Climate Change 2001:
Working Group II: Impacts, Adaptation and Vulnerability
Other reports in this collection Carbon Sink

It has been suggested that water-level drawdown and increased temperature will decrease carbon sequestration in subarctic and boreal peatlands, especially in more southern latitudes (Gorham, 1991). However, conclusions are hampered by the diversity of possible responses.

Impacts of climate change on wetland carbon sink can be measured directly by using the eddy covariance method or chamber techniques (Crill et al., 1988; Fowler et al., 1995; Alm et al., 1997b; Aurela et al., 1998) and modeling the balances (Frolking et al., 1998). In recent years, several high quality, continuous, snow-free season time series of net ecosystem exchange have been obtained for wetlands and peatlands in the subarctic zone (Burton et al., 1996; Friborg et al., 1997; Schreader et al., 1998), the boreal zone (Shurpali et al., 1995; Jarvis et al., 1997; Kelly et al., 1997; Lafleur et al., 1997; Roulet et al., 1997; Suyker et al., 1997; Goulden et al., 1998), and the temperate zone (Happell and Chanton, 1993; Pulliam, 1993). Extrapolations of temporally detailed enclosure measurements along with estimates of the export of DOC have been undertaken to derive peatland carbon balances (Waddington and Roulet, 1996; Carroll and Crill, 1997). These studies point out the difficulty of assessing the sink/source strength of a peatland within a reasonable level of certainty because of errors introduced in the scaling process (Waddington and Roulet, 2000).

For northern peatlands, much is known about the effects of water-level drawdown on the carbon balance, based on studies carried out in peatlands drained for forestry (Glenn et al., 1993; Roulet et al., 1993; Laine et al., 1995; Martikainen et al., 1995; Minkkinen and Laine, 1998). These studies may be used cautiously to represent the climate change impact because the effect of drainage on ecosystem structure and functioning is similar to that predicted after drying caused by climate change in northern latitudes (Laine et al., 1996). The change observed in vegetation structure after water-level drawdown directs biomass production to the shrub and tree layer; in most cases, primary production and biomass increased (Laiho and Finér, 1996; Laiho and Laine, 1997; Sharitz and Gresham, 1998). Simultaneously, litter production increased (Laiho and Finér, 1996; Laiho and Laine, 1997; Finér and Laine, 1998), and the litter was more resistant to decomposition (Meentemeyer, 1984; Berg et al., 1993; Couteaux et al., 1998). There is some evidence that part of the carbon from decomposing litter is stored in the peat profile down to 0.5-m depth (Domisch et al., 1998). These alterations to vegetation production and litter flow into soil have been observed to keep the net carbon accumulation rate into the soil of boreal bogs in most cases at the level prior to water-level drawdown, sometimes exceeding this level (Minkkinen and Laine, 1998). However, increased duration and shortened return periods of extreme droughts may have detrimental effects on the peat carbon balance, as indicated by the results of Alm et al. (1997b).

Water-level drawdown will cause a decrease in CH4 emissions as substrate flux to anoxic layers is decreased and consumption of CH4 in the thicker aerobic layer is enhanced (Glenn et al., 1993; Roulet et al., 1993; Martikainen et al., 1995; Roulet and Moore, 1995). It has been suggested that reduced CH4 emissions after water-level drawdown, together with an increase in tree biomass and a fairly small change in carbon sequestration into peat, may even decrease the greenhouse effect of these ecosystems (Laine et al., 1996).

Therefore, it is not immediately clear that a warmer, drier climate necessarily will lead to a large loss of stored peat for all peatland types. The feedback between climate and peatland hydrology and the autogenic nature of peatland development is poorly understood. Clymo (1984) first developed this idea, and it has recently been used to show how surface topography on peatlands is preserved (Belyea and Clymo, 1999). Hilbert et al. (1998) have expanded on the work of Clymo (1984) and developed a model of peatland growth that explicitly incorporates hydrology and feedbacks between moisture storage and peatland production and decomposition. Their studies suggest that some peatland types (e.g., most bogs) will adjust relatively quickly to perturbations in moisture storage.

Future rates of carbon sequestration in swamps on mineral soils will depend largely on the response of trees to changes in hydrology, temperature, and elevated CO2 concentrations. The aboveground productivity of temperate zone swamp forests is strongly regulated by the extent of soil saturation and flooding, and the long-term effect of changes in hydrology on growth will depend on the position of forests along the current hydrological gradient (Megonigal et al., 1997). Drier conditions may increase NPP on extensively flooded sites and decrease it on dry and intermediate sites.

The combination of high soil carbon density and rapid warming in peatlands underlain by permafrost have raised concerns that northern peatlands may become net carbon sources rather than sinks (Lal et al., 2000). Indeed, working in tussock and wet tundra, Oechel et al. (1993) estimate that these systems are now net sources of 0.19 Gt C, caused mainly by melting of permafrost and lowering of the water table. Botch et al. (1995) report that peatlands of the former Soviet Union are net sources of 0.07 Gt C yr-1. Other parts of the boreal zone may have become enhanced sinks as a result of recent warming (Myneni et al., 1997), and continued warming could change the equation to favor net carbon storage—as suggested by warming experiments in arctic ecosystems (Hobbie, 1996).

Higher temperatures may affect carbon cycling of other wetlands as well. Increased photosynthetic activity of deep-rooted wetland plants, such as sedges, may enhance substrate availability for methanogenesis—which, together with higher temperatures, might lead to higher CH4 emissions (Valentine et al., 1994; Bergman et al., 1998; Segers, 1998) where water level would remain near the soil surface.

