LETTERS NATURE Vol 4454 January 2007 The sediment S-ratio and TOC content,which indicate changes in correlation between summer and winter monsoons recognized here bottom water oxygen,are functionally independent from Ti content, is also evident for the Indian monsoon system,with the Younger an indicator of dust input.Thus,their inter-consistency,reflecting Dryas representing a time of weaker summer and stronger winter high lake mixing whenever Ti-rich dust input is high,makes a com- monsoons'.The high-resolution records from East Asia indicate pelling case for interpreting these records as a robust measure of that this summer/winter monsoon anti-correlation also applies on winter monsoon strength.We deduce that the pre-Bolling-Allerod, shorter timescales (Fig.2). the Younger Dryas,and the later Holocene were all characterized by Migration of the annual mean position of the ITCZ provides a strong winter monsoons in East Asia.During those intervals,low single coherent explanation for the observed trends in both winter TOC content and higher magnetic susceptibility and S-ratios indicate and summer monsoons over the past 16 kyr,as well as for the strong enhanced wind mixing of the lake's waters,resulting in a well oxyge- anti-correlation between them.When the ITCZ is displaced north- nated lake floor,good preservation of magnetic minerals,and ward,the summer monsoon should strengthen,while the winter increased degradation of organic matter.In parallel,the higher Ti monsoon will weaken.A northward shift in the ITCZ would be content during these times suggests intensified winter winds in cent- expected during times of Northern Hemisphere warming2,such as ral China leading to an increased input of dense Ti-bearing dust.In the Bolling-Allerod and the early Holocene,times when,indeed,the contrast,during periods of warmer climate such as the Bolling- East Asian summer monsoon was strong'2 and the winter monsoon Allerod and the early Holocene,Ti concentrations are lower,while was weak.This mode of explanation for climate change has recently lower magnetic susceptibility and S-ratios are coupled with higher proven fruitful for interpreting palaeoclimate records from the trop- TOC content,implying reduced winter monsoon winds and stable ical Americass(Fig.2;see also Supplementary Information).The stratification of the lake. data reported here,in concert with existing data'- 3202127 suggest that The Lake Huguang Maar records of winter monsoon strength these ITCZ migrations extended across the Pacific show a remarkable relationship with the radiometrically dated The role of climate and environmental change in the success or 88O records from stalagmites at Dongge cave2 and Hulu cave', failure of societies is a matter of intense debate4322 It would be East Asia(Fig.2).Strong winter monsoon phases reconstructed from simplistic to imagine that all episodes of societal change are driven by Huguang Maar correlate with higher O at Hulu and Dongge caves, climatic events,especially in an advanced and complex society such which indicates lower summer precipitation rates and thus a weaker as dynastic China.Nevertheless,we note that,on the basis of our new summer monsoon'.Thus,our data,when compared with the spe- Huguang Maar data,major changes in Chinese dynasties occurred leothem'2 and South China Sea2 records,argue for an inverse cor when the winter monsoon was strong(Fig.3).The anti-correlation relation between the strengths of the winter and summer monsoons between winter and summer monsoon strength indicated by com- (see also Supplementary Information).Moreover,the similarity of parison of the Huguang Maar data with the cave records would the records implies that monsoon changes during the latest glacial, suggest that dynastic transitions tended to occur when the summer Bolling-Allerod,Younger Dryas and Holocene were synchronous monsoon was weak and rainfall was reduced.Dynastic changes in and common to large regions of coastal Southeast Asia.The inverse China often involved popular uprisings during phases of crop failure Figure 3The Lake Huguang Maar 910 AD810 palaeoclimate records during the past 4,500 yr in the context of major events in the cultural history of China.a,Major changes in Chinese dynasties occurred during dry phases,as indicated by the titanium and magnetic susceptibility records from Lake Huguang Maar and applying the observed anti-correlation between the winter and summer monsoons, na maa while the described 'golden ages' n centun tended to occur during wet phases. b,Comparison of titanium records 000 1.050 1,100 1,150 1,200 1.250 1,300 from Lake Huguang Maar and the Age (cal.yr Bp) Cariaco basin.The shared features of the two climate records as well as the similar timing of Chinese*and ang Fra of Mayan societal changes suggest a disunity role for coherent climate changes 1.000 (that is,ITCZ migration)across the Pacific in the events of widely 2.000 dispersed civilizations. 