Historical tsunamis in South China

Abstract

An accurate assessment of tsunami risk of a region requires a credible record of past tsunami events in the region. Existing surveys on historical tsunamis of South China have not presented a consistent list of events. The current report makes reference to original historical literature and evaluates the validity of suspected tsunami events in published surveys. A set of refined historical data for further investigation of the tsunami hazard in the region was produced. Only two events have been identified as credible reports of tsunami in the current study. Some events previously considered as tsunami, including a few with great reported casualties, are found to be unsubstantiable.

Appendix

1.1 A1. Discussions on selected suspected events

• [1] This event was regarded as a possible tsunami in some studies (see Table 1), probably based on reports like [Q1]. It was described as a storm surge in some other reports [e.g., Q2]. This is an example of an event recorded in different completeness in different literatures (see Sect. 3, and compare [Q1] and [Q2]).

• [2] This was a significant earthquake that affected a wide area of the Fujian province. Reported ground-shaking extended approximately 450 km from northeast to southwest along the coast of the province. Sunken boats/ships were reported during the earthquake [Q3].

  The context of the literatures may provide some more hints about the tsunami. Reports of this event were abundant (c.f. Chinese Academy of Sciences 1956, p 809), implying that sufficient attention of historians was attracted to record the damage in detail. The condition of the sea during the earthquakes was mentioned in only one document, implying that the damage caused by any sea-wave was minor and localized. This was unusual for a tsunami generated by such a large earthquake, presumably originated by a long fault. Based on earthquake damages described in the literature, Zhuo and Chen (1983) suggested a strike-slip motion for the fault responsible for the earthquake, which is consistent with fault-plane solutions for modern earthquakes in the area. Such a fault motion is believed to be inefficient to generate tsunami. It is speculated here that the earthquake did not generate a tsunami directly by vertical fault motion. Instead, it induced a submarine landslide inside a harbour near the capital city, causing a local tsunami that damaged the ships. In any case, because this event is a damaging (Sect. 4.1.5) unusual wave (Sect. 4.1.1) associated with an earthquake (Sect. 4.1.4), it is considered here as a tsunami.

• [3] This was the strongest historical earthquake that occurred in the Hainan region. The fault motion was reported to be strike-slip. Although ground subsidence at the coast was clearly recorded in the literature, no observations of a seismic sea wave were reported. A modern field survey suggested that the ground subsided for 2.5–3.5 m during the earthquake (Chen and Huang 1979).

• [4] Some studies (Lee 1988; NGDC database) noted the date of this event to be 16 September 1640. However, the precise date for this event was not recorded in the historical literature. Instead, it was recorded that the event occurred during the period between 16 September and 14 October (equivalent to the 8th month of the 13th Chongzhen year). Some records described the event as a sea-surge associated with an earthquake [e.g., Q4], which is probably the reason that it was regarded as a possible tsunami by some existing surveys. However, it was clearly documented in the earliest accessible report of this event [Q5] that the earthquake was not associated with the sea-surge but a separate event. Similar to event [1], Keimatsu (1963) misinterpreted this event as a tsunami.

• [5] This earthquake occurred at the same place as event [4]. Zhou and Adams (1986) assigned an approximate date to this event between 16 September and 26 November 1641 but other catalogues (Chinese Academy of Sciences 1956; National Seismological Bureau, 1995; Guangdong Historiography Committee 2003) stated the exact date of the mainshock to be on th November. No observations of the sea were reported.

• [6] There were two different dates given for this event, namely 8 January and 15 February 1661. The studies referred in Table 1 used either or both dates. Fang (1969) gave a detailed review to this event. It was reported in both Chinese and western materials that there was an earthquake in early 1661 at southern Taiwan (Tainan). The records in the Chinese literature were so brief that even the date was not explicitly noted. Two pieces of western material, written in German and Dutch, respectively, described the event in greater but inconsistent detail. The Dutch report stated that the earthquake occurred on 15 February and lasted for 6 days. It did not mention any observations of the sea. The German report stated that the earthquake was somewhere in January and lasted for 6 months. Three ships entering the harbour were severely shaken and one nearly sunk. Seawater was observed “to be drawn into the sky”, but not recorded as damaging.

• [7] This earthquake was briefly reported in a few literatures (c.f. Fang 1969), with one including an observation of the sea [Q6]. Based on the context, it is clear that it was a storm that induced the large sea wave, not ground-shaking. The storm and its induced-damage was recorded by official dynastic document (First Archive of China 1996), while the earthquake was not, implying that it was unlikely to be an earthquake-generated tsunami.

