Occurrence of lemurs in landscapes and their species‐specific scale responses to habitat loss

Studies on the impact of habitat loss on species occurrence consistently find that the amount of habitat (measured as patch area) is a major determinant of species occurrence at a patch‐level. However, patch‐level research may fail to detect important patterns and processes only observable at a landscape‐level. A landscape‐level approach that incorporates species‐specific scale responses is needed to better understand what drives species occurrence. Our aim was to determine the landscape‐level scale of effect of habitat amount on the occurrence of three species of nocturnal lemurs (Cheirogaleus medius, Microcebus murinus, and M. ravelobensis). We surveyed line transects to determine the occurrence of three lemur species within a fragmented landscape of deciduous dry forest and anthropogenic grassland in northwestern Madagascar. To determine the scale of effect of habitat loss on lemur occurrence, we compared logistic regression models of occurrence against habitat amount among eight different landscape scales using Akaike's Information Criterion values. We found differing scale responses among the lemurs in our study. Occurrence of C. medius responded to habitat amount at scales between 0.5–4 ha, M. murinus at scales between 1 and 4 ha and M. ravelobensis at scales between 0.125 and 4 ha. We suggest that the scale of effect for C. medius is mediated by their ability to hibernate. A relatively lower scale‐response for Microcebus spp. likely reflect their omnivorous diet, small habitat requirements, and limited dispersal ability. Differences in scale responses between M. murinus and M. ravelobensis are likely a result of differing dispersal ability and responses to edge effects between these species. Our study is among the first on lemurs to show the value of a landscape‐level approach when assessing the effects of habitat loss on species occurrence. Graphical flowchart of how pseudo‐absence points were determined. (a) Fragment and with survey transect. (b) Addition of detection buffer. (c) Location of presence points. (d) Home range buffer placed on detection point. (e) Pseudo‐absence point placed within detection buffer absence zones. (f) Presence and pseudo‐absence landscape scale buffers. Research Highlights Closely related species show differing landscape‐level scale of effect of habitat loss on occurrence. Differing scale responses to habitat amount may relate to home range size and dispersal ability. Small scale of effect of habitat amount on the occu