Inaccessible ledges as refuges for the natural vegetation of the high Andes

Questions: Have millennia of human land use fundamentally altered the vegetation of a large proportion of the high Andean puna biome, with natural vegetation now restricted to inaccessible areas? Can inaccessible ledges be used as surrogates to infer the potential natural vegetation (PNV) in heavily impacted areas of the puna ecosystem of the high Andes? Is there a difference in plant community composition and diversity between the potential natural puna vegetation, represented by areas inaccessible to grazing and burning, and the anthropogenically disturbed vegetation found on nearby, but accessible, slopes? Location: Abra Maalaga Private Conservation Area, Cusco, southern Peruvian Andes. Methods: Four study habitats were chosen that comprised ledges and slopes from within and outside of the conservation area. For each habitat, vegetation composition was recorded using eight to twelve 2 x 2-m(2) plots studied for species cover and abiotic variables. Results: Analysis of species richness using two-way ANOVAs with Tukey test found that plots from the three habitats inaccessible to anthropogenic disturbance exhibited similar richness levels, whereas plots accessible to grazing and anthropogenic burning had significantly higher species richness. Likewise, CCA separated out plots of the three habitats inaccessible to anthropogenic disturbance from the unconserved slope plots. Species indicator analyses found the three inaccessible habitats to share the largest number of indicator species, with none being shared by the accessible, unconserved slope habitat. The PNV, inferred from the inaccessible vegetation, comprises a mosaic of Polylepis pepei woodland and tussock grassland, dominated by Festuca aff. procera, Luzula gigantea, Valeriana mandoniana and Carex pichinchensis. Conclusions: As both the conserved and unconserved ledge habitats contain a vegetation that approaches that of the conserved slope, ledges can be taken as a surrogate to infer the PNV in heavily impacted areas where no conserved slopes are available. From preliminary data, the presumed PNV of the study area corresponds to a distinct vegetation assemblage including species previously unknown to science. Adjacent disturbed, accessible land contained a higher species diversity, with a flora that may have originated from localized, disturbed natural habitats.