  In the same year there was a confirmed earthquake on 5 January 1792 NTOU [n.d.] pointed to the possibility that it could be tsunamigenic, based on the description of a sea-surge and an earthquake in the same literature [Q7]. However, it is ambiguous whether both the earthquake and the sea-surge occurred coincidentally (see Sect. 3). Fang (1969) summarized that there were three earthquakes in the 59th Kangxi year (between 8 February 1720 and 27 January 1721). No record of any sea-surge was mentioned associated with the three earthquakes, so they are here not considered tsunamigenic.

• [8] It was reported that there was a high tide of 30 feet in the neighbourhood of Canton (Guangdong), China, that swept away 4,000 houses and swallowed up 10,000 inhabitants (Mallet and Mallet 1858; The Annual Register 1766). The Russian database noted that the place of occurrence as Guangzhou (capital city of Guangdong province). No earthquake was reported associated with this event, so it did not pass the validation (see Sect. 4.1.4).

  Due to the huge damage claimed, the discussion here is continued assuming that the event is validated. This event was not recorded in any Chinese historical literature, which was virtually impossible for such a large scale disaster. Crop yield and rice price in Guangdong was reported to be normal in official dynastic documents (National Palace Museum 1982a), implying that it was unlikely that a great disaster has occurred during the period. Therefore, this event did not pass the confirmation, even if it is assumed to be validated.

  The event was marked “doubtful” in the NGDC database. Yet, it was still regarded as a major historical tsunami disaster in the world, probably due to its huge casualty claimed, by some recent researchers (e.g., Bryant 2001).

• [9] The earliest traceable record of this event was a letter sent to a Fellow of the Royal Society (de Visme 1769). It was an indirect record that quoted an account of an earthquake from another letter. It was reported that a major shock and four aftershocks were felt, and ships in harbour were shaken and whirled about. Berninghausen (1969) and Cox (1970) believed that this event actually referred to the 1765 event (event [8]) because it was not found in the catalogue compiled by Iida et al. (1967).

  The report was not verified by any Chinese literature, probably due to the minor damage. It was a contemporary report with considerable details, making it reliable.

• [10] This sea-surge occurred somewhere between 24 April 1781 and 21 June 1781, at Qieteng Bay, Taiwan (near the current Dapeng Bay, Donggang, approximately 30 km southeast to Kaohsiung) [Q8]. The flood was reported to last for a few quarters, which is a typical period of a tsunami. Based on conditions 1 and 2 (see Sect. 4.1.1 and 4.1.2), this event is sufficient to be validated as tsunami. The weather was reported to be fine, eliminating the possibility of a storm surge. No earthquake was reported in the region in this period, suggesting that the tsunami was probably not tectonic in origin. The run-up was reported both quantitatively (tens of zhangs) and qualitatively (about a bamboo high). One woman was reported to be drowned. The reporter’s name was also recorded, which was not a common practice in local historiographies, giving support to the reliability of the report. The event was not recorded in the official dynastic record, possibly because the damage was minor. This event is the only tsunami with explicit damage (one life and several huts) in South China in the history.

• [11] It was reported that Taiwan was shaken by an earthquake, associated with a flood extended 120 km for inland, and 40,000 casualties was claimed (Perrey 1862, cited by Soloviev et al. 1974). The water was reported to remain above ordinary level for 8 h (Mallet and Mallet 1858), which is too long for a tsunami. Yang (1987) incorrectly reported the date of this event to be 22 May 1781, and considered this and event [10] as the same event, probably due to the misread of the dates. The same mistake was made in some later studies (e.g., NTOU, n.d.; Ye et al. 2005). Bryant (2001) regarded the event as the tsunami with the highest death tolls in the Pacific Ocean region (before 2004).

  The validity of this event is highly questionable because the event was totally unrecorded in any Chinese literature, which is virtually impossible due to its significance. From the official dynastic record (National Palace Museum 1982b), there was a storm at Taiwan on 2 June 1782. The storm was reported to have induced some damage, but the crop yield of the whole province (including Fujian and Taiwan) was reported to be not affected. Thus, a great disaster occurring in the region during that period was highly unlikely. The casualty claimed (40,000 lives) was approximately 5% of the whole contemporary population of Taiwan, which could have seriously damaged the stability and economy of the society. This event is invalidated due to its unreasonably long period (see Sect. 4.1.2). Even if it is assumed validated, it still could not pass the confirmation.

• [12] The damage of this significant earthquake was recorded in detail (c.f. National Seismological Bureau 1995). An agitated sea was observed in a harbour (Luermen, near Tainan) (Chen 1968; Bank of Taiwan 1993). Ships were reported shaken. Reports of land erosion occurred at the same year were suggested to be related to the earthquake (NTOU, n.d.). However, it was also reported that a flood occurred due to a heavy rain lasing for about a week before the earthquake (National Seismological Bureau 1995), which is more likely the cause of the land erosion. The damage of the earthquake was reported in detail in an official dynastic document (First Archive of China 1959) but no condition of the sea was reported. The earthquake damage inland was reported to be more severe than that at the coast, suggesting that the epicenter was located onshore. This idea is also suggested in National Seismological Bureau (1995).

