Dynamics of Atmospheres and Oceans 63 (2013) 131– 141

Contents lists available at SciVerse ScienceDirect

Dynamics of Atmospheresand Oceans

journal homepage: www.elsevier.com/locate/dynatmoce

Relationship between the frequency of tropicalcyclones in Taiwan and the Pacific/NorthAmerican pattern

Ki-Seon Choia, Il-Ju Moonb,∗

a National Typhoon Center, Korea Meteorological Administration, Republic of Koreab College of Ocean Science/Ocean and Environment Research Institute, Jeju National University, Republic ofKorea

a r t i c l e i n f o

Article history:Received 10 July 2012Received in revised form 26 May 2013Accepted 29 May 2013

Available online xxx

Keywords:Tropical cyclonePacific/North American teleconnectionpatternEast AsiaTaiwan

a b s t r a c t

The frequency of tropical cyclones (TCs) in Taiwan during June toOctober (JJASO) is found to have a strong negative correlation withthe Pacific/North American (PNA) pattern in the preceding April. Inthe negative PNA phase, the anomalous cyclonic and the anomalousanticyclonic circulations are intensified at low latitudes and mid-latitudes from East Asia to the North Atlantic, respectively, fromApril to JJASO. Particularly in East Asia, the anomalous southeast-erly that converges between the anomalous anticyclone to the eastof Japan and the anomalous cyclone to the east of Taiwan plays adecisive role in moving TCs not only to Taiwan, but also to the mid-latitude coastal regions of East Asia as a result of the steering flow.In addition, a southwestward extension of a western North Pacific(WNP) high during the positive PNA phase also contributed to a fre-quent movement of TCs to southern China without traveling northtoward the midlatitude regions of East Asia. Due to the difference inthe typical tracks of the TC in the WNP according to the PNA phase,the intensity of the TC in the negative PNA phase is stronger thanthat in the positive PNA phase.

© 2013 Elsevier B.V. All rights reserved.

∗ Corresponding author at: College of Ocean Science/Ocean and Environment Research Institute, Jeju National University, Ara1 Dong. Jejusi 690-756, Republic of Korea. Tel.: +82 64 754 3412; fax: +82 64 756 3483.

E-mail address: ijmoon@jejunu.ac.kr (I.-J. Moon).

0377-0265/$ – see front matter © 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.dynatmoce.2013.05.003

132 K.-S. Choi, I.-J. Moon / Dynamics of Atmospheres and Oceans 63 (2013) 131– 141

1. Introduction

The Pacific/North American (PNA) teleconnection pattern is one of the most prominent modes oflow-frequency variability in the Northern hemisphere (NH) extratropics (Wallace and Gutzler, 1981;Wallace, 2000). It tends to be most pronounced in the winter months. The PNA is associated with aRossby wave pattern, with centers of action over the Pacific and over North America. It refers to therelative amplitudes of the ridge over western North America and the troughs over the central northPacific and southern U.S. (Leathers et al., 1991a). That is, the positive PNA phase is characterized by ananomalous anticyclone in the subtropical western North Pacific (SWNP) and over the intermountainregion of North America and by an anomalous cyclone in the North Pacific and over the southeasternU.S., and vice versa in the negative PNA phase.

The positive PNA phase is associated with above-average temperatures over western Canada andthe extreme western United States, and below-average temperatures across the south-central andsoutheastern U.S. The PNA tends to have little impact on surface temperature variability over NorthAmerica during summer. The associated precipitation anomalies include above-average totals in theGulf of Alaska extending into the Pacific Northwestern United States, and below-average totals overthe upper Midwestern United States (Leathers et al., 1991a,b; Leathers and Palecki, 1992; Colemanand Rogers, 1995; Isard, 1999; Renwick and Wallace, 1996).

Although the PNA pattern is a natural internal mode of climate variability, it is also strongly influ-enced by the El Nino/Southern Oscillation (ENSO) phenomenon. The positive phase of the PNA patterntends to be associated with Pacific warm episodes (El Nino), and the negative phase tends to be asso-ciated with Pacific cold episodes (La Nina) (Trenberth et al., 1998; Straus and Shukla, 2002; Lin et al.,2005). The PNA phases are also associated with warm phases of Pacific Decadal Oscillation (PDO) andthe reorganization of the PNA pattern toward a positive mode is strongest when the ENSO and PDOare in phase (Trouet and Taylor, 2009).

