estimation of the magnitudes and epicenters of philippine historical earthquakes

Tectonophysics 317 (2000) 137–169www.elsevier.com/locate/tecto

Estimation of the magnitudes and epicenters ofPhilippine historical earthquakes

Maria Leonila P. Bautista a,*, Kazuo Oike ba Philippine Institute of Volcanology and Seismology (PHIVOLCS), C.P. Garcia Ave., UP Campus Diliman,

Quezon City 1100, Philippinesb Department of Earth and Planetary Sciences Graduate School of Science Kyoto University Sakyoku, Kyoto 606-8502, Japan

Received 17 November 1998; accepted for publication 29 October 1999

Abstract

The magnitudes and epicenters of Philippine earthquakes from 1589 to 1895 are estimated based on the review,evaluation and interpretation of historical accounts and descriptions. The first step involves the determination ofmagnitude–felt area relations for the Philippines for use in the magnitude estimation. Data used were the earthquakereports of 86, recent, shallow events with well-described effects and known magnitude values. Intensities are assignedaccording to the modified Mercalli intensity scale of I to XII. The areas enclosed by Intensities III to IX [A(III ) toA(IX )] are measured and related to magnitude values. The most robust relations are found for magnitudes relatingto A(VI ), A(VII ), A(VIII ) and A(IX ).

Historical earthquake data are obtained from primary sources in libraries in the Philippines and Spain. Most ofthese accounts were made by Spanish priests and officials stationed in the Philippines during the 15th to 19th centuries.More than 3000 events are catalogued, interpreted and their intensities determined by considering the possible effectsof local site conditions, type of construction and the number and locations of existing towns to assess completenessof reporting. Of these events, 485 earthquakes with the largest number of accounts or with at least a minimumreport of damage are selected. The historical epicenters are estimated based on the resulting generalized isoseismalmaps augmented by information on recent seismicity and location of known tectonic structures. Their magnitudesare estimated by using the previously determined magnitude–felt area equations for recent events.

Although historical epicenters are mostly found to lie on known tectonic structures, a few, however, are found tolie along structures that show not much activity during the instrumented period. A comparison of the magnitudedistributions of historical and recent events showed that only the period 1850 to 1900 may be considered well-reportedin terms of magnitude distribution. Each earthquake is evaluated for its ‘quality’ of determination based on thenumber of intensity reports. Earlier than 1850, the data collected are few and most earthquakes had fewer than tenreports. Good quality reports began to be collected from 1850, partly correlative to an increase in the number oftowns and partly to the start of a systematized collection of earthquake accounts by the Manila Observatory.Parameters of these well-reported earthquakes may be used for conducting various seismological studies. Examplesof how the parameters of poorly reported events were arrived at are also discussed. © 2000 Elsevier Science B.V. Allrights reserved.

Keywords: earthquake catalog; felt areas; historical earthquakes; historical epicenters; Philippine earthquakes; seismic hazard

* Corresponding author. Fax: +63-2-926-3225.E-mail addresses: [email protected] (M.L.P. Bautista), [email protected] ( K. Oike)

0040-1951/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved.PII: S0040-1951 ( 99 ) 00272-3

138 M.L.P. Bautista, K. Oike / Tectonophysics 317 (2000) 137–169

1. Introduction and other international traders who conductedcommerce in various parts of the archipelago. Itwas only during the Spanish rule in the PhilippinesVarious earthquake studies depend on the exis-

tence of a robust earthquake catalog. Among these that earthquake reports began to be maintained invarious letters and chronicles. The number ifare studies on seismic hazard and risk analysis,

recurrence intervals and rates, seismic gaps, statis- accounts increased with the development of settle-ments, improved means of communication, andtical analysis, slip rates and studies characterizing

tectonic activity. In the Philippines, the instrumen- the construction of masonry structures to whichdamage was usually ascribed. Although Spain ‘dis-tally derived seismic database only encompasses a

100 year period, whereas a potentially rich source covered’ the country in 1521, actual colonizationbegan only in 1565. The colonization mission wasof additional earthquake data can be gathered

from at least 300 years of written records. During geared towards, among other aims, the conversionof native Filipinos to the Roman Catholic religion.the last 25 years, a significant volume of informa-

tion has been added to the knowledge of the The construction of massive stone churches, some-times the only masonry structure in a town, wasseismotectonics of the Philippines. For example,

there are now research results leading to the under- one of the methods used by the colonizers tofurther the spread of the Catholic faith. Thestanding of the activities of tectonic structures

using seismicity, including the determination of different religious groups spreading their respectivefaiths also managed to keep records of theirthe earthquake focal mechanisms (Fitch, 1970;

Rowlett and Kelleher, 1976; Seno and Kurita, churches’ histories, and these included the timesthat natural phenomena, such as typhoons, vol-1978; Acharya and Aggarwal, 1980, Cardwell

et al., 1980; Hamburger et al., 1983; Lewis and canic eruptions and earthquakes, affected theirrespective structures. Hence, the review of histori-Hayes, 1989; Bautista, 1996a). Results of various

geological (Nakata et al., 1977, 1996; Hirano et al., cal earthquakes in the Philippines is one involvedwith being cognizant of the chronological develop-1986; Suppe, 1988; Pinet and Stephan, 1990;

Ringenbach et al., 1990; Deffontaines et al., 1993) ment of towns and communities that are signifi-cantly centered on a particular church unit.and geophysical studies that included marine

seismic profiles, global deformation system, aero- In this study, we estimate the epicenters ofPhilippine historical earthquakes and their magni-magnetism and paleomagnetism (Ludwig, 1970;

Lewis and Hayes, 1983; Wolfe and Self, 1983; tudes using areas of felt intensities. We furtherpresent additional discussions on how we arrivedHayes and Lewis, 1984; Cole et al., 1989; Bischke

et al., 1990; Duquesnoy et al., 1994) can help at the parameters of sparsely reported events.illuminate the locations of faults and subductionzones. Some recent earthquakes have been studiedin detail (Su, 1969; Morante, 1974; Stewart and 2. Previous studiesCohn, 1977; Acharya, 1978; Solidum and Sabit,1988; Umbal et al., 1990; Bautista et al., 1991; In the Philippines, five catalogs of historical

earthquakes exist: that of the work of PerreyShibutani et al., 1991; PHIVOLCS, 1994), andthose results have added to the understanding of (1860), Maso (1895, 1927), Repetti (1946) and

SEASEE (1985). Perrey’s work, which was writtentheir source regions. Recently, some workers havealso started digging trenches to learn about paleo- in French, described earthquakes and volcanic

eruptions in the Philippines from the 15th to 18thearthquakes (USGS, 1995; Daligdig, 1997). Allthese data could be used for analyzing the histori- centuries. Maso’s work in 1895 was in the Spanish

language and was more comprehensive than thatcal earthquake accounts with new perspectives.The Spanish Empire colonized the Philippines of Perrey (1860) but mostly dealt with earth-

quakes. It cited accounts especially of large earth-in the 15th century. Before this, no description oraccount of any earthquake has yet been found, quakes from 1589 to 1899. Maso also derived data

from the Manila Observatory when it started itsalthough the country was in contact with Asian

139M.L.P. Bautista, K. Oike / Tectonophysics 317 (2000) 137–169

operation in the 1860s using a simple pendulum done using only magnitude-to-maximum epi-central intensity (Io) relations (M–Io). Bautistaand early seismographs. In his 1895 catalog, Maso

also drew isoseismal maps of selected large earth- (1993) estimated the magnitudes of 50 historicalPhilippine earthquakes using an Ms–Io relationquakes using the Manila Observatory Intensity

Scale of I to VI. In 1927, Maso again published and got Ms values of 6.9, 7.4, 7.9 and 8.4 for Iovalues of 6, 7, 8 and 9 (using the adapted Rossi–another catalog, written in the English language.

The contents were basically the same as his 1895 Forel intensity scale of I to X ) respectively.Thenhaus et al. (1994) also used an Ms–Io relationpublication except that smaller events were not

included. After 24 years, Repetti (1946), who was to determine the surface magnitudes of severalhistorical earthquakes using the modified Mercallialso from the Manila Observatory, published the

most comprehensive historical catalog of the intensity (MMI) scale of I to XII. Using Io toestimate magnitudes poses a problem, since inten-Philippines. He had an advantage over Perrey and

Maso because he was able to see their catalogs sity reports may be affected by many factors,including population density, earthquake depthsand was able to crosscheck their references. Repetti

(1946) also provided his comments and insights to and site response. Another drawback is when theestimated epicenter is located offshore. Such casesthe work of the two earlier authors. In addition,

his work also contained more bibliographic refer- are common in the Philippines since it is sur-rounded by seas on all its sides.ences and was better researched. After 39 years,

SEASEE (1985) published another Philippine To overcome the uncertainties in magnitudeestimates arising from the use of magnitude–inten-earthquake catalog that merged Repetti’s historical

data and instrumental data from the Manila sity relations, recent investigators have startedemploying an approach relating magnitude withObservatory, the Philippine Atmospheric, Geo-

physical, Astronomical Services Administration felt areas. Nuttli and Zollweg (1974) related bodywave magnitude with felt areas of 22 earthquakes(PAGASA) and reports of investigations of

post-19th century earthquakes. All previous that occurred from 1962 to 1972 in the centralUnited States. They found that the log of the feltstudies, however, merely concentrated in chroni-

cling the accounts in detail and assigned neither area (Alog) and magnitudes (body and surfacemagnitudes) could be related using second-degreemagnitude nor epicenter systematically.

Some early seismologists (Mallet, 1853, 1854a, polynomial fits, and they suggested that theseequations could be used to estimate magnitudes of1854b; Milne, 1912) also made worldwide catalogs

that included historical Philippine earthquakes. 19th century earthquakes in the same region. Nuttliet al. (1979) determined the relationship of bodyMore recently, Dunbar et al. (1992) listed epicen-

ters, magnitudes and damage attributable to wave magnitudes versus affected area (Af ) andalso of log of areas inside Intensity IV isoseismalPhilippine earthquakes from 1599 to 1990. Utsu

(1990) also listed historical earthquakes of the (AIV ) using 41 earthquakes that occurred in west-ern and central USA. Street and Lacroix (1979)Philippines, mainly derived from the SEASEE

(1985) catalog. Some authors listed tsunamigenic studied 37 recent earthquakes in central northeast-ern USA and derived quadratic least squares fitsearthquakes (Rudolf, 1887; Heck, 1947; Iida et al.,

1967; Berninghausen, 1969; Cox, 1970; Nakamura, between magnitudes of earthquakes with total feltareas (At) between 10 000 and 100 000 km2 using1978) that affected the Philippine shores. Although

many early and even recent authors have multiple regression relationships between magni-tude, felt areas and fall-off of intensity. Tuttle andattempted to catalog Philippine historical earth-

quakes, none, except for Bautista (1996b, 1999), Sykes (1992) calculated magnitudes of severallarge historic earthquakes along the San Andreashas attempted to infer the sizes and source regions

of these pre-19th century earthquakes using felt fault and inferred the magnitudes of three histori-cal earthquakes by relating magnitudes and feltareas.

Meanwhile, previous studies on estimating mag- areas of recent earthquakes along the same seg-ment of the fault. Meanwhile, Ebel (1996) esti-nitudes of historical Philippine earthquakes were


mated the magnitudes of historical earthquakes in idea of the prevailing social history when theearthquake occurred. The first earthquakenortheastern North America using intensity

reports and aftershock characteristics. Bollinger reported in the Philippine was in 1589. Prior tothis time, there are no known confirmed reportset al. (1993) related the moment magnitudes of

109 earthquakes for earthquakes in the USA and of earthquakes from the Philippines. The tsunamiin the Philippines in 35 BC mentioned byCanada with felt areas inside Intensity VI and VII

isoseismals using least squares fits and found that Lockridge (1988) has since been found to beerroneous (Patricia Lockridge, National Oceanicearthquakes with magnitudes from 4.5 to 7.5

occurring in the eastern USA have five times bigger and Atmospheric Administration, written com-munication, 1994).felt areas than those occurring in the western USA.

