JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 110, B12101, doi:10.1029/2005JB003765, 2005 Spatial variations of magnetic susceptibility of Chinese loess for the last 600 kyr: Implications for monsoon evolution Qingzhen Hao and Zhengtang Guo1 Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, China Received 5 April 2005; revised 2 September 2005; accepted 27 September 2005; published 2 December 2005. [1] We examine spatial variations in magnetic susceptibility (MS) over the Loess Plateau in China based on 50 sections in order to identify spatial changes in monsoon climate at key glacial, interglacial, and interstadial intervals for the last 600 kyr. The results indicate strong coherence between MS variations during the interglacial periods and present-day precipitation and temperature patterns. This suggests that the strength of the summer monsoon had a dominant influence on the MS signals in soils, through modulating pedogenic intensity. The distribution of MS during glacial periods is characterized by weak S-N gradients and rough W-E zonal patterns, indicating a negligible effect of the summer monsoon. Interstadial patterns are intermediate between the glacial and interglacial ones. Interglacial patterns for the last 600 kyr are essentially similar, suggesting that the climate regime during these periods has not undergone significant changes and that the east Asian summer monsoon has remained the main moisture carrier. Our estimates of the relative amplitudes of climate oscillations during these time slices are consistent with earlier paleoclimate studies. These data, associated with the available susceptibility-based climofunctions, may be used to estimate spatial changes of paleorainfall and paleotemperature for these key periods and hence to test climate models. Citation: Hao, Q., and Z. Guo (2005), Spatial variations of magnetic susceptibility of Chinese loess for the last 600 kyr: Implications for monsoon evolution, J. Geophys. Res., 110, B12101, doi:10.1029/2005JB003765. 1. Introduction key time intervals, such as marine oxygen isotope stage (MIS) 11 that is usually regarded as an astronomical [2] The spatially correlative Quaternary loess-soil analogue of the Holocene [Loutre and Berger, 2003], and sequences in the Chinese Loess Plateau (Figure 1) have MIS 13 representing a warm, humid climate extreme when long been regarded as near-continuous records of the east compared with the record over the last 2.6 Myr [Guo et al., Asian monsoon climate [Liu, 1985; Kukla and An, 1989; An 1998]. Examination of the spatial climate changes for a et al., 1990a], and the oldest loess-soil sequences have been longer time span would significantly improve our under- dated back to 22 Ma [Guo et al., 2002]. The alternations standing of the east Asian monsoon climate, and would also between soils and loess are commonly interpreted as an provide useful boundary conditions for climate models. indication of the waxing and waning of the east Asian [4] Magnetic susceptibility (MS) in Chinese loess is monsoon circulation, with the soil-forming periods widely used as a climate proxy [An et al., 1990a, 1991] corresponding to a strengthened summer monsoon and loess and is also a stratigraphic marker [Kukla and An, 1989]. Its deposits to a strengthened winter monsoon [An et al., value is higher in soils than in the intervening loess layers 1990a]. [Liu, 1985; An et al., 1990a]. Fine-grained minerals (mag- [3] The temporal changes of the monsoon climate in the netite and maghemite) of pedogenic origin are thought to be Loess Plateau have been reconstructed using various prox- responsible for the higher values in soils [e.g., Zhou et al., ies [e.g., An et al., 1990a; Guo et al., 1991, 1998, 2000a; 1990; Maher and Thompson, 1991]. This interpretation is Heller et al., 1993; Maher et al., 1994; Lu¨etal., 1994, also supported by most recent rock magnetic studies [Maher 1999; Wu et al., 2001; Ding et al., 2002]. However, only a et al., 2003; Deng et al., 2004]. Analyses of various mineral few studies, mainly centered on the last two or younger magnetic parameters, such as temperature-dependent sus- climate cycles, have been carried out to examine spatial ceptibility and hysteresis properties [Deng et al., 2004], changes in the climate patterns [Sun et al., 1996; Nugteren indicate that the magnetic properties of the paleosols youn- and Vandenberghe, 2004], and the spatial patterns for older ger than 600 ka are similar. The temporal changes of MS in time slices remain to be addressed. These include several the sequences of the last 600 kyr are similar to other proxies of pedogenic intensity, such as the chemical weathering 1 Also at State Key Laboratory of Loess and Quaternary Geology, index (FeD/FeT) [Guo et al., 2000a], Rb/Sr ratio [Chen et Institute of Earth Environment, Chinese Academy of Sciences, Xian, China. al., 1999] and soil micromorphology [Guo et al., 1998]. Copyright 2005 by the American Geophysical Union. Because