Conflicting reconstructions of Holocene variability of the East Asian summer monsoon (EASM) from speleothem versus other types of proxy records have yielded widely divergent estimates of its phase relationship with the Indian summer monsoon (ISM) and Northern Hemisphere summer insolation (NHSI). This apparent discrepancy has been partly attributed to the uncertainties in the climatic representation of Chinese speleothem oxygen isotope (δ18O) records. Here we present a composite speleothem δ18O record of the last ∼14 kyr from Shennong Cave in southeastern China and model-simulated data of rainfall and meteoric δ18O over eastern China. Our synthesis of the proxy-model data suggests that the spatial patterns in both speleothem δ18O and paleo-rainfall over eastern China during the Holocene are diverse at orbital and multi-millennial scales. Our findings imply that: 1) speleothem δ18O in the EASM regime is largely controlled by the large-scale circulation and concomitant latitudinal shifts of the monsoon rain belt; notwithstanding the heterogeneous spatiotemporal pattern of Holocene rainfall as inferred from various proxy records, a coherent orbital-scale speleothem δ18O variability across most Asian monsoon regions (except southeastern China) indeed stems from the NHSI-forced changes in overall monsoon intensity; overall monsoon intensity is not equivalent to monsoon rainfall amount but a manifestation of the large-scale atmospheric circulation; 2) divergent phase relationships with NHSI between speleothem δ18O and other proxy records are consistent with—rather than contradictory to—the NHSI forcing mechanism. Speleothem δ18O and rainfall records reflect two different aspects of the monsoon dynamics. These results may thus, largely help to reconcile the divergent views of the Holocene Asian monsoon variability.
Bibliographical noteFunding Information:
To compare the spatiotemporal patterns of the Holocene hydroclimate records in eastern China with CAM5 modelling results (Kong et al., 2017), we divided eastern China into 6 zones (Fig. 1 and Supplementary Fig. S2). Normalized speleothem δ18O records (raw data were presented in Fig. 4, method shown in supplementary information) were compared with modelled convective/annual ratios in zones 1–6 from CAM5 (Figs. 5A–F). Pollen-inferred annual rainfall records in Fig. 4 were used for comparison with modelled annual rainfall anomalies in zones 1–6 (Figs. 5A1-F1). These comparisons reveal a strong consistency between normalized speleothem δ18O and simulated convective/annual ratio from CAM5 (Figs. 5A–F), as well as between reconstructed and simulated annual rainfall amounts from CAM5 (Figs. 5A1-F1) in each zone over the last 9 kyrs. Speleothem δ18O values (convective/annual ratios) from zones 1–4 show a coherent long-term increasing (decreasing) trend from the Early to Late Holocene, in contrast to those in zones 5 and 6 (Figs. 5A–F). With respect to spatial distributions of speleothem δ18O or convective/annual ratio in the Late Holocene, the pattern indicates lower speleothem δ18O values (higher convective/annual ratios) in zones 1–4, but similar or higher speleothem δ18O values (or lower convective/annual ratios) in zones 5–6 for the Early Holocene. This is supported by the simulated anomalies (between 9 kyr, 6 kyr and PI) of convective/annual ratio and δ18Ow data from COSMOS-wiso that suggest a positive anomaly of convective/annual ratio in northern and central China (zones 1–4) but a similar value or a negative anomaly in southeastern China (zones 5–6) (Figs. 6G–I), and a coherent negative anomaly of δ18Ow over the eastern China, except a localized positive δ18Ow anomaly in southeastern China (zones 5–6) (Figs. 6A–C). This muted δ18Ow response in the southeastern China can be also observed in other isotope-enabled simulations (Battisti et al., 2014; Liu et al., 2014b). The consistency between reconstructed and simulated annual rainfall amount over the Holocene is not as strong as that between speleothem δ18O and convective/annual ratio (Figs. 5A1-F1). Together with CAM5 and TRACE simulation results, overall, we can find that zones 1–2 were characterized by a wet climate in the Early-Middle Holocene and a dry climate in the Late Holocene; however, decreased annual rainfall in the Early Holocene and increased annual rainfall in the Middle-Late Holocene characterized zones 3–6 (Figs. 5A1-F1). This is further supported by our simulated annual rainfall result from COSMOS-wiso, showing a positive annual rainfall anomaly over northern China but a negative anomaly in the regions from central to southern China between 9 kyr, 6 kyr and PI (Figs. 6D–F), consistent with our Shennong δ13C record from southeastern China (Fig. 4).Based on a mid-Holocene (∼6 kyr) rainfall maximum documented by paleo-rainfall records in northern and northwestern China, a number of studies (Chen et al., 2016; Li et al., 2020b; Liu et al., 2015, 2020) argued that the EASM does not directly respond to NHSI and the lagged EASM response to NHSI is modulated by high-latitude forcing (e.g., continental ice sheet and meltwater flux into the North Atlantic). Here, we find that the paleo-rainfall records in zone 2 indeed reflect a mid-Holocene rainfall maxima (Figs. 4 and 5B1); however, many records in zone 1 show an early-Holocene (∼9 kyr) rainfall maxima (Figs. 4 and 