Storm Forecast Accuracy

[Dan]

The Storm Forecast Accuracy map is updated usually once a week. It uses a 50 x 50 mile (80 x 80km) grid.

The data is collected from the UK Met Office ATD (Arrival Time Difference) Network and uses the following basis:

• if lightning occurred and was forecast = high accuracy
• if lightning occurred and wasn't forecast = low accuracy
• if lightning was forecast but didn't occur = low accuracy
• if lightning wasn't forecast and didn't occur = high accuracy

Each time 'low accuracy' appears in one of the boxes on the grid, a "1" is added to the cumulative number value. Grid squares which are consistently accurate in our forecasts remain on "0".

The overall accuracy can then be determined by comparing the number value in each grid square to the total number of forecasts issued. The higher the number value in a grid square, the lower the accuracy. The number value in each grid square cannot, and will not, be higher than the total number of forecasts issued, for obvious reasons.

[Dan]

Map issued Monday 8th June 2009

(valid for 17th May - 7th June 2009)

[Dan]

Map issued Wednesday 17th June 2009

(valid for 17th May - 16th June 2009)

[Dan]

So we've started this experiment again, commencing 1st May 2010 and running throughout the summer to see which regions have the highest forecast accuracy. Maps will be attached every day following a storm forecast with the latest accuracy data.
Accuracy since 1st May 2010:

[Dan]

<span style="font-weight: bold;">Accuracy since 1st May 2010:</span>
<span style="font-weight: bold;">
</span>

[Dan]

Accuracy since 1st May 2010, updated 23rd May 2010:

[Dan]

Accuracy since 01 May 2010, updated 24th May 2010:

[Dan]

Accuracy since 01 May 2010, updated 26th May 2010

[Dan]

Accuracy since 01 May 2010, updated 29th May 2010

[Dan]

Accuracy since 01 May 2010, updated 6th June 2010