EXTENDED RANGE FORECAST OF ATLANTIC SEASONAL HURRICANE ACTIVITY AND U.S. LANDFALL STRIKE PROBABILITY FOR 2006

 

We continue to foresee another very active Atlantic basin tropical cyclone season in 2006.  Landfall probabilities for the 2006 hurricane season are well above their long-period averages.

 

(as of 4 April 2006)

 

 

 

By Philip J. Klotzbach[1] and William M. Gray[2]

 

with special assistance from William Thorson[3]

 

 

 

This forecast as well as past forecasts and verifications are available via the World Wide Web at http://hurricane.atmos.colostate.edu/Forecasts


Emily Wilmsen and Brad Bohlander, Colorado State University Media Representatives, (970-491-6432) are available to answer various questions about this forecast

 

 

 

Department of Atmospheric Science

Colorado State University

Fort Collins, CO 80523

Email: amie@atmos.colostate.edu

 

 

 

 

ATLANTIC BASIN SEASONAL HURRICANE FORECAST FOR 2006

 

Forecast Parameter and 1950-2000

Climatology (in parentheses)

Issue Date

6 December 2005

Issue Date

4 April 2006

Named Storms (NS) (9.6)

17

17

Named Storm Days (NSD) (49.1)

85

85

Hurricanes (H) (5.9)

9

9

Hurricane Days (HD) (24.5)

45

45

Intense Hurricanes (IH) (2.3)

5

5

Intense Hurricane Days (IHD) (5.0)

13

13

Net Tropical Cyclone Activity (NTC) (100%)

195

195

 

 

 

 

 

 

PROBABILITIES FOR AT LEAST ONE MAJOR (CATEGORY 3-4-5) HURRICANE LANDFALL ON EACH OF THE FOLLOWING COASTAL AREAS:

 

1)      Entire U.S. coastline - 81% (average for last century is 52%)

 

2)      U.S. East Coast Including Peninsula Florida - 64% (average for last century is 31%)

 

3)      Gulf Coast from the Florida Panhandle westward to Brownsville - 47% (average for last century is 30%)

 

4)      Above-average major hurricane landfall risk in the Caribbean

Notice of Author Changes

 

By William Gray

The order of the authorship of these forecasts has been reversed from Gray and Klotzbach to Klotzbach and Gray.  After 22 years (since 1984) of making these forecasts, it is appropriate that I step back and have Phil Klotzbach assume the primary responsibility for our project’s seasonal, monthly and landfall probability forecasts.  Phil has been a member of my research project for the last five years and has been second author on these forecasts for the last four years.  I have greatly profited and enjoyed our close personal and working relationships.

 

Phil is now devoting more time to the improvement of these forecasts than I am.  I am now giving more of my efforts to the global warming issue and in synthesizing my projects’ many years of hurricane and typhoon studies.

 

Phil Klotzbach is an outstanding young scientist with a superb academic record.  I have been amazed at how far he has come in his knowledge of hurricane prediction since joining my project five years ago.  I foresee an outstanding future for him in the hurricane field.  I expect he will make many new forecast innovations and skill improvements in the coming years.  I plan to continue to be closely involved in the issuing of these forecasts for the next few years. 



ABSTRACT

 

Information obtained through March 2006 continues to indicate that the 2006 Atlantic hurricane season will be much more active than the average 1950-2000 season.  We estimate that 2006 will have about 9 hurricanes (average is 5.9), 17 named storms (average is 9.6), 85 named storm days (average is 49.1), 45 hurricane days (average is 24.5), 5 intense (Category 3-4-5) hurricanes (average is 2.3) and 13 intense hurricane days (average is 5.0).  The probability of U.S. major hurricane landfall is estimated to be about 55 percent above the long-period average.  We expect Atlantic basin Net Tropical Cyclone (NTC) activity in 2006 to be about 195 percent of the long-term average.  This early April forecast is based on a newly devised extended range statistical forecast procedure which utilizes 52 years of past global reanalysis data. Analog predictors are also utilized. We have maintained our forecast from our early December prediction as the Atlantic Ocean, although cooling slightly with respect to climatology, remains anomalously warm and central and eastern tropical Pacific sea surface temperatures anomalies have continued to cool.  Currently, weak La Niña conditions are observed.  We expect either neutral or weak La Niña conditions to be present during the upcoming hurricane season. 

