Spatial autocorrelation in Drosophilidae (Diptera) assemblages sampled in forest and pasture areas of the Cerrado Biome
Oliveira, Luciana Talita da Cunha deBlauth, Monica LanerGottschalk, Marco Silva
Abstract Spatial autocorrelation reflects the non-independence of observation sites due to their physical proximity. Evaluating this factor in a sampling design is essential to avoid pseudoreplication. Drosophilidae sampling, often conducted using banana and yeast traps for taxonomic surveys or ecological analyses, enables data comparison across various studies in different locations and seasons. However, the potential autocorrelation of sampling units due to trap spacing has seldom been tested, despite the widespread use of this methodology. In this study, we assessed the effects of spatial autocorrelation on Drosophilidae sampled using banana-baited traps placed 30 m and 60 m apart in forest and pasture settings during both dry and rainy seasons. The randomized accumulation curve indicated that the rainy season has a higher species richness, and traps spaced 60 m apart capture a broader species range. Species compositions were analyzed using Similarity Analysis (ANOSIM) with Jaccard and Bray-Curtis dissimilarity indexes. Distinct Drosophilidae compositions were observed across different phytophysiognomies and seasons. In evaluating species composition using 30 m and 60 m spacing across various areas and seasons, there is a tendency to sample similar assemblages, but with different dominances. No autocorrelation was detected in the Mantel test with the Jaccard index. These results suggest maintaining a minimum distance of 30 m between traps in taxonomic surveys and ecological studies of Drosophilidae.
Texto completo