Heat stress as an emerging constraint on global dairy systems: CMIP6-based projections and national-scale milk loss exposure

Title Global dairy heat-stress exposure dataset (THI) and processing code based on bias-corrected CMIP6 projections Description This repository provides the data and code used to compute heat-stress exposure indicators for dairy systems based on the Temperature–Humidity Index (THI). The dataset consists of monthly THI-based metrics derived from bias-corrected CMIP6 climate projections (NEX-GDDP-CMIP6) at 0.25° spatial resolution. The indicators are computed from daily maximum air temperature and relative humidity and aggregated to facilitate analysis of long-term exposure patterns. The repository also includes the full processing workflow used to compute THI and derive exceedance metrics, ensuring reproducibility of the results presented in the associated study. Contents THI datasets Monthly THI indicators (NetCDF format) Derived from daily temperature and relative humidity Spatial resolution: 0.25° Temporal coverage: Historical baseline (1991–2020) Mid-century (2041–2070) End-century (2071–2100) Scenarios: SSP1–2.6 SSP2–4.5 SSP5–8.5 Models: GFDL-ESM4 IPSL-CM6A-LR MPI-ESM1-2-HR MRI-ESM2-0 UKESM1-0-LL Code Scripts for THI computation from daily climate variables Processing of exceedance metrics (e.g. THI ≥ 72) Aggregation to monthly indicators Data processing workflow (NetCDF / CDO / Python or R depending on your stack) Methods summary The Temperature–Humidity Index (THI) is computed from daily maximum temperature and mean relative humidity following a standard formulation used in dairy heat-stress studies. Exceedance metrics (e.g. number of days with THI ≥ 72) are derived at daily resolution and aggregated to monthly and annual scales. The dataset represents climate-driven exposure only and does not include livestock data, response functions, or production estimates. Important notes This dataset provides climate-based indicators of heat-stress exposure, not direct estimates of milk production losses. Livestock distribution data (GLW3) and loss-equivalent metrics used in the manuscript are not included in this repository. Users should interpret THI thresholds as comparative indicators, not universal physiological thresholds. Reproducibility All processing steps required to reproduce the dataset are provided in the repository, including: THI computation threshold exceedance calculations temporal aggregation Related publication This dataset supports the manuscript: Heat stress as an emerging constraint on global dairy systems: high-resolution CMIP6 projections and national-scale milk loss exposure