 

 

 

Acknowledgment


We are grateful to the National Science Foundation (NSF) and Lexington Insurance Company (a member of the American International Group (AIG)) for providing partial support for the research necessary to make these forecasts.  We also thank the GeoGraphics Laboratory at Bridgewater State College (MA) for their assistance in developing the Landfalling Hurricane Probability Webpage (available online at http://www.e-transit.org/hurricane).

 

The second author gratefully acknowledges valuable input to his CSU research project over many years by former graduate students and now colleagues Chris Landsea, John Knaff and Eric Blake.  We also thank Professors Paul Mielke and Ken Berry of Colorado State University for much statistical analysis and advice over many years. 

 

DEFINITIONS


1        Introduction

 

This is the 23rd year in which the CSU Tropical Meteorology Project has made forecasts of the upcoming season’s Atlantic basin hurricane activity.  Our research team has shown that a sizable portion of the year-to-year variability of Atlantic tropical cyclone (TC) activity can be hindcast with skill exceeding climatology.  These forecasts are based on a statistical methodology derived from 52 years of past data and a separate study of analog years which have similar precursor circulation features to the current season.  Qualitative adjustments are added to accommodate additional processes which may not be explicitly represented by our statistical analyses.  These evolving forecast techniques are based on a variety of climate-related global and regional predictors previously shown to be related to the forthcoming seasonal Atlantic basin tropical cyclone activity and landfall probability.  We believe that seasonal forecasts must be based on methods that show significant hindcast skill in application to long periods of prior data.  It is only through hindcast skill that one can demonstrate that seasonal forecast skill is possible.  This is a valid methodology provided that the atmosphere continues to behave in the future as it has in the past. 

A variety of atmosphere-ocean conditions interact with each other to cause year-to-year and month-to-month hurricane variability.  The interactive physical linkages between these many physical parameters and hurricane variability are complicated and cannot be well elucidated to the satisfaction of the typical forecaster making short range (1-5 days) predictions where changes in the momentum fields are the crucial factors.  Seasonal and monthly forecasts, unfortunately, must deal with the much more complicated interaction of the energy-moisture fields with the momentum fields. 

We find that there is a rather high (50-60 percent) degree of year-to-year hurricane forecast potential if one combines 4-5 semi-independent atmospheric-oceanic parameters together.  The best predictors (out of a group of 4-5) do not necessarily have the best individual correlations with hurricane activity.  The best forecast parameters are those that explain the portion of the variance of seasonal hurricane activity that is not associated with the other variables.  It is possible for an important hurricane forecast parameter to show little direct relationship to a predictand by itself but to have an important influence when included with a set of 4-5 other predictors. 

In a five-predictor empirical forecast model, the contribution of each predictor to the net forecast skill can only be determined by the separate elimination of each parameter from the full five predictor model while noting the hindcast skill degradation.  When taken from the full set of predictors, one parameter may degrade the forecast skill by 25-30 percent, while another degrades the forecast skill by only 10-15 percent.  An individual parameter that, through elimination from the forecast, degrades a forecast by as much as 25-30 percent may, in fact, by itself, show much less direct correlation with the predictand.  A direct correlation of a forecast parameter may not be the best measure of the importance of this predictor to the skill of a 4-5 parameter forecast model.  This is the nature of the seasonal or climate forecast problem where one is dealing with a very complicated atmospheric-oceanic system that is highly non-linear.  There is a maze of changing physical linkages between the many variables.  These linkages can undergo unknown changes from weekly to decadal time scales.  It is impossible to understand how all these processes interact with each other.  It follows that any seasonal or climate forecast scheme showing significant hindcast skill must be empirically derived.  No one can completely understand the full complexity of the atmosphere-ocean system or develop a reliable scheme for forecasting the myriad non-linear interactions in the full-ocean atmosphere system.

 

 

 

2        Early April Forecast Methodology

 

Our initial early April seasonal hurricane forecast scheme demonstrated hindcast skill for the period of 1950-1995.  Our new, recently developed early April forecast scheme uses more hindcast years (1950-2001) and shows improved hindcast skill and better physical insights into why such precursor relationships have an extended period memory.

