A PREI·TMINARY REPORT - University of Hawaii...Methodology of study and scope of preliminary report • Tsunami of 1813 or 1814 Tsunamis of December 1854. • Tsunami of June 1896.

IAPARBSE TSUHAIDS IN SAWAD

A PREI·TMINARY REPORT

By Doak C. Cox

January 1980

Environmental Center University of Hawaii

CONTENTS

Introduction

Background and purposes of study

Methodology of study and scope of preliminary report •

Tsunami of 1813 or 1814

Tsunamis of December 1854. •

Tsunami of June 1896.

Introduction •

Contemporary accounts •

Subsequent accounts.

Arrival times.

Runup heights

Kailua-Kona. Keauhou, Kaawaloa, and Napoopoo • Hookena and places to the south on the Kona coast • Kaalualu, Honuapo, and Punaluu • Other Hawaii sites. Sites on Maui • Sites on Oahu • • Sites on Kauai. • Summary.

Tsunamis of August 1901

Introduction •

Contemporary accounts • •

Subsequent accounts.

Arrival times.

Runup heights

Tsunami of March 1933

Introduction •

Contemporary accounts •

Subsequent accounts. \. . Warnings, characteristics, and effects •

Arrival times and travel times

Runup heights

Hawaii Maui. Oahu. Kauai

Tsunami of March 1952 •

Introduction •

Hawaiian effects •

iii

. . . .

1

1

1

4

5

8

8

8

9

15

15

15 17 20 21 21 21 22 22 22

23

23

23

26

26

27

32

32

32

32

32

32

36

36 36 36 36

37

37

37

Tsunami of May 1968 •

Introduction • • •

Hawaiian effects •

Travel times and runup heights • •

Summary of historic record •

General comments

Travel times. • Runup heights

Ranges of possible error and most probable values Historic record • • • • •

Longitudinal runup profiles

Needs and schemes for reconstruction •

Reconstruction scheme used in this studY. • Reconstruction principles

Use of historic information generally. • • • • • Datum plane • • • • • • • • • • • • • Most probable runup values and limits of error. Plotting positions and scales. • • • • • • • • Use of geophysical information • • • • • • • Treatment of ranges of uncertainty in runup and of local detail. • Treatment of upper limits. • • • • • • • • • Treatment of questionable values • • • • • Interpolation, extrapolation, and treatment

of runup values of uncertain location Rationale for modifications of principles •

Comparison of model and non-model schemes for profile reconstruction

Site numbering systems and profile orientation •

Hawaii profiles. • • . • • • • • • • • • • • •

1896, 1901, and 1933 profiles, Kona and South Kohala coast • 1896, 1901, and 1933 profiles, Kau coast 1896, 1901, and 1933 profiles, Hilo • • • • •• • • 1896, 1901, and 1933 profiles, elsewhere • • • • • • • • • • Other profiles • • • • • •

Maul profiles.

Kauai profiles

Profiles for other islands. •

Acknowledgements •

References.

iv

. . . . .< . .

Page

40 40 40 40

42

42

42

44 44 45

48

48

48

49

49 49 49 50 50 50 50 51

51 52 52 54

54

54 5'1 5'1 5'1 5'1

5'1 62 62

65

66

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

TABLES

Japanese tsunamis reported to be observed or recorded in Hawaii ••

Characteristics of earthquakes accompanied by Great Ansei tsunamis

Reported times, sizes, and character of waves, tsunami of June 1896 in Hawaii. • • • • • •

Reported effects of waves, tsunami of June 1896 in Hawaii •

Hawaiian arrival times and travel times, tsunami of June 1896

Hawaiian runup heights, tsunami of June 1896. • • • • • •

Data on earthquakes and tsunamis of August 1901 in Japan •

Reported times, sizes, character and effects of tsunami waves of August 1901 in Hawaii. . • • • • • • • • • • • • ',' • • • • •

Hawaiian arrival times and travel times of tsunamis of August 1901 •

Hawaiian runup heights of the tsunami of August 1901

Reported times, characteristics, and effects, tsunami of 4 March 1933 in Hawaii • • • • • • • • . • • •

Hawaiian arrival times and travel times, tsunami of March 1933.

Hawaiian runup heights, tsunami of March 1933. • • • • • • •

Marigraphic characteristics of the tsunami of March 1952 in Hawaii •

Hawaiian runup heights of the tsunami of March 1952 • • • • • • •

5

10

12

16

18

23

24

27

29

30

33

34

37

38

16. Hawaiian marigraphic data and runup heights of tsunami of May 1968 • • •• 41

17. Travel times of Japanese tsunamis to Hawaii • • • • • • • • 43

18.

19.

Runup heights of Japanese tsunamis in Hawaii

Comparison of 1896 and 1933 tsunami runup heights on the island of Hawaii used in National Flood Insurance Program (NFIP) with those considered most probable in this study • • • • • • • • • • • • • • • • • • • . • • • • • • •

v

46

47

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

PIGURES

Chart of the Pacific showing tsunami travel time to Hawaii ••

Kailua-Kona in 1892. . • • • • • • • • • • • • • • • • •

Runup profile of 1946 Aleutian tsunami along north coasts of Maui ••

Map of Kaual showing tsunami runup sites

Map of Maui showing tsunami runup sites •

Map ot northern part of Hawaii showing tsunami runup sites

Map ot southern part ot Hawaii showing tsunami runup sites

Reconstructed runup profiles ot Japanese tsunamis: Northeast and Puna coasts, Hawaii • • • • • • •

Reconstructed runup profiles ot Japanese tsunamis: Kau and west coasts, Hawaii • • • • • • • • • •

Reconstructed runup profiles ot Japanese tsunamis: Maui

Reconstructed runup profiles of Japanese tsunamis: Kauai •

vi

14

53

55

56

58

59

60

61

63

64

1

INTRODUCTION

Background and purposes of study

The study leading to this report was undertaken because of the importance of historic information in tsunami hazard management, and particularly in the identification of tsunami hazard zones in the National Flood Insurance Program whose full implementation in Hawaii is scheduled in the near future. In this program, the hazard zones are intended to represent the extent of tsunami inundation expectable, on the average, once in a hundred years.

The expectable extent of inundation at any coastal site is to be estimated from the measured or estimated runup heights of historic tsunamis at the site. The methodology of inundation zone estimation has been presented elsewhere (Houston et al., 1977; Cox, 1979).

A previous report (Cox and Morgan, 1977) on the history of tsunamis generated or possibly generated along Hawaiian coasts indicated, on the one hand, that earlier accepted reports of some tsunami runups were invalid and, on the other hand, that there had been some evidences of runups, and even some tsunamis, that had been overlooked in compiling the record used in the National Flood Insurance Program. A subsequent study (Cox, 1979) was undertaken to develop, from all of the historical information available, reasonable estimates of the runup profile of each local or possibly local tsunami along each coast where the runup was significant.

It seemed clear from these studies that a thorough review should be made of the historical information available for the tsunamis from distant sources that were known to have been observed in Hawaii or that might be suspected of having significant effects here.

Specific questions have been raised concerning the validity of the highest runups in the record for the Kona coast of Hawaii being used in the National Flood Insurance Program. These were runups of the Japanese tsunami of June 1896. Rather than review the historical information available for that tsunami alone, it seemed best to review the entire record of Hawaiian experience with Japanese tsunamis, so that the runup patterns of all of the tsunamis could be compared. The result of such a study could be considered the first phase of a general review that should, at some point, be extended to cover the record of tsunamis from other distant source regions.

This preliminary report is the first product of the review of the historical record of Japanese tsunamis in Hawaii. It reflects an extensive search for and review of local contemporary documentation of the Hawaiian effects of the Japanese tsunamis that have been reported to be observed or recorded here. It also includes some notes on a tsunami reported in Hawaii early in the 19th century that might conceivably have been of Japanese origin.

No attempt has been made yet to identify, from the Japanese historical record, additional tsunamis of Japanese origin that might have been observed or recorded in Hawaii, and only a few of the marigraphic records of the Japanese tsunamis known to have been recorded in Hawaii have been available to the study at this point. However, it seems unlikely that information of direct significance in hazard zoning will be turned up in the completion of the study. Hence this preliminary report has been prepared in the hopes that the information it presents will be used in the determination of the tsunami hazard zones before they are adopted by the National Flood Insurance Program.

Methodology of study and scope of preliminary report

According to the most recent Hawaiian tsunami catalog (Pararas-Carayannis and Calebaugh, 1977), 13 Japanese tsunamis have been observed in Hawaii or recorded by tide gages in the Islands. One more tsunami observed in Hawaii may have originated off Japan, and one more originating off Japan may have been observed or recorded in Hawaii. The fifteen Japanese or possibly Japanese tsunamis previously reported or suspected of being observed in Hawaii are summarized in Table 1.

The nine tsunamis discussed in this report (indicated by asterisks in Table 1), include the five for which there are reports of Hawaiian observations or of amplitudes exceeding 0.3 meters on Hawaiian tide gage records in the Hawaiian catalog (the tsunamis of December 1896, March 1933, March 1933, March 1952, and May 1968, and one of those of December 1901), a second tsunami of December 1901 for which there was a suspicion of Hawaiian observation, the two great tsunamis of December 1854

2

Date

Table 1. Japanese tsunamis reported to be observed or recorded'in Hawaii

Maxfl1lUlll I'Unup height.

Tsunata) Japan )

( .. ten )b

Hawaiian re~rts xilllUlll

Arrival runup hetje~S date (Illeters

g)*1813-14 Tsunami of unknown ortgin 1

*1854 Dec 23 Great Ansei Tokaido tsunami 10 22 Decd) 1

*1854 Dec 23 Great Ansei Nankaido tsunami 6 23 Deed) 1

*1896 Jun 15 Great Meijt Sanriku tsunami 27 15 Jun 9.1

*1901 Aug 9 Sanriku tsunami 0.6 9 Aug 1.2

*1901 Aug 10e) Sanriku tsunami 0.25 9 Aug 1

1923 Sep 1 Sagami Bay tsunami 12 1 Sep <0. If)

*1933 Mar 3 Great Showa Sanriku tsunami 28.2 2 Mar 2.9

1944 Dec 7 Hankaido tsunami 7.5 7 Dec O.lf }

*1952 Mar 4 (Tokachi Hokkaido) tsunami 3.8 3 Mar 1.2

1952 Sep 26 ~jtn-sho (Izu Is.) Volcano tsunaari 0.3 26 Sep <0. If)

1953 Nov 26 Boso Peninsula tsunami 3 25 Nov <0.1

*1968 'May 16 Tokachi (Hokkaido) tsunami 1 15 May 1.0f )

1973 Jun 18 Hokkaido tsunamrt 1 17 Jun 0:3f }

1975 Jun 10 Hokkaido tsunami l' 10 Jun O.lf }

Notes:

a) Ansei, Me1.1i. and Showa are the nallles of eras. sanriku refers to the three "Riku" provinces of Honshu. Boso is a peninsula farther south on Honshu. Tokaida and Nankaido are regions still farther south on the Honshu coast.

b} From lida et !l .• (1967).

c) From Pararas-Carayannis and Calebaugh. (1977).

d} Calculated from assUNd travel tillle.

e) Not reported as observed in Hawaii by lida !i!l. (1967). Pararas-Carayannis (1969), or Pararas-Carayannis and Calebaugh,(1977). but included because possibly recorded.

f) Maximum amplitude recorded on tide gage.

g) Asterisked events are discussed tn·this report.

3

that were included in the Hawaiian catalog as observed in Hawaii without detail, and the tsunami of 1813 or 1814 whose origin is not known.

The review methodology I have employed in the study is essentially the same as that involved in the local tsunami study (Cox and Morgan, 1977). I have concentrated especially on sources of information in Hawaii, mainly newspapers. However, I have also checked most of the sources of information cited for each of the tsunamis in the Hawailan catalog, and other sources of information to the extent that they were easily available. I have reviewed the the reported runup heights and arrival times of each of the tsunamis and corrected them on the basis of the seeming reliability of the reports. I have ~o estimated additional runup heights where possible from reports of effects.

To complete the study, the records of the Japanese tsunamis for which there is only Hawaiian marigraphic evidence should be checked, the marigraphic evidences of all of the tsunamis should be reviewed, sources of information not easily available should be searched, and a search should be made for possible evidences of Hawaiian observation of other significant Japanese tsunamis that are not known to have been observed here.

The wo"rk to date has demonstrated that there are significant errors in some of the runup heights on the Kona coast of Hawaii now accepted in the National Flood Insurance Program. It seems unlikely that errors of anything like the same significance will be found in the further work contemplated. Hence, in the light of the plans for near-future proposal of the tsunami hazard zone in the National Flood Insurance Program, it seems best to issue the results of the work to date in the form of this preliminary report.

I hope on not only that the report will provide guidance to the National Flood Insurance Program, but that its readers will call to my attention any errors and information that I have overlooked.

4

TSUN AMI OF 1813 OR 1814

According to Ii (1959), a "strange rising of the sea" occurred at Hooken&, South Kona, Hawaii in 1813 or 1814. Ii's report has been discussed at length by Cox and Morgan (1977) who considered the event to represent most probably a tsunami of distant origin, although possibly one of local origin. As will be seen from the pattern of runup heights of known Japanese tsunamis in HawaU, the fact that the "rising" was reported on the Kona coast and not elsewhere would suggest, in the absence of other evidence, that the tsunami of 1813-14 tsunami arrived from Japan it it were not locally generated. However, the historic record of tsunamis in Japan, which probably includes all of those that were significant in the 19th century, does not include any 1813 or 1814.

Hence I consider that the 1813-14 tsunami was probably not of Japanese origin;

5

TSUN AMIS OF OBCEMBER 1854

Tho two II'cult Japan"". tlilunamia of tho AMII ora, lenerated on 23 Oecember 18U, aro roportod 1n tho Catalol of Tlunamll In HawaU (PararU-ClI'lyannia, 19691 Parllr,,"Carayannll and Cal'Mush, 1977) to havo 600n oS,.rna In HawaU. Raoh of these tlunamlll waa .. lIoolated with a Vlry lara' Oll'thquako whOlilolooaUon and mllsnltude II indloated in Table 2.

Tlunlmi

TOklidoc)

Hlnk.ida

No ttl I

Tlbl. 2. Charlctlr11t1es of e.rthquak,s Iccompln1.d bl Greet ~sei tlunamisl

Lltitude LonGitude

137.S' E

135.S' E

M

8.4

8.4

23 Dec 00:00 to 00:15

23 Dec 08:00±

I' From lidl !1 1.1 ... 1967.

b) UMT. Univer'll (Greenwich) mlln time.

c) The lource artl of the Tok,ido tsunami WIS misp.ll.d in lid •• t .1. II Tokk.ido. Through tlPogr.ph1cal misttk. thi5 WI' Iltered to -­Hokkl1do in Plrar •• -ClrI1Inni., 1969. and tht miatake was continued in Plrlrl'.Cirll,nn1. Ina Cllloaugh. 1977.

The .ffeotl of the two oarthquakes and two tsunamis have often been conful,d, oven in Japanese roport., Both tlunamll had very hlrh runupa In Japan, tn. reported maximum for the Tokaldo tJunaml bolnr 10 moteI'I and that for the Nankaldo tllunami being 8 meteI'll. They were rOlpoMible for deltroytnr about 11,000 hOUltll, for wreakinc many .hips and hundreds of small boata, and for over 3,000 doathl.

The orirln 1I'0a of tho Toleaido. tlunami, as mOlt recently estimated by Haton (1916) from Itl Umo. ot arrival at plao.. on the Japan .. e aout, lay mostly in,hore ot the epioenter of the acaompanylnr earthquake and extended about 230 kilometers alone the south coast of BOMhu, .omewhat further north eat than lout~we.t of the epicenter.

The orilin lI'ea of the Nankaldo tlunaml, as e.timated earUer by Hatori (1974, a) from the same kind of information, extended about no kilometers southWelt ot the epioenter ot the aacompanying larthquake, mOltly oft the lOutheat aout of Shikoku.

loth tJunamtl were reoorded on tide gages a.t Astoria, Oreron and at san Pranoisco and San otero, CaUfornla, althoulh laohe (1(158), who ftnt oaUed attention to the marill'aphic recorda, did not realize that they were .eparat, events.

The Tsunami Travel Time chll't for Honolulu (Anon. 1971) (Plr. 1) lune.tI that a t.unami lenerated at the Toleaido earthquake eploenter .howd have reached Honolulu In a 11ttle over 8 hOuri, and ono renerated at the Nankaido urthquake eploenter in a Uttle over 8i hourI, in other words at about 10 p.m, on ZZ Deoember and about 6 a.m. on 23 December, reapeatlvely. The expectable arrival tim .. on the 1.land of Hawaii wotUd have been a halt ot to three quarters of an hour later.

The HawaUan tsunami Catal91 report. for eaoh of the tlunamll merely, "Observed. No details." It oitell, u refereno08 tor the Hawaiian obIervano .. , laohe (1858), and Wilion (1958) tor the Tokaldo tJunamt, and laohe (1856), Milne (1912), and Willon (19&8) for the Nankaldo tsunami. Althourh the oited reportl diloUII or at least mention the tsunamil and rlaocnlze that they were recorded In CaUforn14, nonl of them mentioM any obIervatlons in HawaU, and the t.unamll seem not to have been inoluded In any other 1111t11 ot thOllI oblerved in HawaU (althoulh notl' on the Hawallan obIervanoes were reported In the Paolflc oatalop of lIda !! !l., 1987, and Bolov'ev and Go (1989, an».

120' 140' liII" 180" 160" 140" 120' 1!Ir III'

I·rr~ 'I i "\-~ C,o .. i,. Moun,ain'Gi •• o,. C, •• k "Coll.". ;: I I \ <=:::;:. A lAS I A I II

!, I UNIONOf50YlfTSOCIAIISTIfPUIIICS i.-1 r"')1 {cFJ "P I ,II TSUNAMI.WARNING SYSTEM '.:i " I I ,,---r .1.......1 g. a ... r ~I ~i I ./ . ,:- c'/ I • d' I! rf..xwa!i~ Showing Reporting Stationa and 1 'I' 4~,. ~-.~?:--C----z;:, I,i ~'~ I '" ~M}ddl.::n •• lanc!5~-" ~~' TRnami Travel Time. to Honolulu : .w

S L A / I > B ~ ~ I N ~ S " A I I /~odiak 'J ri'ka ' , i / .' ~ '. Cold lay, 1'-" -', U S DEP E OF CO ERCE I ' . Of I ~!, '" i .... r.:r~",JAI" ... , :'~~.. ARTM NT MM

'H' "'?1"O«H(}TS«~~r._ .,,' ,,'·'~~~'.ka 6~ II""". NATIONAL OCEAN SURVEY < 'E k - , (: ._ .... ," •• ., " A"u~ '~i C~p,.~ ~ I 7., Do .. A, Ja .... , Adin. Di,.dor'

'" 7'-'/1' '-- ill \ ,~~/V! /;" .... Sh.~;al, I·Saroch.f \: 4~ / / ' VJi,o!'a,c __ .~ N ; 0 A I ,I

l. /" Y (I ~+,' .... , I Tofin .!.E-N.wport -~J /.!\. ,,+" .... , I I c....,/'~

r' ", ,/ I ",,'i - .. '';J' , ! I "'/ ~"_... 3.~__ '~.c.,,' City c_f G/ N_ , ...

