Data and code for: Nonuniform resizing of marine life under climate change

This record is the complete data-and-code archive for the paper below. It contains every analysis script, the raw and processed datasets, the fitted models, all result tables, and all figures needed to reproduce the study from beginning to end. The code is also mirrored on GitHub (linked below); this Zenodo deposit is the authoritative, citable version of record and additionally holds the large data and model files that do not fit on GitHub. Trindade-Santos, I., Webb, T. J., Heim, N. A., Knope, M. L., Monarrez, P. M., Payne, J. L., Finnegan, S., & McClain, C. R. (2026). Nonuniform resizing of marine life under climate change. Proceedings of the National Academy of Sciences, 123, e2606099123. https://doi.org/10.1073/pnas.2606099123 Study in brief We project future changes in mean body size across 23,329 species of marine molluscs from five classes (Bivalvia, Cephalopoda, Gastropoda, Polyplacophora, Scaphopoda) across ten global ocean basins, under IPCC Low (SSP1-1.9) and High (SSP5-8.5) emission scenarios. Body-size responses to sea-surface temperature, dissolved oxygen, and primary productivity (with depth and taxonomic family as additional predictors) are quantified with class-specific generalized additive models (GAMs) trained on 2,798,611 species-level occurrences — those with complete environmental data, drawn from the 3,029,688 occurrence records retrieved from OBIS — matched to contemporary conditions from Bio-ORACLE v3.0, and projected decade by decade from the 2000s to the 2090s. The central result is that projected resizing is nonuniform: its direction and magnitude differ markedly among classes and among basins. The GAM formula used for every class is log10(length) ~ s(temperature) + s(dissolved_oxygen) + s(log10(productivity)) + s(log10(depth)) + family, fitted in mgcv with REML and double-penalized component selection. What is included Primary data — the master occurrence dataset (2,798,611 records with complete environmental data, 23,329 species) at the native 0.05° resolution, its 0.5° and 1.0° re-aggregated versions, the ocean-basin shapefile, the class silhouettes, and the raw Bio-ORACLE v3.0 environmental layers. Fitted GAM models — cached model objects for the main analysis and for every sensitivity analysis, so figures and tables can be regenerated in minutes rather than re-fitting all GAMs (which takes hours). Prediction and summary tables — the projection outputs and all seventeen Supporting Information tables (Tables S1–S17) as individual CSV files. Figures — main-text Figures 1–3 and Supporting Information Figures S1–S17 as vector PDF, editable SVG, and high-resolution TIFF. Analysis code — the main Quarto pipeline plus the depth-gradient, grid-resolution, and occurrence-threshold sensitivity analyses, the Figure S3 script, and the coarse-resolution data-preparation script. A README documents every file and explains how to reproduce the study. How to cite Please cite both the paper and this archive. Paper: Trindade-Santos, I., Webb, T. J., Heim, N. A., Knope, M. L., Monarrez, P. M., Payne, J. L., Finnegan, S., & McClain, C. R. (2026). Nonuniform resizing of marine life under climate change. Proceedings of the National Academy of Sciences, 123, e2606099123. https://doi.org/10.1073/pnas.2606099123 Data and code archive (this record): Trindade-Santos, I., Webb, T., Heim, N., Knope, M., Monarrez, P., Payne, J., Finnegan, S., & McClain, C. (2026). Data and code for: Nonuniform resizing of marine life under climate change [Data set]. Zenodo. https://doi.org/10.5281/zenodo.20460201 Related resources Published article (PNAS): https://doi.org/10.1073/pnas.2606099123 Code repository (GitHub): https://github.com/IsaacPesca/Nonuniform-resizing-of-marine-life-under-climate-change Data sources and license This archive is released under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. Third-party data — OBIS (occurrences), Bio-ORACLE v3.0 (environmental layers), WoRMS (taxonomy), Marine Regions (ocean-basin boundaries), and the MOBS body-size database — remain subject to their own licenses and should be cited directly.