China City-Level Thermal Variability and Surface Ozone Analysis Archive: Historical Components and Near-Future Attribution

This repository provides processed data, intermediate products, statistical model outputs, and figure-ready files supporting an analysis of short-term thermal variability and surface ozone responses across 364 Chinese cities. The historical component covers 2000–2024 and includes harmonized daily maximum temperature (Tmax), maximum daily 8-hour average surface ozone (MDA8 O₃), selected meteorological covariates, and their trend, seasonal, and residual components derived using Seasonal-Trend decomposition using Loess (STL). The statistical outputs include full-year Peter–Clark Momentary Conditional Independence Plus (PCMCI+) conditional-dependence screening with a maximum lag of 7 days, corresponding sensitivity analyses for the warm season, alternative lag windows, and regional aggregation, and city-specific distributed lag nonlinear model (DLNM) estimates. The DLNM archive contains city-level model diagnostics, cross-basis coefficients and covariance matrices, national and seven-region multivariate meta-analysis results, exposure–lag–response summaries, and response artifacts used for future attribution. The near-future component covers 2030–2040 and includes STL-derived Tmax residual sequences from five downscaled CMIP6 models under SSP1-RCP2.6, SSP2-RCP4.5, SSP3-RCP7.0, and SSP5-RCP8.5. It also contains city-level counterfactual attribution results, ensemble summaries, exposure metrics, boundary-restriction sensitivity analyses, and raster and tabular data used to generate the future temperature-residual and ozone-increment maps. The future ozone results quantify response-based O₃ residual increments associated with positive Tmax residuals under fixed historical exposure–response relationships. They represent residual-scale statistical attribution rather than projections of total future ozone concentrations. Third-party source data are redistributed only where their original licenses permit; otherwise, source citations, access information, processing scripts, and derived data products are provided.Intermediate grid-cell outputs generated during the STL decomposition are stored in NetCDF format. Owing to their substantial data volume, these intermediate files are not included in the current Zenodo release. They can be reproduced using the archived processing scripts and source-data information or obtained from the corresponding author upon reasonable request.