PaRaMetriC Atmospheric Spectral Tool for Irradiance Calculation using Hourly ERA5 data
- About the Project
- Available Datasets
- Data Structure
- Getting Started
- How to Cite
- References
- Authors
- License
- History
PaRaMetriC is a metrological framework for passive radiative cooling technologies developed as a Joint Research Project within the European Partnership on Metrology Programme.
This repository contains the software used to simulate and evaluate downwelling longwave irradiance using atmospheric states derived from ERA5 reanalysis (Hersbach, 2023) and computed via RRTM_LW (Mlawer, 1997).
Fluxes are calculated over 16 contiguous longwave (infrared) spectral bands from 3–1000 μm wavelength.
ERA5 data points are defined on a regular latitude–longitude grid at 0.25° resolution and 37 fixed pressure levels.
The output fluxes are defined over the time, latitude, longitude, and lw_bands dimensions.
Tip
If you're unfamiliar with NetCDF format, we recommend NASA’s Panoply to explore, plot, and export the data.
- Denver, USA – 9 points (3 lat × 3 lon), 12 months, hourly
- Las Vegas, USA – 9 points (3 lat × 3 lon), 12 months, hourly
- Madrid, Spain – 20 points (4 lat × 5 lon), 12 months, hourly
- Paris, France – 4 points (2 lat × 2 lon), 12 months, hourly
- Rome, Italy – 4 points (2 lat × 2 lon), 12 months, hourly
- Turin, Italy – 4 points (2 lat × 2 lon), 12 months, hourly
- Singapore – 9 points (3 lat × 3 lon), 12 months, hourly
- Tokyo, Japan – 9 points (3 lat × 3 lon), 12 months, hourly
- Las Vegas, USA – 9 points (3 lat × 3 lon), 3 months (JJA), hourly
- Madrid, Spain – 20 points (4 lat × 5 lon), 3 months (JJA), hourly
- Riyadh, Saudi Arabia – 9 points (3 lat × 3 lon), 3 months (JJA), hourly
- Turin, Italy – 4 points (2 lat × 2 lon), 3 months (JJA), hourly
- France – 21 lat × 21 lon (5.25° × 5.25°), 2019–2023, 6-hourly
- Spain – 21 lat × 21 lon (5.25° × 5.25°), 2019–2023, 6-hourly
- Lleida, Spain – 4 points (2 lat × 2 lon), 31 July & 1 August, 1989–2023, hourly
- Sesto Fiorentino, Italy – 12 points (3 lat × 4 lon), 31 July & 1 August, 1989–2023, hourly
Each NetCDF4 file contains the following calculated variables:
sd(time, latitude, longitude, lw_bands)– RRTM-calculated surface downward longwave radiation flux (W·m⁻²)su(time, latitude, longitude, lw_bands)– Surface upward longwave radiation flux (W·m⁻²)sn(time, latitude, longitude, lw_bands)– Surface net longwave radiation flux (W·m⁻²)tu(time, latitude, longitude, lw_bands)– TOA upward longwave radiation flux (W·m⁻²)r(time, latitude, longitude)– Relative humidity calculated from 2 m temperature and dewpoint (%)
Note
- Band 0 contains total infrared flux; bands 1–16 represent spectral subdivisions.
- Band limits are stored in
lw_band_limits(cm⁻¹).
The following fields are copied directly from ERA5:
t2m– 2 m temperatureskt– Skin temperaturecbh– Cloud base heighttcc– Total cloud cover (ascloud_area_fraction)tcwv– Total column vertically integrated water vaporu10,v10– 10 m wind componentsstl3,stl4– Soil temperatures at levels 3 and 4avg_sdlwrf,avg_sdlwrfcs– Time-averaged surface downward LW radiation flux (all-sky / clear-sky)avg_sdswrf,avg_sdswrfcs– Time-averaged surface downward SW radiation fluxavg_snlwrf,avg_snlwrfcs– Time-averaged surface net LW radiation fluxavg_snswrf,avg_snswrfcs– Time-averaged surface net SW radiation fluxavg_tnlwrf,avg_tnlwrfcs– Time-averaged TOA net LW radiation flux
Warning
- ERA5 fluxes are accumulated over one hour and normalized by 3600 s. We treat these as instantaneous values centered at
t - 0.5 h. - ERA5 fluxes correspond to total LW radiation and should be compared to band 0 values from
RRTM_LW. - NaN values may appear over sea regions or where RRTM fails (e.g., north-west corner of the France dataset); further investigation is ongoing.
For a quick start, check out the interactive Colab notebook:
Notebook: RRTM_LW_ERA5_workflow.ipynb
This notebook guides you through:
- Installing the required packages and dependencies
- Loading pre-fetched ERA5 data for an illustrative TMY case
- Running the
RRTM_LWmodel - Producing and plotting longwave irradiance output
- Setting your CDS API key and preparing user-defined configurations
No local installation needed — everything runs in the cloud.
Otherwise, you can clone the repository locally with:
git clone https://github.com/21grd03-parametric/pastiche.git
cd pastiche
python3 -m pip install -r requirements.txtERA5 downloads require a configured CDS API account. See the Copernicus Climate Data Store API setup guide for instructions.
And run a full simulation from a configuration file with:
python3 main_parallel.py config/tmy_paris.jsonPASTICHE expects a working RRTM_LW executable. By default, the Colab workflow uses /content/RRTM_LW/rrtm_v3.3.1_linux_ifx; for local runs, set the executable path with:
export PASTICHE_RRTM_EXE=/path/to/rrtm_v3.3.1_linux_ifxIf you use PASTICHE, please cite the accompanying open-access paper:
Belotti et al. (2026). Spectral longwave atmospheric irradiance determination for site- and date-specific passive radiative cooling modeling. Sustainable Energy Technologies and Assessments 91 (2026): 105046. DOI: 10.1016/j.seta.2026.105046
- Mlawer et al. (1997). Radiative transfer for inhomogeneous atmospheres: RRTM, a validated correlated-k model for the longwave. DOI: 10.1029/97JD00237
- Hersbach et al. (2023). ERA5 hourly data on single levels and pressure levels from 1940 to present, Climate Data Store. DOIs: 10.24381/cds.adbb2d47, 10.24381/cds.bd0915c6
- Beck et al. (2023). High-resolution Köppen-Geiger maps for 1901–2099 based on constrained CMIP6 projections. Scientific Data 10, 724.
Claudio Belotti
Email: claudio.belotti@cnr.it
Lorenzo Pattelli
Email: l.pattelli@inrim.it
This project is GPL-3.0 licensed.
- Initial data release to INRIM and University of Lleida.
- Added relative humidity at 2 m above surface, calculated from ERA5 2 m temperature and 2 m dewpoint temperature.
- Added ERA5 total cloud cover.