Through extensive analysis of NOAA-NCEP reanalysis products, we have recently developed a new set of 1 April extended range predictors which shows superior hindcast prediction skill over our previous 1 April forecast scheme.  The location of each of these new predictors is shown in Fig. 1.   The pool of six predictors for this extended range forecast is given in Table 1.  Strong statistical relationships can be extracted via combinations of these predictors (which are available by the end of March) and the amount of Atlantic basin hurricane activity occurring later in the year. 

 


Figure 1: Location of predictors for the early April forecast for the 2006 hurricane season. A plus (+) means that positive values of the parameter indicate increased hurricane activity this year, and a minus (-) means that positive values of the parameter indicate decreased hurricane activity this year. 

 

 

 

Table 1: Listing of 1 April 2006 predictors for this year’s hurricane activity.  A plus (+) means that positive values of the parameter indicate increased hurricane activity this year, and a minus (-) means that positive values of the parameter indicate decreased hurricane activity this year.  The combination of these six predictors calls for an active hurricane season.

 

Predictor

Values for 2006 Forecast

1) February 200 mb U (5°S-10°N, 35-55°W) (-)

+0.6 SD

2) February-March 200 mb V (35-62.5°S, 70-95°E) (-)

+0.3 SD

3) February SLP (0-45°S, 90-180°W) (+)

-1.4 SD

4) February SST (35-50°N, 10-30°W) (+)

+1.2 SD

5) Previous November 500 MB Ht. (67.5-85°N, 50°W -10°E) (+)

+0.6 SD

6) Previous September-November SLP (15-35°N, 75-95°W) (-)

-1.4 SD

2.1              Physical Associations among Predictors Listed in Table 1

 

Brief descriptions of our early April predictors follow:

 

Predictor 1.  February 200 mb U in Equatorial East Brazil (-)

 

(5°S-10°N, 35-55°W)

 

Easterly upper-level zonal wind anomalies off the northeast coast of South America imply that the upward branch of the Walker Circulation associated with ENSO remains in the western Pacific and that cool ENSO or La Niña conditions are likely to be present in the eastern equatorial Pacific for the next 4-6 months. El Niño conditions shift the upward portion of the Walker Circulation to the eastern Pacific and cause 200 mb westerly wind anomalies over the tropical Atlantic. These anomalies inhibit Atlantic hurricane activity. 

 

 

Predictor 2.  February-March 200 MB V in the Southern Indian Ocean (-)

 

(35-62.5°S, 70-95°E)

 

Anomalous winds from the north at 200 mb in the southern Indian Ocean are associated with a northeastward shift of the South Indian Convergence Zone (SICZ) (Cook 2000), a more longitudinally concentrated upward branch of the Hadley Cell near Indonesia and warm sea surface temperatures throughout most of the Indian Ocean. This also implies that warm ENSO conditions have likely been prevalent throughout the past several months due to the lag teleconnected effect of a warm Indian Ocean with a warm eastern Pacific Ocean. Strong lag correlations (r > 0.4) with this predictor indicate that a change in phase of ENSO from warm to cool is likely during the latter part of the spring/early summer.

 

Predictor 3.  February SLP in the Southeast Pacific (+)

 

(0-45°S, 90-180°W)

 

High sea level pressure in the eastern Pacific south of the equator indicates a positive Southern Oscillation Index (SOI) and stronger-than-normal trade winds across the Pacific. Increased trades drive enhanced upwelling off the west coast of South America that is typical of La Niña and hurricane-enhancing conditions in the Atlantic. Cool sea surface temperatures in the eastern Pacific are associated with these higher surface pressures and tend to persist throughout the spring and summer thereby reducing vertical wind shear over the tropical Atlantic and providing more favorable conditions for tropical cyclone development.

 

 

Predictor 4.  February SST off the Northwestern European Coast (+)

 

(35-50°N, 10-30°W)

 

Warm sea surface temperatures off the northwest coast of Europe correlate quite strongly with warm sea surface temperatures across the entire North Atlantic Ocean. A warm North Atlantic Ocean indicates that the thermohaline circulation is likely stronger than normal, the subtropical high near the Azores is weaker than normal and consequently trade wind strength across the Atlantic is also reduced. Weaker trade winds induce less upwelling which keeps the tropical Atlantic warmer than normal. This pattern tends to persist throughout the spring and summer implying a warmer tropical Atlantic during the hurricane season which is an enhancing factor for developing tropical waves.

