Besides the essential heading correction, we apply also a true airspeed correction.

A model equivalent of the airspeed is determined using the ground vector together with the model wind vector. The angle with respect to true north is used for the heading correction and the length of this vector is compared to the reported true airspeed.

Statistics show that some aircraft exhibit a rather large offset between the observed and model equivalent airspeed. A lookup table is created based on differences between airspeed observations and model equivalents over very long period.

We re-processed one month of observation without (v2.2-noTAS) and with the true airspeed correction (v2.2) and the results, while comparing to ECMWF, are shown below.

The table below shows the mean difference in wind speed, and the relative change defined by the difference of the absolute biases divided by the standard deviation of the reference. The impact of the true airspeed correction on the bias is neutral.

The next table shows the standard deviation of the difference of the windspeed with ECMWF. An average improvement of around 4% is observed.

The statistics for wind direction show a similar behaviour as can be seen in the two tables below.