While many studies related to the PNA pattern have paid attention to its influences on regional tem-perature and precipitation in North America, fewer studies have focused on possible remote impactsof the PNA pattern on Indian monsoon and Asian climate. Peings et al. (2009) suggested a winter-to-spring PNA index as a reasonable basis for multiple linear regression scheme for the prediction of theIndian summer monsoon rainfall. Gong et al. (2007) found that the frequency of dust storms in northernChina was positively associated with the PNA pattern on an interannual time scale during 1962–2002.Wang et al. (2000) explained how ENSO linked to the PNA affects the East Asian climate through thePacific-East Asian teleconnection. Until now, however, there have been few studies on the relationshipof the PNA pattern with the activity of tropical cyclones (TCs) in the western North Pacific (WNP).

Klotzbach and Gray (2004) used the PNA pattern as one of the potential predictors for 6–11-monthpredictions of seasonal hurricane activity in the Atlantic basin. Our study is based on the idea that apositive PNA phase in the preceding winter is usually related to a cold ENSO phase in the current sum-mertime and that, eventually, a cold ENSO phase can enhance the activity of TCs in the Atlantic. SinceENSO linked to the PNA is a factor influencing TC activity over the WNP (Chan, 2000; Wang and Chan,2002) and the TC variability between the North Atlantic and the North Pacific is related (Wang, 2010),it is interesting to examine whether the PNA pattern is related to the activity of TCs in the WNP. In par-ticular, the present study attempts to identify a possible remote relationship between the PNA patternin April and the frequency of TCs that affect Taiwan from June to October (JJASO) in the same year.

Section 2 describes the data and the methodology used. Section 3 investigates the relationshipbetween the PNA index and the frequency of TCs that affect Taiwan, as well as the associated large-scaleenvironments. The conclusion and summary appear in the final section.

2. Data and methodology

2.1. Data

Data on the frequency of TCs in the WNP from 1979 to 2011 (33 years) were obtained from thebest track archives of the Regional Specialized Meteorological Center, Tokyo Typhoon Center. Thedata consist of the names of the TCs, their central positions (latitude and longitude), their minimum

K.-S. Choi, I.-J. Moon / Dynamics of Atmospheres and Oceans 63 (2013) 131– 141 133

Fig. 1. The monthly frequency of TCs that influence Taiwan and the percentage frequency of TCs in each month compared withthe total frequency of TCs during the period 1979–2011 (33 years).

surface central pressures, and their maximum sustained wind speeds (MSWS) (MSWS: 10-min aver-aged maximum winds to the nearest 5 kts) measured every 6 h. In this study, the term TC includestropical depressions (MSWS <34 kts), tropical storms (34 kts ≤ MSWS ≤ 47 kts), severe tropical storms(48 kts ≤ MSWS ≤ 63 kts), typhoons (MSWS ≥ 64 kts), and extratropical cyclones that have transitionedfrom TCs.

Atmospheric circulation data from 1979 to the present were obtained from the National Cen-ters for Environmental Prediction/Department of Energy (NCEP/DOE) Global Reanalysis (R-2) datasetwith horizontal resolutions of 2.5◦ × 2.5◦ and 17 vertical pressure levels (Kanamitsu et al., 2002). Thedata include geopotential height (gpm) and horizontal and meridional wind (m s−1). We also usedextended monthly sea–surface temperature (SST) data, which have been reconstructed using theInternational Comprehensive Ocean–Atmosphere Data Set (ICOADS), to analyze large-scale oceanicenvironments (Reynolds et al., 2002). The National Oceanic and Atmospheric Administration’s (NOAA)interpolated outgoing long-wave radiation (OLR) data were also used in our analysis. The data,retrieved from the NOAA’s satellite series, are provided by the NOAA’s Climate Diagnosis Center (CDC,http://www.cdc.noaa.gov) and are available for June 1974 until the present, except a missing periodfrom March to December of 1978 (Liebmann and Smith, 1996).

The analyses in this study are limited to the period 1979–2011; this is because of two reasons:(1) a significantly higher density of satellite observation, which is a critical factor for TC tracking inthe open ocean and the production of reanalysis data, was accomplished from the late 1970s and (2)NCEP/DOE R-2 dataset, an improved version of the NCEP-NCAR reanalysis dataset R-1 that fixed errorsand updated parameterizations of physical processes, is available from 1979.

