Age of the Tibetan ice cores

An accurate chronology is the essential first step for a sound understanding of ice core records, however, dating of ice cores drilled from the high elevation glaciers is challenging and often problematic, leading to great uncertainties. The Guliya ice core, drilled to bedrock (308.6 m in length) from the northwestern Tibetan Plateau (TP) and widely used as a benchmark for paleoclimate research, is believed to 20 reach > 500 ka (thousand years) at its bottom. Meanwhile other Tibetan ice cores (i.e., Dasuopu and East Rongbuk in the Himalayas, Puruogangri in the central TP, and Dunde in the northeastern TP) are mostly of the Holocene origin. In this study, we drilled ice cores to bedrock from the Chongce ice cap ~30 km from the Guliya ice core drilling site. We performed measurements of C, Pb, tritium and β-activity for 25 the ice cores, and used these values in a two-parameter flow model to establish the ice core depth-age relationship. The modeled ages of two Chongce ice cores at the ice-bedrock contact are 8.3±3.6 6.2 ka B.P. and 9.0 ± 3.6 7.9 ka B.P. respectively. The significant discrepancy between the Guliya and all other Tibetan ice core chronologies calls for a revisit of this legend ice core record. 30 2 The Cryosphere Discuss., https://doi.org/10.5194/tc-2018-55 Manuscript under review for journal The Cryosphere Discussion started: 11 April 2018 c © Author(s) 2018. CC BY 4.0 License.


Introduction
Ice cores from the Tibetan Plateau (TP) provide a wealth of information for past climatic and environmental conditions that extends beyond the instrumental period (e.g., Thompson et al., 1989;1997;2000).An accurate chronology is the essential 35 first step for a sound understanding of such ice core records.However, ice core dating is always a challenging task because seasonal signals suitable for annual layer counting are usually only observable in top sections of ice cores.For deeper (older) sections, annual cycles cannot be identified due to rapid thinning of ice layers.If sufficient organic matter (e.g., plant or insect fragments) is found inside the ice cores, 40 the conventional radiocarbon ( 14 C) dating can be used (Thompson et al., 2002).
Unfortunately, the presence of such material is far from guaranteed, which limits its application for ice core dating.Recently, a novel method was developed to extract water-insoluble organic carbon (WIOC) particles at microgram level from carbonaceous aerosol embedded in the glacier ice for Accelerator Mass Spectrometry 45 (AMS) 14 C dating (Jenk et al., 2007;Uglietti et al., 2016).Carbonaceous aerosol is constantly transported to the glaciers, where it is deposited and finally incorporated in 3 The Cryosphere Discuss., https://doi.org/10.5194/tc-2018-55Manuscript under review for journal The Cryosphere Discussion started: 11 April 2018 c Author(s) 2018.CC BY 4.0 License.glacier ice.Consequently, carbonaceous aerosol in ice cores can provide reliable dating at any given depth when the samples contain sufficient carbon mass (> 10 µg).
Here we applied this recently established technique for dating the Tibetan ice cores.50

