Climate during the Holocene

3.2.3.a PALYNOLOGICAL TIME SERIES The climate variability during the Holocene was investigated using palynological time series and archaeological and historical records. Sakaguchi (1982) described climatic variability during the last 7600 years based on a continuous section of peat samples taken from the wall of a pit in the raised bog of the Ozheghara moor, located near the head of the Tadami river, which flows into the Sea of Japan, 150 km north of Tokyo (site P73). The age was determined by the corrected 14C ages of the peat taken at 50 cm intervals and, additionally, by the dates of the tephra beds, which are intercalated in the peat layers. The age of the latter is known either through ancient manuscripts or archaeological findings. For ages younger than 2.45 kaBP, the non-corrected 14C readings were used because the difference between the corrected and the non-corrected readings are small (the corrected being younger by 20-80 years). At 450 cm deep, the non-corrected age is 7.03kaBP and is slightly over the correction limit given by the tree ring of Pinus aristata, dated at 4760 BC. Sakaguchi then suggested a correction factor of +600 years. Thus, the corrected age of 7.03kaBP should be 7.6kaBP. The Ozegahra moor lies 1400 m above sea level. The trees in this zone are mainly deciduous. Conifer trees increase in number above 1550 m of which some species cannot withstand snow cover, so their decline indicates cold temperatures and snow cover.

Sakaguchi (1982) took the percentage of Pinus as the characteristic for climate change as it reflects the flourishing or decline of the sub-alpine conifer forest zone, which, in turn, reflects the downward or upward shifting of the vegetation zones (i.e., the fall or rise in air temperatures). In a later paper (Sakaguchi, 1983), he generalized the palynological diagram of site P73 and compared it with paleo-climatic diagrams from all over the world. A correlation with Japanese culture was also performed. The following sequence was obtained:

1. 5660-4500 BC: a warm stage (Early Jomon warmest stage);

3. 2450-2270BC: cold (Middle/Late Jomon cold stage);

4. 2270-2090 BC: warm (first Late Jomon warm stage);

5. 2090-1460BC: warm;

6. 1460-1250 BC: warm (second late Jomon warm stage);

7. 1250-870 BC: warm/cold (transitional stage);

8. 870-400 BC: cold (latest Jomon cold stage);

10. 20-240 AD: warm (cold transitional stage);

11. 240-730 AD: cold (Kofun cold stage);

12. 730-1300 AD: warm (Nara-Heian-Kamakura warm stage);

13. 1300-1900 AD: cold (Little Ice Age).

Sakaguchi (1983) regarded the period between 870 and 400 BC (his dates 866-398 BC) as the first long-term cold stage since

7.6 ka BP. During this period, according to various reports quoted by the author, the water level of Lake Ogawara sank to below the present level (c. 3kaBP) and a regression phase occurred in the Tsugaru plain, evidenced by peat layers and buried trees (reaching a peak c. 2.5 kaBP).

Another cold stage (following the Yayoi warm stage, which began c. 400 BC), which Sakaguchi (1983) calls the Kofun cold stage, started c. 2 ka BP (17 AD) and lasted until 730 AD with cold climaxes appearing around 270 AD and 510 AD. A warm episode occurred c. 390 AD. The author quoted a few sources that present geomorphologic evidence showing that the sea was 2-3 m below that of the present during the cold stages. Sakaguchi (1983) came to very interesting conclusions regarding the impact of the climate of the Kofun cold stage, the intensity of which, he claimed, exceeded that of the Little Ice Age. This period was called Kofun because huge earthen grave mounds, called kofuns, were constructed for the ruling class during this period. These, with other historical evidence, tell of a highly organized centrally controlled society. The cold climate was not favorable for a rich rice harvest (irrigated rice cultivation in Japan is considered to have begun in the latest stage of the Jomon period in North Kyushu and then to have spread to eastern Japan in the Yayoi period). Moreover, the period was characterized by frequent floods. These conditions led to the development of elaborate agricultural techniques, especially of a type of irrigation of the paddy fields. The culmination of the Kofun culture corresponded with the warmer interval c. 400 AD.

