MONTHLY TROPICAL WEATHER SUMMARY NATIONAL WEATHER SERVICE MIAMI FL 2 PM EST MON NOV 29 1993 FOR THE NORTH ATLANTIC...CARIBBEAN SEA AND THE GULF OF MEXICO... ...SUMMARY OF TROPICAL CYCLONE ACTIVITY FOR 1993... THE HURRICANE SEASON OFFICIALLY ENDS TOMORROW AND...TO DATE...THIS HAS BEEN A RELATIVELY INACTIVE YEAR FOR TROPICAL CYCLONES. THERE WERE 8 TROPICAL STORMS IN 1993...AND 4 OF THOSE WENT ON TO BECOME HURRICANES. THESE TOTALS ARE BELOW THEIR RESPECTIVE LONG-TERM AVERAGES BY 2. ASIDE FROM EMILY...THE 1993 HURRICANES WERE VERY SHORT-LIVED. THE LAST NAMED SYSTEM ENDED ON 21 SEPTEMBER. BARRING ANY FURTHER DEVELOPMENTS...THIS IS THE EARLIEST THAT TROPICAL STORM AND HURRICANE ACTIVITY HAS CONCLUDED SINCE 1930. SEVERAL OF THE TROPICAL CYCLONES...INCLUDING ONE OF THE TWO TROPICAL DEPRESSIONS THIS YEAR...MADE LANDFALL. THEY EACH PRODUCED LOCALLY HEAVY RAINS AND THEIR ASSOCIATED FLOODS RESULTED IN A SEASON TOTAL OF 280 DEATHS ACROSS THE ATLANTIC HURRICANE BASIN. THE YEAR'S ONLY MAJOR HURRICANE... EMILY...CLIPPED THE OUTER BANKS OF NORTH CAROLINA AT NEAR PEAK STRENGTH. 1993 ATLANTIC HURRICANE SEASON STATISTICS NUMBER NAME CLASS DATES HIGHEST LOWEST U.S. DEATHS 1-MIN WIND PRESSURE DAMAGE (M.P.H.) (MB) ($MILLIONS) 1 ARLENE TS 18-21 JUN 40 1000 22+ 6 2 BRET TS 4-11 AUG 60 1002 184 3 CINDY TS 14-17 AUG 45 1007 4 4 DENNIS TS 23-28 AUG 50 1000 5 EMILY H 22 AUG-6 SEP 115 960 35 3 6 FLOYD H 7-10 SEP 75 990 7 GERT H 14-21 SEP 100 970 76 8 HARVEY H 18-21 SEP 75 990 + TOTAL DOES NOT INCLUDE DAMAGE TO CROPS IN SOME TEXAS COUNTIES. ARLENE FORMED JUST NORTHWEST OF THE YUCATAN PENINSULA ON 18 JUNE AND HEADED SLOWLY TOWARD THE LOWER TEXAS COAST. IT REMAINED POORLY ORGANIZED UNTIL REACHING TROPICAL STORM STRENGTH ABOUT 100 MILES EAST OF BROWNSVILLE. THE CENTER MADE LANDFALL OVER NORTHERN PADRE ISLAND TEXAS WITH SUSTAINED WINDS NEAR 40 MPH ON THE MORNING OF THE 20TH. ARLENE THEN WEAKENED. WIDESPREAD HEAVY RAINS AND LOCAL FLOODS OCCURRED FROM SOUTHEASTERN MEXICO THROUGH COASTAL TEXAS. IN ADDITION...THE PRE-ARLENE DISTURBANCE APPARENTLY INDUCED A LANDSLIDE IN EL SALVADOR THAT KILLED 20 PEOPLE. TROPICAL STORM BRET FORMED ON 4 AUGUST EAST OF THE WINDWARD ISLANDS. IT MOVED ALMOST DUE WESTWARD ACROSS THE NORTH COAST OF SOUTH AMERICA AND THEN HIT NICARAGUA ON THE 10TH. IT CROSSED FROM CENTRAL AMERICA INTO THE EASTERN PACIFIC ON THE 11TH...WHERE IT LATER BECAME EASTERN AND CENTRAL PACIFIC HURRICANE GREG. IN THE ATLANTIC IT REACHED A MAXIMUM INTENSITY OF ABOUT 60 MPH. MOST OF THE 184 FATALITIES RELATED TO BRET CAME FROM FLOODS AND MUD SLIDES IN THE ENVIRONS OF CARACAS VENEZUELA. TROPICAL STORM CINDY FORMED NEAR THE ISLAND OF MARTINIQUE ON AUGUST 14. AFTER REACHING A WIND SPEED OF 40 MPH...IT DISSIPATED OVER THE ISLAND OF HISPANIOLA ON THE 17TH. TWO PEOPLE DIED ON MARTINIQUE AND TWO FATALITIES OCCURRED IN THE DOMINICAN REPUBLIC. TROPICAL STORM DENNIS MOVED NORTHWESTWARD OVER THE EASTERN ATLANTIC DURING ITS SIX DAY LIFETIME. ITS MAXIMUM WIND SPEED IS ESTIMATED TO HAVE BEEN 45 MPH. EMILY FORMED ON 22 AUGUST ABOUT 600 MILES EAST-NORTHEAST OF THE LEEWARD ISLANDS. IT FIRST REACHED HURRICANE STRENGTH ON THE 26TH. AFTER TEMPORARILY WEAKENING...IT REGAINED HURRICANE INTENSITY THE FOLLOWING DAY. EMILY FOLLOWED A COURSE THAT VARIED BETWEEN WEST AND NORTHWEST...EVENTUALLY BRINGING THE WESTERN PORTION OF THE EYEWALL OVER THE OUTER BANKS OF NORTH CAROLINA ON 31 AUGUST. THREE DEATHS IN ROUGH SURF WERE REPORTED...TWO ALONG THE NORTH CAROLINA COAST AND ONE IN VIRGINIA. THE HURRICANE CAUSED SERIOUS DAMAGE OVER A PORTION OF THE OUTER BANKS. EMILY TURNED AWAY FROM THE UNITED STATES ON THE 1ST OF SEPTEMBER AND WEAKENED. IT DISSIPATED WHILE BECOMING EXTRATROPICAL