Padrões de crescimento em floresta tropical sazonalmente seca em relação à variabilidade ambiental no semiárido pernambucano

Abstract

This study aims to evaluate the relation of the vegetation to its environmental conditions, in order to understand its behavior, in the seasonally dry tropical forest of Caatinga, in northeastern Brazil. Initially, the annual and monthly variability of rainfall and average air temperature were evaluated and the water deficit was quantified through Climatological Water Balance (CWB). It identified high water deficit, high rainfall variability, an average rainfall of 520.70 mm, and an average annual air temperature of 30.66 ºC. Because of the cyclical drought, there was no soil water surplus in all years studied. The values of potential evapotranspiration of 1563.44 mm.year-1 were higher than the values of precipitation, characteristics of a climate of the type (DdA'a'). Then, the structure, dynamics and growth in basal area of two distinct zones in a Caatinga environment located in Floresta – PE, Brazil were evaluated. They were: A1 (the least preserved), that suffered loss of vegetation in 1986, and A2 (the most preserved), with no substantial history of disturbance. Forty plots (20 m x 20 m) were implanted for the continuous forest inventory – CFI, measuring individual plants with a circumference at 1.30 m from the ground ≥ 6 cm. There was a decrease in the number of families, gender, and species, with mortality rates higher than those of entrances. However, an increase was noticed in the basal area for (A1), initially (2008) with 3.1 m2.ha-1 and in 2019 with a value of 4.6 m2.ha-1. While (A2) came from 8.3 m2.ha-1 to 7.0 m2.ha-1. Thus, a greater influence of drought can be seen on A2, because of a period of low rainfall from 2012 to 2015. Finally, the influence of environmental variables on the periodic annual increment - PAI of the basal area was analyzed using a mixed linear model. Spearman's correlation was used to verify the existence of correlation with environmental variables. The analysis of variance with mixed effects was performed considering period (time) and species as a random effect, while the fixed effects variables were considered as dendrometric: number of shafts, inlets, mortality, equivalent diameter and environmental variables, then the following mixed linear model was created. The mixed ANOVA assumptions were performed based on the analysis of residues. It was noticed that there is a correlation between precipitation and seasonal factor, associated with temperature and drought indices. However, the accumulated annual and semiannual rainfall do not show a strong relationship with the PAI. There was an influence for (A1) in relation to environmental variables in fixed effects parameters, whereas in (A2), only significant influences were noted in the fixed effects for dendrometric variables. In general, the mixed linear model showed good performance for both areas, presenting good residual distributions, without the presence of outliers. It is concluded that vegetation growth is influenced by environmental variables, although this fact was perceived only for A1 (less preserved), for A2 (preserved) such influences were not noted, possibly because of intrinsic individual traits. However, in the long term it is necessary to continue monitoring the vegetation through the studied variables.