Estimation of the magnitudes and epicenters of Philippine historical earthquakes

The magnitudes and epicenters of Philippine earthquakes from 1589 to 1895 a re estimated based on the review, evaluation and interpretation of historic al accounts and descriptions. The first step involves the determination of magnitude-felt area relations for the Philippines for use in the magnitude estimation. Data used were the earthquake reports of 86, recent, shallow ev ents with well-described effects and known magnitude values. Intensities ar e assigned according to the modified Mercalli intensity scale of I to XII. The areas enclosed by Intensities III to IX [A(III) to A(IX)] are measured and related to magnitude values. The most robust relations are found for ma gnitudes relating to A(VI), A(VII), A(VIII) and A(IX). Historical earthquake data are obtained from primary sources in libraries i n the Philippines and Spain. Most of these accounts were made by Spanish pr iests and officials stationed in the Philippines during the 15th to 19th ce nturies. More than 3000 events are catalogued, interpreted and their intens ities determined by considering the possible effects of local site conditio ns, type of construction and the number and locations of existing towns to assess completeness of reporting. Of these events, 485 earthquakes with the largest number of accounts or with at least a minimum report of damage are selected. The historical epicenters are estimated based on the resulting g eneralized isoseismal maps augmented by information on recent seismicity an d location of known tectonic structures. Their magnitudes are estimated by using the previously determined magnitude-felt area equations for recent ev ents. Although historical epicenters are mostly found to lie on known tectonic st ructures, a few, however, are found to lie along structures that show not m uch activity during the instrumented period. A comparison of the magnitude distributions of historical and recent events showed that only the period 1 850 to 1900 may be considered well-reported in terms of magnitude distribut ion. Each earthquake is evaluated for its 'quality' of determination based on the number of intensity reports. Earlier than 1850, the data collected a re few and most earthquakes had fewer than ten reports. Good quality report s began to be collected from 1850, partly correlative to an increase in the number of towns and partly to the start of a systematized collection of ea rthquake accounts by the Manila Observatory. Parameters of these well-repor ted earthquakes may be used for conducting various seismological studies. E xamples of how the parameters of poorly reported events were arrived at are also discussed. (C) 2000 Elsevier Science B.V. All rights reserved.