With 20-year (1975-1994) climatological data, we demonstrate that the tropical storm track over the western North Pacific (0-40oN, 100-180oE) exhibits prominent subseasonal variations on a time scale of about 40 days from May to November. The storm track variability is regulated by the conspicuous Climatological Intraseasonal Oscillation (CISO) in the strength of the western North Pacific summer monsoon and the associated position of the western Pacific Subtropical High. The CISO cycle regulates the number of tropical storm formation during the Pre-Onset and Withdraw Cycles but not during the Onset and Peak Monsoon Cycles (from mid-June to mid-September).
Harr and Elsberry (1991) demonstrated that the tropical cyclone track types exhibit evident intraseasonal variations during the year of 1980: The straight-moving and recurving tracks occurred alternatively during that summer from June to October. Other years appear to have similar persistent periods of alternating track types. Normally, the intraseasonal variations of tracks are subject to year-to-year variability. Therefore, it is not clear to what extent the tropical cyclone tracks have subseasonal variation in the climatological sense. Recently, Uedo et al. (1995) noticed a dramatic change in tropical cyclone tracks in late July and early August, accompanying the end of Baiu, implying that the intraseasonal variation of tropical cyclone tracks may be phase locked to annual cycle.
In a recent paper, Wang and Xu (1996) (hereafter WX96) have demonstrated that the Northern Hemisphere (NH) summer monsoon, especially the western North Pacific summer monsoon (WNPSM), displays statistically significant climatological intraseasonal oscillation (CISO). Four CISO cycles were identified in the NH summer monsoon domain from May to November. The extreme phases of CISO cycles characterize subseasonal variations of the NH summer monsoon (i.e., the onset, peak, and withdraw, as well as active/break monsoon periods) and link monsoon singularities (the weather events which occur on a fixed pentad with usual regularity) at various monsoon regions: WNPSM, Indian summer monsoon, and East Asian subtropical monsoon.
If the WNPSM has prominent climatological intraseasonal oscillation, the corresponding circulation changes would affect tropical storm activity. The purpose of the present paper is to explore the climatological intraseasonal variability of tropical storm activity. What we discovered is that the storm tracks indeed display significant climatological intraseasonal variations on a time scale of about 40 days. In section 2 we briefly describe the data used for the present study. In section 3 we present evidence of the CISO of WNPSM and define four primary CISO cycles using outgoing longwave radiation (OLR) pentad mean time series. In section 4 we contrast tropical storm tracks during the extremely wet and dry phases of each CISO cycle, describe the prominent changes in storm tracks, and explore how the circulation change affect the storm tracks.
Our analysis focus on the western North Pacific region (0-45oN, 100-180oE). The original tropical storm (defined as the tropical cyclones with maximum wind speed exceeding 17 m/s) track data were obtained from the Joint Typhoon Warning Center (JTWC) at Guam. The data contains a summary of the best-track position and the maximum sustained wind at 6-h intervals for the tropical storms from 1945 to 1994. It is found that at the formation time, the maximum sustained wind of some storms before 1973 is much larger than 17 m/s. This may add uncertainty to the formation positions. We made a tropical storm track climatology using 20 years of data from 1975-1994. The reason for choosing this period are three-fold. First, during this period the track data are most reliable. Second, The CISO revealed by WX96 used OLR observations for the same period and European Centre for Medium-range Weather Forecast circulation data from 1979-1994. Third, there was a notable interdecadal shift of the climate in the Pacific SST after 1976 (Nitta and Yamada 1990, Trenberth 1990) which have affected El Nino (Wang 1995) and rainfall in the east Asia. The 1975-1994 period are not affected by this interdecadal climate shift.
CLIMATOLOGICAL Intraseasonal Oscillation Cycles in the western North Pacific
The WNPSM is characterized by heavy rainfall and southwesterlies associated with the monsoon rains. WX96 used climatological pentad mean OLR averaged over the core region of WNPSM (10-20oN, 120-140oE) to represent the intensity of the WNPSM. They have shown that this climatological pentad mean OLR can be decomposed into a smoothed annual cycle (the sum of the first three Fourier harmonics) and a CISO component which reflects variations on a time scale of 20-72 days. They have demonstrated the significance of the CISO by using three statistical tests and by checking the physical consistence between OLR and circulation. This CISO component was used as an index to measure the intensity change of WNPSM. We will use this CISO Index of WNPSM to define subseasonal cycles and associated wet and dry phases.
