Direct measurements of deglacial monsoon strength in a Chinese stalagmite

Chinese speleothems (cave deposits) preserve a remarkable paleoclimate record in their oxygen isotope ratios (d18O); the precise interpretation of this record has been the subject of stimulating discussion. Most studies link the d18O variability in Chinese speleothems to regional summer monsoon rainfall and/or rainfall integrated between tropical sources and cave sites. Discussion has centered on mechanisms behind this link as well as the location and seasonality of hypothesized rainfall changes. Until now, these hypotheses were not directly tested in speleothems because conventional drill sampling techniques are insufficient for measuring speleothem d18O at seasonal resolution. Here we use an ion microprobe to analyze seasonal d18O variability in an annually banded stalagmite from Kulishu Cave (northeastern China) that grew during the last deglaciation. The new seasonal resolution data show that the stalagmite d18O values record two aspects of regional monsoon dynamics: (1) changes in the isotopic fractionation of water vapor sourced from both the Indian and Pacific Oceans, and (2) the annual proportion of summer monsoon rainfall, which was systematically greater during the Holocene and Bølling-Allerød than during the Younger Dryas. Both relate to regional rainfall; the isotopic fractionation changes also relate to rainfall integrated from tropical sources. INTRODUCTION Speleothem d18O profiles from many locations in China follow the inverted signal of Northern Hemisphere summer insolation and record the millennial interstadial events observed in Greenland ice cores (Wang et al., 2001, 2008; Yuan et al., 2004; Dykoski et al., 2005; Kelly et al., 2006; Cheng et al., 2009, 2012a, 2012b). For a majority of the records, speleothems are sampled by conventional drilling techniques (0.5 mm spot drilling) whereby individual drill spots average d18O from multiple years of growth. Beginning with one of the earliest of these speleothem records (Wang et al., 2001), the d18O variability is interpreted as a proxy for East Asian Summer Monsoon (EASM) intensity. The interpretation is explained in two ways. In the first explanation (Wang et al., 2001; Cheng et al., 2009; herein called the WangCheng interpretation), speleothem d18O variability reflects a mixing model with two seasonal precipitation components; one is low d18O summer (monsoon) rainfall and the other is high d18O winter and spring rainfall. In this model, the annual fraction of monsoon rainfall dictates the drill sample d18O values. When boreal summer insolation is high, the EASM strengthens, leading to an increase in the annual proportion of low d18O monsoon rainfall, and so the drill sample d18O value is relatively low. The second explanation proposes that the d18O of Chinese precipitation is dictated by Rayleightype fractionation of water vapor originating from Indian Ocean and Pacific Ocean sources (Yuan et al., 2004; herein referred to as the Yuan interpretation). In this model, increased summer insolation leads to an increase of upstream rainfall from air masses destined for Chinese cave sites, which ultimately results in lower d18O(rain) values at the sites. This explanation can also be framed in terms of a mixing model with two seasonal precipitation components, but in this version the d18O of summer monsoon precipitation is the key variable that changes. The two hypotheses have somewhat different climatic implications. In the Wang-Cheng explanation, speleothem d18O values respond directly to changes in the amount of summer monsoon rainfall at the cave site. In the Yuan interpretation, d18O changes relate to changes in the integrated amount of monsoon rainfall from tropical ocean sources to cave sites. By comparison, the model results of Pausata et al. (2011) suggest that d18O changes in Chinese speleothems result solely from changes in the amount of summer rainfall sourced from the Indian Ocean, explicitly ruling out analogous changes over the Pacific. Here we test to what extent each hypothesis explains the d18O record of stalagmite sample BW-1 (14.0–10.4 k.y. before present, A.D. 1950) from Kulishu Cave, China (39.7°N, 115.7°E, 610 m above sea level; Fig. 1), using an ion microprobe to perform 10-mm-diameter, seasonal resolution d18O analyses. SAMPLE SELECTION Sample BW-1 is ideal for four reasons. (1) Earlier workers established a linear age model for BW-1 based on 24 230Th dates along the vertical growth axis (Ma et al., 2012). The sample contains banding that is corroborated as annual by the 230Th dating. (2) Given its average growth rate (76 mm/yr) and the 10 mm resolution at the Wisconsin Secondary Ion Mass Spectrometer Laboratory (WiscSIMS), it is possible to make several d18O analyses in most annual bands. (3) Drill sample measurements of d18O in BW-1 show a large, rapid change across the Younger Dryas (YD)–Holocene transition (at 11.53 k.y. before present; Ma et al., 2012). (4) The cave is near the northern extent of monsoon influence, where the d18O value of EASM rainfall likely correlates with summer rainfall amount (Liu et al., 2014). METHODS The WiscSIMS ion microprobe is capable of high-precision (±0.3‰ 2 standard deviation [s.d.].) and accurate in-situ d18O measurements in 10-mm-diameter spots in speleothem calcite (Kita et al., 2009; Orland et al., 2009). Analysis of BW-1 was completed using a standard-sample–standard bracketing technique. Two groups of ~4 analyses of calcite standard UWC-3 (d18O = 12.49‰, Vienna standard mean ocean water; Kozdon et al., 2009) bracket each set of 10–15 sample analyses. The 2 s.d. of the bracketing standards represents the spot to spot reproducibility of the intervening samples. (See the GSA Data Repository1 for detailed results.) Screening based on pit condition and ionization yield (see the Data Repository) removed 77 of the original d18O analyses (6%) from the plots and discussion presented here. Each sample analysis was positioned based on reflected light and confocal laser fluorescent microscope (CLFM) imaging. Prior work demonstrates that CLFM imaging and ion 1 GSA Data Repository item 2015193, supplementary information including data tables, figures, and supplementary description of methods, is available online at www.geosociety.org/pubs/ft2015.htm, or on request from [email protected] or Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301, USA *E-mail: [email protected] GEOLOGY, June 2015; v. 43; no. 6; p. 555–558; Data Repository item 2015193 | doi:10.1130/G36612.1 | Published online 7 May 2015 © 2015 eological Society of A erica. For permission to copy, contact [email protected]. on May 27, 2015 geology.gsapubs.org Downloaded from