 

 

Predictor 5.  Previous November 500 MB Geopotential Height in the far North Atlantic (+)

 

(67.5-85°N, 50°W-10°E)

 

Positive values of this predictor correlate very strongly (r = -0.7) with negative values of the Arctic Oscillation (AO) and the North Atlantic Oscillation (NAO). Negative AO and NAO values imply more ridging in the central Atlantic and likely also a warm North Atlantic Ocean (50-60°N, 10-50°W). Also, on decadal timescales, weaker zonal winds in the sub-polar areas are indicative of a relatively strong thermohaline circulation which is favorable for hurricane activity. Positive values of this November index are negatively correlated with both 200 mb zonal winds and trade winds the following September in the tropical Atlantic. The associated reduced tropospheric vertical wind shear enhances conditions for TC development. Other features that are directly correlated with this predictor are low sea level pressure in the Caribbean and a warm North and Tropical Atlantic the following summer. Both of the latter are also hurricane-enhancing factors.

 

 

Predictor 6.  Previous September-November SLP in the Gulf-SE USA (-)

 

(15-35°N, 75-95°W)

 

Low pressure in this area during September-November of the previous year correlates quite strongly with the positive phase of the PNA. According to Horel and Wallace (1981), the PNA is positive during the final year of most warm ENSO events. Therefore, a change to neutral or cool ENSO conditions is to be expected the following year. This feature is also strongly correlated with the following year's August-September sea level pressure in the tropical and subtropical Atlantic. August-September SLP in the tropical Atlantic is one of the most important predictors for seasonal activity, that is, lower-than-normal sea level pressures in the tropical Atlantic provide more favorable conditions for TC activity. In addition, easterly anomalies at 200 mb throughout the tropical Atlantic are typical during the following year's August-September period with low values of this predictor.

 

2.2              Hindcast Skill

 

Table 2 shows the degree of hindcast variance explained by our 1 April forecast scheme based upon our 52-year developmental dataset (1950-2001). To reduce over-fitting, we use no more than five predictors. Note that there is substantial skill for predictions of HD and NTC.

 

Table 2:  Variance explained based upon 52 years (1950-2001) of hindcasting.

 

Variables Selected

Variance (r2) Explained

Jackknife (r2)

NS – 1, 2, 4, 5

0.45

0.34

NSD – 1, 2, 4, 5 

0.59

0.50

H – 2, 3, 5, 6

0.53

0.41

HD – 1, 2, 5, 6

0.65

0.57

IH – 2, 3, 4, 5

0.61

0.53

IHD – 1, 2, 4, 5, 6

0.55

0.46

NTC – 1, 2, 4, 5, 6

0.71

0.64

 

 

3                   Analog-Based Predictors for 2006 Hurricane Activity

 

Certain years in the historical record have global oceanic and atmospheric trends which are substantially similar to 2006.  These years also provide useful clues as to trends in activity that the upcoming 2006 hurricane season may bring.  For this early April extended range forecast, we project atmospheric and oceanic conditions for August through October 2006 and determine which of the prior years in our database have distinct trends in key environmental conditions which are similar to current February-March 2006 conditions.  Table 3 lists our analog selections.

 

We select prior hurricane seasons since 1949 which have similar atmospheric-oceanic conditions to those currently being experienced.  Analog years for 2006 were selected primarily on how similar they are to conditions that are currently observed such as warm tropical and North Atlantic sea surface temperatures and cool ENSO conditions.  In addition, we look for analogs with similar conditions to what we project for August-October 2006 including warm Atlantic sea surface temperatures, neutral to cool ENSO conditions and west phase QBO conditions. 

 

There were four hurricane seasons since 1949 with characteristics similar to what we observe in February-March and what we project for August-September.  The best analog years that we could find for the 2006 hurricane season are 1964, 1996, 1999 and 2003.  We anticipate that 2006 seasonal hurricane activity will have slightly more activity than what was experienced in the average of these four years.  We believe that 2006 will be a very active season in the Atlantic basin. 

