The Met Office uses a configuration of the Unified Model (UM) to produce air quality forecasts, known as AQUM - Air Quality in the Unified Model. It is a limited area configuration, covering the UK and some of nearby European countries, on a 0.11 degree (~12km) grid. It uses UKCA with the Regional Air Quality (RAQ) chemistry mechanism, and the CLASSIC aerosol scheme.
The model is initialised at 18Z daily, to produce a 126-hour forecast: midnight-to-midnight on the next 5 days. Initial meteorological conditions are taken from the T+6 dump of the 12Z global model run, and initial chemistry and aerosol conditions are taken from T+24 of the previous AQUM run. Lateral boundary conditions are taken from the Copernicus Atmosphere Monitoring Service (CAMS) C-IFS model at ECMWF. It also makes use of GFAS emissions from ECMWF.
Operational forecasts also include a post-processing step. This concerns only the five main pollutants used by the Daily Air Quality Index (DAQI). The concentrations are regridded to a 2 km OSGB grid, a simple statistical bias correction technique based on observations is applied, and finally the DAQI is calculated.
The daily air quality index is an integer from 1-10 (inclusive) representing the overall short-term risk to health, based on a few key pollutants. These are:
An index from 1-10 is derived for each of these pollutants, according to whether certain thresholds are exceeded during the given 24 hour period. Different aggregation methods are used depending on the pollutant:
The daily air quality index is then the maximum of these individual indices. Indices 1-3 are considered "Low", 4-6 "Moderate", 7-9 "High", and 10 "Very High".
For further details, including the threshold values, see Defra's report.
The resulting data archived to MASS includes both the raw model output in pp format, and post-processed data in nimrod format. On this platform, we expose only the post-processed data, converted to netCDF for convenience.
Mass concentrations of pollutants are given as hourly means. These are concentrations at surface level; there is no vertical component. A representative timestamp is chosen to be the end of the hour. All units are µg/m^3.
The daily air quality index is given for each day, with a representative timestamp of 12Z. The individual pollutant indices that contributed to the DAQI are not retained. Note that in order to maintain a consistent naming convention between all daily datasets, an aggregation method of "mean" is used for the purposes of naming the file (eg
aqum_daily_daqi_mean_yyyymmdd.nc), although this is an arbitrary choice as it is not (directly) derived from hourly data at all.
Each file contains some additional attributes:
field_code: code used internally to identify this variable.
short_name: field code mapped to a user-friendly name (one of: DAQI, NO2, O3, SO2, PM10, PM2p5).
data_type: the string "sppo", meaning Statistical Post-Processing of Observations, as opposed to "raw".