Pollinator redundancy gives plants little insurance against warming because partners share a climate: code and derived data

This archive contains the code and derived data for the article "Pollinator redundancy gives plants little insurance against warming because partners share a climate" (submitted to Ecology Letters). We combined 1,508 standardised plant–pollinator networks from EuPPollNet v1.3.0 with the realised thermal niches of 1,749 wild pollinator taxa, estimated from GBIF occurrence records (2000–2025) and ERA5-Land climate data. The analyses quantify how many wild-pollinator visits come from taxa beyond their realised warm limit under recent climate and under 2 and 3 °C of warming, test whether partner redundancy insures plants against this exposure using independent-risk and within-network null models, validate the thermal limits against independent network surveys, and estimate changes in pollinator detections across thermal ranges between 2000–2010 and 2016–2025. Contents- code/pipeline contains all Python scripts needed to reproduce the results, in pipeline order, including upstream data preparation.- code/run contains the Windows PowerShell runner that calls each step.- code/gee contains the Google Earth Engine scripts used to extract ERA5-Land climate data.- code/archive_exploratory and code/archive_superseded hold earlier exploratory analyses and earlier script versions, kept for transparency and not needed to reproduce the results.- data/derived contains the derived tables (thermal niches, interaction-level thermal safety margins, plant and network exposure, validation, sensitivity analyses and insurance null models).- data/summaries contains the status and summary files of every analysis stage, from which the numbers in the article are taken.- provenance lists the GBIF download DOIs and data sources. Raw data are not redistributed. Network data are available from EuPPollNet, occurrence data from GBIF through the download DOIs listed in provenance/GBIF_DOWNLOADS.md, and climate data from ERA5-Land (Copernicus Climate Change Service). The figures of the article can be regenerated with the scripts in code/pipeline. See README.md for details and MANIFEST.csv for file checksums. Code is released under the MIT licence and derived data under CC BY 4.0.