• [13] Earthquakes occurred at the South China Sea or the Philippines, outside the study region. The 1872 event was reported to cause a tsunami and invaded a town near the epicenter (Lee 1988). No sea-surge was recorded in South China at those times.

• [14] An unusual wave associated with an earthquake was reported to sink boats in Tongan county, Fujian [Q9]. An ebb tide preceding a flood, which is characteristic of a tsunami but not a storm surge, was reported.

  This event did not pass the confirmation because of two reasons. First, the description of this event in all later studies (e.g., NGDC and Russian databases; Iida et al. 1967; Berninghausen 1969; Cox et al. 1980; Lee 1988) were ultimately based on Keimatsu (1963), who was the first to notice the report in [Q9], which is the unique record of the event. The event was not verified by local historical literature of nearby counties, contemporary newspapers or instrumental seismic reports. Second, seismographic stations have been installed by Japanese scientists in Taiwan since the 20th century (National Seismological Bureau 1995). The nearest station to the Tongan county was no more than 400 km. It is unusual that the earthquake claimed was not recorded.

• [15] A docked ship at Shantou was reported to touch ground during the earthquake [Q13], which could be due to an ebb tide or a risen sea bottom. If it was the former, the event is validated.

  A sea wave with 7-meter high was reported to be associated with this earthquake (c.f. NTOU, n.d.; Ye et al. 1993). However, the source of information is considered as unreliable for two reasons. First, the period of the wave was reported to be in seconds, which is far too short for a tsunami. Second, none of the sources of information are contemporary reports.

  Based on contemporary reports, this event did not pass the confirmation due to two reasons. First, it was reported that the rumor about a catastrophic tidal wave in Amoy was not true [Q14]. The damage in Amoy was reported to be slight and no sea-surge was observed (Shen Bao, 23rd January 1918). Second, it was reported that westerners in Shantou took refuge on ships (Shen Bao, 26th January 1918), which is an unusual behaviour if the coast was just attacked by a tsunami.

  Note that the region is relatively high in seismicity. Significant earthquakes have occurred at this region in 1067, 1600, 1640 (event [4]) and 1641 (event [5]). No tsunami was reported associated with these events.

• [20] A sea-surge in 1795 was reported to affect eight counties, covering 300 km of coastline of the Fujian province [Q10]. Further investigation revealed that it was actually a flood occurred in autumn, 1794 (the 8th month of the 59th Qianlong year); affecting about four counties [Q11], caused by a heavy rain and river-flooding [Q12]. This was again an example that an incomplete historical record could lead to misunderstanding (see Sect. 3).

1.2 A2. Quotations of historical literatures

Notes (in italic text) were added to facilitate understanding. The source of the quotation (in bold text) was given in its original language.

• Q1 “In the 10th month, a sea-surge occurred at Haiyang and Chaoyang counties.”

  十月, 海陽、潮陽二縣海潮溢。〈〈宋史.五行志〉〉

• Q2 “In the 10th month, 9th Xining year, Song Dynasty, a typhoon and a sea-surge occurred, damaged houses and crops. The situation of the Haiyang and Chaoyang counties were the same.”

  宋神宗熙寧九年十月, 颶風, 海潮漲溢, 壞民居及禾稼, 海陽、潮陽二縣同。〈〈光緒十年潮陽縣志〉〉

• Q3 “On the 8th of the 11th month, 32th Wanli year, an earthquake occurred; In the night of the ninth, a great earthquake occurred, (ground shaking) started from the northeast (and migrated) to the southwest. More than ten tremors occurred in this night. Mountain, rock and the sea were shaken. There were some ground-cracks, especially at the capital city (Quanzhou) (... Houses collapsed inside and outside the city. A lot of boats/ships were sunk.”

  (萬曆) 三十二年十一月初八日地震, 初九夜大地震, 自東北向西南, 是夜連震十餘次, 山石海水皆動, 地裂數次, 郡城尤甚, 開元東鎮國塔第一層尖石墜, 第二第三層扶欄因之併碎, 城內外廬舍圯、覆舟甚多。〈〈萬歷泉州府志〉〉

• Q4 “In the 8th month, 13th year, there was a sea-surge and frequent earthquakes.”

  (明崇禎) 十三年八月, 海溢, 地屢震。〈〈康熙二十三年潮州府志〉〉 〈〈乾隆揭陽縣〉〉

• Q5 “In the 8th month, 13th year, there was a sea-surge. In winter, there were frequent earthquakes.”

(明崇禎) 十三年八月, 海溢 ; 冬, 地屢震。〈〈順治十八年潮州府志〉〉

• Q6 “In the 8th month, there was rain and storm. It was sultry at night. Sea water rose high. Ships in harbour collided with each other. The ground shook. Not many houses were left undamaged in the county. Thousands of inhabitants were killed by collapsing houses or drown.”

