Modern stalagmite oxygen isotopic composition and its implications of climatic change from a high-elevation cave in the eastern Qinghai-Tibet Plateau over the past 50 years

An oxygen isotope record of a stalagmite from Huanglong Cave in the eastern Qinghai-Tibet Plateau dated with 230 Th and 210 Pb methods provides variations of the Asian monsoon with an average resolution of 1 year over the past 50 years. This study shows that the δ 18 O of dripwater in the cave represents the annual mean δ 18 O of local meteoric precipitation and the stalagmites were deposited in isotopic equilibrium. A comparison of the stalagmite δ 18 O record with instrumentally meteorological data indicates that shifts of the δ 18 O are largely controlled by the amount effect of meteoric precipitation con-veyed through the southwest monsoon (the Indian monsoon) and less affected by temperature. Therefore, the variations of δ 18 O record reflect the changes in monsoon precipitation on inter-annual time scales under the influence of the southwest monsoon. Like many other stalagmite δ 18 O records in the Asian monsoon regions, the δ 18 O record of the stalagmite from Huanglong Cave also reveals a gradually enriched trend during the past 50 years, i.e. relatively enriched in 18 O. This trend may indicate the decline of


ARTICLES
GEOLOGY mite δ 18 O values are controlled by many factors, such as δ 18 O in precipitation [5,6] , depositing process of calcite, cave temperature during depositing period, especially in the regions strongly influenced by summer monsoon [7,8] . Therefore, it is necessary to study oxygen isotopic composition and its implications of modern stalagmite systematically. Currently, great progress in history, amplitude and driving factors of the Asian monsoon variation has been made by the stalagmite records on glacial/interglacial and millennial time scales [9 -11] . Although, Burns et al. [12] discussed the relation between the stalagmite δ 18 O records and instrumental monsoon precipitation in southern Oman on inter-decadal scales, much research work is still needed on modern stalagmite depositing process and the changes of oxygen isotope composition in responses to the climate change on short time scales. Thus, based on the research of modern carbonate-water oxygen isotope system and the instrumental observation data, this paper probes into the stalagmite oxygen isotopic composition and its implications from a high-elevation cave-Huanglong Cave in the eastern Qinghai-Tibet Plateau, and discusses the relationship between monsoon precipitation and upper stratospheric temperature changes over the past 50 years, which would help us improve understanding the variations of the Asian monsoon and the future climate.

Study site and sample description
Huanglong Cave (32°43′N, 103°49′E, 3588 m a. s. l.) is formed in Triassic limestone at Huanglong Ravine with high-elevation mountainous climate in the eastern Qinghai-Tibet Plateau [13] . This site is sensitive to monsoon change at fringe zone of the eastern Qinghai-Tibet Plateau influenced by interaction of the southwest monsoon and the East Asian monsoon with mean annual temperature (MAT) of 4℃ and mean annual precipitation (MAP) of 759 mm. Therefore, it is an attractive place to explore relationship between the stalagmite δ 18 O records and the Asian monsoon precipitation.
We sampled vertically along the growth axes of two huge stalagmites using portable core drilling and obtained two stalagmite cores in length of 300 mm (HL021) and 135 mm (HL022) with diameter of 30 mm, respectively. Field investigations showed that there was plenty of dripwater in the cave and a great deal of water membrane on the top of stalagmites, indicating that the two stalagmites were actively growing in May, 2002 when samples were collected. The following 210 Pb dating and study of modern carbonate deposit also confirmed that the two stalagmites were still growing. HL021 and HL022 were compactly crystallized with light white and some gray laminas in the top, and with no apparent hiatus. In order to understand the process of modern stalagmite deposition and to compare the stalagmite δ 18 O records with instrumental meteorological data, the paper mainly presents the results from the top 6 mm of HL021 and the top 10 mm of HL022. We also systematically collected some samples of cave dripwater from entrance to the end of the cave and some actively dripping, uncrystal growing soda straw stalactites sagged from the roof of the cave in October, 1999 and May, 2002. Meanwhile several glass slides were also put in the cave at the dripping location to collect depositing carbonate. The data-logger for detecting the cave temperature and relative humidity (RH) was set in the deepest site from the entrance in the cave, which automatically and continuously worked from May, 2001 to May, 2002.
In order to obtain high-resolution oxygen isotopic records, we firstly halved the stalagmites vertically from the bottom to the top along the growth axes, polished their cutting surfaces, and then, with about average interval of 50 μm, scraped sub-samples off along successive laminations. To avoid cross-contamination of sub-samples, alternative sub-samples were selected for analyzing. 52 δ 18 O data were obtained along the growth axis of HL021, resulting in an average resolution of ~1 year, and 22 δ 18 O data were obtained from HL022 with a time resolution of ~2 years. One sample is taken off to parallel the growth banding of stalagmite from HL021 and HL022 respectively, for MC-ICPMS 230 Th-dating; 8 samples from HL021 and 7 samples from HL022 for 210 Pb-dating were symmetrically collected downwards from the top at intervals of 2.5 mm with knife (Table 1).

