The Temperature (MaxT and MinT) Index measures the accuracy of automated post-processed model maximum and minimum temperature forecasts for a fixed list of UK sites.
The index measures a sequence of maximum and minimum temperatures, covering 2 days.
Temperature forecasts for each element are issued for each site for both days such that in effect 4 sets of forecast data are issued.
A rolling 36-month score is produced using the combined available data from the quality controlled site list for all forecasts.
Maximum and minimum temperature errors are treated as categorical variables. The scoring system uses the "percentage correct" score to compile a simple index value for temperature.
Verifying observations are taken from National Climate Messages (NCMs) at 2100 UTC for maximum temperatures and 0900 UTC for minimum temperatures which are stored in the quality-controlled climate database.
For each forecast (fc) and corresponding observation (ob), a score is awarded to the following criteria:
For the first maximum and minimum temperatures
And the second maximum and minimum temperatures
The percentage correct (PC) is defined as: PC = (100 x score)/(max score), where score is the total points scored during the period and max score is the maximum points possible in the period.