ON THE 6TH OF SEPTEMBER. FLOYD DEVELOPED ON 7 SEPTEMBER OVER THE SOUTHWESTERN ATLANTIC. IT ACCELERATED TO ABOUT 50 MPH OVER THE FOLLOWING THREE DAYS ON A TRACK WHICH TURNED FROM NORTHWARD TO EAST-NORTHEASTWARD. FLOYD REACHED HURRICANE STRENGTH LATE ON THE 9TH SEVERAL HUNDRED MILES TO THE SOUTHEAST OF NOVA SCOTIA. IT LOST ITS TROPICAL CHARACTERISTICS ON THE 10TH...AND EVENTUALLY HIT FRANCE AS A VIGOROUS EXTRATROPICAL STORM. GERT DEVELOPED OVER THE SOUTHWEST CARIBBEAN SEA ON 14 SEPTEMBER. IT BECAME A TROPICAL STORM...AND HIT NICARAGUA...ON 15 SEPTEMBER. FOR THE NEXT THREE DAYS...IT MOVED OVER THE LAND MASSES OF CENTRAL AMERICA AND THE YUCATAN PENINSULA. IT THEN ENTERED THE OPEN WATERS OF THE BAY OF CAMPECHE. GERT STRENGTHENED INTO A HURRICANE ON THE 20TH AND STRUCK THE COAST OF MEXICO JUST TO THE NORTH OF THE CITY OF TUXPAN THAT SAME DAY WITH SUSTAINED WINDS TO NEAR 100 MPH. GERT CLAIMED A TOTAL OF 76 LIVES IN MEXICO AND CENTRAL AMERICA... PRIMARILY DUE TO VERY HEAVY RAINS...AND THE ASSOCIATED FLOODING. HARVEY REMAINED AT SEA...MOVING NORTHEASTWARD ACROSS THE CENTRAL ATLANTIC FROM 18-21 SEPTEMBER. IT BECAME EXTRATROPICAL ON THE 21ST. TROPICAL DEPRESSION ONE FORMED IN THE NORTHWESTERN CARIBBEAN SEA AT THE END OF MAY AND CROSSED WESTERN CUBA AND THE NORTHWESTERN BAHAMAS ON 1 JUNE...BEFORE LOSING ITS TROPICAL CHARACTERISTICS. FLOODS CAUSED 7 DEATHS IN CUBA. TROPICAL DEPRESSION TEN WAS A SHORT-LIVED CYCLONE OVER THE WEST- CENTRAL ATLANTIC AT THE END OF SEPTEMBER. RAPPAPORT/PASCH ------------------------------------------------------------------------------- ******************************************* 1993 Atlantic Hurricane Season summary and 1994 Atlantic Hurricane Season Forecast ******************************************* Dr. William Gray, Colorado State University This message summarizes the tropical cyclone (TC) activity which occurred in the Atlantic Basin during 1993, and verifies the author's seasonal forecasts of this activity which were issued on 24 November of last year and updated on 5 June and 4 August of this year. The 1993 hurricane season was relatively inactive with eight named storms (average 9.3) and four hurricanes (average 5.7). There was only one major (or intense) Saffir/Simpson category 3-4-5 hurricane (average is 2.2) and this storm (Emily) was only of marginal category 3 intensity for slightly less than one day. The seasonal total of hurricane days was only 10 (average is 23) and the seasonal total of named storm days was just 30 (average is 46). Only two systems (tropical cyclone Arlene in south Texas and Hurricane Emily near Cape Hatteras) affected the US mainland. The author had forecast a substantially more active hurricane season. This forecast failure was primarily the result of the misjudgment of the remarkably persistent El Nino conditions which occurred this year and of the high Caribbean basin surface pressure which became established in August. The Caribbean pressure changes are believed to be unrelated to the El Nino and are a result of the intertropical convergence line being shifted southward over the South American continent. Additional unfavorable hurricane conditions included a continuing West Sahel drought. It is quite unusual for an El Nino to last through three consecutive seasons as the current El Nino event has. Actually, warm water has persisted in the equatorial Pacific through the last four hurricane seasons. The El Nino is the most dominant of the climate influences affecting Atlantic hurricane activity, tending to strongly suppress this activity. All tropical cyclone activity ended after the 21st of September, a date which is normally only 60 percent of the way through the Atlantic season. Table 1 lists statistical information for the author's 1993 seasonal forecast. All parameters were overforecasted. Notably, the number of named storm days, hurricane days, and HDP were all badly overcast. As is discussed throughout this report, these errors were largely the result of the unanticipated August reintensification of the El Nino plus the unexpected strong rise in August-October Caribbean basin surface pressure. TABLE 1: Verification of 1993 hurricane activity forecast from various dates of forecast. 