Later, Sakaguchi (1989) revised his diagram (Fig. 3.3). In his paper of 1989, Sakaguchi also described a pollen section of a core 940 cm long taken from the Oshima peninsula, in the southern part of Hokkaido Island, northern Japan (site K87). The age of the bottom layer is 23.7 kaBP. On the basis of ratios between the different pollens, the author divided the coldness levels into a scale ranging from 1 (warmest) to 5 (coldest), when 4 to 5 was the range during the last glacial period, 4-2 from 13ka to 8.8kaBP, with short cold periods of grade 4 from 10.2 ka to 10 ka and from 9.4 ka to 8.8 kaBP. The warm climate continued up to the present, with an intercalation of a cooler climate from 1.55 ka to 0.89kaBP.

Generalizing Sakaguchi's (1989) pollen diagram of site K87 by calculating a running average diagram for the abundance of Betula (birch), Alnus (Alder) and Quercus (oak) pollen (Fig. 3.4), one can see that, in the lower part of the Holocene (and the same goes for the pre-Holocene period, which is not shown on the present diagram), one can observe an abundance of Betula pollen, while that of Quercus is very low. In other words, the Betula and Quercus curves change inversely to each other. For this period, it can be concluded that the abundance of Betula and scarcity of Quercus were mainly a function of a colder period and vice versa. During the period between 7 ka and 4 ka BP, however, this relationship cannot be observed, as both Betula and Quercus were relatively rather low. During the same period, the non-arboreal pollen was

Pollen time series Japan P73 PolIen time series ^ K87

Modified according to Sakaguchi (1983)* Modified according to Sakaguchi (1989)*

Pollen lime series Japan Modified according to Tsukada (1986)*

Levant History Paleo-climate

Ottoman Mamluk


Byzantine Roman Hellenistic Persian

Iron Age

Middle Bronze Age

Early Bronze Age

Pottery Ncolilhic A+B

Pre-potteiy Neolithic B

Prc-pottcry Ncolilhic A

Little lee Age cold-hum id cold-humid cold-humid cold-humid warm-dry warm-dry 2 1

vvarm-drv warm-diy cold-humid w arm-dry cold warm cold-

Fig. 3.3. Holocene pollen time series for Japan. *Adjusted to scale and streamlined (3-5) points by the running average method.




Little Ijce Age Nara-Heian Kamakurï _ warî istagë * î TheiKofun j cold stage

Lateft Jomon cofd stage

Secon<j Jomon i warni stage p> f"irst LatefJomon ¿cc > 1 warm Staee .Middle Late Jomon \ cold stage


Fig. 3.4. Holocene pollen time series for Japan generalized from the P73 site data in Sakaguchi (1983) and K87 site data in Sakaguchi (1987). (See text for details.) *Adjusted to scale and streamlined (3-5) points by the running average method.

relatively abundant. The fact that Pinus was also at a minimum may suggest that, during this period, the environmental situation, although relatively warm, was for other reasons not favorable for Pinus or Quercus.

Consequently, when trying to correlate the diagrams of sites P73 and K87 (Fig. 3.3), and taking into account the different climatic and, thus floral, environments, it is suggested that we look at what happened during the extreme and sufficiently lengthy climatic phases, which can be assumed to have had some impact on the floral environment all over Japan. (One should also take into account variations in the dating system between the two sites: the dates of site P73 were carried out at 50 cm intervals and these age measurements were corrected, while at site K87, the density of dating was not as precise, with no report of age correction.) With these considerations in mind, one can nevertheless see that one of the main long cold periods (i.e., the

Kofun cold stage) caused an increase in the pollen of Betula and a reduction in Quercus. Later, the Nara-Heian-Kamakura warm stage was characterized by more Quercus pollen. Yet, Sakaguchi (1989, p. 1) came to the conclusion that "K87 possibly belongs to a different climatic region from P73, at least after 8.8kaBP". I would reinterpret this conclusion by suggesting that, while the geographical environments of the two sites went through the same climatic changes, the process had distinct effects during the extreme and lengthy changes in climate but less distinct effects on the flora when, climatologically speaking, the changes were less distinct.

Miyoshi and Yano (1986) investigated the pollen of a core drilled in a moor in the Chugoku mountains of western Japan. The section was divided into zones and sub-zones as follows:

zone A1, 20 ka to 14kaBP: sub-Arctic and dry; zone A2, 14ka to 11.5kaBP: sub-Arctic with cool-temperate/ dry-wet;

zone B, 11.5 ka to 8kaBP: cool temperate/wet; zone C1, 8 ka to 4 ka BP: temperate/wetter;

zone C2, 4ka to 1.7kaBP: cool temperate/wetter; zone D, 1.7kaBP to present: cool temperate/wetter.

