Study sites
The study was conducted during the months of May and June 2021 on Lipsi Island, Greece (approximate area: 17 km2), which is located within the eastern Aegean Sea (37° 17’ 44.7” N, 26° 46’ 45.5” E) and used as a model small island ecosystem. While the island experiences low-intensity urbanisation on a global scale (building density = 15-20% in urban areas), the impacts of urban development on the natural landscape of small island ecosystems can be comparatively greater than in built-up areas and warrant investigation (Fernandes and Pinho, 2017). Sampling over an entire year was not logistically feasible, so we chose this timescale because previous studies in Mediterranean regions indicated that peak butterfly activity and abundance should occur in May and June (Fileccia et al., 2015; Hantson and Baz, 2013). The lower abundance and richness observed in early spring and late summer is mainly due to a reduction in flower diversity and, thus nectar sources for pollinators (Hantson and Baz, 2013).
Nine urban and nine rural sites with clear separation were selected (Figure 1). The minimum distance between study sites in urban areas was 75.3 m, which limited the double counting of individuals. Due to the lack of trees, shade is restricted on Lipsi Island, thus, locations with high light intensity and low shade were utilised to give an accurate representation of the urban and rural habitats used by butterfly species. Sites were chosen to represent the predominant land-use types utilised by butterflies during one or several stages of their life cycle (Grill and Cleary, 2003). The chosen rural habitats were shrubland, olive groves, and meadows, while the urban habitats included agricultural meadows, abandoned land, parks, roadsides, and olive groves. The sites were similar in size to keep the sampling effort consistent.