~

'.; :: AUSTIALI'A \;

REPOITING ~T A TlONS )

• T~ •• totionl Ma,.d~~ 'oi", • Tid •• tation. hayi ... automatic Way. d.tec'.,. ~ r A S.i ... ollroph .tatio... I........' SOU r H

..... ' No'.: Tid •• 'alio.1I lacal.d al Pu,,'a ",."a., Chil. a"d' I I~.I I Pu.,'o Willia ••• Chil •• a"d •• iu,og,aph .,o,ion ailla Plo,a, "r,."'i,,o, "0' .hown,

I \! T ,,',lIS Ii! I ! ( r! j- i" ;I ,- lfO" W.II,""on I.,. 160"' 140" 120' 1011" .,.

PACIFIC

120'

fUJWARY,1971

Figure 1. Chart of Pacific showing tsunami travel times to Honolulu.

7

There is no mention of the tsunamis in the POlrIiesian or in the New Era and Weekly ~ the only English-language newspapers published in Hawai n 1854, and no reference to it has beenToUnd in contemporary Hawaiian language newspapers. Their inclusion in the Hawaiian Catalog may have resulted from a juxtaposition in Wilson (1928) of a discussion of the trans-Pacific effects of the 1854 tsunamis and a discussion of the tsunamis recorded at Hilo on 4 November and 28 December 1927.

The maximum ranges of the waves of the tsunamis as recorded on the Ban Francisco and Ban Diego tide gages were as follows:

San Francf seQ

San Dfego

Tokaido tsunami

0.7 feet

0.40 feet

tlakaido tsunamf

0.45 feet

0.60 feet

Considering these ranges recorded in California, it is quite probable that both tsunamis coul/j easily have been recorded in Hawaii if there had been tide gages in operation here at the time. There is, however, no real evidence that either of the tsunamis was observed in Hawaii.

8

TSUNAMI OF JUNE 1896

Introduction

The great Meiji-era tsunami of 15 June 1896 was generated off the Sanriku coast of Honshu. The associated earthquake, which occurred at 10:33 Universal Mean Time (UMT), is estimated to have had a Richter magnitude of 7.6 and an epicentral location at 39.6 N, 144.2 E (Tokyo Astronomic Observatory, cited in Hda et al., 1967). The tsunami had runups as high as 27 meters on the coast of the northern part of the Rikuzen province. It caused the death of over 27,000 persons and injured more than 9,000 more. Over 10,000 houses were swept away and an additional 2,500 were damaged (nda et al., 1967). . -

The origin area of the tsunami, as estimated by Hatori (1974b) from its arrival times at places on the Japanese coast, was elongated north-northwest to south-southeast and had a length of about 270 km. The earthquake epicenter was on the seaward edge of this area somewhat north of the center.

The tsunami was observed on several Hawaiian islands and was reported to have had high runups along the west coast of the island of Hawaii (nda et al., 1967; Pararas Carayannis, 1?69; Pararas­Carayannis and Calebaugh, 1977). The validity orthe highest runuP values has, however, been questioned.

Contemporary accounts

Accounts of the tsunami on the island of Hawaii were carried by the Hila Tribune, a weekly newspaper whose publication had been initiated Just seven months earlier. More general accounts of the tsunami in Hawaii were carried in five Honolulu newspapers: The Hawaiian Gaze2ifu an evening semi-weekly; the Evening Bulletin, the Hawaiian Star, and the Independent, all evening . es; and the Pacific Commercial Advertiser, a morning daily (the Gazette and the Advertiser had the same editor).

The Bulletin and the Independent carried accounts of the tsunami on Monday, 15 June 1896, the day of its arrival. The other three newspapers carried accounts the next day. Identical accounts in the Gazette and the Advertiser were based on notes by Walter Wall on the marigraphic recording of the tsunami at Honolulu Harbor and included copies of the marigram. The account in the Star carried the same information, but in more abbreviated form, and, in addition, notes that the waves had been observed by bathers at Waikiki and that thousands of fish had been stranded at the mouth of Nuuanu stream during the recessions.

News of the effects of the tsunami on Kauai reached Honolulu in time for publication in the 16 June Star article and the second article in the Bulletin published the same evening. More extensive. identical accounts of these effects were published by the Advertiser and the Gazette the next day. The Kauai news was based principally on a report by Captain Peterson of the inter-island steamer "James Makee" which had been grounded by the tsunami at Kapaa. Peterson also relayed information that had been telephoned from Kilauea and from Nawiliwili. The Advertiser and Gazette accounts earried additional information supplied by the purser of the "Makee".

The Hila Tribune published on Saturday, 20 June, an account of the effects of the tsunami on the . island of HawaIi. News of these effects had reached Honolulu the previous day by way of an inter­island steamer and had been published by the Bulletin, the Star and the In~ndent that evening (19 June). More extensive reports of characteristics and effec~ the tsunam onHawaii, particularly in the Kau and Kana districts, were published by the Advertiser on 20 June and the Gazette on 23 June. The latter accounts, essentially identical to each other, consisted, in the main, of a report by Captain Simersen of the "W.G. Hall", the inter-island steamer, and quotations from letters by J. Kaelemakule, a store owner and government agent in Kailua-Kana, and a Miss Paris of Honolulu, who was visiting in Kailua.

The "Hall" had been at Honuapo when the tsunami arrived, and had been forced to recall the boats that were lightering freight to or from shore and to steam to sea. She returned to Ronuapo when the danger was over, finished the transfer of cargo, and returned to Honolulu, stopping on the way at two or more points including Kailua-Kana and Makena, Maul, where Miss Paris successively embarked and disembarked.

9

The Advertiser and Gazette articles are confusing in that the excerpts from the Kaelemakule and Paris letters, which pertained to Kailua, were inserted between the parts of Simersen's account that pertained to Kawaihae and Kailua and the parts pertaining to Keauhou and other points. The Advertiser account is especially confusing in that it omitted a subtitle that was supplied by the Gazette, distinguishing Simersen's Keauhou information from Paris's Kailua information. However, the first quotations from Simersen were attributed to him at the end of the article with the note that Simersen's remarks were corroborated by other members of the crew of the "Hall", Simersen probably obtained the information pertaining to places not visited by the "Hall" on the return trip to Honolulu by way of the telephone system that then linked Kona, Kau, and Hilo (Star, 19 June; Gazette, 23 June; Hilo Tribune, 27 June).

Some additional information was provided in the Star account of 19 June by a correspondent in Kona. Additional accounts were carried in the 23 July Advertiser on the tsunami effects in Kona, in the 24 June Star on the effects in Hilo, and in the 27 June Tribune on the effects in Kau.

Identical editorials in the Advertiser on 17 June and the Gazette on 19 June expressed the opinion that the waves came from the west or southwest because they had greater force on Kauai than Oahu, but also speculated that the waves might have come from Mexico. The 17 June Advertiser published the opinion of a visitor with volcanologic interests that the waves were generated the leeward part of the Hawaiian archipelago. An editorial in the 20 June Advertiser noted that the tsunami had been observed at Kaanapali, Maui (although no damage seems to have been reported there), but not at Lahaina. The editor commented that Maul would be sheltered by Oahu and other islands from waves from Japan. The opinion that the waves were from Japan was expressed by the Star on 23 June when news of the effects of the tsunami at Santa Cruz, California had been received TIle' opinion was confirmed when news was received from Yokohama concerning the disastrous effects of the tsunami on the Sanriku coast of Japan was published by the Advertiser on 27 June.

Information on the times, sizes, character, and effects of the waves in Hawaii is summarized from the contemporary accounts and collated in Tables 3 and 4. Several of the places in Kailua to which reference was made in the accounts may be identified in an 1892 map of Kailua on file in the State Survey Office (Fig. 2). The wharf and wharf warehouse, the Kaelemakule store, and a courthouse are shown. The Paris Hotel may have occupied either the house owned by Kalakaua that is now known as the Hullhee Palace or that 200 feet northwest of the wharf. The only structures identified with McDougall on the map are a small one seaward of the courthouse that McDougall had transferred to the Minister of the Interior before 1892 and a still smaller one 200 feet north of the wharf. McDougall's store may have been constructed at the site of the latter structure subsequent to 1892 (or elsewhere), McDougall may have taken over the property identified on the map as Kapukui's, or McDougall's store may be the unidentified one seaward of that owned by Akana. The lumber yard is not identified on the map.

There is now a seawall along much of the Kailua waterfront, and there has been a good deal of filling behind it, especially in the several hundred feet east of the wharf. As estimated by William Russell and Elizabeth Cunningham (personal communication), who made a rough survey at Kailua for this study, the original ground level at the building identified in 1892 with Kaelemakule was probably originally only about 4 feet above mean sea leveL At the small building identified with McDougall it was probably about 9 feet msl, at the unidentified store 8treet, and at the courthouse 21l feet.

Subsequent accounts

The occurrence and effects of the tsunami were mentioned in the missionary journal The Friend (Anon., 1896) and in the Hawaiian Annual for 1897 (Thrum, 1897). The occurrence and effects were discussed in some detail in a report of the Imperial Earthquake Investigation Committee of Japan (Anon., 1897). Although not included in a list of Hawaiian tsunamis compiled by Jaggar (1931), the tsunami was included in a list of Hawaiian tsunamis compiled by Shepard (1950), and a list of tsunamis affecting Hilo compiled by Horikawa (1961). Hatori (1963) compared the runup pattern of the 1896 tsunami around the island of Hawaii with the pattern of the 1933 Japanese tsunami and the patterns of tsunamis from other source regions. Times, heights, and other data on the tsunami in Hawaii were tabulated by lida et aL, 1967, Pararas-Carayannis (1969), and Pararas-Carayannis and Calebaugh (1977).

10

Place

~

Kilauea Lridg.

[(apaa Lndg.

rlawn.fwl1f

Oahu

Honolulu

Hawaff

Hawaii t unspec.

[(awafhae

Koha 1 a or leona

ICona. unspec.

Kal1ua

Keauhou

Table 3. Reported times. sizes, and character of waves. tsunami of June 1896 in Hawaii

Runup (+) and Inundation (+) Time ) drawown ( - ) • and

hr:mfn HSrI feet b) recessfon (-). 15 Jun Feature (and dat .. ) feet

-40 07:30t Recession

8t waves +4 (t1?)

-3 (tl) _40d)

+<90

07:38 Arrfval +0.1 (t1)e) 07:45 1st crest

08:00 Trough 08:05 to

+0.2 (tne) 08:33 2nd crest 08:30 Strong current 08:48 Trough

+0.3 (tl )~) 09:00 3rd crest (2 to 4 range)

+15 to +40

O8:00t Arrival 08:15 Arrhal 11:25 Crest

<08:00 Arrhal

08:15 Beg.fnnfng of recessfon

+20± 15 waves

->300

perceptfble, 3rd • largest +40±

08:00t Arrival 08: 1St Crest +20 -1000 08:3Ot Arrival or crest

09:00t Crest Recession

+8± ( .. )

09: 15 Trough . 2 waves follOWing trough were successively larger.

and a later wave stf11 larger <09:35 Wive period

10 mfn. +10, -20 10 waves before

hfghest +20 10 successfvely hfgher crests

4 notable Wives. level ranaining hfXh for 1 hr. after 4th. Another hf!h Wive fn evening'of 1 Jun

rest +18 -$000 08:3Ot Arrival

or crest +35

Referencesa)

Petersen } EB 16 <Jun

PCA 17 Jun. HG 19 Jun

HS 16 Jun

EB 16 Jun

} '.11 EB 15 Jun

} Wall IN 15 Jun

HS 19 Jun

PCA 23 Jun Sflltrsen HT 20 Jun

HT 20 Jun

HS 19 Jun EB 19 Jun

EB 19 Jun

PeA 23 Jun. HS 19 Jun S111trsen !Cael_kule Parfs PeA 23 Jun

Paris

PCA 23 Jun

E8 19 Jun HT 20 Jun

HS 23 Jun IN 19 Jun

Sflltrsen

Table 3. (continued)

PlaGe

Hawaff (cont.)

Kaawaloa

Napoopoo

Hookena

Time hr:mfn HSTa)

15 Jun

09:00±

09:00±

09:30±

Feature

Arrival or crest

Arrhal or crest

Arrival or crest

Runup (+) and drawdown (-) •.

feet ) (and datum)b

+30

+30

+8 (hw) +12g)

Inundation (+) and

recession (-). feet Referencesa)

Sfmersen

Sfmersen

Simersen HS 19 Jun

Kau. unspec. 12 waves perceptfble EB 19 Jun HS 19 Jun

Kaalualu 09:40± Arrival or crest

Honuapo 09:4O± Arrhal or crest

3 large waves

Punaluu 09:40± Arrhal or crest

3 large waves

Hilo 04:00± Arrfval 10:00± Arrival

or crest Wave period

6-7 min. 21 :OO± Crest

Notes:

a)

b)

HST fn 1896 was UMT - 10 hours 30 minutes

tl • tfde level hw • high water

c) References: Parfs Petersen Kaelemakule Simersen Wall EB HG HS HT IN

PCA 20 Juni HG 23 Jun PCA 17 Juni HG 19 Jun PCA 20 Juni HG 23 Jun PCA 20 Juni HG 23 Jun PCA 16 Jun; HG 16 Jun Evening Bulletin Hawaiian Gazette Hawaiian Star

+20

+12

+12 +10±

+12 +10±

+5 to 8

+8±

PCA

Hilo Tribune Inde¥endent P8ci ic Commercial Advertfser

Sfmersen

Sfmersen PCA 23 Jun

Sfmersen PCA 23 Jun

EB 19 Jun

-20 (wafan~enue wharf) Sfmersen

HT 20 Jun HT 20 Jun

d) Reportpf recession at Kapaa may be confused reference to recession reported at Kilauea by Petersen.

e) Marigraphic height

f) Word received in Hilo. presumably by telephone from somewhere in Kana. that "a series of tidal waves were in progress"

g) May not refer to Hookena

11

12

Place

Kauai

Kilauea lndg.

Kapaa

Nawfliwl1i

Oahu ~uanuu estuary

Honolulu Harbor

Waikfki

Maui

Kaanapalf

lahaina

.!1!!!!ll Kawaihae

Kailua

Table 4. Reported effects of waves, tsunami of June 1896 in Hawaii

Effects

Bottolll bared at landing

. Boats stranded at landfng, stealller grounded at anchorage .

Flooding nearly to warehouse

Road and bridge flooded

Waves rushed up Water backed up Fi sh stranded Several waves Boats affected Recorded on tide gage

Observed by bathers

Recession noted

No effects

Wharf deIDlished

Wharf: Dulged

DHDl f shed Warehouse flooded and freight damaged Scott·, Clarke llJllber yard flooded; lllllber

washed to sea and later back on shore McDougall Store:

Flooded Fl ooded to wi ndow sill s 6 .,les hitched to wagon fn vicfnity

swept off thefr feet Kaele.akule Store:

Flooded Flooded to 2It ft. over lanai Suffered a1ADst as II.ICh as McDou9al Store

COurthouse (or courthouse vicinit1?): Flooded Nearly flooded

Paris Hotel (former Kalakaua residence). near11 flooded

Building occupied by Parfs (Paris Hotel?). flooded and walls in vicinit1 broken

Slall bufldings and tanks, Chinese storehouse. and 2 native houses swept away

Wharf and McDougal, Kaelemakule, ICupta. Scott. and Clark properties ~ged

Referencea )

Petersen

EB 16 Jun, Petersen. HS 16 Jun

EB 16 Jun, Petersen

IN 15 Jun EB 15 Jun HS 16 Jun IN 15 Jun EB 15 Jun PtA 20 Jun

HS 16 Jun

PtA 20 Jun

PtA 20 Jun

PeA 20 Jun

Sflllersen, Kaelemakule; HS 19 Jun

PeA 20 Jun Silllersen Silersen, HS 19 Jun,

HT 20 Jun

Kae'e.akule. HT 19 Jun PeA 23 Jun

PeA 23 Jun.

HT 19 Jun Kael_kule PeA 23 Jun

HT 20 Jun EB 19 Jun

HS 19 Jun

Paris

HS 19 Jun

IN 19 Jun

Place,

.l!!.!!!ll (cont.)

Keauhou

lCaawaloa

Napoopoo

Hookena

lCaalualu

Honuapo

Punaluu

Hilo

ICafwflahl1ahf

Note:

Table 4. (continued)

Effects

water entered cave on palf; 9 houses demolished; fish caught in lantana

Beach washed out leaving rocks and debris Chfnese store and 3 or 4 houses demolished;

Post Office shifted; all barrels of oranges awaftfng shfpment swept away

6 houses and Post Office washed away. wharf demolfshed; fish caught fn lantana

3 frame and 3 grass houses washed away At least 3 houses destroyed, wharf not damaged Wharf and 1 house destroyed; warehouse flooded Wharf badly damaged; 2 houses washed away;

lumber scattered Wharf destroyed 3 houses and lumber from lumberyard washed away Landfng wrecked; 2 storehouses damaged Wharf destroyed Wharf, Chfnese store; and bridge demolfshed Wharf damaged, all houses inundated; lumber

scattered Wharf and several houses wrecked; road fnundated

B or 9 feet Scarcely felt Only damage was washing away of small wharf Wharf destroyed Wharf and 4 houses washed away

W.G. Hall forced to halt transfer of cargo. Large tfmber from scow washed into pond.

Wavehouse almost reached warehouse

No damage

Scow alternately stranded and refloated at Wafluku Rfver. Two yachts damaged

No damage Bottom exposed at Wafanuenue wharf Yacht driven under Waiakea bridge; launch

carrfed ashore; boat overturned; scow nearly wrecked at Wailuku Rfver

Wharf and house destroyed

a) See note c). Table 3.

