Cold Periods During the Last Millennium

Studies of temperature change during the last millennium in China and in other regions of the globe were reviewed. Cold periods were identified for East Asia, USSR, Europe, North America, Polar Region and Southern Hemi­ sphere. Most of them were grouped into the same time intervals: first half of · 12th, second half of 13th, second half of 15th, ,17th �d 19th century. Temper­ atures in the last two cold periods were 0.5° C-1.0° Clower than the average of the 20th century. The first three cold periods were less cold and less uniform in geographical distribution and in temporal variations. Therefore, the first three cold periods may be regarded as the transition from the Medieval Warm Period to the. Little Ice Age. Long-term variations of solar activity and vol­ canism were compared with the changes of temperatures. The second· to the fourth cold periods seem to relate to the Wolf, Sporer, and Maunder Minima of solarmctivity. The intensification of explosive volcano eruptions in early 15th and 17th century may also have contributed to the occurrence of the third aitd fourth cold periods. However, only a slight increase of volcanism and a weak decrease in solar activity in the early 19th century can hardly fully intepret the severe cold period throughout the 19th century. Perhaps neither one nor both of the aforementioned factors can be responsible for all of the five cold periods identified during the last millennium. It is suggested· that the changes in cryosphere, biosphere and the oceanS may interact with the atmosphere, controlling (together with the external factors) the occurrence of the cold periods. Of course, anthropogenic factors should also bf! considered, especially for the last one and a half centuries. Finally, factor or factors other than those aforementioned, such as changes in orbital parameters, may also take part in regulation of the climate for the last millennium.


INTRODUCTION
Recently, climatic change during the last millennium was studied with greater and greater interest, because it is important not only for examining the mechanism and causes of natural variability of the climate in the past, but also for detection of the greenhouse effect at the present time and for prediction of climatic change in the future. Most notably, an International Symposium on the Little Ice Age climate was held at Tok yo, Japan 25-27 September 1991 and a Workshop on the Medieval Warm Period was convened at Tucson, the U.S. 5-8 November, 1991. Meanwhile, a series of papers has been published in the last decade. Present paper aims at reviewing recent research works, comprising the state of the art in reconstruction of the palaeo-climate, solar activity and volcanism, and finally providing a schematic picture of the temperature changes, especially the cold periods within the last millennium, and a preliminary interpretation of the latter.

COLD PERIODS SINCE A.D. 1000
East Asia: Curves in Figure 1 show the temperature changes in East Asia. The curve (d) was reproduced from the famous paper of Chu (1973). Curve (e) shows the 50-year running mean temperature anomalies in China referred to the average of the period from the 1880s-1970s. Horizontal lines correspond to the average of the last millennium (Wang 1991). Both the curves (d) and (e) were reconstructed on the basis of documentary records. However curve (e) was found from the average of four seasons, whereas curve (d) has a bias towards winter time, so shows a bigger amplitude than (e). Figure 1f was based on oxygen isotopes in an ice core from the Dunde ice cap iil the Qinghai-Tibetan Plateau (Thompson et al. 1989). Good. parallelism between curves (e) and (f) demonstrated the reliability of the reconsl!ll ction based on documentary records. The.two curves of Japan (Figure lb and le) show some resemblance to China, but the cold period in 19th century seems stronger than that in China. Figure la indicates that the cold periods over the Himalayas occurred in the 13th, 15th, and 18th· to 19th century. It is generally in accordance with that in China and in Japan. Finally, five cold periods over East Asia are schematically shown on the bottom· of Figure 1 with blank or shaded checks, which represent the cold and severe cold periods respectively. (The former or latter has an anomaly about -0.5° or -0.5°C to -l.0°C.) USSR: Klimanov (1992) has carried out a comprehensive review of studies on climate change in the former USSR. Various proxy data were applied in the analysis: glacier fluc tuations, lake level changes, documentary records, dendrochronological records, pollen data, lichenometry data and isotopic carbon data. The area examined started in the west from Europe: A lot of papers have been contributed to the study on the climate change in historical time over Eu.rope. Two curves are given in Figure 3; one was the classical temperature series of central England, which is based on both instrumental observations and documentary records (Lamb 1977), another is the schematic presentation. of fluctuations of the Grosser Aletsch Glacier (Holzhauser and Zumbuhl 1988). Pfi ster (1992) has compared the latter with the results of the study based on documentary records. Very good accordance was found. Four cold periods are outlined at the bottom of Figure 3.  North America: Scuderi (1987) has correlated the tree-ring indices of California with glacier fluctuations. Six cold periods were found. The first two in the 13th and 14th centuries were called "early Matthes;" the name of the author who invented the terminology of the Little Ice Age. The last four cold periods were concentrated between A.D. 1450 and 1850 ( Figure 4a). Another tree-ring series of California is shown in Figure 4c (LaMarche 1974). The narrowing of the tree-ring widths in the middle of the 13th century, from the middle to late 15th century and in the 17th and early 19th century were synchronous in both curves of the tree-ring series. However, a study on pollen data of north western Lower Michigan (Bernabo 1981) shows somewhat different characteristics to the tree-ring data (Figure 4b). It shows a prolonged cold period between A.D. 1450 and 1850. The discrepancy may depend on the poor time resolution of pollen data. This notwithstanding, most of the troughs in  Polar Region: Temperature data were very scarce in Polar Regions. Fortunately, data of sea ice and of oxygen isotopes in ice cores supplied the gaps. Ice cores from Devon Island (Patterson et al. 1977) and from Greenland (Robin 1983) provide a quantitative measure of the temperature changes ( Figure 5). Te mperature series reconstQ.!cted from tree-ring data (Briffa 1992) and from sea-ice records (Bryson 1974) present evidences for the occurrence of cold periods in the 12th to 13th and 17th to 19th centuries. The cold period in the second half of the 15th century was unstable and relatively weak in the two latter curves.
Southern Africa: Until very recent time, climate change during the last millennium in Southern Africa was rarely known. Tyson and Lindsay (1992) have reviewed studies on oxygen isotope analysis of cave speleothems and mollusk remains in shell middens, on the foraminifera of inshore marine deposits, palynological and micromammalian fossils, and on tree-ring data. However, one can hardly fi nd any characteristics which are common to the Northern Hemisphere. A prolonged cold period was found from A.O. 1300 to 1650, and since 1650 temperature generally increased. That is contrary to the trend found elsewhere on the globe. Another group of proxy temperature series showed that the 11th, 14th and 18th centuries were colder than the others. However, these centuries were relatively warmer in the Northern Hemisphere. Does this discrepancy depend on the regionality of the climate change, or on the difference of data sources? This has yet to be understood.
Other Regions of the Southern Hemisphere: Despite the signifi cant discrepancy found between temperature changes in Southern Africa and in the Northern Hemisphere, a series· of temperature reconstructions for other regions of the Southern Hemisphere showed good accordance with the Northern Hemisphere. For example, Thompson et al. (1986) demon strated close correlation between decadal temperature anomalies in the Northern Hemisphere (which were reconstructed by Groveman and Landsberg 1979), and the oxygen isotope record of an ice core from Quelccaya Ice Cap, Peru. Figure 6 shows three series of oxygen isotope and one of a temperature reconstruction for New Zealand based on stalagmite anomaly mea surements (Wilson et al. 1979). The cold periods are summarized and schematically shown at the bottom of Figure 6. It confirms the general contemporaneity of cold periods in the Northern Hemisphere and in some areas of the Southern Hemisphere.
New evi�ence on temperature changes during the last millennium are being found all the time. Therefore, the results outlined above should be considered as preliminary, and will be improved with more data sources. Antarctica, Low Dome ice core 6'180 (Morgan 1985); d. New Zealand, stalagmite (Wilson et al. 1979))