3.000 100 000 120 5.000 140 160 180 500 1.000 1,500 2.000 2.500 3.000 3.500 4.000 Age (cal.yr Bp) 16 82007 Nature Publishing GroupThe sediment S-ratio and TOC content, which indicate changes in bottom water oxygen, are functionally independent from Ti content, an indicator of dust input. Thus, their inter-consistency, reflecting high lake mixing whenever Ti-rich dust input is high, makes a com￾pelling case for interpreting these records as a robust measure of winter monsoon strength. We deduce that the pre-Bølling–Allerød, the Younger Dryas, and the later Holocene were all characterized by strong winter monsoons in East Asia. During those intervals, low TOC content and higher magnetic susceptibility and S-ratios indicate enhanced wind mixing of the lake’s waters, resulting in a well oxyge￾nated lake floor, good preservation of magnetic minerals, and increased degradation of organic matter. In parallel, the higher Ti content during these times suggests intensified winter winds in cent￾ral China leading to an increased input of dense Ti-bearing dust. In contrast, during periods of warmer climate such as the Bølling– Allerød and the early Holocene, Ti concentrations are lower, while lower magnetic susceptibility and S-ratios are coupled with higher TOC content, implying reduced winter monsoon winds and stable stratification of the lake. The Lake Huguang Maar records of winter monsoon strength show a remarkable relationship with the radiometrically dated d18O records from stalagmites at Dongge cave2,3 and Hulu cave1 , East Asia (Fig. 2). Strong winter monsoon phases reconstructed from Huguang Maar correlate with higher d18O at Hulu and Dongge caves, which indicates lower summer precipitation rates and thus a weaker summer monsoon1 . Thus, our data, when compared with the spe￾leothem1,2 and South China Sea24 records, argue for an inverse cor￾relation between the strengths of the winter and summer monsoons (see also Supplementary Information). Moreover, the similarity of the records implies that monsoon changes during the latest glacial, Bølling–Allerød, Younger Dryas and Holocene were synchronous and common to large regions of coastal Southeast Asia. The inverse correlation between summer and winter monsoons recognized here is also evident for the Indian monsoon system, with the Younger Dryas representing a time of weaker summer and stronger winter monsoons10. The high-resolution records from East Asia indicate that this summer/winter monsoon anti-correlation also applies on shorter timescales (Fig. 2). Migration of the annual mean position of the ITCZ provides a single coherent explanation for the observed trends in both winter and summer monsoons over the past 16 kyr, as well as for the strong anti-correlation between them. When the ITCZ is displaced north￾ward, the summer monsoon should strengthen, while the winter monsoon will weaken. A northward shift in the ITCZ would be expected during times of Northern Hemisphere warming25, such as the Bølling–Allerød and the early Holocene, times when, indeed, the East Asian summer monsoon was strong1,26 and the winter monsoon was weak. This mode of explanation for climate change has recently proven fruitful for interpreting palaeoclimate records from the trop￾ical Americas17,18 (Fig. 2; see also Supplementary Information). The data reported here, in concert with existing data1–3,20,21,27, suggest that these ITCZ migrations extended across the Pacific. The role of climate and environmental change in the success or failure of societies is a matter of intense debate4,5,28,29. It would be simplistic to imagine that all episodes of societal change are driven by climatic events, especially in an advanced and complex society such as dynastic China. Nevertheless, we note that, on the basis of our new Huguang Maar data, major changes in Chinese dynasties30 occurred when the winter monsoon was strong (Fig. 3). The anti-correlation between winter and summer monsoon strength indicated by com￾parison of the Huguang Maar data with the cave records would suggest that dynastic transitions tended to occur when the summer monsoon was weak and rainfall was reduced. Dynastic changes in China often involved popular uprisings during phases of crop failure Age (cal. yr BP) Age (cal. yr BP) 5,000 4,000 1,000 1,500 2,000 2,500 3,000 3,500 4,000 1,000 1,050 1,100 1,150 1,200 1,250 1,300 500 AD1644 AD1368 AD1279 AD960 AD1127 AD907 AD618 AD9-24 AD1271-1368 AD618 202BC 256BC 1752BC 1752BC 2183BC 0 3,000 2,000 1,000 0 Wet Wet Dry Dry Ti (counts s–1) Ti (counts s–1) Ti (counts s–1) Era of disunity W Shang Xia Zhou Ming Suang Tang Han E Zhou South NorthYuan Qin a b Eighth century: decline of AD907: collapse of Tang dynasty Peak of Tang dynasty Tang dynasty Major phases of collapse of ‘Classic Maya civilization’ Peak of ‘Classic Maya period’ AD~910 AD~860 AD~810 Magnetic susceptibility (10–6 SI) Lake Huguang Maar Cariaco basin Figure 3 | The Lake Huguang Maar palaeoclimate records during the past 4,500 yr in the context of major events in the cultural history of China. a, Major changes in Chinese dynasties30 occurred during dry phases, as indicated by the titanium and magnetic susceptibility records from Lake Huguang Maar and applying the observed anti-correlation between the winter and summer monsoons, while the described ‘golden ages’30 tended to occur during wet phases. b, Comparison of titanium records from Lake Huguang Maar and the Cariaco basin. The shared features of the two climate records as well as the similar timing of Chinese30 and Mayan4,5 societal changes suggest a role for coherent climate changes (that is, ITCZ migration) across the Pacific in the events of widely dispersed civilizations. LETTERS NATURE| Vol 445| 4 January 2007 76 ©2007 NaturePublishingGroup