2.2. Methodology

In this study, TCs that influence Taiwan (hereafter, TW-TC) are defined as those that pass throughthe area of 21–26◦ N, 119–125◦ E (embedded map in Fig. 1) (Chu et al., 2007). A total of 159 TCs struckTaiwan over the 33 years, about 90% of which occurred in June–October (hereafter, JJASO) (Fig. 1).Thus, this study focuses on the TCs that affected Taiwan in JJASO.

The lifetime of the TC is defined as the period from its occurrence to its disappearance. To calculatethe TC’s passage frequency (TCPF), each position of a TC is binned into a 5◦ × 5◦ grid box, and a TC isonly counted once, although it may enter the same grid box several times. The TC’s genesis frequency(TCGF) is calculated by the same method as for the TCPF. This study also used the Student’s t test todetermine significance (Wilks, 1995).

3. Relationship between the frequency of TW-TCs and PNA

3.1. Definition of a new PNA index

Wallace and Gutzler (1981) defined a PNA index based on the distribution of the normalized coldseason’s (December to May) 500-hPa geopotential height anomalies. Another PNA index used by the

134 K.-S. Choi, I.-J. Moon / Dynamics of Atmospheres and Oceans 63 (2013) 131– 141

Fig. 2. (a) Regression (eigenvector) maps and (b) PC time series (solid line with a closed circle) of the second leading REOFs inApril, showing the geopotential height at 500 hPa in the region north of 20◦ N; the data are compared with the Pacific/NorthAmerican (PNA) index derived from the CPC (dotted line with a open circle) for the period 1979–2011. The eigenvector ismultiplied by 100.

NOAA’s Climate Prediction Center (CPC) is defined as the time coefficient of the second leading modeof the rotated empirical orthogonal function (REOF) analysis of the 500-hPa geopotential heights in theregion between 20–90◦ N (http://www.cpc.noaa.gov/products/precip/CWlink/pna/pna.shtml). In thepresent study, the latter definition was used to obtain the PNA index using April’s 500-hPa geopotentialheight anomalies (Fig. 2), but the calculation was made only for the 33 years. Our analysis reveals thatthe time series of the second mode were well correlated with the PNA indexes provided by the CPC(r = 0.76), which were significant at the 99% confidence level (Fig. 2b).

Using the principal component (PC) of the second leading mode, we found that there was a strongnegative correlation between the PNA index and the frequency of TW-TCs in JJASO (r = −0.65, signif-icant at the 99% confidence level) (Fig. 3). A highly negative relation between them was maintained(r = −0.64, significant at the 99% confidence level) when the trends in two times series are removed.

To examine more clearly the relation between the two variables, we conducted one-point corre-lation analysis of the frequency of TW-TCs in JJASO and the 500-hPa geopotential heights in April(Fig. 4a). Overall, the spatial distribution was similar to the spatial distribution of the eigenvector ofthe second leading mode in Fig. 2a, but the sign was reversed. This result indicates that the frequencyof TW-TCs in JJASO is influenced by large-scale circulations related to the PNA pattern in the precedingApril. Using April’s 500-hPa geopotential heights for points revealing high correlations (significant atthe 95% confidence level) in Fig. 4a, a new PNA index was developed and defined as follows:

PNA = −[HGT500(42.5N, 102.5E) + HGT500(40.5N, 100.0W) + HGT500(32.5N, 60.0W)

− HGT500(57.5N, 65.0W)], (1)

where the HGT500 represents April’s 500-hPa geopotential heights. The formula was multiplied by thenegative value because the correlation distribution shows a spatial pattern of the opposite polarity tothe eigenvector of the second leading mode in Fig. 2a.

The new PNA index obtained from this definition indicated strong positive correlations with thePC of the second leading mode in Fig. 2a (upper in Fig. 4b) and the index derived from the NOAA’s

K.-S. Choi, I.-J. Moon / Dynamics of Atmospheres and Oceans 63 (2013) 131– 141 135

Corr = -0.65

1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009

Prin

cip

le C

om

po

ne

nt

(PC

)

-4000

-2000

0

2000

4000

TC

nu

mb

er

0

2

4

6

8

10

PC (2 nd mode)

TC n umber

Trend (PC 2nd mode)

Trend (T C n umber)

Fig. 3. Comparison between the PC time series of the second leading mode in Fig. 2b (solid line with a closed circle) and thefrequency of TCs that influence Taiwan (TW-TC) (dotted line with an open circle) during JJASO.