Referencea)

Simersen Sime.rsen

HS 19 Jun

EB 19 Jun HT 20 Jun IN 19 Jun Simersen

HS 19 Jun IN 19 Jun Simersen; HS 19 Jun HS 19 Jun HI 19 Jun Sfmersen

HS 19 Jun

HS 19 Jun EB 19 Jun HT 20 Jun IN 19 Jun Simerien

Sfmersen HS 19 Jun

Simersen

EB 19 Jun Simersen HT 20 Jun

HT 20 Jun, HS 24 Jun HZ 19 Jun

13

14

100,", I

Figure 2. Kailua-Kona in 1892

15

Arrival times

The times reported for the arrival of the tsunami at various places in Hawaii have a range of nearly 2l hours (Table 5). The range is much too great to be accounted for by the differences in travel times from a source off the Sanriku coast. The one arrival time that is well established is that indicated on the Honolulu marigram, 07:38 Hawaiian Standard Time (HST>. In 1896, HST was lOt hours behind UMT. Hence the travel time of the tsunami to Honolulu, assuming it was generated at the time of the earthquake, may be calculated as 7 hours 35 minutes. This agrees closely with the travel time from the earthquake epicenter to Honolulu indicated on the Tsunami Travel Time Chart (Anon., 1971) (Fig. 1).

The travel time of the tsunami to the west coast of Hawaii should have been expected to be a little more than half an hour more than the travel time to Honolulu. The 08:15 HST arrival time at Kawaihae reported by Simersen is equivalent to a travel time of 8 hours 12 minutes, which agrees reasonably well with the expectation. The travel times of the tsunami to Kailua and Keauhou could not have differed by more than a few minutes from that to Kawaihae. The arrival times reported by Simersen and Kaelemakule are reasonably consistent if they referred to the first crest rather than the beginning of a rise. The beginning of a recession at 08:15 reported in a Hawaiian Star article (29 June) is inconsistent with the arrival of a crest at the same time reported in the same article. An arrival as early as 08:00 as reported by the Pacific Commercial Advertiser (23 June) is unlikely, although the Hilo Tribune referred indirectly to an even earlier arrival time.

According to the Tribune (20 June), in Hilo:

Just before 8 a.m •••• word was rapidly passed over town that a series of tidal waves were in progress. Crowds were seen running toward the Wailuku Bridge where a Tribune reporter had alreadY placed himself on duty. The first appearance of a tidal wave was noticed by the unusual heaving and activity in the bar of the river.

Because the "word ... rapidly passed" preceeded by some time the "first appearance," the former must have been information telephoned from some place on the west coast, possibly Kawaihae but probably Kailua. The Tribune report does not suggest a long delay, however-certainly nothing approaching the 1 t or 2 hours implied by Simersen's report of an arrival time at Hilo of 10:00 HST. The arrival of the tsunami at about 04:00 HST, reported in the Bulletin is absurd. From the travel time of the 1933 tsunami to Hilo, 8 hours 35 minutes, it may be judged that the arrival time of the 1896 tsunami in Hilo was about 08:38, which seems consistent with the Tribune report.

The arrival of the tsunami at coastal villages in Kona south of Keauhou could not have been more than a few minutes later than the arrival at ~awaihae, Kailua, and Keauhou. Its arrival at ports on the Kau coast could not have been more than 10 or 15 minutes later. Simersen's reports of arrival times at these places seem quite erroneous unless they refer to the times of arrival of particularly high crests late in the wave train.

The report in the Hawaiian Star of the arrival at Kailua-Kona of another high wave during the evening of 15 June, perhaps 10 to 15 hours after the first arrival, is of interest in the light of a reported observation of high waves at Lahaina, Maui about 15 hours after the first arrival in Hawaii of the 1933 tsunami from Japan. The 1933 tsunami had apparently not been observed earlier at Lahaina (Pararas-Carayannis and Calebaugh, 1977), and, as noted earlier, the 1896 tsunami was apparently not observed there at all.

Runup heights

Kailua-Kona

The 8-foot runup at Kailua-Kona that was reported in recent lists and catalogs (Hatori, 1963; lida et aL, 1967; Pararas-Carayannis, 1969; and Pararas-Carayannis and Calebaugh, 1977) (Table 6) is based On Kaelemakule's account in the Pacific Commercial Advertiser (20 June) and Hawaiian Gazette (23 June). Kaelemakule's estimate was based on "ordinary high water" datum. The account suggests that Kaelemakule and his associates were not aware of any earlier wave than the one that reached that

16

Table 5. Hawaiian arrival times and travel times. tsunami of June 18~

Ti.es. h:m, 15 Junb) Riported

Referencea) Arrival Arrival

Place HST IJMT Travel

Kauat

Kapaa EB 16 Jun 07:301: 18:00 7:271:

Oahu

Honolulu WI 11 • IIIIri gram 07:38 18:08 7:35

Hawatt

Kawaihae PeA 23 Jun 08:00:1: 18:30 7:57 Sf_rsen 18:15 18:45 8:12

!Cohala or Kona HT 20 Jun <08:00 <18:30 <7:57

{ PCA 23 J ... 08:00:1: Katlua HS 19 Jun <08:15 <18:45 <8:12

St .. rsen • Kae18llllkule <08:301: <19:001: <8:271:

Keauhou Sf .. rsen 08:301: 19:001: 8:271:

Napoopoo Sf .. rsen 09:00:1: 19:301: 8:571

Kaawaloa Stallrsen 09:00:1: 19:301: 8:571:

Hookena St .. rsen 09:301: 20:001: 9:271:

Kaalualu S1_rsen 09:4Ot 20: lOt 9:371:

Honuapo Sf .. rsen 09.:4G.t 20: lOt 9:371:

Punaluu Sf .. rsen 09:4G.t 20: lOt 9:371: r 19J ... 04:001: 14:30t 3:571:

HT 20 Jun >08:00 >18:30 >7:57 Hflo Si_rsen 10:00:1: 20:301: 9:571:

Pararas-tarayannis • talebaugh 7:54

Notes:

a) See note c). Table 3 for identification of newspaper references cited. 1933 travel ti .. s reported in Pararas-tarayannis (1969) and Pararas-tarayannis and Calebaugh (1977) are identical.

b) HST • Hawa11an Standard ri .. in 1896 was UMT - 10 hours 30 minutes, 1896 tsunami travel ti .. s based on earthquake tt .. , 10:33 UMT 15 June 1896.

17

height and suggests that no later wave reached higher, or at least much higher, at his store. However, Paris's account in the same newspapers indicates that at least three waves reached successively greater heights in Kailua.

Although the church at Kailua was constructed of stone, the buildings at which tsunami effects were reported were propably timber structures on posts with at least crawl spaces and generally several feet below the floors.

The approximately 8-foot runup height above ordinary high water reported by Kaelemllkule (9 ft± msl.), would have covered the lanai of Kaelemakule's store (if it were the building identified with him on Fig. 2) to the reported depth of 2t feet if the lanai floor were only 2t feet above ground level, and would have flooded the store itself to a depth of half a foot if the floor of the store were 4t feet above ground. The above estimates are reasonable, although usually the floors of lanais were at the same level or no more than a toot below the levels of the buildings to which they were attached. If the 10-foot runup reported by the Advertiser had occurred at that site, the floor levels would have had to have been higher above ground or the building would probably have been shifted, at least, by the waves. The 10-toot runup estimate would certainly be compatible with the eUects reported elsewhere.

Inundation even on the ground as low as the building identified as a courthouse in 1892, as reported by the Bulletin. would be consistent only with the 20-toot runup height reported in the Star. and the report that the last of ten waves "came up and invaded the sanctum sanctorium ot the Kailua courthouse" implies a runup even higher. The 10-toot report in the Bulletin seems to have been drawn from the same sources that led to the 8 foot report in other newspapers and hence seems to be an exaggeration. It is possible that the waves merely had a significant inundation seaward of the building identified as a courthouse in 1892, or that in 1896 the courthouse was at some place with lower ground elevation.

The runup height of 30 feet listed in parenthesis by Hatori as an alternate to the 8-feet seems not to be based on any contemporary account and out of line with all ot the reports of the effects of the waves.

Considering the possibility that the runup of the highest wave may have been slightly higher somewhere in Kailua than the runup reported by Kaelemakule, it would appear that the runup at Kailua was somewhere in the range from 9 to 11 teet above mean sea level.

Keauhou, Kaawaloa, and Napoopoo

The most extensive record of runup heights is that pertaining to Keauhou, Kaawaloa, Napoopoo, and Hookena on the Kona coast; to Kaalualu, Honuapo, and Punaluu on the Kau coast; and to Hilo; that was provided in Simersen's accounts in the Pacific Commercial Advertiser (20 June) and Hawaiian Gazette (23 June). According to the Hawaiian Star (19 June), Simersen reported runups ranging 15 to 40 teet, but as will be seen from the column of Simersen's estimates in Table 6, they actually ranged from 8 to 35 feet. Simersen's runup estimates for places from Kaawaloa counter-clockwise to Punaluu have been included in subsequent tabulations, but somehow his estimate ot the runup at Keauhou, 35 teet, was reduced to 30 feet in the report of the Japanese Imperial Earthquake Investigation Committee (Anon., 1897) and in later accounts.

It seems safely assumed that Simersen's estimate ot the runup at Honuapo, where the "Hall" was transferring freight at the time of the tsunami arrival, was reasonably accurate, even though the "Hall" put to sea for the duration ot the high waves. It may also be assumed that his estimates of the runup at other ports visited by the "Hall" on her return trip to Honolulu were reliable. Unfortunately there is no way of knowing which ports these were, other than Kailua.

The only information provided by Simersen that can be used to check his runup estimates is his report that, at Keauhou "the water rose about thirty-five feet entering the cave on the side of the pall." There is a pall (cliff) paralleling the general coast near the shore at the head of Keauhou Bay. Inland of the southern edge of the bay, the top of the pall is about 50 feet above sea level, but it's height decreases both to the north and to the south. There is now but one cave notable on this cliff, and that is at its base. The floor of this cave is about 10 feet and its root about 15 feet above sea level. If this is the cave referred to by Simersen, its inundation would not seem at all noteworthy in reporting waves with a runup height ot 35 teet.

18

Table 6. Hawaiian runup heights. tsunami of June 1896

Ree2rted heights I ft. News2!~r accountsa' SubSequen~)

Accepted hed,ht. DatumS)

feet 1151. Place Sil11ersen Others Reference accounts Min. Prob. Max.

!!!!!! Hanalei 9

Kilauea Lndg. 3 5 a

Kapaa (3 drawdown) tl PeA 16 Jun 4 E8 16 Jun 3 4 5

Nawl1iw111 tl 3 5 a

Koloa Lndg. 5

Hanapepe 6

Wa1 .. a a

oahu

Wl1k1k1 1 2 4

Honolulu Hbr. 0.3e) tl 0.3 1 1" 2 (2 to 4, range) IN 15 Jun

Mau1

Makena 4

Lahaina 4

Kaanapali 2 6

Kahului 4

Hawaii

Unspec1 f1 ed 15 to 40 HS 19 Jun

Laupahoehoe 10

Ka1wl1ahl1ah1 a 10 13

Hakalau a

H110 8± 5toa EB 19 Jun a 6 a 10

Kalapana 10

Table 6. (continued)

Newspaper accountsa) Si .. rsen Others oatumbl Reference

SUbsequent) accounts

Accepted hed,ht. feet 1IIS1.

Place Min. Prob. Max.

Hawaif (cant.)

Kau. unspec.

Punaluu

Honuapo

Kaal ualu

KDna l unspec.

Hookena

Kaawaloa

Napoopoo

Keauhou

Kailua

Kawaihae

No~s:

12

12

12

8

30

30

35

20

101:

101:

40

127

8 20

81:

hw

hw

HS 19 Jun

HS 19 Jun

HS 19 Jun

EB 19 Jun

HS 19 Jun

I 15 Jun HS 19 Jun

12

12

12

8 to 30

8

30

30

30

8.

a) See note C)I Table 3 for identifications of newspaper references cited.

b) Datum: tl • tide level hw • high water msl • .. an sea level

10

10

10

9

10

10

11

9

6·

12

12

12

11

17Jt

17'1

19'1

10

8'1

c) Identica' hefghts were tabulated by Anon. (1897). Hatorf (1963) l lida et a1. (1967). Pararas­Carayannis (1969). and Pararas-Carayannis and Calebaugh (1977} except'that:

1. Ifda !1!JL ... rely summarized the heights on the KGna coast. 2. Hatori indicated in parenthesis an additional Kailua height report of 30 ft •• possibly a

misprint for the (probably erroneous) 20 ft .• height reported in the Hawaiian Star. 3. The heights reported in the last three accounts were converted to .. ters.

d) ~n. • Lower limit of range of reasonable values. Prob •• "Most probable" value. Max. • Upper limit of range of reasonable values.

e) The "height" indicated at Honolulu was the marigraph1c amplitude.

14

14

14

14

30

30

35

11

12

19

20

About 500 feet north of the head of Keauhou Bay there are two more caves representing collapses into lava tubes (Elizabeth Cunningham and William Russell, personal communication). 'Ibey are only about 6 feet long, and they are now partly fiDed with rubble so as to be only about 4 feet deep. 'Ibe ground in their vicinity has an altitude of about 24 feet, but they are not caves in a palL In a report by Loomis (1975) on the November 1975 locally generated tsunami, a runup height is referred to an "empty lot right side of cave", but this is clearly a typographical error. 'Ibe site is on the right side of the cove (Keauhou Bay) looking seaward, and there is no cave at the site. It is possible that there was a higher cave on the cliff at the head of the bay, about 50 feet south of the cave at its base, and that this higher cave has since colJapsed. The bedrock at the base of the pall to the south of the cave at its base, aboVe the point now marked as the birthplace of King Kamehameha In, is concealed by large blocka of talus. However, I have found no one who remembers a cave there and no evidence of its existence. .

A report that, at Keauhou, "all the orange barrels belonging to the orange groves were carried away" (Star. 19 June), suggests that the interisland steamers called there. However, there is no direct evidence that the "Hall" stopped at Keauhou on her trip from Honuapo to Honolulu.. Hence it cannot be ascertained whether Simersen actually saw the runup evidence there. He recosmzed that the runup height he reported there was unusual, commenting: "'Ibe cause of the great height of thirty feet is easily explained when it is remembered that Keauhou is in a very narrow bay." Still, considerable doubt must attach to his estimate, and it can be regarded only as an upper limit of the possible range of error.

To judge from the reference to the inundation of the cave, if it is assumed that the cave was that at the base of the cliff, the runup might be estimated at 11 feet msL To judge from the damage caused by the tsunami at Keauhou and the destruction of the beach, the runup would be estimated somewhat higher.

Fortunately, the Simersen's estimate published in the Advertiser and Gazette is not the only estimate of the runup height at Keauhou. 'Ibe Independent on June 19 quoted a communication addressed to the Makaainana (the House of Commons);earrteC:fby the "Hall," that described the effect at Keauhou. According to the communication: "'Ibe sea rose about 18 feet and then receeded with terrible swiftness, laying bare the bed of the ocean for about a mile." 'Ibe houses destroyed were identified as those of: "J.N. Koomoa, A.K. HoapiU, Aid (Chinese), C. Kaiaild, Manuiwa, Keawi, Sam Keawe, and Kahu." 'Ibe Kaiaiki house was described as a two-story bui.ldl.ng that was carried to sea. 'Ibe communication indicated that no damage was done to the wharf, and attributed this fact to the more substantial construction of the wharf than others on the coast.

Although the bed of the ocean could not possibly have been layed bare for more than a very small fraction of the one-mUe recession reported in the communication, the IS-foot runup height estimate seems more reliable than Simersen's 35-foot estimate or the 11-foot minimum estimate.

'Ibe fact that Simersen's estimate of the runup at Keauhou is doubtful raises questions also about his estimates of 3O-foot runup heights at Kaawaloa and at Napoopoo, which have been relied on in subsequent accounts. 'Ibe "Hall" might have put in to Kealakekua Bay on her return trip, but there is no direct evidence that she did so. 'Ibe effects of the waves reported at Kaawaloa and Napoopoo might easily have resulted from much smaller waves, and at Napoopoo, at least, much greater damage should have been expected with a runup as high as 30 feet. 'Ibe runup at both places is here estimated only as between 10 and 30 feet.

Hookena and places to the south on the Kona coast

The runup of 8 feet above high water estimated by Simersen at Hookena and reported in· recent lists and catalogs could have resulted in the washing away of the small wharf which was the only . damage there according to the HUo Tribune (20 June). The Bulletin reported that the tsunami was scarcely felt at Hookena. However, Simersen and the Hawaiian Star (19 June) reported more extensive damage that could probably not have occurred unless the runup height were greater than 8 feet above high water (9 feet above mean sea level). The Star reported that "A Mr. Davis who was in a wagon at the time of the wave, had an escape which seems almost miraculous, the road being eight or nine feet under water". This report suggests a runup even higher than the estimate of "twelve feet above high water mark" seeming to relate in the Star article to Hookena. 'Ibere is, thus a significant range of uncertainty as to the runup height at Hookena, perhaps 9 to 14 feet above mean sea leveL

21

It is not surprising that there were no reports of the effects of the waves at villages on the Kona coast south of Hookena, because communications with these villages were poor. However, there would probably have been reports of some damage at Hoopuloa and MiloUi if the runups at those villages had exceeded 10 feet. Hence it is unlikely that the runups at those villages were that high.

KaaIualu, Honuapo, and Punaluu

The runup heights indicated in the recent lists and catalogs for the Kau villages of KaaIualu, Honuapo and Punaluu are those reported by Simersen (e.g. Hawaiian Gazette, 23 June), 12 feet at each place. The local tsunami of 1868 is estimated to have had a runup height of about 16 feet at Kaalualu (Cox and Morgan, 1977). Hence the damage reported there in 1896 would not be expectable unless the wharf and houses that were wrecked had been constructed during the 29 year period between the two tsunamis. The 1868 local tsunami runup heights were also considerably higher than the 1896 runup heights at Honuapo and Punaluu where the later tsunami resulted in little or no damage. There seems little reason, then to question the .Simersen estimates of Kau runup heights, especially as Simersen and the "Hall" were at Honuapo when the tsunami arrived and put into Honuapo again before leaving for Honolulu.

The runup heights of about 10 feet at Honuapo and Punaluu reported in the Hawaiian Star (19 June) do not differ materially from Simersen's values. The greater height of 20 feet, also reported in the Star at some Kau locality not clearly identified but apparently Honuapo, is clearly inconsistent and may "be assumed erroneous. I consider the runup heights at Kaalualu, Honuapo, and Punaluu, as to have been between 10 and 14 feet msL

Other Hawaii Sites

Considering the lack of any reported observation at Kalapana or Kaimu or the Kau coast, it seems that the tsunami could not have had a runup of more than 10 feet there.

