ITCZ- and ENSO-Controlled Tropical Rainfall Variations

ITCZ- and ENSO-controlled tropical rainfall variations over therecent two millennia

Liangcheng Tan^1,2,3*, Chuan-Chou Shen^4*,Ludvig Loewemark⁴,R. Lawrence Edwards⁵, Sakonvon Chawchai^6,7, Yanjun Cai^1,2, Xianfeng Wang⁸, Sebastian F.M. Breitenbach⁹, Hai Cheng^2,5, Zhisheng An¹,Akkaneewut Chabangborn^6,7, Yongli Gao¹⁰, Ola Kwiecien¹¹, Chung-CheWu⁴, Barbara Wohlfarth⁶

1. State Key Laboratory of Loess and Quaternary Geology, Institute of Earth Environment, Chinese Academy of Sciences, Xi’an 710061, China

2. Institute of Global Environmental Change, Xi’an Jiaotong University, Xi’an 710054, China

3. Joint Center for Global Change Studies (JCGCS), Beijing 100875, China

4. High-Precision Mass Spectrometry and Environment Change Laboratory (HISPEC), Department of Geosciences, National Taiwan University, Taipei 10617, Taiwan ROC

5. Department of Earth Sciences, University of Minnesota, Minneapolis 55455, USA

6. Department of Geological Sciences and Bolin Centre for Climate Research, Stockholm University, Stockholm 10691, Sweden

7. Department of Geology, Faculty of Science, Chulalongkorn University, Bangkok 10330, Thailand

8. Earth Observatory of Singapore, Nanyang Technological University, Singapore 639798

9. Geological Institute, Department of Earth Sciences, ETH Zürich, Zürich 8092, Switzerland

10. Department of Geological Sciences, University of Texas at San Antonio, San Antonio 78249, USA

11. Institute for Geology, Mineralogy & Geophysics, Bochum D-44801, Germany

*To whom correspondence should be addressed to: C.-C.S ([email protected]) or L.T ([email protected])

Tropical rainfall plays an important role in global water cycle and energy balance¹. Its variability is closely associated with the meridional shift of the Intertropical convergence zone (ITCZ) and the zonal move of the Walker circulation, linked to the El Niño/Southern Oscillation (ENSO)^2,3. However, the detailed characteristics and mechanisms on multidecadal-to-centennial scales are still unclear, largely because of the controversial views of ITCZ^4-8 and ENSO^9-11 dynamics during historical times.Here we reconstruct a new index of ITCZ dynamics for the past two millennia using absolutely-dated stalagmite-inferred rainfall records from southern Thailand and eastern Indonesia⁴ in tropical deep convective zone. Results show that the ITCZ governed the millennial-scale tropical rainfall. On multidecadal-to-centennial scales, low-latitude rainfalls are dominated by ITCZ shift in Atlantic and western Indian Ocean, instead, strongly influenced by ENSO in the eastern Indian Ocean and Pacific. The drying 20th-century trend in the northern tropics is similar to that in the past warm periods, suggesting that the possible anthropogenic forcing may remain undistinguishable.

Tropics is the most vigorous atmospheric convection realm in the world, where a large part of moisture to middle and high latitudes originates^1,3. The deep convection in tropics can also transfer energy by surface evaporation and water vapor condensation, redistributing the global energy¹. Furthermore, tropical rainfall is crucial for supporting the vastest biodiversity and the livelihoods of more than 40% of the world’s population¹². There was a declining trend in northern tropical rainfall in the 20th century¹³. Whether it was caused by natural (volcanic eruption¹⁴, internal oceanic and atmospheric oscillations¹⁵) or anthropogenic (sulfate aerosol loading¹⁶ and green house gas emission¹⁷) forcing is still in debate. Lacking comprehension of these changes limits our understanding of its future trend ^16,18,19.

Both seasonal and long-term variations of tropical rainfall are closely linked to the position of the Intertropical convergence zone (ITCZ)³. However, the ITCZ dynamics during historical times is far from resolved. A southward mean positioning of the ITCZ has, for example, been suggested for the Little Ice Age (LIA: 1400-1850 AD)^4-6, followed by a northward shift during the 20th century⁵. Other studies^7,8 argued for a contracted ITCZ during the LIA and an expanded ITCZ during the 20th century in East Asia-Australia sector. Tropical rainfall is also influenced by El Niño/Southern Oscillation (ENSO)^2,9. However, the tropical Pacific was dominated by either La Niña- or El Niño-like conditions during the LIA and historical warm periods remains controversial^9-11. The interplay of ITCZ meridional shift and ENSO variability on tropical rainfall was still poorly addressed. Such understanding gaps are mainly due to the lack of millennial-long, accurately-dated, and high-resolved rainfall records from the deep convective zone. Here, we compare a 2700-yrs rainfall record, based on three replicated stalagmites from Klang cave in the southern Thailand peninsular, with different tropical sequences to clarify the aforementioned issues.

