There remain many challenges in using historical documents to reconstruct a reliable history of earthquakes in Japan. Previous catalogs of historical earthquakes in Japan are not conclusive and contain uncertainties about date, hypocenter, magnitude, and tectonic interpretation of each earthquake. There is no database of digital texts of historical documents, which describe each earthquake. Since the density of historical records in Japan is temporally and spatially heterogeneous, seismologists should carefully remove apparent changes of earthquake frequency, which are caused by the heterogeneity of record density. There is, however, no detailed database of the density variation of historical records. The number of researchers, who are interested in historical earthquakes, is small. The situation stated above is caused mainly by the multi-disciplinary character of historical seismology. Japanese seismologists, who usually have little knowledge of history and classical literature, are not qualified to read a historical document and evaluate its reliability. The environment for research on historical seismology is, however, getting better. Japanese historians have published and are still publishing many historical documents, sometimes with translations into modern language. Evaluations of the reliability of each document can easily be done by referring to historical dictionaries or other databases. All these publications and information are available in many libraries. It is now easy and stimulating for many seismologists to read, evaluate, and interpret historical documents.
Understanding historical and pre-historical earthquakes is essentially important in estimating site, magnitude and timing of future earthquakes. Records are supplied mostly from historical documents we have in Japan dating from about 1, 300 years ago and from natural records found in topography and geological formations in earthquake areas. The usefulness of collating historical and natural records is shown from some historical earthquakes. For example, the Tsukushi Earthquake of AD. 679 was reasonably inferred to have occurred from the Minoh fault zone in Fukuoka Prefecture, Kyushu, based on historical documents and on data from excavations of the ground just on the fault line. Dating the last activity of an active fault is indispensable for the long-term prediction of earthquakes. For example, the AD. 1596 earthquake was revealed by ground excavations to have occurred from the Arima-Takatsuki fault zone. This means a significant decrease in the possibility of a large earthquake in the near-future at the location. The concept of the characteristic nature of the size and the recurrence timing of the earthquake from a fault, on which long-term prediction is currently based, was derived from good examples of fault activity recorded on Japanese historical documents and nature.
Japan has a database of earthquake damage, which includes the damage to buildings, and the direction of overturned chimneys, gravestones, and that of collapsed wooden houses. These data reflect the strong ground motions generated from the earthquake source process. The spatial distributions of damage and directions of overturned bodies are simulated by a 3D dynamic rupture model, and the possibility of inferring the earthquake source process from the historical data is discussed.
Liquefaction features from paleoearthquakes are common at archaeological sites of Japan. The most remarkable features are dykes of sand and gravel in which the clasts become finer approaching the surface. Liquefaction at some sites occurred not only in sand but also in gravel. Some archaeological features have been entrained and moved in sand intrusions.
A large volume of historical documents in Japan show that great subduction earthquakes have repeatedly occurred along the Suruga-Nankai trough off southwest Japan since A.D. 684 with an interval of 100-200 years. They occurred as pairs of M8 events, one in the eastern half (Tokai earthquake) and another in the western half (Nankai earthquake), as was the case for the 1854 Ansei earthquakes, while sometimes occurring as single giant events like the 1707 Ho'ei earthquake. Although the space-time pattern of their recurrence is the best-known in the world, we should study more past events in order to understand the tectonophysical bases of their recurrence. In this respect I review the present understanding of historic Tokai and Nankai earthquakes and discuss related problems from the viewpoint of historical seismology. In this paper, the first of the three in all, I review the events until the early half of the 14th century. The keys to identifying older events are strong ground motion and damage in Kyoto, Nara, and Osaka, those in wider area of southwest Japan, tsunamis along the Pacific coasts of southwest Japan, typical coseismic vertical crustal movements of the Kochi plain, the Muroto and Oma'ezaki points, and the Ise and Suruga Bay coasts, temporal inactivity of specific hot springs, and aftershock activities recorded in Kyoto. The 684 Hakuho earthquake was definitely a Nankai event, and possibly included a Tokai event simultaneously (possibly Ho'ei type). The 887 Nin'na earthquake was also a definite Nankai event and was probably a Tokai event as well (Ho'ei type). The 1096 Eicho earthquake was clearly a Tokai event, but the following 1099 Kowa earthquake has some discrepancies that prevent it from being regarded as a M8 Nankai event. It is not clear yet whether great earthquakes occurred or not in the ca. 200 year intervals of 684-887 and 887-1096. It seems probable that great Tokai and Nankai earthquakes took place in the mid-13th century, but a more detailed investigation of historical seismology is required to discover the missing event.
Possible tsunami deposits were found at two archeological sites along the Pacific coast near Hamana Lake in central Japan. This region has frequently suffered heavy damage from large earthquakes along the Nankai trough. Three sand deposits covering layers that include archeological remains were found at one of the sites. The ages of the sand deposits were constrained by remains just below the sand deposits, which showed good correspondence to the ages of three major historic earthquakes that seriously damaged this area, i.e., A.D. 1707, 1605, and 1498. These features strongly suggest that the sand deposits were caused by tsunami from the historic earthquakes. Core samples obtained near this site showed that a similar sequence continues into a deeper part, suggesting the existence of tsunami deposits of events older than A.D. 1498, before which few historical records are available in the Tokaido side. Further investigation of tsunami deposits in this area would yield useful information on the reconstruction of large earthquakes along the Nankai trough.