The recent lists and catalogs show for Hilo the 8-foot runup height reported by Simersen. Although the "Hall" clearly did not visit Hilo on the return trip from Kau to Honolulu, and Simersen must have been informed of the Hilo runup by telephone, no other estimate of the Hilo height is available, and considering the effects of the waves at Hilo, a runup between 6 and 10 feet is reasonable.

The destruction of a wharf and a house at Kaiwilahilahi, a landing about 2 miles southeast of Laupahoehoe peninsula near Papaaloa, suggests a runup height between 8 and 13 feet there. However, it seems very doubtful that the runup of the tsunami could have been as great as 8 feet at Hakalau or 10 feet at Laupahoehoe without their observation at those places being reported in the newspapers.

Considering the tsunami effects at Kawaihae reported by Simersen, the runup there may be estimated at between 6 and 12 feet.

Sites on Maui

Although the "W.G. Hall" put into Makena, Maui, on her return to Honolulu from Hawaii, Capt. Simersen did not include any report of the effects of the tsunami there in the account he gave to the Honolulu newspapers (Pacific Commercial Advertiser, 20 June; Hawaiian Gazette, 23 June). It is probable either that residents of Makena did not observe the tsunami or that Simersen considered their reports inconsequential compared with those from Hawaii, and in any case that there was no damage at Makena. It seems unlikely that the runup at Makena can have been greater than 4 feet. It is also unlikely that the tsunami hErd a runup greater than 4 feet at Lahaina or Kahului considering the lack of any reports of its observation at that places.

The Pacific Commercial Advertiser (20 June) reported that the waves were not observed at Lahaina, and it is probable that their runup there also was no more than 4 feet. However, the Advertiser reported that at Kaanapali recession was noted. Although the newspaper did not mention any inundation, it seems likely that the recession would be accompanied by a runup of between 2 and 5 feet.

/

22

Sites on Oahu

From the fact that the tsunami was observed at Waikiki by bathers (Hawaiian Star, 16 June), it seems probable that the runup there was between 1 and 4 feet.

The stranding of fish in the Nuuanu estuary (Hawaiian Star. 16 June) suggests that there must have been a runup of at least a foot in Honolulu Harbor. However, the fact that its marigraphic amplitude was only 0.3 feet suggests that the maximum runup was not greater than about 2 feet. A I­to 2-foot runup is eonsistent with the 2- to 4-foot estimates reported in the Independent if the latter are considered to refer to trough-to-crest ranges rather than runup heights.

Sites on Kauai

The effects of the tsunami at Kilauea Landing and at Nawiliwili reported by Petersen (Pacific Commercial Advertiser. 17 June) suggest runups between 3 to 8 feet at both those places. The i-foot runup reported by the Evening Bulletin at Kapaa seems reasonable considering the descriptions of the waves there.

Summary

Table 6 summarizes the ranges in runup heights that were reported for the tsunami from place to place, estimated values based on the described effects, and most probable values representing, with three exceptions, the geometric means of the reasonable ~tremes. The reasons for selecting the geometric means as the most probable value are discussed in a section on the "Ranges of possible error and most probable values" in the summary of the historical runup record of all of the Japanese tsunamis. The exceptions apply to the most probable values of the runup at Keauhou, Kaawaloa, and Napoopoo.

If there were no other estimates than the maximum and minimum at Keauhou, a value approximating either the maximum, 35 feet or the minimum, 11 feet, would seem to be more probable than the geometric mean, 19i feet, the choice depending on whether or not there was in 1896 another wave higher than the one at the base of the pall. However, in the absence of other estimates, it could not be determined whether the greater probability attached to the upper end or the lower end of the reasonable range, and for hazard management purposes use of the geometric mean would seem most appropriate at Keauhou as well as other sites. The geometric mean of the range of reasonable values at Keauhou, 11 to 35 feet, would be 19i feet. However, because the 18-foot runup estimate at Keauhou provided by the IndiOindent. is very little lower than the 19i-foot geometric mean, this estimate has been substituted or the geometric mean at Keauhou.

At both Kaawaloa and Napoopoo, the upper limits of the ranges of reasonable runup values are Simersen's estimates of 30 feet, and the lower limits are the minimum values estimated from the damages reported, 10 feet. The probable values adopted at both these places are 17 feet, as at Keauhou slightly lower than the geometric means of the limit, 17i feet.

23

TSUNAMIS OF AUGUST 1901

Introduction

Two tsunamis were generated in August 1901 off the Sanriku coast, one during the evening of 9 August, the other early the next morning. Neither had significant effects in Japan. Information pertinent to the tsunamis as recorded in Japan and to the two accompanying earthquakes is presented in Table 7.

Time Japan UHf. date 9 June

9 Aug 09:24

10 Aug 18:34

Table 7. Data on earthquakes and tsunamis of August 1901 in Japan

Earthquakes

Latbtude Longitude Richter magnitude

40.5 N 141.5 E 7.7

40.5 N 141.5 E 7.8

Tsunamis

Maximum amplitude. m. Place

0.6 Miyako

0.25 Ayukawa

As estimated by Hatori (1975) from arrival times of the tsunamis at places on the Japanese coast, the source areas of both of the tsunamis lay about 100 km off the coast at Hachinohe, and each was elongated northwest-southeast, having a length of about 100 km. The origin areas of the two tsunami overlapped, but that of the second lay a little farther northwest than that of the first.

Tsunami waves were observed on 9 August on the west coast of the island of Hawaii and were recorded by the tide gage at Honolulu.

Contemporary accounts

Accounts of the Hawaiian evidence and effects of the August 1901 tsunami were carried in the same Hawaiian newspapers that had reported the effects of the 1896 Sanriku tsunami.

The first account, published by the Evening Bulletin on 9 August, the day of the arrival of the tsunami in Hawaii, was only a brief note based on a wireless telegraph dispatch that a tidal wave had arrived at Kailua-Kona, that no damage was yet reported, but that the water was still rising. The Pacific Commercial Advertiser carried a more extensive account of the Kailua occurrence the following morning.

The Star of 10 August carried an account based on the recording of the tsunami on the Honolulu tide gage and on a report telephoned from Kailua to Hilo by telephone and carried by ship from there to Honolulu.

More complete accounts of the tsunami and its effects on Hawaii were published on the Bulletin and the Star on 13 August, the Star on 14 August, the Advertiser on 14 and 19 August, the Hawaiian Gazette on 16 and 20 August, and the Hilo Tribune on 16 August. A second Hilo paper, the Hawaii Herald apparently did not cover the event.

Professor A.B. Lyons was quoted in the Star on 10 August as initially considering a local source possible, and according the Bulletin residents on Hawaii feared the waves might be precursors to some local volcanic or seismic disaster. However, Lyons later considered a source to the west more likely on the basis of the marigraphic record and the fact that the waves had their greatest effects at Kailua and no reported effects at Hilo. In the same issue the Star considered an Asiatic source most probable. Dr. Sereno Bishop, quoted in the Advertiser on 10 August, the Advertiser of 12 August, and· the Gazette of 13 August, considered a source in either Japan and South America as possible. However, the Tribune of 16 August questioned whether the waves were indeed those of a tsunami, noting the occurrences of unusually high astronomic tides at the time, and the Advertiser of 19 August and the Gazette of 20 August echoed the doubt.

Table 8. Reported times, sizes, character, and effects of tsunami waves of August 1901 in Hawaii

Time Runup (+) or HST drawdown (-) Inundation

Place 9 Aug Features feet, (and datum) feet Effects

oahu

Honolulu 11 :30 Arrival or 1st crest +0.2 ~t:li:~ Recorded on tide gage 21:30-23:30 3 larger waves +0.4 tl RecO,rded on tide gage

Hawaii

Kohala No evidence of 9 Aug waves

Mahukona Inundated house on 10 Aug

Kailua 11:00 Arrival or 1st crest +4 (ground?) Inundated back verandah. Hackfe1d store

2 ft. over wharf, Inundated front porch Hackfe1d store

+75 Several houses inundated;

3 waves +25 one washed away

3 waves 3 ft. over wharf. Inundation to Hackfeld back verandah

Keauhou Arrival 3 or 4 minutes after that at Kailua; one hut washed away

One hut demolished

Napoopoo Arrival after that at Keahou. before that at Hookena

t:tookena Arrival 15 adn. after that at Kailua

Kau No evidence

References b)

HS 10 Aug HS 10 Aug

HS 13 Aug, PCA 14 Aug. HG 16 Aug

HS 14 Aug. HG 16 Aug

PeA 10 Aug

EB 13 Aug

HS 13 Aug peA 14 Aug.

HG 16 Aug PeA 19 Aug.

HG 20 Aug

PeA 10 Aug EB 13 Aug

HS 13 Aug. PeA 14 Aug. HG 16 Aug

PeA 14 Aug. HG 16 Aug

EB 13 Aug HS 13 Aug

IS) ....

Place

HawaU (cont.)

Hilo

Kukuihaele

buai

Notes:

Ti_ lIST

9 Aug Features

a) Mlr1graph1c height. datlll presllllbly tide level.

b) References: E8. Evening Bulletin HS • HawaUan Star

PeA • Pacific CoaBercial Advertiser HG • HawaUan Gazette

Table 8. (continued)

Runup (+) or drat«kMI (.)

feet. (and datua) Inunclltion

feet Effects

No evidence

Ship at anchor swung around on 11 Aug

No evidence

Referencesb)

PeA 10 Aug

HS 14 Aug

lIS 14 Aug

to.:! (JI

26

When the first ship to Honolulu left Yokohama after the occurrence of the earthquake and tsunami, the importance of the event had apparently not been recognized at Yokohama, because the Honolulu newspapers did not pick up any news of the tsunami source after it arrived on 19 August. By the time the next ship from Japan arrived on 27 August, the event seems to have been forgotten in the Honolulu papers.

Information on the times, sizes, character and effect of the waves are summarized rom the contemporary accounts and collated in Table 8.

Subsequent accounts

The tsunami was listed among those occurring in Hawaii by Shepard et aL 1950. The Hawaiian evidences were coupled in the preliminary Pacific tsunami catalog (lida, .!! aL, 1967) with the 9 August tsunami recorded in Japan, and this coupling was assumed in the Hawaiian tsunami Catalog (Pararas­Carayannis, 1969; Pararas-Carayannis and Calebaugh, 1977). However, the Hawaiian evidences were coupled by Watanabe (1968) with the 10 August tsunami in Japan.

Arrival times

The tsunami, or perhaps the first crest, was reported to have arrived at Kailua-Kona on 9 August at 11:00 HST (21:30 UMT) (Evening Bulletin, 13 August 1901). The arrival of the tsunami, or perhaps the arrival of the first crest, at Honolulu was reported to have been recorded on the Honolulu tide gage at 11:30 HST (22:00 GMT) (Hawaiian Star, 10 August 1901). Judging from the difference between the arrival times at Honolulu and Kailua of the June 1896 tsunami from Japan, the arrival at Honolulu should have been about 40 minutes earlier than the arrival at Kailua rather than 30 minutes later. It seems probable, therefore, that the first wave identified on the Honolulu marigram was a later wave than the first observed at Kailua.

The Bulletin (13 August) reported that the tsunami, or perhaps the first crest, arrived at Kailua­Kona at 11:00 HST. The Advertiser (14 August) reported that it arrived at Keauhou 3 or 4 minutes after its arrival at Kailua, and the Star (13 August) and Advertiser (14 August) that it arrived successively at Keauhou, Napoopoo, and Hookena--at the later place 15 minutes after the arrival at Kailua. These reported intervals are more or less consistent with expectable differences in the travel times to these places.

A puzzling problem is presented by the travel times of the waves from Japan. These are computed in Table 9 assuming that the waves observed in Kona and detected on the Honolulu marigram were, alternatively, those of the earlier and the later of the two 9 August tsunamis recorded in Japan. It will be seen that, if it is assumed that the first detected waves in Hawaii represented the earlier tsunami, their reported arrival at Kailua was nearly 4 hours later than expectable from the 1896 travel time and the time of their detected effects on the Honolulu marigram about 5 hours later than expectable. If it is assumed that the later, larger waves recorded at Honolulu represented the later tsunami, their detection was nearly 6 hours later than expectable.

The alternative assumption as to the identity of the first detected waves, that they represented the later tsunami, is untenable because the implied travel times, about 3 hours to Kailua and 3t hours to Honolulu, would be much shorter than possible.

As will be seen later, the travel times of the 1896 tsunami to Honolulu and Kailua are consistent with the travel times of other Japanese tsunamis to those places. It is thus the travel times of the 1901 tsunami that are anomalous. The possibility that the source of the waves was not where it has been estimated and the possibility that their generation was not synchronous with the either of the earthquakes should be entertained.

There are hints of the second possibility in the Japanese records. The most recent estimates by Hatori (1975) of the seaward limitl of the source areas of the two tsunamis were based on post­earthquake arrival-time estimates of 54 and 58 miButes, respectively, at a tide gage at Ayukawa, Miyagi Prefecture, Honshu, and 56 and 62 minutes, r~pectively, at a tide gage at Hanasaki, Hokkaido. Previously, Hatori (1974) had estimated that the source areas of both tsunamis lay immediately along the shoreline between Miyako and Hachinohe, 'Honshu, on the basis of an estimate of 84 minutes for the Ayukawa travel time for one of the tsunamis. Reproductions of the Ayukawa marigrams (Imamura and

Table 9. Hawaiian arrival times and travel times of tsunamis of August 1901

Place observed or recorded

oahu

Honolulu First detected waves Later larger waves

~

Kailua

Hookena

Note:

Reported arrival time HST UMT

9 Aug. lli30 9 Aug, 22:00 9 Aug, 21:30 10 Aug, 08:00

9 Aug, 11:00 9 Aug, 21:30

9 Aug, 11:15 9 Aug, 21:45

Calculated ~rAyel time hr:mln

(1) (2)

12:36 22:30

12:06

12:21

03:26 13:26

02:56

03:21

a) Alt.emate calculated travel times are based. on the assumption that the tsunami originated at the times of the earthquakes, on 9 August 1901 at. respectively: (1) 09:24; (2) 18:34.

27

Moriya, 1939) indicate, in the train of waves after each of the earthquakes, about 1.7 hours after the first arrivals used by Hatori in his later estimation, the arrival of a set of waves having amplitudes as great or greater than those of the first arrival, and a longer period. Another brief increase in amplitude (but not period) occurred about 4 hours after the initial arrival of the first wave train, but no corresponding increase is evident in the second wave train. The possibility of repeated disturbances of the sea fioor in the source areas following each of the earthquakes thus cannot be eliminated. Waves originating at the time of the second repetition following the first earthquake would arrive in Hawaii at about the arrival time reported at Kailua and only an hour earlier than the arrival time "­reported at Honolulu.

No evidence seems to support the alternative of a very different source area for the waves. No other large North Pacific earthquakes occurring on 9 or 10 August 1901 seem to haV'e been reported. Waves generated at the time of the first of the Japanese earthquakes that would have arrived at the times indicated in Japan and Hawaii would have had to be generated west of the Japanese islands and should have been obvious on their west coasts.

It is therefore assumed here that the tsunami waves observed and recorded in Hawaii were assOCiated, at least indirectly with the first of the Japanese earthquakes, that of 9 August.

Runup heights

The only runup heights on the island of Hawaii indicated in contemporary accounts pertain to Kailua. The runup height at Kailua reported in the Hawaiian tsunami catalog (Pararas-Carayannis and Calebaugh, 1977), equivalent to 4 feet, is probably derived from the Pacific Commercial Advertiser account (10 August). This height was probably estimated above ground level, because the effects reported could not have resulted from a runup only 4 feet above sea level. The waves were reported to have covered the wharf by 2 feet according to the Evening Bulletin (13 August) or 3 feet according the Advertiser (19 August).

28

The inundation was reported to have reached the front of the Hackfeld store near the waterfront according to the Bulletin (13 August), reached the back of the store according to the Advertiser (19 August), or covered the back verandah according to the Advertiser (10 August). These conructing reports are paralleled by the reports that Kailua was inundated to a distance of 25 yards by the Hawaiian Star (13 August) and 25 feet by the Advertiser (14 August).

COnsidering the Conflicting reports and the terrain, the runup at Kailua seems to have been somewhere in the range between 5 and 10 feet msl.

The destruction of one hut at Keauhou suggests a range of possible runup values there between 6 and 12 feet.

The reports of arrival times of the tsunami relative to the arrival time at Kailua, if they may be trusted, indicate that the tsunami was observed at Napoopoo and Hoopuloa, and that it had runups of at least Ii feet at Napoopoo and 2 feet at Hoopuloa. The fact that no effects were reported suggests that the runups were not more than 7 feet at either place.

There is no evidence of even observations of the tsunami along the Kau coast or at Hilo, other than a questionable Hilo runup height equivalent to 4i feet listed in the Hawaiian tsunami catalog (Pararas-Carayannis and Calebaugh, 1977) for which no contemporary confirmation has been found. The lack of evidence suggests that the runup height was not more than 5 feet at Honuapo and 3 feet at Hilo.

The report that a house at Mahukona was inundated and that the tsunami was observed elsewhere on the west coast of Kohala is suspect. The report was based on an account carried to Honolulu by the master of the "Ke Au Hou," a vessel which had been moored at Kukuihaele on the Hamakua coast. It was coupled with a report of strong currents at Kukuihaele on 11 August. Kukuihaele is an exposed landing on a cliff coast. Bven high tsunami waves would probably not have cause strong currents there, and it is unlikely that the waves of the tsunami would have been high there two days after its arrival, or even at Mahukona the day after the arrival. It is possible, however, that the report was garbled, that the reported effect at Mahukona was on 9 August and that the tsunami had runup on the order of 5 feet at Mahukona.

The 2-inch marigraphic amplitude of the first wave of the tsunami identified on the Honolulu marigram and 3- or 4-inch amplitude of waves arriving 10 to 12 hours later, reported in the Star (10 August), suggest that there may have been a runup somewhere in Honolulu harbor in the range between Iuilf a foot and 1 i feet.

No observations of the tsunami were reported on Kauai, a fact that was specifically noted by the Star (14 August). Hence, there were probably no significant runups along the Kauai coast.

The runup heights estimated above are summarized in Table 10.