Klang Cave (8°1′ N, 98°48′ E), located in the core region of the ITCZ, is in the Ao Luk district, 40 km northwest of Krabi town, the western part of southern Thailand (Fig. 1). The annual mean precipitation is 2760 mm (1901-2011 AD). During the rainy season, May-October, the southwest Indian summer monsoon (ISM) delivers more than 75% of annual rainfall from Indian Ocean. The dry season, November-April, experiences the northeast winter monsoon, delivering ≤25% of annual rainfall from the South China Sea and Western Pacific (supplementary materials, Fig. S1). Evapotranspiration surpasses winter precipitation and reduces the rainfall water into the regional karst system (Fig. S2). Consequently, recharge of the aquifer of Klang cave is mainly derived from ISM rainfall.

The main passage of the cave, 1400 m long, developed downward under the ground. The narrow tunnel is only 0.5-2 m wide. This close cave maintains relative humidity of 95-100% for the tunnels and chambers >100 m from the entrance. A stable cave air temperature of 23.5±0.5 °C (1σ) was recorded from April 2011 to August 2012 (Fig. S3). Thus, cave air temperature is always lower than surface temperature, which suppresses density-driven ventilation²⁰. Three columnar-shaped aragonite stalagmites, TK16, TK131, and TK133, with lengths of 29.8, 51.7 and 36.3 cm, respectively, were collected 700-900 m from the entrance (Fig. S4). Instrumental analyses (Table S1) show that the stalagmites are characterized with high uranium (up to 40 ppm) and low thorium contents (as low as 10 ppb), which can offer precise U-Th dates. Stalagmite TK16 grew from 2 to 998 AD, with dating errors of 2-4 years. TK131 deposited from 1733 to 2005 AD, with most dating uncertainties as small as 0.4-2 years. U-Th ages, with most dating errors <10 years, show that TK133 deposited from 706 BC to 1867 AD (Table S1, Fig. S5).

δ¹⁸O records of the three stalagmites mutually match over contemporaneous deposition periods of 1733-1867 AD for TK131 and TK133, and 2-998 AD for TK 16 and TK133, despite systematical offsets of 0.5-2‰ between their absolute δ¹⁸O values (Fig. S6, Table S2). The offsets can be caused by prior carbonate precipitation through different water flow paths and different degassing rates above stalagmites²¹. The observed degree of replication suggests minimal kinetic isotope fractionation effects and that the stalagmite δ¹⁸O can be attributed to changes in cave dripwater δ¹⁸O. A spliced δ¹⁸O record, covering the recent 2700 years (706 BC-2005 AD) was built with TK131 data and adjusted TK133 data (Fig. 2 and Fig. S6). The composite TK record consists of 1,201 δ¹⁸O measurements, with an average resolution of 2.9 years from 706 BC to 920 AD and 6.8 years during 920-1733 AD. An extremely high resolution of 0.5 year is for data after 1733 AD (Table S2), making it an ideal candidate for comparison with the meteorological records.

δ¹⁸O signal in rain, cave dripwater, and thus in stalagmite carbonate, can serve as a proxy for pan-regional moisture source dynamics and/or regional-local rainfall amount^15,22,23. Previous observation and model results showed an inverse relationship between δ¹⁸O values and rainfall amount in the tropical convective region^22,24,25. Indeed, significant negative correlations between monthly rainfall amount and precipitation δ¹⁸O (δ¹⁸O_precip) (r = -0.64, p < 0.01), as well as annual rainfall amount and weighted mean ¹⁸O_precip(r = -0.58, p< 0.01) (Fig. S7), are observed in Bangkok, Thailand. A significant negative correlation (r = -0.52, p<0.01, 5-point running average) between the composite TK δ¹⁸O time series and local rainfall record between 1901 and 2005 is also expressed in Figure 2. These observations strongly support the interpretation of the composite TK δ¹⁸O as a paleo-rainfall proxy.

