• The goal of the test was to evaluate Thompson microphysics with HWRF model.

• DTC performed large-scale statistical evaluation of the model output for HWRF with Thompson microphysics and compared them to operational configuration run by EMC.


• The AL mean track errors showed statistically significant (SS) pairwise differences indicating HDT6’s mean track errors were smaller than those for the control (H16A) after 66 hours.


• Mean along- and cross-track errors indicated a slow bias and a tendency to produce storm locations right of the actual storm for both the HDT6 and H16A configurations.

• Mean absolute intensity errors for HDT6 in the North Atlantic basin were slightly larger than those of H16A, with SS differences at two lead times (18, 84 hrs).Mean along- and cross-track errors indicated a slow bias and a tendency to produce storm locations right of the actual storm for both the HDT6 and H16A configurations.

• Mean intensity error distributions revealed lower mean intensity biases for HDT6 at longer lead times.

• The mean track errors for HDT6 in the eastern North Pacific basin were larger than those for the baseline configuration, with SS differences indicating degradations in track for HDT6 beyond 96 hours.

• Mean along-track errors indicated both configurations had a tendency to be too fast out to two days in the eastern North Pacific basin.

• Mean absolute intensity errors for all lead times beyond 72 hours showed SS degradations for the HDT6 configuration. Mean intensity errors demonstrated that, while both configurations exhibited a large negative intensity bias, the HDT6 configuration yielded larger negative biases beyond 36 hours.

• Results show some improvement in the spatial distribution of clouds with the HTD6 configuration. But it did not translate into improvement of track and intensity forecasts.