Place

Oahu

Honolulu

Hawatt

Makukona

Kailua-lCona

Keauhou

Napoopoo

Hoopuloa

Honuapo

Hflo

Note:

Table 10. Hawaiian runup heights of the tsunami of August 1901

Relnrted het ght Referencesa- Feet Datum

HS 0.4 (IIIar1frapht c alq) ttude)

PeA 4 ground EB 2 wharf PeA :c)

wharf

4?b)

a) Newspaper references: EB· Eventng Bulletin HS • Hawaitan Star

Accepted hetghts,b) feet 1IS1.

A~n. Prob. Max.

1 1~

51

5 7 10

6 8Ii 12

lit 3 7

2 4 7

5

3

PeA • Pactftc Commercial Advertiser

b) Min. • Lower limit of range of reasonable values Prob. • MOst probable value Max. • Upper limtt of range of reasonable values

c) Reduced from metric values reported by Pararas-Carayannts and Calebaugh (1977)

29

Table 11. Reported times, characteristics, and effects, tsunami of 4 ~rch.1933 in Hawaii Co) CI

Runup (+), d~awdown l-)' Inundation (-),

Time ) or range R) or recession (+) Referencesb) Place (hranin)a (feet) (feet) Features, effects, notes

Midway

11:05 to 13:00 Train of waves Jaggar, 1933

Kauai

Arrival times (2 Mar) not reported but considered by newspaper to be too early to represent Japan tsunami

Naw111wl1i Sa~ 1 eft anchorage

~kuiula Sa.ns forced to leave harbor ~ SI 7 Mar

lawai Beach eroded by waves and stream flood

Pakala SaqJIns forced to leave anchorage

Oahu

Honolulu 14:40 Arrhal recorded by tide gage HSB 3 Mar, Jaggar 1933

15:00 to 19:00 -0.6 to +0.8 20 waves } HSB 3 Mar -0.5 First trouoh

+0.7 First crest

Beaches (unspec.) 15:00 Exceptionally low tide followed rapidly' by exceptionally high tide HA 3 Mar

Maui

lahaina (2 Mar) 3 small tidal waves HSB 3 Mar (3 Mar) 3 miniature tidal waves:

} AA'~, 05:00 -35 to -45 First 05:30 Second

These disturbances also reported to last 10 minutes apiece

Kaanapali (3 Mar) Shipping activities cancelled due to "high afternoon seas rurining in the wake of a ICona stonn" MN 6 Mar

Hawaif

Hilo

Kaalualu

West coast (unspec.)

Napoopoo

keauhou

kailua

Kawaihae

Notes:

11:00

10:00:t

>15:00 15:36

15:20

<16:00 17:40

15:20

15:20

R 2 to 3

+10

-8 +9It

Yarning issued by Hllwn. Yolc. Obs.

Warning issued Sa ... n IIIOved I nto harbor and Coconut

Island cleared .

First arrival First arrival Height at Yailuku River, several feet

greater at Walloa River

Sea bottom bared 300 ft. inundation, fishpond wall destroyed

First arrival Series of waves at 10 minute intervals Tenth wave largest, observed from

Kawaihae to Hanamalo Pt. (S. of M110i1)

Yarning issued by Hawn. Yole. Obs. Cargo reaoved from wharf. Seventh wave largest. Sea bottom bared.

Arrival Inundation across road around trees

(2 feet deep), and around a house (7 feet deep?); boathouse and water tank moved; stone walls damaged, sampans,

Hnt 3 Mar, SB 3 Mar

Jaggar 1933 Hnt 3 Mar

SB 3 Mar

HnI 3 Mar Jaggar 1933

Hnt 3 Mar Jaggar 1933

Jaggar 1933 Hnt 3 Mar

Jaggar 1933 HA 3 Mar

HSB 3 Mar

Jaggar 1933

canoes, and a motorboat washed ashore Hnt 3 Mar MaxllIUII drawdown (1851) MaxillUll runup (IRS1) Jaggar 1933

Yarning issued by Hawn. Yolc. Obs. KGna Inn reta1ning wall damaged. Water

topped porch and splashed into lobby, Wall broken and caretakers house moved

at palace _rican Factors It..ber yard inundated

and lumber scattered Sampan on marine railway near KGna Inn

tossed over seawall Doats set adrift ' Series of 10 large waves. last largest;

10 minute wave period. Considerable damage.

10 minute wave period, some damage

Jaggar 1933

Hnt 3 Mar HA 3 Mar

HSB 3 Mar Pararas-Carayannis

(1969); Pararas­Carayanni sand Calet>augh (1977)

Pararas-Carayannis (1969); Pararas­Carayannis and Ca lebaugh (1977)

a) Time Is Hawaiian Standard Time (- UHf - 10:30), 2 March unless otherwise indicated. b) Newspaper references: 11. _ lJ .... __ 'u1 •• a ..... __ ........ __ HSB • Honolulu Star-Bulletin

UTU _ U.c1". T ........ ' .... J!II_Ua ... s1,f ~ • Maui News GI • Garden Island

w ....

32

TSUNAMI OF MARCH 1933

Introduction

Another great tsunami was generated off the Sanriku coast on 3 March 1933, in the Showa era. It accompanied an earthquake of magnitude 8.3 occurring 2 March at 17:31 UMT whose epicenter was at 39.1 N, 144.7 E. The tsunami had a maximum measured runup height of 28.2 meters on the Rikuchu coast. It caused the deaths of over 2,700 persons in Japan and injuries to over 1,000 more. Nearly 8,000 houses were destroyed and over 8,000 vessels of various sizes were wrecked.

The origin area of the tsunami, as estimated by Hatori (1974) from its arrival times at places on the Japanese coast, was elongated north-south and had a length of 300 km. The earthquake epicenter lay near the center of the seaward edge. As thus estimated, the origin area of the 1933 tsunami overlapped that of the 1896 tsunami but extended farther south and not so far northwest.

The tsunami arrived in Hawaii on the afternoon of 2 March (HST). Its effects here were similar to, but not so severe as those of the 1896 tsunami, and its runup heights were not so great.

Contemporary accounts

The character and effects of the tsunami on the island of Hawaii were reported on 3 March and 4 March 1933 by the Honolulu Advertiser. the Honolulu Star-Bulletin, and the Hilo Tribune Herald. The incidence of high waves on the southwest coast of Maui was reported by the Maui News on 3 March and 6 March, and very brieny by the Star-Bulletin on 3 March. Wave effects on the south coast of Kauai were reported by the (weekly) Garden Island on 7 March.

A brief account of the tsunami was published by Jaggar (1933) in the report of the Hawaiian Volcano Observatory for March.

Subsequent accounts

The occurrence of the tsunami in Hawaii was noted by Neumann (1935), by Jaggar (1946, 1947), by Powers (1946(a», and by Shepard ~ al. (1950). Hatori (1963) noted its runup at Napoopoo. Information on its Hawaiian effects was summarized in the Hawaiian tsunami catalog (Pararas­Carayannis, 1969; Pararas-Carayannis and Calebaugh, 1977).

Warnincs. characteristics and effects

The earthquake was recorded by Hawaiian Volcano Observatory seismographs at Kilauea, Hilo and Kona, on the island of Hawaii. On the basis of the seismographic records and radio reports of the effects of the earthquake and tsunami in Japan, the Observatory issued warnings of the expected arrival of the tsunami to the Hilo harbor master and to representatives of American Factors, Ltd., at Kailua and Napoopoo. As a result of these warnings, boats in the Wailoa estuary at Hilo were moved out into the harbor, and cargo was removed from the wharf at Napoopoo. No damage resulted trom the waves at these places.

Some damage occurred at Kaalualu, Kau, and at Keauhou and Kailua, Kona. According to the Hawaiian tsunami catalog (Pararas-Carayannis, 1969; Pararas-Carayannis and Calebaugh, 1977), the waves caused some damage at Kawaihae, but no reports of damage at Kawaihae have been found in contemporary sources of information. The waves were observed on Maui, Oahu, and Kauai, and were recorded on the tide gage at Honolulu.

Characteristics of the waves, including arrival times and sizes, and their effects, as reported in contemporary. reports, are summarized in Table 11.

Arrival times and travel times

The times of arrival of the tsunami reported in contemporary sources of information are summarized in Table 12. According to the Hawaiian tsunami CatalO( (Pararas-Carayannis, 1969;

Table 12. Hawaffan arrfval times and travel times. tsunami of March 1933

Times (hr: mia.) calculated

Reference a) Reported Arrival Arrival

Place HST lIfT

Midwy 2 Mar 11:057

Kauai 2 Mar 7

Oahu

Honolulu Jaggar.1933 2 Mar 14:40 3 Mar 01:10

Mauf

Lahaina HS8 3 Mar 2 Mar (17:007) 3 Mar 03:301 HA 4 Mar 3 Mar 05:00± 3 Mar 15:30

~

West coast Jaggar. 1933 2 Mar 15:20 3 Mar 01:50

Hflo Jagger. 1933 2 Mar 15:36 3 Mar 02:06

Notes:

a) Newspaper references HA • Honolulu Advertiser HS8 • Honolulu Star-Bulletfn

33

Travel

7:39

09:591 21:59±

8:19

8:35

Pararas-CarayannJs and Calebaugh, 1977), the tsunami arrived at both Kawaihae and Kailua at 15:20. This seems to have been based, however, on Jaggar's report that the waves arrived at that time on the west coast of Hawaii, without specifying the place. The report may have referred to the arrival at either of the Kona coast places where the Hawaiian Volcano Observatory issued warriings, Kailua and Napoopoo, but it is very unlikely to have referred to Kawaihae.

The major possible errors in the reported times at the tide gages at Midway and Honolulu are probably clock errors. Jaggar's report of the arrival time in Hilo is probably very nearly as accurate, since it was given to the nearest minute.

The reports of the arrival times at Lahaina, Maul, are confusing. Both the Honolulu Star-Bulletin and the Honolulu Advertiser reported the observation there of three "small" or "miniature" tidal waves. The Star-Bulletin reported that these waves were observed "Thursday night", I.e. on the second of March, but the Advertiser reported that their first observation was the next morning. It is possible that both reports refer to the same waves, that the report to the Advertiser was confused, and that the waves were first observed 2 March at 5:00 p.m. rather than 3 March at 5:00 a.m. Even if this were the case, the first observation was about 2 hours later than the expectable time of the first arrival. However, the highest waves in the train observed on Hawaii occurred over an hour after ·the initial arrival, and it may be that the early waves of the train were not observed at Lahaina.

Waves occurring still later, on 3 March, were responsible for cancelling shipping activities at Kaanapali, north of Lahaina, but these waves seem to have been the waves of a Kona storm that was affecting all of th_e islands by then.

w ... Table 13. Hawaiian runup heights, tsunami of March 1933

Press Reported runup he~hts. feet Accepted runup reightS.

feet msl c Place referencea) Datum

.COnt~rary Jaggar lIa r1 pararas-b) Min. Prob. Mix. press (1933) (1967) Carayannis

Hawaii

Mahukona <5

Kawaihae <6

W. coast unspec. HSB 10

Kailua-Kona 9 10 11

Keauhou HTH ground? 7 HTH ground 23 7Js 101t 14

Napoopoo 9It 17Js 9It 9 9It 10

Kaalualu HTH a 11 15

Honuapo <6

Punaluu <a

Kalapana <a

Hilo HTH (t .. )d) 1 1" 2

Miui

lahaina HSB. HA 1 2 5

Oahu

~nolulu HSB (o.a)e)

Place

lCaua1 (cont.)

Naw111w111

Kuku1ula

Pakala

Notes:

Press ref.a )

61

61

61

Table 13. (continued)

Reported runup he1 atlts. feet

DIlt .. COri1:ellporary Jiggar URatori ~ pararas-b)

press (1933) (1967) Carayannis

a) Press references (which refer to reports of effects as well as heights).

61 • Garden Island HA • Honolulu Advertiser HS8 • Honolulu Star-Bulletin Hllt· H110 Tribune Herald

b) Pararas .. Carayann1s (1969) and Pararas-Carayannisand Calebaugh (1977)

c) Min • Lower limit of range of reasonable values Prob • Most probable value Mix • Upper limit of range of reasonable values

d) Half of reported .x1 .. range

e)_ Mil.r1graph1c amplitude

Accepted runup heights. feet 1IS1C,

M1 n ~riiti;·Max.

3

3

3

4

4

4

5

5

5

.. ..

36

Travel times of the tsunami to various places have been computed, assuming the tsunami was generated at the time of the earthquake, and added to Table 12.

Runup heights

The runup heights that have been reported for the tsunami at places on Hawaiian coasts are summarized in Table 13.

Hawaii

It will be noted that Hatori (1963) reported a 17t foot runup at Napoopoo, but this figure obviously represents the range from trough to crest reported by Jaggar (1933). Jaggarts report of a runup height of 9i feet may be assumed accurate.

No runup height was reported at Kailua-Kona, but from the reported effects of the waves it appears that the runup must have been about the same as that of the 1896 tsunami there, somewhere between 9 and 11 feet.

Heights of 2 to 3 feet at Keauhou reported in the Hilo Tribune Herald (3 March) were clearly depths of inundation above ground leveL A height of 7 feet reported there in the same newspaper as if it were a depth of inundation would imply a runup height much greater than is consistent with the reports of damage. However, it is unlikely that the damage could have resulted from a wave with a runup height of only 7 feet. From the various reports it appears that the runup at Keauhou was somewhere in the range from 7 or 8 to 14 feet. .

The 300-foot inundation at Kaalualu reported by the Herald, and the baring of the sea bottom there reported by Jaggar (1933), suggest that the runup might have been as great as 15 feet, but this is probably the upper limit of the range of reasonable values. The lower limit might be as little as 8 feet.

It is unlikely that the runup exceeded 8 feet at Kalapana or Punaluu, 6 feet at Honuapo or Kawaihae, or 5 feet at Mahukona, considering the lack of reports that the tsunami was observed at those places.

The maximum runup at Hilo may be estimated from the crest-ta-trough ranges reported by Jaggar as about 1 i feet.

Maui

The observation of the tsunami at Lahaina (Honolulu Star-Bulletin. 3 March, Honolulu Advertiser. 4 March) suggests that it must have had a runup height there of at least a foot, but its description as ttsmalltt or ttminiaturett suggests that the runup was no more than 5 feet.

Oahu

The marigraphic amplitude of the tsunami reported at Honolulu by the Star-Bulletin (3 March) is double the amplitude reported for the 1901 tsunami. Accordingly, its runup in Honolulu Harbor may be estimated as twice that of 1901, or between 1 and 4 feet.

Kauai

The fact that sampans were forced by the tsunami to leave their moorings in Nawiliwili Harbor, in Kukuiula Harbor, and at Pakala (Garden Island, 7 March) suggests that the waves had a runup of at least 3 feet at those places. The fact that the tsunami had no other reported effects at the three places suggests that the runup8 did not exceed 5 feet.

The beach retreat reported by the Garden Island at Lawai would suggest a substantial runup there, except that it appears that the loss of sand may have been resulted primarily from a stream fiood rather than the tsunami.

37

TSUNAMI OF MARCH 1952

Introduction

A tsunami of moderate size in Japan accompanied a major earthquake that occurred on 4 March 1952 at 01:23 UMT. The epicenter of the earthquake was at 42.2 N, 143.8 E, off the southeast coast of Hokkaldo, about 200 km northeast of the estimated source areas of the tsunamis of August 1901. Its magnitude was 8.1 (lida et al., 1967).

The tsunami had runup heights of 3.3 m. in the Tokachi district and 3.6 m. in the Kushiro district of Hokkaido. It damaged port facilities and swept 91 houses away at Kushiro. To the southeast it had runups of much as 2.6 m. on the Sanriku coast of Hoshu.

A tsunami warning was issued by the Seismic Sea Wave Warning System on the basis of the occurrence and location of the earthquake and early observations of the waves.

The tsunami was recorded by tide gages on a number of Pacific islands and on the coasts of North and South America. It arrived in Hawaii on 3 March very late at night.

Hawaiian effects

The marigraphic. recording of the tsunami in Hawaii was noted in the Volcano Letter (Macdonald, 1952), in United States Eart uakes 1952 (Murphy and Cloud, 1954), and in the preliminary Pacific tsunami catalog lida et al., 1967 and Hawaiian tsunami catalog (Pararas-Carayannis, 1969; Pararas­Carayannis and Calebaugh, 1977). (A 1963 paper by Iida is cited in the Hawaiian tsunami catalog as a reference, but this paper does not contain information on the tsunami in KawaiO.

Characteristics of the tsunami as recorded on Hawaiian tide gages are indicated in Table 14. Hawaiian standard time in 1952 was 10 hours behind UMT, rather than 10* hours as it had been in 1933 and earlier. This time difference, and tsunami generation at the time of the earthquake, are assumed in calculating the travel times of the tsunami.

Table 14. Marigraphic characteristics of the tsunami of March 1952 in Hawaii

Arriva 1 time. Travel time, Place hr:m1n, HST hr:m1n

Port Allen. Kauai

Pearl Harbor. Oahu

Honolulu. Oahu 2 Mar 23:15a) 7:52a) 7:30b)

Kahului, Maui

H110, Hawa11 7:42b)

. Notes:

a) From Macdonald (1952)

b) From Pararas-carayannis and calebaugh (1977)

c) From Cox (1964)

Maximum a..,litude. feet

0.3b)

0.3b)

0.5~~ 0.3

1.0b)

0.4c)

In addition to the marigraphic data, the Hawaiian tsunami Catalog reported runup heights for the tsunami for a number of places on the Kona coast of Hawaii, ranging (when converted from metric units) from less than 2 feet at Kawaihae to 3t feet at Napoopoo and Hookena (Table 15). However, the references cited in the catalog for the Hawaiian records of this tsunami do not contain this runup information, and contemporary issues of the newspapers provide no indications that the waves were observable on Hawaiian coasts.

Place

Kauli

Port Allen

Oahu

Honolulu

Pearl Harbor

Maui

Kahului

Hawaii

Hilo

Kalapana

Hookena

Nlpoopoo

KeauhOu

Kailua

Kawaihae

Notes:

a) From Table 14.

Table 15. Hawaiian runup heights of the tsunami of March 1952

MIl ri graphi c amplitude.

feeta)

0.3

0.3

<0.3

1.0

0.4

Reported runup height. feetb)

3

3Js

3Js

3

2

<2

Accepted runup ) heights. feet msl c Min. Probe Max.

1

1

1 lis 2

Is 1d) lis

Reported rvnup hei ghts cons.i dered not credible

b) From Pararas-Clrayannis (1969) and Pararas-Clrayannis and Cllebaugh (1977).

c) Accepted runup heights: Min • lower limit of range of reasonable values Prob • JItIst probable value Max • Upper limit of range of reasonable values

d) From Cox (1964).

