This sample dataset contains extracted hourly forecasts up to three days ahead of the main pollutants:
- ozone (O3)
- nitrogen dioxide (NO2)
- sulphur dioxide (SO2)
- carbon dioxide (CO)
- particulate matter > 10 µm (PM10)
- particulate matter > 2.5 µm (PM25)
- air quality index (AQI)
at almost 3000 locations all over Europe.
All values are in µg/m3 except AQI which is dimensionless index in the range between 1 and 5 indicating the level of air quality (1 - best, 5 - worst)
The details about locations can be found in this file.
Sample data covering period from Nov 2017 to May 2018 is available on the public FTP server:
ftp://ftp.ecmwf.int/pub/macc/atmoshack
Files are named as
CAMS_WEB_FORECAST_${YYYYMMDD}_D+${DPLUS}.csv.gz
where
YYYYMMDD - date on which a forecast was generated DPLUS - forecast range in days; (T0 = YYYY-MM-DD 00UTC)
D+0 files contain forecast steps T0+0h..T0+24h D+1 files contain forecast steps T0+25h..T0+48h D+2 files contain forecast steps T0+49h..T0+72h D+3 files contain forecast steps T0+73..T0+96h
For example, file CAMS_WEB_FORECAST_20180101_D+1.csv.gz with contain forecasts valid from 2018-01-02 01:00 to 2018-01-03 00:00.
- Data is encoded in gzipped CSV files
- Files contain air quality forecasts for ~3000 of major European cities
- Times are expressed in UTC
ID,Date,Time,AQI,O3,NO2,SO2,CO,PM10,PM25
ALXX0001,2018-05-10,00:00,1,65.79,2.01,1.10,152.63,9.23,8.69
ALXX0001,2018-05-10,01:00,1,67.77,1.89,1.01,151.53,9.63,9.04
ALXX0001,2018-05-10,02:00,1,66.07,1.81,0.90,150.62,9.94,9.30
See this Jupyter notebook: https://github.com/OpenDataHack2018/Available-Open-Data/blob/master/Reading_CAMS_regional_data.ipynb