The most notable feature of TK δ¹⁸O record is the long-term increasing trend, which signifies decreasing regional rainfall over the past 2,700 years. This millennial-scale rainfall decline is recognizable in paleoclimate records from numerous locations in the northern tropics, such as Southeast Asia²⁶, Mesoamerica^27,28, and the Caribbean⁴. An opposite increasing rainfall trend is documented in the southern tropics, including east Africa²⁹, the Western Pacific Warm Pool³⁰, the eastern Pacific³¹ and South America³² (Fig. S8). This tropical inter-hemispheric precipitation see-saw pattern on millennial timescale is similar to the previous observation for the last glacial³³.

Tropical rainfall is influenced by the zonal ENSO variability and the meridional ITCZ shift^2-4. Modern observations clearly show that rainfalls in the northern and southern parts of the western Pacific (abbreviated to NWP and SWP, respectively) synchronously decrease during El Niño events (Fig. 1). The influence of ITCZ is, therefore, possible to be revealed by the regional rainfall gradient. The instrumental annual rainfall gradients between two regions of 5-10° N, 98-130° E and 5-10° S, 98-130° E during 1983-2012 indicate this feasibility (Fig. S9). This gradient time series shows a significant positive correlation (r = 0.63, p<0.05) with the rainfall record over the Sahel, which is controlled by the north excursion of ITCZ in boreal summer³ (Fig. S9). More regional rainfall gradient corresponds to wetter condition over Sahel, and vice versa. This concurrency supports our argument of that rainfall gradient between NWP and SWP is a reliable indicator for the shift of ITCZ. TK record from southern Thailand and a 2000-yrs composite LL record from Liang Luar cave in eastern Indonesia² (Fig. 1) are negatively correlated with local rainfall amount. As a result, a record of ITCZ shift index (SI) sequence for the last 2,000 years can be established by subtracting the standardized LL record from the standardized TK record (Fig. S10, see Methods). Negative (Positive) ITCZ SI values represent more (less) rainfall in the NWP compared to the SWP, and a northward (southward) shift of the Pacific ITCZ.

The ITCZ SI sequence shows a general southward moving trend over the past two millennia (Figs. 3 and S10E), attributable to the declining boreal summer insolation³. It reveals a relatively northern mean ITCZ position in the first two centuries, followed by a southward drift from the 3rd century to the end of the 5th century. A sharp northward return is observed within the next 40 years, replaced by another multi-centennial southward migration until the early 9th century. The ITCZ moved gradually northward between the 9th and early 12th centuries, followed by an abrupt 30 years northward movement in the mid 12th century. It then retracted south for the next 600 years and reached the most southern position at 1,810 AD. With the beginning of the industrialization in the mid 19th century, the ITCZ generally shifted northward (Fig. 3).

Data analysis by an ensemble empirical mode decomposition (EEMD)³⁴ method (Fig. S11) shows that during the last 2,000 years, the ITCZ was dominated by millennial-scale change, with a variance of 57%. Multicentennial-scale variations (cycles of ~500 and 200-250 yrs) contribute additional 27%. ITCZ migrations on millennial- to multicentennial-scale are coherent with the temperature contrast between northern and southern extratropics (Fig. 3E). A relatively warming in the northern extratropics reduces the meridional temperature gradient in the northern extratropics and resulting in northward ITCZ migration³.

Observations and modeling studies revealed that freshwater input and/or sea ice expansion in the North Atlantic can affect the formation of North Atlantic Deep Water and the Atlantic Meridional Overturning Circulation, decrease temperature in the high northern latitudes, and induce a southward ITCZ shift^35,36. The close correspondence between our ITCZ SI record and a North Atlantic record of ice-rafted debris (IRD)³⁷ (Fig. 3D) suggests a strong coupling between North Atlantic climate and ITCZ over the past 2,000 years. On multidecadal scales, the periodicity at 63 yrs in the ITCZ SI record is also similar with the 55-80 yrs cycle in the North Atlantic sea surface temperature (SST). We find a close resemblance between ITCZ SI and a marine sediment Ti record (r = -0.53, p <0.01) from Cariaco Basin⁴. Relatively negative ITCZ SI values (northward ITCZ) correspond to rainfall-deduced high Ti content (Fig. 3A). The significant correlation corroborates that the ITCZ influences the tropical Atlantic rainfall on centennial to multidecadal scales.

