As discussed by Kittel etal. (2000), terrestrial ecosystems in the Arctic and boreal forest regions are expected to be highly sensitive to climate change and may play a strong role in biospheric feedbacks to global climate. This sensitivity arises from complex interactions among ecosystem structure and function, soil and permafrost processes, and regional climate. Biophysical and biogeochemical dynamics of these landscapes in turn impact the global climate system through control over surface-atmosphere exchanges of energy and radiatively active trace gases. Arctic and boreal ecosystems are characterized by large carbon stocks. Tundra ecosystems contain approximately 11% of the world's soil carbon that might react to near-term climate change (e.g., McGuire et al., 1995). There is indeed evidence that recent warming in the Arctic (Chapter 11) has affected both the structure and function of Arctic ecosystems (e.g, Chapin et al., 1995; Oechel et al., 1995). Functional responses include changes in the biogeochemical cycling of carbon, nutrients and water in ecosystems while structural responses represent changes in the species composition across a landscape, that may further modify the function of ecosystems (Clein et al., 2000).
The issues are complex. For example, the analysis of Betts (2000) indicates that a change from tundra to boreal forest could increase carbon uptake by trees, and help to mitigate the warming effects of atmospheric greenhouse gas loading. However, such changes could be offset by the lower albedo of boreal forest as compared to tundra. A substantial amount of carbon and nitrogen from the tundra could be released in inorganic forms. As these soils warm, decomposition rates increase and the growing season lengthens. A large release of carbon dioxide from these soils could enhance the rate and magnitude of climate warming. On the other hand, a large release of inorganic nitrogen could also result in increased Net Primary Production (NPP), which could buffer carbon loss from the soil or cause Arctic ecosystems to act as a net sink for carbon dioxide (Clein et al., 2000). NPP is the net carbon gain by vegetation. It equals the difference between gross primary production (GPP, carbon gain through photosynthesis at the scale of ecosystems) and carbon loss through plant respiration (also known as autotrophic respiration). This contrasts with heterotrophic respiration (HR), which is the respiration by organisms (animals and microbes) that (like the authors) gain their carbon by consuming organic matter rather than producing it themselves. The key to the carbon balance is the net ecosystem production (NEP), which is the net carbon flux of the ecosystem (NPP and the carbon loss through HR) (see the book by Chapin etal., 2000).
It should come as no surprise that modeling Arctic ecosystems and ecosystem change is a vibrant research area. Ecosystem models have historically been either models of ecosystem structure (biogeography models) or models of ecosystem function (biogeochemistry models). Recently, some models have been developed that can simulate changes in both ecosystem structure and function. There has also been an evolution from equilibrium models (VEMAP Members, 1995) to transient models (e.g., McGuire et al., 2000) that incorporate time changes in forcings such as temperature, carbon dioxide and precipitation. Transient models, also known as "dynamic vegetation models" have been coupled to global climate models (e.g., Cox et al., 2000) and can now simulate how changes in ecosystem structure and function influence climate at time scales of minutes to centuries. Important issues that these models need to represent include the responses of soil and vegetation carbon, changes in shrubbiness of tundra and invasion of trees into tundra regions. A major challenge of ecosystem models is to represent exchanges of both carbon dioxide and methane. There are often tradeoffs between carbon dioxide and methane exchanges. An expansion of wetlands enhances carbon storage in anaerobic soils but releases more methane, which (molecule for molecule) is 23 times more effective than carbon dioxide in trapping heat near the surface of the Earth.
A good example of an Arctic application is the effort of McGuire et al. (2000) (also see other papers in the same journal), who examined carbon dynamics over the entire Arctic tundra and Kuparuk river basin (Alaska) for historical and future conditions (1921-2100) using the Terrestrial Ecosystem Model (TEM). TEM is designed to simulate the carbon budget of terrestrial ecosystems across the Earth at high spatial resolution for time periods of a century or more. Input variables include temperature, precipitation, cloud cover, atmospheric carbon dioxide, elevation, soil texture, vegetation and nitrogen inputs. Temperature and precipitation were based on observations (1921-94) and output from a run of the Hadley Centre CM2 GCM (19952100), which included the combined effects of radiative forcing from changes in greenhouse gases and sulfate aerosols. The evolution of atmospheric carbon dioxide was based on ice core records, observations (1921-94) and an assumed 0.7% per year increase thereafter. To simulate the influence of atmospheric nitrogen inputs, McGuire et al. annually added 0.06 gN m-2 y-1 to the inorganic nitrogen pool of TEM.
Figure 9.14 Historical and projected changes in (a) net primary production (NPP) (b) heterotrophic respiration (HR) and (c) net ecosystem production (NEP) for the entire Arctic tundra area (solid line) and the Kuparuk Basin (dotted line) as simulated by the Terrestrial Ecosystem Model (TEM) during the historical period (1921-94) and projected period (1995-2100). The shading represents the standard deviation in carbon fluxes simulated by TEM across the entire tundra (from McGuire et al., 2000, by permission of Blackwell Publishing).
In this simulation (Figure 9.14), carbon storage of tundra increases substantially across the Arctic during the projection period (seen in the time series of NEP). But as pointed out by McGuire et al. (2000) and Clein et al. (2000), these estimates represent the suite of assumptions about the function and structure of tundra ecosystems that have been incorporated into the model. The model was calibrated to observed tundra carbon dynamics in the Kuparuk Basin and then applied across the Arctic. A conclusion that McGuire et al. (2000) drew from their "upscaling" study is that carbon dynamics really needs to be simulated in a spatially explicit fashion - tundra carbon dynamics in the Kuparuk Basin cannot be taken as representative of those across the Arctic. A major challenge is to incorporate topographic controls over soil moisture. They also identify the need for more spatially explicit forcing data (e.g., precipitation and temperature).
Was this article helpful?