Ecosystem C exchange has been measured in a few experiments, mostly in Alaskan wet and moist tundra (Fig. 4). In the wet tundra both gross ecosystem production, i.e., the photosynthetic gains, and respiratory C losses increased with nutrient addition. The increases were particularly pronounced in P and NP addition treatments with a strong N X P interaction, which was similar to the response pattern in the biomass (see above). Warming, in contrast, had smaller effects on C02 fluxes but still increased or tended to increase the fluxes. The net ecosystem productivity, which is the difference between the C fluxes into and out of the ecosystem, increased strongly after fertilizer addition and also tended to increase after warming. Hence, the increase in photo-synthetic carbon sequestration was more pronounced than the increase in respiration with warming only. However, as with the biomass response to combined warming and fertilizer addition, there was a negative interaction with decreased net ecosystem production due to pronounced respiratory C losses when the two treatments were combined.
Increased gross C fixation was found early in the growing season in tussock tundra after 3.5 years of warming, but the net ecosystem production still was negative because of a higher growing season respiratory C loss (Hobbie and Chapin, 1998). Also, measurements in the Swedish treeline heath after seven years of treatment showed a mid-season C loss in warmed plots relative to
ill i a
CL LU O
N P NP Treatment
FIGURE 4 Mean (±SE) net ecosystem production (NEP), ecosystem respiration RE and gross ecosystem production (GEP), in two Alaskan wet sedge tundra ecosystems subjected to N, P, and NP addition, greenhouse warming (GH) and combined NP and warming (GHNP). CT is untreated controls. (Redrawn from Shaver et al., 1998, with kind permission from Ecological Society of America).
controls. Flowever, the methods did not allow the inclusion of more than very low-growing vascular plants into the chambers for the flux measurement. Because the vascular plant biomass had increased strongly, it is likely therefore that the net ecosystem C balance for the system as a whole was positive, as after fertilizer addition (Christensen et al., 1997).
Modeling of the C balance for tussock tundra based on responses to experimental treatments (McKane et al., 1997) showed that warming first is likely to decrease the ecosystem C pool as a consequence of increased respiration. However, as soil N is mineralized and taken up by the vegetation, growth increases and offsets the respiratory losses and the model predicts a slight long-term increase in ecosystem C stock. Part of the increased sequestering of C is because the C-to-N ratio in the vegetation is higher than the ratio in the soil. Hence, the system as a whole can increase its C content without any increase in the N content by redistribution of N from the soil to the plant biomass (Shaver et al, 1992). The effect on the C sequestration is particularly large if the N is incorporated in woody tissue with a C-to-N ratio that is several times higher than the soil C-to-N ratio. If additional N is supplied, as in the fertilizer treatment, the model showed an even stronger increase of the C stock as a combined effect of the addition and a priming effect on N mineralization.
Overall, the results from ecosystem experiments, which have been conducted across a broad variety of ecosystem types in the Arctic, have shown both large similarities and dissimilarities. Within all manipulated ecosystem types, it appears that N or P addition has led to the greatest response, followed by a lower and much more variable response to warming, while water addition generally has led to small responses. Most responses, regardless of type of treatment, have occurred through the direct or indirect effect of the treatment on the N (or P) cycle (Shaver et al, 1992), which also feeds back to the ecosystem carbon balance in a variety of ways (McKane et al, 1997).
There are also great differences in responses between ecosystem types. These differences are mostly associated with the magnitude and not the direction of the responses, perhaps most evident in the range of biomass accumulation in response to warming. From the experiments done at various sites with similar vegetation types of, e.g., wet sedge tundra (Shaver et al, 1998), tussock tundra (e.g., Chapin et al, 1995; Ilobbie and Chapin, 1998) and dwarf shrub vegetation (fonasson et al, 1999b), the responses are similar within each ecosystem type, suggesting that the responses can be scaled up to represent large-scale heterogeneous vegetation assemblages. This should allow realistic modeling of longer-term responses to environmental change across broad regions of tundra.
Was this article helpful?