The low-level cross-equatorial flow in the western Indian Ocean caused by the seasonal ITCZ shift is an essential component of the Indian summer monsoon (ISM). A significant positive correlation (r = 0.50, p <0.01) between our ITCZ SI record and a 2000-yrs stalagmite δ¹⁸O time series from northern India¹⁵ supports these links (Fig. 3B). A northward ITCZ shift corresponds with strong ISM rainfall in northern India, while a southward shift leads to a weakened ISM. In contrast, we observed dry (wet) conditions in east Africa in the southern tropics during periods with a northward (southward) ITCZ shift (Fig. 3C). These correlations suggest that on centennial to multidecadal scales, ITCZ shifts control spatial rainfall distribution in the western Indian Ocean realm.

Comparison of detrended TK stalagmite-inferred rainfall with published sequences, including a lacustrine sediment record from tropical south China³⁸, a multi-proxy synthesized stalagmite record from eastern Indonesia in the southern tropics², stalagmite records from Panama²⁸ and Belize²⁷ in Mesoameric, and a varved lacustrine record from the central Peruvian Andes³², in the tropical eastern Indian Ocean and Pacific Ocean are given in Figure 4a-f. All records are featured with dry periods during the 20th century (Current Warm Period, CWP) and 950-1150 AD in the Medieval Warm Period (MWP, 950-1300 AD)³⁹. In contrast, relative humid conditions are expressed during the LIA (1400-1850 AD) and the Dark Age Cold Period (DACP, 400-800 AD)³⁹. This warm/dry-cool/wet climate pattern is also observed in the southeast Tibetan Plateau⁴⁰. Moreover, our TK record shows decreased rainfall in another warm period during 150 BC to 80 AD, corresponding to the “Roman Warm Period”(RWP)³⁹. The synchronous rainfall variations over tropical eastern Indian Ocean and Pacific indicate that the ITCZ migration was not the primary control of rainfall in this region on centennial scale over the past two millennia^2,11. Otherwise, an antiphase relation would be observed between records in the northern and southern tropics.

The observed dry conditions during the RMP and MWP correspond well with intensified El Niño activities recorded in lake sediments from tropical eastern Pacific region^31,41 (Fig. 4j and 4k). An annually resolved ice core inferred Niño 4 SST from Peru⁴² (Fig. 4l) expresses enhanced El Niño activities during MWP and CWP, concurrent with the droughts documented in our TK record. TK record also coincides with the Southern Oscillation Index (SOI)¹¹ (Fig. 4i), with increased δ¹⁸O values (decreased rainfall) corresponding to decreased index (more El Niño-dominated conditions), and vice versa (Fig. 4i). We observe two multidecadal droughts during 255-335 AD and 1220-1300 AD in TK record concurred with enhanced El Niño activities in the tropical eastern Pacific record^31,41 (Fig. 4k, 4j) and reduced SOI¹¹ (Fig. 4i). The spatial distribution of droughts in the tropics was similar to those during El Niño events today (Fig. 1). All evidence suggests that ENSO variability, rather than ITCZ shifts, controlled the rainfall variability in tropical eastern Indian Ocean and Pacific regions on centennial to multidecadal scales. The impact of ENSO on tropical rainfall can be through the east-west displacement of the ascending and descending branches of Walker circulation⁴³. During El Niño (La Niña) conditions, the ascending branch of Walker circulation move eastward (westward), and increased descent (ascend) broadly distributed over northeastern India, southwest China, and southeast Asia, suppressing (enhancing) monsoon rainfall^43,44.

Modern observations and model results indicate that the Indo-Pacific warming caused a weakening of the Walker circulation since the mid-19th century⁴⁵. Model simulations also supports La Niña-dominated conditions and enhanced Walker circulation during the LIA⁴⁶ and DACP⁴⁷, and El Niño-dominated conditions during the CWP⁴⁶, MWP and RWP⁴⁷ (Fig. 4m). It is worthy to note that the drought peaks during the RWP and the period of 1220-1300 AD coincide with the two largest volcanic eruptions in 53 BC and 1257 AD^48-50 during the past 2700 years (Fig. 4n) . Decreased rainfall in TK record is also observed during another two large eruptions in 1600⁵¹ and 1815 AD⁵², suggesting a quick response of tropical rainfall to volcanic eruption⁵³. However, the annual-scale volcanic aerosol’s effect⁵³ should not be the primary driving force of decadal and longer timescales droughts in the tropics. As shown in Figure 4n, the intensities of the volcanic eruptions in the 20th century and MWP were smaller than those in the 13th century, but the droughts during these two periods were much more severe.

Our record, with accurate dates and high resolution, suggests that the drying trend in the northern tropics in the 20th century is similar to those during the historical warm periods of RWP and MWP, and the anthropogenically forced rainfall changes may remain difficult to detect.

