Process-explicit models reveal pathway to extinction for woolly mammoth using pattern-oriented validation

Authors: Stuart C Brown, Sean Haythorne, Damien A Fordham.

This respository contains the R code and some example data (for 10 unique niche samples) to run the Mammoth model paleopop simulations.

In order to run the example simulations, paleopop v 1.0.0 must be installed. paleopop v 1.0.0 is packaged in this respository and the code checks to make sure it is installed. If a newer version is installed, it is reverted to v 1.0.0 due to code-breaking changes introduced in later versions.

The code has a number of package dependencies listed below:

• "doParallel", min version = "1.0.14"
• "foreach", min version = "1.4.4"
• "gdistance", min version = "1.2.2"
• "geosphere", min version = "1.5.7"
• "lhs", min version = "1.0"
• "metRology", min version = "0.9.28.1"
• "R6", min version = "2.4.0"
• "raster", min version = "2.8.19"
• "sf", min version = "0.7.7"

The script checks the above dependencies and advises if a newer version of a package is required. If using sf > 1.0.0 please set sf_use_s2 to FALSE.

The scripts are designed to run in parallel across n sessions - please set n accordingly. For a guide, using 95 cores on a workstation uses ~ 90GB RAM to run the example simulations. Example simulations run in approximately 3 minutes.

There are two scripts contained in the repository. They need to be run in the following order:

• mammoth_full.R
• mammoth_scenarios.R

The first script build, runs, and saves the output from a multi-simulation manager (see paleopop documentation for details). The second script, imports the multi-simulation manager and then runs three counter-factual scenarios each with differing rates of human harvest.

The second set of counter-factual simulations run in ~11 minutes across 95 cores.

Generally, the second script would only be run following model selection with, for example, Approximate Bayesian Computation [ABC]. This would ensure that the counter-factual scenarios are only run on the models that best fit observations. See the main manuscript for details on our ABC model selection process.

All outputs are saved in the results/ directory.

The second script also produces plots of temporal abundance for the example simulations, and creates maps of extirpation time and the difference in extirpation time for a scenario run with and without human harvesting.

Additional niche samples are available on reasonable request.

Additionally, scripts found in extras/ directory can be used to recreate the sampling effort map in the supporting information of the main manuscript.

Directory structure:

Mammoth/
  - Data/ # data necessary to build the models
  - extras/ # additional data to create maps of fossil sampling effort
  - k_cuts/ # niche samples required for the models
  - results/ # output directory for simulation results

This code is released under the following licence:

GNU GENERAL PUBLIC LICENSE Version 3, 29 June 2007

Copyright (C) 2007 Free Software Foundation, Inc. https://fsf.org/ Everyone is permitted to copy and distribute verbatim copies of this license document, but changing it is not allowed.