One of the intriguing events accompanying large tsunamis, especially those that strike the Pacific coasts of Japan, is the luminous phenomenon; many historical Japanese documents have described how fire balls or pillars of fire seemed to come out from the sea when tsunami approached. Among 17 events of large tsunamis from 684 to 1946, where the surface wave magnitude was greater than 8 or the tsunami magnitude was greater than 3, nine tsunami events accompanied luminous phenomena. In spite the event's high probability, few explanations have been offered as to the source mechanism except luminescent planktonic organisms, which is hardly plausible because luminous tsunamis have been eye-witnessed even in the winter season when such planktonic organisms are less active, and even in the daytime when the intensity of light emitted from planktonic organisms is unlikely to exceed the day-time brightness. Most tsunami earthquakes are thought to be associated with sediments in the accretionary prism. One recent important finding is that large volumes of stable methane hydrate are present within ocean-floor sediments at water depths exceeding about 500m at lower temperatures. When the equilibrium conditions of coupled low temperature and moderate hydrostatic pressure are disturbed by an earthquake, the hydrate abruptly decomposes. Conversely, a breakdown of hydrate may cause a further mass movement, and a cascading chain of events may occur. Some eye-witness reports in historical documents strongly suggest that luminous phenomena associated with tsunami are attributable to methane hydrate disruption, not others causes such as luminescent planktonic organisms.
Immediately after the 1923 Kanto Earthquake (M=7.9), two large aftershocks of M=7.2 and 7.3 occurred in succession somewhere in the southern Kanto district. The first aftershock occurred about 3 minutes after the main shock and the second about 4.5 minutes after the main shock. Strong ground motions from these events and locations of their epicenters were examined mainly from data of 548 descriptions of personal experiences. It was deduced from them that shaking due to the first aftershock was severe in the Tokyo Metropolis and eastern Kanagawa prefecture. In consequence, there were many descriptions in the Tokyo Metropolis that the shaking was as strong as that due to the main shock and caused extensive damage. On the other hand, there were few descriptions for the first aftershock in the western area of the southern Kanto district, which is western Kanagawa prefecture, Yamanashi prefecture, and eastern Shizuoka prefecture. However, shaking due to the second aftershock was strong in this area. Some people living near the boundary of the three prefectures described that its shaking was as strong as that due to the main shock. Comparing the facts described in these personal experiences with the distributions of seismic intensities from other M=7 class earthquakes occurring in the southern Kanto district after the 1855 Ansei Edo Earthquake, it was concluded that the epicenter of the first aftershock was located in and around northern Tokyo Bay and that the epicenter of the second aftershock was in eastern Yamanashi prefecture. The result for the second aftershock was consistent with the epicenter determined from the seismic records by the Kumagaya Meteorological Observatory and with the epicentral distance estimated from data of S-P time at the Gifu Meteorological Observatory. The characteristics of the sequence of strong shakings within 5 minutes after the 1923 Kanto Earthquake could be elucidated in the southern Kanto district from the results of the present study.
This paper reviews the contributions to disater management made by seismology based on historical documents and artifacts. The main task was to predict when, where, and how large future earthquakes would be based on knowledge of past earthquake diasters and geographical observations for earthquake prediction. They have been the main information source for medium and long-term earthquake predictions for occurrence. As for location and maginitude, it has been the main information source for identifying maximum credible earthquakes in each region. It also helps to recommend wise land use planning by avoiding active fault zones to mitigate future damage. As a further potential application, it could be used to give us an account of how people and society responded to the catastrophic disasters in the past to give us a lessons for improving preparedness for the next disaster.
It is well known that, in Japan, very high seismic activity can be traced back to more than one thousand years ago based on old written documents. Many seismologists and tsunami researchers have been receiving benefits from collected and published documents on historical earthquakes and related phenomena (hereafter, collected documents). However, it had been stated for long time that it is difficult to use the collected documents properly because of their huge amount. In fact, only a few “experts” could use them. Besides, it was also a fact that some studies made mistakes with careless reading of the collected documents. To improve this situation, a committee was established in the National Research Center for Disaster Prevention (at present, National Research Institute for Earth Science and Disaster Prevention) to construct an electronicdata-base system of the whole collected documents. The committee is now sleeping after constructing a small data-base system and its handling software. Recently, a new movement occurred to make a data-base. Them, it is not meaningless to summarize the activities of the committee about 10 years ago and clarify the problems that arose through the constructing processes of the small data-base system. Major problems pointed out were; 1) huge amount, 2) no indices except time, 3) mixture of documents of different qualities. Whereas the seismic (geological, geophysical) time scale is long compared to that of a human life, we much owe to historical records when we discuss seismicity in some regions. In other words, the collected documents are a privilege resources of our Japanese scientists. Accomplishing easy access and easy usage of collected documents based on a data-base system is a key to future developments in seismology and related studies.