DYNAMICS OF VEGETATION IN AN ATLANTIC FOREST FRAGMENT IN NORTHEASTERN BRAZIL

To test the hypothesis that the physiognomy, structure and dynamics of woody species are influenced by the spatial gradient edge-interior, data were collected from these parameters in 2004, 2007 and 2010 in an Atlantic forest fragment. Species were sampled for the Canopy (DS), with circumference > 15 cm to 1.3 m soil and understory (SB) in circumference between ground level> 3 cm and CAP <15 cm on edge-gradient formed within 35 years. In 2004, we assessed three positions in the fragment: edge, intermediate and inner (> 150 m from the forest boundary), and analyzed 1,000 m(2) per site, divided into plots of 10 m(2), to measure individuals of higher class. In each plot a sub-plot of 5 m(2) for sampling understory was installed. Branches were collected from three adult individuals from the species to obtain the density of the wood. The number of individuals (NI), basal area (BA), mortality (TM), recruitment rate (TR), periodic annual diameter increment (IPA), loss rate (P) and gain rate (G) Area and basal turnover rates in number of individuals (TN) and basal area (TA) class tree component in the three locations showed no significant edge effect. At the edge, the understory had lower values of NI and AB and higher TM than in the forest interior. Differences between the strata occur regardless of edge sealing, since young plants present in most individuals in the sample class SB, which are more sensitive than adult trees to changes caused by the creation of the edge. Rates of loss and gain of AB and AB turnover in NI and in SB presented higher values than on the DS, which points to a stratum with larger changes. In DS there were no differences in species richness among environments differs from that observed in SB. The floristic composition of the DS in environments closer to the forest limit was quite distinct from that of the interior, indicating that there has been no recovery of floristic composition. All environments SB had low similarity. The edges of the DS and SB had a higher concentration of individuals in class intervals smaller density values timber. The DS has already recovered biomass, although the edge and intermediate environments have not recovered species composition late. The SB still suffers the edge effect. Possibly, the highest concentration of individuals is in class with lower density wood result on edge.