sign in sign up
<<
Home , Atlantic hurricane, Pacific hurricane

THE INTERANNUAL VARIABILITY OF HURRICANE ACTIVITY IN THE ATLANTIC AND EAST PACIFIC REGIONS

Anthony R. Lupo Tamara K. Latham Trenton H. Magill Joseph V. Clark Christopher J. Melick Patrick S. Market Department of Soil, Environmental, and Atmospheric Sciences University of Missouri – Columbia Columbia, Missouri

Abstract

The investigation of the interannual and interdecadal variations in hurricane activity has been an important topic of study lately, especially with regard to their implications for change issues. On the interannual time-scale, the El Niño and Southern Oscillation (ENSO) phase has

been correlated with hurricane activity in the Atlantic and Eastern Basins. For example, various atmospheric and oceanic parameters that influence hurricane development Thisbecome study significantly examined alteredthe interannual during an variability El Niño event, of hurricane leading intensityto suppressed (measured easterly as windwave speed development and growth in the Atlantic, but more activity in the Eastern Pacific Ocean Basin.

and interpreted through the Saffir-Simpson Scale) from 1938 through 2007 in the Atlantic and 1970 through 2007 in the Pacific basins, respectively. These data were then compared with the Decadaloccurrence Oscillation of El Niño/La (PDO) Niña was eventsalso examined as defined here. using Not the surprisingly, Japan Meteorological during an El Association Niño year the intensity(JMA) index. of Atlantic El Niño/La hurricanes Niña variability was found superimposed to be weaker onthan variability during a associated neutral year with or athe La Pacific Niña

year, but these conclusions were opposite in the Eastern Pacific Ocean Basin. There were also Thissignificant study alsodifferences examined found the interannualin hurricane variation intensity inbetween hurricane El intensityNiño and byLa genesisNiña years region when (i.e. the PDO was in phase 1 (warm phase), rather than when the PDO was in phase 2 (cool phase).

Atlantic: the eastern and western Basins, the Caribbean, and the Gulf of ; Eastern Pacific: divided into quadrants using 20o N and 125o W as the quadrant intersection point). Finally, the utility of this information in a long-range forecast application is demonstrated. Department of Soil, Environmental, and Atmospheric Sciences Corresponding author address: Dr. Anthony R. Lupo

302 E Anheuser BuschEmail: Natural [email protected] Resources Building, University of Missouri Columbia, Columbia, Missouir 65211 Lupo et al.

1. Introduction activity has been studied extensively over the last few an active tropical (e.g., Zuki and Lupo 2008). decades,While there the has interannual been a recent variability increase in of the hurricane interest and Thus,place the it into purpose context of this with work respect is three-fold: to the earlier 1) Examine work of the interdecadal variability of hurricane activity Atlantic Ocean Basin hurricane activity from 2000 – 2007 season provides a case study, and a rationale for the of LJ00. In performing this analysis, the 2005 hurricane been(e.g., prompted Henderson-Sellers by several et active al. 1998; hurricane Emanuel, years in 2005; the Webster et al. 2005; Schultz, 2007). This interest has excessive activity is given. 2) Re-analyze Atlantic Ocean basin activity published in LJ00. Some changes in the Atlantic Ocean Basin, as well as concern about changes classification of ENSO years have taken place recently since in hurricane frequency that may be a result of changes Predictionthe publication Studies, of LJ00 available (e.g., 1974 online has at: been http:// re-classifiedwww.coaps. as in Earth’s climate. These active years included the 2004 fsu.eda La Niñau year, see the Center for Ocean and Atmospheric season during which several struck Florida, and the 2005hereafter season LJ00 which was the most active hurricane ). There were also changes in the classification of interdecadalseason since 1933.variability Some in studies, the occurrence (e.g., Lupo and and intensity Johnston of changessome hurricanes would not(e.g., make Andrew, substantial 1992, was changes reclassified in the as Atlantic2000, region hurricane), suggested activity, thatand thatthere this is significantvariability a category 5 ). While it is hypothesized that these

LJ00 , an update is necessary to preserve the inmay the be linked behavior to the of Pacific interannual Decadal variability Oscillation of (PDO). tropical integrity of the previously published results. 3) Apply the cycloneMore occurrences specifically, inLJ00 the found North that Atlantic there was as related a change to methodologies from LJ00 to hurricane activity in the East the El Niño and Southern Oscillation (ENSO). During the ofPacific the Ocean hurricane Basin, activity as a similar into sub-basins, comprehensive which study is not of availableEast Pacific elsewhere. Ocean hurricane Lastly, an activity, analysis including of tropical a partition storm ENSO related variability in the number and intensity of hurricanesnegative phase basin of the wide. PDO However, (1947-1976), during they the found positive little here. activity for both basins following Schultz (2007) is added was strong interannual variability between El Niño and 2. Data and Methods Laphase Niña of years. the PDO During (1977 La Niña – 1999) years, they there found were thatmore there and The data and methods are similar to those used by stronger hurricanes overall even though there were fewer can be found there and in references therein. Thus, only LJ00, and more detail regarding some of these subjects Lamonths Niña with years. significant Additionally, hurricane studies activity. have shown Also, there that were more Caribbean and Gulf region hurricanes during a brief review is given here and differences between this a.study Data and LJ00 will be highlighted. negativesince 1999 phase (see of Lupo the PDO.et al. 2007 and references therein), The tropical occurrence and intensity data the Pacific Ocean Basin region has reverted back to the

AtlanticRecently, between Vimont accumulated and Kossin cyclone (2007) energyas well (whichas Kossin is [available(1938 through online 2006 at: –http://weather.unisys.co Atlantic, 1970 though m 2006], and – positivelyand Vimont, correlated (2007) found with a tropicalstrong correlation cyclone activity) in the North and East Pacific) were downloaded via the UNISYS website complete description of these data sets and their reliability these data can also be found in other archives (e.g., LJ00). A the Atlantic Meridional Mode (AMM). In their work, they also find that the Atlantic Multi-decadal Oscillation (AMO) (e.g., Landsea 1993) can be found in Jarvinen et al. (1984) state,may excite correlate the AMM with on tropical a decadal cyclone time scale. activity Thus, on these the for the Atlantic, and Davis et al. (1984) for the East Pacific. indexes, which quantify the climatological background tropicalIn this study, storms we in use their the analysis, term tropical and here cyclone tropical to includestorms referboth hurricanesonly to those and entities tropical that storms. obtained LJ00 maximum did not include wind variabilitydecadal time as related scale (AMO) to the orPDO both in the the tropical interannual and mid- and latitudedecadal background (AMM) time-scales. climatological Additionally, state have interdecadal also been speeds between 35 – 64 kt. data in the Atlantic basin before 1938 were not used in the study as studiesdemonstrated have in discussed other studies the synergistic (e.g., Deser relationshipet al. 2004; explained in LJ00; and following their lead, this study only betweenDeser et oceanical. 2006; and Deser atmospheric and Phillips conditions 2006), and needed several for uses Saffir-Simpson (Simpson 1974) hurricane intensity data. In order to eliminate problems with some of the data 2 National Digest as discussed in LJ00, we combined hurricane intensity The Interannual Variability of Hurricane Activity in the Atlantic and East Pacific Regions