Methods summary

U-Th dating: Stalagmites, TK16, TK131 and TK133, were cut into halves along their growth axes and polished. Powdered subsamples, 50-100 mg each, of 61 layers were drilled along the growth axis on the polished surface for U-Th dating. We followed the chemical procedure described in ref.S1 and ref. S2 to separate uranium and thorium. U-Th isotopic compositions and ²³⁰Th dates were determined by a multi-collector inductively coupled plasma mass spectrometer (MC-ICPMS), Thermo Fisher Neptune, at the High-Precision Mass Spectrometry and Environment Change Laboratory (HISPEC), National Taiwan University (ref. S3) and Isotope Laboratory, Xi’an Jiaotong University (ref. S4). Age models were established by using 5000 Monte-Carlo simulations and a polynomial interpolation procedure in the COPRA routine (ref. S5).

Stable isotope analysis: Stalagmite subsamples for oxygen stable isotope analyses were drilled at intervals of 2 mm, 1 mm and 0.5 mm, for TK16, TK131 and TK133, respectively. A total of 1388 subsamples were analyzed on an IsoPrime100 gas source stable isotope ratio mass spectrometer equipped with a MultiPrep system in the Institute of Earth Environment, Chinese Academy of Sciences. Reported ¹⁸O values are calculated with respect to the Vienna Pee Dee Belemnite (VPDB). An international standard NBS 19 and a laboratory standard HN were analyzed every 10-15 samples to monitor instrumentation and reproducibility. The replicates showed that the external error for ¹⁸O was better than ±0.06‰ (1σ).

Construction of the ITCZ shift index: First, both TK ¹⁸O and LL composite records were normalized for the contemporary period, 7-1997 AD, by using the equation as follows:

,

where is the original value, the maximum and minimum value of the time series, respectively, and the normalized result. Secondary, due to different age uncertainties and proxy resolutions, we applied 20-years moving average to the annually interpolated results of the normalized TK (TK_n) and LL (LL_n) records. Finally, the ITCZ SI was constructed by subtracting the smoothed LL_n record from the smoothed TK_n record (Fig. S13). Relative negative (positive) ITCZ SI values represent more (less) rainfall in the northern tropics relative to the southern tropics, indicating northward (southward) shift of the Pacific ITCZ.

References:

1Waliser, D. E. & Gautier, C. A satellite-derived climatology of the ITCZ. J. Clim. 6, 2162-2174 (1993).

2Griffiths, M. L. et al. Western Pacific hydroclimate linked to global climate variability over the past two millennia. Nat Commun 7, doi:10.1038/ncomms11719 (2016).

3Schneider, T., Bischoff, T. & Haug, G. H. Migrations and dynamics of the intertropical convergence zone. Nature 513, 45-53 (2014).

4Haug, G. H., Hughen, K. A., Sigman, D. M., Peterson, L. C. & Röhl, U. Southward migration of the intertropical convergence zone through the Holocene. Science 293, 1304-1308 (2001).

5Sachs, J. P. et al. Southward movement of the Pacific intertropical convergence zone AD 1400-1850. Nature Geosci 2, 519-525 (2009).

6Eroglu, D. et al. See-saw relationship of the Holocene East Asian-Australian summer monsoon. Nat Commun 7, 12929, doi:12910.11038/ncomms12929 (2016).

7Yan, H. et al. Dynamics of the intertropical convergence zone over the western Pacific during the Little Ice Age. Nat. Geosci. 8, 315-320 (2015).

8Denniston, R. F. et al. Expansion and contraction of the Indo-Pacific tropical rain belt over the last three millennia. Scientific Reports 6, 34485, doi:10.1038/srep34485 (2016).

9Cobb, K. M., Charles, C. D., Cheng, H. & Edwards, R. L. El Nino/Southern Oscillation and tropical Pacific climate during the last millennium. Nature 424, 271-276 (2003).

10Tan, M. Circulation background of climate patterns in the past millennium: Uncertainty analysis and re-reconstruction of ENSO-like state. Sci. China Earth Sci. 59, 1225-1241, doi:10.1007/s11430-015-5256-6 (2016).

11Yan, H. et al. A record of the Southern Oscillation Index for the past 2,000 years from precipitation proxies. Nat. Geosci. 4, 611-614 (2011).

12Dirzo, R. & Raven, P. H. Global state of biodiversity and loss. Annu. Rev. Env. Resour. 28, 137-167 (2003).

