Comunidades de visitantes florais e redes de interações em mosaico vegetacional do Parque municipal de Mucugê e arredores, Chapada Diamantina

Abstract

(SPATIAL VARIATIONS IN INTERACTION NETWORKS OF PLANTS AND FLOWER VISITORS AT MULTIPLE SCALES) Most studies on interactions between plants and flower visitors with a network approach are descriptive and lack empirical tests of their properties. Although many studies point out that communities of animals and plants respond to habitat variations and to landscape characteristics, there are still a few evidences that demonstrate the effect of these factors on interaction networks. Aiming at investigating factors that explain spatial variation in networks of visitors and flowers, I selected 15 sampling units of 0.25 ha, from 1.5 km to 8.0 km apart from each other, in the surroundings of Parque Municipal de Mucugê (12°59'18"S and 41°20'22"W), Chapada Diamantina, Bahia, where the dominant vegetation is composed of rocky fields intercalated with Cerrado patches, and the regional climate is tropical semi-humid. The plants and their visiting species were sampled in March, April and May (main flowering period), totalizing a sampling effort of 33 hours/area. Using visit data, I built local interaction networks, described with the following parameters: number of vertices, number of links, diameter, number of components, average shortest path, centralization, degree of nestedness and degree of agglomeration. Sampling units were described according to their habitat characteristics, similarity of floral resources, similarity of flower visitor communities, vegetation type, distance between sampling units and landscape characteristics (obtained from the characterization of the surroundings of each unit in buffers of 250, 500 and 750 m, based on a Ikonos satellite image with 1-m resolution, and using parameters of composition and configuration). Using a principal component analysis (PCA) I extracted explanatory axes for network, habitat and landscape parameters. In order to test for the association between explanatory factors and network characteristics I estimated correlations with adjusted significance levels. None of the factors analyzed showed a significant relationship with the network properties. Hence, results indicate that even if communities vary spatially, their network properties are constant. Those findings advanced knowledge in this field and may be important for conservation.