Analytical methods
The δ 18 O and δD values of cave dripwater were measured at the Center of Isotope Geochemistry, University of California, Berkeley, where δ notation represents hydrogen and oxygen data, respectively, δ = [(R sample /R std ) −1] ×1000, in which R stands for D/H and 18  Carbon dioxide for isotopic analysis was produced with McCrea's phosphoric acid method [14] . Modern stalagmite oxygen isotope analysis was carried out with Finnigan-Delta-Plus mass spectrometer in the Key Laboratory of Western China's Environmental Systems, Ministry of Education, Lanzhou University. Oxygen data are reported in δ notation relative to the VPDB, where δ = [(R sample /R std ) −1] ×1000, in which R stands for 18 O/ 16 O. The precision for each analysis is ±0.1‰ for δ 18 O. The 230 Th dating was finished in Isotopic Chronology Laboratory of Department of Geology and Geophysics, University of Minnesota, USA. The 230 Th dating procedures and methods were described in ref. [15], and the 2σ errors are reported. The 210 Pb activity of stalagmite was measured by low background multi-channel alpha spectrometry in the Radioisotope Lab of the Academia Sinica, Taiwan, China. The 210 Pb dating procedures were described in refs. [16,17].

Stalagmite age model
Stalagmite HL022 and HL021 are young with dense texture and no hiatus; they have high uranium concentration and appear neither recrystallization, nor erosion phenomena. So they are suitable for 230 Th and 210 Pb dating methods. 210 Pb has a half-life of 22.3 years, so 210 Pb dating method method is suitable for constructing a precise timescale of stalagmite in the cave where 210 Pb is relatively enclosed during the deposition for a short period [16] . The top 20mm of HL021 and HL022 is in single color, which implies relatively smooth growth rates during this interval. The determination of 210 Pb age and the calculation of stalagmite growth rates are based on the 210 Pb activity decay with time for the last 50 years. We plotted the 210 Pb activity against the depth (Table 1, Figure 1). The stalagmite 210 Pb activity exhibits an exponential decay until the depth of 16.25 mm. The 210 Pb activity remains a constant below this depth, indicating that the 210 Pb activity below this depth is equilibrated with its supported source (background values), so we can draw a conclusion that the sample is younger than 100 years and the 210 Pb activity mostly comes from excess 210 Pb of water and atmosphere above this depth. The growth rates of HL021 and HL022 were obtained by exponentially fitting of the decrease of the 210 Pb activity with the depth. The maximum growth rates of HL021 and HL022 determined by total 210 Pb(y tot ) activity are 0.246 mm/a and 0.144 mm/a, respectively. Assuming that the 210 Pb activity below 16.25 mm represents the radioactive background value of 210 Pb, the average growth rates of the two stalagmites determined by excess 210 Pb(y ex ) activity are 0.104 mm/a (HL021) and 0.143 mm/a (HL021), respectively. The 230 Th dating results of HL022 and HL021 are shown in Table 2. The average growth rates of HL022 and HL021 are estimated by the 230 Th dating in order to test the reliability of the 210 Pb dating results. According to the 230 Th dating data, the average growth rates are   0.105 mm/a for HL021 and 0.145 mm/a for HL022, which are almost in agreement with the average growth rates based on excess 210 Pb. Moreover, the 210 Pb dating results show a marked peak value in the vicinity of 1963 which probably resulted from a great scale proliferation of radioactive material in nuclear tests of the 1960s [18,19] , which indirectly proved the reliability of the 210 Pb dating results of stalagmite in Huanglong Cave. Therefore, using the result of the excess 210 Pb (y ex ) dating, we established an absolute-dated oxygen isotope record from Huanglong Cave over the past 50 years.

