Holocene Indian Summer Monsoon variations inferred from end-member modeling of sediment grain size in the Andaman Sea

Abstract

Whether the Indian Summer Monsoon (ISM) Maximum started in the early Holocene or middle Holocene remains a controversial issue. Grain-size measurements were conducted on 157-cm-long core ADM-C1 from the Andaman Sea; the core spanned the last 11.2 kyr BP. Four end-members (EMs) were unmixed using lognormal parametric end-member modeling. Genetic analyses of the grain-size EMs suggested the EM1 and EM2 may correlate with suspension in the upper layers and transportation in the benthic nepheloid layer, respectively, while the EM3 and EM4 were tied with deposition affected by monsoonal currents and sedimentation under extreme events, respectively. The EM3 was dominated by ISM intensity and transported by summer monsoon currents, with increase in the proportion of EM3 reflecting increasing ISM intensity, and vice versa. Variations in EM3 were used to define three stages of ISM evolution, as follows: (1) During 11.2–9.1 kyr BP, EM3 was relatively low, indicating a weak ISM; (2) during 9.1–4.5 kyr BP, EM3 increased substantially, indicating a strengthened ISM; (3) during 4.5-0 kyr BP, EM3 decreased gradually, suggesting a gradual weakening of the ISM. The Holocene ISM Maximum started at ~9.1 kyr BP rather than in the early Holocene, which may have been partly due to the slowdown of the Atlantic meridional overturning circulation (AMOC) during the early Holocene that resulted in a decreased land-sea thermal contrast between the landmass and the Indian Ocean. An additional cause may have been the remnant Northern Hemisphere ice sheets that impeded the northward shift of the Intertropical Convergence Zone (ITCZ). Seven relatively brief decreases in the ISM intensity (events 7, 6, 5, 4, 3, 2, 1) occurred at ca. 10.3, 9.7, 7.3, 5.7, 4.2, 2.4, and 0.8 kyr BP, respectively. They corresponded, within the age uncertainties, to the increased supplies of ice-rafted detritus to North Atlantic sediments, implying that changes in the ISM intensity on millennial scale were dominated by climatic processes in northern high latitudes.

Introduction

As a major component of the global monsoon system, the Indian Summer Monsoon (ISM) affects more than one billion people across the Indian subcontinent and Southeast Asia, and has a major influence on the rise and decline of agricultural civilizations in South Asia via its effect on rainfall (An et al., 2011; Hillman et al., 2017; Rawat et al., 2015; Wang, 2009). Accordingly, determining the timing, duration, magnitude and forcing mechanisms of past variations in the ISM is critical to future environmental prediction. However, the initial time of the Holocene ISM Maximum are debated. Some had argued that the time of Holocene ISM Maximum initiated in the early Holocene, during 11-10 kyr BP (Fleitmann et al., 2003; Hong et al., 2003; Rashid et al., 2007; Wu et al., 2018; Zhang et al., 2017a). The reconstructed ISM from the δ¹⁸O in carbonate of Lake Xingyun (Wu et al., 2018) along with <16 μm particles percentage and Ti content in the Lake Tengchongqinghai (Zhang et al., 2017a) suggested that the initial time of Holocene ISM Maximum started at ~11 kyr BP. In addition, the low seawater salinity (δ¹⁸O_sw) in the Andaman Sea during 10.8–5.5 kyr BP (Rashid et al., 2007), more negative δ¹⁸O in stalagmite during 10.3–9.6 kyr BP (Fleitmann et al., 2003) and low values of δ¹³C from the cellulose of Carex mulieensis in Hongyuan Peatland between 10.8 and 5.5 kyr BP (Hong et al., 2003) all indicated the Holocene ISM Maximum started at ca. 11-10 kyr BP. However, others considered the initial time of the Holocene ISM Maximum occurred in the middle Holocene, during 9-8 kyr BP (Prins and Postma, 2000; Schulz et al., 1998; Cao et al., 2015a). The low percentage of eolian dust from core Makran-469 indicated an intensified ISM during ca. 9-8 kyr BP (Prins and Postma et al., 2000). Besides, both the high biological productivity based on total organic carbon (TOC) in core SO90-111 KL from the northeastern Arabian Sea during 9.1–4.4 kyr BP (Schulz et al., 1998) and the high sensitive grain-size percentage during 9-7 ka BP (Cao et al., 2015a) suggested the Holocene ISM Maximum started at ca. 9-8 kyr BP. The origin of this discrepancy remains unclear.

Due to the existence of an island arc chain, the Andaman Sea has great potential for preserving high quality, undisturbed sedimentary records and therefore the region is well suited for sediment-based paleoenvironmental reconstruction. Sediment grain-size distributions contain abundant information on paleoenvironmental variations and have been widely applied to paleoclimatic reconstruction in various geological settings (Ashley, 1978; Porter and An, 1995; Weltje, 1997; Xiao et al., 1995). Numerous numerical unmixing methods have been proposed to infer paleoenvironmental conditions from sediment grain-size distributions (Darby et al., 2009; Dietze et al., 2012; Paterson and Heslop, 2015; Qin et al., 2005; Sun et al., 2002; Weltje, 1997; Yu et al., 2015; Zhang et al., 2017b). In this study, we present a well-dated Holocene paleoclimatic record based on grain size end-member analysis from a marine sediment core obtained from the Andaman Sea. Our main aims are:

• to characterize past ISM variability;

• to explore the possible forcing mechanisms of ISM evolution during the Holocene and

• to probe into the possible causes for the discrepancy of the onset of Holocene ISM Maximum.