CPC (lower panel in Fig. 4b), in which the coefficients were r = 0.68 and r = 0.65 (significant at the99% confidence level), respectively. In addition, the new PNA index had a strong negative correlation(r = −0.75, significant at the 99% confidence level) with the frequency of TW-TCs in JJASO (Fig. 5). Thiscorrelation was higher than that (r = −0.65) between the PC of the second leading mode in Fig. 2aand the frequency of TW-TCs in JJASO. The linear trends of the two time-series in Fig. 5 also revealedan out-of-phase relationship. A highly negative relation between them was maintained (r = −0.73,significant at the 99% confidence level), when the trends in two times series are removed.

3.2. Changes in the frequency of TW-TCs according to the PNA pattern

To examine the change in the frequency of TW-TCs in relation to the new PNA pattern and itsrelation to the large-scale atmospheric environment, eight of the highest PNA years (hereafter, positivePNA phase) and eight of the lowest PNA years (hereafter, negative PNA phase) were selected from thelast 33 years (Table 1).

It is well known that the PNA pattern is linked to tropical Pacific SST anomalies (SSTA), i.e., ENSO(Trenberth et al., 1998). Therefore, none of the summer El Nino and La Nina years were selected inorder to exclude the effect of ENSO, even though there was a low correlation (r = −0.08) between thePNA and the Nino-3.4 index in the same month. In this study, the test years were selected from the18 years remaining after excluding 10 El Nino (SSTA ≥ 0.5 ◦C) years (1982, 1986–87, 1991, 1994, 1997,2002, 2004, 2006, 2009) and 5 La Nina (SSTA ≤ −0.5 ◦C) years (1988, 1998, 1999, 2007, 2010), which

Table 1Comparison of the frequency of TW-TCs in JJASO in the positive PNA phase and in the negative PNA phase.

Positive PNA Negative PNA

Year TC number Year TC number

1979 3 1981 41980 3 1985 71983 1 1989 31984 3 1990 51993 1 1992 41995 4 2000 71996 2 2001 62008 4 2003 5Total 21 Total 41Average 2.6 Average 5.1Climatology 4 Climatology 4

136 K.-S. Choi, I.-J. Moon / Dynamics of Atmospheres and Oceans 63 (2013) 131– 141

Fig. 4. (a) Correlation map between the frequency of TW-TCs in JJASO and the geopotential height at 500 hPa in April and (b)comparisons of the new PNA index obtained from the present study, showing the PC of the second leading mode and NOAA’sCPC. In (a), the shaded areas are significant at the 95% confidence level.

were defined using SSTA in the Nino-3.4 region (5◦ S–5◦ N, 120◦ W–170◦ W) in JJASO. Even with theremoval of the trend in the new PNA index, the selected test years were not changed.

The remaining parts of this section analyze the differences in the frequency of TW-TCs betweenthe two PNA phases, as well as the large-scale environmental factors that cause such differences. Theanalysis revealed that the frequency of TW-TCs during the negative PNA phase (41 TCs) was twiceas high as that during the positive PNA phase (21 TCs). This difference in the frequency of TW-TCsbetween the two PNA phases was significant at the 99% confidence level. This result clearly indicatesthat TCs affect Taiwan more frequently in the summer during the negative PNA phase than during thepositive PNA phase. Actually, in the negative PNA phase, all years except 1989 (7 of 8 years, or 87.5%)exceeded the 33-year mean value (4 TCs), but during the positive PNA phase, only the years 1995 and2008 (2 of 8 years, or 25%) exceeded it (Table 1).

K.-S. Choi, I.-J. Moon / Dynamics of Atmospheres and Oceans 63 (2013) 131– 141 137

Corr = -0.75

1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009

Ne

w P

NA

in

de

x

-400

-200

0

200

400

TC

Nu

mb

er

0

2

4

6

8

10New P NA index

TC number

Trend (new PNA index)

Tren d (TC nu mber)

Fig. 5. Time series of the new PNA index (solid line with a closed circle) and the frequency (dotted line with an open circle) ofTW-TCs.

To investigate what causes the difference in the frequency of TW-TCs between the two PNA phases,the differences in the 500-hPa streamline pattern between the two PNA phases in April and JJASOwere analyzed (Fig. 6). Overall, the spatial distribution in April was similar to that of negative PNApattern (Fig. 6a). The spatial distribution in April also showed that Taiwan was under the influenceof an anomalous southwesterly from an anomalous cyclonic circulation, centered in the IndochinesePeninsula, and an anomalous easterly from an anomalous anticyclonic circulation, centered to thesoutheast of Lake Baikal.