These studies all suggest that equations relating When the Spaniards came they started to orga-nize isolated and small Filipino communities formagnitudes and felt areas are site dependent, and

different areas must be characterized by different various purposes, including the facilitation of taxor tribute collection and to spread their new reli-empirical relationships. Hence, in this study, we

have derived our own empirical relations for the gion. Philippine towns started to sprout by thelate 15th to early 16th century. In the center ofPhilippines using recent Philippine earthquake

data. these towns were found the local governmentbuildings, schools and the church. By the end ofThis paper shall serve as the first attempt to

determine the parameters of Philippine historical the Spanish rule, in the late 18th century, manyPhilippine towns had their own Spanish church.earthquakes using felt areas. Aside from deviating

from the use of Io to estimate magnitudes, it is Hence, in this context, we now can focus on thegeneral conditions of the country when earth-also an improvement on the previous work of

Bautista (1993, 1996b) as it evaluates more histori- quakes struck. Fig. 1 shows the flow of informationwhen an earthquake occurred during the Spanishcal earthquake data. The quality of determination

of each event shall also be classified to differentiate time. When a damaging earthquake occurred, twopersons usually wrote a report about it, the mayorwell-estimated from poorly estimated historical

epicenters. This is important to make future users (called gobernadorcillo) and the local priest. Themayor sent his report to the governor of theaware of their accuracy, to make sure that these

quality assessments become an integral part of the province, who, in turn, submitted it to the centralgovernment based in Manila, where the Governor-historical catalog and also to prevent the possible

indiscriminate use of the data without future users General was based. At times, these reports reachedthe Spanish king, especially when it involved mone-knowing how really accurate the parameters are.

Musson (1998) had challenged whether it is, at tary losses for the colonial government. For areaslike Mindanao, which was under stricter political–all, advisable to parameterize these historical

events precisely for the reason that future users military rule until the end of the Spanish era owingto its refusal to be converted to the Christiantend to get only the resulting data without con-

sidering their accuracy. Since we want to avoid religion, another report was prepared by the localmilitary commandant; this was first forwarded tosuch tendencies, we urge that rankings we have

attached to each event be considered carefully by the governor in Mindanao before being relayed tothe government in Manila. Aside from these, therefuture users.were also reports made independently by travelersand historians and which are also found in historybooks. In our review, we have never found any3. Overview of the social history during the Spanish

era in the Philippines in relation to earthquake report written by a native Filipino or from aFilipino’s point of view. Aside from official reports,reportingaccounts were usually written from a foreigner’sviewpoint and consist of descriptions, impressionsBefore one can proceed with evaluation of

historical earthquakes, it is important to have an and ideas.


Fig. 1. Flow of how earthquake information was relayed during the Spanish period in the Philippines.

Another factor that may be important to take reports even for strong earthquakes. Sketchy anduncertain reporting of early earthquakes com-note of is the lack of an efficient means of com-

munication. Until the end of the Spanish era there pound this problem. The factors discussed inSections 4.1–4.7 were considered in the determina-were still isolated villages in areas like in northern

Luzon and central Mindanao and in many of the tion of epicenters.7100 islands of the country. Therefore, unless, itwas a big event, small earthquakes were usually 4.1. Location of faults and subduction trenchesnot reported from these places because of thedifficulty of relaying information. One of the first and important steps is the

identification of tectonic structures to which earth-quakes could be attributed. During the last twodecades a significant leap in our understanding of4. How the historical epicenters were determinedthe Philippine seismotectonics was brought aboutby results of new research detailing informationIn most cases, the epicenter is assumed to be

located in the area of highest intensity. However, on active faults and subduction trenches. A thor-ough review of the tectonic setting of thesome actual epicenters have been as far as 100 km

from the area of most intense damage, as in the Philippines, including the locations of known faultsand trenches, was undertaken during the early partJuly 16, 1990 earthquake (Ms 7.8) of Luzon. In

some cases, the presence of tectonic structures and of this study. This knowledge is particularly impor-tant in deciding which historical earthquake isrecent seismicity were used to decide the possible

location of a historical epicenter. In drawing the related to which source from among the differentearthquake source zones. The results of the reviewisoseismals, generalized maps were done. This is

because seas bound the Philippines on all sides, of Philippine tectonics, seismicity and focal mecha-nism data were used to delineate the active faultsmaking it difficult to get well-distributed intensity


and subduction zones. The location of trenches Survey (USGS) was also used to delineate thefaults. The resulting map was used as the basewas further improved using digital 2◊×2◊ bathy-map in drawing the isoseismal maps.metric data from the Scripps Institution of

Oceanography (SIO). The data were downloadedfrom the anonymous ftp site (topex.ucsd.edu). The 4.2. Locations of recent earthquakesdescription on how the database was derived isfound in Smith and Sandwell (1996). In addition By plotting recent seismicity (Fig. 2), the loca-to new tectonic information, a 30◊×30◊ digital tions of active structures (Fig. 3) could be con-

firmed. Recent earthquake data were derived fromelevation model obtained from the US Geological

Fig. 2. Seismicity map of the Philippines. Plot of recent, shallow earthquakes (magnitude 5 and above) from 1960 to 1995. Thelocations of recent earthquakes were considered in the determination of historical epicenters. [Earthquake data from the NationalEarthquake Information Center (NEIC), International Seismological Centre (ISC) and Harvard University.]


Fig. 3. Tectonic map of the Philippines. The locations of active faults and subduction zones were delineated based on recent seismicity,focal mechanism data, topographic and bathymetric data. Heavy, hachured, saw-toothed lines represent subduction zones. Lines arestrike-slip faults, whereas small, saw-toothed hachured lines are thrust faults with teeth towards the dip direction. Dashed lines arecollision zones. (The digital bathymetry dataset is from the SIO and digital topographic dataset is from the USGS).

the National Earthquake Information Center and 4.3. Completeness of reportingthe International Seismological Centre. Separatecatalogs by Gutenberg (1956) and Gutenberg and One aspect in the evaluation of historical earth-

quakes is the assessment of the completeness ofRichter (1954), Seismological Notes of the Bulletinof Seismological Society of America, Manila reporting. To achieve this, it is important to know

the locations of reporting towns. Dates when townsObservatory, PAGASA (1976) and SEASEE(1985) were also reviewed and included in the started to be organized provided information on

where reports could be expected to come fromrecent earthquake catalog.


during an earthquake. During the Spanish colonial were used to pinpoint the existence of settlementsrelative to the occurrences of specific earthquakes.rule in the Philippines, towns developed from a

particular church unit. The Spanish settlers orga- This allowed for an assessment of the completenessof reporting and made the non-reporting of inten-nized the settlements they found, and they made

it a point to build a church in these thriving sities as equally important as the presence ofintensity reports. Hence, a non-reporting couldcommunities. It is difficult to determine the chrono-

logical development of these settlements as they mean a low intensity or that damage is non-existent or not so significant in a particular town.evolved into established towns. The dates when

churches were built, which were readily available, Aside from a church, other structures that may be

Fig. 4. The locations of Philippine churches from 15th to 19th centuries are shown by the white squares. Early Philippine townsdeveloped from a church as a central focus. The locations of early churches were used to pinpoint locations of early Philippinetowns that, in turn, were used to assess the completeness of reporting of earthquake effects.


found in a typical, early Philippine town, or plaza use of tile roofs was banned; instead, galvanizediron sheets were used. This type of roofing contin-as it was then called, were a government house,

school, tribunal house and other important govern- ues to be used to the present day. Hence, it canbe seen that only by understanding the historicalment offices. Fig. 4 shows the location of these

15th to 19th century church sites. It can be seen context and scenario when these historical earth-quakes occurred could their proper evaluationthat the northern Philippines had a greater number

of churches compared with the southern part, be done.which was an area that was more successful inresisting Spain’s colonization efforts and remained

4.5. Effects of site conditionsmostly under Muslim rule. This distribution con-tributed to controlling the distribution of historical

Using the geologic map published by the Southepicenters, since the reports used in this study

East Asia Association of Seismology andcome mainly from church and government records.

Earthquake Engineering (SEASEE, 1985), thegeneral locations of Quaternary alluvium depos-its — sites of soft sediments — were delineated.

4.4. Types of construction during the pre-20thIntensity reports from towns underlain by soft

century Philippinessediments were carefully assessed for the possibleeffect of site amplification. This possibility was

Filipinos have always lived in wooden structuresalso carefully considered in evaluating both histori-

consisting of bamboo with nipa or cogon grass ascal and recent earthquakes. For example, the towns

roofing materials. This type of construction canof Pangasinan, especially those along the river

resist strong shaking caused by strong earthquakes.channels, were severely affected by liquefaction

This could have been another reason for theduring the 1990 Luzon earthquake despite being

absence of earthquake reports before the arrivallocated several tens of kilometers away from the

of the Spaniards, besides the possible destructionepicenter.

of pre-Spanish rule native writings. The earlystructures, including churches, were made of localmaterials, like nipa and cane (Liporada and 4.6. Other factors considered in evaluating

historical accountsAtanacio, 1988). When the Spanish authoritiesand priests started building tall and masonry struc-tures, the damage began to be documented. In a Factors such as possible exaggerations, biases,

misquotes, changes of names and jurisdiction, andtypical town, the first masonry structure that wasbuilt was the church. A church usually began as a local history were also taken into consideration

when reviewing the archives for historical earth-structure made of wood similar to the local houses.As years went by, the churches were improved or quakes. For example, a mountain was said to have

been leveled during the Northern Luzon earth-rebuilt by the use of local materials like adobe andlime. Really massive Spanish era churches are still quake of 1627. In later accounts, the number of

mountains leveled has increased to two (Perrey,found in many towns of the Philippines. Some ofthese may be as long as 70 m and as wide as 80 m 1860; Repetti, 1946). The number of documented

casualties sometimes also depends on the historian;(National Media Production Center, 1980). Atypical church had a belltower and an adjacent oftentimes only Spaniards were counted, whereas

any ‘person of no account’ was disregarded, as inrectory or chapel. In Manila, local houses duringthe 18th century were made of wood but the use the 1645 earthquake. The transfer of information

from one historian to another also led to mis-of tile roofs became prevalent. This was the situa-tion when the June 3, 1863 earthquake occurred. quoted names. For example, the place ‘Cagayan’

was later miswritten as ‘Camarines’ (Nieremberg,The strong shaking caused the heavy tile roofs tocave in and was mainly the reason for the heavy 1635; Perrey, 1860; Repetti, 1946) when the 1627

earthquake was described, creating a confusiondeath toll in Manila. After the 1863 quake, the


between two places which are about 400 km apart. not fall but merely get damaged, they are still inthe Intensity VII level. Intensity VIII level isChanges of territorial jurisdiction compared with

present times also presented some minor confu- considered to result in some limited parts of astructure collapsing, but the serious collapse of asion. For example, many historians referred to a

place called Nueva Segovia, which at present is masonry structure ( like at least half of the struc-ture) is deemed to be at Intensity IX level. Mostnot the name of any of the present-day towns. A

further search showed that the former Nueva of the early churches were poorly built, exceptthose found in more developed areas like ManilaSegovia is now named Lal-lo.and the surrounding provinces. Construction laterimproved in terms of sturdiness and solidness, and4.7. Techniques in intensity evaluationthis fact is oftentimes mentioned in earthquakeaccounts or church histories. Since much impor-For intensities when no damage is reported, keytant information on the state of masonry structureswords were searched for in the accounts to aid infor each town is not known with certainty when aintensity assignment. As a general rule, thespecific earthquake occurred, structures were allfollowing phrases and their corresponding inten-assumed to be poorly built unless information onsities were searched for:their structural design is available and could beevaluated. From Intensity X to XII, the descrip-tions in the MMI scale are more about the effectsPossible description Intensity assignedon the environment and the extent to which

Not felt I destruction took place. Hence, if damage is judgedVery slightly felt II to be quite extensive, including the occurrence ofSlightly or weakly felt III a tsunami, extensive fissuring or ground cracking,Felt; moderately or regularly felt IV

then intensities may be raised to these highestStrongly felt Vvalues. Reports of liquefaction were carefully eval-Very strongly felt VIuated, as liquefaction is more dependent on groundconditions, and were used with caution in evaluat-ing earthquake intensities.The Intensity VI level is also the boundary

between the non-damaging and damaging inten-sities. When slight damage is mentioned, unlessother observations or aggravating factors are men-tioned, Intensity VI is assigned. Meanwhile, to 5. Isoseismal maps of the historical earthquakessimplify assignment of intensities from VIIupwards, we have also studied the differences in Using all the gathered information for each

individual event, isoseismal maps were drawn. Thethe descriptions in the structural damage betweenthe damaging intensities from Intensity VII to IX. drawing software used is the Generic Mapping

Tool version 3 (GMT 3.0). A terrain map wasTo differentiate between Intensities VII and VIII,for example, we found that ‘significant damage’ used to facilitate the relating of historical accounts

with topography and possible site effects. The basebegins to be reported at Intensity VII level. At thislevel, too, less serious cracks that are repairable map also includes the location of tectonic struc-

tures that were reviewed. For each event, themay appear and structures do not collapse exceptfor the fall of loosened bricks, tiles, cornices or locations of towns that were already in existence

when the earthquake occurred are also plotted toweak chimneys (Wood and Neumann, 1931).Serious fall of tall structures such as the ‘‘twisting, guide in assessing the completeness of reporting.

