Microbial growth rate is a stronger predictor of soil organic carbon than carbon use efficiency

This repository contains the full R analysis pipeline for the manuscript “Microbial growth rate better predicts soil organic carbon than carbon use efficiency”. The code includes all data processing, statistical analyses, and figure generation used in the study. The script integrates observational ¹⁸O–H₂O microbial-growth data with outputs from four microbial-explicit land surface models (ORCHIDEE-CENTURY, ORCHIDEE-MIMICS, CABLE-CASA, and JULES-RothC). It performs: preprocessing of growth, respiration, CUE, and SOC datasets computation of microbial physiological metrics (specific growth rate, turnover time, etc.) analysis of temperature sensitivities (incubation temperature and MAT) global mapping of microbial growth and CUE fields correlation assessments among modeled and observed variables Random Forest analyses of SOC drivers across models partial dependence analyses for microbial growth and CUE visualization of all analytical figures (Figures 2–5 and Supplementary Figures) The code is structured modularly (Files 01–10) and can be run sequentially or via a master script. All figure panels in the published manuscript are generated directly from these scripts using R (v4.3.0) and the ggplot2 ecosystem. No third-party or proprietary code is included. All scripts rely solely on publicly available R packages, and parameter settings are fully specified to ensure reproducibility.