6
GeoJournal 5.2 117-122/1981 Akademische Verlagsgesellschaft- Wiesbaden 117 Reconstruction of the Summer Rainfall Regime for the last 500 Years in China Wang Shao-wu, Zhao Zong-ci and Chen Zhen-hua, Department of Geophysics, Peking University, Beijing, China Abstract: The anomalies of summer rainfall (June-August) for the last 500 years in China have been estimated according to the relationship between the instrumentally observed rain- fall and the degrees of floods and droughts for the last 25 years. It is shown that calculated anomalies of average rainfall for 10, 20 and 30 year periods have sufficient accuracies. The long-term climatic variations for the last 500 years are discussed based on the estimated anomalies of summer rainfall of 25 stations over the E part of China for the last S00 years. For example, there were prolonged drought in the N of China at about the end of Ming dynasty, but there were extensive floods in China for the early times of Qing dynasty. * The anomalies of rainfall for the last 500 years were averaged along latitude zone and their decade running mean were formed. It was found that the long-term variations are very clear. The 80 year cycle is predominating. Besides, an about 200 year cycle is also shown in the charts of 50-year mean rainfall anomalies. It is indicated that both above mentioned cycles relate closely to the solar activity. *The Ming dynasty lasted from 1368 to 1643 AD and was followed by the Qing dynasty (1644--1911). Introduction The temporal and spatial characteristics of floods and droughts for the last 500 years have been examined 1). Concerning the periodity, 36 years and 2-3 years cycles are found 2) 3), longer period cycles were less researched. Now we discuss mainly the long-term variations of floods and droughts. Investigating the long-term variations of floods and droughts it is often convenient to use the decade average, but this is difficult to do with the frequency of types or with the degrees of droughts and floods. Therefore, it is desirable to find a method to transform the degrees of floods and droughts into rainfall values, so that decade averages are easy to form. This paper is an attempt to reconstruct the summer rainfall values for the last 500 years. Then an analysis of the long-term variations in the reconstructed rainfall series is carried out. The Relationship between the Floods or Droughts and the Summer Rainfall (June-August) Firstly, it must be indicated that there is a close relation between the floods and droughts on the one hand and the summer rainfall on the other. Fig 1 shows the percentage of summer rainfall as related to the annual total for 100 stations in China. It can be seen from Fig 1 that in general the summer rainfall amounts to 40-50% of the annual rainfall, varying from 60-70% for many stations in N and N-E China, and to 30 % for some stations in S China. Formerly, the degree of floods and droughts was classified mainly according to the summer data, we also associate the degree of floods and droughts rather with the summer rainfall, than with annual rainfall. The summer rainfall data for the last 25 years were used as a base of analysis, because more or less complete and continual records of rainfall are available only since 1951. For comparison, 100 stations E of lO0~ for which the degree of floods and droughts are

Reconstruction of the summer rainfall regime for the last 500 years in China

Embed Size (px)

Citation preview

GeoJournal 5.2 117-122/1981 �9 Akademische Verlagsgesellschaft - Wiesbaden

117

R e c o n s t r u c t i o n o f t he S u m m e r R a i n f a l l Reg ime

f o r t he last 5 0 0 Years in C h i n a

Wang Shao-wu, Zhao Zong-ci and Chen Zhen-hua, Department o f Geophysics, Peking University, Beijing, China

Abstract: The anomalies of summer rainfall (June-August) for the last 500 years in China have been estimated according to the relationship between the instrumentally observed rain- fall and the degrees of floods and droughts for the last 25 years. It is shown that calculated anomalies of average rainfall for 10, 20 and 30 year periods have sufficient accuracies. The long-term climatic variations for the last 500 years are discussed based on the estimated anomalies of summer rainfall of 25 stations over the E part of China for the last S00 years. For example, there were prolonged drought in the N of China at about the end of Ming dynasty, but there were extensive floods in China for the early times of Qing dynasty. * The anomalies of rainfall for the last 500 years were averaged along latitude zone and their decade running mean were formed. It was found that the long-term variations are very clear. The 80 year cycle is predominating. Besides, an about 200 year cycle is also shown in the charts of 50-year mean rainfall anomalies. It is indicated that both above mentioned cycles relate closely to the solar activity.

*The Ming dynasty lasted from 1368 to 1643 AD and was followed by the Qing dynasty (1644--1911).