The pattern in April continued to JJASO, particularly in the Taiwan area, where an anomalous east-erly in April maintains until October due to an anomalous cyclonic circulation centered in the east ofTaiwan, although there were some regional changes (Fig. 6b). This anomalous flow may play a role inmoving TCs to Taiwan as a result of the steering flow (George and Gray, 1976) in the negative PNAphase in JJASO.

As the difference in the large-scale atmospheric circulation between the two PNA phases mightaffect the activity of TCs in the WNP, we also analyzed the difference in the genesis frequency of TCs

Fig. 6. Difference in the 500-hPa streamlines the two PNA phases (negative minus positive) in (a) April and (b) JJASO. Theshaded areas are significant at the 95% confidence level.

138K

.-S. Choi,

I.-J. M

oon /

Dynam

ics of

Atm

ospheres and

Oceans

63 (2013) 131– 141

Fig. 7. Differences in the (a) TCGF and (b) the TCPF the two PNA phases (negative minus positive). In (a), the cross and the multiplication marks indicate the mean locations of the genesisof the TCs in the positive and the negative PNA phase, respectively. In (a) and (b), the small boxes inside the circles are significant at the 95% confidence level. In (b), the solid and thedashed lines represent 5875 gpm contours in the positive and the negative PNA phase, respectively.

K.-S. Choi, I.-J. Moon / Dynamics of Atmospheres and Oceans 63 (2013) 131– 141 139

TC l ifet ime (day)

Posit ive Negat ive0

6

12

18

24TC c entr al pre sur e (h Pa)

Posit ive Negat ive900

930

960

990

1020

10.38.5

965.8

973.7

Fig. 8. The TC’s lifetime (left panel) and its central pressure (right panel) at landfall in Taiwan between the negative and thepositive PNA phases. The boxes show the 25th and 75th percentiles, with the lines in the boxes marking the median and thecircles the values below (above) the 25th (75th) percentiles of the distributions. The numbers in the figure denote the averagesin each phase (cross marks).

(TCGF, Fig. 7a), the TCPF (Fig. 7b), and the TC’s intensity (Fig. 8) between the two PNA phases in JJASOin the WNP. The spatial distributions of the TCGF for each 5◦ × 5◦ latitude–longitude grid box revealedthat the genesis location of TCs generally tends to move northwestward in the positive PNA phase.This can be seen by comparing the average genesis locations of TCs for the two PNA phases (+ and× symbols in Fig. 7a). Differences of latitude and longitude in the genesis location of the TCs for thetwo PNA phases were significant at the 95% confidence levels. For total TC genesis in the WNP, thefrequency during the negative PNA phase was 20 TCs higher than that during the positive PNA phase(positive: 122, negative: 143), which was significant at the 90% confidence level. This may be becausethe anomalous cyclonic circulation intensified over the SWNP in the negative PNA phase (see Fig. 6b)and provided more favorable conditions for the genesis of TCs than in the positive PNA phase.

In TCPF, the TCs in the positive PNA phase moved mainly westward from the east of the Philippinesto southern China, whereas TCs in the negative PNA phase tended to move farther to the northeastalong the coast of East Asia, i.e., from the sea far to the east of the Philippines to Taiwan, Korea, andJapan through the East China Sea. These patterns are related to the difference in the circulation patternbetween the two PNA phases (Fig. 6b), i.e., anomalous southeasterlies in the negative PNA phase arestrengthened in the midlatitude areas of East Asia, including the East China Sea, Korea, Taiwan andJapan. In particular, the anomalous flows during the negative PNA phase can steer airflow that movesTCs to the midlatitude regions of East Asia.

On the other hand, the anomalous anticyclonic circulation in the east of Japan in Fig. 6b indicatesthat the WNP high (WNPH) during the negative PNA phase developed more to the north than theWNPH during the positive PNA phase. This can be seen by the 5875-gpm contour averaged for each ofthe phases (Fig. 7b). Since TCs tend to move along the western periphery of the WNPH (Lander, 1994;Chan, 2000), the differences in spatial distribution of the WNPH seems to influence the differences inboth TCPF and TCGF according to the PNA phase.

