Avaliação da regeneração natural em áreas atingidas pelo rejeito da barragem de Fundão, Mariana, MG

Abstract

Environmental impacts cause a series of damages to forest ecosystems. In this context, ecological restoration can be used as a tool to reverse damage to the environment. In Brazil, the biggest environmental disaster occurred in the last decade, the collapse of the Fundão dam. As a solution, several initiatives were taken to promote the restoration of environments affected by the tailings. The technique of natural regeneration is still little used in restoration projects. However, it is an efficient, low-cost ecological strategy with large-scale applicability. This study aimed to evaluate the natural regeneration process in areas degraded by tailings and to evaluate which environmental variables most influence the density and richness of natural regeneration. For this, ten areas of the Doce river basin, in the state of Minas Gerais, were selected. Of this total, eight areas present tailings deposition and two were not impacted by the mud, being evaluated as reference ecosystems. In six areas restoration techniques were implemented, in others this process has been occurring naturally. Initially, an inventory of the regenerating layer was carried out, including all shrub and arboreal individuals with a height equal to or greater than 10 cm and with DBH (diameter at 1.30 m from the ground) less than 5.0 cm. In addition, data on soil cover and canopy, distance of the sample units to the watercourse and the nearest forest fragment, weight of living and dead biomass, percentage of surrounding area anthropized and conserved, as well as recording the presence or absence of cattle at each location. Statistical analyzes were carried out using the non-parametric Kruskal-Wallis test and generalized additive models for position, scale and shape (GAMLSS). The floristic survey identified 607 individuals in the areas affected by the tailings and 908 individuals in the reference ecosystems. In general, the Fabaceae family presented the greatest richness, and the species Vernonia polyanthes was accounted for in all degraded regions. The Kruskal-Wallis test revealed that, between areas, density and richness differ statistically. The Shannon and Pielou index ranged from 0.24 to 2.84 and 0.35 to 0.78, respectively. Among the areas under restoration, pioneer individuals and species were predominant, and in a few areas plants belonging to late secondary succession were recorded. The areas with tailings and without restoration actions presented a quantity of zoochoric species close to the reference ecosystems and the proportion of zoochoric individuals greater than in the anthropic areas. In other degraded areas, anemochoric individuals were more abundant. Statistical analysis by GAMLSS indicated that several environmental variables influence natural regeneration. However, canopy cover, presence or absence of livestock and distance to the forest fragment are the predictor variables that best explain the variation in the density of individuals and species richness.