Associated paper: Adapting to a shifting planet: The future of Drosera species amidst global challenges and conservation imperatives
Journal: Anthropocene, 49, 100466
Article DOI: https://doi.org/10.1016/j.ancene.2025.100466
Dataset DOI: https://doi.org/10.5281/zenodo.20631938

This repository contains the data, metadata, and source code associated with the paper Adapting to a shifting planet: The future of Drosera species amidst global challenges and conservation imperatives.

The study evaluates current and future habitat suitability for 39 South American Drosera species using species distribution models, bioclimatic predictors, MaxEnt, and the KUENM workflow.

The paper describes projections for 2050 and 2070 under SSP5–8.5 using HadGEM2-AO and MRI-CGCM3 general circulation models.

Olivares-Pinto, U., Santiago Lopes, J. C., Ruiz-Aguilar, C., Oki, Y., & Fernandes, G. W. (2025). Adapting to a shifting planet: The future of Drosera species amidst global challenges and conservation imperatives. Anthropocene, 49, 100466. https://doi.org/10.1016/j.ancene.2025.100466

This repository is organized for reproducibility of the species-level ecological niche modeling workflow.

It includes standardized occurrence CSV files, metadata, and generic scripts that avoid hardcoded local paths or species-specific filenames.

Drosera_Species_Climate_Impact/
├── README.md
├── PAPER.md
├── LICENSE.md
├── CITATION.cff
├── data/
│   ├── README.md
│   ├── species_metadata.csv
│   └── species/
│       ├── 01_intermedia/
│       │   ├── occurrences_clean.csv
│       │   ├── occurrences_independent.csv
│       │   ├── precomputed/
│       │   └── M_variables/
│       ├── 02_communis/
│       └── ...
├── scripts/
│   ├── run_kuenm_species.R
│   ├── run_all_species.sh
│   └── prepare_species_csvs.py

Each species folder uses the following naming convention:

occurrences_clean.csv          # cleaned occurrence records
occurrences_independent.csv    # independent occurrence records for final evaluation
background_points.csv          # rare-species background points, when available
precomputed/occurrences_*.csv  # uploaded split files preserved for traceability
M_variables/                   # selected environmental variable sets, e.g. Set_1/

The original uploaded split files (drosera_joint.csv, drosera_train.csv, drosera_test.csv) were preserved under precomputed/.

They are not used by default because the generic script regenerates paper-aligned 75% calibration / 25% testing splits from occurrences_clean.csv.

The generic script follows the paper-level workflow:

1. Read cleaned occurrence records.
2. Split occurrences into 75% calibration and 25% testing records.
3. Use selected calibration-area variables from M_variables/.
4. Calibrate candidate MaxEnt models using KUENM.
5. Evaluate candidate models with partial ROC, omission rate, and AICc.
6. Generate final models with bootstrap replicates and logistic output.
7. Project models to current and future scenarios when G_variables/ is available.
8. Evaluate final models with independent occurrence records.
9. Summarize projections, estimate projection changes, and run MOP extrapolation-risk analyses.

Example:

Rscript scripts/run_kuenm_species.R \
  --species_dir="data/species/03_montana" \
  --species_name="Drosera montana" \
  --species_code="sp03" \
  --maxent_path="data/species/03_montana" \
  --replicates=500

For a quick test that only validates files and does not run MaxEnt/KUENM:

Rscript scripts/run_kuenm_species.R \
  --species_dir="data/species/03_montana" \
  --validate_only=true

To regenerate only the occurrence split files in paper-aligned 75/25 format:

Rscript scripts/run_kuenm_species.R \
  --species_dir="data/species/03_montana" \
  --split_occurrences=true \
  --overwrite_split=true \
  --calibrate=false \
  --evaluate_candidates=false \
  --run_final_models=false \
  --run_final_evaluation=false \
  --run_summaries=false \
  --run_mop=false

bash scripts/run_all_species.sh --validate_only=true

or, once M_variables/, G_variables/, and maxent.jar are available:

bash scripts/run_all_species.sh --replicates=500

• R
• Java Runtime Environment
• MaxEnt 3.4.4 or compatible maxent.jar
• KUENM R package
• Optional: devtools to install KUENM

Install KUENM with:

install.packages("devtools")
devtools::install_github("marlonecobos/kuenm")

The complete research data archive is available through Zenodo:

Zenodo DOI: 10.5281/zenodo.20631938
Zenodo record: https://doi.org/10.5281/zenodo.20631938

The source code, metadata, validation notes, and reproducible KUENM/MaxEnt scripts are maintained in this GitHub repository:

Repository: https://github.com/HpcDataLab/AdaptingToaShiftingPlanet

If you use this repository or dataset, cite the associated paper and the archived Zenodo dataset DOI.

Olivares-Pinto, U., Santiago Lopes, J. C., Ruiz-Aguilar, C., Oki, Y., & Fernandes, G. W. (2025). Adapting to a shifting planet: The future of Drosera species amidst global challenges and conservation imperatives. Anthropocene, 49, 100466. https://doi.org/10.1016/j.ancene.2025.100466