It has been estimated that up to 50% of the earth's land surface has been transformed by human action (Vitousek et al., 1997b). Key amongst many changes have been: (i) conversion of forest to agricultural land; (ii) the subsequent abandonment of agricultural land and the natural or managed re-establishment of woody vegetation; (iii) fire suppression; and (iv) expansion of woody species. All these land-use change have altered ecosys tem productivity through changes in species composition, above- and belowground allocation of carbon, rooting depth and soil faunal communities (Nepstad et al., 1994; Jackson et al., 2000) and consequent changes in nutrient cycling and carbon storage (Trumbore, 1997; Jackson et al., 2002). The net consequence of the sum of land-use changes is thought to be a loss of carbon from terrestrial ecosystems to the atmosphere of 100 Pg C since 1850 (Houghton, 1995). The following subsections discuss in more detail the key land-use changes highlighted earlier. The specific changes discussed are not an exhaustive list but instead reflect our understanding of those changes currently thought to have been most significant from the perspective of recent changes in carbon storage in terrestrial vegetation.
The clearing of forest to make way for agriculture has been the most significant historical human-induced land-use change. Currently ~1800 million hectares are cultivated globally (Turner et al., 1993). In the USA cropland has increased from 50 million hectares to 200 million hectares in the last 150 years (Richards, 1990), with 140 million hectares of forest cleared during this time (Ramankutty and Foley, 1999). In Europe ~50% of the land surface is currently used for agriculture, with ~30% devoted to the production of crops; in addition, grasslands cover 55 million hectares (CarboEurope GHG, 2004a). Prior to 1950, land-use change was predominantly in the temperate zone; however, since then and arguably for the last century, forested regions in the tropics have seen the highest and most uncertain rates of conversion (Houghton, 2003). In general, conversion of forest to agricultural land removes a large stable aboveground and belowground carbon pool, both of which will likely be accumulating carbon. Depending on the harvest technique, and the use to which the harvested wood is put, forest clearance will result in the oxidation of a major proportion of the aboveground biomass and belowground soil carbon pools (Guo and Gifford, 2002). Once established, croplands and grasslands exhibit idiosyncratic carbon sequestration patterns. These patterns are strongly dependent on the exact use to which they are put, the intensity of their management and their geographical location (Grunzweig et al., 2004). Janssens et al. (2003) estimated that croplands in Europe lose 300 million tonnes of carbon per year, and are consequently the largest annual biospheric source of CO2 to the atmosphere. European grasslands were, conversely, thought to be a sink for atmospheric CO2 of 101 million tonnes of carbon per year (CarboEurope GHG, 2004b). However, it must be acknowledged that both these estimates were bounded by very large uncertainties.
From the perspective of carbon accumulation in biomass, abandonment of agricultural land that results in managed or natural regeneration of secondary forest is of great consequence. Nabuurs et al. (2003) explained that the current age structure of European forests is the consequence of large-scale afforestation of lands cleared for grazing or other agricultural practices in the early 20th century, and past management practices that centred on replanting after clear-cutting, followed by regular thinning of mono-specific even-aged stands. The result is that previous agricultural land is now covered by young forest, which is around 57 years, with an average above-ground carbon stock in stemwood of 40-50 t C/ha and an age structure in which there is an almost complete absence of forest cover that is more than 150 years. The age structure of these young European forests is important, because it is considered as the main reason that they are still actively sequestering carbon at a high rate. Valentini et al. (2000) highlighted how 25 of 27 forest study sites located throughout Europe were accumulating carbon, with a mean annual NEP for the 27 sites of 3 t C/ha/year. Within the Valentini study the oldest forest site was 110 years, while the mean age was 60 years. Significantly, the accumulation of stemwood in European forests has been predicted to continue until 2025 when age-related declines in productivity will become important (Nabuurs, 2004). In comparison with Europe, the average age of forests in the USA and European Russia is older at 76 and 80 years, respectively (Nabuurs, 2004). In the USA this is because there has not been such a pronounced abandonment of agriculture, with only 6% of the forestland lost to agriculture being reforested since 1920 (Houghton and Hackler, 2000). However, although managed afforestation of abandoned agricultural land in the USA has been small, the natural re-establishment of woody species has been significant. Casperson et al. (2000) calculated that the encroachment of native vegetation on abandoned agrcultural land in eastern USA constituted almost the entire regional sink for carbon.
While not the focus of this chapter, changes in vegetation biomass induced by land-use change cannot be considered independently of changes in soil carbon stocks. However, there is much specificity in soil carbon responses to land-use change. For example, the conversion of pastureland to pine plantation forestry resulted in losses of soil carbon sufficient to offset increases in vegetation; but conversion to broad-leaved or naturally regenerated secondary forest had no effect on soil carbon (Guo and Gifford, 2002). Conversely, the conversion of cropland to forest or plantation typically increases soil carbon in addition to the expected increases in aboveground biomass (Guo and Gifford, 2002).
Fire is a critical component of many ecosystems and has profound instant and long-term implications for carbon cycling and storage. Both historical reconstructions and current understanding of global fire frequency and extent are limited. However, in what the authors describe as a 'first approximation', Mouillot and Field (2005) estimated that 608 million hectares burned per year at the end of the 20th century, 86% of which were in tropical savannahs. Fires in forests have more impact on global carbon stocks, and Mouillot and Field (2005) estimated that 70.7 million hectares of forest burned annually at the beginning of the century, the majority being in the boreal and temperate forests of the northern hemisphere. As a consequence of fire suppression policies the area of temperate and boreal forest subject to fires decreased to 15.2 million hectares per year in the 1960s and a current figure of 11.2 million hectares per year. Over the same time period fires in tropical forests have increased exponentially to 54 million hectares per year. Although it is thought that fire suppression has certainly resulted in significant carbon accumulation in terrestrial ecosystems over the last century (Houghton et al., 2000), the quantitative consequences of fire, and thus fire suppression, on the terrestrial carbon cycle are hard to know. This is because difficult-to-determine factors, such as intensity of burn, vegetation type and standing biomass, control the amount of carbon emitted during a fire. However, experiments have demonstrated the impact of fire on ecosystem carbon pools. In a Minnesota oak savannah, fire suppression led to an average increase of 1.8 t C/ha/year, with most of the carbon stored in woody biomass and on the forest floor (Tilman et al., 2000).
Unmanaged encroachment of woody species on primarily herbaceous ecosystems, either as a consequence of climate change, human activity or the interaction of both, typically increases carbon sequestration in above-ground biomass. The literature increasingly documents carbon sequestration in above-ground biomass due to woody encroachment in diverse locations. These include the Arctic, where Sturm et al. (2001) document a doubling of woody vegetation in some Alaskan Arctic tundra study plots over the last 50 years. Increasing size of individual trees, filling in of gaps and increased density of forests at tree line sites were all observed. Historical photos were used to document increases in woody vegetation cover of up to 500% over a 63-year period in unmanaged Texas rangelands (Asner et al., 2003). Knapp and Soule (1998) observed a 59% increase in the cover of western juniper in central Oregon over a 23-year period. Similar trends have been documented in savannah ecosystems (Scholes and Archer, 1997) and grasslands (Jackson et al., 2002). However, the study of Jackson et al. (2002) was unique in that it clarified the need for consideration of soil carbon changes when assessing carbon sequestration due to woody encroachment, as at wetter study sites losses of soil carbon completely offset carbon accumulation in woody biomass.
Was this article helpful?