Development of a Rapid Intensification Index for the Eastern Pacific Basin

1
Development of a Rapid Intensification Index for
the Eastern Pacific Basin

• John Kaplan
• NOAA/AOML
• Hurricane Research Division
• Miami, FL
• and
• Mark DeMaria
• NOAA/NESDIS
• Fort Collins, Colorado

Acknowledgements Joint Hurricane Testbed
2
Background

• The present capability to predict rapid
  intensification (RI) remains inadequate (Keith
  (2000), Lilly (2002), and Kenna (2003)) and
  predicting RI has been ranked as one of the top
  forecasting priorities by the NHC
• Since the present intensity prediction models
  have not yet shown the capability to predict RI
  adequately, Kaplan and DeMaria (2003) have
  developed a simple index for predicting the
  probability of RI using SHIPS model output
• The RI index was run in real-time in the Atlantic
  basin from 2001-2003 as part of the JHT
• The Atlantic version of the RI index will be run
  operationally starting with the 2004 hurricane
  season
• An RI index analogous to the one developed in the
  Atlantic basin is being developed for the E.
  Pacific basin with support from the JHT

3
Cumulative Frequency distributions
(1989-2002) All over-water tropical and
subtropical cyclones
4
Definition of Rapid Intensification (RI)

• Rapid intensification (RI) is defined as the 95th
  percentile of all over-water 24-h intensity
  changes of the subtropical and tropical cyclones
  that developed from 1989-2002
• This equates to a 24-h maximum sustained wind
  increase of gt 15.4 ms-1 (30 kt) (Atlantic
  Basin)
• This equates to a 24-h maximum sustained wind
  increase of gt 18.0 ms-1 (35 kt) (E. Pacific
  Basin)

5
24-h tracks of the E. Pacific RI cases
(N85) (24-h change in maximum wind gt 35 kt)
6
Development of the SHIPS RI index(Kaplan and
DeMaria Wea. Forecasting 2003)

• Determine SHIPS predictors for which
  statistically significant differences existed
  between the RI and non-RI cases
• Determine RI thresholds for all statistically
  significant predictors (average of all the RI
  cases)
• Compute probability of RI for each 24-h period by
  comparing the SHIPS predictor magnitudes to the
  corresponding RI thresholds
• Employ the 5 (Non-GOES version) and 7 (GOES
  version) SHIPS predictors that yielded the
  highest individual RI probabilities to compute a
  combined probability of RI
• Provide real-time estimates of the probability of
  RI with each SHIPS forecast

7
Predictors used in the 2004 Atlantic and E.
Pacific versions of the RI index
Non-GOES Version

• Previous 12 h intensity change (t0 h)
• Observed sea-surface temperature (24-h mean)
• Maximum potential intensity - initial storm
  intensity (24 h mean)
• 850-200 hPa vertical shear from 200-800 km at (24
  h mean)
• 850-700 hPa relative humidity from 200-800 km
  (24 h mean)
• GOES Version
• All predictors used in Non-GOES version PLUS
• Area-averaged inner-core brightness temperature
  (t0 h)
• Standard deviation of inner-core brightness
  temperature (t0 h)

8
Performance of the GOES version of the RI index
for the dependent sample (1995-2002)
9
N
Brier Score 1/N ? (F-E)2
(Wilks 1995)
i1
N number of forecasts F forecast probability
(where 50 is expressed as 0.5) E Event
probability (where E1 when RI occurred and 0
when it did not) The Brier score was computed
both for the RI index using the forecast
probability of RI values and for climatology
using the climatological probability of RI
Brier skill score 1 - (Brier score RI
index/Brier score climatology)
Thus, the RI index has skill (no skill) if it
has a Brier score that is less (greater) than
climatology
10
Brier Skill scores for the GOES versions of the
Atlantic and E. Pacific RI indices for the
dependent samples (95-02)
11
Real-time performance of the non-GOES and GOES
versions of the Atlantic RI index (2001-2003)
Non-GOES
GOES
12
Brier Skill scores of the real-time Atlantic RI
index forecasts
13
Performance of the non-GOES version of the E.
Pacific RI index for the 2003 re-run forecasts
Non-GOES
GOES
14
Brier skill scores for the EPAC RI index for the
2003 re-run forecasts
15
Change in 2003 SHIPS model skill obtained by
replacing Reynolds SSTs with AMSR-E SSTs
(Gentemann 2004)
E. Pacific
Atlantic
16
Yearly Brier skill scores for the GOES version of
the EPAC RI index for the 1995-2002 dependent
sample
17
Performance of the GOES version of the scaled RI
index for the 1995-2002 E. Pacific dependent
sample (N1248)
18
Summary

• The Atlantic RI index showed some skill when
  run in real-time in 2001 and 2003 (GOES version
  only), but performed poorly in 2002

• The E. Pacific version of the RI index showed
  more skill than the
• Atlantic version for the dependent samples

• The 2003 EPAC re-run forecasts were much less
  skillful than
• climatology, but appear to be within the range
  of skill that can be
• expected in any given year

• The GOES version of the RI index was found to be
  more skillful than the non-GOES version in both
  basins

• The possibility of developing a scaled version
  of the E. Pacific RI index is also being
  investigated

• The E. Pacific RI index will be tested in
  real-time for the first time
• during the upcoming 2004 Hurricane season

19
24-h mean intensity changes for the non-GOES and
GOES versions of the Atlantic RI index
(2001-2003)
Non-GOES
GOES
20
24-h mean intensity changes for the E. Pacific
2003 re-run forecasts
Non-GOES
GOES