1993 Fcst 1993 1993 made on Fcst Fcst 1993 24 Nov as of as of Verifi- 1991 5 June 4 Aug cation ____________________________________________________________________ Named Storms (NS) 11 11 10 8 Named Storm Days (NSD) 55 55 50 30 Hurricanes (H) 6 7 6 4 Hurricane Days (HD) 25 25 25 10 Intense Hurricanes (IH) 3 2 2 1 Intense Hurricane Days (IHD) 7 3 2 0.75 Hurricane Destruction Potential (HDP) 75 65 55 23 Net Tropical Cyclone Activity (NTC) NOT MADE 95 85 55 Other considerations contributing to overforecasts of 1993 Activity were the inappropriate qualitative alteration of our own recently developed statistical forecast values. Our objective forecasts proved to be superior to the adjusted forecasts, particularly for the parameters most in error. Statistical forecasts with verification are given in Table 2. TABLE 2: Statistical forecast model results for 1993 from Gray, Landsea, Mielke, and Berry (1992, 1993a,b). Forecast Parameter and Forecast Observed its last 43 yr. ave. ----------------------- -------- Verification 24 Nov. 4 June 5 Aug. 1993 1992 1993 1993 Name Storms (NS) (9.3) 9.7 11.7 9.8 8 Name Storm Days (NSD) (46.1) 45.8 53.6 52.1 30 Hurricanes (H) (5.7) 5.6 9.0 6.5 4 Hurricane Days (HD) (23) 18.1 12.5 23.9 10 Intense Hurricanes (IH) Cat. 3-4-5 (2.2) 1.8 -0.3 1.7 1 Intense Hurricane Days (IHD)(4.5) 3.1 -2.1 1.1 0.75 Hur. Destruction Potential (HDP) (68.1) 48.8 21.8 50.4 23 Net Tropical Cyclone Activity (NTC) (100%) 82% 73% 72% 55% --------------------------------------------------------------------------- ------------------------------------------------------------------------ EXTENDED RANGE FORECAST OF ATLANTIC SEASONAL HURRICANE ACTIVITY FOR 1994 ------------------------------------------------------------------------ This message discusses details of a 6-11 month extended range seasonal forecast of the tropical cyclone activity likely to occur in the Atlantic Ocean basin during 1994. This forecast is based on new research by the author and his colleagues which allows estimates of next season's Atlantic tropical cyclone activity to be made by November of the prior year. The forecast scheme is based on a 10-month forward extrapolation of the Quasi-Biennial Oscillation (QBO) of equatorial stratospheric zonal wind, two measures of West African rainfall through mid-November 1993 and an extended range forecast of El Nino conditions for August to October 1994. Information through mid-November 1993 indicates that the 1994 Atlantic hurricane activity is likely to be somewhat above average with about 6 hurricanes (average 5.7), 10 named storms (average 9.3), 25 hurricane days (average 23), 2 intense hurricanes (average 2.2) and a hurricane destruction potential of 85 (average 68). The 1994 season should be more active than the three recent 1991, 1992 and 1993 hurricane seasons and, in particular, more active in the tropical regions (at latitudes south of 25 deg N) where only one short lived hurricane has occurred during the last three years. The 1994 season should be more like the hurricane seasons of 1988 and 1989 which produced a total of five major hurricanes. The