Simpsoncategories intensity (category scale 1 and precluded 2 - weak; thecategory use of3, 4, tropical and 5 Arithmetic means and correlations were analyzed, and – intense) following Landsea (1993). The use of the Saffir- means were tested for statistical significance using a since before this date, most storms were simply assigned two-tailed z-score statistical test, assuming test. These the distributions null hypothesis were cyclone information prior to 1970 in the East Pacific basin tested(e.g., Neter using et the al.2 total1988). sample Intensity climatology distributions as the were expected also tested using a χ b.as Methodologya category 1 (e.g., see Schultz 2008). The complete archive for the years discussed above frequency and a sub-period as the observed frequency. thanIt has using been anhypothesized approximated that distributionusing the climatological since such analyticalfrequency as distributions the “expected” (e.g., frequency Poisson is distribution) more appropriate may for the Atlantic Ocean Basin (Fig. 1a) and the East Pacific OceanFig. 1a. Basin (Fig. 1b) were examined in this study. real-world distributions not adequately represent Atlantic Basin Hurricane Tracking Chart, National Hurricane Center, Miami, FL. (e.g., Lupo et al. 1997; LJ00). revealsIt should strong be cautioned relationships that betweenwhile statistical two variables, significance it does not

are found imply to cause be strong, and but effect. not Conversely, relationships that still have underlying causes duestatistically to some significant atmospheric may forcing process or mechanism

(e.g., Lupo et al. 1997; LJ00). since Finally, the nominal the data tropical were stratified by calendar year

Fig. 1a. East Tracking Chart, National Hurricane Center, Miami, FL cyclone season is recognized as starting on 1 June and ending on 30 November (15 May and 31 October) in the theAtlantic annual (East Pacific)and monthly Ocean distributionsBasin. We then of analyzed tropical cyclone occurrence in order

term trends in season length withinto find both both the long total and sample short- and each intensity category.

by sub-basin in the Atlantic The sample was also stratified

Ocean Basin following LJ00. o o Fig. 1. The East Pacific Ocean Basin N. Hurricaneswas stratified (tropical into quadrants storms) The maps of a) the Atlantic Ocean Basin (G = Gulf, C = Caribbean, W (E) = West weredivided assigned by 125 toW and the 20 basin (East) Atlantic), and the b) Pacific Ocean Basin (NW = Northwest, NE = Northeast, SW= Southwest, and SE= Southeast) used in this study and provided by the National Hurricane hurricane (tropical storm) Center. The heavy lines represent the boundaries between sub-basins used in this study. Their status.in which they first reached boundaries are outlined in section 2b and LJ00. The diagonal line across Mexico represents a dividing line between the Atlantic and Pacific Basins. Volume 32 Number 2 ~ December 2008 3 Lupo et al. c. El Niño and Pacific decadal oscillation definitions La Niña (LN) Neutral (NEU) El Niño (EN)

As in LJ00 and references therein, the Japan this study. A list of El Niño (EN), La Niña (LN), and Neutral 1938 1939 1940 Meteorological Agency (JMA) ENSO Index was used in 1942 1941 1951 (NEU) years used here is displayed in Table 1. In addition 1944 1943 1957 summary,to consulting the LJ00, index auses more long-term detailed running description mean of seathe 1949 1945-1948 1963 JMA ENSO Index can be found on the COAPS website. In 1954 - 1956 1950 1965 surface temperature (SST) anomalies from the Niño 3 and 1964 1952 – 1953 1969 3.4 regions in the central and eastern tropical Pacific (e.g.,o 1967 1958 – 1962 1972 Pielke and Landseao 1999). The SST anomaly thresholds 1970 - 1971 1966 1976 used to define EN years are those greater than +0.5 C, 1973 - 1975 1968 1982 less than -0.5 C for LN years and NEU otherwise. These thresholds are also similar to those used by Schultz (2007) 1988 1977 – 1981 1986 – 1987 for his study of East Pacific hurricane activity. The JMA 1998 - 1999 1983 – 1985 1991 ENSO criterion defined the El Niño year as beginning on 1 (2007) 1989 – 1990 1997 October of the previous year. This definition, however, was modified following LJ00 so that the El Nino year coincided 1992 – 1996 2002 with the Atlantic and East Pacific hurricane . 2000 – 2001 2006 Table 1. A list of years examined in this study separated by The PDO (also known as the North Pacific Oscillation ENSO phase. The current2003-2005 year is in parenthesis and is not [NPO] in some studies) is a 50 to 70 year oscillation recently described (e.g., Minobe 1997; Mantua et al. 1997) within the eastern Pacific Ocean Basin. As defined by Gershanov officially classified yet. and Barnett (1998), the positive (warm) phase of the PDO (Fig. 2) is characterized by an anomalously deep Aleutian in Gershanov and Barnett (1998) and Lupo et al. (2007). Low. Cold western and central north Pacific waters, warm tied Landsea to changes (1993), in the Gray long-term et al. pressure (1997), and patterns Landsea in the et eastern Pacific coastal waters, and warm tropical Pacific al. (1999) have demonstrated that hurricane activity is waters also characterize this phase of the PDO. Following LJ00, this phase is referred to as PDO1. The reverse levelAtlantic pressure Ocean Basin and other(e.g., the background North Atlantic variables Oscillation, (e.g., conditions characterize the negative (cool) phase of PDO windor NAO). shear) Deser have et al. displayed(2004) also interdecadal demonstrated variations, that sea (Fig. 2) and these conditions are referred to as PDO2 (e.g., LJ00). Each PDO period, or era, is defined using calendar years (see Table 2) and thisPacific information Decadal can Oscillation also be found which Vimont and Kossin (2007) correlated to Atlantic manifestedtropical cyclone by changes activity. in LJ00 the ENSO-found that the influence of the PDO was Postive (warm) phase (PD01) Negative (cold) phase (PD02) that there was little or no ENSO related variability, specifically,

andrelated more) variability intense hurricanesduring PDO2 during and significantly fewer and less (more

variationsEl Niño (La in Niña) tropical years cyclone during activity PDO1. canAs stated make by interpretation LJ00, the NAO-related of PDO- related hurricane variability more

time scales between the PDO and Fig. 2. The phases of the PDO as described in the text. SST anomalies are shown in long-termdifficult given variations the overlap in the in NAO. the

