Hurricane Notes, Part 2

1
Hurricane Notes, Part 2

• ATS 553

2
Bill Grays Forecast

• Forecasts in December, June and August
• Forecasts
• Named Storms, Named Storm Days
• Hurricanes, Hurricane Days
• Intense Hurricanes, Intense Hurricane Days
• Hurricane Destruction Potential
• Maximum Potential Destruction
• Net Tropical Cyclone Activity

3
Techniques

• Uses a combination of SUBJECTIVE and OBJECTIVE
  techniques.

4
Factors in Grays Forecast
5
1. Phase of ENSO

• The Walker Circulation shifts eastward during El
  Nino.
• Tropical Atlantic experiences upper-level
  westerlies during El NinoSHEAR
• El Nino means FEW hurricanes.
• La Nina means comparatively MORE hurricanes.

6
An El Nino Year
7
A La Nina Year
8
2. The Quasi-biennial Oscillation

• An oscillation in
• The zonal winds
• Along the equator
• In the Stratosphere
• Period is about 23 months.

9
(No Transcript)
10
2. QBO

• When lower stratospheric or upper tropospheric
  winds are westerly, thats SHEAR, which results
  in fewer hurricanes.
• However, the QBO is hard to measure near the
  tropopause. Its easy to measure higher, like 30
  mb

11
2. QBO

• 30mb winds
• From the east
• Tropopause winds are from the west
• SHEAR, fewer hurricanes
• From the west
• Tropopause winds are from the east
• NO SHEAR, more hurricanes

12
3. Sahelian Rainfall

• In drought years in the Sahel, Atlantic hurricane
  activity is suppressed.
• Why?
• Maybe the same shear that is bad for hurricanes
  is also bad for AEWs.
• Maybe dust from the Sahel is a hurricane
  suppressant (there is evidence of this)

13
4. Caribbean Sea Level Pressure Anomalies

• When MSLP in the Caribbean is lower than normal,
  you get MORE hurricanes.

14
H
When the pressure is unusually high in the
Caribbean, it usually means that the ITCZ is far
to the south, usually in a tradewind trough.
15
H
When the pressure is unusually low in the
Caribbean, it usually means that the ITCZ is far
to the north, usually in a monsoon trough.
16
5. Caribbean 200mb Zonal Wind Anomaly

• A very direct way to include shear in the Gray
  forecast
• U from the west means FEWER hurricanes
• U from the east means MORE hurricanes
• Can be due to any cause ENSO, QBO, a ridge, a
  trough, whatever

17
6. SSTs in 3 regions of the Atlantic

• Warmer water means MORE hurricanes.
• Colder water means FEWER hurricanes.
• SST can change rapidly in the Atlantic, due to a
  generally shallow thermocline. Therefore, it is
  hard to predict how high the SSTs will be far in
  advance.

18
MANY other variables

• A total of 15 variables
• Also includes ANALOG YEARS in the forecast.

19
How Well Does Gray Do?
20
Year August Forecast Observed
1984 10 12
1985 10 11
1986 7 6
1987 7 7
1988 11 12
1989 9 11
1990 11 14
1991 7 8
1992 8 6
1993 10 8
1994 7 7
1995 16 19
1996 11 13
21
How Well Did Gray Do This Year?
22
NOAA Forecast 2004

• 12-15 tropical storms
• 6-8 hurricanes
• 2-4 intense hurricanes

23
Observations 2004

• H. ALEX (3)
• T.S. BONNIE
• H. CHARLEY (4)
• H. DANIELLE (2)
• T.S. EARL
• H. FRANCES (4)
• T.S. GASTON
• T.S. HERMINE

• H. IVAN (5)
• H. JEANNE (3)
• H. KARL (4)
• H. LISA (1)
• T.S. MATTHEW
• T.S. NICOLE
• T.S. OTTO

24
Observations 2004

• H. ALEX (3)
• T.S. BONNIE
• H. CHARLEY (4)
• H. DANIELLE (2)
• T.S. EARL
• H. FRANCES (4)
• T.S. GASTON
• T.S. HERMINE

• H. IVAN (5)
• H. JEANNE (3)
• H. KARL (4)
• H. LISA (1)
• T.S. MATTHEW
• T.S. NICOLE
• T.S. OTTO

15 tropical storms!
25
Observations 2004

• H. ALEX (3)
• T.S. BONNIE
• H. CHARLEY (4)
• H. DANIELLE (2)
• T.S. EARL
• H. FRANCES (4)
• T.S. GASTON
• T.S. HERMINE

• H. IVAN (5)
• H. JEANNE (3)
• H. KARL (4)
• H. LISA (1)
• T.S. MATTHEW
• T.S. NICOLE
• T.S. OTTO

8 hurricanes!
26
Observations 2004

• H. ALEX (3)
• T.S. BONNIE
• H. CHARLEY (4)
• H. DANIELLE (2)
• T.S. EARL
• H. FRANCES (4)
• T.S. GASTON
• T.S. HERMINE

• H. IVAN (5)
• H. JEANNE (3)
• H. KARL (4)
• H. LISA (1)
• T.S. MATTHEW
• T.S. NICOLE
• T.S. OTTO

6 intense hurricanes!
27
Problems with Grays Technique
28
1. Multicollinearity

• Grays forecast has MANY problems with
  multicollinearity.
• Forecasting something using multiple variables
  that themselves are highly correlated.
• Leads to artificially high correlations, but
  doesnt improve the forecast at all.

29
200mb u
Sahel drought
ENSO
MSLP
30
200mb u
QBO
31
MSLP
Low latitude SSTs
Higher latitude SSTs
32
Guinean Precip
Sahel drought
33

• Only ENSO, QBO, and SST anomalies in the Atlantic
  are truly independent. These are probably the
  only variables that add real value to the
  forecast.
• These factors change slowly and are relatively
  easy to forecast, which is good!

34
2. Subjective Elements

• QBO
• ENSO
• Analog years

35
3. Overforecasting

• Forecasts of of hurricanes, of hurricane
  days, etc. are pretty good.
• Many other s are just junk.

36
4. Garbage in, garbage out

• Forecast equations were trained using over a
  hundred years of hurricane data.

37
1852
38
1866
39
5. Forecast VALUE?

• How do we act on this information?

40
Why Does Grays Technique Work So Well?
41

• 1. ENSO, QBO, and SST anomalies dominate the
  equations, and there is no doubt that they ARE
  good predictors.

42

• 2. Gray is a skilled forecaster. He makes
  excellent forecasts of ENSO, QBO, SST anomalies,
  etc.

43

• 3. Subjectivity at the National Hurricane Center.