The difference in the TCGF and the TCPF between the two PNA phases may affect the intensity of theTCs. To examine this connection, the TC’s lifetime (duration) and the TC’s central pressure at landfallin Taiwan in each phase was analyzed (Fig. 8). The analysis revealed that TCs during the negative PNAphase have longer lifetimes and lower minimum central pressure than TCs during the positive WPphase, indicating that TCs that occur in the negative PNA phase are more intense and longer lastingthan those that occur during the positive PNA phase. The differences in the mean TC’s lifetime andcentral pressure between the two PNA phases were significant at the 90% and 95% confidence level,respectively. These differences are possibly due to a difference in the amount of energy supplied fromthe ocean between the two PNA phases. As shown in Fig. 7b, TCs that occurred during a negativePNA phase moved mainly over the sea, where they obtained enough energy from the ocean to inten-sify, whereas TCs that occurred during a positive PNA phase mostly made landfall in southern China,dissipating quickly due to the effect of the terrain.

140 K.-S. Choi, I.-J. Moon / Dynamics of Atmospheres and Oceans 63 (2013) 131– 141

Fig. 9. Differences in the OLR the two PNA phases (negative minus positive). The contour interval is 2 Wm−2. The shaded areasare significant at the 95% confidence level.

Fig. 10. A schematic illustrating large-scale environments related to TC activity in Taiwan in the negative PNA phase. Theabbreviations of ‘AC’ and ‘AA’ indicate ‘anomalous cyclone’ and ‘anomalous anticyclone’.

Characteristics of the occurrence and the intensity of TCs in the two PNA phases also can beexplained by the difference in OLR distribution between the two PNA phases (Fig. 9). Considering thatthe smaller the OLR is, the stronger the convective activity becomes, the main development regions forTCs along the western periphery of the WNPH, where the negative OLR dominated during a negativePNA phase, maintained more favorable conditions during this phase for the genesis of TCs in JJASO.Overall, these results support the idea that TCs during a negative PNA phase were more active alongthe coast of East Asia, including Taiwan, Korea, and Japan.

4. Summary and conclusions

This study found a strong negative correlation between the frequency of TW-TCs in JJASO andthe PNA index in April over a 33-year period. The PNA index was redefined using April’s 500-hPageopotential heights for points revealing high correlations between the frequency of TW-TCs in JJASOand 500-hPa geopotential heights in April. The new PNA index showed a high correlation with the PCof the second leading mode (frequently defined as the PNA index) obtained from REOF analysis usingApril’s 500-hPa geopotential height anomalies.

From April to JJASO in the negative PNA phase, anomalous cyclonic and anomalous anticycloniccirculations were intensified at low latitudes and midlatitudes, respectively, from East Asia to theNorth Atlantic. This implies that the anomalous southeasterly converged between the anomalouscyclonic circulation in low latitudes and that the anomalous anticyclonic circulation in midlatitudesin East Asia played a decisive role in moving TCs not only to Taiwan, but also to the midlatitude coastalregions in East Asia as a result of the steering flow (see schematic diagram in Fig. 10). In addition, asouthwestward extension of a WNPH during the positive PNA phase also contributed to a frequentmovement of TCs to southern China without traveling north toward the midlatitude regions of East

K.-S. Choi, I.-J. Moon / Dynamics of Atmospheres and Oceans 63 (2013) 131– 141 141

Asia, resulting in a decrease of the average TC intensity due to the effect of the terrain when TCslanded in southern China. Different characteristics in the activity of TCs in the two PNA phases werealso confirmed from the difference in the OLR distribution between the two PNA phases.

From the above results, we propose that the PNA pattern in the preceding April can be a majorfactor in the seasonal prediction of the frequency of TW-TCs in JJASO. We will develop a statisticalmodel for the prediction of TW-TCs using this factor in the future.

Acknowledgments

This research was supported by the 2013 scientific promotion program funded by Jeju NationalUniversity. The authors would like to thank the anonymous reviewers for their valuable commentsand suggestions.

References

Chan, J.C.-L., 2000. Tropical cyclone activity over the western North Pacific associated with El Nino and La Nina events. J. Climate13, 2960–2972.

Chu, P.-S., Zhao, X., Lee, C.-T., Lu, M.-M., 2007. Climate prediction of tropical cyclone activity in the vicinity of Taiwan using themultivariate least absolute deviation regression method. Terr. Atmos. Oceanic Sci. 18, 805–825.

Coleman, J.S.M., Rogers, J.C., 1995. Ohio River Valley winter moisture condition associated with the Pacific-North Americanteleconnection pattern. J. Climate 16, 969–981.