Chronology of previous ice cores
The Dunde ice cores are the first ones ever drilled on the TP (Fig. 1).In 1987, three ice cores (139.8 m, 136.6 m and 138.4 m in length) were drilled to bedrock at an altitude of 5325 m a.s.l.(above sea level) from the Dunde ice cap (38°06′ N, 96°24′ E) 55 in the Qilian Shan mountains on the northeastern TP (Fig. 1).Their original chronology was suggested to be 40 ka B.P. (before present, i.e., before 1950 AD) at the depth of 5 m above the ice-bedrock contact, and potentially more than 100 ka B.P. at the ice-bedrock contact (Thompson et al., 1989).Later, Thompson et al. (2005) provided a single 14  of 6200 m a.s.l.from the Guliya ice cap (35°17′ N, 81°29′ E) on the northwestern TP (Fig. 1).Top 266 m of the Guliya core was dated to a period spanning 110 ka B.P., 65 and the ice below 290 m was suggested to be >500 ka B.P. based mainly on 36 Cl-dead ice at the bottom section (Thompson et al., 1997).
The Dasuopu ice cores: In 1997, three ice cores were drilled from the Dasuopu glacier (28°23′ N, 85°43′ E) in the Himalayas.The first core (159.9m in length) was drilled at an altitude of 7000 m a.s.l., and two more cores (149.2 and 167.7 m in length, 70 respectively) were drilled to bedrock 100 m apart on the col at an altitude of 7200 m a.s.l.(Thompson et al., 2000).It was suggested that the Dasuopu ice field accumulated entirely during the Holocene (Thompson et al., 2005).
The Grigoriev ice core is drilled at the top of the Grigoriev ice cap in the west Tien Shan (41°59′ N, 77°55′ E; Fig. 1).In 2007, an ice core (86.87 m in length) was drilled to bedrock at an altitude of 4563 m a.s.l..The 14 C dating of organic soil collected from 90 the bottom of the ice core borehole showed that the age of the soil was 12.656-12.434ka B.P., coincident with the beginning of the Younger Dryas cold period (Takeuchi et al., 2014).
In 2012, we drilled two ice cores to bedrock with length of 133.8 m (Core 1) and 135.8 m (Core 2) and a shallow core (Core 3) of 58.8 m at an altitude of 6010 m a.s.l.
from the Chongce ice cap on the northwestern TP (35º14′ N, 81º7′ E; Fig. 1).The direct distance between the Chongce and the Guliya ice core drilling sites is ~30 km (Fig. S1).In 2013, two more ice cores were recovered to bedrock with the length of 100 removing the ~3 mm outer layer with a bandsaw in a -20 °C cold room and rinsing with ultra-pure water in a class 100 laminar flow box.The water-insoluble organic carbon (WIOC) fraction of carbonaceous particles in the sample was filtered onto freshly preheated quartz fiber filters (Pallflex Tissuquartz, 2500QAO-UP), then combusted stepwise (10 min at 340 °C; 12 min at 650 °C) using a thermal-optical 120 carbon analyzer (Model4L, Sunset Laboratory Inc., USA) for separating organic carbon (OC) from elemental carbon (EC), and the resulting CO 2 was measured by the Mini Carbon Dating System (MICADAS) with a gas ion source for 14 C analysis at the University of Bern LARA laboratory.Details about sample preparation procedures and analytical methods can be found in previous studies (Jenk et al., 2007(Jenk et al., , 2009;;Sigl 125 et al., 2009;Uglietti et al., 2016).The overall procedural blanks were estimated using artificial ice blocks of frozen ultra-pure water, which were treated the same way as 8 The Cryosphere Discuss., https://doi.org/10.5194/tc-2018-55Manuscript under review for journal The Cryosphere Discussion started: 11 April 2018 c Author(s) 2018.CC BY 4.0 License.real ice samples.The average overall procedural blank is 1.34±0.62µg carbon with a F 14 C of 0.69±0.13(Uglietti et al., 2016).Conventional 14 C ages were calibrated using OxCal v4.2.4 software with the IntCal13 calibration curve (Bronk Ramsey and Lee, 130 2013;Reimer et al., 2013). 210