13Zhang, X. et al. Detection of human influence on twentieth-century precipitation trends. Nature 448, 461-465 (2007).

14Winter, A. et al. Persistent drying in the tropics linked to natural forcing. Nature communications 6 (2015).

15Sinha, A. et al. Trends and oscillations in the Indian summer monsoon rainfall over the last two millennia. Nature communications 6, 6309, doi: 6310.1038/ncomms7309 (2015).

16Hwang, Y. T., Frierson, D. M. & Kang, S. M. Anthropogenic sulfate aerosol and the southward shift of tropical precipitation in the late 20th century. Geophys. Res. Lett. 40, 2845-2850 (2013).

17Xie, S.-P., Lu, B. & Xiang, B. Similar spatial patterns of climate responses to aerosol and greenhouse gas changes. Nature Geosci 6, 828-832 (2013).

18Friedman, A. R., Hwang, Y.-T., Chiang, J. C. & Frierson, D. M. Interhemispheric temperature asymmetry over the twentieth century and in future projections. J. Clim. 26, 5419-5433 (2013).

19Lau, W. K. & Kim, K.-M. Robust Hadley Circulation changes and increasing global dryness due to CO2 warming from CMIP5 model projections. Proceedings of the National Academy of Sciences 112, 3630-3635 (2015).

20Breitenbach, S. F. et al. Cave ventilation and rainfall signals in dripwater in a monsoonal setting-a monitoring study from NE India. Chem. Geol. 402, 111-124 (2015).

21Fairchild, I. J. et al. Modification and preservation of environmental signals in speleothems. Earth-Science Reviews 75, 105-153 (2006).

22Dansgaard, W. Stable isotopes in precipitation. Tellus 16, 436-468 (1964).

23Myers, C. G. et al. Northeast Indian stalagmite records Pacific decadal climate change: Implications for moisture transport and drought in India. Geophys. Res. Lett. 42, 4124-4132 (2015).

24Bony, S., Risi, C. & Vimeux, F. Influence of convective processes on the isotopic composition (δ18O and δD) of precipitation and water vapor in the tropics: 1. Radiative‐convective equilibrium and Tropical Ocean-Global Atmosphere-Coupled Ocean‐Atmosphere Response Experiment (TOGA‐COARE) simulations. Journal of Geophysical Research: Atmospheres 113 (2008).

25Moore, M., Kuang, Z. & Blossey, P. A moisture budget perspective of the amount effect. Geophys. Res. Lett. 41, 1329-1335 (2014).

26Partin, J. W., Cobb, K. M., Adkins, J. F., Clark, B. & Fernandez, D. P. Millennial-scale trends in west Pacific warm pool hydrology since the Last Glacial Maximum. Nature 449, 452-455 (2007).

27Kennett, D. J. et al. Development and disintegration of Maya political systems in response to climate change. Science 338, 788-791 (2012).

28Lachniet, M. S. et al. A 1500‐year El Niño/Southern Oscillation and rainfall history for the isthmus of Panama from speleothem calcite. Journal of Geophysical Research: Atmospheres 109 (2004).

29Verschuren, D. in Past Climate Variability Through Europe and Africa. Developments in Paleoenvironmental Research,vol. 6. Kluwer Academic Publishers, Dordrecht, The Netherlands Chapter 8.   (eds R.W. Battarbee, F. Gasse, & C.E. Stickley)  (2004).

30Oppo, D. W., Rosenthal, Y. & Linsley, B. K. 2,000-year-long temperature and hydrology reconstructions from the Indo-Pacific warm pool. Nature 460, 1113-1116 (2009).

31Conroy, J. L., Overpeck, J. T., Cole, J. E., Shanahan, T. M. & Steinitz-Kannan, M. Holocene changes in eastern tropical Pacific climate inferred from a Galápagos lake sediment record. Quat. Sci. Rev. 27, 1166-1180 (2008).

32Bird, B. W. et al. A 2,300-year-long annually resolved record of the South American summer monsoon from the Peruvian Andes. Proceedings of the National Academy of Sciences 108, 8583-8588 (2011).

33Wang, X. et al. Interhemispheric anti-phasing of rainfall during the last glacial period. Quat. Sci. Rev. 25, 3391-3403 (2006).

34Wu, Z. & Huang, N. E. Ensemble empirical mode decomposition: a noise-assisted data analysis method. Advances in adaptive data analysis 1, 1-41 (2009).