For each town, the evaluated intensity is affixedfall, of chimneys, columns, monuments, alsofactory stacks, towers’’ (Wood and Neumann, on the map. Then, a generalized isoseismal map is

drawn for all earthquakes that may be evaluated.1931) may occur at Intensity VIII level. Usingthese, we assumed that when church belltowers do Although many earthquakes had numerous


Fig. 5. Map of the July 18, 1880 (12:40 p.m.) earthquake, the historical earthquake that had the largest number of reports. Romannumbers are intensities. Squares represent locations of towns existing at that time. Solid circle is the epicenter estimated at 14.5 NLat., 121.6 E Long. Felt areas (in square kilometers) at Intensities X to V and their calculated Ms values are: A( X )=5715 sq. km(Ms 8.0); A(IX )=21 050 sq. km (Ms 7.9); A(VIII )=35 544 sq. km (Ms 7.8); A(VII)=63 314 sq. km (Ms 7.5); A(VI)=150 116 sq. km (Ms 7.3); A(V)=359 521 sq. km (Ms 7.1). Ms (ave) is 7.6.


Table 1Parameters of recent (1911–1995) earthquakes used in determining Ms–felt area relations. Earthquake parameters are from theNEIC. Time is in Greenwich Meridian Time (GMT). Areas in square kilometers. A(III )…A(IX ) are the areas inside Intensities IIIto IX

Date E. N. Origin time Depth Ms A(IX ) A(VIII ) A(VII ) A(VI) A(V ) A(IV ) A(III )Long. Lat. (GMT ) (km)

6/10/69 121.408 13.232 17:15:29 37 4.6 3062 13 743 36 344 84 4729/6/78 122.549 10.371 14:46:38 53 4.8 30 456 66 267 126 361 235 9057/4/71 121.876 15.604 11:30:51 30 5.1 14 922 95 568 240 8915/17/92 126.882 6.946 15:36:21 41 5.1 24 609 101 645 244 4358/10/68 121.575 15.504 16:41:25 33 5.2 9536 49 1968/12/84 120.811 18.394 16:51:20 13 5.2 42345/23/87 125.410 8.047 17:09:04 32 5.2 27 1746/13/77 124.645 13.309 11:47:41 35 5.3 9916 50 447 130 942 449 74312/17/80 126.192 7.969 11:41:45 40 5.3 9894 58 0022/25/87 121.080 18.808 15:20:18 16 5.3 12 722 21 842 34 8125/1/70 121.833 15.666 3:22:13 29 5.4 19 460 52 130 113 4374/12/79 122.543 10.543 15:37:59 48 5.4 7954 23 818 104 57412/27/79 124.502 13.954 18:23:03 33 5.4 10 780 43 715 109 2808/11/83 120.957 18.828 12:02:59 14 5.4 66382/27/84 121.273 19.216 18:06:34 34 5.4 48 380 157 6941/3/85 121.172 19.134 3:37:20 52 5.4 27 300 86 88511/9/75 125.098 13.793 20:34:50 33 5.5 27 417 89 211 194 2712/13/76 121.703 15.670 8:07:33 47 5.5 8868 50 412 90 0199/13/83 126.625 7.529 2:03:27 46 5.5 19 0044/5/88 120.393 13.295 15:36:57 30 5.5 42 0043/26/90 125.606 9.253 22:47:17 39 5.5 29 295 145 0727/23/90 122.348 11.269 18:02:51 16 5.5 16 085 54 422 169 8325/6/91 125.253 10.353 0:26:13 32 5.5 17 311 130 0467/2/71 123.892 12.266 5:34:27 35 5.6 4136 18 350 47 156 93 4708/25/76 124.445 13.045 12:29:54 23 5.6 27 378 44 867 73 7565/6/83 121.711 15.439 18:24:18 33 5.6 24 734 144 8779/22/84 120.923 18.424 21:18:58 46 5.7 9784 56 058 189 98011/27/84 124.388 7.601 23:34:01 11 5.7 22 314 75 740 215 9966/19/61 121.970 12.720 1:45:27 56 5.75 18 719 57 59011/22/68 122.332 16.275 8:59:23 26 5.8 23 054 127 4842/11/87 120.981 18.831 6:12:53 23 5.8 12 450 39 508 74 6085/8/88 120.135 14.942 19:44:55 49 5.8 9080 44 548 30 44748/3/68 122.309 16.479 6:25:06 37 5.9 156 06112/28/82 121.397 19.945 13:49:29 34 5.9 49 245 213 6265/18/87 125.362 8.302 7:27:00 16 5.9 2162 5594 11 882 22 353 44 1476/22/65 123.500 7.200 23:48:07 56 6 14 577 32 477 62 414 113 5314/15/70 122.665 15.069 13:14:21 12 6 106 1767/5/73 124.699 13.225 22:46:16 38 6 11 206 46 154 108 722 172 831 298 2926/18/87 121.356 17.291 10:01:07 42 6 8227 61 406 260 3568/28/68 122.012 15.550 20:42:17 15 6.1 40 6304/24/85 120.815 16.498 1:07:15 33 6.1 1452 11 993 62 6758/4/85 123.500 7.496 2:36:24 35 6.2 48 258 93 67411/13/70 123.973 11.946 14:16:18 15 6.3 34 523 86 1847/25/71 123.685 12.363 12:51:42 33 6.3 24 049 81 732 222 5903/31/80 121.962 16.130 12:41:48 43 6.3 4256 36 137 146 042 285 43811/18/87 124.770 12.845 16:27:05 22 6.3 32 3542/3/69 127.384 4.901 21:41:42 33 6.4 273 782 811 7996/7/76 124.829 14.087 7:36:55 33 6.4 2922 22 284 75 163 180 886 765 0166/19/88 121.067 12.376 19:52:07 16 6.4 2850 17 836 81 118 184 164 374 50411/13/91 126.371 8.361 12:13:02 35 6.4 40 414 69 042 109 982 181 756


Table 1 (continued )

Date E. N. Origin time Depth Ms A(IX ) A(VIII ) A(VII ) A(VI) A(V ) A(IV ) A(III )Long. Lat. (GMT ) (km)

11/22/81 120.839 18.752 15:05:21 24 6.5 13 870 55 082 133 625 241 770 365 4108/17/83 120.860 18.231 12:17:56 28 6.5 17 181 46 720 169 164 568 7302/5/70 122.117 12.598 22:05:58 11 6.6 10 592 32 109 85 660 151 49211/17/88 124.537 12.339 6:55:46 19 6.6 6538 38 1862/8/90 124.694 9.775 7:15:32 25 6.6 6212 84 947 141 1122/18/91 126.480 8.870 2:37:25 23 6.6 251 752 447 7685/11/93 126.570 7.219 18:26:51 59 6.6 34 6943/12/15 124.000 12.000 14:48:30 40 6.6 8038 34 370 266 5806/18/80 126.657 9.475 17:14:54 54 6.8 97 9295/22/72 122.290 16.599 6:04:00 34 6.9 124 530 214 4647/21/77 122.361 16.882 13:45:54 33 6.9 2264 29 886 67 108 146 9744/12/70 122.052 15.064 4:01:44 24 7 57 1683/18/77 122.327 16.773 21:43:52 37 7 42 779 118 404 242 991 424 2232/24/88 124.616 13.477 3:53:03 24 7 2136 120 363 421 1205/6/24 119.000 16.000 16:19:30 T 7 1183 5007 8178 23 687 66 936 216 2486/15/28 121.500 12.500 6:12:36 T 7 18 151 79 440 212 7881/11/82 124.358 13.752 6:10:06 40 7.1 38 014 169 440 356 776 538 1086/14/90 121.899 11.760 7:40:56 18 7.1 2854 20 108 56 480 122 903 195 636 369 188 643 2405/17/92 126.645 7.239 9:49:19 32 7.1 4659 16 848 61 755 278 335 745 47811/14/94 121.087 13.532 19:15:31 33 7.1 862 8524 26 710 78 421 165 596 337 9884/25/72 120.309 13.370 19:30:09 50 7.2 12 250 63 880 157 55310/31/75 125.993 12.540 8:28:03 50 7.2 99 300 277 954 621 494 1 125 5083/2/23 124.000 6.500 16:48:52 T 7.2 24 156 50 062 111 372 209 0876/13/29 127.000 8.500 9:24:34 T 7.2 2452 8546 72 345 187 674 357 658 558 4798/1/68 122.201 16.522 1:19:22 37 7.3 11 122 42 270 65 890 248 908 594 6764/7/70 121.717 15.761 5:34:06 37 7.3 8143 75 404 202 007 418 8383/17/73 122.787 13.372 8:30:52 33 7.3 8089 14 916 26 530 48 086 138 571 311 07212/15/89 126.729 8.337 18:43:45 24 7.3 14 092 81 997 192 468 382 723 669 3874/21/95 125.800 12.200 0:30:10 23 7.3 24 194 60 604 178 858 371 656 567 67711/13/25 125.000 13.000 12:14:45 T 7.3 3251 8307 28 500 82 574 253 106 567 64012/2/72 126.601 6.473 0:19:47 33 7.4 55 266 246 260 497 110 816 5287/12/11 126.000 9.000 4:07:36 50 7.5 20 454 48 490 101 955 194 920 377 188 677 9552/14/34 119.000 17.500 3:59:34 T 7.6 6838 32 786 91 632 196 000 416 7367/16/90 121.172 15.679 7:26:35 25 7.8 8413 28 345 87 5088/16/76 124.023 6.262 16:11:10 33 7.9 10 616 55 767 174 815 316 524 503 488 758 523 1 763 5584/14/24 126.500 6.500 16:20:23 T 8.3 57 726 158 934 330 670 1 028 780 2 185 551 3 847 430 6 248 854

reports, some had only one or two reports; in spite was done for each of the 485 historical eventsstudied.of this, they were also evaluated if some damage

was attributed to those places and if the reportswere proximal to an established active structure,e.g. like the Philippine Fault or Philippine Trench.The large uncertainty due to such under-reportedevents is addressed in the succeeding sections. 6. How magnitudes were determined

Fig. 5 shows the isoseismal map of one of thewell-described and largest historical earthquakes, After drawing the isoseismal maps, their magni-

tudes were estimated using the newly determinedthe July 18, 1880 earthquake of Luzon. Thisearthquake had more than 100 intensity reports. magnitude–felt area relations, as described in

Section 7.The same style of drawing the isoseismal maps


(a) (b)

(c) (d)

Fig. 6. Ms versus felt area relations for Intensities III to IX for Philippine earthquake data from 1911–1995. Ms values from NEIC.Intensity values were evaluated from reports of investigations and previous earthquake catalogs. All depths are shallow. Refer toTable 1 for the detailed earthquake parameters used. Lines shows the linear fits.