these decade-to-decade variations color in accord with the scale shown (degrees C). The arrows showu anomalies in wind inMany Atlantic studies hurricanes have speculated are forced that by speed and direction. Reproduced with permission from the Joint Institute for the Study of and Ocean (JISAO– http://jisao.washington.ed ) at the University of 4 National Weather Digest Washington. The Interannual Variability of Hurricane Activity in the Atlantic and East Pacific Regions a combination of both. However, no studies separating hurricane activity by phase of the PDO have been from 1995 – 2000, there were 49 hurricanes (81 tropical ), while during the years from 2001 – 2006, there describedperformed inon thisEast paragraphPacific Ocean would basin be hurricane the opposite activity of season.have been Two 47 otherhurricanes six-year (94 periods tropical werecyclones), compared in spite to as of yet, and it is hypothesized here that the relationships theseof the twofact recentthat 15 eras hurricanes and both occurred of these during earlier the periods 2005 is based on the spatial pattern of the relative warmth those found in the Atlantic Ocean Basin. Our hypothesis occurred before the satellite era. In particular, 43 (37) similarof the Eastern to that ofPacific the correlation SSTs for each of thephase ENSO of thephase PDO with as tohurricanes consider andthe activity62 (74) of tropical the pre-satellite cyclones occurredera periods from as described above (and shown in Fig. 2). This hypothesis is comparable1884-1889 (1932-1937).to the modern Thus,eras, since it would it is likely be reasonable that some others). tropical cyclones were missed during the pre-satellite this region’s tropical cyclone activity (see Shultz 2008 and era. PDO Phase Period of Record

Additionally, during the 2000 – 2007, there were 29 Phase 1 1933 - 1946 intense hurricane (category 3 – 5), which represented Phase 2 1947 - 1976 47% of all storms (Table 3). This compares with the LJ00 thestudy ratio in which of intense 42% storms of all storms overall. were The intense increased during activity the Phase 1 1977 - 1998 previous 62 years. This indicates only a slight increase in TablePhase 2. 2 1999 - of the latest period, however, is less than 50%, which was Phases of the Pacific Decadal the number of intense storms for the 30-year period of 3. AtlanticOscillation Ocean (PDO) Basin since Hurricane 1933. Activity Revised and Extended to 2000 - 2007 years1940-69 does (Table not make 3). Thus, any noteworthyas shown in changeTables 3in and the 4ratio for ofthe hurricaneentire 69-year intensities record, for the the inclusion overall of Atlantic the last Oceanseven As stated in the introduction, there have been some basin activity. The addition of tropical storm data does not the latest seven hurricane seasons have been added since changes to the original data sets used by LJ00. Additionally, significantly change the overall results of LJ00 either. and tropical storm information isthe included original here. publication Thus, since of LJ00, this category 1940 - 69 1970 - 89 1990 - 99 2000 - 07 All TS this section will focus only on work is partially an update of LJ00, 3.7 4.3 4.5 7.9 4.4 Cat. 1, 2 2.9 3.4 3.8 4.1 3.4 mostmajor recent changes activity. in the Atlantic basin Cat. 3 - 5 2.9 1.6 2.6 3.6 2.5 climatology since LJ00, as well as the Tot hur. a. activity: 2000 – Cat. 4, 5 1.4 0.9 1.4 2.3 1.3 2007 5.8 5.0 6.4 7.8 5.9 TableTS+Hur. 3. The mean annual9.5 number of9.3 hurricane and10.9 tropical storm15.7 (TS) events 10.3within

AtlanticDuring hurricanes, the 2000-2007 which tropical means the Atlantic Ocean Basin over decadal time periods corresponding with similar studies. cyclone seasons, there were 62 Hurricanes are stratified by weak hurricanes (Cat. 1 and 2), intense hurricanes (Cat. 3, that the years since the 4, and 5), and Cat. 4 + Cat. 5. Categories are based on the Saffir-Simpson scale. All TS Cat 1 Cat 2 Cat 3 Cat 4 Cat 5 Thur LN LJ00 study were the most Atlanticactive during tropical the cyclone1938 – 11.9 5.4 2.1 1.5 1.2 1.0 0.7 6.5 2007 record. However, the EN NEU 10.4 4.2 2.4 0.9 1.4 1.0 0.4 6.2 been relatively active. Total Duringseasons the since six year 1995 period have 8.7 3.9 2.2 0.9 1.1 0.4 0.2 4.7 Table 4. 10.4 4.4 2.3 1.1 1.2 1.0 0.4 5.9 The average annual occurrence of Atlantic Ocean Basin tropical storms and hurricanes (from 1938-2007) versus El Niño / La Niña phase forVolume each Saffir-Simpson 32 Number 2category. ~ December 2008 5 Lupo et al. b. Trends and interannual variability

none of these trends were statistically significant (not shown) at the 95% confidence level. The trend in tropical LJ00 reported that there were no statistically storms and category 1 hurricanes (not shown) would show stormsignificant occurrences trends in does the not Atlantic result Ocean in any basin change hurricane in the an increase stronger than that in Fig. 3, which shows the activity. Adding the 2000-2007 hurricane and tropical isoverall real trendis clouded for the by total two number issues. of The hurricanes trend could (significant be the at the 90% confidence level). Whether this upward trend trends for the occurrence of category 2 – 5 storms, in that placement starting and ending points in a longresult time of: 1)series the (theproblem recent of activitythe “fortuitous” could be

the peak pre-satellite in a long-term era and oscillation); storms may or 2) have that the earlier part of the 69-year period was in these factors are contributing to the overall been undercounted. It is likely that both of only was used, there was an increase in the occurrenceincrease. Nonetheless, of Atlantic if basin the 1970-2007 tropical storms period and category one hurricanes and the trends Fig. 3

