Highly Variable Densities and a Decline in Critically Endangered Golden-Crowned Sifaka (Propithecus tattersalli) Abundance from 2008–2018

International audience

Inglés. Animal abundance is determined by a number of factors, including vegetation structure, food availability and quality, human activities, predation risk, and disease. Vegetation structure, food availability, and human activity often are used to guide conservation efforts, such as protected area zoning and reforestation, especially for primates. We sought to determine whether Critically Endangered golden-crowned sifaka (Propithecus tattersalli) densities could be predicted across a heterogeneous landscape as a function of vegetation structure, food availability, and human activity. We conducted walking transect surveys across the sifakas' entire global range in Loky-Manambato Protected Area of Madagascar from 2016-2018, expanding upon a study conducted in 2006/2008. Potential predictors of sifaka density included metrics of vegetation structure (e.g., tree density, forest type), food availability (e.g., food tree basal area, normalized difference vegetation index (NDVI)), and human activity (e.g., tree cutting, livestock grazing). Low-intensity tree cutting and wet season NDVI were the best, positive predictors of sifaka densities. Sifaka densities within study units across their restricted range (880 km(2)) were highly variable (range: 6.8-78.1 sifakas/km(2)), emphasizing the importance of large-scale study designs across all suitable land cover types for assessing a species' abundance, regardless of its area of occupancy. We estimated that 10,222-12,631 sifakas remain. Based on previous surveys, this indicates that populations either remained stable over the past 20 years or have declined by 30-43% in the past 10 years; we argue that a decline is most likely based on our updating of forest cover estimates for both of the prior studies. We also found that wet season NDVI is a positive predictor of sifaka densities, which will aid managers in prioritizing conservation actions in this region using widely available remotely sensed data.