Iberian Bombus extirpation projection at HEALPix nside=128 — full GLMM refit on the DestinE Climate DT substrate. Methodological extension of
weatherxbiodiversity-projection(HEALPix nside=64). Reference paper: Soroye et al. (2020)
This repository fits Soroye et al.'s (2020) thermal-niche extirpation GLMM at HEALPix nside=128 (~46 km) — the native pixelisation of DestinE Climate DT IFS-NEMO standard — for Iberian Bombus, then projects to SSP3-7.0.
It is a sibling resolution-doubling extension of weatherxbiodiversity-projection (nside=64). The combined results across substrates feed the methodological diagnostic in weatherxbiodiversity-substrate-sensitivity.
Headline result — Tier 1¶
![Tier-1 GLMM coefficient summary at HEALPix nside=128 — sc_TEI_delta = +0.347, 95% HDI [+0.139, +0.533], substrate-robust against the canonical CEA replication](/weatherxbiodiversity-projection-nside128/build/main_result_healpix-721a71d4868ef131f576cf09df17f067.png)
The GLMM coefficient on sc_TEI_delta at HEALPix nside=128 is +0.347, 95% HDI [+0.139, +0.533] — within ±30% of the original CEA replication (+0.479) and the nside=64 replication (+0.454). The TEI-based extirpation mechanism is substrate-robust.
Headline result — Tier 2¶

Tier-2 projects the substrate-matched GLMM forward to DestinE Climate DT SSP3-7.0 at the same native pixelisation it was fit on (no parent-aggregation deviation). Per-species rankings should be cross-checked against the substrate-sensitivity diagnostic (sibling repo).
Why this repo exists¶
The canonical replication (weatherxbiodiversity-projection) fits the GLMM at HEALPix nside=64 and projects DestinE data aggregated to that grid. This sibling repo refits the entire GLMM at nside=128, the native DestinE substrate, so that:
Tier 1 estimates the GLMM coefficients at three substrates (CEA, nside=64, nside=128) — the substrate-robustness check
Tier 2 projects without any cross-substrate aggregation step
The combined output supports the cross-substrate diagnostic in
weatherxbiodiversity-substrate-sensitivity
Quick start¶
git clone https://github.com/annefou/weatherxbiodiversity-projection-nside128.git
cd weatherxbiodiversity-projection-nside128
mamba env create -f environment.yml
mamba activate weatherxbiodiversity-projection-nside128
snakemake --cores 1 # Tier 1 (HEALPix nside=128 GLMM refit)
snakemake --cores 1 tier2 # Tier 2 (DestinE) — needs DestinE credentialsPipeline structure¶
Tier 1 — HEALPix nside=128 GLMM refit (notebooks 01, 02h–04h)
Tier 2 — DestinE Climate DT SSP3-7.0 projection (notebooks 05–08)
Citation¶
If you use this work, please cite:
This software:
CITATION.cff→ DOI Fouilloux (2026).The original paper: Soroye et al. (2020).
The canonical nside=64 sibling and the substrate-sensitivity diagnostic via their own DOIs (see CITATION.cff
references).
- Soroye, P., Newbold, T., & Kerr, J. (2020). Climate change contributes to widespread declines among bumble bees across continents. Science, 367(6478), 685–688. 10.1126/science.aax8591
- Fouilloux, A. (2026). weatherxbiodiversity-projection-nside128. Zenodo. 10.5281/ZENODO.20113780