. The annual frequency of hurricane occurrence in the Atlantic Ocean were also statistically significant (not shown) Basin for all categories for the period 1938 – 2007. All Crbn Gulf W Atl E Atl at the 95% confidence level. However, an examination of Fig. 3 suggests significant LN hurricaneinterannual occurrencesvariability as found and by tropical LJ00. The addition of the 2000-2007 6.5 (5.4) 1.4 (0.8) 1.2 (1.3) 2.9 (2.5) 1.1 (0.8) storm activity did not impact greatly EN NEU 6.2 (4.2) 0.9 (0.5) 0.9 (1.1) 3.0 (1.9) 1.3 (0.5) Total terms of where these storms occur and 4.7 (3.9) 0.5 (0.2) 0.6 (1.1) 2.9 (1.9) 0.8 (0.7) the overall results presented by LJ00 in Table 5. The average annual occurrence of Atlantic hurricanes and tropical 5.9 (4.4) 1.0 (0.6) 0.9 (1.2) 3.0 (2.0) 1.1 (0.6) activetheir interannual when considering variability tropical (Table storm 5). occurrences,Only the Gulf and region there was were slightly a few more andstorms East (from Atlantic 1938-2007) (E Atl) sub-basins. by sub-ocean basin as stratified by ENSO phase. The regions are the Caribbean (Crbn), (Gulf), West Atlantic (W Atl), western Atlantic tropical storms during La Niña years. The Table 6a. interdecadal variability with respect to the PDO All TS Cat 1 Cat 2 Cat 3 Cat 4 Cat 5 Thur is likewise unchanged by LN the addition of tropical 12.0 5.4 2.3 1.3 1.2 1.2 0.7 6.6 EN since there were not NEU 10.5 3.8 2.4 0.9 1.7 1.3 0.5 6.7 storm data (Table 6), Total 9.7 4.1 2.3 1.1 1.4 0.4 0.5 5.6 variations in the tropical Table 6b. significant year-to-year 10.9 4.5 2.3 1.1 1.5 1.0 0.5 6.5 All TS Cat 1 Cat 2 Cat 3 Cat 4 Cat 5 T years are marginally hur storm numbers. PDO2 LN attributedmore active to weaker than PDO1 and 10.8 5.2 1.8 1.8 0.8 0.8 0.4 5.6 EN years, which LJ00 NEU 10.0 4.4 2.5 1.1 0.9 0.8 0.3 5.6 events. Also, the strong Total 7.1 3.8 2.0 0.5 0.5 0.3 0.0 3.3 ENSO-relatedless frequent variations El Niño Table 6. in total tropical cyclone 9.6 4.4 2.3 1.1 0.8 0.7 0.3 5.2 The average annual occurrence of Atlantic hurricanes stratified by ENSO phase and Category (Saffir-Simpson scale) during a) PDO2 (1947 - 1976, 1999-present), and b) PDO1 6 National Weather Digest numbers during PDO1 (1938-46, 1977-98). The Interannual Variability of Hurricane Activity in the Atlantic and East Pacific Regions years is still a feature found in this study when tropical storms were added to the data set. tropicalincreased development. over the Atlantic (e.g., Goldenberg c. The active 2005 hurricane season, a case study and Shapiro 1996). Strong wind shear is not favorable to

Of all the years in active.the Atlantic This Ocean is likely Basin, the result2005 of was the convergence the most of several factors which were favorable for the occurrence of Atlantic hurricanes. Tropical SSTs were warmer than normal within the key regions of the Atlantic

Fig. 4. A sample anomaly (o hurricane(see Fig. 4). season,We also note the tropicshere that experienced during the 2005 the C) map from the month of September 2005 strongest easterly phase (source: Climate Prediction Center: Climate Diagnostics Bulletin). The anomalies are relative to of the Quasi-biennial Oscillationthe 1971 – 2000 base period. found(QBO) to in be recent correlated decades favorably (Fig. 5). to increasedThe easterly Atlantic phase hurricane QBO has activity been

(e.g. Gray, 1984a, 1984b; Elsner et wellal.1999) to Atlantic in the hurricane past. Recently, activity, the or atQBO least has not been to enhance found not or to inhibit correlate it in La Niña years (e.g., The Tropical Fig. 5. o o online at http://hurricane.atmos. colostate.edu/forecastsMeteorology Project / [available A time series of 30 hPa (red) and 50 hPa (blue) winds for the (5 N to 5 S) covering approximately4. a East 20 year Pacific period (source:Ocean Climate Activity: Prediction 1970 Center).– 2007 tropospheric wind shear ];over and the Ricks Atlantic basin during the et al.2007). Nonetheless, due to the decreased upper a. Climatology and long term trends heart of the 2005 hurricane season, the upper level winds were favorable for supporting tropical cyclones (see TMP Including tropical storms, there were 619 East Pacific site). The warmer SSTs and the QBO phase informationhttp:// tropical cyclone events included in the 38-year sample. www.cpc.ncep.noaa.gocan be found by accessingv) publishedthe monthly in Climate hard copy Prediction by the This resulted in an average of 16.3 tropical cyclones per Center (CPC) Climate Diagnostics Bulletin (see season (Table 7). This includes 276 tropical storms (7.3 per year) and 343 hurricanes (9.0 per year). As expected, National Centers for Environmental Prediction (NCEP). the number of category 1 storms was the largest (120), beFinally, expected 2007 to was weaken a year the with Atlantic La Niña-type Ocean basin SSTs subtropical (but was and there were 159 intense storms (category 3-5), not classified as a La Niña year by COAPS), and that would whichThe represented overall trend 46.3% demonstrates of the hurricane that thereactivity. was This a Niña years were more active overall in the Atlantic (Table compares to 42.7% overall in the Atlantic Ocean Basin. during that year as well (not shown). Recall, La correlate with decreased hurricane activity because of the wasdecrease opposite in the to East the Pacific Atlantic Ocean basin (Fig. where 6). However, there was this a 4). A strong subtropical Atlantic trough has been shown to trend was not statistically significant. This overall trend Volume 32 Number 2 ~ December 2008 7 Lupo et al.

combined basins this means there was no statistically statistically significant increase in activity, but for the bynegatively the interannual (correlation variability coefficient in hurricane of -0.35, activity significant for both at basinsthe 90% being confidence opposite level), one anotherand this aswill discussed likely be explainedbelow. An significant trend upward or downward. It is noted here that trend lines shown in Fig. 3 and Fig. 6 are simple linear best fit (regression) models. The two time series (from examination of the trends in each East Pacific hurricane category would demonstrate that only the category 3 (5) 1970 for the All Atlantic region)TS correlatedCat 1 Cat to each 2 otherCat 3 Cat 4 Cat 5 Thur downward (upward) trend, LN whilestorms the showed rest of the a significantcategories

15.1 7.3 3.0 1.7 1.6 1.4 0.1 7.8 A breakdown by month EN showed no significant trend. NEU 16.4 7.1 3.3 1.7 1.6 2.6 0.2 9.3 Total 16.9 7.3 3.0 1.7 2.1 2.2 0.7 9.7 (Table 8) indicates that Table 7. 16.3 7.3 3.2 1.7 1.7 2.2 0.2 9.0 activityMay is the begins first in month the East that significant tropical cyclone The mean number of East Pacific tropical storms and hurricanes separated by ENSO phase and intensity (Saffir – Simpson scale) in the East Pacific from 1970 – 2007. Pacific. August was the most active month with 152 total cyclonesstorms occurring per year), over and the 38 by November,year climatology this (4 region tropical is relatively inactive. Thus,

roughly one calendar month aheadthe East of the Pacific Atlantic season season, was a result similar to Davis et al.

