Potential Update to TCMWindTool: Wind Interpolation Method

As has been discussed, the TCMWindTool currently uses the Modified Rankine Vortex method for wind speed interpolation from the TCM product.  For those unfamiliar with the Modified Rankine Vortex, here are a few of the key features:

  • The method requires the following input:  radius of maximum winds (assumed in the tool to be the radius of the eye + 8.0 nautical miles), maximum winds speeds at this radius,  and four quadrant data (maximum radius of 34, 50, and 64 knot winds).
  • The wind speeds are interpolated linearly from the storm center to the radius of maximum winds.
  • Outside of the radius of maximum winds, the wind speeds are interpolated using exponential decay between each maximum wind radius (radius of maximum winds, 64, 50, and 34 knots).
  • Outside the radius of 34 knots, the wind speeds are “blended” with a background wind field.

A recent discussion with several of the CSTAR TC Winds collaborators motivated a literature review to see if there are other possible methods of wind speed interpolation.  In Holland et al. (2010), the authors develop an alternate method for this interpolation.  A few key features of this methodology:

  • The method requires the following input:  radius of maximum winds, maximum winds at this radius, surface air density in this location (though the authors show that if this data is not available, the resulting error is small),  storm central pressure, the pressure at a defined external location from the cyclone center, the wind speed at this external location, and the radius of this external location.
  • The authors show in the paper that the error in wind speed interpolation with this method is much less sensitive to the radius of maximum wind than the Modified Rankine Vortex method.
  • The authors show in their sensitivity analysis that “archived values of the subjectively defined radius of gale-force winds can likely be used” for defining the characteristics of the external data from the cyclone center

Recently, Bryce coded the Modified Rankine Vortex method along with the Holland et al. (2010) method.  Shown in the attached images are plots for Irene (2011) out to a radius of 400 km from the storm center.  The TCM product from 2100 UTC 24 August was used to create these images.  Some of the key TCM data used in the analyses:

Maximum sustained winds:  100 kt
64 KT……. 70NE 60SE 25SW 50NW.
50 KT…….110NE 100SE 50SW 75NW.
34 KT…….250NE 200SE 125SW 160NW.  (used for outer wind information for Holland et al. interpolation)
Dvorak technique used to estimated pressure at the 34 kt wind radii (1005 hPa)

You can see that the two interpolation methods result in significantly different radial wind speed distributions.  The next step will be to show objectively that the Holland et al. model indeed is a better interpolation method.  This can be done by taking the interpolated wind fields (with no forecaster manipulation or considering land reduction) at several times near the time of landfall and comparing to HWind analyses near these times.

As always, comments and suggestions are welcome!

Advertisements
This entry was posted in TC Inland and Marine Winds. Bookmark the permalink.

2 Responses to Potential Update to TCMWindTool: Wind Interpolation Method

  1. Jonathan Blaes @ WFO RAH says:

    Bryce, thanks for exploring this and investigating. The developers of the TCMWindTool acknowledge that the Rankine Vortex model was a weakness of the design of the tool. My hunch is that they used a simple model because of the limited inputs available to the tool via the TCM or other product.

    Regarding the external location variables needed in the Holland et al. model (the pressure at a defined external location from the cyclone center, the wind speed at this external location, and the radius of this external location), do you have to obtain these for each forecast hour in the TCM product? I can envision this could be a problem since this data is not provided centrally, and could be burdensome, perhaps some others will have some comments about this concern.

    I noticed that the wind distribution “appears” more reasonable and the gradients more relaxed. I am not sure if looking better actually translates into verifying better, but the appearance looks great. One thing I did notice is that there is a much larger area of stronger winds in the Holland model, the area of red seems to be much greater. Given the consistent bias of the NDFD forecasts, this TCMWindTool output would potentially require even larger reduction factors.

    • bptyner says:

      Jonathan, thanks for the feedback. Regarding the external location variables, as mentioned in the Holland et al. paper, the 34kt/gale force wind radii should be sufficient for improved wind interpolation over the rankine vortex. This data is available in the TCM product, as far as I can tell.

      Very good point regarding the possibility of needing to increase the reduction factors if this interpolation method proves to be beneficial. I will update when I have a change at comparing to HWind analysis for verification.

      All others are encouraged to comment as well.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s