Back to Intensity Change projects
Back to Main Projects Page
Tropical cyclone wind radii estimation utilizing an empirical
inland wind decay model
John Kaplan (NOAA/AOML/HRD)
Collaborating Scientists ::
Mark DeMaria (NOAA/NESDIS/ORA)
Jason Dunion (NOAA/AOML/HRD)
Peter Dodge (NOAA/AOML/HRD)
Jose Salazar (NOAA/NHC)
Nicholas Carrasco (AER)
A technique for estimating the wind radii and maximum wind of landfalling
tropical cyclones in the Atlantic and eastern North Pacific basins utilizing
an empirical decay model (Kaplan and DeMaria 1995, 2001) has been developed as
part of the NOAA Joint Hurricane Testbed (JHT) (Kaplan et al. 2007). The
aforementioned technique employs a revised version of the decay model (DeMaria
et al. 2006) to generate estimates of the maximum wind and 34,50 and 64 kt
wind radii of landfalling storms where the input required to run the model is
obtained from the latest official NHC interpolated forecast. This particular
application of the decay model was accepted for operational use prior to the
2008 Hurricane season and is currently running on a server at NHC where the
output is available for use by forecasters for landfalling Atlantic and eastern
North Pacific basin storms. A more completed description of this particular
application of the decay model can be found in Kaplan et al. (2007).
Figure 1 shows an example of the decay model forecasted maximum wind that was
obtained for Hurricane Katrina (2005) using the operational NHC forecasted
input information that was available at 1200 UTC 29 August. Although this
forecast was not issued in real-time it is an independent example of the
model's performance since the revised version of the decay model described
above was derived using only landfalling cases from 1967-1993. As can been
in Fig. 2, both the OFCI and decay model forecasted intensities were too high
although the decay model values were in somewhat better agreement for this
particular forecast. Since the overland intensity forecast is very sensitive
to a stormÕs landfall intensity, the finding that the inland intensities were
too high might be due, in part, to the fact that the initial intensity used to
generate both the OCFI and decay models forecasts was about 10 kt higher that
that which was observed.
Figure 3 shows the decay model forecast errors as a function of radius from
the storm center where the decay model errors were computed for all available
wind observations that were collected as part of the HRD Hurricane Field
Program (HFP) landfall experiment. It is important to note that prior to
computing the model errors all of the wind observations were first standardized
to a 1-minute 10-m wind assuming open terrain exposure using the methods of
Powell et al. (1996). Also, only observations that occurred during the time
period when Katrina was still tropical were verified. Inspection of the figure
indicates that the decay model forecasted wind speeds had a high bias of 2 to
7 kt and that the mean error was about 14 kt. The finding of a high bias at
the individual observation locations appears to be consistent with the results
in Fig. 2 that indicated that the decay model forecasted maximum winds were
also higher than those that were observed. In the future, similar comparisons
between the model forecasted and observed winds will also be made for the 3-D
numerical models like HWRF and GFDL using other suitable landfall datasets that
were collected during the HRD HFP to ascertain how well these models predict
Since the updated decay model was not available for real-time testing
until the middle of the 2006 hurricane season, the model was verified for
an independent sample of landfalling Atlantic and E. Pacific basin
hurricanes. Figure 1 shows the errors between the National Hurricane
Center best track maximum wind and 64, 50 and 34 kt wind radii estimates
and those from the decay model for the 11 hurricanes that made landfall
in the Atlantic and E. Pacific basin during the period from 2004 -2006.
For comparison, the errors between the NHC best track maximum wind and
wind radii estimates and those from the GFDL, AVNO and NGPS models are
also depicted. The errors were obtained by comparing the model and NHC
best track estimates for initial (t=0 h) forecast times that were within
12 h of landfall. Errors were computed every 6 h until each system
either dissipated or became extratropical. The figure shows that the
maximum and wind radii estimates obtained using the new version of the
decay model were generally in better agreement with the NHC best track
estimates than both the old version of the decay model and the other
numerical guidance for this sample. The lone exception was the 34 kt wind
radii for which the AVNO model estimates were in better agreement.
- Finalized the software routines that are required to run the decay model
- Developed code to run an updated version of the Kaplan/DeMaria decay
model to obtain real-time estimates of the maximum wind and
the radii of 64,50 and 34 kt winds
- Ran the updated version of the decay model in real-time during the 2006
- Evaluated decay model performance on an independent sample of 11
hurricanes that made landfall in the Atlantic and E. Pacific basins
during the 2004-2006 hurricane seasons
- Presented decay model results at the 61st Interdeparmental Hurricane
- Provided revised version of the inland decay model to JHT facilitator for
- Developed software for stratifying decay model errors by radial and
azimuthal distance and time after landfall.
- Developed method for overlaying decay model maximum wind forecast on geography for real-time display purposes.
- Test the decay model in real-time for landfalling cases during the 2011
and 2012 Atlantic and eastern North Pacific hurricane seasons.
- Evaluate decay model performance for suitable landfalling storms.
- Develop method for evaluating the overland wind forecasts of
three-dimensional numerical models.
DeMaria, M., J. A. Knaff, and J. Kaplan, 2006: On the decay of tropical
cyclone winds crossing narrow landmasses., J. Appl. Meteor. and Clim.,
Kaplan, J. , and M. DeMaria, 1995: A simple empirical model for predicting the
decay of tropical cyclone winds after landfall. J. Appl. Meteor., 34, 2499-2512.
Kaplan, J., and M. DeMaria, 2001: On the decay of tropical cyclone winds after
landfall in the New England region. J. Appl. Meteor., 40, 280-286.
Kaplan , J, J. Dunion, and N. Carrasco, 2007: Estimating tropical cyclone wind
radii utilizing an empirical inland wind decay model. Joint Hurricane Tesbed
Powell, D., P.P. Dodge, and M.L. Black, 1996: Hurricane Andrew's landfall in
south Florida. Part I: Standardizing measurements for documentation of surface
wind fields. Wea. Forecasting, 11, 304-328.
Back to Intensity Change projects
Back to Main Projects Page
Last modified: 7/14/2011
Fig. 1. Decay model estimated maximum sustained wind estimated (green) obtained based upon input from the 1200 UTC 29 August NHC official interpolated forecast cycle for Hurricane Katrina (2005).
Fig. 2. Decay model (Decay) and official interpolated (OFCI) maximum sustained wind estimates for the 1200 UTC 29 August forecast cycle for Hurricane Katrina. The NHC best track estimates (Best) for this time period are also depicted.
Fig. 3. Decay model wind error (blue) and biases (red) as a function of radial distance from the storm center. The number of observations in each of the two radial bins (N) is also depicted along the x-axis. The mean time after landfall of all of the observations that were verified was about 6 h.