(1984) and Schultz (2008). By geographic region (Table 9), the southeast part of the Fig. 6. hurricaneEast Pacific (tropical was the storm) most occurrencesactive, and 76% happen (82%) within of the As in Fig.May 2, except forJune the EastJuly Pacific OceanAug Basin andSept the yearsOct 1970 -Nov 2007. Dec LN cyclones, including two hurricanesthis region. Only occurred five tropical in the 0.5 2.0 4.2 4.0 2.6 2.6 0.3 0.0 northwest region. EN NEU 0.7 2.3 3.5 3.7 3.9 2.2 0.3 0.1 Total b. ENSO and PDO variability 0.6 1.6 3.5 4.8 3.8 2.1 0.6 0.1 During El Niño years, there Table 8. 0.6 2.0 3.6 4.0 3.6 2.0 0.4 0.1 were more tropical cyclones outside of The these mean months number are ofnot East included Pacific on tropical the table). cyclones (tropical storms and hurricanes) by month from May through December from 1970-2007 (two tropical cyclones occurring Phase NE SE SW NW Total storms reaching hurricane in strength, the East especially Pacific intense (Table 7), and this included more LN hurricanes, in this region than in La Niña years. This might

1.0 12.4 1.4 0.3 15.1 EN ofbe expected El Niño as year the tropicalwaters in cyclones the East isPacific not are statistically warmer NEU 1.1 12.7 2.6 0.1 16.4 during El Niño years. While the greater overall frequency Total 0.5 13.3 3.2 0.0 16.9 significant, it is opposite of the ENSO variability in the 1.0 12.8 2.4 0.1 16.3 Atlantic Ocean Basin. When separating by PDO phase Table 9. Theº stratificationº of East Pacific Basin tropical cyclone (Table 10), it can be shown that there was little overall activity (yearly means from 1970-2007) by geographic region ENSO variability in PDO2 years, while the ENSO variability 8(along National 20 N andWeather 125 WDigest or as in Fig. 1b.) and ENSO phase. The Interannual Variability of Hurricane Activity in the Atlantic and East Pacific Regions

were more intense storms during La Niña years, especially and four more tropical cyclones overall during El Niño was accentuated in PDO1 years (two more hurricanes in the PDO1 phase. years) as it was in the Atlantic region (LJ00), except here during Comparing El Niño years, the length tropical of thecyclone East activity Pacific is season greatest by byEl Niño stronger years warm were more ENSO active SST (a result significant at the examining the monthly occurrence (Table 8), shows that Table 10a. anomalies90% confidence located level). closer These to were years characterized the Americas (see Lupo et All TS Cat 1 Cat 2 Cat 3 Cat 4 Cat 5 Thur LN southeastal. 2007), portion or over of our the basin East 15.4 7.7 3.1 1.9 1.6 1.1 0.1 7.7 ofPacific study. region, This especially provided the EN NEU 15.4 8.6 3.0 2.0 0.6 1.2 0.0 6.8 rationale for the breakdown Total 15.8 6.8 2.0 2.0 2.3 2.3 0.4 9.0 sub-basins. Table 10b. 15.6 7.8 2.8 2.1 1.4 1.4 0.2 7.8 of the East Pacific Region into All TS Cat 1 Cat 2 Cat 3 Cat 4 Cat 5 T if there were variations hur in theIn order intensity to of determine storms, LN compared their distributions, 14.0 6.0 2.5 1.0 2.0 2.5 0.0 8.0 EN testingLJ00 used these histograms for statistical and NEU 16.7 6.6 3.4 1.5 1.9 3.1 0.1 10.1 Total 18.4 8.2 3.8 1.6 2.0 2.2 0.6 10.2 Table 10. significance. For the overall 16.8 6.9 3.4 1.5 2.0 2.8 0.2 9.9 East Pacific sample, there The average annual occurrence of East Pacific hurricanes stratified by ENSO phase were no significant differences and category (Saffir-Simpson scale) during a) PDO2 (1970 - 1976, 1999-present), and b) in the distributions (Fig. 7), PDO1 (1977-98). even though the differences in the hurricane frequencies the(Table occurrence 7) demonstrate of intense that storms. there During were differences in the overall averages for

PDO2 (there was little ENSO-related interannual variability), and PDO1 (EN years significantly more active), the theresults distributions were different (which even can if be there inferred were no statistically significant differences in intense storms during El Niño years from Table 10). There were clearly more

(5.0 versus 2.8 for La Niña and 1.8 for hurricaneNeutral years) occurrences for PDO2 in years.El Niño However, and La the real substantive differences between of a greater number of weaker hurricanes Niña years, during PDO1, were the result and tropical storms. The PDO1 result butis consistent found that with during the more results active of Schultz years, there(2007), were who aused greater a different number methodology of weaker

Fig. 7 storms. This is, however, different from category for a) the total sample, b) El Niño years, c) neutral years, and d) La Niña LJ00, who found that in the Atlantic there years.. Histograms of the average annual hurricane frequency separated by Volume 32 Number 2 ~ December 2008 9 Lupo et al.

seasonalfrom May variations through in November, tropical storms while inversus La Niñahurricanes years, the final conclusion of LJ00 was that the next few years werethe season similar is overall. shorter However, (primarily the Juneseasonal – October). peak in each The after 1999 would be more active hurricane seasons in a manner similar to that of the 1940s through 1960s. Table years, the peak in tropical cyclone activity was clearly 3 verifies their assertion of a more active 2000-2007 August,ENSO phase while occurred in neutral in yearsdifferent the peakmonths. was Duringlater (August El Niño wasperiod. an El During Niño theseyear. However, eight years, there only were during more the tropical 2002 stormsand 2007 than seasons a typical were year there during few these hurricanes seasons. andThus, 2002 the cycles/ September), between and the earlier phases (July of the / August)PDO here in (notLa Niña shown). years. in hurricane activity. ThereA breakdown was little difference of tropical cyclonein the activityENSO-related by geographic seasonal work of LJ00 successfully projected long-term increases region demonstrated that the southeast region was hurricane seasons, which implies that SSTs may have the most active, and varied similarly to that of the total been In a thestronger East Pacific,factor than El Niño the atmospheric years were morebackground active sample discussed above. The northeast region, however, state in the interannual variability of the occurrence and was relatively more active in La Niña and neutral years, intensity of hurricanes in this area of the world. This while the southwest region was more active in El Niño

correlationmay be especially to Atlantic true Ocean during basin PDO1 activity years. and, This thus, also to and neutral years (Table 9). During the El Niño years, implies that East Pacific hurricane activity has an inverse this reflects the fact that warmer SSTs were located over intensethe East hurricanes Pacific formation were in regions the southwest (see Lupo and et southeastal. 2007), the QBO. However, the atmosphere may also be more especially in the southern sectors. While nearly all of the atmosphericconducive to East background Pacific hurricane state is an activity important during factor El Niño in seasonalyears. Zuki tropical and Lupo cyclone (2008) occurrence demonstrate and that warma favorable SSTs regions, neither regiono had a significantly higher by themselves will not guarantee an active season. separationproportion ofof thesecategory data 4 byand phase 5 storms. of the Only PDO one (not hurricane shown) wouldwhich formedreveal that north the of ENSO 20 N variability achieved category described 3 status.above isA b. Hindcasting the 2000-2005 seasons