pedogenic intensity in the Loess Plateau is mainly 0148-0227/05/2005JB003765$09.00 controlled by the strength of the east Asian summer mon- B12101 1of10 B12101 HAO AND GUO: SPATIAL CHANGES OF CHINESE LOESS B12101 Asian summer and winter monsoons [L. X. Chen et al., 1991]. In winter, the cold dry northwest winter monsoon of subarctic origin prevails across the region. In summer, the region is under the control of the warm humid southeast summer monsoon of tropical/subtropical origin, leading to abundant precipitation. In spring, the summer monsoon begins to penetrate progressively further inland, and its northern front constitutes an important rain belt. From early June to early July of each year, this belt usually reaches and is contained in the Yangtze–Huaihe River basin between 26–34° N in eastern China, resulting in a long and continuous rainy season (the so-called ‘‘Meiyu belt’’ or ‘‘plum rain belt’’). Later, as the monsoon strengthens, the front penetrates into northern China, including the Loess Plateau, where most of the annual rainfall is concentrated in July, August and September [L. X. Chen et al., 1991]. At present, there is a strong climate gradient across the plateau (Figure 1). [7] In Quaternary loess-soil sequences in China, major interglacial soils and interbedded loess units of the last 600 kyr are labeled as S0, L1, S1, L2, S2, ..., and S5 from the top to the bottom according to Liu [1985]. Some soil units are indeed polycyclic, and can be subdivided into subunits [Guo et al., 1996]. For example, S2 consists of two subsoils, named S2SS1 and S2SS2, and S5 consists of three subunits (S5SS1, S5SS2 and S5SS3). The Holocene soil S0 differs from the older soils by its dark color, weak rubifi- cation, and the absence of clear calcitic horizons [Guo et al., 1991, 1994]. In contrast, the older soils are generally characterized by a brownish or reddish color with an angular blocky structure. Loess units usually have a light brown or yellow brown color with a massive structure. The Figure 1. Chinese Loess Plateau, location of the studied boundaries between loess and soil units are well reflected by sites, (top) mean annual temperature (°C, dashed line) and MS (Figure 2). The stratigraphy is correlative across the whole Loess Plateau and can be readily correlated with the (bottom) mean annual precipitation (mm, dashed line) for 18 the interval 1951–1990. Solid (half shaded) circles marine d O record [Kukla, 1987], as was also confirmed by represent the newly measured (published) sections. The the eolian dust record from the North Pacific [Hovan et al., codes of sections and cities are as follows: BX, Binxian; 1989]. The loess-soil sequences above the L6 unit are CW, Changwu; DB, Dingbian; DX, Dingxi; HN, Huning; correlative with marine oxygen isotope stages (MIS) 1 through 15 [Kukla, 1987], and the correlation scheme with HX, Huanxian; JB, Jingbian; LC, Luochuan; LF, Linfeng; LT, 18 Lantian; LZ, Lanzhou; JX, Jixian; SMX, Sanmenxia; the SPECMAP d O record [Imbrie et al., 1984] is shown in XA, Xi’an; XN, Xining; XF, Xifeng; YC, Yinchuan; Figure 2. YCH, Yichuan; YL, Yulin; YS, Yongshou. [8] Here we use MS data from 50 sections (Figure 1) for the study of the spatial distribution of MS during selected time slices; 29 of the sections were newly measured for this soon, the temporal and spatial changes of MS should study, and data for the other 21 sections were taken from primarily reflect the effects of the summer monsoon [An published literature (Table 1). Among these, 15 sections et al., 1990a, 1991]. provide MS records younger than L3. These sites cover a [5] On the basis of the MS data from 50 sections large area with the annual mean precipitation (MAP) spanning the last 600 kyr, this study aims (1) to address varying from 300 to 650 mm and the annual mean the spatial characteristics of MS variations in the Loess temperature (MAT) from 6 to 13°C (Figure 1). Among Plateau region and their relationship with present-day cli- the 29 newly measured sections, seven (Xining, Xifeng, mate conditions; (2) to examine the spatial changes in the Changwu, Weinan, Yichuan, Dali and Linfen) were contin- monsoon climate during glacial, interglacial and interstadial uously sampled at 10 cm intervals. For the other 22 sections, times; and (3) to estimate the approximate amplitude of we first measured the MS in the field at 10 cm intervals displacement of the climate belts during different time using a portable susceptibility meter, in order to locate the slices. positions with maximum (for soils) and minimum MS (for loess units), and then took three parallel samples at these positions for measurements of mass MS in the laboratory. 2. General Setting and Methods The Holocene soils at Dali, Puxian, Weizhuang, Xixian and [6] The climate in the Loess Plateau is mainly controlled Yichuan were not successfully sampled due to strong by two seasonally alternating monsoon circulations: the east disturbance by agricultural activity.