In his study of the vegetation of Japan during the last 20,000 years, Tsukada (1986) used data from 40 separate locations. Figure 3.3 presents five out of his twenty-four Quercus pollen time series together with the climate change profiles based on the pollen diagrams presented by Sakaguchi (1983). As can be seen, a clear reduction in Quercus is evident at sites 5 and 6, which are at a latitude of c. 33° N, and at site 39 at a latitude of c. 44° N. However, one should consider that Quercus responds negatively mainly to a cold setback in temperate regions, while in a warmer climate a positive response depends on a number of factors: plenty of humidity and environmental factors, such as other competing floral societies.

As discussed above, the sequence of climatic changes suggested by Sakaguchi, (1983, 1989) based on the palynological section of site P73 is the most detailed and could be correlated, in terms of major changes, with other time series in Japan. Therefore it is suggested that Sakaguchi's division be regarded as the base curve for Japan. Unfortunately, this section does not extend prior to c. 8kaBP.

3.2.3.b GEOMORPHOLOGICAL OBSERVATIONS Being situated in a region of pronounced tectonic activity, Japan is also a region where the movement of sea terraces, owing to tectonic activity, has to be taken into consideration. Japanese scientists examined the sea shores of Japan and concluded that sea transgressions were responsible for high terraces dating from 6.5 ka to 5 ka BP (peaking between c. 6.5 ka and 6 ka BP). This post-glacial transgression is called the Omon transgression. Many deposits, known as middens, were left by this transgression. Between 3 ka and 2kaBP, the so-called Yayoi regression occurred. This was inferred from the occurrence of shallow buried valleys and the development of coastal dunes. Another regression, called the Middle Jomon minor regression took place c. 5 ka to 4 ka BP (Yonekura and Ota, 1986). Pirazzoli and Delibrias (1983), investigating Late Holocene and recent sea-level changes and crustal movement in Kume Island, claimed that the maximum Holocene mean sea level was probably reached c. 4 ka BP, while a gradual fall occurred between at least 3 ka and 1.2kaBP. Sakaguchi (1983) stated that the Middle/Late Jomon stage (2450-2270 BC) was characterized by a fall in seawater temperature, as evidenced by the 18 O composition of Meterix lamarcki (Sakaguchi, 1983, refering to data from Chinzei et al., who investigated the paleo-oceanography of the Pacific along the east coast of Honshu) which indicated a weak but rather long cold episode from 5 ka to 4kaBP. Sakaguchi (1983) also quoted Ota et al., who investigated the Holocene sea levels in Japan and found a minor regression between 5 ka and 4 ka BP, andMatsushima, who described a decrease in the number of warm water/shallow water molluscan assemblages in the deposits in the southern Kanto region from 5 ka and their disappearance c. 4.5kaBP.

However, high sea levels were recorded at 4ka and 3.5 kaBP. Sakaguchi (1983) correlated the climatic changes at site P73 with the number of other archaeological sites and shell middens at high levels and found that these both occurred from c. 4.3 BP to c. 3.3 BP.

Yonekura and Ota (1986) cite Furukawa, who dated shells in middens to c. 4 ka BP in the Nobi plain of central Honshu, and Hirai, who observed that Lake Ogawara was at a high level between 5 ka and 4 ka BP. As this lake is connected to the sea, it should be considered as a high sea-level mark.

Sakaguchi (1982) suggested that the period around the twelfth and thirteenth centuries was colder and humid, based on finding a layer of detritus. However, in a revised paleo-temperature curve (Sakaguchi, 1989), these periods were interpreted as being warmer. During the warmperiod between 730 and 1300 AD, called by Sakaguchi (1983) the Nara-Heian-Kamakura warm stage, the sea level was higher in the area of Fukuyama city (Sakaguchi, 1983; citing Kuwashiro (1965)).

During the Little Ice Age, which started in Japan c. 1300 AD, there was a warm episode at ca 1840 and an extremely cold period around 1890. During this last episode, in which storms were abundant, rice harvests were low, leading to starvation, panic and rebellions. Sakaguchi (1983) concluded that, overall, the Little Ice Age was relatively moist in Japan.