Figure 1 presents the OLR CISO Index for WNPSM. Obviously, there are four apparent cycles with contrasting wet and dry phases as shown on the figure. The first cycle spanning from May 11 to June 14 is termed as Pre-Monsoon Cycle. Its wet phase (Pentad 28-29 or May 16-25) marks the onset of the South China Sea summer monsoon, while its dry phase (P31-32, May 31-June 9) characterizes pre-monsoon dry weather over the vast area of the western North Pacific. The second cycle lasting from June 15 to July 24 is called Onset Cycle because during its wet phase (P34-35, June 15-24) the WNPSM begins. Its dry phase (P38-39, July 5 to 14) represents the first break monsoon after the onset of WNPSM. During the wet phase of the cycle III (P46-47 or August 14-23), the WNPSM reaches its height and for this reason we call it Peak Cycle. The dry phase of the Peak cycle (P49-50, or August 29-September 7) indicates the second, also the strongest break monsoon. The fourth major cycle consists of a wet phase (P58-59, or Oct. 13-22) that corresponds to the last wet spell of the WNPSM and a dry phase (P62-63, or November 2-11) which marks the withdraw of the WNPSM. In between Cycle II and III there is a short special transitional period. So is in between Cycle III and IV. We will not discuss these two special transition periods in the present paper.
WX96 have shown that the occurrence of each of the wet or dry phases are accompanied by abrupt changes in OLR and large-scale circulation systems. The transitions between wet and dry phases are in a discontinuous manner. They are relatively quick adjustment between two relatively steady evolution stages.
SUBSEASONAL variations of the tropical storm tracks
With 20-year (1975-1994) climatological data, we have demonstrated that the tropical storm track over the western North Pacific (0-40oN, 100-180oE) exhibits pronounced subseasonal variations on a time scale of about 40 days. From mid-May to mid-November, four distinct Climatological Intraseasonal Cycles (CISOs), each with contrasting storm track patterns are revealed.
During the Pre-Monsoon Cycle, the TS activity concentrated in the wet phase (May 15-24) with recurvature track dominant (recurving latitude is about 20oN) (Fig. 2). Little activity occurs in the dry Phase (May 31-June 9). During the Onset Cycle, The wet phase (June 15-24) is characterized by recurving tracks with a recurvature latitude at 25oN, whereas the dry phase (July 5-14) is dominated by northwestward straight movers (Fig. 3). During the Peak Cycle, although the track difference south of 30oN are not apparent , the tracks north of 30oN have striking patterns between the wet and dry phases (Fig. 4). The wet phase (August 14-23) tracks are predominantly northward and concentrated in the longitude band between 120-140oE, whereas the dry phase (August 29-September 7) tracks are primarily northeastward and shifts eastward by about 10o longitudes. During the Withdraw Cycle, the tracks exhibit co-existence of recurver and westward straight movers, but the wet phase (October 13-22) recurvature occurs at about 22-25oN and more westward while the dry phase recurvature occur at 17-20oN and more to the east. Moreover, the wet phase straight tracks are located more northward than their counterparts in the dry phase, resulting a notable difference in the number of TS invading northern South China Sea.
The storm track variability is regulated by the conspicuous Climatological Intraseasonal Oscillation (CISO) in the strength of the western North Pacific summer monsoon and the associated changes in the position of the western Pacific Subtropical High. The CISO cycle regulates the number of tropical storm formation during the Pre-Onset and Withdraw Cycles but not during the Onset and Peak Monsoon Cycles (from mid-June to mid-September).
This research is supported by the Marine Meteorology Program of the Office of Naval Research under the grant N00014-96-1-0796. This is the School of Ocean and Earth Science and Technology publication number XXXX.
Harr, P. A., and R. L. Elsberry, 1991: Tropical cyclone track characterics as a function of large-scale circulation anomalies. Mon. Wea. Rev., 119, 1448-1468.
Nitta, T., and S. Yamada, 1989: Recent warming of tropical surface temperature and its relationship to the Northern Hemisphere circulation. J. Meteor. Soc. Japan, 67,375-383.
Trenberth, K. E., 1990: Recent observed interdecadal climate changes in the Northern Hemisphere. Bull. Amer. Meteor. Soc., 71, 988-993.
Ueda, H., T. Yasunari and R. Kawamura, 1995: Abrupt seasonal change of large-scale convection activity over the western Pacific in northern summer. J. Meteor. Soc. Japan, 73, 795-809.
Wang, B., 1995: Interdecadal changes in El Nino onset in the last four decades. J. Climate, 8, 267-285.
Wang, B. and X. Xu, 1996: Northern Hemisphere summer monsoon singularities and climatological intraseasonal oscillation, submitted to J. Climate.
Back to top | Back to menu