Ports.com

HAKODATE

TROPICAL CYCLONES AFFECTING HAKODATE

Tropical Cyclone Climatology at Hakodate

For the purpose of this study, any tropical storm or typhoon approaching within 180 nmi of Hakodate is considered to represent a threat to the port. Table V-42 contains a descriptive history of all 46 tropical storms and typhoons passing within 180 nmi of Hakodate during the 48-year period 1945-1992. Unless otherwise indicated, all of the tropical cyclone statistics used in this report for storms passing within 180 nmi of Hakodate are based on the data set used to compile Table V-42.

Tropical cyclones which affect Japan generally form in the area bounded by 5ºN to 30ºN and 120ºE to 165ºE. The latitudinal boundaries shift poleward during the summer months and then equatorward in winter in response to the seasonal changes of the synoptic environment.

Considering the entire western North Pacific basin, about two-thirds of the tropical cyclones reach at least typhoon intensity at some point in their life cycle. In specific reference to the 46 tropical cyclones listed in Table V-42, all except Ruby of 1985 reached typhoon intensity at some point in their history. Indeed, the intensity of 16 of the listed storms reached what is referred to as super-typhoon intensity (at least 130 kt) earlier in their life cycle. Although there is a positive correlation between the maximum storm intensity and the eventual intensity of the storm at CPA to Hakodate, the relationship is very weak. Much depends on the track of storm before it reaches Hakodate.

Tropical cyclones are nurtured by a warm marine environment. In this basin maximum storm intensity typically occurs between 20ºN and 25ºN where sea-surface temperatures average near 84ºF (29ºC) during the month of August. However, after recurvature under the influence of the mid-latitude westerlies and a colder environment, tropical cyclones lose their tropical characteristics. In this situation, the size of the circulation usually expands, the speed of the maximum wind decreases, the translational (forward) speed of motion increases and the distribution of winds, rainfall and temperature becomes increasingly asymmetric.

The primary tropical cyclone season for Hakodate extends from early August through late September. As shown in Table V-43, tropical storms have passed within 180 nmi of Hakodate as early as June, and as late as December. There were no occurrences recorded during the five-month period of January through May, and during November. It can be seen in the table that few typhoon-strength storms penetrate the 180 nmi threat radius around Hakodate. Only 4 of the 46 storms occurring during the 48-year period of record have been of typhoon strength when at their closest point of approach (CPA) to Hakodate. All four occurred during the last half of September.

Table V-43 also shows the motion history of the 46 tropical storms and typhoons which passed within 180 nmi of Hakodate during the period 1945-1992. The average storm speed at CPA to Hakodate is 35 kt. The reason for this relatively rapid speed of movement is because of in the location of Hakodate. As shown in Appendix A, most tropical cyclones that pass close to Hakodate have already recurved and are moving northeastward under the influence of upper level westerlies. Since most storms accelerate after recurvature, the storms affecting Hakodate are in the acceleration phase, and rapid movement is common. The average movement for all storms is 041º/35 kt. Storms occurring during September have an average movement speed of 38 kt at CPA. As shown in Table V-42, it is not uncommon for late season storms (September and October) to exceed 50 kt.

During the 48-year period from 1945 through 1992 there were 46 tropical storms and typhoons that met the 180 nmi threat criterion for Hakodate. Figure V-174 shows the monthly distribution of the 46 storms by 7-day periods. The figure clearly shows that the period of peak activity extends from early August through late September.

Figure V-175 depicts the annual distribution of tropical storms and typhoons passing within 180 nmi of Hakodate during the 48-year period 1945-1992. Although the occurrence of 46 storms in 48 years suggests an average of nearly one storm per year passing within 180 nmi of Hakodate, there were many years when this event did not occur and likewise many years when there were multiple occurrences. For example, the figure shows that 30 tropical cyclones entered the 180 nmi threat radius around Hakodate during the 19-year period 1949 through 1967, an average of approximately 1.6 per year. Only one year during the 19-year period had no storm activity recorded. However, the 21-year period 1968 through 1988 had only 9 storms enter the 180 nmi threat radius, an average of only 0.4 per year. No tropical cyclone entered the 180 nmi threat radius of Hakodate during two-thirds (14 of 21) of the years in the latter period.

