Summary Page Explanation
The summary page displays AQI difference 'highlights' covering the time period specified at the top of the grid.
In the example image we are shown that the maximum difference between the forecasted AQI and measured AQI for Lima was 5 - when the forecast was below the actual produced value. In addition the maximum difference and the times they occurred are shown per pollutant. For example the time between 3rd-8th July with the maximum divergence between the PM2.5 forecast and measurements was 03 July, 12:00. For PM10 the maximum divergence was 05 July, 09:00.
The page is automatically ordered by the AQI difference column, starting with the absolute value and then preferring positive differences over negative ones, i.e. the order would be 6, 6, -6, 5, -5, -4, 2 etc
Its should be noted that the in-situ measurements on the summary page are gathered across a 3 hour window; the pollutant/AQI value for 06:00 encompasses data from 04:30 - 07:30. This is due to the 3 hourly times used by the forecast, to ensure the figures are comparable.
Pollutant difference rules
- Where there are multiple times the data diverges for a particular pollutant with the same overall AQI difference, we will prefer items where the forecast is higher than the measured. I.e if at 12:00 we have forecast 4, measured 6, and at 15:00 we have forecast 5, measured 3, we will display 15:00 values for the pollutant. This is despite the fact that the overall values were higher at 12:00.
- Where there is no difference here (i.e. either the forecast is always higher, or the measured is), the time with the highest overall AQI will be shown. I.e if at 12:00 we have forecast 5, measured 3, and at 15:00 we have forecast 6, measured 4, we will display 15:00 values for the pollutant.
- Where there is no difference, we will usually return the earliest time. I.e if at 12:00 we have forecast 6, measured 4, and at 15:00 we have forecast 6, measured 4, we will display 12:00 values for the pollutant.
AQI difference rules
The individual AQI values with the highest difference are those shown in the left most columns. These are the overall AQI values (i.e. across all pollutants) where the difference was at it's maximum. So looking just at the PM pollutants, if we assume the following example:
| Pollutant | 12:00 | 15:00 |
|---|---|---|
| PM2.5 | forecast 2, measured 1 | forecast 2 measured 2 |
| PM10 | forecast 2, measured 2 | forecast 6 measured 3 |
The difference will have an value of 3, and show forecast 6, measured 3. This is extended out for all pollutants in reality. In effect it is displaying the 'highlight' of the maximum divergence between forecasted and measured AQI values.
The difference is dictated by the forecast AQI minus the measured AQI. If the measured is higher than the forecast AQI we have a negative difference.
In cases where multiple pollutants have identical differences we will follow similar rules to the pollutant rules above:
- Where there are multiple pollutants with the same overall AQI difference, we will prefer items where the forecast is higher than the measured. I.e if for PM2.5 we have forecast 4, measured 6, and PM10 we have forecast 6, measured 4, we will display forecast 6, measured 4, difference +2. Effectively we preference a positive difference over a negative difference.
- Where there is no difference here, the AQI's with the highest overall AQI value will be preferred. I.e if for PM2.5 we have forecast 5, measured 3, and for PM10 we have forecast 6, measured 4, we will display forecast 6, measured 4, difference +2.
- When there is no further difference it no longer matters, as the displayed value would be the same.