Climate plays a major role in shaping the distribution of plant species. In the central Andes, heterogeneous geography and harsh environmental conditions restrict plant species distributions. The abrupt geography, however, includes valleys formations where comparatively warmer conditions are found, along with associated flora. The genus Polylepis (Rosaceae) forms some of the highest elevation forests in the world (~4,000m). We studied two species of Polylepis to examine and compare the climatic factors influencing their distribution, and to assess the effects of geography and climate on the niche of the species. One of the species, P. pepei, has a small range and inhabits exclusively cloud forests. In contrast, P. tomentella is commonly distributed in variety of ecosystems throughout its range (e.g., altiplano, dry valleys, and tucuman forest). We developed niche models for P. pepei and P. tomentella using the Maxent algorithm combining geo-referenced data with climate layers. We derived the effects of climate-envelop and geography for each species using a multivariate approach.
Results/Conclusions
Results showed that the climate-envelop of each species for the most part does not overlap. The ecological niche of P. pepei is much smaller than that of P. tomentella. In both species, maps of potential distributions over-predicted their current distribution. Reconstruction of paleovegetation concur that at least up to 600 years ago, Polylepis spp. had a much broader distribution in the high Andes. Spatial correlation analysis showed significant correlations between the ecological niche of the species and geography, suggesting that the niche model has some error due to spatial dependency. Our results indicate that 1) P. tomentella is a more plastic species than P. pepei, allowing it to inhabit a larger variety of ecosystems, and 2) both geography and climate play important roles in the distribution of Polylepis. Niche modeling also suggests that Polylepis distribution is influenced by other environmental factors that deserve to be identified and included in future prediction models.