39

The Honolulu Advertiser, Honolulu Star-Bulletin, and Hilo Tribune Herald reported on 4 March the occurrence of the earthquake and tsunami in Japan, the issuanCe of an alert by the warning system, the recording of waves with only 0.5-foot amplitude at Wake Island that led to a decision not to issue a more formal warning, and the lack of a significant disturbance at Midway Island that led to ending the alert. The article in the Star Bulletin also noted the recording of the tsunami on the Honolulu tide gage, reporting that its amplltude was 6 inches. The Garden Island on Kauai (5 March) discussed the alert. The Maul News did not mention the event.

It appears that the runup heights reported in the Hawaiian tsunami catalog as if referring to the Japanese tsunami of 3 March 1952 actually refer to the Kamchatka tsunami of 4 November 1952. There seems to be no credible evidence that the tsunami was actually observed on any Hawaiian coast, and the only credible evidences of its size in Hawaii are its marigraphic amplitudes.

Considering the amplitude at Kahului, the runup there was probably between 1 to 2 feet, and it would seem that the tsunami could have been visually observed in the quiet waters of the harbor if it had not arrived in the middle of the night.

The warning of the tsunami issued in Hawaii may not have been intended as a formal warning by the Seismic Sea Wave Warning System. Civil Defense warning sirens were not sounded. However, the warning was called to public attention by radio announcements, and at least on Kauai by the police (Garden Island, 5 March). Shortly after 20:00, announcement was made that the waves had had an amplitude at Wake Island of only half-a-foot. The alert was cancelled shortly after the expected arrival of the tsunami in Hawaii (Honolulu Advertiser, 4 March) on the basis of reports that its effects at Midway Island had been negligible.

40

TSUNAMI OF MAY 1968

Introduction

A Sanriku tsunami of moderate size accompanied a very large earthquake occurring on 16 May 1968 at 00:49 UMT. The earthquake epicenter was reported at 40.8 N, 143.2 E by the U. S. Coast and Geodetic Survey (Pararas-Carayannis and Calebaugh, 1977) and at 40.7 N, 143.6 E by Japanese authorities (Hatori, 1974b). Its magnitude was 7.9. From the arrival time of the tsunami at various places on Japanese coasts, Hatori (1974) has estimated its origin area as an ellipse, elongated northwest-southeast and about 230 km long, approximately centered at the earthquake epicenter. This area lies mostly seaward of, but overlaps, the estimated origin areas of the August 1901 tsunamis; its southern end overlaps the origin areas of the June 1996 and September 1933 tsunamis; and its northern end very likely overlaps the origin area of the March 1952 tsunami.

The tsunami was recorded by tide gages on a number of Pacific coasts, including those of the Hawaiian Islands, where it arrived late at night on 15 May.

Hawaiian effects

The tsunami was recorded on tide gages at Midway atoll, at Nawiliwili on Kauai, at Honolulu and Mokuoloe Island {Kaneohe Bay on Oahu; at Kahului on Maui; and at Hilo on Hawaii. Values identified as amplitudes reported in the' Hawaiian tsunami catalog (Pararas-Carayannis and Calebaugh, 1977) appear to be ranges. Arrival times and amplitudes read from the tide gage records are indicated in Table 16.

The tsunami seems not to have been observed visually in Hawaii. The only mention of its Hawaiian effects in the newspapers were an account in the Hawaii Tribune Herald (16 May) noting the marigraphic recording of the tsunami at Nawiliwili and describing the marigraphic record at Hilo and a note in the Honolulu Advertiser (17 May) that the waves were barely noticeable, probably referring to their marigraphic record at Honolulu.

An alert was issued by the Tsunami Warning System, but this was cancelled at 18:56, shortly after the tsunami reached Midway without development into a formal warning. A second alert was issued early the next morning (16 May) on the basis of the recording of an aftershock of the Sanriku earthquake, but this alert was also cancelled, at 06:45. The Hawaii Tribune Herald reported that the aftershock " ... at 2:50 a.m. today and of 7.2 magnitude, touched off minor tsunamis at Pacific tide stations ... ". However, the wave motion reported was probably that of the original tsunami.

Travel times and runup heights

Travel times of the tsunami to tide stations in Hawaii, calculated assuming that the tsunami was generated at the time of the earthquake, have been added to Table 16. Also added to the table are runup hei~hts estimated from the marigraphic amplitudes of the tsunami.

Table 16. Hawaiian marigraph1c dAta and runup heights of tsunami of May 1968

Times i hr:m1n Arrival

Mar1yraPh1c amp 1tude

Est1mated runuPa,e1ghts, feet 11151

Place HST,15 May Trayel feet

Midway Island 18:42 3:59 0.8

kaua1

Naw111wt 11 22:03 7:14 0.2

o.hu

Honolulu 22:23 7:34 0.4

Coconut Is. 22: 26tb) (kaneohe Say) 7:37t 0.1

Mau1

kahului 22:29 7:40 1.6

Hawaii

Hflo 22:45 7:56 0.9

Notes:

a) Estimated runup heights: '-in • Lower limit of range of reasonable yalues Prob • Most probable yalue Max • Upper limit of range of reasonable yalues

b) Chart speed erratic

Rln. proS. Mix.

1

1" 2Js

1 2 3

41

42

SUMMARY OF HISTORIC RECORD

General comments

From the evidences of the historic record presented and discussed in previous chapters, it is clear that the Japanese tsunamis of June 1896 and March 1933 caused considerable damage in the Hawaiian Islands, especially on the Kona coast of the island of Hawaii, and that the tsunami of August 1901 also caused some damage there. It also appears that reported Hawaiian observations of the two tsunamis of December 1854 and the tsunami of March 1952 are erroneous.

The amplitudes of the tsunamis of March 1952 and May 1968 as recorded by tide gages were so large it would seem those tsunamis could have been visually observed at Kahului and Hilo except that they arrived at night. The Japanese tsunamis of September 1923, December 1944, November 1953, June 1973, and June 1975, and a Bonin Island tsunami of September 1952, were also recorded on tide gages in Hawaii (Pararas-Carayannis and Calebaugh, 1977). However, their marigraphic amplitudes were so small that their observation would have been difficult, and they had no significant Hawaiian effects. '

It is possible that, between 1813 or 1814 and 1896, some Japanese tsunamis were observed in Hawaii, especially on the Koria coast. However, it is unlikely that any others were observed at Hilo or Honolulu between 1847 and 1896. It is doubtful that, during the last century at least, any tsunamis other than those of 1896, 1901 and 1933 have had effects of practical significance anywhere in Hawaii.

Travel times

Travel times for the Japanese tsunamis of 1896, 1901, 1933, 1952, and 1968, computed on the assumption that each of these tsunamis was generated at the time of the accompanying earthquake, are compiled in Table 17. Computation of additional travel times should be possible from marigraphic records of the 1923, 1944, 1953, 1973 and 1975 Japanese tsunamis when they are available. However, even in this preliminary report, certain comments are appropriate.

All of the tsunamis for which travel times are indicated were generated off the Sanriku coast of Honshu or the adjacent coast of Hokkaido. The Tsunami Warning System chart of tsunami travel times to Honolulu (Fig. 1) indicates that a tsunami from this general source area should be expected to arrive at Honolulu about 7 t hours after generation. The minimum-travel-time path to Honolulu would lie north of the Leeward Islands of the Hawaiian archipelago, or most of them. It might pass through the archipelago between a seamount southeast of Nihoa and the seamount capped by Kauai, Niihau and Kaula; through the Kauai channel, northeast of Oahu; or possibly even through the Kaiwi channel, southeast of Oahu.

Waves from this general source area should be expected to arrive at places on the northeast coast of anyone of the major islands earlier than at places on the south coast -of the same island. Otherwise their travel times should be expected to increase successively at places on Kauai, Oahu, Maui, and Hawaii. No more than about an hour should be expected to elapse from the earliest arrival on Kauai to the latest arrival on Hawaii. Between the two historic travel times to Hilo reportedly based on marigraphic evidences, and among four out of five of the historic travel times to Honolulu, there are agreements within less than a quarter of an hour. As earlier noted, the travel time to Honolulu of the 1901 tsunami is anomalous. The other reported travel times appear consistent with the majority of the Honolulu travel times and with the Hilo travel times based on marigraphic evidences, except as follows:

1) The reported arrival of the 1968 tsunami at Kaneohe Bay on the northeast coast of Oahu was a few minutes later rather than earlier than its reported arrival at Honolulu on the south coast; and the reported arrival of the 1896 tsunami at Hilo on the northeast coast of Hawaii was more than 20 minutes later than its reported arrival at Kailua on the southwest coast. However, the arrival time of the 1968 tsunami at Kaneohe is unreliable; and as will be seen there are reasons for supposing that the first waves of the 1896 tsunami were not observed at Hilo.

Place

Kauai

Kapaa

Haw111w11 i

Port Allen

Oahu

Kaneohe Bay

Honolulu

Maui

Kahulu1

Laha1na

KawaU

Hilo

Punaluu

Honuapo

ICaalualu

Hookena

Kealakekua Bay

Keauhou Bay

Kaf1ua

Kawa1hae

Notes:

Table 17. Travel times of Japaoese tsunamis to Hawa11 (hr:min}aJ

1896 1901 1933 June August March

7:27

7:35 12:36 (1) 7:39

9:59 ~??) 21:59 11}

8:35 :I: O:20b)' (1) 8:35 (1)

9:37 :t (1)

9:37 :t (1)

9:37 :t (1)

9:27 :t (1) 12:21 (1)

8:57 :t (1)

8:27 :I: (1)

8:12 ~ O:15b} 12:0& (1) 8:19

8:12 :t O:10b)

