Amazon-wide distribution patterns of understory plants: Effects of soils and changing climate

Résumé fourni par la source

Distributions of Amazonian plants are poorly known. Here, we analyze to what degree current Amazon-wide species distribution patterns reflect general aspects of the species’ environmental niches, and how distributions may be affected by climate change. Occurrence records from quantitative field surveys were combined with records from the Global Biodiversity Information Facility (GBIF) after verification of taxonomical identifications. The importance of 40 environmental predictors (ranging from bioclimatic to edaphic variables) was assessed for each species separately. Ensemble species distribution models (SDMs) were trained using current bioclimatic variables, and potential range shifts were then modelled by SDMs using bioclimatic variables from future climatic scenarios. Six different distributional types were identified on the basis of the shape and extent of the predicted distributional area and the most important environmental variables in the SDMs. In order of increasing area of occupancy, the types were: Guianan, Andean, southern, circum-Amazonian, bimodal, and widespread. The edaphic niche was an important predictor of distribution for all species, especially for the Andean and bimodal ones, which had high soil base cation optima. All widespread species had low base cation optima, and their most important predictor was climatic. We observed that, of the 18 species, 14 would significantly lose habitat under the medium and extreme future climate scenarios. We suggest that Amazon-wide plant distribution patterns reflect edaphic optima, with species adapted to cation-rich soils generally having the narrowest distributions and species adapted to poor soils, potentially the broadest. Projected climate change will reduce the suitable habitat area, especially for the core rainforest species, while species currently found in the drier southern Amazonia might be able to expand northwards. We identified six different distributional types for understory ferns (Guianan, Andean, southern, circum-Amazonian, bimodal, and widespread), and since species in each category shared a unique combination of niche properties in relation to soil nutrient availability and climate, these are likely relevant for other plants as well. Soil base cation concentration emerged as an important variable in all the species distribution models, highlighting the importance of reliable soil data for understanding Amazonian plant distributions. Most species (14 out of 18) were predicted to suffer a net loss of habitat by 2080, even with moderate climate change scenarios; 8 would lose more than 50% under the pessimistic scenario even without considering future deforestation. Species of the core rainforest area were predicted to be especially hard hit by climate change; the only species whose habitat was predicted to expand by more than 25% has a southern distribution that currently extends beyond Amazonia to drier biomes.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.