7. Material studied NEIC. Intensities were evaluated from field investi-gation reports and earthquake catalogs.

The generalized isoseismals were drawn for each7.1. Recent eventsof these earthquakes using the 1931 MMI scale.Then, the areas enclosed by various intensities wereA total of 86 well-described earthquakes from

1911 to 1995 were selected as the calibrating earth- calculated. Isoseismals were drawn as generalizedlines in order to compensate for areas withoutquakes for this study. Table 1 lists these earth-

quakes and their parameters. All events are shallow intensity reports, especially in the offshore areas.Ms values were found to be adequate for thisand had an Ms between 4.6 and 8.3. Surface

magnitude values and depths were taken from the study’s purposes, since only one event had Ms>8,


(e) (f)

(g)

Fig. 6. (continued )

which avoids the problem of Ms saturation atR2=0.63 n=25 for Int. III

values of 8 and above. Relations between Ms andfelt areas enclosed by Intensities II to X [termed Ms=1.48[ log A(IV )]−1.53A(II ), A(III ) … to A(IX ) in this study] were

R2=0.61 n=60 for Int. IVplotted. From among the nine relations, the bestfits are those relating Ms with A(VI), A(VII),

Ms=1.36[ log A(V )]−0.40A(VIII) and A(IX ). However, relations for Ms–A(IV) and Ms–A(V ) include most data points. R2=0.63 n=71 for Int. VRelations for Ms–A( X) and Ms–A(II ) were notincluded because of too few data points. Fig. 6a–g Ms=1.30[ log A(VI )]+0.48shows plots of Ms–felt area relations for Intensities

R2=0.70 n=67 for Int. VIIII to IX. The resulting equations are:

Ms=1.42[ log A(III )]−1.48 Ms=0.83[ log A(VII )]+3.43


Table 2Estimated epicenters and magnitudes of Philippine historical earthquakes. Time is local time (GMT+8). A blank entry means thedata are unknown

Year Month/day Hr Min. N. Lat. E. Long. Ms No. of reports Quality

1608 1203 11.10 124.80 5.0 2 C1619 1130 16 0 18.50 121.60 8.0 9 B1621 1200 11.70 122.10 7.2 4 C1627 900 18.70 121.80 7.5 3 C1636 1221 4 0 7.40 123.60 6.5 4 C1645 1130 12 30 15.60 121.20 7.9 5 B1658 819 19 0 14.65 121.10 5.7 3 C1665 325 5 30 12.75 125.60 6.7 1 C1675 300 13.20 121.35 6.0 3 C1677 1207 11 30 14.50 119.50 7.3 4 C1687 919 16.90 122.20 6.9 2 C1688 1019 16.95 122.20 7.5 3 C1721 114 11 0 18.65 121.70 6.9 5 B1730 14.10 122.10 6.9 2 C1743 112 9 0 14.10 121.45 6.6 9 B1750 310 11 0 13.50 123.10 4.5 1 C1771 201 12 0 14.55 121.15 5.0 2 C1787 712 22 45 11.05 122.25 7.4 11 A1796 1105 6 0 16.10 120.50 6.9 9 B1808 7.35 121.65 5.7 1 C1811 1005 13.55 123.10 6.7 9 B1824 1026 14.20 121.90 7.4 5 B1828 1109 10 30 13.70 119.50 6.6 2 C1830 118 9 15 14.30 121.85 6.3 3 C1839 227 12 0 16.95 120.65 6.9 5 B1840 322 0 30 12.95 123.85 6.8 5 B1852 916 10 30 14.30 120.05 7.6 35 A1852 925 13.35 123.30 4.8 3 C1852 1224 14 0 13.85 120.90 5.1 8 B1861 630 16 30 8.25 127.05 6.0 1 C1861 723 12 25 13.50 123.10 4.1 2 C1861 812 23 15 13.50 123.10 4.6 1 C1861 830 15 30 13.30 123.05 5.9 5 B1861 925 20 57 13.30 123.45 4.5 2 C1862 304 9 30 13.75 120.40 6.1 5 B1862 531 13 5 17.95 121.15 5.3 2 C1862 713 8 0 15.65 121.50 5.8 2 C1862 804 18 10 18.10 120.80 4.5 3 C1862 807 18 30 17.60 121.50 5.1 5 B1862 908 19 0 17.90 120.50 6.2 8 B1862 908 19 5 17.60 121.10 3.8 4 C1862 908 21 30 17.60 121.10 4.5 4 C1862 908 22 0 17.60 121.10 4.5 4 C1862 909 1 0 17.50 121.05 4.7 4 C1862 912 14 45 17.65 121.65 5.2 4 C1862 922 16 0 13.40 122.65 4.9 2 C1862 922 16 17 13.35 122.75 5.1 4 C1862 922 20 58 13.05 123.00 5.4 2 C1862 923 16 10 13.30 122.95 6.0 8 B1862 1029 16 0 14.10 121.20 4.0 2 C1863 603 11 25 14.55 120.90 6.5 39 A1863 611 7 0 13.10 123.00 4.9 1 C




1863 712 23 30 16.35 120.75 4.4 1 C1863 729 18 30 17.60 120.75 5.3 5 B1863 731 18 0 17.15 120.60 4.6 3 C1863 811 22 40 16.65 120.95 4.7 3 C1863 927 3 5 11.25 124.55 6.1 3 C1863 928 9 0 12.20 123.90 4.6 2 C1863 1011 13.10 123.05 3.7 1 C1863 1119 18 0 7.25 125.90 4.9 1 C1864 103 9 0 6.00 124.10 6.9 5 B1864 712 4 45 14.20 121.90 5.0 2 C1864 713 20 20 17.15 121.05 4.4 3 C1864 1013 16 0 11.25 124.85 5.6 2 C1864 1206 3 45 12.35 123.65 4.8 2 C1865 415 12 30 17.15 121.05 4.4 2 C1865 1003 6 20 16.60 120.75 4.2 2 C1865 1019 14 30 13.60 123.60 6.5 16 A1865 1126 20 21 13.50 121.25 4.2 3 C1865 1224 23 30 17.60 120.25 4.2 2 C1866 114 20 30 12.45 123.70 4.3 2 C1866 413 17.50 120.25 3.9 1 C1866 415 19 50 13.50 121.25 4.0 1 C1866 628 8 0 18.05 120.45 4.8 2 C1866 709 6 15 15.35 119.90 3.8 1 C1866 727 13 0 17.60 120.45 3.4 1 C1866 1027 9 30 13.10 123.75 3.2 1 C1866 1118 17.60 120.60 4.3 2 C1866 1124 2 44 18.25 120.45 3.7 1 C1866 1228 19 0 18.10 120.35 4.6 1 C1867 107 3 0 13.50 123.10 4.8 2 C1867 110 10.70 122.05 3.2 3 C1867 204 5 8 13.55 121.20 4.2 1 C1867 223 21 15 7.35 125.95 4.7 1 C1867 226 12 45 17.25 120.60 4.9 2 C1867 316 1 6 18.15 120.55 5.5 2 C1867 719 9 0 14.30 119.60 5.8 2 C1867 815 7 15 13.55 121.25 3.4 2 C1867 1226 18 30 11.95 125.65 5.5 1 C1868 114 5 0 13.50 120.50 5.0 2 C1868 219 6 48 8.50 126.90 6.0 2 C1868 226 13 30 15.20 119.50 4.9 2 C1868 404 1 0 11.30 124.55 5.4 2 C1868 605 21 45 10.45 122.20 4.8 1 C1868 629 0 11 10.45 122.25 5.5 2 C1868 730 5 30 18.30 120.70 5.8 5 B1868 822 19 55 13.50 121.25 4.5 2 C1868 1020 5 30 17.20 121.10 4.4 2 C1868 1115 11.90 125.90 5.8 2 C1869 215 7.40 121.55 5.1 1 C1869 429 2 0 7.05 123.80 5.8 1 C1869 710 8 45 14.70 120.00 4.9 3 C1869 816 7 0 12.40 123.85 6.5 7 B1869 1001 3 15 14.25 120.45 6.6 42 A1869 1004 19 32 15.70 119.90 5.6 8 B1869 1023 8 30 14.05 121.80 5.6 7 B




1870 106 18 45 13.55 121.20 4.0 2 C1870 108 20 45 13.55 120.50 4.9 4 C1870 211 19 0 12.45 124.60 4.8 1 C1870 314 2 0 18.10 120.35 4.8 3 C1870 317 14 50 13.50 120.85 4.5 3 C1870 510 17 45 13.20 123.60 4.3 1 C1870 516 11 45 13.55 121.05 3.7 2 C1870 517 19 0 13.55 121.00 3.7 2 C1870 523 15 30 18.20 120.85 6.1 7 B1870 1104 7 0 7.00 127.10 6.8 1 C1871 220 19 0 9.10 124.70 5.0 1 C1871 323 18 48 16.95 120.70 5.1 4 C1871 528 3 0 7.25 125.95 5.5 1 C1871 613 16 30 13.35 123.45 3.9 2 C1871 711 13 0 17.05 120.45 5.2 8 B1871 711 14 0 16.60 120.35 3.4 2 C1871 722 13.20 123.65 4.1 2 C1871 801 5 0 13.95 122.00 4.2 2 C1871 1009 13 30 18.10 120.10 5.7 4 C1871 1105 0 45 8.90 126.90 7.3 8 B1871 1129 8 30 7.40 121.60 5.9 1 C1871 1208 9 30 7.30 123.75 7.0 10 A1871 1219 11 0 8.50 126.90 7.2 3 C1872 126 12 30 15.80 119.45 6.8 1 C1872 204 19 45 13.20 123.50 5.2 3 C1872 305 0 45 13.50 121.25 4.5 4 C1872 905 19 0 12.90 124.70 6.2 5 B1872 1219 14 0 15.95 120.05 5.8 2 C1872 1229 3 48 14.40 120.35 6.4 15 A1873 116 23 45 15.05 121.50 5.6 10 A1873 118 8 20 14.70 122.20 5.4 3 C1873 118 9 56 15.25 121.30 5.5 3 C1873 118 19 50 15.00 122.55 5.7 3 C1873 217 17 0 11.40 126.05 6.0 1 C1873 303 10 17 16.50 120.80 5.8 3 C1873 330 17 58 18.10 120.55 5.0 3 C1873 611 15 15 13.20 121.20 4.8 6 B1873 714 7 11 12.90 121.85 4.7 2 C1873 809 2 30 14.30 121.70 5.0 4 C1873 1114 9 30 13.60 122.25 6.5 14 A1873 1114 17 15 13.90 121.90 5.0 4 C1873 1216 19 45 11.85 124.00 5.7 3 C1874 116 20 0 12.50 123.50 5.3 3 C1874 202 21 55 15.50 119.60 6.1 14 A1874 203 5 15 14.50 119.65 5.9 6 B1874 228 12 45 12.40 123.70 5.3 3 C1874 413 22 20 18.10 120.45 5.1 4 C1874 708 2 32 16.10 122.25 6.1 9 B1874 824 4 15 18.35 120.50 5.3 4 C1874 824 22 15 7.20 121.50 6.4 4 C1874 1016 2 9 15.40 121.40 5.9 16 A1874 1016 2 15 15.40 121.45 6.1 23 A1875 118 14 15 11.70 125.90 6.7 2 C1875 126 23 30 14.70 122.30 5.4 3 C




1875 307 19 0 17.05 121.00 5.6 3 C1875 308 17 22 17.00 120.50 5.6 6 B1875 312 13 0 17.65 120.65 5.2 4 C1875 518 7 0 15.75 119.40 6.0 8 B1875 519 3 30 13.60 123.05 6.1 12 A1875 519 4 0 13.60 123.05 4.9 3 C1875 519 6 40 13.90 122.40 5.4 3 C1875 522 18 0 12.55 122.10 5.9 1 C1875 602 12.30 123.50 5.3 4 C1875 811 23 55 7.20 124.00 6.4 2 C1875 911 10.90 124.80 5.2 2 C1875 916 19 0 16.95 120.50 5.7 8 B1875 1101 8 25 17.00 120.10 6.1 9 B1875 1104 19 0 11.10 124.65 5.6 2 C1876 217 16 0 16.35 120.05 5.3 3 C1876 311 4 25 16.35 120.70 4.6 1 C1876 612 15 0 13.35 123.35 4.5 2 C1876 713 14 34 15.10 120.00 5.0 4 C1876 725 20 0 10.10 126.40 7.2 2 C1876 1117 15 20 17.25 120.35 5.0 1 C1877 226 12 27 15.30 119.60 6.2 10 A1877 602 15 6 15.55 120.45 5.8 16 A1877 606 0 22 18.45 120.70 5.1 4 C1877 704 17 7 14.10 123.10 6.6 18 A1877 723 8 20 11.10 124.65 5.5 1 C1877 912 8 0 17.45 120.70 5.6 8 B1877 1017 22 45 13.95 119.85 4.5 3 C1877 1129 5 57 14.40 119.65 5.6 3 C1878 112 13 5 15.10 119.60 5.8 11 A1878 116 11.05 124.65 5.5 1 C1878 729 2 0 12.50 122.10 5.9 1 C1878 803 4 14 14.60 119.80 6.2 25 A1878 916 4 15 6.95 125.75 6.5 1 C1879 103 1 29 15.60 119.50 5.1 10 A1879 514 10 57 13.95 119.90 5.6 10 A1879 622 18 55 18.15 120.60 4.3 3 C1879 630 15 5 9.35 125.60 6.9 10 A1879 1018 13 28 13.80 120.20 5.5 11 A1879 1214 12 10 17.65 120.65 4.6 5 B1880 227 9 20 17.05 121.05 5.2 1 C1880 714 16 53 14.90 121.85 6.2 35 A1880 718 4 40 14.90 121.55 7.6 166 A1880 720 7 40 14.40 121.70 5.9 31 A1881 323 7 30 8.90 125.70 4.9 4 B1881 612 23 0 9.95 126.60 6.0 1 C1881 711 4 35 10.45 121.90 6.3 2 C1881 716 6 13 18.10 120.60 5.4 5 B1881 727 8 10 16.35 120.90 5.3 1 C1881 730 8 15 13.90 120.10 4.9 4 C1881 814 13 46 14.40 119.70 5.6 10 A1881 815 1 15 15.25 121.70 6.1 10 A1881 901 4 10 16.45 120.90 4.6 2 C1881 917 20 55 16.35 120.95 4.9 3 C1881 918 15 40 16.40 120.90 5.0 3 C