Pb
The accessible time range using radioactive isotope 210 Pb dating is ~150 years due to the 22.3-year half-life of 210 Pb, a product of the natural 238 U decay series.Here 210 Pb 135 dating was performed on the Chongce 216.6 m Core 4, with a total of 52 samples collected from the depth of 0-76.6 m.Each sample ( ~100 -200 g) was cut parallel to the drilling axis in a -20 °C cold room.The samples were processed according to the standard method established by Gäggeler et al. (1983).The samples were melted for 24 hours after adding 0.05% (V:V) analytical reagent HCl (30%).Afterwards, 100 µL 140 ages due to fossil fuel ( 14 C dead) contribution (Jenk et al., 2006).Only 2% fossil contribution would shift the mean of the calibrated age ranges for these samples by up to 200 yrs towards younger ages, resulting in a smaller age range close to the ages 210 estimated by 210 Pb dating.The 14 C age profile in the depth range of 80-180 m shows large scatter and no clear increase in age (Fig. 4).This is likely caused by the relatively young age of samples in combination with relatively large analytical uncertainties due to the presence of high mineral dust load in the Chongce ice core.
We made use of the 14 C ages (excluding the top four samples for the reasons 215 discussed above), the 210 Pb results (Fig. 3), and the tritium horizon (Fig. 2) to establish the depth-age relationship for the Chongce 216.6 m Core 4 (Fig. 4), by applying a two-parameter flow model (2p model) (Bolzan, 1985  For the Chongce 135.8 m Core 2, a depth-age relationship using the 2p model was also attempted (Fig. 5).In this case the model is constrained by the 14 C cal. ages and the β-activity horizon of the Chongce 58.8 m Core 3 (Fig. 2), assuming a similar 230 depth-age relationship for the upper parts of Core 2 and Core 3, which is reasonable given that their drilling sites are only several meters apart (Fig. S2).Although the derived annual accumulation rate of 137±54 mm w.e./year is in good agreement with the 140 mm w.e./year derived from the tritium horizon (Fig. 2), we find the model to be poorly constrained for this simulation.For instance, the derived age at the depth of 235 the oldest 14 C sample is 9.1 ± 4.0 7.2 ka B.P., much older than the actual 14 C age (6.3±0.2 ka B.P.) at that depth.The large uncertainty (Fig. 5) further indicates that the model is additional constraint (Fig. 6).With the additional age constraint at the bottom, the 1σ confidence interval for the model is significantly reduced.The ice age at the bedrock for the Chongce 135.8 m Core 2 is thus estimated to be 9.0 ± 3.6 7.9 ka B.P..This seems to be a reasonable estimate considering 1) the so derived accumulation rate (103±34 mm w.e./year) is in relative agreement with the tritium based estimate (140 mm 245 w.e./year); and 2) the modeled age at the depth of the oldest 14 C sample is now 5.2 ± 1.4 1.9 ka B.P., similar to the actual 14 C age of 6.3±0.2 ka B.P. given the uncertainty range.Although this result is far from satisfying, it is much better than the result obtained from the model without the additional bottom age constraint.
We have noticed that the reconstructed average annual accumulation rate from the 250 Chongce 216.6 m Core 4 roughly doubles the rate reconstructed from the Chongce 58.8 m Core 3. It is possible that the Core 4 drilling site may receive extra snow supplies, such as snow drifting, whereas part of the snow deposition at the Core 3 drilling site may be blown away due to wind scouring (Fisher et al., 1983).A full 16The Cryosphere Discuss., https://doi.org/10.5194/tc-2018-55Manuscript under review for journal The Cryosphere Discussion started: 11 April 2018 c Author(s) 2018.CC BY 4.0 License.be of Holocene origin (Thompson et al., 2005;Hou et al., 2004).The oldest cal.It is apparent that the temporal scale of the Guliya ice core is at least an order of magnitude older than other TP and the Tien Shan cores.Thompson et al. (2005) considered this as evidence that the growth (glaciation) and decay (deglaciation) of 275 large ice fields in the lower latitudes are often asynchronous.Our new understanding of the chronology of the Chongce ice cores that were drilled only ~30 km away from the Guliya ice core drilling site cannot back up this evidence.Though the validity of the Guliya chronology has been assumed repeatedly since its publication (Thompson et al., 2005(Thompson et al., , 2017)), Cheng et al. (2012) argued that the Guliya ice core chronology 280 should be shortened by a factor of two in order to reconcile the difference in the δ 18 O variations between the Guliya ice core and the Kesang stalagmite records (Fig. 1 and Supplement).It is worth pointing out that the 36 Cl-dead in the bottom section of the Guliya core provided the key evidence for the existence of ice older than 500 ka (Thompson et al., 1997).Since 36 Cl is presumably present as water-soluble ion, it can 285 be easily removed from the snow or firn layer by meltwater percolation.Thus, the 18 The Discuss., https://doi.org/10.5194/tc-2018-55Manuscript under review for journal The Cryosphere Discussion started: 11 April 2018 c Author(s) 2018.CC BY 4.0 License.absence of 36 Cl could have resulted from its leaching with glacier meltwater, as indicated by the abnormally low chloride concentrations in the bottom section of the Guliya record (Thompson et al., 1997), although a thorough explanation is beyond the scope of the current work.Nevertheless, much work is necessary for conviction of the 290 exceptional length of the Guliya ice core.