5.independent Discussion of the phase of the PDO in the East Pacific. When forecasting (and hindcasting) hurricane activity a. Climatology in each basin for the upcoming season (Table 11) from to2000-2007 produce in reliable a manner forecasts. similar Since to that a forecast demonstrated of the As a result of this work, it was shown that by geographicalby LJ00 for 1999, distribution it is shown of stormthat this genesis work regionscan be used was given using their method, we evaluated the forecast using overall in the climatology of the Atlantic Ocean basin for the statistical extending the results of LJ00, there was little change 2 it was shown that several factors known to correlate with betterthe same forecast, calculation or shows used toan determine annual distribution χ closer to hurricane activity. By examining the active 2005 season, thattest (Neter of the et climatological al. 1988). Thus, contingency a lower score forecast represents used by a were favorable during that hurricane season. Thus, this veryincreased active hurricane season canactivity be explainedin the Atlantic using Ocean traditional Basin analyses and the convergence of favorable atmospheric LJ00. The years 2000 – 2005 presumed that the forecast of the ENSO phase was correct as LJ00 did for 1999. In necessarily invoking climate change as the cause of the largethe Atlantic, errors the in hurricanepoorest hindcast occurrence was werethat for recorded the 2005 in activeand oceanic season. processes However, (e.g., it is concededZuki and Lupo that climate2008) without change eachseason, category. in which However, every basin an active was yearquite did active not guaranteeand, thus, as a factor cannot be ruled out as changes in climate could conditions which give rise to tropical cyclone formation. a high score as the 2003 and 2004 seasons were similarly be reflected in the background atmosphere and oceanic andactive, this but is likely the hindcast due to the for fact the that 2004 the season southeast was region much dominatesbetter. In the the East activity Pacific, in the this hindcasts basin. Thus, were thebetter threshold overall Also, it is likely that, while this season was quite active, for what constitutes a good forecast should probably be the disparity in the numbers between 2005 and the next lower in this region, and several more years would be satellitemost active observations year, 1933, early was in probably the record. not Additionally, as great as needed in order to establish a recommendation. the observed numbers (27 versus 21) due to the lack of 10 National Weather Digest The Interannual Variability of Hurricane Activity in the Atlantic and East Pacific Regions

Atlantic Region during the height of the Year Carribean Gulf West Atl East Atl Total Score season was predicted to be a La 2006 hurricane season. The 2007

2000 1 (1) 3 (1) 8 (3) 3 (1) 6 (3.30) goodNina year. compared In the Atlantic, to the the previous 2006 2001 4 (2) 2 (0) 8 (3) 1 (1) 6 (4.50) years,and 2007 and forecasts this is likely were due still to fairly the 2002 2 (0) 3 (1) 6 (1) 1 (1) 3 (2.20) distribution of tropical cyclones 2003 1 (1) 6 (4) 4 (1) 5 (3) 9 (13.20) wasfact that fairly during similar PDO2 across years, each the 2004 1 (1) 2 (0) 6 (1) 5 (3) 5 (5.20) phase of ENSO. The distribution of the tropical cyclone intensities 2005 7 (5) 6 (4) 11 (6) 4 (2) 17 (31.70) 2006 1 (1) 1 (1) 7 (2) 1 (1) 5 (2.30) 2007 2 (0) 4 East(2) Pacific7 Region(2) 2 (0) 4 (2.80) TS,(2006: three five weak, TS, three and two weak, strong two Year Northeast Southeast Southwest Northwest Total Score hurricanes)strong hurricanes; was also 2007: similar ten to

and three strong hurricanes). the expected (five TS, four weak, 2000 0 (1) 15 (3) 3 (1) 1 (1) 6 (3.25) 2001 1 (0) 9 (3) 5 (3) 0 (0) 6 (5.25) three ofIn the the previous East Pacific, hindcasts, the 2002 0 (1) 11 (2) 4 (1) 0 (0) 4 (1.64) but2006 worse forecast than was the otherbetter three. than 2003 3 (2) 12 (0) 1 (1) 0 (0) 3 (4.50) was not very good even though 2004 0 (1) 10 (2) 2 (0) 0 (0) 3 (1.33) theConversely, overall number the 2007 of storms forecast was 2005 1 (0) 13 (1) 1 (1) 0 (0) 2 (0.58) close to what would be expected. 2006 1 (0) 16 (3) 2 (1) 0 (1) 5 (2.69) There were more (fewer) tropical cyclones in the southwest 2007 1 (0) 10 (3) 3 (2) 0 (1) 6 (5.69) (southeast) part of the East Table 11. The annual hindcasts for tropical cyclone frequencies based on the work of LJ00 (for the 2000-2005 seasons), and forecasts (for 2006 – 2007) based on the results shown thehere. total A simple score pointis a scoring score for system that season (|predicted (a lower – observed|) score is a isbetter used forecast,here. The or table is a better forecastsPacific Basin for than this regionexpected. (seven The displays the observed2 value (|difference from observed|). The value in parenthesis beside 2006 tropical cyclone intensity χ somewhat good, except that there “fit”). TS, five weak, three strong) was weak hurricanes, and strong hurricanes was similar to For each of these years, the ratio of tropical storms, were more strong storms observed (eight TS, five weak, 6 strong) than expected. In 2007, the intensity forecasts what should be expected (Tables 6 and 10). This would aswere well better, forecast even as if most the overall of the hindcastsnumbers wereshown lower in Table (10 result in a forecast of a decreasing frequency in each TS, three weak, one strong) Thus, 2006 and 2007 were not category of stronger winds, except for PDO2 El Niño years in the East Pacific region where more strong 11. The 2007 forecasts for both regions were similar to storms than weak storms are expected. In the Atlantic, that of 2006, since weak El Niño conditions collapsed in only 2004 did not conform reasonably to the expected February, 2007 and neutral to La Niña conditions could intensitiesintensity distribution were reasonably (we observedhindcast (not 3 weak shown). hurricanes 6.reasonably Summary be expected and Conclusions to develop (Fig. 8). and 6 strong). In the Pacific and for each of these years, The climatological behavior of tropical cyclone activ- c. Forecasting the 2006 – 2007 seasons -

ity in the Atlantic and East Pacific Ocean Basins was ex For the 2006 and 2007 seasons, a true forecast was amined using the methodologies of LJ00. In the Atlantic, made in January and February assuming that the 2006 an update to include the 2000 – 2007 seasons, as well as Niñoand 2007 SST patternseasons setwould in during be a La the Niña-type late -to-early season based included.a re-categorization Tropical storm of hurricanes activity was(e.g., also Andrew added – 1992)to the on the SSTs (see Lupo et al. 2007, their Table 1). A weak El or ENSO years (e.g., 1974 becomes a La Niña year) was Volume 32 Number 2 ~ December 2008 11 Lupo et al.

