Dinâmica de uma floresta de terra firme na Estação Experimental de Silvicultura Tropical, Manaus, Amazonas

Abstract

The dynamics of the tree community was studied in a terra firme Tropical Forestry Experimental Station, Manaus Amazonas during the period 1996 to 2010. The objectives of the study was to:(1); describe changes in floristic composition between 2004 to 2010 (2); report the changes in floristic composition along an environmental soil gradient (3); analyze changes in mortality and recruitment between 1996 to 2010 (4); assess the relationship in tree mortality and recruitment with soil; (5); evaluate the effectiveness of the probabilistic transition matrix (the first-order Markov Chain) for different monitoring intervals (two, four and six years); (6) estimate and evaluate the behavior of the variables: fresh and dry biomass and carbon content throughout the monitored period. The study was carried out over two transects for the tree community (diameter at breast height DBH > 10 cm) measuring 20 x 2500 m each (125 plots of 20 x 20 m systematically allocated in each of the transects). Temporal changes in floristic composition were small, as indicated by a Sørensen index of similarity of 97.2%, there were gains of 38 species (4.23%) and two families and losses of 20 species (2.27%) suggesting that the sudied tree community is adapted to the environment. These small changes in floristic composition were due to slight increase and decrease in richness of rare or unusual species. The soil variables were condensed into two ordination axes. The first axis explained 42.73%of data variation, represented by a gradient of soil texture, while second axis explained 23.73% of data variation, being largely explained by clay and sand, to a lesser extent by the concentrations of nitrogen (N) and carbon (C). The soil gradient was a good predictor of floristic composition, explaining 71.9% and 71.4% in qualitative and quantitative data variation respectivaly. This relation supports the deterministic theory that environmental heterogeneity is the main cause for the gradual replacement of species, suggesting a structure for niches. Recruitment rates (1.25 ± 0.66%. yr-1) equalized mortality rates (1.24 ± 0.45%. yr-1), giving the community a state of dynamic equilibrium. Over the period monitored both recruitment and mortality ranged in the intervals, the highest mortality rate observed was 1.94 ± 0.84%. yr-1 during the years 2004 to 2006, wich included the 2005 drought and blowdown. Although there is an environmental heterogeneity that provides different niches there was no significant effect (p> 0.05) of soil gradient on the events of death and recruitment of individual trees, suggesting that both occurred in a stochastic way along the gradient. For time scale considered there was no relationship between mortality and recruitment on environmental heterogeneity, which may be explained by the low variation of the predicted variables (mortality and recruitment) and the selective pressure that the lowland species suffer. Markov Chain approach is a reliable tool to project the forest dynamics on a short-term basis, of two years. Intervals of four and six years have not produced good estimates, showing that projection periods should be short enough to capture the changes between states and in demographics data of vegetation in order to allow the prognosis for the following period. We observed the same pattern for the variables: density of individuals, dry biomass, fresh biomass and carbon content. Confirming what was found for mortality rates and recruitment, these variables remained in balance if we consider the whole period, although fluctuations were observed in these values in different intervals. Despite the apparent dynamics observed over the years the tree community has remained stable in the range considered.