Sorry, you need to enable JavaScript to visit this website.

Workshop: An introduction to the CREST r-package to reconstruct climate from palaeoecological datasets

San Carlos de Bariloche, Río Negro, Argentina
Contact person
Fernanda Charqueno Celis
E-Mail address
Working groups

WORKSHOP PAGES-IPA: An introduction to the CREST r-package to reconstruct climate from palaeoecological datasets
By Dr. Manuel Chevalier


Date: Saturday, 3 December 2022
Time: 9:00 – 14:00 (UTC -3)
Venue: TBC


The CREST (Climate REconstruction SofTware) technique is a flexible probabilistic tool for reconstructing climate
from palaeoecological datasets (e.g. pollen, rodents, forams)1 that has been successfully employed to reconstruct multiple climate parameters (e.g. temperature, precipitation, moisture index) from different regions (e.g. South Africa, Colombia, Turkey). 

In contrast to the more commonly-used reconstruction techniques, such as the Modern Analogues or all the regression-based techniques that require extensive collections of modern proxy samples to produce reliable reconstructions, CREST uses modern proxy occurrence data to estimate the proxy-climate relationships (e.g. using sources like the GBIF database (Global Biodiversity Information Facility). Capitalizing on these abundant and freely available data, CREST is thus more broadly applicable, making it an effective tool to produce climate quantifications in regions where these remain unmistakably absent2.

To facilitate its use by the global community, CREST is now encapsulated in a dedicated R package crestr that includes calibration data for four terrestrial and two marine proxies3

In particular, the depth and richness of the plant calibration dataset enable its use on virtually any pollen record on the planet4.

About the workshop

To introduce the package and provide guidelines to produce high-quality climate reconstructions, an in-person workshop will be organized. 

In detail, this workshop aims to: 
1) to describe the fundamental theory that underpins the model;
2) demonstrate how the different R functions of the package can be parameterised to best model the data; and 
3) help participants use their datasets and produce preliminary reconstructions. Basic knowledge in R is preferable but not necessary to participate in the workshop. 

Details about the method and package can be found here:

This in-person one-day workshop, organized in collaboration by PAGES-ECN and IPA-ECN, is intended to reach ECRs interested in quantitative paleoclimate reconstructions just after the IAL-IPA meeting which will be carried out in Bariloche. 

The workshop will be held on Saturday, 3 December with a duration of 5 hrs (9:00 – 14:00), with a theory and practice part for the participants. 


Deadline for pre-registration: 16 September

Register here:

More information

Fernanda Charqueno Celis:

> View pdf for more details

1. Chevalier, M., Cheddadi, R. & Chase, B. M. CREST (Climate REconstruction SofTware): a probability density function (PDF)-based quantitative climate reconstruction method.
Clim. Past 10, 2081–2098 (2014). 
2. Chevalier, M. et al. Pollen-based climate reconstruction techniques for late Quaternary studies. Earth-Sci. Rev. 210, 103384 (2020). 
3. Chevalier, M. crestr: an R package to perform probabilistic climate reconstructions from palaeoecological datasets. Clim. Past 18, 821–844 (2022). 
4. Chevalier, M. Enabling possibilities to quantify past climate from fossil assemblages at a global scale. Glob. Planet. Change 175, 27–35 (2019)