As part of the creation of the climatology/statistical paper for the TC Inland Winds Group, NDFD forecasts were verified by comparing the latest forecast cycle prior to verification to H*Wind surface analyses. NDFD sustained wind speed forecasts are available since December 2005, when they began to be created in operations. Atlantic TCs impacting the region Dec. 2005-2011 include:
Ernesto (2006)
Barry (2007)
Gabrielle (2007)
Cristobal (2008)
Hanna (2008)
Earl (2010)
Irene (2011)
No H*Wind analyses were created by HRD for Barry (2007). The remaining 6 storms were used in the analysis. Summary plots of the normalized average wind speed bias relative to the average H*Wind analyzed speed in the region are shown in the plots here. Note the overprediction of wind speeds for most storms and over most WFOs. Feel free to comment on these finalized images as well as provide feedback to make the figures more clear. I will attempt to provide some additional summary points over the upcoming days.
Bryce, these are nice charts that once again confirms the message that your verification work and subjective experience has provided us and that is that the wind forecasts in the NDFD possess a high bias in tropical cyclone events.
The only real exception to the charts (https://cimmse.files.wordpress.com/2012/06/hwind_ndfd_comparison.ppt) which are just smothered with red suggesting a high bias is Gabrielle which was a tropical storm (peak winds were 50 kts) that moved across far eastern NC and the Outer Banks. Note the wind gust map across NC (http://www4.ncsu.edu/~nwsfo/storage/cases/maps/windgust.20070909.gif) which shows that wind gusts of 35 mph or less occurred across nearly all locations except the immediate coast and Outer Banks.