while more storms formed in the southeast and southwest part of the East

likely due to the fact that ENSOPacific years Ocean bring Basin. warmer This is

downwaters tothe the ENSO East Pacific years byregion. phase When of the PDO, breaking the

in occurrence and intensity andENSO-related geographic differences formation Fig. 8. region are accentuated As in Fig. 4, except for April 2007. - down of the hurricane and tropical storm activity from variability is similar to that in occurring PDO1 years, in the but Atlantic were Atlantic database. In the East Pacific, a thorough break blurred in PDO2 years. This ENSO and PDO related in La Niña years and they were more intense. 1970 – 2007 was examined in a manner similar to LJ00. (LJ00), except that in the Atlantic more storms occurred The major findings for East Pacific tropical cyclone aactivity greater demonstrated amount of activity that there than were found 16.3 in the storms Atlantic per of theAdditionally, recent increase it is hypothesized in hurricane here activity that the which Atlantic may Oceanyear (9.0 basin. hurricanes The long and term 7.3 trendtropical showed storms) only which a slight was behurricane associated season with of 2005 the PDO, was so but active, also not possibly only because due to decreased upper tropospheric shear over the Atlantic tropical cyclone activity, and this was opposite to that which may have been associated with a stronger easterly decrease (not statistically significant) in East Pacific increase. The combined basin trend was not statistically of the Atlantic which showed a statistically significant forecastingphase of the utility QBO along by using with the warmer same thanhindcasting normal method SSTs. likely. There were a slightly higher percentage of intense Finally, it is shown that this work may have some significant, so that either trend for these basins is equally even though there were a few more intense hurricanes in calculation.of LJ00, except This here approach we assign included each season assuming from that 2000 the – hurricanes in the East Pacific than in the Atlantic region, 2005 a forecast score based on a statistical distribution these years were reasonable, more seasons will be needed wasthe Atlantic the most region active, during and 2000 the seasonal– 2007 than activity that found (season in inENSO order forecast to evaluate was correct. what would While constitute the hindcasts a good for score most inof LJ00. The southeast part of the East Pacific Ocean Basin

beginning,An examination peak, and end) of was the about interannual 15 to 30 days variability earlier each basin. Two true forecasts were made for the 2006 than that in the Atlantic Ocean Basin. forecastsand 2007 wereseasons; relatively and in poor, the Atlantic, which may the haveforecasts been were due cyclones during El Niño years, and that this was mainly relatively good. In the Pacific, the tropical cyclone season accounteddemonstrated for that by there more were storms more East becoming Pacific tropical intense preliminary and data from additional seasons will be used hurricanes than during La Niña years. The tropical cyclone to verifya poor our ENSO approach. forecast in 2006. This work, however, is season was one or two months longer in El Niño years,

Acknowledgments

The authors would like to thank the reviewers, Mr. Justin Arnott (Binghamton, NY Forecast Office) and Mr. Raymond Zehr (Colorado State University/NOAA/NESDIS, Cooperative Institute for Research in the Atmosphere - CIRA) for their helpfulu comments in making this work stronger. We would like to also thank Dr. Nathan Mantua at the Joint Institute for the study of Atmosphere and Ocean (JISAO) [available online at: http://jisao.washington.ed ] for providing us with Fig. 2. 12 National Weather Digest The Interannual Variability of Hurricane Activity in the Atlantic and East Pacific Regions

Authors

Anthony R. Lupo Christopher J. Melick received his Ph.D. from the Associate Professor of the Atmospheric Sciences in the Soil, Environmental, andis Atmospheric the Department Sciences Chairman Department and anat at the University of Missouri- been in the areas of large-scale atmospheric dynamics, Columbia in May 2008. Under the supervision of Dr. climatethe University dynamics, of and Missouri climate - Columbia.change. His His current research teaching has dissertationPatrick Market, focused Chris has on been the involved synoptic with and the mesoscale Research stabilityon Convective aspects of Group for eventsthe last across 4 years the and central his assignments include the dynamic sequence, the climatology sequence, and general circulation. He . He received both a Bachelor of Science has been a Fulbright Scholar, studying climate change include(2000) and diagnostics Master of examining Sciences (2003) the development in Atmospheric of at the Russian Academy of Sciences in Moscow. He is extratropicalScience from cyclones, Purdue University.heavy convective His research snowfall interests events, also a contributing scientist and reviewer for the World andMeteorological PhD degrees Organization in Atmospheric (WMO) Science Intergovernmental from Purdue Panel on Climate Change (IPCC). He received his MS as well as . Dr. Melick is a member of the NWA, AMS, Sigma Xi, Chi Epsilon Pi, Golden Key Honor University in West Lafayette, IN. In service to the National Society, Phi Beta Kappa, and Missouri Academy of Science. Weather Association, Dr. Lupo is an Associate Editor of In addition, he has also been three times nominated and the National Weather Digest and serves on the Weather Patrickawarded S. three Market times is an to Associate The Chancellor’s Professor List. of Atmospheric Analysis and Forecasting Committee, the Education TameraCommittee, K. Latham and the Publications Committee. Science at the University of Missouri-Columbia. He of Science Degree inis Soil a senior and Atmosphericat the University Science. of Missouri She is earned his Bachelor of Science degree in meteorology - Columbia. She will graduate in May 2009 with a Bachelor from Millersville University of Pennsylvania in 1994, and currenthis Master teaching of Sciences assignments and Ph.D. degrees include from the Saint synoptic Louis currently a member of the Storm Chase Team and takes University in 1996 and 1999, respectively. Dr. Market’s part in MU campus forecasting. She is planning to pursue the daily forecasting practicum. His research explores Trentona career inMagill Broadcast Meteorology. issuesmeteorology in heavy sequence, and mesoscale forecasting, meteorology, including and the occurrence of thundersnow, and the graduated from the University of Missouri in May 2008 with a Bachelor of Science Degree in Soil and Atmospheric Science and a Minor in Mathematics. He was efficiency of mesoscale convective systems. In service to a member of the Storm Chase Team and took part in MU the National Weather Association, Dr. Market has served campus forecasting. Trent also held an internship at KMIZ as a Council member, Co-editor of the National Weather in Broadcast Meteorology. He plans to pursue a career in Digest, a tape evaluator for the Broadcast Meteorology JosephBroadcast V. ClarkMeteorology. is a graduate of the Department of Soil, Committee, and a Co-Chair of the Weather Analysis and Forecasting Committee.