1952 1968 March May

7:14

WJ.±

~~~~ 7:34

7:40

l:..!2. l.ii§.

a) All travel t1mes are cGq)uted assuming the tsunamis were generated at the time of the assoc1ated earthquakes. Travel times underlined are based on .. r1graph1c 1nforwtion. Others are based on visual obServation.

b) Mean and .. xi mum deviations of,.ultiple estimates.

c) From Pararas-Carayannis and Ca1ebaugh (1977).

d} Based on arrival ti. reported by Macdonald (1952).

43

44

2) The Hilo travel times of the 1896 and 1933 tsunamis, although in agreement with each other, are three quarters of an hour longer than the mean of the 1952 and 1968 travel times based on marigraphic evidences. The discrepancies may have resulted from failures to observe the first few waves in the train at Hilo. Reasons have been presented for supposing that the time when the 1933 tsunami was thought to arrive was accurately noted.

3) Travel times of the 1896 tsunami to ports on the southern part of the Kona coast and on the Kau coast are too long by as much as 1 t hours. The discrepancies may have resulted in the same way as that at Hilo. One would expect an accurate report in the case of the time or arrival of notable waves at Honuapo where the "W. O. Hall" was transferring cargo when they arrived.

4) Even after correction for a possible 12-hour error, the reported travel time of the 1933 tsunami to Lahaina was 2 or 2t hours too long. The report is questionable, but the discrepancy may have resulted in the same way as that at Hilo.

5) The travel times of the 1901 tsunami to Kailua and Hookena are about 4 hours too long, although an hour shorter than the travel time indicated by marigraphic evidence at Honolulu. The anomalous travel times of the 1901 tsunami have been discussed at some length in the chapter on that tsunami.

Runup heights

Ranges of possible error and most probable values

Some range of possible error is associated with even the most precisely measured runup heights. This range may be no more than about a foot, but the range in the case of an estimate based on the reported effects of a tsunami, especially one made many decades later, may be a major fraction of the most probable value of the runup. Because of the great variability in the accuracy of runup height measurements and estimates, it is important that estimates of probable limits to the errors in them be noted and taken into account in their use.

Two different kinds of error should be distinguished. On the one hand, in the determination of the level of the datum plane from which a runup height is measured, or in the measurement of the difference in level between the datum plane and in evidence of the runup, the greatest overestimate likely is approximately equal to the greatest underestimate likely. If:

Ho most probable value

Hmin = minimum 1 ike 1 y value

Hmax maximum 1 ike 1 y val ue

lIH Ho Hmin = H

max - H 0

Ho = (Hmax, - Hmin)/2 = arithmetic mean of Hmin and Hmax

In this case the most probable value and limiting values of the runup may be reported as:

H = H :t: lIH o

(1)

(2)

(3 )

On the other hand, in the estimation of a runup height without measurement, as from a description of the effects of the waves, it is more likely that the ratio of the limiting minimum value to the most probably value is approximately equal to the ratio of the most probably value to the limiting maximum value than that the arithmetic differences between the limiting values and the most probably value are equal. In this case:

HO = ;fHmax Hmin ; geometric mean of Hmin and Hmax

and the probable limits of the runup might be reported as:

H = H f±l o

[

H 'k

where f = m~xJ 2 Hm1n

45

(4 )

(5)

(6)

Cox and Morgan (1977) reported measurements and estimates of the runup heights of local tsunamis and possible local tsunamis in the form ·of equation (3). Many would better have been reported in the form of equation (7). However, where either the most probable runup value, H~ is small, or the ratio t.H/Ho_ and the value f are small, the difference in the values of the most problkble runup as computed from equations (2) and (5) will be small. Hence most probable, maximum likely, and minimum likely values of runups have been assumed in this study to be related in accordance with equations (4) through (6). '

Because estimation of the probable limits of error of previously reported runup heights, and the estimation of the probable limits to ranges of the values of runups based on descriptions of effects, have' neccessarily been judgmental rather than statistical, rigorous interpretation of the accepted limits as corresponding to precise statistics is not possible. However, if the results are to be used in further statistical analysis, the limits should be considered approximately the equivalent of limits of 95 percent confidence.

Historic record

The Hawaiian runup heights of the June 1896, August 1901, March 1933, March 1952, and May 1968 tsunamis are summarized in Table 18 as they have been estimated in this study.

Runup heights for the 1896 and 1933 tsunamis have been used by Houston et aL (1977) in estimating tsunami runup frequency distributions and 10o-year tsunami runup heights on the Konacoast of Hawaii, but not elsewhere, for the National Flood Insurance Program. As may be seen from Table 18, the 1896 tsunami, at least, had significant runup heights on Kauai, as well, and also on parts of the coast of Hawaii other than in Kona.

In Table 19, runup heights for the two tsunamis used in the National Program, supplied me by the Corps of Engineers, are compared for places on the coast of Hawaii, with the values that seem most probable from this study.

As will be seen, the 1896 values employed by Houston!!.!!- for the coast from Napoopoo to Kailua, which were based on the 30-foot reports in Pararas-Carayannis and Calebaugh, greatly exceed the most probable values. At other places, the runup was assumed insignificant even where there are significant most probable values.

The agreement between the 1933 values employed by Houston et.!!- and the most probable values is fairly good on the Kona coast, but again there are plaCes where there are significant most probable values where it was assumed that there were none.

Although considerable uncertainty remains as to the actual runup heights at many of the places where wave estimates have been made in this study, it seems best that the values here accepted as most probable should be substituted for the values previOUSly used and that the runup frequency distributions and 10o-year runup heights to be used in the National Flood Insurance Program should be recalculated accordingly.

46

Table 18. Runup heights of Japanese tsunamis in Hawaii

June August March to'.arch May Place 1896 1901 1933 1952 1968

H-f ~ H~ H~ H-o-f 0

Kaual

Hanalel <9 Kl1auea Lndg. 5 1.6 Kapaa 4 1.3 Nawiliwill 5 1.6 4 1.3 <~ Koloa Lndg. <5 Kukuiula 4 1.3 Hanapepe-Port Allen <6 Pakala 4 1.3 Waimea <8 4 1.3

Oahu

Waikiki 2 2.0 Honolulu Habor lit 1.5 1 .1.5 2 1.0 <1

!!!!!.!. Makena <4 Lahaina <4 2 2.5 Kaanapali lit 1.7 Kahului <4 lit 1.4 21t 1.5

Hawaii

Laupahoehoe <10 Kaiwl1ahl1ahi 10 1.3 Halealau <8 Hilo 8 1.3 <3 lit 1.4 2 1.7 Kalapana <10 <8 Punaluu 12 1.2 <8 Honuapo 12 1.2 <5 <6 Kaalualu 12 1.2 11 1.4 Hoopuloa <10 4 1.8 <8 Hookena 11 1.3 Napoofoo 17 1.7 3 2.3 9It 1.0 Kaawa oa 17 1.7 Keaahou 18 1.8 BIt 1.4 10lt 1.3 Kailua 10 1.1 7 1.4 10 11.1 Kawaihae BIt 1.4 <6 Mahukona 51 <5

Note: H • most probable value 10 • factor for estimating limits to reasonable ranges Qf values

Table 19. Comparison of 1896 and 1933 tsunami runup heights on the island of Hawaii used in

National Flood Insurance Program (NFIP) with those considered most probable in this study

RUllup heights. feet msl.

Site numbera) 1896 June 1933 March

This This Place WE~ HIG NFIP study NFIP study

Hawaii

laupahoehoe 82.3 8-1 0 <10 0

Kaiwllahil ahi 83.5 0 10 0

Hakalau 88.7 { 9-1 or} 9-2 0 < 8 0

Hilo 103.8 11-5 0 8 0 lit

Kalapana 129 16-3 0 <10 0 <8

Punaluu 145.7 18-1 0 12 0 <8

Honuapo 148.3 e8•3

• } 19.1 0 12 0 <6

Kaalualu 152 20.3 0 12 0 11

Hoopaloa 11.2 22-3 0 <10 0 <8

Hookena 16 23-1 7.0 11 9.5

Napoopoo 19.5 24-2 30.0 171t 9.5 9lt

Kaawaloa 20.1 24-3 30.0b) 171t 9.3b)

Keaahou 24.5 25-1 30.0 191t 9.2 10lt

Kailua 27.4 23.4b) 10 5.5b) 10

Kawai hoe 47.7 1-2 0 SIt 0 <6

Mahukona 54.6 2-1 0 0 <5

Hotes:

a) See section on site numbering systems.

b) lnterpollated.

47

48

LONGITUDINAL RUNUP PROFILES

Needs and schemes for reconstruction

There is direct historic evidence for estimating the runup height of the 1952 Japanese tsunami at only one coastal site in Hawaii, and there are such evidences in the case of the 1896 tsunami, which had the highest runups in Hawaii of any historic Japanese tsunami, at only 17 sites. The number of sites where the runup may be estimated for other historic Japanese tsunamis varies from 2 in the case of the 1968 tsunami to 11 in the case of the 1933 tsunami.

It would be absurd to assume that the runups of these tsunamis were significant at these sites alone, and for the purposes of tsunami hazard zoning is necessary to estimate somehow the runups of the tsunamis at coastal sites generally, in other words to reconstruct near-shore runup profiles of these tsunamis along the coasts. Such profiles may be termed longitudinal runup profiles to distinguish them from the transverse runup profiles extending from the shoreline inland to the inundation limits.

Two fundamentally different schemes might be considered applicable to the reconstruction of longitudinal runup profiles, and two other schemes representing modifications of these have been found useful in practice.

The simplest scheme would be to interpolate the runup heights of a tsunami in some simple fashion between the sites where the heights are known or may be estimated from historic evidence, and to extrapolate the runup heights in some simple fashion as neccessary beyond the range of such sites.

Although it might seem that no geophysical understanding would be needed in such a scheme, it is neccessary to adopt certain criteria for the interpolation and extrapolation that must be based, at least, on general characteristics of the runup pattern of the tsunami. In practice it has been found desirable to utilize the characteristics of the runup patterns of other tsunamiS, particularly others from the same source region, and other geophysical information, to guide and modify the results of the interpolation and extrapolation. In a practical scheme of this sort no numerical model is used Hence such a scheme will be referred to as a non-model scheme.

Theoretically it may be possible some day to determine the characteristics of the source of a tsunami and the behavior of the tsunami as it propagates across the ocean and runs up on shore with sufficient accuracy to allow calculation of its runup height at any point on the coast, using information about the source and a numerical model but no historical evidences of the runup.

However, it will probably never be possible to determine the source characteristics of most historic tsunamis with sufficient precision, and when the best of present numerical models are applied, even to recent tsunamis, the runup patterns they indicate differ substantially from the patterns insofar as these are defined by historic evidence. Hence, in practice, even if a numerical model is used in the reconstruction of the runup profile of a tsunami, it is neccessary to adjust its results to the runup heights indicated by the historic record of the tsunami and, generally, to the runup patterns of other tsunamis from the same source region. A scheme using a numerical model in this way will be referred to as a model scheme.

Model and non-model schemes have been used in previous reconstructions of Hawaiian longitudinal runup profiles of tsunamis as follows:

Reconstruction scheme used in this study

A non-model scheme has been adopted in this study. It inVOlVes, essentially the revision of previous reconstructions of the profiles of Japanese tsunamis by Houston et a1 (1977), using more extensive records of the tsunamis and of evidences of their runups. As pointed out by Cox (1979), considerable judgment is inescapable in reconstructing the runup profile of a tsunami between the sites of known runup heights. It is important to adopt a set of principles for the reconstruction that, in spite of some inherent arbitrariness, are reasonable and as objective as possible. It is also important that these principles be stated explicitly so that others can investigate the effects of other reasonable principles and form their own judgments.

49

The principles used in this study were in most respects those developed for application to the local tsunamis (identified as "criteria" by Cox, 1979) with certain generalizations to permit their application to distant tsunamis as well as local tsunamis. However, certain generalizations are necessary to allow for permitting their application to distant tsunamis as well. As thus generalized, and with some reorganization and revision for the sake of clarity, the principles are presented below. The rationale for the original principles has been discussed elsewhere (Cox, 1979). Principles significantly modified are indicated by asterisks in the following presentation. The rationale for the modifications will be discussed after the presentation.

Reconstruction principles

a. Use of historic information generally

(i) All previously reported measurements or estimates of the runup of a tsunami are taken into consideration in the reconstruction of its longitudinal runup profile. Each measurement or estimate is identified as either reliable or questiollable.

(ii) A range of possible runup values is estimated for each site where marigraphic information is available.

(iii) A range of runup heights is estimfl,ted for each additional site where the estimation may be based on the reported effects of the tsunami.

b. Datum plane

(i) Runup heights used in profile reconstruction are adjusted to mean-sea-Ievel datum where the adjustment is appropriate and bases for the adjustment are available.

(ii) In the reconstruction of the profile of a locally generated tsunami accompanied by coastal subsidence, runups at sites on subsiding coasts are adjusted to pre-subsidence mean sea level

c. Most probable runup values and limits of error

Except in the case of a site at which only a questionable runup value, or only an upper limit to the possible runup, is reliably estimable from a description of effects, not only the most probable value of the runup but probable upper and lower limits to the range of uncertainty of the runup are estimated as well as possible from all available historical information as follows:

• (i) At the site of a previously reported measurement or estimate that is considered reliable:

r = 0

previously reported value

rmax = r/

rmin = rolf

• (ii) At the site of an estimated range of uncertainty of the runup:

ro = /rmax . rmin

where r = historically most probable value o rmin = probable lower limit of range of uncertainty

rmax ,= probable upper limit of range of uncertainty

f = (rmin I rma~)~

50

d. Plotting positions and scales

m A logarithmic scale is used in plotting runup values.

(ii) A linear scale is used in plotting coastal distances, which are measured along straight lines connecting coastal nodes in the Waterways Experiment Station model (Houston et aL, 1977) neglecting details of coastal configuration between the nodes. --

e. Use of geophysical information

Within constraints indicated by criteria f through j the longitudinal runup profile of a tsunami is reconstructed:

(0 Assuming the runup pattern of the tsunami is similar to that of other tsunamis from the same source region; and \

(m Taking into account runup effects that are independent of the direction of tsunami approach.

(iii) In the case of a tsunami event involving two possible tsunamis with different sources, both possible sources are taken into account.

f. Treatment of ranges of uncertainty in runup and of local detail

(i) The profile is drawn through the historically most probable value at each site except as indicated in (ii).

• (ii) If the historically most probable value at a site is clearly inconsistent with the reported reliably reported effects of the tsunami at adjacent sites or with geophysical criteria em or e(ii), the profile is drawn through a runup value that is between the upper and lower limits of the range of uncertainty for the site and that is more nearly consistent with the reported effects at adjacent sites or the geophysical principles.

(iii) Where there are multiple reported measurements or estimates of the runup within a horizontal distance of 1000 feet or less along the coast, the profile is drawn through whichever historically most probable value is highest.

• (iv) Where there are multiple reported measurements or estimates of the runup at various distances normal to the ,coast at a site, the profile is drawn through the value pertaining most closely to a distance 200 feet inland from the shoreline.

g. Treatment of upper limits

At a site where only the upper limit of the possible runup is available, the profile is drawn through a value assumed as follows:

Runup estimated from Aasumed Cue adjacent values runup

(t) Considerably higher than Limiting value .Umltlng value

(U) SUghtly higher than or equal SUghtly lower to Umlting value than Umltlng value

(Ill) Lower than Umltlng value As estimated from adjacent values

, >

i. Treatment of questionable values

At the sites of a questionable runup value, the runup is estimated as follows:

Cu.

(t)

(ii)

(iii)

Runup estimated from adjat'~nt values

In close agreement with questionable value

Not In acreement, if slte of an adjacent value is cloae

Not in acreement, If site of nearest adjacent value is distant

Assumed runup

Questionable value

As estimated from adjacent values

As estimated by compromise

j. Interpolation, extrapolation, and treatment of runup values of uncertain location

In the absence of historical and geophySical evidence a profile is reconstructed assuming:

51

(i) Constant log-runup gradients for i!ferpolating between sites of previously available runup values separated by no more than about 2 x 10 feet.

• <U) A standard log-runup gradient for extrapol!Sing from a single runup value or a value separated from adjacent values by more than about 2 x 10 feet.

H - H e-a Ix - XII - I

• (iii) In the vicinity of a value of uncertain location:

For 0:; Ix'i :( " H a H e-2aA/3

o

For A:; Ix'i :; 2A H a H e-2a (X' - A)/3

o

For Ix'i > 2A

HaHe-2alx'l o

where H a runup

H1 a previously available runup value at site 1

H2 • previously available runup value at site 2

H a previously available runup value of uncertain location o

Xl • coastal distance to site 1, in feet

x2 • coastal distance to site 2. in feet

Ix'i ~ coastal distlnce fro. center of range of uncertainty in location

11 • indicates Ibsolute value of quantity between vertiCil blrs

A • 1/2 rlnge of uncertlinty in locltion

e a blse of nltural loglrithms

I • 1. 75 X 10-5

I • 1.25 X 10-5 for I tsunami of distant (non-Hlwaiiln) origin

a • 1.75 X 10-5 for I locll or possibly locil tsunami

52

Rationale for modifications of principles

cm and cOi): In the study of local tsunamis, the historically most probable runup at a site was considered to be the arithmetic mean of the upper and lower limits of the range of uncertainty. The reasons for considering in this study that the historically most probable value is the geometric mean of the upper and lower limits have been discussed in the summary of the historical information.

f(ii): In the application of the criteria to the reconstruction of local tsunami runup profiles, there are no coasts on which differences from the historically most probable values would have been suggested systematically by the criterion of assumed similarity of runup patterns of tsunamis from common source regions or consideration of the effects of such coastal features as reefs and narrow bay entrances. However, in the more general application of the criteria,it seemed wise to allow for the choice of runup values, within the ranges of uncertainty, on the basis of these criteria.

i(ii) and i(Ui); In the formulas applied to extrapolation and to profile reconstructi!?p in the vicinity of runup values of uncertain location, Cox (1979) assumed of a value of 1.75 x 10- for the coefficient "a". This value was derived from the average log-runup gradient of the tsunami of November 1975 between Keauhou Landing and Kalapana on the southeast coast of the Island of Hawaii. This tsunami was generated immediately off that coast. So steep a gradient would be atypical for the runups of tsunamis of distant origin. In accordance with a suggestion by Harold Loomis of the Joint Institute of Marine and Atmospheric Research at the University of Hawaii (personal communication), the value of the coefficient "a" used in this report has been derived from a study of the runup pattern of the April 1946 Aleutian tsunami. The northeast coast of Maul and adjacent parts of the northwest and southeast coasts (Fig. 3) were selected for the study because the northeast coast is oriented not far from normal to the direction of tsunami approach, as is the Kona coast of Hawaii in the case of waves from Japan, and because 1946 runup heights were measured along it at fairly short intervals.

The coast was divided into six segments, in each of which the runups pattern could be approximated by straight lines (the long-dashed lines shown in Fig. 3). Each line was drawn between a point representing a particularly high runup measurement (or Ilverage of closely spaced measurements) and a point representing a particularly low measurement (or average). The natur-,~. log-runup ~'$dients of these lines, if coastal distances were measured in feet, ranged from 0.90 x 10 to 1.38 x 10 . The value of the coeffiC!.iSnt "a" selected for use in this study was the average of the six natural log runup gradients, 1.25 x 10 .

The effects of this selection in the case of the 1946 tsunami are indicated by the short-dashed lines added to Fig. 3. Runups of the 1946 tsunamis at the sites of measurement along the coast are underestimated by the dashed lines, but by an average of only 4 percent.

Deleted principles: One of the principles applied in the reconstruction of the runup profile of a local tsunami was that a runup at any site was considered insignificant if it were smaller than the tenth highest runup in the record of other historic tsunamis at that site. The reason for adopting the principle was that, \ in determining site-specific runup frequency distributions for the Federal Flood Insurance Program, Houston et al. (1977) used only the ten highest runups at a site. The principle determines where the reconstruction of a longitudinal runup profile will not serve a particular purpose, but it does not guide the reconstruction as such. For this reason it has been deleted from the generalized principles. The longitudinal runup profiles reconstructed in this study have been drawn to 1-foot minimum runups.

Comparison of model and non-model schemes for profile reconstruction

There are more similarities between the model scheme used by Houston et al. (1977) and the non­model scheme used by Cox (1979) and in this study than might be assumed. Primary reliance is placed in both schemes on historical evidences of runup heights. In both schemes it is assumed that tsunamis from the same source region have had similar runup patterns. The model scheme is capable, at least in theory, of determining the influences of sea floor topography from the source region of a tsunami to a Hawaiian coast on the runup pattern. However, even in that scheme, adjustment is necessary to the runup patterns indicated by historic evidence.