  是年 (康熙六十年) 八月, 臺灣怪風暴雨, 流火天竟夜, 海水皆立, 港中諸船互相撞壞。地又大震, 郡治幾無完屋, 居民壓、溺死者以數千計。〈〈台灣小志〉〉 (龔柴　著)

• Q7 “... because of earthquake and sea-surge, people jointly organized a play to worship the god...”

……續因地震、海水冷漲, 眾百姓合夥謝神唱戲……〈〈明清史料戊篇.第一本, p.21;.朱一貴供詞〉〉

• Q8 “In around the 4th to 5th month, 46th Qianlong year, the weather was fine. Suddenly the sea roared like thunder. Giant wave appeared. Water rose for tens of zhang high (1 zhang approx. equals to 3–1/3 m). Villagers nearby were submerged. They climbed upwards, expecting to die. After a few quarters, it ebbed. People were swinging on top of bamboos, crying for help. One strong man jumped to ground, and helped others getting down. Gazing lands, farmlands and gullies were full of leaping fishes. Villagers nearby rushed to collect by baskets. It was heard that one woman was drown ... Fishmen sailed on top of bamboos on raft, watching their homes submerged from far. (reported by Confucian scholar Lin Shi-seng on the twentieth of the 2nd month, 10th Daoguang year )”

  乾隆四十六年四、五月間, 時甚晴霽, 忽海水暴吼如雷, 巨湧排空, 水漲數十丈, 近村人居被淹, 皆攀援而上至尾, 自分必死。不數刻, 水暴退, 人在竹上搖曳呼救, 有強力者一躍至地, 兼救他人互相引援而下。間有牧地甚廣及附近田園溝壑, 悉是魚蝦, 撥刺跳躍, 十里內村民提籃挈筒, 往爭取焉。聞只淹死一婦 ......漁者乘伐從竹上過, 遠望其家已成巨浸, 茅屋數椽, 已無有矣。(道光十年二月二十日恩貢生林師聖報) 〈〈台灣采訪冊〉〉

• Q9 “Earthquake occurred; sea ebbed then flooded. Many fishing boats were sunk.”

  地大震, 海潮退而復漲, 漁船多遭沒。〈〈民國同安縣志〉〉

• Q10 “In the 60th (Qianlong) year, the coastal low-lying farmlands in the Putian, Haicheng, Zhaoan, Longxi, Huian, Jinjiang, Tongan and Mahang counties were flooded by sea-surge. Some of them (were significantly damaged and) became disasters. The rent was waived or deferred to next year.”

  (乾隆) 六十年, 以莆田、海澄、詔安、龍溪、惠安、晉江、同安及馬巷近海低窪田禾猝被海潮淹浸, 分別成災不成災, 蠲免本年錢糧及緩至次年徵收, 各有差。 [陸人驥, 1984, 引述 〈〈同治重刊重纂福建通志.卷五十二〉〉]

• Q11 “... the rent of the 59th year of the four counties Longxi, Nanjing, Changtai and Haicheng under Zhangzhou prefecture, Fujian province; as well as the Huafeng sub-county, were deferred because of the flood in the 8th month last year. Also due to bad crop, the rent of the 59th year of the two counties Jinjiang and Nanan, as well as the Luoxi sub-county, was deferred to this year ...”

  ......福建漳州府屬龍溪、南靖、長泰、海澄四縣, 並華封縣丞所轄應徵五十九年份地錢糧, 因上年八月被水, 所有應徵銀兩, 及泉州府屬之晉江、南安二縣, 並羅溪丞所轄五十九年份錢糧, 亦因收成歉薄, 業經加恩, 俱著緩至本年徵收, 以紓民力 ......〈〈乾隆六十年九月二十七日上諭〉〉

• Q12 “Began at 5 pm on the tenth of the 8th month, until 1 pm on the thirteenth, there was a heavy rain at the capital city of the Zhangzhou county. Together with the river flooding, the water inside the city was a few zhang deep.”

  ......漳州郡城於八月初十酉刻起, 至十三日午刻大雨傾盆, 加以溪河漲發, 城內水深丈餘 ......〈〈乾隆五十九年九月十五日上諭〉〉

• Q13 “... a certain ship, 30 miles to sea off Amoy (Xiamen), and sailing in 15 fathoms of water felt the shock distinctly. Another ship which was in Swatow (Shantou) at the time, lying alongside the wharf, touched bottom.” [The China Mail, 15th February 1918]

• Q14 “Reports were current in the Colony (Hong Kong) yesterday, that Amoy had suffered as much, if not more than Swatow, and also that a tidal wave had caused tremendous damage. We are happily able to state that very little damage has been done in that city.” [The China Mail, 17th February 1918]