I n t r o d u c t i o n

The temporal and spatial characteristics o f f loods and droughts for the last 500 years have been examined 1). Concerning the periodi ty, 36 years and 2 - 3 years cycles are found 2) 3), longer period cycles were less researched. Now we discuss main ly the long-term variations of f loods and droughts.

Investigating the long-term variations of f loods and droughts it is often convenient to use the decade average, but this is d i f f icu l t to do wi th the frequency of types or wi th the degrees of droughts and floods. Therefore, i t is desirable to f ind a method to transform the degrees of f loods and droughts into rainfall values, so that decade averages are easy to form.

This paper is an attempt to reconstruct the summer rainfall values for the last 500 years. Then an analysis of the long-term variations in the reconstructed rainfall series is carried out.

T h e R e l a t i o n s h i p b e t w e e n the F l o o d s or D r o u g h t s

and t he S u m m e r R a i n f a l l ( J u n e - A u g u s t )

First ly, i t must be indicated that there is a close relation between the f loods and droughts on the one hand and the summer rainfall on the other. Fig 1 shows the percentage of summer rainfall as related to the annual total for 100 stations in China. I t can be seen from Fig 1 that in general the summer rainfall amounts to 4 0 - 5 0 % of the annual rainfall, varying from 6 0 - 7 0 % for many stations in N and N-E China, and to 30 % for some stations in S China. Formerly, the degree o f f loods and droughts was classified mainly according to the summer data, we also associate the degree of f loods and droughts rather wi th the summer rainfall, than wi th annual rainfall. The summer rainfall data for the last 25 years were used as a base of analysis, because more or less complete and continual records of rainfall are available only since 1951. For comparison, 100 stations E of lO0~ for which the degree o f f loods and droughts are

118 GeoJournal 5.2/1981

deviation from the degree 3 mean. Obviously, the deviation for degree I is generally more than 50 %. In the Yang-tze River basin it is more than I00 % and so in some regions in the North. For almost all stations the deviation for degree 2 is more than 25 %. Many stations deviate in degree 4 and 5 by about - 25 % and - 5 0 % respectively.

This analysis shows that the degrees 1-5 , which fairly well coincide with the degree of floods and droughts, roughly equal to 50 %, 25 %, 0 %, -25 %, -50 %. We may regard these degrees of mean deviation from average sum- mer rainfall as the anomalies corresponding with floods and droughts.

Fig I

available, were chosen. The missing records for individual stations were supplemented using the percentage charts ** of rainfall anomalies.

Secondly, the data of observed summer rainfall for 100 stations for the last 25 years were grouped into five degrees, so as the frequency distributions of the degrees agree roughly with frequency of floods and droughts, that is I /8, I /4, I /4, I /4 and I /8 for I to 5 degrees respectively. In analogy to the terminology floods and droughts 4), I = wet, 2 = subwet, 3 = normal, 4 = subdry, 5 = dry. The mean rainfall values for each degree were calculated, and so the specific degree means of rainfall obtained for I00 stations in China. Then the departure charts for each degree (I, 2, 4, 5) from degree 3 were estimated respective- ly. The degree 3 mean was regarded as the normal of sum- mer rainfall. It can be seen from these charts that for the degree I the deviation of rainfall in the Yang-tze River basin is generally larger than 400 mm, and in the North about 400-500 mm. For degree 2, in the Yang-tze River basin the deviation is generally I 0 0 - 1 5 0 m m and in the North about 100 ram. For degree 4, we observe the opposite, in the Yang-tze River basin and the North of China the deviation is about -100 ram, south of the Yang-tze River the negative departure is much larger. For degree 5, the negative departures S of the Yang-tze River go down below -200 ram, and on the coast as far as -300 mm.

To achieve a concept of degree mean summer rainfall, Fig 2 shows the percentage distribution of the rainfall

** kindly provided by the Long-range Forecasting Section, Central Meteorological Office.

Recons t ruc t i on o f the Rainfa l l Regime f o r the last

500 Years

When the correlation between deviation degrees and disaster degrees proves to be correct, we may obtain the anomalies of summer rainfall for the last 500 years in China.