Based on historical and other relevant data from 1650 to 1983 AD, Takahashi (1987) analyzed the long-term variation in storm damage in Japan. He found a 70 year cycle, which tended to increase in frequency when the climate was warm. The analysis indicated the considerable influence of volcanic activity on climatic change and, hence, on storm damage over a time scale of several decades.

Mikami (1987) reconstructed the climate of Japan from 1781 to 1790 AD, in comparison with that of China, in order to derive data from processing Chinese historical documents. The Japanese climate in the late eighteenth century was estimated to have been very cool and wet, as severe famines caused by crop failure were frequently recorded. From 1782 to 1787, Japan experienced the most severe famine in its history, causing a decrease in its population by approximately one million. The statistical examination of the relationships between Japan and China show simultaneity of rainfall in Japan and central China, with opposite indications in north and south China. These are considered to be linked to global circulation patterns. For example, in 1978, when it was extremely dry and hot in Japan and middle China, the axis of the sub-tropical high (North Pacific high) was located to the north. Such a situation most probably also occurred in 1785. In 1980, when it was exceedingly wet and cool in Japan and middle China, the axis of this high was shifted to the south and the polar frontal zone extended from middle China to Japan. Such a situation most probably also existed in 1783.

Oguchi (1997) suggested that since the transition from the Pleistocene to the Holocene, which led to the northward shift of frontal zones, typhoons started to hit Japan more frequently, causing frequent and heavy rainstorms and leading to extensive erosion by running water. The resultant incised channels were cut into regolith as well as consolidated bedrock because of the strong erosive force during storms. Channel formation is still active in most basins. Accordingly, erosion rates and sediment yields subsequent to rapid channel expansion in these areas will increase significantly if future climatic changes lead to increased storms.


From the research studies surveyed in this chapter, it can be concluded that during the last glacial period, between 20 ka and 18 ka BP, the climate all over Japan was drier than today. However, the degree of difference between then and now varied from north to south. It was very dry in Hokkaido as well as in Honshu. This was so because the Siberian high-pressure zone was very strong in winter, causing a strong winter monsoon. Around 10 ka BP, while it was still colder than today, the northern regions of Japan were drier because the strong Siberian high still prevailed, while the central region was wetter. In the northern Japanese Alps, for instance, strong rains and floods occurred. Around 6kaBP, it was very warm all over Japan. The island of Hokkaido was dry, while central Japan was rather wet. The dry Hokkaido conditions may have been caused by the weak winter monsoon, which resulted in decreasing snow accumulation. After c. 4kaBP towards 3 ka BP, the climate became cooler, reaching a cold maximum around 2.5kaBP. During this period, most of Hokkaido was dry

(except in the deep south), while most of Japan was wet. The formation of sand dunes in Kyushu shows that this region was also dry (Yoshino and Urushibara, 1979).

As was discussed by Issar (1995b, pp. 55-65), a good correlation was found between China and the Levant and, therefore, it is only logical that such a correlation should also include Japan, as shown in Figure 3.4.

The most outstanding correlation is between the Kofun cold stage in Japan and the Roman-Byzantine cold stage in the Levant, as well as between the subsequent Nara-Heian-Kamakura warm stage in Japan and the Moslem-Arab warm period in the Levant.

From Sakaguchi's curves for the P73 and K87 sites (Sakaguchi, 1983, 1989), it can be concluded that the latest Jomon cold Stage, c. 3 ka BP, was actually composed of three sub-stages, at c. 3.3 ka, 3 ka and 2.6kaBP, which, in general, correlate with the Iron Age cold period of the Levant, which started c. 3.2ka and continued until c. 2.7 kaBP (Issar, 1995a, b). Suzuki (1979) maintained that the most extreme spell occurred at 3.5 kaBP.

The First Late Jomon and Second Jomon warm stages can be correlated with the MB and Late Bronze warm periods of the Levant, while the Middle Late Jomon cold stage can most probably be correlated with the EB cold period of the Levant.

An abrupt bend in Sakaguchi's curves towards a colder trend, which was not given a special name by him, correlates with the Lower and Middle Chalcolithic cold period of the Levant, if one agrees to push this phase a few hundred years back on the time dimension. As the accuracy of the age determinations of the climatic changes during the lower part of the Holocene is in the order of magnitude of a few centuries, (because of the difference between corrected and non-corrected 14C dates), such a correlation looks feasible. The same can be said with regard to the Middle Neolithic cold period of the Levant and the cold period of the Base Neolithic.

0 0

Post a comment