Figure V-176 depicts, on an 8-point compass, the octants from which the 46 tropical cyclones in the data set approached Hakodate. Sixty-one percent (28 of 46) of the storms approached Hakodate from the southwest octant, with the remainder evenly split between the west and south octants. The approach direction is determined at CPA, and may not represent the initial approach direction of the tropical cyclone toward Hakodate.

Because of climatological considerations, there are preferred areas of the western North Pacific basin from which tropical cyclones eventually affect Hakodate. However, there are some tropical cyclones, which, even though they traverse these preferred areas, do not affect Hakodate. Also, as might be expected, there are seasonal shifts to these preferred areas.

Figures V-177 and V-178 address the probability of tropical cyclones affecting Hakodate. Using a grid system, a tabulation was made of the total number of tropical cyclones passing through a given grid area regardless of whether they eventually passed within 180 nmi of Hakodate. A further tabulation was made of those storms which did eventually pass within that distance from Hakodate. After smoothing, the two tabulations were converted into probabilities and contours were drawn to connect points of equal probability.

The solid lines on the figures represent a "percent threat" for any tropical cyclone location within the depicted area. The heavy, dashed lines represent the approximate time in days for a system to reach Hakodate. For example, in Figure V-177, during the months of July and August a tropical cyclone located at 30ºN 135ºE has an approximate 25% probability of passing within 180 nmi of Hakodate and would reach Hakodate in about 1-1/2 to 2 days.

A comparison of Figures V-177 and V-178 shows that the threat axes for different times of the year are similar north of 35ºN. South of 35ºN, the July and August axis is broader, and the maximum threat axis is farther east than the one for the September through June period. Because of Hakodate's northerly location, essentially all of the storms that enter its 180 nmi threat radius have recurved and their movement is being influenced by upper level westerly winds. Consequently, they are moving in a general north to northeasterly direction. To pass within 180 nmi of Hakodate, the majority of the storms cross southern Honshu between 131ºE and 140ºE and approach Hakodate through the eastern Sea of Japan. A few pass through the Tsushima (Korea) Strait or across Korea before entering the Sea of Japan. North of 35ºN, some travel northward over land, paralleling the Japanese Alps along the length of Honshu.

Wind and Topographical Effects

The wind observation site for the Port of Hakodate is located on a mast atop the Maritime Safety Administration building on Central Wharf (Figure V-172). The building is eight stories high. The mast on which the anemometer is located is approximately 8 ft tall and located on top of the building. Consequently, the anemometer is estimated to be between 85 and 95 ft above the water level of the harbor. Winds measured by the anemometer are considered by local authorities to be representative of harbor winds except in easterly wind situations. When easterly winds are observed, the authorities state that the Inner Harbor anchorage areas experience winds about 3 to 5 kt higher than the anemometer indicates, and the Outer Harbor anchorages experience winds about 6 to 10 kt higher. The differences in velocity are believed to be caused by the turbulence and blocking effects of the many buildings located adjacent to the east side of the harbor. The effects of the buildings on easterly winds would decrease with increased westward distance from the buildings. Consequently when easterly winds prevail, the anchorages would experience stronger wind velocities than the east side of the inner harbor, and the westernmost anchorages would experience the strongest winds.

Local Weather Conditions

Observational data are available for the Hakodate area only for the period 1965 through 1992. Two observation sites recorded wind data, and neither is located at the harbor. One site is located in the city of Hakodate approximately 1.9 nmi northeast of the north part of Hakodate's Inner Harbor and at an elevation of 118 ft (36 m). The exposure of the anemometer is not known. A second observation site is located at Hakodate Airport, approximately 4.5 nmi east of the east side of the Inner Harbor near the north coast of Tsugaru Strait. The exact elevation of the airport observation site is not known, but it appears to be low lying due to its close proximity to the coast. Data from the airport site is available for the period 1975-1992 only. Observational data from each site during the passage of selected tropical cyclones are listed in Table V-44. The data contained in Table V-44 have been selected from observations recorded at Hakodate during the passage of the tropical cyclones listed in the table.

