Fruit feeding butterflies as indicator taxon, pitfalls and concerns demonstrated in the Atlantic Forest

•Fruit feeding butterflies are poorly utilized when quantifying traditional measures.•Fruit feeding butterflies are extremely sensitive to certain disturbances.•New approaches to grouping of samples can yield great results. In the search for the ideal indicator taxon, fruit feeding butterflies have attracted increased focus for use in ecosystem assessments in the tropics. This study aims to shed light on the usability of fruit feeding butterflies in ecosystem assessments by comparing the effect of two different measures of ecosystem degradation on the fruit feeding butterfly community. The Interior biogeographical sub region of the Atlantic Forest, southern Brazil. In this study, the impact of tree cover and urban area on the abundance, richness and diversity of fruit feeding butterflies were investigated. Tree cover was measured within 200 and 1000 m of each trap while urban area was just measured within 1000 m. This was accomplished with the use of standardized trapping methods and geospatial tools that allowed for spatial analysis of the area surrounding each trap. In contrast to the common approach, the grouping of sampling units where based on the specific research question addressed, thus enhancing the statistical significance of the results. Within 200 m a moderate amount of tree cover (~35–50% tree cover), were found to be positively correlated with all community metrics, while a high amount of tree cover within both 200 and 1000 m (~>50% tree cover) displayed a pronounced negative correlation with all community metrics. Any amount of urban area within 1000 m (~0–30% urban area) showed a pronounced negative correlation with all community metrics. Considering the effect of tree cover, several commonly used community metrics for fruit feeding butterflies were shown to respond contradictorily to what might be expected. However, regarding urban area, the fruit feeding butterfly community were found to respond as expected and was extraordinarily sensitive to disturbances. These results highlight the need to be critical in the choice of community metrics when utilizing fruit feeding butterflies as indicator taxon, while also demonstrating their sensitivity to habitat changes and potential in future habitat assessments. In addition, grouping trap data in accordance to specific research questions was found to yield very robust results, thus allowing for reduced survey costs in future studies.