1881 920 1 10 16.50 120.95 6.1 7 B1881 920 6 54 16.40 120.90 4.6 2 C1881 922 21 10 16.45 120.80 4.3 2 C1881 927 8 42 16.65 120.30 4.5 2 C1881 928 7 34 16.85 121.00 5.5 4 C1881 928 15 15 16.45 120.75 5.6 2 C1881 929 21 22 16.25 120.95 5.3 7 B1881 930 2 40 16.40 121.00 6.0 3 C1881 1002 8 24 13.60 120.35 4.6 4 C1881 1004 22 33 17.00 121.00 5.0 2 C1881 1023 16 35 13.60 120.35 4.3 2 C1881 1023 17 48 13.65 120.35 4.8 2 C1881 1023 22 5 13.60 120.40 4.9 4 C1881 1208 1 5 16.00 121.00 5.6 2 C1881 1216 19 57 13.15 123.60 4.5 3 C1881 1231 9 20 15.20 119.55 5.8 8 B1882 110 22 25 13.65 122.95 3.8 3 C1882 120 21 15 15.40 119.30 4.9 2 C1882 202 8 45 16.20 119.25 5.7 1 C1882 311 6 5 7.50 121.60 5.6 1 C1882 312 4 38 13.70 120.30 5.1 7 B1882 318 13 0 7.65 124.50 5.1 2 C1882 430 2 8 15.10 119.70 5.5 10 A1882 430 12 52 14.85 119.50 5.4 2 C1882 501 7 32 15.30 119.45 5.6 3 C1882 510 11 30 7.60 124.50 5.7 1 C1882 522 3 54 14.00 121.00 5.1 2 C1882 621 14 34 17.15 121.90 5.4 4 C1882 724 22 34 14.10 120.10 5.5 5 B1882 725 2 45 14.05 119.65 5.7 3 C1882 728 7 29 18.70 121.90 5.6 6 B1882 911 15 58 13.50 120.45 6.4 14 A1882 912 0 30 14.10 119.80 6.3 14 A1882 916 0 32 14.75 119.75 5.1 4 C1882 917 13 46 13.50 122.90 5.5 8 B1882 922 18 0 16.75 120.95 4.6 2 C1882 1004 6 32 13.70 121.60 5.0 5 B1882 1010 8 46 13.90 123.10 5.6 12 A1882 1010 15 45 13.75 123.30 5.2 4 C1882 1102 22 45 13.80 121.50 4.1 3 C1882 1128 21 3 15.05 119.60 5.6 9 B1882 1206 11.80 124.15 6.0 1 C1883 205 19 45 15.40 121.30 4.6 3 C1883 206 4 19 16.30 120.75 5.5 11 A1883 209 19 28 16.30 121.00 5.6 8 B1883 209 19 33 16.30 120.90 5.4 3 C1883 211 4 38 14.25 121.80 4.5 3 C1883 426 3 26 14.30 121.75 4.2 2 C1883 506 11 2 14.60 121.40 5.1 6 B1883 605 8 59 15.90 119.50 5.3 3 C1883 713 20 25 13.40 123.20 5.8 8 B1883 714 3 25 13.50 123.10 4.4 8 B1883 720 8 0 13.70 122.90 5.7 7 B1883 727 6 51 13.60 121.85 5.1 13 A




1883 809 4 17 14.25 121.85 4.0 1 C1883 810 10 50 17.65 120.25 3.9 2 C1883 1017 19 17 18.60 121.65 4.7 1 C1884 109 23 22 13.55 122.65 5.2 5 B1884 110 2 47 13.70 122.80 4.2 3 C1884 110 22 51 13.70 122.75 5.1 2 C1884 322 8 49 16.20 119.10 5.9 4 C1884 419 22 43 14.85 119.55 5.7 4 C1884 510 0 45 17.60 120.30 3.9 2 C1884 512 15 14 14.30 119.70 5.1 2 C1884 605 0 0 8.60 124.80 5.1 1 C1884 730 6 15 16.00 121.00 5.2 3 C1884 803 0 57 17.75 120.20 5.0 2 C1884 815 15 45 13.75 122.80 4.4 4 C1884 817 20 28 14.05 123.00 3.9 1 C1884 818 1 30 13.80 122.80 3.6 2 C1884 920 22 5 13.95 122.10 4.9 1 C1884 925 17 0 18.15 120.50 4.4 1 C1884 1011 16 12 16.60 120.20 4.2 1 C1884 1028 20 6 14.95 121.85 5.5 12 A1884 1111 8 26 13.00 123.10 6.1 14 A1884 1111 12 0 13.00 123.10 5.4 3 C1884 1113 19 0 7.15 125.60 4.3 1 C1884 1217 15 56 15.75 120.95 5.2 8 B1884 1217 16 14 15.50 121.25 5.1 1 C1884 1220 12 39 15.00 121.90 5.4 5 B1884 1223 21 0 11.40 124.50 5.1 1 C1884 1224 16 30 11.40 124.50 5.1 2 C1884 1227 21 30 9.00 126.80 5.7 1 C1885 121 5 7 8.95 126.80 5.7 1 C1885 217 8 0 8.60 124.85 4.5 1 C1885 218 5 30 18.65 120.90 4.8 2 C1885 222 7 30 7.90 127.10 7.2 6 B1885 302 16 52 18.50 120.00 5.9 2 C1885 303 9 37 13.55 123.15 3.6 2 C1885 403 16 32 14.55 120.20 4.9 4 B1885 514 23 15 17.20 120.05 5.9 14 A1885 521 2 0 8.95 126.85 5.7 1 C1885 710 14 30 18.55 121.60 4.8 2 C1885 721 4 59 17.60 120.10 4.6 2 C1885 723 2 45 8.85 123.10 7.0 12 A1885 724 8 34 15.30 122.00 5.9 14 A1885 804 8 40 8.80 123.10 4.1 1 C1885 804 19 0 13.10 123.70 4.2 1 C1885 909 8.80 123.10 4.8 1 C1885 923 8.80 123.10 4.8 1 C1885 929 22 0 10.10 126.60 6.2 1 C1885 1013 9 6 14.90 122.30 5.6 2 C1885 1016 1 0 9.05 126.80 5.8 1 C1885 1026 10 45 17.05 120.10 5.7 3 C1885 1030 22 15 9.05 126.75 5.8 1 C1885 1116 15 21 14.60 121.80 5.9 13 A1885 1119 3 15 16.65 120.90 6.0 14 A1885 1119 13 31 15.95 120.75 4.1 2 C




1885 1208 15 18 17.55 120.65 4.1 1 C1885 1227 15 40 16.55 120.80 4.7 1 C1886 101 8 26 16.20 119.60 4.9 1 C1886 111 8 0 9.05 126.80 5.8 1 C1886 410 1 0 10.50 122.20 5.4 2 C1886 414 15 43 16.35 120.65 5.3 5 B1886 521 4 18 13.75 122.35 5.4 7 B1886 523 11 0 13.60 120.80 4.3 2 C1886 603 3 21 13.65 120.40 5.0 2 C1886 606 13 56 15.35 119.95 5.0 2 C1886 728 22 50 16.65 121.95 4.8 1 C1886 801 19 58 13.50 120.05 6.3 22 A1886 902 15 54 16.00 121.00 5.6 7 B1886 1103 1 25 18.10 120.50 4.6 2 C1886 1126 9 0 10.45 122.20 5.1 1 C1887 124 17 30 10.40 122.20 5.1 1 C1887 201 5 0 17.40 121.20 5.8 4 C1887 202 15 0 11.45 122.05 7.3 7 B1887 301 10.45 123.80 4.4 1 C1887 304 12 42 9.10 126.80 5.8 1 C1887 324 13 14 13.45 123.25 5.6 10 A1887 324 13 16 13.70 122.90 4.3 3 C1887 324 13 28 13.60 122.95 4.4 2 C1887 324 14 30 13.40 123.25 4.6 3 C1887 325 4 45 13.95 122.60 5.3 8 B1887 325 5 5 13.60 122.80 3.7 2 C1887 325 5 45 13.70 122.85 3.7 2 C1887 325 7 10 13.65 122.85 5.4 3 C1887 401 11 30 13.65 122.85 4.3 4 C1887 402 1 0 13.55 122.85 4.1 3 C1887 402 11 35 13.35 123.35 4.5 3 C1887 412 19 0 9.05 126.80 5.8 1 C1887 422 13.75 122.75 4.1 2 C1887 521 7 15 13.55 123.05 3.9 1 C1887 602 8 42 18.40 120.95 4.9 2 C1887 611 2 23 13.65 122.45 4.8 5 B1887 619 4 32 15.35 121.55 5.4 18 A1887 822 11 0 9.10 126.75 5.8 1 C1887 915 14 0 14.00 124.00 5.2 2 C1887 919 2 25 16.15 119.80 4.8 1 C1887 1001 5 5 13.90 123.75 5.7 3 C1887 1001 14 5 9.05 126.80 5.8 1 C1887 1115 19 45 15.35 119.85 4.4 3 C1887 1130 21 34 15.65 121.20 5.0 5 B1887 1203 8 0 8.75 123.25 4.7 1 C1888 124 3 30 10.45 123.75 4.4 1 C1888 126 19 45 8.85 126.75 5.8 4 C1888 307 20 10 13.60 123.15 3.9 1 C1888 503 5 40 18.65 120.85 5.4 3 C1888 507 5 3 5.60 124.85 4.7 2 C1888 517 13 30 5.65 124.85 4.7 2 C1888 611 23 59 13.20 123.55 4.6 1 C1888 723 15 0 13.55 123.05 3.9 1 C1888 729 7 30 8.40 127.10 5.7 1 C




1888 815 9 9 16.65 120.90 5.0 1 C1888 818 22 25 18.80 121.90 6.1 5 B1888 819 2 0 17.15 122.15 6.5 6 B1888 819 6 10 10.40 122.10 5.1 1 C1888 819 6 39 18.70 121.90 5.2 2 C1888 819 6 55 10.40 122.10 5.8 1 C1888 819 7 15 10.45 122.10 5.8 1 C1888 819 8 0 17.30 122.25 6.0 1 C1888 915 9 9 16.65 120.90 5.0 1 C1888 916 0 41 17.20 122.15 5.4 3 C1888 1007 11 45 7.25 123.65 5.2 1 C1888 1008 9 10 17.40 120.60 4.2 2 C1888 1130 0 30 9.10 126.75 5.8 1 C1888 1203 19 0 10.40 122.25 5.1 1 C1888 1214 14 45 15.05 122.05 5.1 3 C1888 1222 20 54 9.00 126.75 5.8 1 C1889 101 2 20 9.35 126.30 6.7 4 C1889 112 13 5 9.40 126.60 5.9 4 C1889 112 13 40 9.10 126.75 6.2 2 C1889 121 9.05 126.80 5.5 1 C1889 129 21 30 18.30 121.05 5.3 2 C1889 205 8 0 7.30 123.60 6.6 7 B1889 211 3 45 7.55 121.65 5.6 1 C1889 420 8 51 15.55 119.40 4.7 2 C1889 425 13 10 13.65 123.05 3.9 1 C1889 512 1 28 15.65 119.25 4.9 3 C1889 515 0 20 18.35 121.55 5.2 4 C1889 525 2 23 13.15 121.10 6.4 21 A1889 526 5 30 12.15 121.20 4.6 1 C1889 717 22 55 13.10 123.65 4.0 1 C1889 804 17 45 7.50 121.60 5.6 1 C1889 819 4 55 17.95 120.30 4.0 2 C1889 1006 3 15 5.90 127.00 7.3 7 B1889 1006 6 10 6.30 126.15 5.1 1 C1889 1020 14 25 13.10 124.45 5.6 2 C1889 1210 23 10 17.95 120.15 4.2 2 C1890 113 21 14 15.80 119.85 5.8 11 A1890 118 11.10 124.75 4.6 1 C1890 206 16 30 11.35 124.85 6.4 8 B1890 206 21 0 11.35 124.85 4.9 1 C1890 413 6 4 18.75 120.25 5.9 6 B1890 413 12 0 17.75 120.20 4.4 3 C1890 413 12 45 18.45 120.00 6.1 3 C1890 503 7 24 13.55 122.00 5.1 7 B1890 523 13 55 13.60 120.50 4.4 2 C1890 604 10 14 14.75 121.95 5.2 3 C1890 714 6 45 6.20 120.75 5.0 3 C1890 929 22 45 17.90 121.20 4.9 3 C1890 1222 10 52 9.20 121.80 5.5 1 C1890 1229 7 9 16.15 119.60 4.9 1 C1891 106 22 8 13.95 123.05 4.4 1 C1891 201 23 31 18.70 120.75 5.4 2 C1891 315 2 4 13.60 121.40 5.0 5 B1891 329 12 15 7.30 123.60 5.2 2 C