probability of hurri- cane destruction along the US East Coast, Peninsula Florida, and within the Caribbean basin for 1994 is projected to be somewhat greater than the mean probability for the last 40 years and is distinctly higher than the probabilities for the last three years. It is expected that the long running 1990-1993 El Nino-like warm water conditions will finally dissipate by the next hurricane season and that comparatively cool surface water conditions will be present in the east and central equatorial Pacific; this trend should enhance hurricane activity. It is also expected that West African Sahel drought will abate somewhat next season. This upward rainfall trend should also diminish the inhibiting influences on hurricane activity than have occurred in recent years. Stratosphere QBO conditions will be in the easterly phase which will tend to inhibit tropical cyclone activity. But in the net hurricane activity should be enhanced over recent years. Table 1 gives our statistical and adjusted prediction. TABLE 1: Statistical prediction as obtained with final adjusted forecast and comparison with the last 43-year average. Qualitative Percentage Forecast Statistical Actual of last 43 Parameter Forecast Forecast Year Mean Named Storms (NS) 9.72 10 108% Named Storm Days (NSD) 60.56 60 130 Hurricanes (H) 6.40 6 105 Hurricane Days (HD) 24.95 25 107 Intense Hurricanes (IH) 1.92 2 92 Intense Hurricane Days (IHD) 7.19 7 150 Hurricane Destruction Potential (HDP) 84.83 85 122 Net Tropical Cyclone Activity (NTC) 109.10 110 110 Discussion Implicit in this forecast is the anticipated dissipation of the long running equatorial Pacific warm water event which has now persisted for nearly four consecutive years. Historical records going back to the 1850s (see Wright, 1989) indicate that above average water temperatures in the equatorial Pacific have never persisted for more than four consecutive years and that three and four year warm events of the sort we have just had are very rare. Our extended range ENSO prediction scheme forecasts a NINO-3 sea surface temperature anomaly of -0.33 deg C for the August-October 1994 period. This ENSO forecast showed a measure of agreement of 0.72 in 34 years of cross-validated hindcast forecasts for the years of 1959-1992. This ENSO forecast supports an end to the long running El Nino-like warm water conditions by next hurricane season. The inhibiting influence on hurricane activity of warm equatorial Pacific water temperature for next year is believed to be very low. Consequently, hurricane activity should be higher. In addition, our extended range estimate of West African precipitation for next year gives slight positive precipitation values (+0.12 S.D.). This West Sahel rainfall forecast has a measure of agreement of 0.67 (or two-thirds of the variability) in 34 years of cross-validated hindcast forecast tests for the years of 1959-1992. This indicates that we should likely see a break in the long running Western Sahel drought for next year. Rainfall amounts should be higher than have occurred during most of the last 25 years. Recent August to mid-November 1993 Gulf of Guinea rainfall has been slightly above average for the first time since 1988. This is another indication that next year's seasonal hurricane activity should be above what has been observed for the last three years. Thus, of our three primary extended range predictor, two (ENSO and West African rainfall) appear favorable for enhanced hurricane activity with only the QBO as an unfavorable indicator for increased activity. Next year's hurricane activity should be enhanced over that of the last three years, at least.