Land uses such as agriculture and forestry always change carbon fluxes in ecosystems. High carbon losses have been reported for agricultural crop production on drained wetland in Europe and North America, as much as 10-20 t C ha-1 yr-1 (Armentano and Menges, 1986). Studies have shown that agriculture on peat soils may contribute significantly to nitrous oxide (N2O) emission (Nykänen et al., 1995a). Kasimir-Klemedtsson et al., (1997) conclude that agricultural practices on organic soils lead to a net increase in emissions of GHGs because of large fluxes of CO2 and N2O, over decreases in emissions of CH4.

As agricultural management fundamentally alters the processes of wetlands and gradually leads to decreases in wetland area (Okruszko, 1996). Large areas of wetlands have been lost in Russia, Europe, and North America by complete drainage and conversion to other land uses. It has been estimated that 53% of the original 89 Mha of wetlands in the coterminous United States were lost by the 1980s (Shepard et al., 1998), much of it to agricultural conversion. Development for agriculture also can have offsite effects—for example, reduced water quality that impacts fisheries (Notohadiprawiro, 1998). Arable agriculture always transforms wetlands into sources of GHGs to the atmosphere (Armentano and Menges, 1986; Okruszko, 1996), with the exception of CH4.

Consequences of the development of tropical peatlands include lowering of the water table, which promotes peat oxidation and decomposition. Peat loss and subsidence can occur at very fast rates—as much as 0.9 cm per month (Dradjad et al., 1986). Eventually, shrinkage and oxidation may lead to loss of the entire peat profile and exposure of underlying nutrient-poor substrates or potential acid sulphate soils (Maltby et al., 1996; Rieley et al., 1996). In subcoastal situations, this may be followed by marine inundation.

When wetland use for forestry involves only management of existing tree stands, the impacts on functions and processes may be small, and the ecosystem may remain within the wetland concept (Aust and Lea, 1991; Minkkinen et al., 1999). However, if artificial drainage is included, decay of organic matter is enhanced, with consequent increases in CO2 emissions from peat (Glenn et al., 1993; Silvola et al., 1996). The results of the carbon balance change reported are highly variable, depending on methods used and climatological differences. Losses of peat carbon have been reported by Braekke and Finer (1991) and Sakovets and Germanova (1992), whereas increased post-drainage carbon stores have been reported by Anderson et al. (1992) and Vompersky et al. (1992). Based on a large cross-sectional data set from forest drainage areas in Finland, it was shown that carbon accumulation in peat soil increased in southern parts of the country (annual mean temperature 3-4.5°C) but decreased in northern Finland (with mean temperature of 0-1°C). The carbon accumulation increase was clearest for nutrient-poor bog sites (Minkkinen and Laine, 1998). Cannell and Dewar (1995) have concluded that drainage and planting of conifers on organic soils produces little long-term change in soil carbon stores because enhanced organic matter oxidation is compensated by increased litter production of the tree stand.

Water-level drawdown after drainage decreases CH4 emissions from peatland (Glenn et al., 1993; Roulet et al., 1993; Martikainen et al., 1995; Roulet and Moore, 1995). Increased consumption in the surface soil may even form a small CH4 sink in some cases (Glenn et al., 1993; Roulet et al., 1993; Fowler et al., 1995; Martikainen et al., 1995; Roulet and Moore, 1995; Komulainen et al., 1998). The effect of drainage on N2O emissions has been reported to be fairly small and restricted to fen sites (Martikainen et al., 1993).

Forest harvesting in tropical swamp forests can result in changes to the quality and quantity of organic matter inputs from vegetation, and—as the work of Brady (1997) has shown—if tree root mats decline, net accumulation of peat also may decline. Where selective logging is combined with artificial drainage, decomposition and subsidence of peat may proceed at rates of 3.5—6.0 cm yr-1 (Brady, 1997). In contrast with peatlands of the temperate and boreal zone, there has been poor success with establishing forestry plantations on tropical peatlands.

Peat harvesting totally changes the structure and functioning of the original ecosystem by removing the vegetation and finally most of the accumulated peat deposit. This has a fundamental impact on the GHG balances of harvesting sites: CH4 emissions almost stop (Nykänen et al., 1995b), but the whole accumulated carbon store forms a CO2 source to the atmosphere during harvesting and combustion (Rodhe and Svensson, 1995), even if peat combustion may replace imported energy in countries with no other major domestic energy sources.

Recent findings have shown that restoration of cut-away peatlands after harvesting soon initiates colonization of peatland plants (Smart et al., 1989; Tuittila and Komulainen, 1995; Campeau and Rochefort, 1996; Wheeler, 1996; Wind-Mulder et al., 1996; Ferland and Rochefort, 1997; LaRose et al., 1997; Price et al., 1998) and may restart carbon accumulation (Tuittila et al., 1999).

Elevated CO2 is likely to stimulate CH4 emissions in a wide variety of wetland ecosystems, including freshwater marshes (Megonigal and Schlesinger, 1997) and rice paddies (Ziska et al., 1998). Studies in northern peatlands have been equivocal: One study reports a maximum increase of 250% (Hutchin et al., 1995) and another no increase (Saarnio et al., 1998). Because CH4 is a more powerful GHG than CO2, wetlands amplify the greenhouse effect of elevated CO2 by converting a portion of this gas to CH4.
Increased nitrogen deposition may alter the species composition of wetland communities (Aerts et al., 1992) and their production, leading to higher production and net accumulation rates in peatlands where production is nitrogen limited (Aerts et al., 1995). There is some indication that nitrogen inputs may affect trace gas emissions from peat soils by increasing emissions of N2O and sometimes decreasing those of CO2 and CH4 (Aerts, 1997; Aerts and Ludwig, 1997; Aerts and Toet, 1997; Regina et al., 1998).

Other reports in this collection