5
The Cryosphere Discuss., https://doi.org/10.5194/tc-2018-55Manuscript under review for journal The Cryosphere Discussion started: 11 April 2018 c Author(s) 2018.CC BY 4.0 License.drilled on the col of East Rongbuk Glacier (28º1′ N, 86º58′ E, 6518 m a.s.l.) on the north slope of Qomolangma (Mount Everest) in the Himalayas.In 2002, two more ice cores (108.8 m and 95.8 m in length, respectively) were drilled to bedrock nearby the previously drilling site.In a previous study, we matched the CH 4 /δ 18 O atm phase record of both the East Rongbuk 117.1 m and 108.8 m cores to the GRIP CH 4 and the GISP2 85 δ 18 O atm of the Greenland summit ice cores, and the results suggest a Holocene origin of the East Rongbuk ice cores 216.6 m (Core 4) and 208.6 m (Core 5) at an altitude of 6100 m a.s.l. on the Chongce ice cap (35°15′ N, 81°5′ E).The detailed positions of the five Chongce ice cores are shown in Fig. S2.All the ice cores were transported frozen to the cold room in the Nanjing University for further processing.The basal sediment collected from the bottom of Core 4 was measured for the first luminescence dating, resulting in an age 105 of 42±4 ka B.P., which was regarded as an upper constraint for the age of the bottom ice at the drilling site (Zhang et al., 2018).., https://doi.org/10.5194/tc-2018-55Manuscript under review for journal The Cryosphere Discussion started: 11 April 2018 c Author(s) 2018.CC BY 4.0 License.We performed 14 C measurements on WIOC extracted from 22 samples collected discretely along the 216.6 m Chongce Core 4 and 9 samples along the 135.8 m Chongce Core 2, as well as 5 samples collected from the East Rongbuk 95.8 m ice core.The 14 C sample decontamination was performed at Paul Scherrer Institute by 115