35Broccoli, A. J., Dahl, K. A. & Stouffer, R. J. Response of the ITCZ to Northern Hemisphere cooling. Geophys. Res. Lett. 33 (2006).

36Zhang, R. & Delworth, T. L. Simulated tropical response to a substantial weakening of the Atlantic thermohaline circulation. J. Clim. 18, 1853-1860 (2005).

37Bond, G. et al. A pervasive millennial-scale cycle in North Atlantic Holocene and glacial climates. Science 278, 1257-1266 (1997).

38Chu, G. et al. The ‘mediaeval warm period’drought recorded in Lake Huguangyan, tropical South China. The Holocene 12, 511-516 (2002).

39Lamb, H. H. Climate, history and the modern world.  (Routledge, 2002).

40Xu, H. et al. Hydroclimatic contrasts over Asian monsoon areas and linkages to tropical Pacific SSTs. Scientific Reports 6, 33177, doi:10.1038/srep33177 (2016).

41Moy, C. M., Seltzer, G. O., Rodbell, D. T. & Anderson, D. M. Variability of El Nino/Southern Oscillation activity at millennial timescales during the Holocene epoch. Nature 420, 162-165 (2002).

42Thompson, L. et al. Annually resolved ice core records of tropical climate variability over the past~ 1800 years. Science 340, 945-950 (2013).

43Ropelewski, C. F. & Halpert, M. S. Global and regional scale precipitation patterns associated with the El Niño/Southern Oscillation. Monthly weather review 115, 1606-1626 (1987).

44Kumar, K. K., Rajagopalan, B., Hoerling, M., Bates, G. & Cane, M. Unraveling the mystery of Indian monsoon failure during El Niño. Science 314, 115-119 (2006).

45Vecchi, G. A. et al. Weakening of tropical Pacific atmospheric circulation due to anthropogenic forcing. Nature 441, 73-76 (2006).

46Man, W. & Zhou, T. Forced response of atmospheric oscillations during the last millennium simulated by a climate system model. Chin. Sci. Bull. 56, 3042-3052 (2011).

47Liu, Z. et al. Evolution and forcing mechanisms of El Niño over the past 21,000 years. Nature 515, 550-553 (2014).

48Mann, M. E., Cane, M. A., Zebiak, S. E. & Clement, A. Volcanic and solar forcing of the tropical Pacific over the past 1000 years. J. Clim. 18, 447-456 (2005).

49Zielinski, G. A. & Mershon, G. R. Paleoenvironmental implications of the insoluble microparticle record in the GISP2 (Greenland) ice core during the rapidly changing climate of the Pleistocene-Holocene transition. Geol. Soc. Am. Bull. 109, 547-559 (1997).

50Lavigne, F. et al. Source of the great A.D. 1257 mystery eruption unveiled, Samalas volcano, Rinjani Volcanic Complex, Indonesia. Proceedings of the National Academy of Sciences 110, 16742-16747, doi:10.1073/pnas.1307520110 (2013).

51De Silva, S. L. & Zielinski, G. A. Global influence of the AD 1600 eruption of Huaynaputina, Peru. Nature 393, 455-458 (1998).

52Stothers, R. B. The great Tambora eruption in 1815 and its aftermath. Science 224, 1191-1199 (1984).

53Iles, C. E., Hegerl, G. C., Schurer, A. P. & Zhang, X. The effect of volcanic eruptions on global precipitation. Journal of Geophysical Research: Atmospheres 118, 8770-8786, doi:10.1002/jgrd.50678 (2013).

54Harris, I., Jones, P., Osborn, T. & Lister, D. Updated high‐resolution grids of monthly climatic observations-the CRU TS3. 10 Dataset. Int. J. Climatol. 34, 623-642 (2014).

55Moberg, A., Sonechkin, D. M., Holmgren, K., Datsenko, N. M. & Karlen, W. Highly variable Northern Hemisphere temperatures reconstructed from low- and high-resolution proxy data. Nature 433, 613-617 (2005).

56Kobashi, T., Goto-Azuma, K., Box, J., Gao, C.-C. & Nakaegawa, T. Causes of Greenland temperature variability over the past 4000 yr: implications for northern hemispheric temperature changes. Climate of the Past 9, 2299-2317 (2013).

Acknowledgements:

We gratefully acknowledge the National Basic Research Program (2013CB955902) and NSF (41372192; 41290254) of China. This study was also supported by Taiwan ROC MOST (104-2119-M-002-003,105-2119-M-002-001) and NTU (101R7625) grants to C.-C.S. H.C. and R.L.E. received financial support from the U.S. NSF (EAR-0908792 and EAR-1211299), S.C. received support from the TRF (MRG5980080) and NRC of Thailand, and S.F.M.B. received funding from the European Union’s Horizon 2020 Research and Innovation programme under the Marie SkÅ‚odowska-Curie grant agreement No 691037.