George, J.E., Gray, W.M., 1976. Tropical cyclone motion and surrounding parameter relationships. J. Appl. Meteorol. 15,1252–1264.

Gong, D.-Y., Mao, R., Shi, P.-J., Fan, Y.-D., 2007. Correlation between east Asian dust storm frequency and PNA. Geophys. Res.Lett. 34, L14710, http://dx.doi.org/10.1029/2007GL029944.

Isard, S.A., 1999. Zones of origin for Great Lakes cyclones in North America, 1899–1996. Mon. Weather Rev. 128, 474–485.Kanamitsu, M., Ebisuzake, W., Woolen, J., Yang, S.-K., Hnilo, J.J., Fiorino, M., Potter, G.L., 2002. NCEP–DOE AMIP-II reanalysis

(R-2): dynamical seasonal forecast system 2000. Bull. Am. Meteorol. Soc. 83, 1631–1643.Klotzbach, P.J., Gray, W.M., 2004. Updated 6–11-month prediction of Atlantic basin Seasonal hurricane activity. Weather Fore-

cast. 19, 917–934.Lander, M.A., 1994. An exploratory analysis of the relationship between tropical storm formation in the western North Pacific

and ENSO. Mon. Weather Rev. 122, 636–651.Leathers, D.J., Yarnal, B., Palecki, M.A., 1991a. The Pacific/North American teleconnection pattern and United States climate. Part

I: regional temperature and precipitation associations. J. Climate 4, 517–528.Leathers, D.J., Yarnal, B., Palecki, M.A., 1991b. The Pacific/North American teleconnection pattern and United States climate. Part

II: temporal characteristics and index specification. J. Climate 4, 517–528.Leathers, D.J., Palecki, M.A., 1992. The Pacific/North American teleconnection pattern and United States climate, Part II: temporal

characteristics and index specification. J. Climate 5, 707–716.Liebmann, B., Smith, C.A., 1996. Description of a complete (interpolated) outgoing longwave radiation dataset. Bull. Am. Mete-

orol. Soc. 77, 1275–1277.Lin, J.L., Zhang, M.H., Mapes, B.E., 2005. Zonal momentum budget of the Madden–Julian oscillation: the source and strength of

equivalent linear damping. J. Atmos. Sci. 62, 2172–2188.Peings, Y., Douville, H., Terray, P., 2009. Extended winter Pacific North America oscillation as a precursor of the Indian summer

monsoon rainfall. Geophys. Res. Lett. 36, L11710, http://dx.doi.org/10.1029/2009GL038453.Reynolds, R.W., Rayner, N.A., Smith, T.M., Stokes, D.C., Wang, W., 2002. An improved in situ and satellite SST analysis for climate.

J. Climatol. 15, 1609–1625.Renwick, J.A., Wallace, J.M., 1996. Relationship between North Pacific wintertime blocking, El Nino, and the PNA pattern. Mon.

Weather Rev. 124, 2071–2076.Straus, D.M., Shukla, J., 2002. Does ENSO force the PNA? J. Climate 15, 2340–2358.Trenberth, K.E., Branstator, G.W., Karoly, D., Kumar, A., Lau, N., Ropelewski, C., 1998. Progress during TOGA in understanding

and modeling global teleconnections associated with tropical sea surface temperature. J. Geophys. Res. 103, 12324–14291.Trouet, V., Taylor, A.H., 2009. Multi-century variability in the Pacific North American. Clim. Dyn.,

http://dx.doi.org/10.1007/s00382-009-0605-9.Wallace, J.M., 2000. North Atlantic oscillation/annular mode: two paradigms-one phenomenon. Q. J. R. Meteorol. Soc. 126,

791–805.Wallace, J.M., Gutzler, D.S., 1981. Teleconnections in the geopotential height field during the Northern Hemisphere winter. Mon.

Weather Rev. 109, 784–812.Wang, C., 2010. Is hurricane activity in one basin tied to another? Eos 91, 93–100.Wang, B., Chan, J.C.-L., 2002. How strong ENSO events affect tropical storm activity over the western North Pacific. J. Climate

15, 1643–1658.Wang, B., Wu, R., Fu, X., 2000. Pacific-east Asian teleconnection: how does ENSO affect east Asian climate? J. Climate 13,

1517–1536.Wilks, D.S., 1995. Statistical methods in the atmospheric sciences. Academic Press, pp. 467.