干旱区生态环境知识资源中心

Climate (drought, rainfall), geology (habitat availability), land use change (provision of artificial waterpoints, introduction of livestock), invasive species (competition, predation), and direct human intervention (lethal control of top-predators) have each been identified as processes driving the sustainability of threatened fauna populations. We used a systematic combination of empirical observational studies and experimental manipulations to comprehensively evaluate the effects of these process on a model endangered rodent, dusky hopping-mice (Notomys fuscus). We established a large manipulative experiment in arid Australia, and collected information from relative abundance indices, camera traps, GPS-collared dingoes (Canis familiaris) and dingo scats, along with a range of related environmental data (e.g. rainfall, habitat type, distance to artificial water etc.). We show that hopping-mice populations were most strongly influenced by geological and climatic effects of resource availability and rainfall, and not land use, invasive species, or human effects of livestock grazing, waterpoint provision, or the lethal control of dingoes. Hopping-mice distribution declined along a geological gradient of more to less available hopping-mice habitat (sand dunes), and their abundance was driven by rainfall. Hopping-mice populations fluctuated independent of livestock presence, artificial waterpoint availability or repeated lethal dingo control. Hopping-mice populations appear to be limited first by habitat availability, then by food availability, then by predation. Contemporary top-predator control practices (for protection of livestock) have little influence on hopping-mice behaviour or population dynamics. Given our inability to constrain the effects of predation across broad scales, management actions focusing on increasing available food and habitat (e.g. alteration of fire and herbivory) may have a greater chance of improving the conservation status of hopping-mice and other small mammals in arid areas. Our study also reaffirms the importance of using systematic and experimental approaches to detect true drivers of population distribution and dynamics where multiple potential drivers operate simultaneously. (C) 2017 Elsevier B.V. All rights reserved.