A Long-Term Summer Climate Livability Index (SCLI) Dataset for China from 2005 to 2024

To eliminate spatial inconsistencies, a bilinear interpolation method was uniformly applied to resample all multi-source datasets to the standard 0.25° grid.Methodologically, extreme events were defined using local dynamic thresholds based on the 1991–2020 climatological baseline. For instance, an effective heatwave was strictly defined as a sustained period of at least three consecutive days during which Tmax exceeded the local 90th percentile (P90) and the UTCI simultaneously surpassed 26°C. The six core sub-indicators were then normalized into suitability scores (0–1) using non-linear scaling approaches, where inverted S-shaped penalty functions were specifically applied to climate hazards to capture capacity thresholds.

The final annual SCLI was synthesized using a hierarchical weighting mechanism combining data-driven Principal Component Analysis (PCA) weights and theoretically informed macro weights. Each annual file is provided in the standard Network Common Data Form (.nc). Within each file, users can access the overall SCLI variable alongside the normalized scores of the six sub-indicators (e.g., score_utci, score_tp, score_ndvi, score_day_hw, score_extreme_precip, score_dem) to facilitate spatiotemporal attribution analysis.The standard netCDF format ensures broad compatibility with common data analysis environments (such as Python's xarray, R, and MATLAB) and spatial analysis software (including QGIS and ArcGIS).

When utilizing this dataset, several limitations should be acknowledged. First, because the SCLI is calculated using ERA5 and satellite-based data, any systematic biases present in these upstream products will propagate into the final suitability scores. Second, the 0.25° spatial resolution is designed for large- to meso-scale regional assessments; it cannot resolve micro-scale urban features such as block-level heat islands or localized topographic variations.

Finally, the SCLI reflects a generalized environmental baseline and does not account for the varying degrees of vulnerability across specific demographic groups.