Diazgranados, Mauricio , Barber, Janet .
Effects of climate change on the future distribution of frailejones (Espeletiinae Cuatrec., Asteraceae) in the Tropical Andes.
The grasslands above tree line in the tropical Andes, called paramos, are characterized by the presence of distinctive plants with stemmed rosettes known as frailejones. The dramatic radiation of the 140+ species of frailejones (subtribe Espeletiinae Cuatrec., Asteraceae) provides a model system to study speciation mechanisms and biogeographic patterns in sky islands. Frailejones are widely distributed and locally abundant in the paramos of Colombia, Venezuela and Ecuador and the preservation of this unique ecosystem is tightly linked to the plants' presence and persistence. Like the paramos themselves, which are shrinking rapidly, frailejones are highly threatened by climate change, mining, agriculture and livestock. Our goal was to estimate potential changes in species' geographic distributions under different scenarios of climate change. A comprehensive database containing 5094 geo-referenced plants was compiled. Predictors for modeling were obtained from WorldClim and included 19 bioclimatic variables plus elevation. In total, 23 Global Circulation Models (GCM) were used with three SRES scenarios (A1B, A2A and B2A) for the years 1990 to 2080. Species distribution models (SDMs) were estimated for a total of 133 species with MAXENT v3.3.3.k. After initial tests, auto-correlated predictors and variables of low contribution to the SDMs were excluded. The algorithm was run with 25% of the data for testing the models. SDMs for current conditions were replicated 88 times. We also performed 100 replications with cross-validation. Potential areas of distribution were thresholded using the 'maximum training sensitivity plus specificity criterion'. Fifteen landscape metrics were estimated using FRAGSTATS 3.4. The credibility of the SDMs was assessed based on ROC curves, AUC values, tests of normality, coefficient of variation of the mean thresholded area between replications, and variability between cross-validation outputs. SDMs for 60 species were retained. In most of the geographic range, isothermality was the most important predictor for modeling. Results showed that there are significant differences between predictions from different GCMs and SRES scenarios. Most of the species are predicted to have a reduction in suitable area, with fragmentation and simplification (or extinction) of their populations. Species at the extremes of the elevation gradient will suffer the most drastic effects, either increasing or decreasing their suitable areas, with the first extinctions predicted to occur by 2020. The most affected geographic areas and potential migration paths were also identified.
Log in to add this item to your schedule
The Espeletia Project
1 - Saint Louis University & Missouri Botanical Garden, Biology Department, 3507 Laclede Ave., St. Louis, MO, 63103, USA
2 - Saint Louis University, Biology Department, 3507 Laclede Avenue, St. Louis, MO, 63103, USA
species distribution modeling
ecological niche modeling
Presentation Type: Oral Paper:Papers for Sections
Location: Union B/Hyatt
Date: Tuesday, July 10th, 2012
Time: 9:30 AM