TC Winds Conference Call

1

TC Winds Conference Call

Wednesday, July 17, 2012

11:00 AM

2

Agenda

• Holland et al. (2010) model vs. Rankine

• Decay over land

• NDFD Verification update

• Discussion

3

TCMWindTool

• “The tool then interpolates down from the radius of maximum wind to the 34 knots wind radii using the Rankine vortex interpolation”—TCMWindTool Documentation

• Question: Is there something better than can be incorporated into the TCMWindTool?

4

Holland et al. 2010

• “A Revised Model for Radial Profiles of Hurricane Winds” (MWR, 2010)

• Develops new profile similar to Rankine and other previous practiced methods, but empirically modified to incorporate data from outside the flow field as well as data at radius of maximum winds

5

• “Input data consists of an outer wind and surface pressure, sea surface temperature, radius of maximum winds, and central pressure. All of these parameters are readily available from most hurricane archives or warnings.“

• “If no direct external wind observations are available, then an estimate of the radius of gale-force winds will suffice…as shown, errors in this radius will not have a marked effect on the accuracy of the wind field, whereas the anchoring of the profile in the outer domain will produce a marked improvement”

• Main issue with Modified Rankine Vortex: too sensitive to TCM product

Holland et al. 2010

6

• Both modified rankine vortex (as in TCMWindTool) and Holland et al. 2010 have been coded in NCL

• Goal: show improvement in interpolated wind field for Holland et al. model

Holland et al. 2010 Update

7

Example Plot: Irene (2011)2100 UTC 25 August• TCM Data:

Max winds: 100 KT64 KT: 70NE 60SE 25SW 50NW.50 KT: 110NE 100SE 50SW 75NW.34 KT: 250NE 200SE 125SW 160NW.

(Quadrant 1 Cross Section)

8

Example Plot: Irene (2011)2100 UTC 25 August

9

Initial Results• Modified Rankine leads to discontinuity in wind profile near radius

of maximum winds

• Much smoother gradients for wind profile in Holland et al. model

• Modified Rankine Vortex completely sensitive to various maximum wind radii– Not a good NHC product

• Holland et al. not as good near radius of maximum winds– Does not take into account 64 kt wind radii, so gradients off in

this region

• Sensitivity analysis to various parameters currently being conducted to show overall improvement in Holland et al. model given TCM uncertainties

10

Discussion

• What are suggested ranges of errors for TCM maximum wind radii to test sensitivity with? Any documentation for these suggested errors?

11

Decay over Land

• TCMWindTool motivation: tool linearly interpolates 6 hour wind speed data to the hourly grids

• Is this linear interpolation ok?

12

Decay over Land

• RUC hourly analyses used in this study and analyzed for all landfalling storms 2002-2011– Kyle (2002) -Isabel (2003)– Charley (2004) -Gaston (2004)– Ernesto (2006) -Gabrielle (2007)– Hanna (2008) -Irene (2011)

• Quadrant averaged decay analyzed by quadrant• Biggest errors in linear interpolation associated

with large changes in storm strength between the six hour TCM times

13

Isabel (2003)

14

Irene (2011)

15

NDFD Verification Process• NDFD forecast and HWind analysis is interpolated to a common grid

• Difference calculated at each analysis time between forecast and HWind analysis wind speed, plots created for this difference at each time

• Average difference is also calculated and is normalized relative to the average HWind analysis wind speed at that grid point

• H*Wind analyses also plotted on the common grid at all times

• All images have been created and will be available at by Thursday: – http://www4.ncsu.edu/~bptyner

• Forecaster input welcome over next several weeks

16

Focuses Next Few Weeks

• TCMWindTool immediate focuses for improvement– New model suggestion(s) for land decay once

over land to account for rapid changes in storm strength

– Holland et al. (2010) sensitivity to TCM product analysis vs. modified rankine vortex

• NDFD write up for paper• Gust factor analysis and write up for paper• Goal: draft of paper end of summer, move on to

LES simulations– Continue conference calls for forecaster input

17

Open Discussion

• Questions/Comments• Discussion points:

– TCM suggested errors to examine sensitivity for 34, 50, 54 knot radii as well as maximum wind radius?

– Suggestions for land decay function? Other NHC products that can be used for this decay?

– NDFD other plots to include? Email to Bryce who can then compile and write summary report?

• Next Conference Call: Wednesday, August 8th at 11:00 am