Florística e estrutura de florestas inundadas na Floresta Nacional de Caxiuanã, Pará, Brasil

Abstract

In community ecology one of the most intriguing questions is how plant communities coexist. The study objective was to evaluate the species composition of the natural regeneration community in relation to geographic distance, soil substrate type and soil physicochemical composition. The study was carried out in the Caxiuanã National Forest, Pará, Brazil. Thirty plots of 2m x 2m and 5m x 5m were systematically distributed, fifteen plots in each type of forest, floodplain and igapó. Within each plot, all plants of natural regeneration with more than 20 cm in height were marked and identified. The similarity of species from natural regeneration between igapó and várzea forests was tested by Principal Component Analysis - PCA. To identify the possible co-structure between PCAs and environmental variables, we used co-inertia analysis, followed by a Monte Carlo permutation test (1,000 interactions). Eighty-five species were identified, with sixty-six species in the forests of Igapó in thirty-nine families and fifty species in the forests of Várzea, in twenty-three families. Regarding species richness (t=0.575; p=0.57) and the number of individuals (t=1.07; p=0.29) of natural regeneration between forests, we did not find a significant difference. The PCA analysis clearly separated the várzea and igapó species both for the 2mx2m strata plots, the first axis explains 10.2% of the data variance and the second axis explains 7.4%, and for the plots of the stratum of 5mx5m, the first axis of the order explains 10.5% of the data variance and the second axis explains 6.9%. Co-inertia showed significant covariance between the soil matrices and vegetation structure both for the 2mx2m stratum, sharing 48% of the data variance (Monte Carlo test, p = 0.001), and for the 5mx5m stratum, sharing 50% of the data variance (Monte Carlo test, p = 0.001). Co-inertia also showed significant covariance between the geographic distance and vegetation structure matrices for the 2mx2m stratum, sharing 39% of the data variance (Monte Carlo test, p = 0.001) and for the 5mx5m stratum, sharing 41% of the data variance (Monte Carlo test, p = 0.001). The plant community structure of natural regeneration in both igapó and várzea forests is influenced by both geographic distance and soil environmental variables.