cRacle: R Tools for Estimating Climate from Vegetation

Abstract

Premise of the studyThe Climate Reconstruction Analysis using Coexistence Likelihood Estimation (CRACLE) method for the estimation of climate from vegetation is a robust set of modeling tools for estimating climate and paleoclimate that makes use of large repositories of biodiversity data and open-access R software. MethodsHere we implement a new R library for the estimation of climate from vegetation. The cRacle library implements functions for data access, aggregation, and models to estimate climate from plant community composition. cRacle is modular and features many best-practice features. ResultsPerformance tests using modern vegetation survey data from North and South America shows that CRACLE outperforms alternative methods. CRACLE estimates of mean annual temperature (MAT) are usually within 1{degrees}C of the actual when optimal model parameters are used. Generalized Boosted Regression (GBR) model correction is also shown here to improve on CRACLE models by reducing bias. DiscussionCRACLE provides accurate estimates of climate from modern plant communities. Non-parametric CRACLE modeling coupled to GBR model correction produces the most accurate results to date. The cRacle R library streamlines the estimation of climate from plant community data, and will make this modeling more accessible to a wider range of users.