Modern cave carbonate-water isotopic system
A key issue is whether Huanglong Cave stalagmite δ 18 O values can be interpreted by the δ 18 O in meteoric precipitation and cave temperature during calcite precipitation [20,21] . Huanglong Cave is a relatively closed cave with low MAT and RH of 100%. Systemic research on modern stalagmite deposition conditions and cave dripwater isotope composition was performed to test whether the cave dripwater came from meteoric precipitation or not. Identical replication tests of two stalagmites in the same cave are virtually criteria to check stalagmite equilibrium deposition in the cave temperature over the contemporaneous growth period [20,21] . To characterize modern cave dripwater, we plotted the values were directly plotted on LMWL, suggesting that the cave water had not been significantly affected by evaporative processes and cave dripwater originated from meteoric precipitation. Therefore, the isotopic composition of the dripwater in Huanglong Cave could reflect the isotopic composition of meteoric precipitation.
Another robust test is the comparison of δ 18 O from contemporaneous stalagmite records in the same cave.
The δ 18 O records of HL021 and HL022 are virtually identical during the period when the two stalagmites grew contemporaneously ( Figure 3). According to Hendy test criteria [21] , the stalagmites of Huanglong Cave were deposited in isotopic equilibrium during the whole growth period, and the δ 18 O values of stalagmite mainly depend on cave temperature and δ 18 O of meteoric precipitation [21,22] .

Stalagmite δ 18 O record
The δ 18 O values of HL021 range from −11.55‰ to the amplitude of only 0.35‰ at the same growth period [24] . In Kahf Defore Cave of southern Oman (17°07′N, 54°05′E,150 m a. s. l.), located in tropical zone with low-altitude [12,25] , the stalagmite δ 18 O values range from 0.20‰-−1.01‰, with the mean value of −0.38‰ in the same growth period [12,25] . Previous studies verified that the stalagmite δ 18 O records from the Asian monsoon regions mainly reflected the variations of the Asian monsoon intensity on the glacial/interglacial or stadial/interstadial scales and represented the temporal oxygen isotope information of the precipitation [11,20,[24][25][26][27] . But there are many uncertainties about climatic implications of the stalagmite oxygen isotopic composition in the Asian monsoon regions on shorter time scales and inter-annual resolution. Based on correlative analysis, we found that HL021 δ 18 O record of Huanglong Cave significantly and negatively correlates to summer precipitation from the Songpan Meteorological Station in Sichuan Province during the period

