CO₂ over the century.

The difference between NPP and Rh determines how much carbon is lost or gained by the ecosystem in the absence of disturbances that remove carbon from the ecosystem (such as harvest or fire). This carbon balance, or net ecosystem production (NEP), can be estimated from changes in carbon stocks, or by measuring the fluxes of CO₂ between patches of land and the atmosphere (see Box 3.1). Annual NEP flux measurements are in the range 0.7 to 5.9 MgC/ha/yr for tropical forests and 0.8 to 7.0 MgC/ha/yr for temperate forests; boreal forests can reach up to 2.5 MgC/ha/yr although they have been shown to be carbon-neutral or to release carbon in warm and/or cloudy years (Valentini et al., 2000). Integration of these and other results leads to an estimated global NEP of about 10 PgC/yr, although this is likely to be an overestimate because of the current biased distribution of flux measuring sites (Bolin et al., 2000).

When other losses of carbon are accounted for, including fires, harvesting/removals (eventually combusted or decomposed), erosion and export of dissolved or suspended organic carbon (DOC) by rivers to the oceans (Schlesinger and Melack, 1981; Sarmiento and Sundquist; 1992), what remains is the net biome production (NBP), i.e., the carbon accumulated by the terrestrial biosphere (Schulze and Heimann, 1998). This is what the atmosphere ultimately “sees” as the net land uptake on a global scale over periods of a year or more. NBP is estimated in this chapter to have averaged -0.2 ± 0.7 PgC/yr during the 1980s and -1.4 ± 0.7 PgC/yr during the 1990s, based on atmospheric measurements of CO₂ and O₂ (Section 3.5.1 and Table 3.1).

By definition, for an ecosystem in steady state, Rh and other carbon losses would just balance NPP, and NBP would be zero. In reality, human activities, natural disturbances and climate variability alter NPP and Rh, causing transient changes in the terrestrial carbon pool and thus non-zero NBP. If the rate of carbon input (NPP) changes, the rate of carbon output (Rh) also changes, in proportion to the altered carbon content; but there is a time lag between changes in NPP and changes in the slower responding carbon pools. For a step increase in NPP, NBP is expected to increase at first but to relax towards zero over a period of years to decades as the respiring pool “catches up”. The globally averaged lag required for Rh to catch up with a change in NPP has been estimated to be of the order of 10 to 30 years (Raich and Schlesinger, 1992). A continuous increase in NPP is expected to produce a sustained positive NBP, so long as NPP is still increasing, so that the increased terrestrial carbon has not been processed through the respiring carbon pools (Taylor and Lloyd, 1992; Friedlingstein et al., 1995a; Thompson et al., 1996; Kicklighter et al., 1999), and provided that the increase is not outweighed by compensating increases in mortality or disturbance.

The terrestrial system is currently acting as a global sink for carbon (Table 3.1) despite large releases of carbon due to deforestation in some regions. Likely mechanisms for the sink are known, but their relative contribution is uncertain. Natural climate variability and disturbance regimes (including fire and herbivory) affect NBP through their impacts on NPP, allocation to long- versus short-lived tissues, chemical and physical properties of litter, stocks of living biomass, stocks of detritus and soil carbon, environmental controls on decomposition and rates of biomass removal. Human impacts occur through changes in land use and land management, and through indirect mechanisms including climate change, and fertilisation due to elevated CO₂ and deposition of nutrients (most importantly, reactive nitrogen). These mechanisms are discussed individually in the following sections.

IPCC Homepage