5A1) (Goldsmith et al., 2017; Ming et al., 2020). In addition, paleo-rainfall records from loess-paleosol/lake sediments and speleothem δ18O records in the monsoonal fringe of northwestern China (Fig. 8) also show an early-Holocene rainfall maxima, similar to the Holocene pattern of northern China in zone 1. Therefore, this timing offset cannot be fully attributed to chronology uncertainties or sampling resolution of proxy records. In contrast, these observations are further supported by our simulation results that show a wetter climate in the monsoonal fringes of northern (zone 1) and northwestern China during the Early Holocene (9 kyr BP) relative to the Middle Holocene (6 kyr BP) (a positive anomaly between 9 kyr and 6 kyr in Figs. 6F and 7F). An early-Holocene rainfall maxima can be also observed in other Holocene transient climate simulations (Liu et al., 2014b; Zhang et al., 2021). Some records from the arid/semi-arid region in zone 1 showing a mid-Holocene rainfall maxima, however, may reflect a mid-Holocene maxima of effective moisture (precipitation minus evaporation) presumably because of higher NHSI and thus stronger summer evaporation condition in the Early Holocene. This is supported by a recent study involved a series of experiments including transient climate and lake energy/water balance models (Li et al., 2020d).It has been suggested that changes in U1429 δ18O are controlled by local rainfall and freshwater discharge from the Yangtze River (Clemens et al., 2018), changes in NGHP17 δ18O were also considered to be controlled by the local rainfall amount (Gebregiorgis et al., 2018). Our model results indeed show small precession changes in rainfalls at the U1429 site as well as in the middle-lower reaches of the Yangtze River (Figs. 6D–F and 7D-F). The core NGHP17 is located in a region where reduced (increased) precipitation occurs when NHSI is high (low), approximately opposite to that in the ISM fringe (Figs. 6D–F and 7D-F), which can be further supported by the transient simulation (Jalihal et al., 2020) and extreme experiments (Bosmans et al., 2018) from fully coupled model simulations. Therefore, the ∼9 kyr lag of the NGHP17 ISM rainfall record to NHSI is a manifestation of the unique local rainfall feature at the precessional scale. In contrast, the seawater δ18O record obtained from marine core SO 17286–1 in the northern Bay of Bengal (Lauterbach et al., 2020) ( Figs. 1, 6D-F and 7D-F), where lower δ18O values are related to more freshwater discharged from the Ganges-Brahmaputra-Meghna river system associated with stronger ISM rainfall, is consistent with ISM speleothem δ18O records from Qunf (Fig. 9 A) and Bittoo (Kathayat et al., 2016) caves and other marine proxy records from the northern Bay of Bengal (Kudrass et al., 2001). It is also consistent with the speleothem δ18O records, as well as lake carbonate δ18O and paleo-rainfall reconstructions from southwestern China and the Tibetan Plateau (Figs. 8 and 9C), where the ISM prevails (Cai et al., 2015; Zhang et al., 2011). It can be also supported by a coherent negative anomaly (9 kyr minus PI) in δ18Ow over the overwhelming majority of the continental Asian monsoon region from our modelling results (Figs. 6A–C). These observations and simulations reinforce the notion that the ISM intensity is driven by the NHSI on a subcontinental scale and in line with the EASM (Figs. 9C–D) (Cai et al., 2015; Kathayat et al., 2016; Zhang et al., 2011), characterized by a declining trend after the mid-Holocene associated with a gradual southward shift of the Intertropical Convergence Zone (ITCZ) (Fig. 9B), consistent with a large set of modelling works (Battisti et al., 2014; Bosmans et al., 2018; Kutzbach, 1981; Lee et al., 2019; Liu et al., 2014b).We would also like to thank Editor Miryam Bar-Mattews and three anonymous reviewers for their comments and suggestions. Thanks to Dr. Wenwen Kong and Dr. Jun Hu for their suggestions of interpretting. We would like to thank Editor Miryam Bar-Mattews and three anonymous reviewers for their comments and constructive suggestions. We also thank Dr. Wenwen Kong for her help of using snapshot experiments data and Dr. Jun Hu for his suggestions of the interpretation of convective precipitation. This study was supported by grants from the National Science Foundation of China (41888101 and 41972186), the National Key Research and Development Program of China (2017YFA0603401) and the China Postdoctoral Science Foundation (2019T120894).
We would also like to thank Editor Miryam Bar-Mattews and three anonymous reviewers for their comments and suggestions. Thanks to Dr. Wenwen Kong and Dr. Jun Hu for their suggestions of interpretting. We would like to thank Editor Miryam Bar-Mattews and three anonymous reviewers for their comments and constructive suggestions. We also thank Dr. Wenwen Kong for her help of using snapshot experiments data and Dr. Jun Hu for his suggestions of the interpretation of convective precipitation. This study was supported by grants from the National Science Foundation of China ( 41888101 and 41972186 ), the National Key Research and Development Program of China ( 2017YFA0603401 ) and the China Postdoctoral Science Foundation ( 2019T120894 ).
© 2021 Elsevier Ltd
- East Asian summer monsoon (EASM)
- Northern hemisphere summer insolation (NHSI)
- Spatiotemporal pattern