GEOLOGY
of 1951-2002. The correlation coefficient for the full record is −0.53 (significant at the 99.95% confidence level), which indicates that the oxygen isotopic composition of HL021 is significantly affected by the amount of precipitation (known as "amount effect") ( Figure 4). The regions near Huanglong Cave are controlled by the southwest monsoon. More than 80% of total annual precipitation falls during the summer monsoon months (May to October) when the convergence between the low-level southwest monsoon winds off the Bay of Bengal and the cold, dry Siberian air off to southward brings more rain amount and stronger monsoon.  [28,29] , suggesting that the East Asian monsoon and the southwest monsoon have consistent variations over the past half-century. Moreover, the δ 18 O records from Dongge Cave and Hulu Cave [27,30] also show that the southwest monsoon and the East Asian monsoon have similar variations on millennial time scales.
HL021 δ 18 O record has a significant negative correlation with the East Asian summer monsoon index (EAMI) constructed by Jiang and Wang [31] using reanalysis data derived from the United States National Centers for Environmental Prediction and National Center for Atmospheric Research (NCEP/NCAR) through 1951 to 2002, and the correlation coefficient is −0.58 (significant at the 99.95% confidence level) ( Figure 6). HL021 δ 18 O record also has a significant negative correlation with the global monsoon index (GMI) constructed by Wang and Ding [33] , using four sets of rain-gauge precipitation data sets compiled for the period of 1948-2003 by climate diagnostic groups around the world, with the correlation coefficient of −0.52 (significant at the 99.95% confidence level). That is to say, as the δ 18 O values of the stalagmite turn lighter, the intensity of the Asian monsoon gets stronger and vice versa. Based on the intensity of the normalized seasonality of wind field, Li and   [12,25] ; (c) Shihua Cave δ 18 O record [28,29] . Oblique lines indicate their trends by linear fitting.
Zeng [32,36] constructed the monsoon indices to describe both seasonal variations and inter-annual variability of monsoon in different monsoon regions, and discovered that there existed an overall weakening of the monsoon intensity in South Asian, East Asian, North American and West African monsoon regions since the 1950s. In summary, the above analyses show that the southwest monsoon and the East Asian monsoon, which make up of the Asian monsoon, have a synchronous weakening since the 1950s, and the modern monsoon is globally weakening, but the trend has become inconspicuous after 1980, and these analyses also support the idea that the inter-annual changes in the stalagmite δ 18 O of Huanglong Cave indicate the inter-annual variations of the Asian monsoon. Yuan et al. [27] and Dykoski et al. [30] pointed out that the δ 18 O values of the stalagmite, deposited under the isotopic equilibrium, could be interpreted mainly in terms of the δ 18 O of precipitation and cave temperature in the Asian monsoon regions [27,30] . The effect of temperature on the stalagmite δ 18 O is quite complex.
Changes in the temperature-dependent fractionation between calcite and water are small [37] (−0.24‰/ ). On 0.69‰/℃ [38] . Other factors, however, may obscure this relation in the Asian monsoon regions especially. Even if temperature positively correlates to the oxygen isotope composition of stalagmites in low-mid-latitude monsoon regions, the temperature effect is relatively weak for it is counteracted by the strong negative relationship between the stalagmite oxygen isotope and the summer precipitation [39,40] . As to Huanglong Cave, we have found that Lots of researchers have demonstrated that the Asian monsoon was weakening since the 1950s [31][32][33]36] . The changes of the Asian monsoon on millennial-scale were driven by orbitally induced changing of Northern Asian monsoon index (EAMI) [31] ; (c) South Asian monsoon index (SAMI) [32] ; (d) global monsoon index (GMI) [33] ; (e) solar insolation anomalies in South China (SRF down-SW flux anomalies) [34] . Oblique lines indicate their trends by linear fitting.
Hemisphere summer solar insolation [11,26,27,[41][42][43] . However, the causes of the monsoon changes are relatively complex on inter-decadal time scales. Based on instrumental meteorological data provided by NCEP and Climate Data Center of Chinese Meteorological Administration (CDC/CMA), Xia et al. [34] found that insolation declined significantly in China from 1961 to 2000, and the largest decline appeared in southern China, where solar insolation decreased by 5% per decade. The variations of insolation is negatively correlated to the δ 18 O variations of Huanglong Cave during the same period, and the correlation coefficient for the full record is −0.36 (significant at the 95.00% confidence level), which suggests that monsoon changes were affected by the changes of insolation on inter-decadal time scales. He et al. [44] pointed out that the variations of the summer monsoon were closely related to the middle and upper tropospheric temperature variations, and the weakening of the summer monsoon happened in the mid-1960s mainly due to the obvious decline of the troposphere temperature in the East Asian monsoon regions and the African monsoon regions. Based on Microwave Sound-ing Unit satellite observations and model simulation, Ramaswamy et al. [45] revealed that the gradual cooling of the global lower stratosphere over 1979-2003 could be attributed to natural and anthropogenic influences on the evolution of the cooling. Box and Cohen [46] have recently confirmed that the temperature of the upper troposphere and the stratosphere over Greenland was gradually lowering in 1964 -2005. Vecchi et al. [47] pointed out that global warming gradually reduced the sea level pressure gradient of tropical Indo-Pacific, which consequently led to the weakening of the Pacific Walker circulation and influenced the monsoonal circulations over adjacent continents since the mid-nineteenth century. In summary, instrumental observation data indicate that the solar radiation in China monsoon regions is weakening, and the temperature of the stratosphere and the upper troposphere is also decreasing [44 -48] . Therefore, it is a natural cause that possibly leads to the weakening of the Asian monsoon over the past 50 years. Up to now, exploring the reasons of the weakening of the Asian monsoon intensity on short scales is still in a preliminary stage, and better understanding of factor candidates and consequent mechanisms related to the weakening is of great significance for future climate forecast.

Conclusions
The changes of the stalagmite δ 18 O of high-elevation Huanglong Cave, located on the eastern margin of the Qinghai-Tibet Plateau, were sensitive to the changes of the Asian monsoon.
The study on modern stalagmite deposition process and replication tests for isotopic equilibrium suggest that cave dripwater mainly originated from meteoric precipitation and the oxygen isotope composition of stalagmite could indicate the climate change outside Huanglong Cave. The high resolution stalagmite δ 18 O record over the past 50 years allows us to compare this record with instrumental meteorological data, and we find that the