209
Po tracer was added to the solution to determine the yield of the separation.Spontaneous deposition of Po on an Ag disk (15 mm diameter), which was fixed on a wire and immersed in the liquid, was achieved during ~7 hours at 95 °C in 500 mL 9 The Cryosphere Discuss., https://doi.org/10.5194/tc-2018-55Manuscript under review for journal The Cryosphere Discussion started: 11 April 2018 c Author(s) 2018.CC BY 4.0 License.Erlenmeyer flasks using a magnetic stirrer.After drying, the disks were measured by α-counting at the Paul Scherrer Institute.The samples were positioned in vacuum 145 chambers at a distance of 1mm from silicon surface barrier detectors (ORTEC, ruggedized, 300 and 450 mm 2 ) having an α-energy resolution of ~23 keV full width at half-maximum at 5.3 MeV.The yield of 209 Po tracer was measured via its 4.9 MeV α-line.Typical chemical yields were ~75%.150 Tritium Tritium measurements were performed on the Chongce 216.6 m Core 4, with 51 samples collected successively from the depth range of 6.7-11.8m (corresponding to a sampling resolution of ~ 0.1 m per sample), and 42 samples from the depth range of 11.8-32.0m (corresponding to a sampling resolution of ~ 0.5 m per sample).Each 155 sample is ~10 g.Samples were analyzed at the Paul Scherrer Institute using liquid scintillation counting (TriCarb 2770 SLL/BGO, Packard SA).β-activity 10 The Cryosphere Discuss., https://doi.org/10.5194/tc-2018-55Manuscript under review for journal The Cryosphere Discussion started: 11 April 2018 c Author(s) 2018.CC BY 4.0 License. the Chongce 58.8 m Core 3.Each sample is ~1 kg.The β-activity was measured using Alpha-Beta Multidetector (Mini 20, Eurisys Mesures) at the National Key Laboratory of Cryospheric Sciences, China.More details can be found in An et al. (2016).5 Results 165 The β-activity profile of the Chongce 58.8 m Core 3 is shown in Fig. 2a.A β-activity peak at the depth of 8.2-8.4 m was referenced as 1963 AD, while a second β-activity peak at the depth of 4.8-5.1 m was set as 1986 AD, corresponding to the 1986 Chernobyl nuclear accident.Both β-activity peaks were also observed in the Muztagata ice core from the eastern Pamir (Tian et al., 2007).The calculated mean 170 annual accumulation rate is 140 mm w.e.(water equivalent) /year for the period of 1963-2012 AD.The tritium profile of the Chongce 216.6 m Core 4 is shown in Fig. 2b.The tritium activity was not corrected for decay to the time of deposition, because our purpose is to identify the apparent tritium peak (3237±89 TU) at the depth of 21.4 m, which was 175 11 The Cryosphere Discuss., https://doi.org/10.5194/tc-2018-55Manuscript under review for journal The Cryosphere Discussion started: 11 April 2018 c Author(s) 2018.CC BY 4.0 License.attributed to the thermonuclear bomb testing during the period of 1962-63 AD.The calculated mean annual accumulation rate is 297 mm w.e./year for the period of 1963-2013 AD.The 210 Pb activity profile of the Chongce 216.6 m Core 4 is shown in Fig. 3, which shows an exponential decrease as a function of depth in line with the radioactive 180 decay law.The 210 Pb activity concentrations are in the range 7.5-317 mBq/kg, but keep relatively stable for the lower 16 samples, with an average of 11.2±2.1 mBq/kg (not shown).This average was taken as background 210 Pb (BGD) from the mineral dust contained in the ice core and was subtracted from the measured 210 Pb activity concentrations.From the linear regression of the logarithmic 210 Pb activities (BGD 185 subtracted) against depth (Fig. 3), the value of the axis intercept (236±33 mBq/kg) corresponds to the 210 Pb activity at the surface of the Chongce ice cap.Thus we applied the following function to calculate the ice age, assuming a constant initial concentration (CIC) model.We calculated 1891±15 AD at the depth of 44.09 m (i.e.34.36 m w.e.), resulting in a mean annual net accumulation rate of 280±47 mm 190 w.e./year for the period of 1891-2013 AD.This value is in very good agreement with 12 The Cryosphere Discuss., https://doi.org/10.5194/tc-2018-55Manuscript under review for journal The Cryosphere Discussion started: 11 April 2018 c Author(s) 2018.CC BY 4.0 License. the 297 mm w.e./year for the period of 1963-2013 AD derived from the tritium profile of the same ice core (i.e., the Chongce 216.6 m Core 4, Fig. 2b).t s =λ -1 ln( C 0 C S ) Where, t s stands for the age of ice at a certain depth with 210 Pb activities (subtracted) Cs, λ for the decay constant of 210 Pb (0.03114 a -1 ), and C 0 for the 210 Pb surface 195 activity.The 14 C age profile of the Chongce 216.6 m Core 4 is shown in Fig. 4. We collected the 14 C samples taking into consideration of the chronology of the Guliya ice core, but finally realized that most of the samples, especially those collected from the upper sections, are too young to be dated with an acceptable uncertainty.For instance, we 200 obtained 1891±15 AD at the depth of 44.09 m from the 210 Pb measurements (Fig. 3), and the 14 C ages are 0.013-0.269ka cal B.P. at the depth of 40.11-40.97m, and modern to 0.430 ka cal B.P. at 50.06-50.82m.Even though all obtained calibrated age ranges of the uppermost four samples include the expected ages based on the 210 Pb dating results, they have large uncertainties due to the young age and the relatively 205 flat shape of the calibration curve in the past 500 yrs.Furthermore, anthropogenic 13 The Cryosphere Discuss., https://doi.org/10.5194/tc-2018-55Manuscript under review for journal The Cryosphere Discussion started: 11 April 2018 c Author(s) 2018.CC BY 4.0 License.contribution for samples younger than 200 yrs is likely introduce an old bias in 14 C poorly constrained.Given the close proximity between the Chongce 216.6 m Core 4 The Cryosphere Discuss., https://doi.org/10.5194/tc-2018-55Manuscript under review for journal The Cryosphere Discussion started: 11 April 2018 c Author(s) 2018.CC BY 4.0 License.and the Chongce 135.8 m Core 2 and similar bottom altitude of their drilling sites, we used the estimated age at bedrock derived for the Chongce 216.6 m Core 4 as an 240 14 C age for a sample collected down to the ice-bedrock contact of the East Rongbuk 95.8 m ice core, is 6.72±0.43ka B.P., confirming its Holocene origin.The ice cores from Puruogangri in the central TP, and, to a less degree, Dunde in the northeastern TP are of Holocene origin, too (Thompson et al., 2005, Fig. 1).For the Chongce ice cores, 265 our estimated ages at the ice-bedrock contact (8.3± 3.6 6.2 ka B.P. for the Chongce 216.6 m Core 4 and 9.0 ± 3.6 7.9 ka B.P. for the Chongce 135.8 m Core 2 respectively) are either of Holocene origin, or possible origin of late deglaciation period, similar to the result of the Grigoriev ice core in the west Tien Shan (Takeuchi et al., 2014, Fig. 1).In both cases, the results confirm the upper constraint of 42±4 ka B.P. derived from 270 17 The Cryosphere Discuss., https://doi.org/10.5194/tc-2018-55Manuscript under review for journal The Cryosphere Discussion started: 11 April 2018 c Author(s) 2018.CC BY 4.0 License. the luminescence age of the basal sediment sample collected from the bottom of the Chongce 216.6 m Core 4 (Zhang et al., 2018).