An examination of the data in Table V-44 shows that there is an apparent, significant difference in the exposure of the two observation sites to wind. Two tropical cyclone passages listed in the tables occurred during the period when observations were available from both sites. The first, Thad in August 1981, caused east-southeasterly winds of 26 kt at the city site, while the airport site recorded easterly winds of 38 kt with gusts to 59 kt. The second passage, Mirielle in September 1991, brought southwesterly winds of 23 kt to the city site, but the airport site recorded southwesterly winds of 40 kt with gusts to 67 kt.

Figures V-179 and V-180 show tracks and track segments of tropical cyclones considered to have a high probability of having produced sustained winds of 22 kt and 34 kt over water near Hakodate during the 49-year period 1945-1993.

It should be noted that the two figures are derived from theoretical calculations because of the limited observational data that are available. The figures are based on: (1) storm intensity (maximum wind near center), (2) distance of the storm center from Hakodate, (3) bearing of the storm from Hakodate, (4) translational speed of the storm, and (5) frictional characteristics of the terrain between the storm and Hakodate.

The beginning and end points of the arrows in the right panel of each figure show the positions of the tropical cyclone centers when sustained winds 22 kt and 34 kt would, theoretically, have begun and ended at Hakodate.

Wave Motion

The Outer Harbor is exposed and vulnerable to wave motion when winds from southeast through southwest are experienced. Specific wave heights experienced at the port were not identified by local harbor authorities, but the Outer Harbor would essentially experience the same open ocean wave heights predominating in Tsugaru Strait. Consequently, combined (wind wave plus swell) wave heights of 6.5 to 10 ft (2 to 3 m) or higher could be experienced by ships in the outer anchorage, especially when southwesterly winds prevail. Local harbor authorities indicate that maximum wave height in the Inner Harbor is limited to about 3.3 ft (1 m).

The size and orientation of Hakodate Bay limits fetch and precludes large wind waves from generating and reaching the port when winds are from west clockwise through east.

Storm Surge

Storm surge may be visualized as a raised dome of water, moving with the storm, and centered a few miles to the right of its path. The dome height is related to local pressure (i.e., a barometric effect dependent on the intensity of the storm) and to wind stress on the water caused by local winds. Other significant contributing factors are storm speed, direction of approach, bottom topography, and coincidence with astronomical tide.

The worst combination of circumstances (Harris, 1963, and Pore and Barrientos, 1976) would include:

  (1) An intense storm approaching perpendicular to the coast with the harbor within 30 nmi to the right of the storm's track.

  (2) Broad, shallow, slowly shoaling bathymetry.

  (3) Coincidence with high astronomical tide.

Local harbor authorities state that increased water levels due to storm surge have not been a significant problem in the past. Local records obtained during a 1993 port visit indicate that the maximum storm surge height experienced in recent years occurred in August 1981 when tropical cyclone Thad passed within 8 nmi of the port. A water rise of 2.4 ft (72 cm) was recorded for that storm. Other rises of 1.15 to 1.54 ft (35 to 47 cm) have been recorded since that occurrence, but some were due to extratropical low pressure systems rather than tropical cyclones. Tsugaru Strait is configured so that the only wind flow that would likely cause damaging storm surge at Hakodate would be from south clockwise through southwest. A strong, slow moving storm passing just west of Hakodate could potentially cause a storm surge at the port. However, the generally rapid movement of most tropical cyclones as they approach Hakodate would likely prevent the sustained wind flow necessary to cause a significant storm surge.

Source: http://www.nrlmry.navy.mil/port_studies/thh-nc/japan/hakodate/text/sect5.htm