| dc.description.abstract | The increase in the concentrations of atmospheric CO2 in the last century must impact ecosystems primary productivity. This effect may be direct and positive, due to a fertilization mechanism, or indirect (positive or negative) through climate changes. Evidences on the variation of carbon balance in forests in the last decades indicate neutral responses in boreal regions, positive in temperate regions and negative in the tropics. There is little evidence about how primary productivity responds to these global changing factors in subtropical forests. In this study, we investigate how the primary productivity of a dominant conifer in subtropical moist forests of the Southeast region of South America (Araucaria angustifolia) has been affected throughout the last century by elevated CO2 concentrations and consequent climate alterations. For a population of this species, we have tested the validity of the following hypothesis: (i) the increase of atmospheric CO2 indirectly affects the rate of carbon fixation, through the impact of global warming in the local temperature and rainfall regimes; and (ii) besides this indirect effect, the increment in CO2 directly affects the rate of carbon fixation through a fertilization mechanism. In an initial sample of 25 A. angustifolia trees from 0,33 ha of mature subtropical forest in the south of Brazil, we estimated annual series of wood carbon content for 14 trees using series of dendrochronologically dated growth-ring widths (from a preexisting study) and growth-ring densities (by X ray densitometry), and hypsometric (adjusted to local population) and volumetric (general for the species) equations. The individual series were combined in an average index series of carbon increment (I vector), previously removing ontogenetic tendencies (through Regional Curve Standardization) and time autocorrelation (by autoregressive models). By comparing I to regional series of mean temperature and total rainfall (estimations of CRU TS4) through a Correlation Function, we selected several climatic variables related to carbon fixation (P and T matrices). Instrumental data of atmospheric C concentration (NOAA EARL) and estimations of global temperature (CRU TEM4) composed matrices C and A, respectively. Finally, the validity of causal models describing the relations among matrices I, T, P, A e C according to the different hypothesis of the study, was tested through Path Analysis. The resulting chronology of residual indexes of carbon increment comprehended the period of 1890 to 2014, with averages of rbar = 0,27 and EPS = 0,77. In the subsequent analysis, we considered the period of 1901 to 2008 (common period with the climate series). Both models were not rejected in the Path Analysis (P > 0,1), and the model which represents the i hypothesis was considered the most plausible by the greater p-value for the Fisher’s C-statistic (Fisher-C = 9.25, gl =10, P = 0.508). In this model, global warmth comes from the elevation of CO2 which negatively affects the increase in carbon content in A. angustifolia through regional temperature elevation (in previous) may. Besides this climatic factor, rainfalls in previous march and june affected carbon fixation in a positive way, although both are not influenced by CO2 through global warming. In summary, climatic variables explained 19% of temporal variation of carbon increase in the trees. These results demonstrate that, despite the known positive impact of CO2 in water use efficiency in this species (and site), this does not translate into a greater primary productivity, possibly by the ombrophilous character of the regional climate. On the other hand, the change into autumns with higher temperatures has resulted in lower primary productivity of A. angustifolia throughout the last century. As this is the dominant species and plays a key role in the trophic structure, this negative effect of the climatic change in its productivity may, possibly, indirectly impact the structure and the operation of the ecosystems in which it is inserted. | en |