Figure 1 :
Figure 1: Map showing the locations of ice core drilling sites.The numbers for each

Figure 2 :
Figure 2: The β-activity profile of the Chongce 58.8 m Core 3 (a) and the tritium

Figure 3 :
Figure 3: 210 Pb activity profile of the Chongce 216.6 m Core 4 and the derived

Figure 4 :
Figure 4: The depth-age relationship of the Chongce 216.6 m Core 4. The dashed 425

Figure 5 :
Figure 5: The poorly constrained depth-age relationship of the Chongce 135.8 m Core

Figure 6 :
Figure 6: The depth-age relationship of the Chongce 135.8 m Core 2 using additional and Supplement).the model approaches infinity as the depth gets close to bedrock, this bottom age was derived by assuming no further thinning with depth for the last 10 cm w.e..The model 14 The Cryosphere Discuss., https://doi.org/10.5194/tc-2018-55Manuscript under review for journal The Cryosphere Discussion started: 11 April 2018 c Author(s) 2018.CC BY 4.0 License.derived annual accumulation rate is 248±36 mm w.e./year and the model derived age at the depth of the oldest 14 C sample is 4.3± 1.1 1.5 ka B.P., both in good agreement with the accumulation rate of 280-300 mm w.e./year deduced from 210 Pb and tritium (Figs 225 2 and 3) and cal. 14C age of 4.5±0.2ka B.P..This indicates reasonable reliability of the model results.