1891 404 21 30 8.50 126.70 6.3 2 C1891 519 11 16 15.00 121.80 5.4 7 B1891 604 10 15 8.20 125.85 5.2 1 C1891 621 8 55 10.40 122.20 5.1 1 C1891 624 19 42 7.60 126.00 7.2 6 B1891 712 5 5 13.25 123.60 3.7 2 C1891 805 15 0 16.40 119.25 5.2 2 C1891 911 16 10 14.00 119.75 5.1 2 C1891 912 23 3 13.50 121.00 4.5 1 C1891 913 12 5 14.75 121.95 5.2 3 C1891 1002 21 45 8.05 125.75 4.5 1 C1891 1008 13.15 123.65 4.5 4 C1891 1027 17 45 8.90 125.70 4.5 1 C1891 1029 22 10 13.90 120.00 5.8 7 B1891 1102 21 35 13.70 120.20 4.3 3 C1891 1104 0 30 13.50 121.00 3.8 2 C1891 1114 2 30 17.30 120.10 4.1 2 C1891 1121 9 8 13.20 123.55 4.7 1 C1891 1209 17 12 13.60 120.50 4.2 2 C1891 1221 2 14 18.00 121.20 4.8 2 C1891 1228 18 12 19.00 122.20 5.8 5 B1891 1228 18 20 18.20 121.10 5.6 3 C1892 125 5 50 18.50 120.70 5.4 4 C1892 129 3 24 14.15 122.05 4.3 2 C1892 207 13 0 10.30 123.20 5.0 1 C1892 223 9 56 14.25 121.90 5.2 5 B1892 308 20.40 121.45 6.8 1 C1892 313 17 30 10.10 123.00 6.1 6 B1892 313 18 30 10.30 122.15 6.1 2 C1892 316 12 58 16.40 120.40 6.6 80 A1892 322 16 30 10.50 123.10 5.5 2 C1892 326 7 44 16.30 120.85 5.3 7 B1892 407 8 55 16.15 122.05 5.5 7 B1892 520 22 30 15.35 119.65 5.4 2 C1892 615 22 15 6.65 123.75 5.5 3 C1892 706 22 56 13.60 123.10 4.4 2 C1892 712 22 4 13.80 120.00 4.5 2 C1892 723 16.40 120.70 4.0 1 C1892 728 12 7 15.00 121.85 5.8 11 A1892 1203 9 30 16.60 121.00 5.8 8 B1893 125 12 30 18.70 121.90 5.9 4 C1893 226 18 2 13.65 120.20 5.5 4 C1893 308 16 39 16.85 120.20 5.6 9 B1893 401 3 14 13.45 120.65 4.9 8 B1893 412 5 51 13.10 123.65 5.5 5 B1893 424 7 24 12.90 124.25 5.5 8 B1893 508 16 27 13.10 124.45 5.2 3 C1893 602 22 30 7.60 123.80 5.9 5 B1893 621 7 30 7.65 126.10 7.3 14 A1893 914 13 18 15.30 122.00 5.7 5 B1893 1116 4 31 15.30 121.45 5.8 7 B1893 1223 16 20 12.00 124.10 5.6 2 C1894 209 17 5 6.30 125.85 6.7 6 B1894 604 10 15 8.20 125.85 5.2 1 C1895 513 22 45 13.50 120.90 5.4 5 B


exhibited during the recent instrumental period.R2=0.73 n=34 for Int. VIIThis may be explained by the lack of settlementsand structures that could be damaged on theMs=0.63[ log A(VIII )]+4.78eastern Philippine shores. Such conditions prevailR2=0.70 n=19 for Int.VIIIeven now, as evidenced by the low damage incurredby earthquakes originating from the PhilippineMs=0.78[ log A(IX)]+4.53Trench region during the 1989 and 1992 earth-

R2=0.80 n=15 for Int. IX.quakes. On the other hand, based on earthquakeeffects of the August 2, 1968 earthquake in

7.2. Historical earthquakes Casiguran and the April 7, 1970 earthquake inBaler, earthquakes from the East Luzon Trough–From the 6679 historical accounts of more thanCasiguran Fault source region are felt widely and3000 earthquakes, isoseismal maps of 485 eventscan cause damage in areas more than 100 kmwere drawn. These historical events were mostlyaway. The possible effect of attenuation for thederived from the excellent catalog on Philippineeast Philippine source region, therefore, has to beearthquakes done by Repetti (1946) and fromstudied further to assess the real reason for themany other references. Primary earthquakelack of reports. Some historical earthquakes areaccounts were, as much as possible, obtained frombeing related in this study to source regions thatPhilippines and Spanish libraries in order to avoidhave not ruptured during the instrumented period.misquotation, inaccuracies and exaggerations.For example, the sparsely reported August 20,1658 earthquake may have been an event possiblyrelated to the activity of the east Valley Fault, a8. Results and analysesfault system nearest to Metro Manila. A local faultwest of Banahaw Volcano probably moved andThe epicenters and magnitudes of 485 historical

earthquakes were estimated. The resulting histori- triggered the earthquake on January 12, 1743,causing the collapse of the volcano’s crater lake.cal seismicity map is shown in Fig. 7. Their param-

eters are listed in Table 2. Most of the earthquakes Intensity was high in the epicentral area but theoverall intensity reports were limited in extent. Inare related to known active structures. Fig. 7 also

shows that there are more events in the northern fact, there was no report from Manila, which isless than 100 km away. Both faults have shownPhilippines compared with the southern part of

the country. It is difficult to ascertain if this a true no recent activity of a similar nature during theinstrumented period.observation rather than an artifact of the dataset,

but it should be considered that a probable reason Meanwhile, by plotting the magnitude distribu-tion of historical events against that of the 1951–for this is because the northern part was more

successfully colonized, and that the main sources 1997 period (Fig. 8) we find that only the 1851–1900 historical period may be considered as a well-of historical documents are from the colonizers’

records, priest historians and Spanish officials. On reported period, as the number of its magnitude5.6 to 7.5 earthquakes approaches the number ofthe other hand, there were fewer churches in the

southern Philippines, and the area remained mostly events found in the 1951–1997 recent catalog. Notsurprisingly, the events with magnitudes betweenunder the non-Christian religion and was not fully

colonized. Records from Muslim mosques and 5.1 and 5.5 are not that well reported, as expectedfor smaller events. Meanwhile, the number offrom nearby Southeast Asian countries could be

potential future sources of additional historical earthquakes with magnitudes greater than 7.5 doesnot approach the same numbers in the recentearthquake information.

There is also a notable lack of historical earth- catalog. This suggests that either some magnitudesin the 1851–1900 period were underestimated duequakes coming from the eastern subduction zone,

especially the Philippine Trench–East Luzon to poor reporting or that there were really fewlarge magnitude events during this period.Trough region, despite the intense seismicity it has


Fig. 7. Locations of the epicenters of the 485 historical earthquakes determined in this study. Circles show the location of epicentersand are sized according to magnitude.

Fig. 9 shows the plot of the number of earth- the expulsion of the Jesuit order from the Philippinechurch scene. The Jesuit order was the main collec-quakes per 50 year interval. It shows a correlative

increase in the number of quakes with the increase tor of scientific information, and in fact was thegroup that organized the Manila Observatory. Thein the number of towns. However, the period from

1701 to 1850 shows a very low number of earth- increase in earthquake reports during the 1851–1900 period, although it may be attributed to thequakes despite the increase in the number of towns.

This could be interpreted as a period of few earth- increase in number of towns starting around 1850,may also be related to the return of the Jesuits toquakes or may be due to few available reports.

Significantly, the period when the decline in the the Philippines and the establishment of the ManilaObservatory in the 1850s.number of earthquakes was reported coincides with


Fig. 8. Plot of the number of earthquake reports and thenumber of towns per 50 year period. A sudden increase inthe number of towns is noted starting between 1851 and 1900(denoted by the arrow) and can be correlated with an increasein the number of earthquakes during that time. Fig. 10. The number of earthquake accounts for each historical

earthquake was evaluated. Earthquakes with ten or more inten-sity reports are assessed as quality ‘A’, those with five to ninereports as quality ‘B’ and those with less than five reports aregiven quality ‘C’. All earthquakes from 1851 and above havequality ‘A’ ranking. With the increase in the number of townsduring the 1801–1850 period, the number of earthquake reportsalso increased starting from 1851. Earthquakes before 1851have fewer intensity reports, possibly due to fewer towns.

earthquakes with time. The plot shows that earth-quakes earlier than 1850 were nearly all ranked asquality ‘B’ or ‘C’ with one quality ‘A’ event. Mostevents with magnitudes 5.0 and below have quality‘C’, very few had quality ‘B’ and not one hadquality ‘A’. Meanwhile, magnitudes of quality ‘A’events varied from 5.1 to 7.6. Earthquakes withgood quality parameters may be used for conduct-ing seismic studies requiring a longer database.Although Musson (1998) has cautioned against

Fig. 9. Plot of the number of earthquakes per 50 year period.the parameterization of poorly reported events,Historical earthquakes are only well reported after 1850. Priorsuch kinds of event were still parameterized in thisto this, very few earthquakes are found. In the 1851–1900 period

the number of magnitude 5.1 to 5.5 events is below the number study but were given a low quality ranking. Fourof similar magnitude events in the 1951–1997 period. of these events are presented in this study.

The earthquakes were also classified accordingto the number of reports. Earthquakes with ten or 9. Examples on how epicenters were determined for

some sparsely reported eventsmore intensity observation points were ranked ashaving quality ‘A’. Those with five to nine intensityobservation points were assessed as quality ‘B’, The following discusses the basis for the esti-

mated epicenters of some sparsely reported events.whereas those with four or fewer points of intensityobservations were assessed as having quality ‘C’. As mentioned earlier, Musson (1998) cautioned

against parameterization of earthquakes with fewFig. 10 shows the quality of ranking of these


accounts and suggested that, instead, it is better great and prolonged that the Indians, young onesand some very old, have no recollection of havingto make the users aware of the inferences used by

the evaluator of historical accounts. About 40% experienced another of such long duration’’ (deAlzina, 1668) and indicates that the local peopleof the analyzed events have fewer than five inten-

sity reports, making them fall into the ‘C’ category were already accustomed to earthquake occur-rences, but that this was the strongest event thatin terms of the quality of determination. Although

the resulting parameters may not be so useful for they had felt in many years.The lone stone structure (a fort) was undam-conducting seismological studies, these initial eval-

uations are very important for future studies like aged; de Alzina (1668) attributed this to its thick-ness and low height. The other houses were madepaleoseismology and for further reevaluation.of light, wooden materials. The limited damagereport suggests a moderate-magnitude event or a9.1. The March 25, 1665 earthquakesource region that is of a relative distance fromthe town that experienced the intensity. It is alsoThe earthquake on March 25, 1665 (Fig. 11) is

the first reported earthquake affecting Samar very difficult to assess intensity since there wasonly one masonry structure. Hence, a minimumisland. This island faces the northern part of the

Philippine Trench. Several large magnitude earth- intensity of VI was assigned to the town ofPalapag. This intensity level can account for thequakes (1897, 1975) have been recently felt in the

eastern shore of this island, the most recent of non-damage to the stone fort, to the strong groundshaking and to the non-reporting of towns west ofwhich was on April 21, 1995. The 1665 quake was

actually only reported from the town of Palapag Samar island where development was higher.The most active and nearest earthquake sourcein northeastern Samar Island. The description of

this event said that ‘‘there was an earthquake so region to Samar island is the Philippine Trench,which is about 70 km away from the town ofPalapag. Recent major trench-related earthquakes(Ms of at least 7) in 1975 and 1995 also exhibitedlow damage, mainly because of the low populationand development of Samar island even at present.The characteristic intensity distribution, wherethere is insignificant damage possibly due to lowpopulation and development, was also displayedby this quake. Had the source been due to move-ment of a local fault there would have been moredamage, especially on the western part of theisland where development is slightly higher thanon the east. The trench as the possible sourceregion can explain the moderate intensity andcould also explain the lack of reports from themore-developed towns west of Samar Island.