In order to test the reliability of the reconstructed rainfall values for the last 500 years by the above- mentioned method, we have compared the reconstructed rainfall values with the instrumentally observed records for some stations. Unfortunately, stations with long records are very scarce. Fig 3 shows the decade running mean curves for thc, reconstructed and the observed (dotted) rainfall for Peking and Shanghai * * * . The reliability seems satisfactory.

In the following a comparison between the recon- structed and observed rainfall distributions was made for the period from 1901 to 1920 AD. In Fig4 both observed and reconstructed anomalous distributions are given. The anomalies are the deviation from the 500 year mean (I 471 - I 970) of the reconstructed rainfall.

From Fig 4 it is clear that the reconstructed rainfall distributions agree well with the observed rainfall, that in the Yang-tze River basin and S of it rainfall is above normal, and that the N and the far S of China are obviously affected by droughts.

These examples illustrate that it is possible to trans- form the degrees of floods and droughts to the summer rainfall anomalies with some skill.

The Var ia t ions o f Summer Rainfa l l in Histor ical T imes

The anomalies of summer rainfall for 25 stations in China for the last 500 years have been reconstructed by means of above mentioned method. It was found from the decade running mean curves that the range of variations of rainfall for the historical times is much bigger than that for the last 20 to 30 years. For example, in Peking (Fig3), the anomalies are -161 mm for both 1633-1642 and 1740- 1749, and they are 215 mm and 228 mm for 1780-1789

GeoJournal 5.2/1981 119

Fig 2

and 1885-1894; the differences between the highest and lowest values are 370-380mm. But since 1900, it is -123 mm for 1935-1944 and 123 mm for 1947--1956, their difference is only 240-250 ram. Since 1950, it is 122mm for 1950-1959 and - 6 0 m m for 1960-1969, the difference only about 180 ram. As the range of varia- tions for other stations has similar characteristics, it need not be discussed in length. We may conclude that the summer rainfall for the historical times experienced much bigger oscillations than in the last 20 to 30 years.

In order to illustrate further the obvious and sudden change of rainfall during the historical times, the anomalies of reconstructed rainfall for 25 stations in China at the end of the Ming dynasty (mid-17th century) are given in Fig 5. At the end of the Ming dynasty the North of China suffered from prolonged droughts, but at the beginning of the Qing dynasty, the flood situation was more characteristic.

***The summer rainfall data in Peking were taken from 5), the data for Shanghai are kindly provided by the Long-range Fore- casting Section, Central Meteorological Office.

120 GeoJoumal 5.2/1981

Fig 3

Tab I The frequency distribution of types of floods and droughts:

Fig 4

For the 30 year period 1615-1644, type 2 and 5 appeared in 21 years, making up 70%. 1645-1664 types 4 and I appeared in 16 out of 20 years, making up 80 %. Fig 5a shows that the droughts are predominating in Henan,

la Floods over almost the whole country, mainly in the Yang-tze River basin.

Ib Floods in the Yang-tze River basin, droughts N and S of it. 2 Floods in the S and droughts in the N. 3 Droughts in the Yang-tze River basin, floods N and S of it. 4 Floods in the N and droughts in the S. 5 Droughts over almost the whole country.

Shansi and Guanzheng provinces. For example, the average anomaly of rainfall at Zheng zhou for 30 years peaked at - 5 0 mm. Opposite to this case, it is shown by Fig 5b that 1645-1664 was the time which experienced the most

Fig 5

GeoJournal 5.2/1981 121

Fig 6

exceptional floods in the N of China. For example, the anomaly of rainfall at Taiyuan for 20 years peaked at 110 mm. Comparing with the anomalies of summer rain- fall with other periods, the range of variation for this period was more bigger.

Investigating systematically the long-term variation of summer rainfall during the historical times, the mean anomalies for each 50 years period from 1470 to 1979 AD are formed, besides for 1470-1499 and 1951-1979 31 years and 20 years were taken for lack of data. An about 200 year cycle was found from the frequency variation of drought and flood types. Fig 6 proves this cycle also by means of the 50 year mean rainfall anomalies. For example, for 1551-1600 (Fig 6c), type lb was dominating almost whole China, especially the Yang-tze River basin, however, suffered from floods. For 1601-1650 (Fig 6d), in the N of China there were significant droughts, but in the S