Author contributions:

C.-C.S. and L.T. conceived the project and wrote the first draft of the paper. L.T., H.C., R.L.E. and C.-C.W. contributed to the ²³⁰Th dating. L.T. and Y.C. performed the stalagmite δ¹⁸O analysis. S.F.M.B. conducted the age model. S.C., L.L., A.C., B.W. and C.-C.S. did the field work. X.W., Y.C., R.L.E., H.C., S.F.M.B., Y. G., O.K., and Z.A. contributed to the manuscript revision at different stages. All authors discussed the results, edited and commented on the manuscript.

Additional information:

Competing financial interests: The authors declare no competing financial interests.

 

Figure Captions

Figure 1. Hydroclimatic maps with location of Klang cave and other comparison sites. Maps in (A) and (B) show global mean precipitation in January and July, respectively, during 1988-2004. Bands of heavy precipitation in the tropics indicate the ITCZ. (C) Monthly precipitation anomaly (cm/month) map during El Niño years from 1979-2006 (data source: http://jisao.washington.edu/data/gpcp/). Red star denotes location of Klang cave. Black dots indicate comparison sites mentioned in this paper: 1- ref. 2; 2-ref. 30; 3- ref. 26; 4-ref. 27; 5-ref. 28; 6-ref. 4; 7-ref. 31; 8-ref. 41; 9-ref. 32; 10-ref. 42; 11-ref. 29; 12-ref. 15; 13-ref. 40; 14-ref. 38.

Figure 2. (A) Stalagmite TK composited δ¹⁸O record and (B) its comparison with local rainfall record from 1900 to 2005 AD (data source: CRU TS3.21 grid datasets⁵⁴ ). ²³⁰Th dates with 2σ error bars are shown.

Figure 3. Comparison of reconstructed ITCZ shift index (ITCZ SI, emerald lines) with (A) Ti content in the sediment of Cariaco Basin⁴, (B) Indian monsoon rainfall variations recorded by stalagmite δ¹⁸O series from northern India¹⁵, (C) water depth of lake Naivasha in tropical east Africa (ref), (D) ice-rafted hematite-stained grains (HSG) record from North Atlantic³⁷, and (E) temperature gradient between Northern Hemisphere (NH) and Southern Hemisphere (SH) extratropics³.

Figure 4. (A) Comparison of stalagmite TK record with high-resolution rainfall records in the tropics during the last 2700 years. (a) TK δ¹⁸O record from southern Thailand (this study). (b) TOC/N ratio in sediment from tropical south China³⁸. (c) Multi-proxy synthesized stalagmite record (LLPC1) from eastern Indonesia². (d) Stalagmite δ¹⁸O record from Panama²⁸. (e) Stalagmite δ¹⁸O record from Belize²⁷. (f) Sedimentary calcite δ¹⁸O record from a varved lake in the central Peruvian Andes³². (g) Northern Hemisphere temperature (NHT) variations during the last 2700 years (red and blue lines from refs. 55 and 56, respectively). All records are detrended to emphasize their centennial- to decadal timescales variations. (B) Comparison of stalagmite TK rainfall with proxy records of ENSO and volcanic activities. (h) TK δ¹⁸O record. (i) reconstructed SOI record¹¹. (j) El Niño activities recorded by percent sand in lake sediment of El Junco from Galápagos. Increases in sand abundance represent more intense rainfall events associated with El Niño events³¹. (k) El Niño activities recorded in the red color intensity of sediment from Laguna Pallcacocha, southern Ecuador, with intense red color reflecting increased El Niño activities⁴¹. (l) Niño4 SST reconstructed from δ¹⁸O of annual layered ice core from Peru⁴². (m) Model simulated Niño3 index (gray)⁴⁶ and Niño3.4 SST variability (brown)⁴⁷. (n) Volcanic sulfate recorded in GISP2 ice core⁴⁹. Yellow bars denote dry conditions during Roman Warm Period (RWP), Medieval Warm Period (MWP), and Current Warm Period (CWP). Orange bars show two notably multidecadal droughts in TK record, associated with increased El Niño activities. Gray bars present wet Little Ice Age (LIA) and Dark Age Cold Period (DACP), respectively.