The epicenter of the 1665 earthquake is esti-mated at 12.75 N Lat.; 125.6 E Long. and its Ms(ave) is 6.7. This magnitude is small as comparedwith other trench-related events fronting Samar

Fig. 11. Isoseismal map of the March 25, 1665 (1:30 p.m.) earth- Island. Typical events originating from this seg-quake. The filled circle signifies the estimated epicenter; black ment of the trench have Ms of 7 (for example, Mssquares are locations of known and established towns by this

7.7 in 1897, Ms 7.2 in 1975 and Ms 7.3 in 1995).time. Italicized names are places mentioned in earthquakeIn light of this observation it is possible that theaccounts. Roman numbers are MMI intensities. A(VII ) is

10 858 sq. km (Ms 6.7). 1665 event really was an Ms 7 event, but that the


scarcity of damage reports limited its felt area. et al., 1967; Berninghausen, 1969). This earth-quake probably originated from the ManilaRecent events, such as the 1995 earthquake, have

shown that damage is limited when epicenters are Trench, based on its intensity and tsunami report.Likewise, the generation of a tsunami suggests afrom the offshore areas compared with when the

epicenters were on land. Another way of looking shallow epicenter and a magnitude of around 7.Using the Intensity VI isoseismal, the Ms is esti-at the data is that the 1665 event is a possible

lower limit of large magnitude events in this section mated at 7.3 and the epicenter at 14.50 N. lat and119.50 E long.of the trench.

9.2. The December 7, 1677 earthquake 9.3. The October 26, 1824 earthquake

A not-so-well reported earthquake occurredOnly three sites (Gapan in Nueva Ecija, Bauanin Batangas and Manila) in central to southern with only five intensity reports on October 26,

1824 (Fig. 13). This quake destroyed the churchesLuzon had reports, and these stated that strongground motion caused fissures to form. No damage of Lucban town in Quezon province and that of

Cavinti town in Laguna province. Antipolo townwas significant enough to be reported for the samethree places. Intensity VI was assigned to these in Rizal province and Manila also experienced

high intensities. Slight damage was also reportedthree sites by giving importance to the lack ofdamage report, suggesting that had there been from Cavite province. The quite significant inten-

sity at Manila and Cavite are attributed to siteany damage at all, it was, at worst, only slight(Fig. 12). There was also a report of sea waves response, since both are underlain by thick allu-

vium deposits. The reports from Manila andalong the west coast of Luzon (Repetti, 1946; Iida

Fig. 12. Isoseismal map of the December 7, 1677 (7:30 p.m.)Fig. 13. Isoseismal map of the October 26, 1824 earthquake.earthquake. The filled circle signifies the estimated epicenter;

black squares are locations of known and established towns by The filled circle signifies the estimated epicenter; black squaresare locations of known and established towns by this time.this time. Italicized names are places mentioned in earthquake

accounts. Roman numbers are MMI intensities. A(IX ) is Italicized names are places mentioned in earthquake accounts.Roman numbers are MMI intensities. A(IX ) is 8705 sq. km3590 sq. km (Ms 7.3), A(VIII ) is 16 924 sq. km (Ms 7.4),

A(VII ) is 50 359 sq. km (Ms 7.3) and A(VI) is 124 225 sq. km (Ms 7.6), A(VIII ) is 20 482 sq. km (Ms 7.5), A(VII) is27 046 sq. km (Ms 7.1). Ms (ave) is 7.4.(Ms 7.1). Ms (ave) is 7.3.


Cavite, although damage was mentioned, failed tostate the names of churches or other structuresthat were damaged. This omission is consideredsignificant, since if the damage was serious, chroni-clers would have written down the names of thechurches and affected structures. The fact that theyfailed to do so probably meant that the damagein these two places was not that serious. Thisobservation also gives low probability that thesource region came from the Manila Trench in thewest, to which the two places (Cavite and Manila)are nearer. Meanwhile, Lucban and Cavinti areboth underlain by thick pyroclastic materials andyet sustained high intensities, suggesting that thesource is probably nearer these two areas. It issignificant that there were no reports from theprovinces of Batangas, Camarines, Bulacan,Pampanga and from other towns of Quezon andLaguna. The nearest active structure to the heavily Fig. 14. Isoseismal map of March 22, 1840 earthquake. Time

of occurrence was between 8 and 9:00 a.m. The filled circleaffected areas is the Philippine Fault in Quezonsignifies the estimated epicenter; black squares are locations ofprovince. Based on the severely damaged towns,known and established towns by this time. Italicized names arethe probable source is the area between Lucbanplaces mentioned in earthquake accounts. Roman numbers are

and Cavinti towns. The intensity distribution is MMI intensities. A(VIII ) is 1521 sq. km, (Ms 6.8).similar to that of the March 17, 1973 earthquake,except that the 1824 event possible ruptured asegment north of the 1973 segment. North of the may be due to local site response is strong. Despite

the total collapse, only an Intensity VIII wasproposed epicenter for this quake the area is notso well populated, and hence this could explain assigned to both places, since reports of structural

damage should have been received from the otherthe lack of reports in that area. The estimatedepicenter is at 14.2 N Lat. and 121.90 E Long. and towns if this had been Intensity IX. Milne (1912)

included the island of Masbate in the placesits magnitude is 7.4, a characteristic earthquakesize for this segment of the Philippine Fault. affected. None of the other chroniclers did so,

however. An Intensity VIII in the Sorsogon Bayarea means an Intensity VII or VI on Masbate9.4. The March 22, 1840 earthquakeisland and could also explain the inclusion ofMasbate in the affected area if Milne (1912) wasThis is the first large earthquake in the Sorsogon

area. There are descriptions of only two places correct in his data. The estimated epicenter is12.95 N Lat. and 123.85 E Long. and Ms of 6.8.where the total collapse of churches, the only

existing masonry structures, took place — in the This earthquake was probably related to the activ-ity of a fault on the Bicol peninsula (Fig. 14).Sorsogon towns of Sorsogon and Casiguran.

Liquefaction affected a significant portion of A possible modern analogue is the earthquakeof July 2, 1954 (Ms 6.75). This caused the collapseSorsogon town, when its northern part sank by

1.5 m. The towns of Sorsogon and Casiguran are of a church in Sorsogon town, just as happenedduring the 1840 event. Large fissures and landslidesboth near the bay area. At this time, other towns

of Sorsogon, such as Bacon, Bulusan, Donsol, were reported. The epicenter of the 1954 earth-quake, as determined by the InternationalGubat, Matnog and Juban, were already extant,

but no reports from them were received. Hence, Seismological Service (ISS), was 13.0 N Lat. and123.90 E Long., which is very near the sourcethe possibility that the area affected is limited and


region of the 1840 event. It should be pointed out more data to improve the estimates and also toevaluate more events.that the magnitudes of these two events are

thought to be quite similar. If the 1840 and 1954events ruptured the same segment, the similarityof the magnitude values could indicate a character-

Acknowledgementsistic earthquake for this segment of the fault.During these two events, the town of Sorsogon

We would like to thank the help of many people,experienced liquefaction and Masbate island expe-especially the librarians and PHIVOLCS col-rienced Intensity V on each occasion.leagues who helped in the search for historicalaccounts in the Philippines, Spain and Japan. Inparticular, Mr Fernando Rodriguez of Madrid,Spain, who searched the Biblioteca Nacional of10. Discussions and conclusionsSpain, Ms Ma Theresa C. Pascual for combingthe University of the Philippines and Ateneo deThis study was able to estimate the magnitudesManila University libraries and Mrs Apolonia C.and epicenters of 485 well-reported and damagingPascual for searching the Eugenio Lopez Library.historical earthquakes of the Philippines. As anWe would also like to acknowledge our usage ofinitial step, empirical relations between magnitudethe Generic Mapping Tool version 3.0 (GMT 3.0)and felt areas using recent earthquakes were deter-software (Wessel and Smith, 1995) in making themined. A method to assess historical earthquakefigures. We would like to thank Drs. Raymundoaccounts using various controls and assumptions,S. Punongbayan and Bartolome C. Bautista forlike the locations of established towns when themeaningful discussions. We would also like toearthquake occurred, knowledge of local geologi-thank Professor Dr F. Wenzel of thecal condition, type of construction, etc., wasGeophysikalisches Institut, Universitat Karlsruhe,illustrated. The locations of most of these earth-and two anonymous reviewers whose suggestionsquakes are along known active structures. A fewand insights greatly improved this manuscript.epicenters, however, lie along structures that do

not show much activity during the instrumentedperiod. Good quality estimates for both magni-tudes and epicenters are only available from the References1850 period onwards owing to the increase in thenumber of towns. Earlier than this time, magnitude Acharya, H.K., 1978. Mindanao earthquake of August 16,

1976: preliminary seismological assessment. Bull. Seismol.and epicentral determinations suffer from lowSoc. Am. 68, 1459–1468.numbers of historical accounts. This situation may

Acharya, H.K., Aggarwal, Y.P., 1980. Seismicity and tectonicsbe improved further by searching for possibleof the Philippine Islands. J. Geophys. Res. 85, 3239–3250.

archival materials from Moslem mosques or from Bautista, B.C., 1996a. Seismotectonic implications of recentrecords kept in neighboring countries with which Philippine earthquakes from 1980 to 1994. M.A. Thesis

(unpublished). State University of New York at Bingham-the Philippines was in contact during its earlyton, Binghamton, NY, USA, 220 pp.history. Most of the sources of information of this

Bautista, B.C., Bautista, M.L.P., Rasdas, A.R., Lanuza, A.G.,study relied on church and government archivalAmin, E.Q., De Los Reyes, P.J., 1991. The 16 July 1990

documents. New earthquakes with good quality earthquake and its aftershock activity, Monograph of theparameters may be used for conducting seismic 16 July 1990 Earthquake, 81–118.

Bautista, M.L.P., 1993. Historical earthquakes in the Philip-studies requiring a long time history of the seismi-pines, Proceedings to the Joint Conference of Seismology incity, whereas parameters for poorly reported eventsEast Asia, Oct. 29–Nov. 2, Tottori, Japan, 53, abstract only.should be used with care and uncertainties should

Bautista, M.L.P., 1996b. Estimation of the magnitude and epi-be borne in mind when using the resulting parame- centers of Philippine historical earthquakes. M.S. Thesisters. More in-depth research from among local (unpublished), Kyoto University, Kyoto, Japan.

Bautista, M.L.P., 1999. Seismic hazard assessment of the Philip-and other Philippine libraries is needed to obtain


pines using historical and recent earthquakes. D.Sc. Disser- Hayes, D.E., Lewis, S.D., 1984. A geophysical study of thetation (unpublished), 183 pp. Manila Trench, Luzon, Philippines 1. Crustal structure,

Berninghausen, Wm.H., 1969. Tsunamis and seismic seiches of gravity and regional tectonic evolution. J. Geophys. Res.southeast asia. Bull. Seismol. Soc. Am. 59, 289–297. 89, 9171–9195.