Although the Houston et al. municipal model used a finer grid was used in the near-shore region than offshore, the grid is probably not fine enough to indicate well the effects of such coastal features

ci c: 53 <> <> <> I .. , 'tlll".11 dnuny

·1 ... ... .. ~ '" ... ... N

I/)

I n,nlltd!)! ... I J

I '" ( co /

I I

I I

I (

I I ; I • <>

~ • co

i ... <>

<> <> .U.H .. q <> 2

° '" 0 '" ... q

'" ... ::J IV

'Ilpu, n~!II.N <> ~ ° 2 o. '0 ° on '" -'" IV

0 <> "UU)! <..)

0 q J:;

'" on 1:: .. 0 c:

nUlwouOH t>O c: 0

0 i;

<> 'E IV c: ::J

0 .!!) 0 q c: on IV

'" N '':; ::J

1ft QI ... "< O~!I·W 1.0

.". 0'1 ...

4I!ld '0 • ~ Do ..

.. r!",,"~o.Jds '0 ~ ... I Ii c.

° .!! ... C. 8. I ::J

! '" c:

!nrn ... )! ::J

'Ii e ~

,I Ii M !t:

• § QI ... c: '" ::J I .S "11,1", t>O

J e ~ I '" ... • J : !

~ J! ';

~ ~ 2

i ... .Oln~IIII"

! .... ... .. e ... . 5 c: .. .. 2 2 ~ i I 1 'W ..... ~eN

! ... !t • i nlllO~ouOH

8. • 0 c: c: c: ... • .. • • a: a: III III

0 InlouOH

.nlle~ouOH

mdeN

... N

,.MOllOUOH

".d ...... "

<> <> 0 ~ on ... ... .. ... ... N

IU, '11111!811 dnunw

54

as reef channels and small bays. To some extent the adjustments for such effects in the non-model scheme, although largely judgmental, may represent an improvement that could be grafted onto the model schemes. However, even in the non-model scheme no attempt is made to make such adjustments generally at a scale of less than 1,000 feet.

Evaluation of differences between runups estimated from the model scheme and the non-model scheme must await some application of the non-model scheme to tsunamis frorn Kamchatka, the Aleutian Island<;, Alaska, or South America, because the model scheme has been applied only to tsunamis from these source regions. Differences between the Japanese tsunami runup profiles reconstructed in this study and those previously reconstructed by Houston et al. probably renect the greater utilization of historical information in this study rather than the effects any minor differences between the non-model schemes applied in the two studies.

Site numbering systems and profile orientation

In the figures in Cox (l979) showing the profiles of local and possible local tsunamis, site positions along the coastlines were referred to the serially numbered coastal nodal points in the' Waterways Experiment Station (WES) model developed by Houston!! al. (1977) to help reconstruct the profiles of the historic tsunamis from Kamchatka, the Aleutian Island<;, Alaska, and South America. In the WEB system applied to each island, the numbers begin with 1 at some point on the coast and increase in the clockwise direction around the island. Graphs of the coefficients of runup frequency distributions plotted against WES sites were published in Houston et al. (1977).

In tables of tsunami runup lleights estimated by Houston and his colleagues released by the Waterways Experiment Station, the sites of values for the larger recent tsunamis were referred to by numbers based on the compilation of those values by Loomis (1976) in a report published by the Hawaii Institute ot Geophysics (HIG). Loomis had plotted the sites for each island on a series of charts, numbered sequentially in a clockwise direction around each island, each chart showing the location of one or more sites of reported runups. The WES tables identified each site by Loomis's HIG chart number and the serial number of runup site shown on each chart, the site numbers also increasing clockwise around the island. There were some overlaps of the charts, so some sites appeared in the adjacent charts. In the WES tables, for example the runup measurements at Alealea Point, north of Hilo, Hawaii WES site number 102, were identified as Hawaii HIG site numbers 10-5 and as 11-1 (the fifth site (clockwise) on Loomis's Hawaii chart 10 and the first site on chart 11).

Although Houston et aL (1977) published the coefficients of the runup frequency distributions they computed in the form Of graphs in which the WES numbered nodal points were equally spaced, the numbers in neither system increase evenly with distance, and in general the HIG sites do not correspond to the WES sites.

To facilitate relating horizontal positions in the profiles in this report to both the Houston et!:!­graphs and the WES tables, place names and both WES and HIG numbering systems are indicated on the maps of Kauai, Maui, and Hawaii (Fip. 4 to 7), and place names as well as WES site numbers are indicated on the profiles (Figs. 8 to 11). In accordance with criterion dOl), horizontal distances in the profiles represent straight-line distances between WES nodal points.

The profiles in Cox (1979) were oriented as they would be viewed from the sea, in other word<; a leftward direction on a profile represented a clockwise direction around the island. The profiles in this report are oriented in the reverse direction so that both the WES numbers and the HIG numbers of sites increase from left to right.

Hawaii profiles (Figs. 8 and 9)

1896, 1901, and 1933 profiles, Kona and South Kohala coast

Along the Kona coast from Hookena to Kailua, the reconstruction of the profiles of the tsunamis of 1896, 1901, and 1933 is moderately well controlled by the historical evidences and the profiles (Fig. 9) have been based on the interpolation principle j{ii), with the additional guidance of the similarity principle em in some cases.

Between Kailua and Kawaihae the 1896 profile has been based on the interpolation principle. By the similarity principle, the 1933 profile would be superposed on the 1896 profile there. However, the

Coastal site Identifacations

... ... :. ~

Onshore-oWES system' Numbered nodal points of HOllston et al.. 1977_ model.

Offlhore--HIG system: RunuD mao and lite in lOomil, 1976, Number before dash is loomis map n IImb.r Number "fter dash i. I.rial number of .ite

of reported 'unUl) (counted clOck ..... )

"'y

9-2

1

o ... :. ~

~ iff ~

3-9

3-11

io :­.. ~

22'15'

''< I '$:>'b-_ ~ J 4-18 I 22"00' I ~ ~ . -~~

22*00'

(I SCI Ie

.1~ ~' ~ ..:t-'1i- ~'b-

o ~'1i-~'1i-

5 0 5 10 15 20 Mile. I , t , , I " Iii iii I I

25,000 o 25,000 50,000 75.000 ~

100,000 F .. I :. ~

Figure 4. Map of Kauai showing tsunami runup sites

~ :. ~

<1' <1'

"" .., ;. :!!

~, .... -:, ...... 1.>, "''p 'CI' ~ .... ~~

'-' '" 21' 00.'«>...,/'

12-7 12-6

12-6 12-3

;­~ ~

t/ ., l ... " .. ;-

20' 45' ~

Molokini I.

2 ;., l!!

Scale

o .. ., ~

... ~ l!!

o o ;., l!!

5 0 5 10 15 20 Mil., I I I I I I I I Iii iii I i

25.000 o

Makena

25,000 50,000 15.000 100.000 Feet

I Coastal site identifications

Onshore--WES system: Numbered nodal pointa of Houston et aI., 19n, model.

Offshore--HIG system: Runuo mao and site in loomi,. 1916, Number before dash is loomis map number. Number afler daah i, serial number of site

of reported run up Icountad clcek ... 1

'" ...

",:\> 50

O~1> 1>~ -<-O~ ~~1>""

o o

'" ...

5-2

s=4

Figure 5. Map of Maui showing tsunami runup Sites

(1'1 (I)

5'1

upper limit to the 1933 runup at Mahukona is significantly less than the 1896 runup there, and the distance from Kailua to Kawaihae nearly justifies the use of the extrapolation principle j(li). Hence the limiting-height principle has been applied in reconstructing in 1933 profile from Kailua to Kawaihae. The 1901 profile there has drawn assuming the same log-runup gradient as that of the 1933 tsunami.

1896, 1901, and 1933 profiles, Kau coast

In the reconstruction of the profiles of the same three tsunamis on the Kau coast (Fig. 9), the position of the 1896 profile is well controlled at Kaalualu, Honuapo, and Punaluu. An accepted 1933 runup is available only at Kaalualu but the upper limit at Honuapo has forced a modification of the similarity principle in reconstructing the 1933 profile in Kau. Only an accepted upper limit at Honuapo is available for the 1901 tsunami. Its profile in Kau has been reconstructed considering the ratios of its runups to those of the 1896 tsunami at Hilo and on the Kona coast.

1896, 1901, and 1933 profiles, Hilo

At Hilo (Fig. 8), accepted runups are available for the 1896 and 1933 tsunamis, but only an upper limit to the runup of the 1901 tsunamL Assuming similarity to the ratios among the runups of the tsunami in Kau, the 1901 runups would considerably exceed the upper limit. Hence the upper limit is used as a basis for the reconstructing the 1901 profile in Hilo. North of Hilo, the position of the 1896 tsunami profile is controlled by the estimated runup at Kaiwilahilahi and the estimated upper limit to the runup at Hakalau. The profiles of the 1901 and 1933 tsunamis north of Hilo are based on the similarity principle e(l) disregarding details in the 1896 tsunamf.

1896, 1901, and 1933 profiles, elsewhere

Even if interpolation between Hilo and Punaluu were appropriate, the interpolated runup of the 1896 tsunami would exceed the upper limit at Kalapana. The distance is so great that the profiles of all three tsunamis have been extrapolated downward to the east of Hilo and around Kumukaki and downward to the northeast of Punaluu. They have also been extrapolated downward to the north of Kaiwilahilahi because of the lack of any evidences of runups in the Hilo or Hamakua coasts.

The coastal distance between Hookena and Kaalualu is so great that the extrapolation criterion would be applied to the 1896 and 1933 profiles counterclockwise from. Hookena and clockwise from Kaalualu. However, assuming similarity between these profiles were reconstructed that of 1901, which is controlled by a runup estimate at Hoopuloa, the 1896 and 1933 profiles were reconstructed by simple interpolation over this distance.

Other profo.

The ratios between the runup of the 1968 Japanese tsunami at Hilo that has been estimated on the basis of marigraphic evidence and the estimated runups of the 1896 and 1933 tsunamis there would, in consideration of the 1896 and 1933 runups elsewhere and the Similarity principle, lead to the estimation of significant 1968 runups on the Kau and Kona coasts. However, the estimates would be quite speculative and would at most sites be less than or very little greater than the tenth highest runups of other tsunamis. Hence the -Hawaii runup profile of the 1968 tsunami has not been reconstructed except between Hilo and Kaiwilahilahi where the similarity principle was applied, and north of Kaiwilahilahi and east of Hilo where the extrapolation principle was applied.

No estimate of the runup of the 1952 Japanese tsunami II available for any site on the coast of Hawaii except Hilo and from the marigraphic amplitude of this tsunami of Kahului it seems very doubtful that its runup could have been significant.

Maui profiles (Fig. 10)

LOngitudinal runup profiles of the Japanese tsunamis of 1896, 1933, 1952, and 1968 are shown in Fig. 10. Each of the profiles has been reconstructed on the basis of the extrapolation principle (iii) alone.

20' 15'

20' 00'

~ ;

2-1

3-3

10 ., 10

:!! :1 ~I (:) ? ., !!!

15 a 15 10 111 20 Mlle. I j • , • I I I I

I ii' i I I I I

211.000 a 211.000 110.000 111.000 100.000 Fee.

Kawaihae

Puako

Anaehoomalu

Coastal site Identiftcations

Onsnore .. WES system: Numbered nodal pOints of Houston el aI., 1977, model.

Offsnore··HtG syslem. Runuo mao and 511e In Loomis, 1976. Number before dash is Loomis m.ep number Number 811 Sf dash is serial number of site

01 reporled run up {counted clock ••• ,

~I °1 ., ., ~ -. . . ~ ~ ., ~ ~ ~

Figure 6, Map of northern part of Hawaii showing tsunami runup sites,

Qt 20' 15' .co

20' 00'

19'

30'

19' 15

19' 00'

o o , ... \!'

~I ~I ~I 'I ' '" .. '" ., ~ ~ po

• '"

Coastal site identifications

Onshore-oWES system: Numbered nodal points 01 Houston .t 81 .. 1911, model.

Offshore·-HIG system· Runuo mao and site in Loomis, 19711. Number belore dash is Loomis map number. Number after dash is Sltrial number 01 site

01 reported runup (counted clockwi .. )

if: -t .f-q.l .f-q.l q", ~ ~ 11"

~,~4> ,,1J.~

~~ 0>:). "(,-•

• l<.~7> 1J.>:).~ ~1J. '\' i-~ ,..'Q

135

O~ ~ q~ "" q" q

Oq

II 0 II 10 15 20 Miles

... o"bv '8

1fi~ ... a..,.. .... " ... :.II

... " Qv ... "'If

o M

'" '"

Selle 1 I I , , I I' Iii i I I I I

25,000 o 25,000 50.000 75.000 100,000 Feat

~I gl '" ~

Figure 7. Map of southern part of Hawaii showing tsunami runup sites.

19' 15'

19'00'

I~

60 lUI '1411!e4 alluny

1.1) .... C"1 N

L

! " ..,

It

I l / ' I

~/ I : ! I : ! I : I I

_: i ~_I ~I"I' --I. til!!' 41:1 &1 :1 II:! " I lI' e: ! I ! ! I : I 1

: I l 1 : I [ I

, I

i :

~OI!H

\ \ \, \ \ '4el!lI eIlM,e)l

\ \ tOlleOliedna, , . \ \ '. \ \ \. \ '. , \

\ . \ \

\ \ ,

\ , \

... g

... ..

... ..

.... ...

o ...

... CO

• .! ::Ii

o ...

o

....

.! • .. <II

18'1 'HI6114 dnuny 1Il)" P1 N

\ . . \

; • ... o o q o o

o o q ... ...

o o q o ...

o o q ... ...

o

o o q .... ...

... ..

\ \ \ \

. \

, \

, \ , , .... 0\ ~, , ... \

~\ 'b\

\ .. \'.

\ .. 'J\

.. N

\ \

, , \ \

\ \

188, '1411,ell dnunll

i • ." •

i • ." •

Q,)

'" Q,) C C'II C. C'II --. -o '" ~

:.:: e c. c. :::J c: :::J ...

62

The 1896 and 1933 profiles are based on the runups estimated at Kaanapali and Lahaina, respectively, and the 1952 and 1968 profiles on the runups estimated at Kahului. Considering the similarity principle, e(i) and the runup ratios at Hilo, a high 1896 runup should have should have been expected at Kahului. This is ruled out, however, by the lack of a reported observation of the tsunami at Kahului. The upper limit of the reasonable range of 1896 runups at Miliko and Lahaina provide no guidance to the reconstruction of its profile, and any modifications of the profiles on the bases of the similarity principle would be speculative.

The upper limit of the reasonable range of runups of the 1896 tsunami at Maliko and Lahaina provide no guidance· to the reconstructing of its profile, and any modifications of the profiles on the basis of the similarity principle or any other principle would be speculative.

Kauai profiles (Fig. 11)

The reconstruction of the runup profile of the 1896 Japanese tsunami on Kauai (Fig. 11) is controlled by runup estimates at Kilauea Landing, Kapaa, and Nawiliwili, and the reconstruction of the 1933 tsunami by runup estimates at Nawiliwili, Kukulula, and Pakala.

At Nawiliwili the 1896 runup was somewhat higher than the 1933 runup. However, the same differential is improbable on the Koloa coast considering the upper limit to the range of reasonable runup values at Koloa Landing. Hence the 1896 profile has been reconstructed assuming that the 1896 and 1933 runups were identical at and west of Koloa Landing.

Between Nawiliwili and Kilauea Landing, the 1933 profile has been reconstructed assuming similarity in general but not in detail between the runup patterns of the 1933 and 1896 tsunamis.

On the north coast west of Kilauea Landing and on the south coast west of Pakala and there are no runup estimates, and the upper range of reasonable runup values provide no guidance to the reconstruction of the profiles. The extrapolation principle has therefore been applied in the reconstruction of both 1896 and 1933 profiles on these coasts.

There is no runup information available on which to base the reconstruction of Kauai runup profiles of Japanese tsunamis other than those of 1896 and 1933 and no reason to suppose that any others had runups as significant as the runups of those tsunamis.

Profiles for other islands

Japanese tsunami runup estimates on Oahu are available only at Honolulu Harbor and at Waikiki. At both places the highest of the most probable values were 2 feet (1896 at Waikiki and 1933 at Honolulu Harbor). The tenth highest Waikiki runup in the historic record to which Houston ~ al. (1977) applied their frequency analysis was 3 feet (1957 Aleutian tsunami). Hence a 2-foot runup is of no significance there in the frequency analysis. In the historic record at Honolulu, the nineth and tenth highest runups are 2.3 and 1.7 feet (1877 and 1868 South American tsunamis, respectively). Substitution of the 2.0 runup of the 1896 Japanese tsunami for the latter value would result in no significant change in the estimated frequency distribution. Hence no Japanese tsunami runup profiles have been reconstructed for Oahu.

It is possible that the 1896 tsunami, at least, had significant runups on the west coast of Niihau, on the south coasts of Lanai on Kahoolawe, or on the north coast of Molokai. However, any estimates of Japanese tsunami runups on these islands would be quite speculative, and no profiles have been reconstructed for these islands.

10

5

4 i ; 3 .!!

i. 2 .. Ii s:.

~ :I c: :I C .

! . :I

R Uflup heigh •• :

Upper limit 10 t r •• _bl. ".t ....

Rlnlle of " .. oneble ",Iuft [I M.t probeble "elue

e! Questionable value

~.9. _-

l'-" ~ ..-..- ..-"-

!

~ 15 WES Iii. noa. 10 115 20

15 0 15 10 Hi 20 Mil .. I I I , ! I I I I , Iii' i I I . I I

25.000

4« . ........ ........ ~It '.aa

........ """"

o

i ......... """", • c: · · ¥

25

215.000 50.000 75.000 100.000 Fe,t

~'1~9~''''''' _·t ...... · ---';,.--" .......

-'~ ....... ....... -............... ~~ ....... - ~ ~.~!~~~ ......... I -"'.. """"-."'" ........ '''.

30 315

"S "' ....

:r i ¥

40

Figure 10. Reconstructed runup profiles of Japanese tsunamis: Maui

10

15

4

3-: 2.i ..

•• s:.

~ :I

1 ~ c

CI) w

10

5

4

.! 3

i 2 ,!!, ., .c:

" :> c: :> II:

10

5

4

- 3 ::. ~ 2

'" 'OJ .c:

" :> c: ::I

II:

36

6 WES aite nOl,

June 188,

March 1833

i i :ao:

// /

10

f

ci ,. 0: .!! ....

" • '3 .5! oW 0 " ~ lIIC

I 40 WES site no,.

f

//

45

j ~ • ~

16

10

June 188, --T--'=I i 5_

4 ----/'

//

March 11133

t

.' Go • " • c: • :r;

'" ~ .... i .. ii!

20 26

I : 30

! I

36

5 0 5 10 15 20 Mile. I , , , , I I I I I

I I j i f Ii' I I 25.000 o 25.000 50,000 75.000 100.000 Feel

t Aunup heights:

Aan8.01 [I reaaoneble ",'u ..

Upper limit 10 t r .. lIOnabi. ""uel

MOil probabl. valu.

.! Quellionable ""ue

• .. I .lI .. • J; .. 50 1515 158 6

w

3-::. 2i

'" ' . .c:

a. " I ~

10

5

4

II:

3-eI -2.E

'" ' . s:

a. " 1 ~ II:

Figure 11. Reconstructed runup profiles of Japanese tsunamis: Kauai

CJ)

""

;6$

AC~RWi.dIDGEMENTS

The original impetus to. this study represented in the report was a question raised by Harold Loomis of the Hawaii Institute of Geographies as to the validity of certain of the reported runup AAi~91Ittae'1896 tsunami along the Kona Coast of the island of Hawaii.

The use of the microfilms of Hawaiian newspapers on file in the University of Hawaii Library has been indespensible in the study refiected in this report.

Tokutaro Hatori of the Earthquake Research Institute, University of Tokyo, has provided me with information concerning Japanese reports of the Hawaiian effects of the 1896 tsunami and with information on the sources most of the tsunamis with which theis report deals. Hideo Watanabe of the Japan Meteorological Agency also provided me with information on the source of the 1901 tsunami.

I am much indebted in addition to:

Agnes Conrad for finding and drawing to my attention information in the Hawaii State Archives that was of considerable importance in determing the effects of the 1896 tsunami,

Osama Pukuya for locating in the files of the State Survey Office a map of Kailua-Kona that was helpful in locating the sites of some of these effect;

Lt. John Nelson for supplying me with newly prepared microfilms of marigrams of the 1968 tsunami from World Data Center A for tsunami, NOAA Environmental Data Service, Boulder, Colorado; ,

Tim Lui-Kuan, Elizabeth Cunningham, and William Russell for material assistance in identifying the locations and checking the height of some of the reported runups of the Japanese tsunamis on the Kona coast;

Richard Schmidt for assistance in labeling the figures in this report; and Charlotte Kato and her assistance for typing drafts and preparing the final report for printing.

68

REFERENCES

Anonymous, 1898. The Friend. vol. 54, no. 7, pp. 53-54.

Anonymous, 1897. Summary translation of reports of the tsunami in the Hawaiian Islanck (in Japanese). Reports of the Imperial Earthquake Investigation Committee, voL 11, pp. 37-40.

Anonymous, 1977. Tsunami travel time charts for use in the Tsunami Warnin! System. National Oceanic and Atmospheric AdminIStration, National Ocean Survey, Revised e tion, June 1971.

Cox, D.C. and J. Morgan, 1977. Local tsunamis and possible local tsunamis in Hawaii. University of Hawaii, Hawaii Institute of Geophysics Report HIG 77-14, 118 pp.

Hatori, Tokutaro, 1963. On the tsunamis along the Island of Hawaii, Bulletin Earthquake Research Institute, voL 41, pp. 49-59. I

Hatori, Takutaro, 1974a. Sources of large tsunamis in southwest Japan (in Japanese). Zinshin (Journal' of the Seismological Society of Japan), ser. 2, vol. 27, pp. 10-24.

Hatori, Takutaro, 1974b. Tsunami sources in the Pacific side of northeast Japan (in Japanese), Zishin. sere 2, voL 27, pp. 321-337.

Hatori, Takutaro, 1975. Tsunami magnitude and wave source regions of historical Sanriku tsunamis in northeast Japan (in Japanese). Bulletin Earthquake Research Institute, voL 50, pp. 397-414.

Hatori, Takutaro, 1977. Field investigation of the Tokai tsunamis in 1707 and ta54 along the Shizuoka coast: (in Japanese) Bulletin Earthquake Research Institute. voL 52, pp. 407-439.

Horikawa, K., 1981. Tsunami phenomena in the light of engineering viewpoint. Report on the Chilean Tsunami. Field Investigation Committee, Chilean Tsunami, p. 136.

Houston, J. R., R. D. Carver, and D. G. Marble, 1977. Tsunami Wave Elevation Frequency of Occurence for the Hawaiian Islanck. U.S. Army Experiment Waterways EXperiment Station, Hydraulics Laboratory, Technical Report H-77-18, 83 pp. plus plates and appendix.

lida, K., 1963. Magnitude, energy, and generation mechanism of basins, and catalogue of earthquake associated with tsunamis. proceedi~ of Tsunami Meetings Associated with the Tenth Pacific Science Congress. Internat. Union 0 Geology and Geophysics Mono. 24, pp. 7-18.

lida, K., D. C. Cox, and G. Pararas-Carayannis, 1987. Preliminary Catalog of Tsunamis Occuring in the Pacific Ocean. Hawaii Institute of Geophysics, Data Report 5, HIG-87-10.

Imamura, A. and M. Morlya, 1939. Mariographic observations of tsunamis in Japan during the period. Jap~n Jour. of Astronomy and Geophysics. voL 17, p. 125.

Jaggar, T. A., 1933. Tsunami or earthquake tidal wave, March 2, 1933. Volcano Letter (Hawaiian Volcano Observatory), no. 397, pp. 1-2.

Loomis, H.G., 1976. Tsunami Wave Heights in Hawaii. Hawaii Institute of Geophysics Report HIG 76-5.

Macdonald, G. A., 1952. The south Hawaii earthquakes of March and April, 1952. Volcano Letter (Hawaiian Volcano Observatory), no. 515 (January - March, 1952), p. 7

Murphy, L. and W. Cloud, 1954. U.S. Earthquakes. 1952. U.S. Coast and Geodetic Survey, sere 782, 42 pp.

Pararas-Carayannis, G., 1989. Catalog of Tsunamis in the Hawaiian Islanck. U.s. Coast and Geodetic Survey. World Data Center A: Tsunami, WDCA-T 89-2. 94 pp.

Pararas-Carayannis, G., and J.P. Calebaugh, 1977. Catalog of Tsunamis in Hawaii (Rev. ed). World Data Center A: Tsunami, NOAA Environmental Data ServIce Data Center A for Solid Earth GeophysiCS, Report SB-4, 78 pp.

!

67

Shepard, F. P., G. A. Macdonald, and D. C. Cox, 1950. The tsunami of April 1, 1946. Bulletin Scripps Institute Oceanography. vol 5, no. 6., pp. 391-528.

Thrum, T. G. , 1897. Retrospect of 1896. Hawaiian Annual voL 23, p. 131.

Watanabe, H., 1968. Descriptive table of tsunamis in or near Japan (in Japanese). Zishin (Journal of the Seismological Society of Japan), sere 2, voL 21, pp. 293-313.