Data and code from: Functional trait modules link climatic niche geometry to projected range change in global butterflies

Guo, Facheng 1 ; McKirdy, Simon J.2 3; Gao, Lingling4; Gao, Guizhen1

Published Aug 05, 2026 on Dryad. https://doi.org/10.5061/dryad.wh70rxx4b

Data files

Aug 05, 2026 version files 14.73 MB

FE-2026-00186_Dryad_missing_values_revision_20260731_202634.zip

14.73 MB
README.md

3.44 KB

Abstract

This dataset supports the functional trait-space PCA/TPD and generalised additive model (GAM) components of the study "Functional trait modules link climatic niche geometry to projected range change in global butterflies". The archive contains an analysis-ready seven-trait PCA matrix and species-family mapping; a derived species-level table containing species, family, PC1, PC2 and the GAM response E; relative-path R scripts for PCA/TPD, A50/A99 summaries and pooled and family GAMs; compact prediction and core-edge outputs; and file-level metadata, a manifest and SHA-256 checksums. E is a derived response from the study's upstream species-distribution-model workflow. Raw occurrence records, climate rasters, SDM objects, phylogenetic files, taxonomic reconciliation, imputation workflows, climatic-niche analyses, and SEM workflows are outside the scope of this deliberately minimal archive. The package is intended to support inspection and reproduction of the archived PCA/TPD and GAM components and contains no human-subject information.

Manuscript: FE-2026-00186.R3
Status: AUTHOR_APPROVED_FINAL

Dryad file

FE-2026-00186_Dryad_missing_values_revision_20260731_202634.zip: compressed archive containing the data, code, outputs, documentation, manifest and SHA-256 checksums described below.

This archive contains the analysis-ready data, R scripts and compact outputs for the functional trait-space PCA/TPD and GAM components of the study Functional trait modules link climatic niche geometry to projected range change in global butterflies.

The archive includes:

the complete analysis-ready seven-trait PCA matrix and species-family mapping;
the complete species-level GAM input containing species, family, PC1, PC2 and the response E;
relative-path R scripts for PCA, trait-space A50/A99 summaries, pooled and family GAMs, prediction grids, and core/edge summaries;
compact CSV outputs and file-level metadata.
Each major directory also contains a short README describing the files within that directory and any directory-specific conventions.

The closest recovered raw trait source table is not included in this revision because it is not read by the archived PCA or GAM scripts. Raw occurrence records, GBIF downloads, climate rasters, species-distribution-model objects, phylogenetic files, taxonomic reconciliation, imputation workflows, climatic-niche analyses, SEM workflows, manuscript files and internal audit materials are also outside this archive's scope.

E is the derived species-level response used by the archived GAM scripts. Its upstream occurrence, climate and SDM workflow is outside this minimal archive.

Directory guide

data/analysis_ready_traits/: PCA matrix and species-family mapping.
data/derived_GAM_input/: complete GAM input.
code/: relative-path PCA/TPD and GAM scripts.
outputs/: compact tables written by the archived scripts.
metadata/: data dictionary, source-role metadata, missing-value definitions, scope and software list.
docs/: reproduction instructions and inclusion/exclusion statement.

Missing values and empty cells

The three published analysis input tables contain no empty cells and no missing values. The archived PCA and GAM result tables also contain no empty cells. In outputs/GAM/pooled_GAM_summary.csv, the family field is recorded as Not_applicable because the pooled model combines records across all butterfly families. A value of 0 is an observed or encoded zero and must not be interpreted as missing.

No values were replaced with estimates, means, zeros or other inferred values during this curation revision. File- and column-specific information is provided in metadata/MISSING_VALUE_DEFINITIONS.csv and metadata/DATA_DICTIONARY.csv.

Run

From the archive root:

Rscript code/run_PCA_and_GAM.R

The scripts check dependencies and stop with a clear message if a package is missing. They never install packages automatically. No analysis formula, variable selection, imputation rule or model parameter was changed during this curation revision.

Public release

The authors confirm that the files in this revised archive may be redistributed through Dryad under the CC0 1.0 Universal waiver. The incorporated data roles and provenance are identified in the accompanying metadata.