floods occurred. For 1651-1700 (Fig 6e), the rainfall regime was opposite: Floods in the N and droughts in the S of China. 200 years later, 1751-1800, these anomalies reappeared indicating a 200 year cycle. 1751-1800 (Fig 6g), the E of China suffered from floods, and 1801 1850 (Fig 6h), almost whole China experienced floods, especially the Yang-tze River basin (type 1 a). 1851-1900 (Fig 6i), S of the Yang-tze River basin and N of the Yellow River there were floods and droughts in the region between the Yang- tze River and the Yellow River with very high frequency of type 3. In the 1901-1950 period, southern China suffered from prolonged floods. In the 1951-1979 period, the rain- fall regime was reversed, S of the Yang-tze River basin there were droughts and floods in the N. Unfortunately, the data were not sufficient to prove more completely the cycle of about 200 years, it only reappeared two times. If it is true, China would be facing a prolonged flood period, say, for 50 years, in the future.

122 GeoJoumal 5.2/1981

the 80 year cycle of the solar activity. Fig 7b shows maxima in the sunspot 11 year cycle, the extremes are numbered (2)-(7). It may be understood from Fig 7 that the 80 year cycle of solar activity is usually followed by a prolonged flood period, starting from the N of China and lasting for some decades. The six flood periods for the last 500 years related closely to the peak of the sunspot maxima. The 7th flood period may have started already.

Fig 7

By the way, it is interesting to note that at the middle of t h e 16th and 18th centuries the solar activity was intensified significantly (Fig 7). In the middle of the 20th century, solar activity had reached its peak value for the last 2 -3 centuries. For this period China experienced floods, therefore, the about 200 year cycle of rainfall may be linked with the solar activity.

The 80 Year Cycle o f Rainfa l l in Re la t ion to the

Solar A c t i v i t y

Analysis of the variations of the rainfall anomalies for the last 500 years indicates that in addition to the 200 year cycle a clear 80 year cycle exists. The anomalies of the reconstructed rainfall for 25 stations were averaged for 2.5 degree latitude zones and then the decade running mean of the latitude average anomalies calculated (Fig 7a). The arrows in Fig 7a stand for the period of the floods. In the last 500 years extensive flooding happened six times (in Fig 7 noted by numbers (1)-(6)). It is found from Fig 7a that the time interval between the arrows is about 80 years. It is worth to note that in general the flood period began in the N of China and shifted to the S with time. Even more interesting is that the clear 80 year cycle related closely to

Discussion

It has been indicated above, that there was some skill in using the observed rainfall data to estimate and reconstruct the anomalies of summer rainfall for the historical times from flood and drought data. There are some weaknesses of this method. Firstly, the frequency distribution of the degrees of floods and droughts for the last 25 years was slightly diverted from I /8, I /4, I /4, I /4, I /8, with which the degrees of rainfall were estimated. But it is inconvenient to use the real frequency of floods and droughts for the last 25 years, therefore we have still to use the normal distribution of frequencies. Naturally, it may have brought some errors into the calculation. Secondly, the same degree of floods or droughts for one station must not be implied by the same rainfall anomalies for different years. There- fore, we feel that it was more suitable to reconstruct the anomalies for a period, say for the decades, and not for individual years. Thirdly, when the degrees of floods and droughts were termined, we used mainly, but not only the summer data. Relating the annual degrees of floods and droughts only to the summer rainfall may also cause some errors. But if the annual or half year rainfall anomalies were reconstructed, the circumstances would be more complex.

However, it can be said that it is worth to reconstruct the summer rainfall anomalies in historical times and to provide a quantitative concept of them to some extent.

References

l) Wang Shao-wu and Zhao Zong-ci: The investigation of the droughts and floods in the last 500 years in China. (Presented in the Conference of Climate and History, held in East Anglia University, July 1979).

2) Wang Shao-wu and Zhao Zong-ci: The climatic change in China and the atmospheric circulation for the last 100 years. (To be published in the "Proceedings of the Climatic Change Con- ference" held in 1978, China). (In Chinese)

3) Wang Shao-wu and Zhao Zong-ci: The 36-yr. wetness oscilla- tion in China and its mechanism. Acta Meteorologica Sinica 37, I (]979) (In Chinese)

4) The Chinese flood and drought charts from 1470 to 1978 AD. Chinese Meteorological Office, Nangking University and Peking University etc. (to be published). (In Chinese)

5) Chinese Meteorological Office: The rainfall in Peking for 250 years (1724-1973). (1975) (In Chinese)