Daily Human Thermal Index Dataset for India (HiTIC-India) at 1-km Spatial Resolution (2003–2020).

This repository provides access to the Human Thermal Index Collection for India (HiTIC-India), a high-resolution (1 km), gridded dataset of twelve human-perceived temperature (HPT) indices covering the period 2003–2020. The dataset is developed to support research on thermal stress, climate–health interactions, labour productivity, urban heat exposure, and climate adaptation across India. HiTIC-India is generated by downscaling ERA5-derived meteorological variables using a Light Gradient Boosting Machine (LightGBM) framework and multiple high-resolution covariates.Dataset Overview Spatial resolution: 1 km × 1 km Temporal coverage: 2003–2020 Temporal resolution: Monthly mean (complete archive) Daily (sample files only) Spatial extent: India Coordinate reference system: WGS 1984 Geographic (EPSG:4326) Data format: NetCDF (.nc)The dataset contains the following 12 Human Perceived Temperature indices: Indoor Apparent Temperature (ATin), Outdoor Apparent Temperature (ATout), Discomfort Index (DI), Effective Temperature (ET), Heat Index (HI), Humidex (HMI), Modified Discomfort Index (MDI), Net Effective Temperature (NET), Surface Air Temperature (2 m) (SAT), Simplified Wet-Bulb Globe Temperature (SWBGT), Wet-Bulb Temperature (WBT), Wind Chill Temperature (WCT).Repository Contents.Monthly Mean HPT Datasets (Complete Archive) This repository hosts monthly mean datasets for all twelve HPT indices, aggregated from daily predictions for the period 2003–2020. File naming convention- _monthly_mean_2003_2020.nc.Examples: ATin_monthly_mean_2003_2020.nc ET_monthly_mean_2003_2020.nc SWBGT_monthly_mean_2003_2020.nc Each file contains gridded monthly mean values at 1 km resolution covering the entire study region.The complete monthly archives occupies approximately 24 GB.Sample Daily Predicted Files Due to the very large data volume of the daily products, only a limited set of representative daily files is provided here for transparency and reproducibility. File naming convention-Predicted_Year__Day_DOY.nc.Example: Predicted_2003_ATin_Day_001.nc The complete daily dataset (all indices, all years) is approximately 7 TB and cannot be hosted on GitHub.Full Daily Dataset Access The complete daily HPT datasets at 1 km resolution will be made publicly available through the ISRO–NICES data portal prior to publication. Until then, the full daily products are available from the authors upon reasonable request.Methodology Summary HPT indices are first computed from ERA5 reanalysis meteorological variables at coarse spatial resolution. These indices are subsequently downscaled to 1 km spatial resolution using a Light Gradient Boosting Machine (LightGBM) framework. The downscaling incorporates high-resolution covariates that capture surface, atmospheric, and topographic controls on near-surface thermal conditions, including land surface temperature, precipitable water vapour, elevation, slope, aspect, and population density. Model performance is evaluated using standard statistical metrics, including the coefficient of determination (R²), root-mean-square error (RMSE), and mean absolute error (MAE). A detailed description of the methodology and model configuration is provided in the associated manuscript.Contact For questions, feedback, or access to the full daily datasets, please contact: Dr. Saket Dubey Email: [email protected]