Forecast Methodology

How are the forecasts produced?

Our forecasting methodology is based not on any single computer model but is grounded in what is happening in the real world and expert knowledge.

Our first step is to carefully examine the pollution concentrations and trends from all the automatic monitoring networks we manage across the southeast.

Next, we examine why this occurred. Custom back-trajectory models are run which give us both the path and height the air has been at before arriving with us. We then judge what pollution it is likely to have picked up on route, based on the location of major pollution sources such as cities and using our 20 years experience.

Our next step is to ask what will be different in the next few days. Our back-trajectory models also forecast air movement over the coming days. Once we have forecast air movement, and hence the passage over emission sources, we look at how this will be different from what is currently arriving, and how this will impact on our pollution. Other factors include any polluted air over the UK and Europe that might move into the AirAlert area.

We then look for supporting evidence to back up our analysis. There are several European-wide pollution models which produce outputs each day. As with all modelling different models can produce different answers so we examine several looking for agreement between models, and also which models correctly predicted recent conditions recorded across the networks. More credence is given to models which are correct for current levels.

Finally, having gained the overall picture we apply local knowledge and understanding of how local conditions may impact pollution levels, e.g. land-sea breezes moving coastal urban emissions inland, the south side of the Downs being warmer and impacting on ozone generation.

Our team at King’s College London has been examining how pollution levels behave on a 15 minute or hourly basis every day for the past 20 years, and from this comes an inherent understanding of how things behave and interact. We never just believe without questioning outputs from computer models, they have to make sense to our experienced team.


Band Index Ozone Nitrogen
Running 8
hourly mean
15 minute
24 hour
24 hour
µg m-3 µg m-3 µg m-3 µg m-3 µg m-3
  1 0-33 0-67 0-88 0-11 0-16
2 34-66 68-134 89-177 12-23 17-33
3 67-100 135-200 178-266 24-35 34-50
  4 101-120 201-267 267-354 36-41 51-58
5 121-140 268-334 355-443 42-47 59-66
6 141-160 335-400 444-532 48-53 67-75
  7 161-187 401-467 533-710 54-58 76-83
8 188-213 468-534 711-887 59-64 84-91
9 214-240 535-600 888-1064 65-70 92-100
Very High
  10 241 or more 601 or more 1065 or more 71 or more 101 or more

Pollution concentrations recorded by the LAQN analysers have been classified according to Defra’s Air Pollution Index system. This classes levels into bands from ‘low’ to ‘very high’. Each band is subdivided into three to produce an Air Pollution Index, from 1 to 10, 1 being ‘low’, 10 being ‘very high’. Measurements are rounded to the nearest whole number.
The concentrations are classified according to the table above. Note that different pollutants have different concentrations and averaging periods, related to the Read a technical explanation of how the new index bands (1st January 2012 onwards) are calculated here.