 

 

Table 3:  Best analog years for 2006 with the associated hurricane activity listed for each year.

 

Year

NS

NSD

H

HD

IH

IHD

NTC

1964

12

71.25

6

43.00

5

9.75

160

1996

13

79.00

9

45.00

6

13.00

192

1999

12

78.50

8

41.00

5

14.25

182

2003

16

79.25

7

32.75

3

16.75

174

Mean

13.3

77.0

7.5

40.4

4.8

13.4

177.0

2006 Forecast

17

85

9

45

5

13

195

 

 

4                   ENSO

 

We believe that neutral or weak La Niña conditions are likely to be present during August-October 2006.   A La Niña event is now in place in the tropical Pacific according to the Climate Prediction Center.  SOI values continue to remain positive, and trade winds in the central Pacific are anomalously strong.  In addition, oceanic heat content in the tropical Pacific remains below normal.  These features will likely contribute to keeping waters from becoming anomalously warm over the next few months.  In addition, most forecast models call for either neutral or La Niña conditions to persist for the next 4-6 months.  When the tropical Atlantic is warm and neutral or La Niña conditions are present, Atlantic basin hurricane activity is greatly enhanced. 

 

5                   Adjusted 2006 Forecast

 


Table 4 shows our final adjusted early April forecast for the 2006 season which is a combination of our derived full 52-year statistical forecast, our analog forecast and qualitative adjustments for other factors not explicitly contained in either scheme.  We foresee another very active Atlantic basin hurricane season.  We anticipate that ENSO will likely be somewhat cool and will therefore play an enhancing role for the 2006 season.  Warm sea surface temperatures are likely to continue being present in the tropical and North Atlantic during 2006, due to the fact that we are in a positive phase of the Atlantic Multidecadal Oscillation (AMO) (i.e., a strong phase of the Atlantic thermohaline circulation).

 

Table 4:  Summary of our new early April statistical forecast, our analog forecast and our adjusted final forecast for the 2006 hurricane season. 

 

Forecast Parameter and 1950-2000 Climatology (in parentheses)

New Statistical Scheme

 

Analog

Scheme

Adjusted

Final

Forecast

Named Storms (9.6)

10.6

13.3

17

Named Storm Days (49.1)

54.0

77.0

85

Hurricanes (5.9)

6.2

7.5

9

Hurricane Days (24.5)

28.2

40.4

45

Intense Hurricanes (2.3)

2.3

4.8

5

Intense Hurricane Days (5.0)

7.8

13.4

13

Net Tropical Cyclone Activity (100%)

125.9

177.0

195

 

 

 

6                   Skill and Verification of 1 April Forecasts

 

We define forecast skill as the degree to which we are able to predict the variation of seasonal hurricane activity parameters above that specified by a long-term climatology.  The latter is expressed as the ratio of our forecast error to the observed difference from climatology or:

 

Forecast Error / Seasonal Difference from Climatology

 

For example, if there were a year with five more tropical storms than average and we had predicted two more storms than average, we would give ourselves a skill score of 2 over 5 or 40 percent.  By this measure, each of the seven parameters of our seasonal forecasts has shown some degree of skill from 1 April.  Table 5 shows our skill based on 52 years of hindcasts from 1950-2001, and Table 6 displays our skill score in real-time forecasting for the last seven years.  All parameters of our real-time forecasts have shown skill from 1 April.

 

Table 5:  Average percent of variation explained of 1 April hindcasts above that of climatology (in percent) for the 52-year period 1950-2001.  A value of 40 means that we hindcast 40 percent of the variability from climatology or that we were unable to explain 60 percent of the variability from climatology. 

 

Tropical Cyclone

Parameter

Early April

Hindcast Skill

NS

31

NSD

38

H

36

HD

40

IH

40

IHD

34

NTC

47

 

Table 6:  Last seven years’ (1999-2005) average percent of variation explained of our ‘real-time’ forecasts issued on 1 April above that of climatology (in percent).  A value of 30 means that we hindcast 30 percent of the variability from climatology or that we were unable to explain 70 percent of the variability from climatology. 

 

Tropical Cyclone

Parameter

Early April

Forecast Skill

NS

35

NSD