Bischke, R.E., Suppe, J., del Pilar, R., 1990. A new branch of Heck, N.H., 1947. List of seismic sea waves. Bull. Seismol. Soc.the Philippine Fault system as observed from aeromagnetic Am. 37, 269–279.and seismic data. Tectonophysics 183, 243–264. Hirano, S., Nakata, T., Sangawa, A., 1986. Fault topography

Bollinger, G.A., Chapman, M.C., Sibol, M.S., 1993. A compari- and Quaternary faulting along the Philippine Fault Zone,son of earthquake damage areas as a function of magnitude central Luzon, the Philippines. J. Geogr. 95, 1–23.across the United States. Bull. Seismol. Soc. Am. 83, Iida, K., Cox, D.C., Pararas-Carayannis, G., 1967. Preliminary1064–1080. catalog of tsunamis occurring in the Pacific Ocean. Report

Cardwell, R.K., Isacks, B.L., Karig, D.E., 1980. The spatial no. 5. HIG-67-10. International Tsunami Informationdistribution of earthquakes focal mechanism solutions and Center. The State of Hawaii and Office of Naval Researchsubducted lithosphere in the Philippines and northeastern under contract no. Nonr-3748 (03).Indonesian islands. In: Hayes, D.E. (Ed.), Tectonic and Lewis, S.D., Hayes, D.E., 1983. The tectonics of northwardGeologic Evolution of the Southeast Asian Seas and Islands. propagating subduction along eastern Luzon, PhilippineAGU Geophys. Monograph Series 23, 1–24. Islands. In: Hayes, D.E. (Ed.), Tectonic and Geologic Evo-

Cole, J., McCabe, R., Moriarty, T., Malicse, J.A., Delfin, F.G., lution of the Southeast Asian Seas and Islands. AGU Geo-Tebar, H., Ferrer, H., 1989. A preliminary Neogene paleo- phys. Monograph Series 27, 57–78.magnetic data set from Leyte and its relation to motion on Lewis, S.D., Hayes, D.E., 1989. Plate convergence and deforma-the Philippine Fault. Tectonophysics 168, 205–220. tion, North Luzon Ridge, Philippines. Tectonophysics 168,

Cox, D.C., 1970. Discussion of ‘‘Tsunamis and seismic seiches 221–237.of Southeast Asia’’ by W.H. Berninghausen. Bull. Seismol. Liporada, A., Atanacio, P., 1988. Philippine churches: gems ofSoc. Am. 60, 281–287. religious architecture. The Philippines Pearl of the Orient.

Daligdig, J.A., 1997. Recent faulting along the Philippine Fault revised ed., Islas Filipinas.in north central Luzon. unpubl. D.Sc. Thesis (unpublished), Lockridge, P.A., 1988. Historical tsunamis in the Pacific Basin-Kyoto University, Japan.

Natural and Man-Made Hazards. Reidel, pp. 171–181.De Alzina, F.I., 1668. Historia Natural del Sitio, Fertilidad y

Ludwig, W.J., 1970. The Manila Trench and West LuzonCalidad de las Islas e Indios de Bisayas.

Trough — III. Seismic-refraction measurements. Deep SeaDeffontaines, B., Pubellier, M., Rangin, C., Quebral, R., 1993.

Res. 17, 553–571.Discovery of an intra-arc transform zone in Mindanao

Mallet, R., 1853. Catalogue of recorded earthquakes from 1606(Philippines) using morpho-neotectonic data. Z.B.C. to 1850. Third report on the facts of earthquake phen-Geomorp. 94, suppl., 261–273.omena. In: Report of the 22nd Meeting of the British Associ-Dunbar, P.K., Lockridge, P.A., Whiteside, L.S., 1992. Catalogation for the Advancement of Science, London. Johnof significant earthquakes 2150 B.C.–1991 A.D. incl. quanti-Murray, Albermarle Street, pp. 1–176.tative casualties and damage. Report SE-49, World Center

Mallet, R., 1854a. Catalogue of recorded earthquakes fromA for Solid Earth Geophysics. National Geophysical Data1606 B.C. to 1850 (continued). Third report on the facts ofCenter, Boulder, CO, USA. 320 pp.earthquake phenomena. In: Report of the 22nd Meeting ofDuquesnoy, Th., Barrier, E., Kasser, M., Aurelio, M., Gaulon,the British Association for the Advancement of Science,R., Punongbayan, R.S., Rangin, C., French–PhillipineLondon. John Murray, Albermarle Street, pp. 118–213.Cooperation Team 1994. Detection of creep along the Phil-

Mallet, R., 1854b. Catalogue of recorded earthquakes fromippine Fault: first results of geodetic measurements on Leyte1606 B.C. to 1850 (continued). Third report on the facts ofIsland, central Philippines. Geophys. Res. Lett. 21, 975–978.earthquake phenomena. In: Report of the 22nd Meeting ofEbel, J.E., 1996. The seventeenth century seismicity of north-the British Association for the Advancement of Science,eastern North America. Seismol. Res. Lett. 67, 51–68.London. John Murray, Albermarle Street, pp. 2–326.Fitch, T., 1970. Earthquake mechanisms and island arc tecton-

Maso, M.S., 1895. La Sismologia en Filipinas. Establicimientoics in the Indonesian–Philippine region. Bull. Seismol. Soc.Tipo Letograpico de Ranuirez y Compania, Manila. 121Am. 60, 565–591.pp.+figures.Gutenberg, B., 1956. Great earthquakes 1896–1903. Trans. Am.

Maso, M.S., 1927. Philippine earthquakes. In: Weather BureauGeophys. Union 37, 608–614.Seismological Bulletin for 1926, 67–103.Gutenberg, B., Richter, C.F., 1954. Seismicity of the Earth.

Milne, J., 1912. Catalogue of destructive earthquakes, Reportsecond ed., Princeton University Press. 310 pp.of the 80th Meeting of the British Association for theHamburger, M.W., Cardwell, R.K., Bryan, I.L., 1983. Seismo-Advancement of Science, Portsmouth, 649–740, App. 1.tectonics of the Northern Philippine island arc. In: Hayes,

Morante, E.M., 1974. The Ragay Gulf earthquake of MarchD.E. (Ed.), Tectonic and Geologic Evolution of the South-17, 1973, southern Luzon, Philippines. J. Geol. Soc. Phils.east Asian Seas and Islands. AGU Geophys. Monograph

Series 27, 1–21. XXVIII, 1–30.


Musson, R.M.W., 1998. Inference and assumptions in historical earthquake, the Philippines, using aftershock observation.J Natural Disaster Sci 13 (1), 29–38.seismicity. Surv. Geophys., 189–203.

Nakamura, S., 1978. On statistical tsunami risks of the Philip- Smith, W.H.F., Sandwell, D.T., 1996. New global seafloortopography from satellite altimetry. EOS Trans. Am. Geo-pines. South East Asian Stud. 15 (4), 581–590.

Nakata, T., Sangawa, A., Hirano, S., 1977. A report on tectonic phys. Union 77, 46 Suppl., 315.Solidum Jr, R.U., Sabit, J.P., 1988. Report on the June 20, 1988landforms along the Philippine fault zone in the northern

Luzon, Philippines. Sci. Rep. Tohoku Univ. 7th Ser. earthquake and its aftershocks offshore of San Jose, Occi-dental Mindoro. PHIVOLCS unpublished Quick Response(Geogr.) 27 (2), 69–93.

Nakata, T., Tsutsumi, H.R.S., Rimando, P.R.E., Daligdig, J.A., Team (QRT ) reports, 39 pp.Southeast Asian Association of Seismology and EarthquakeDaag, A.S., Besana, G.M., 1996. Surface Fault Ruptures of

the 1990 Luzon Earthquake, Philippines. Research Center Engineering (SEASEE), 1985. In: Series on Seismology(Philippines) IV. 843 pp.for Regional Geogr. Hiroshima University Spec. Publ. no.

25. 86 pp. Stewart, G.S., Cohn, S.N., 1977. The August 16, 1976 Minda-nao Philippine earthquake (Ms 7.8) — evidence for a sub-National Media Production Center, 1980. Directory of Philip-

pine Churches. duction zone south of Mindana. EOS. Trans. AGU 58,1194 abstract.Nieremberg, J.E., 1635. Historiae Naturae, Antwerp XVI.

Nuttli, O., Zollweg, J.E., 1974. The relation between felt area Su, S.S., 1969. The Luzon earthquake of 1 August, 1968: apreliminary report. Bull. Seismol. Soc. Am. 59, 459–472.and magnitude for central United States earthquakes. Bull.

Seismol. Soc. Am. 64, 73–85. Suppe, J., 1988. Tectonics of arc-continent collision on bothsides of the China Sea: Taiwan and Mindoro. Acta Geol.Nuttli, O.W., Bollinger, G.A., Griffiths, D.W., 1979. On the

relation between modified Mercalli intensity and body-wave Taiwanica. Sci. Rep. National Taiwan Univ. 26, 1–18.Street, R., Lacroix, A., 1979. An empirical study of New Eng-magnitude. Bull. Seismol. Soc. Am. 69, 893–909.

Perrey, A., 1860. Documents sur les tremblements de terre et land seismicity: 1727–1977. Bull. Seismol. Soc. Am. 69,159–175.les eruptions volcaniques dans l’archipel des Philippines.

Mem. Ac. Imp. Sc. Arts et Belles-Lettres de Dijon VII, 85. Thenhaus, P.C., Hanson, S.L., Algermissen, S.T., Bartolome,B.C., Bautista, M.L.P., Punongbayan, B.J., Rasdas, A.R.,Philippine Atmospheric Geophysical Atmospheric and Services

Administration (PAGASA), 1976. Catalogue of Philippine Nillos, T.E., Punongbayan, R.S., 1994. Estimates of theregional ground-motion hazard in the Philippines. In: Proc.earthquakes, 1585–1976. Earthquake Catalog. PAGASA,

Quezon City, Philippines. National Conf. on Natural Disaster Mitigation, QuezonCity, Oct 19–21. PHIVOLCS, Quezon City, pp. 45–60.PHIVOLCS Quick Response Teams, The 15 November 1994

Mindoro earthquake, PHIVOLCS Special report no. 2, Tuttle, M.P., Sykes, L.R., 1992. Reevaluation of several largehistoric earthquakes in the vicinity of the Loma Prieta andQuezon City, 38 pp.

Pinet, N., Stephan, J.F., 1990. The Philippine Wrench Fault peninsular segments of the San Andreas Fault, California.Bull. Seismol. Soc. Am. 82, 1802–1820.System in the Ilocos Foothills, Northwest Luzon, Philip-

pines. Tectonophysics 183, 207–224. Umbal, J.V., Masigla, L.M., Medrano, R.N., Diolata, G.P.,1990. Report of investigation on the February 8, 1990 earth-Repetti, W.C., 1946. Catalogue of Philippine earthquakes,

1589–1899. Bull. Seismol. Soc. Am. 36, 133–319. quake in Bohol. PHIVOLCS unpublished report,27 pp.+figures.Ringenbach, J.C., Stephan, J.F., Maleterre, P., Bellon, H., 1990.

Structure and geological history of the Lepanto–Cervantes US Geological Survey (USGS), 1995. Evaluation of the earth-quake potential of the Marikina Fault. Final report to theReleasing Bend on the Abra River Fault, Luzon Central

Cordillera, Philippines. Tectonophysics 183, 225–241. Agency for International Development, 13 pp.+figures.Utsu, T., 1990. Table of Disastrous Earthquakes in the WorldRowlett, H., Kelleher, J., 1976. Evolving Seismic and tectonic

patterns along the western margin of the Philippine Sea from Ancient Times to 1989.Wessel, P., Smith, W.H.F., 1995. New version of the genericPlate. J. Geophys. Res. 81, 3518–3524.

Rudolf, E., 1887. Ueder submarine erdbeben und eruptionen. mapping tools released. EOS. Trans. Am. Geophys. Union76 (33), 329.Gerl. Beitr. Z. Geophysik. Stuttgart I, 360–365.

Seno, T., Kurita, K., 1978. Focal mechanisms and tectonics in Wolfe, J.A., Self, S., 1983. Structural lineaments and Neogenevolcanism in southwestern Luzon. In: Hayes, D.E. (Ed.),the Taiwan–Philippine region. J. Phys. Earth 26, 249–263.

Shibutani, T., Ohkura, T., Iio, Y., Kanao, M., Nihigami, K., Tectonic and Geologic Evolution of the Southeast AsianSeas and Islands. AGU Geophys. Monograph Series 27,Iwata, T., Kakehi, Y., Hirano, N., Ando, M., Bautista, B.C.,

Puertollano, J.S., Lanuza, A.G., Melosantos, A.A., Chu, 157–172.Wood, H.O., Neumann, H., 1931. Modified Mercalli intensityA., Pigtain, R., dela Cruz, E., Punongbayan, R.S., 1991.

Search for the buried subfault(s) of the 16 July 1990 Luzon scale of 1931. Bull. Seismol. Soc. Am. 21, 277–283.