Environmental, and Atmospheric Sciences at the University Asof Missouri-Columbia an undergraduate, and he investigatedis currently a theMaster climatology of Sciences of blockingstudent at the University in theof Illinois northern at Urbana-Champaign. hemisphere. He is currently investigating the relationship between various processes. wind regimes in the Arctic and their effect on

Volume 32 Number 2 ~ December 2008 13 Lupo et al.

References

Davis, M.A.S., G.M. Brown, and P. Leftwich, 1984: A tropical Henderson-Sellers, A., H. Zhang, G. Berz, K. Emanuel, W. cyclone data tape for the Eastern and Central North Gray, C. Landsea, G. Holland, J. Lighthill, S.L. Shieh,Bull. P. Pacific Basins, 1949-1983: Contents, limitations, and Amer.Webster, Meteor. and K.Soc. McGuffie,, 1998: Tropical cyclones and uses. NOAA Technicalth Memorandum. NWS NHC 25, global climate change: A post – IPCC assessment. 20 pp. [Available from the National Hurricane Center, , 79 19 – 38. 11691 SW 17 St., Miami, Florida 33165-2149]. Jarvinen, B.R., C.J. Neumann, and M.A.S. Davis, 1984: A Deser,Sensitivity C., and to A.S. tropical Phillips, forcing. 2006: J. Climate Simulation, of the tropical cyclone data tape for the North Atlantic Basin, 1976/1977 climate transition over the North Pacific: 1886 - 1983: Contents, limitations, and th uses. NOAA , 19 6170 – Tech. Memo. NWS NHC 22, 21 pp. [Available from the ______,6180. National Hurricane Center, 11691 SW 17 St., Miami, climate variability: Linkages between the tropics and Florida 33165-2149]. , and J.W. Hurrell, 2004: Pacific interdecadalJ. framework for understanding Atlantic hurricane Climate , Kossin,variability J.P., and and D.J. trends. Vimont, Bull. 2007: Amer. AMeteor. more Soc., general, North Pacific during the boreal since 1900. , 17 3109 – 3124. 88 1767-1781. ______, A. Capotondi,J. Climate A.R. , Saravanan, and A.S. Phillips, Atlantic hurricanes. Mon. Wea. Rev. 2006: Tropical Pacific and Atlantic climate variability Landsea, C.W., 1993: A climatology of intense (or major) in CCSM3. , 19 2451 – 2481. , 121, 1703 - 1713. J. Climate Elsner, J.B., A.B. Kara, and M.A. Owens, 1999: Fluctuations in North Atlantic hurricane frequency. , 12, ______,Climatic R.A. PielkeChange, Jr., A., Mestas-Nuez,. and J. Knaff, 1999: 427 – 437. Atlantic basin hurricanes: Indicies of climate changes. 42 89 - 129 Emanuel, K., 2005: Increasing destructiveness of tropical cyclones over the past 30 years. , 436, 686. Lupo, A.R., and G.J. Johnston, 2000: The variability in modulation of ENSO teleconnections. Bull. Amer. Oscillation.Atlantic Ocean Nat. Wea. Basin Dig. hurricane occurrence and Gershanov,Meteor. Soc A., and T.P. Barnett, 1998: Inderdecadal intensity as related to ENSO and the North Pacific , 24:1,2, 3 – 13. ., 79, 2715 - 2725. features of blocking anticyclones: A study of Northern ______, R.J. Oglesby, and I.I. Mokhov, 1997: Climatological Goldenberg, S.B., and L.J. Shapiro, 1996: Physical . Climate Dyn. mechanisms for the association of El Nino and West Hemisphere CCM1 model blocking events in present- J. Climate , 2 African rainfall with Atlantic major hurricane activity. day and double CO , 13, , 9 1169 - 1187. 181-195.

Gray, W.M., 1984a:Mon. AtlanticWea. Rev. season hurricane frequency. ______, A.R., Kelsey, E.P., D.K. Weitlich, I.I. Mokhov, F.A. Part 1: El Nino and 30 mb quasi biennial oscillation Akyuz, Guinan, P.E., J.E. Woolard, 2007: Interannual influences. , 112, 1649 - 1668. localand interdecadal climate. Atmosphera, variability in the predominant. Pacific Mon. Wea. Rev. Region SST anomaly patterns and their impact on a _____, 1984b: Atlantic season hurricane frequency. Part 2: 20, 171-196 Forecasting its variability. , 112, 1669 - 1683. climateMantua, oscillation N.J., S.R. withHare, impacts Y. Zhang, on salmon J.M. Wallace, production. and trends associated with multidecadal variability of Bull.R.C. Francis, Amer. Meteor. 1997: Soc.,A Pacific interdecadal _____,Atlantic J.D. Sheaffer, hurricane activity.and C.W. Hurricanes, Landsea, Climate, 1997: and Climate Socioeconomic Impacts 78, 1069–1079.

, B. Springer, H.F. Diaz and R.S. 14 Pulwarty,National WeatherEds. Springer-Verlag, Digest New York, 15-53. The Interannual Variability of Hurricane Activity in the Atlantic and East Pacific Regions

Geophys. Res. Minobe,Lett. S., 1997: A 50 - 70 year climatic oscillation over the North Pacific and North America. , 24, 683 - 686. Applied Statistics. 3d ed Neter, J., W. Wasserman, and G.A. Whitmore, 1988: . Allyn and Bacon, 1006 pp. Bull. Pielke,Amer. R.A., Meteor and C.N.. Soc. Landsea, 1999: La Nina, El Nino, and Atlantic hurricane damages in the United States. , 80, 2027 - 2033.

Ricks, R.J., R. Wagner, D.F. Navarro, J.M. Maldonado, M.J. Abstracts,Santiago, 32 2007:nd Annual A Gulf Meeting of Mexico of the National tropical Weather cyclone Association,climate study using ENSO and QBO as determinants.

Reno, NV, 16 October 2007, 61. Nat. Wea. Dig. Schultz, L.W., 2008: Some climatological aspects of tropical cyclones in the eastern north Pacific. ,32, 45-54. scale. Weatherwise Simpson, R.H., 1974: The hurricane disaster potential , 27, 169 and 186. mode and hurricane activity. Geophys. Res. Lett., 34, Vimont,L07709, D.J., doi:10.1029/2007GL029683 and J.P. Kossin, 2007: The Atlantic. meridional

Webster,intensity P.J., G.J. in Holland, a warming J.A. Curry, environment. and H.R. Science,Chang, 2005:, Changes in tropical cyclone number, duration, and 309 1844 - 1846. variability of tropical cyclone activity in the southern Zuki, Md. Z., and A.R.J. Lupo, Geophys. 2008: Res., The 113 interannual

South China Sea. , D06106, doi:10.1029/2007JD009218.

Volume 32 Number 2 ~ December 2008 15