THE nLITTLE ICE AGEn, SOLAR ACTIVITY AND VOLCANISM
The cold periods identifi ed for East Asia, USSR, Europe, North America, Polar Region and the Southern Hemisphere are reproduced in Figure 7. However, Southern Africa showed quite different cold periods, so those were not included in Figure 7. Five cold periods were synthesized for the globe; the first half of 12th, second half of 13th, second half of 15th, 17th and 19th centuries. The first three cold periods were relatively weak, with temperature anomalies probably less than -0. 5°C compared to the average of the last millennium. The last two cold periods are characterized by severe climate, and temperature anomalies may have varied between -0.5°C to -l.0°C. Lamb (1977) had suggested that the "Little Ice Age" lasted from A.O. 1550 to 1850. However, the climate deterioration began much earlier, after the "Medieval Warm Period" (AD. 900-1300) so some authors suggested that the "Little Ice Age" was started by the end of Middle Ages at around A.O. 1250 (Porter 1986). Considering that the first three cold periods identified above were relatively weak, and the timing of them varies significantly from region to region, they could be regarded .. as the flrecursor of, or the transition to, the "Little Ice Age," rather than a stage of it. The planetary scale of the cold periods during the last millennium suggests that they were controlled by the same scale of forcing factor or factors, for example, solar activity and volcanism. Figure Sa and Figure Sb outline the solar activity during the last millennium. The fonner was based mainly on documentary records of sunspots observed by the naked eyes (Schove 1955). The latter came from quantitative carbon isotope data (Stuiver and Quay 19SO). Both of them show very similar secular changes, including the famous Wolf, Sporer, and Maunder Minima. Figure Sc shows volcanism detected from the acidity of an ice core from the Greenland Ice Cap (Porter 19S6). The cold periods identified above are shown on the bottom of Figure S for comparison. The second cold period in 13th, the third one in 15th and the fourth one in 17th century were more or less concomitant with the minima of solar activity. The two maxima of volcanism (scale of ordinate is reversed downward) seem also to contribute to the occurrence of the. third and fourth cold periods. However, the first cold period is difficult to interpret either from the record of solar activity or from the volcanism. A weak minimum of solar activity was found in the 11th rather than in the 12th century. The maximum of volcanism in the 12th century was too weak to be responsible for the cold period. The fifth cold period in the 19th century can hardly be attributed to the weak minimum of solar activity and maximum of volcanism in the early 19th century, for the last cold period was· quite severe, and was the coldest period for the last millerulium in some area. So, neither just solar activity nor volcanism, nor both of them, can account for all of the five cold periods during the last millennium.
It is suggested that factor or factors other than those above mentioned have taken part in regulation of the earth climate. First of all, internal interaction between the components of early climate system may be considered. The cryosphere, biosphere and the oceans may actively react to climatic change and significantly modulate the latter on different time scales. Secondly, other external factors may also be important in interpreting climatic changes during the last millennium. For example, recently Loutre et al. (1992) emphasized the role of the the earth's orbital parameters in modulation of the climate on the decadal to century time scale. Finally, the cold periods identified above seem also to need modification, and the average temperature should be found with greater accuracy with improved numbers of data sources. Then, the climatic change during the last millennium can be interpreted on a new basis. 1 4 C production rate (l::!,. Q M) (Stuiver and Quay 1980); c. Greenland Ice Sheet acidity (Porter 1986))