Participants: Jonathan Blaes-RAH, Reid Hawkins, ILM, John Billet-AKQ, Bob Bright-CHS, David Glenn-MHX, & Bryce Tyner-NCSU.
-Bryce presented an update to the Holland et al. (2010) wind interpolation method vs. the modified Rankine vortex method. Previous radial plots shown on the blog for Irene (2011) seemed to show a large area of higher wind speeds near the storm center in the Holland et al. model than the modified Rankine vortex model. After further inspection and plotting cross sections from the storm center to the storm perimeter, it was determined that this was only an illusion from the polar stereographic projection used in the plotting algorithm. Bryce is currently working with Anantha to turn the four quadrant data output from the two models into cylindrical coordinates to remove this distortion.
-Bryce suggested using two different methods to show improvement when using the Holland et al. model to interpolate wind speeds. First, he suggested comparing both models to H*Wind surface analyses to see if there is marked improvement in the wind field. He also suggested a sensitivity study where the 30, 50, and 64 knot wind radii are altered according to NHC’s suggested range of errors. As noted in the Holland et al. (2010) paper, the method is much less sensitive to errors in the wind radii and hence should show improvement over the modified Rankine vortex. When complete, plots of this verification will be added to the blog. As always, comments/observations will be welcome.
-The NDFD verification is complete and has been posted to www4.ncsu.edu/~bptyner. This includes a verification of 24 and 48 hour forecasts prior to landfall for 30 select ASOS stations in the study region. Collaborators are encouraged to provide their observations of the NDFD data and email them to Bryce. A form is being created with select questions for collaborators to focus on. These will be distributed over the next week or so.
-Jonathan presented results of a study conducted by several undergraduate students at NC State looking at gust factors of recent TCs affecting the region. Using 14,000+ hourly station observations, the results indicate a logarithmic fit can be made to the gust factors, with high gust factors (approx. 1.5) at lower wind speeds gradually decreasing asymptotically to near 1.15-1.2 at wind speeds >70 mph. A gust factor grid has been coded into smart tools to be used by forecasters based on the logarithmic fit and provides a great alternative to constant gust factors applied to the WFO domain. While tweaks are still being made to the smart tool, it will likely serve as a good starting point for a more scientifically based and collaborative wind gust forecast in the region.
-Slides from the conference call are available here. The next conference call is scheduled for Wednesday, September 12th at 11 AM.