**2. Materials and analytical methods**

#### **2.1 Core HDP-06**

138 International Perspectives on Global Environmental Change

of sediments from the two great Asian lakes (Hovsgol for Ti/Ca and Br; Baikal for diatoms and Br) with a synchrotron radiation (SR) induced μ-XRF line-scanning technique, and then found a 300-500 kyr long cycle, possibly associated with Earth's orbital eccentricity. Moreover, Phedorin et al. (2008) suggested that the variations with periods of 300-500 kyr

Fig. 1. Maps of (a) Lake Hovsgol in (b) northern Mongolia. (a) Bathymetric map of the lake showing drill sites for cores X106 and HDP-06. HDP-06 was used for analyses in the present study. The gray shading and lines in (b) indicate the Lake Baikal catchment area and the

In the present study, we investigated the distributions and their origins of 11 major and trace elements in the Holocene section of Hovsgol's sediment using the SR μ-XRF mappings. With respect to the Hovsgol's sediment, the two-dimensional distribution of the elements has not been studied until now. A visualization of distribution of the elements enables us to recognize detailed sedimentary or geochemical structures such as laminations, local elementally enriched grains and layers, and so on (Katsuta et al., 2007). Moreover, it makes possible to distinguish between primary structures and secondary structures deposited in the sediment. As a result of this study, we discovered an alternation of terrigenous and biological elements in the Hovsgol Holocene sections. In this chapter, we first describe methods for capturing the SR μ-XRF data of the sediment surfaces. After investigating distributions and features of the detected 11 elements, we assess the sources and natures of each element. We discuss the paleoenvironmental implications for the observed alternated pattern in the sedimentation, together with a record of the detritus input into Lake Hovsgol

influenced the evolution of terrestrial ecosystem.

borders of China, Mongolia, and Russia, respectively.

(Murakami et al. 2010).

In the present study, we used core HDP-06 which was collected at the base of the southeast slope of Lake Hovsgol, in northwest Mongolia (Fig. 1). The HDP-06 core drilling took place in March 2006 at 50°54'25" N, 100°27'03" E at a water depth of 235.5 m using an improvised push-coring type technology (Hovsgol Drilling Project Members, 2009). Core HDP-06 was about 26 m long and the sediment recovery was on the order of 80%. In the present study, we analyzed the upper section (about 18 cm) of HDP-06 core 1-1 (about 140 cm), which was undisturbed and fully recovered. The lithologic observation revealed that the core 1-1 bears sedimentary layers distinctive of Hovsgol's sediment, which is composed of diatomaceous clayey ooze in the Holocene section and calcareous clayey silt to silty clay in the last glacial section (Prokopenko et al., 2005).

#### **2.2 Sample preparations**

In order to acquire two-dimensional XRF images of the sediment surface by the SR μ-XRF techniques, we prepared thin sections of the resin-embedded samples. Subsamplings from an open core surface were carried out using aluminum U-channels (Fig. 2). After carefully retrieving the sediment slab from the U-channel, the slabs were impregnated with epoxy resin following the procedure of Tiljander et al. (2002). The U-channels were produced from a 0.2 mm-thick aluminum sheet using a sheet bender (Fig. 2a), which we produced ourselves.

A resin-embedded sediment block was shaped to a thickness of 0.3 mm for SR μ-XRF mapping. This was done to reduce the X-ray scattering from the inside of the sample. First, a pair of sediment slabs (Fig. 3a) was separated with a band saw because several sediment slabs were embedded together with the epoxy resin. Each sediment slab was cut into three pieces to produce approximately 5 cm × 5 cm samples at an angle of 45° to the bedding plane (Fig. 3b), in order to maintain the vertical continuity of the sediment. One face of each sample was polished and glued to a glass slide. After polishing the other face to achieve a thickness of about 0.3 mm, the thin section sample (Figs. 3c and d) was removed from the glass slide using a solvent.

#### **2.3 SR μ-XRF mapping**

The SR experiment was carried out at undulator beamline 37XU of SPring-8, Hyogo, Japan (Fig. 4a; Terada et al., 2004). A Si (111) double-crystal monochromator was used to acquire the incident X-ray beams at 37 keV. The X-ray beam size was adjusted using an XY slit and was 1 mm (V) × 0.5 mm (H). A thin-section sample was fixed on an acrylic holder (Fig. 3c). The sample holder was set on the X-Y computer-controlled step stage, which was rotated by 10º in the rectangular direction of the incident beam toward the detector (Fig. 3d). The stage had step sizes of 1 mm (V) and 0.5 mm (H). The XRF spectra from the sample were measured using a Ge solid-state detector coupled with a multichannel pulse-height analyzer (Fig. 4a). The measurement time was 25 s per step.

In Hovsgol's sediment, fluorescence X-rays of the 11 elements (Ti, Mn, Fe, Cu, Zn, As, Br, Rb, Zr, and Nb) were successfully detected (Figs. 4b and c). Based on the obtained XRF spectra, we determined the distribution map and profile of each element (Fig. 5). The XRF spectra were collected at each position on a sample. At each position, we computed the integrated numbers of X-ray photons with energy near each *Kα1* line (energy window within *Kα1* ± 0.10 keV), and consequently produced XRF maps of the 11 elements. Moreover, the XRF profiles were acquired by averaging three vertical lines of the XRF maps, which were appropriately selected.

Continental Erosion/Weathering Changes in Central Asia Recorded in the Holocene

Sediment from Lake Hovsgol, Northwest Mongolia, by Synchrotron μ-XRF Mapping Analyses 141

Fig. 3. Schematic diagrams of resin-embedded sediment sample and its experimental setup.

(d) is indicated by red rectangle in Fig. 4a.

Fig. 2. Tools for subsampling open core surface. (a–d) Photograph and schematic sketches showing a sheet bender made of aluminum. (e) Schematic sketch showing an aluminum Uchannel. An aluminum sheet (0.2 mm thick) was first cut into 120 mm × 36 mm rectangular pieces with a shearing machine, which in turn was set on a male die A with rods E (b). After pressing a female punch B and plank with rods D downward (c and d), the U-channel (e) was produced. Finally, the sample number and direction were inscribed onto the back of the U-channel using a scriber. The produced aluminum U-channels are pressed into halved cores, with adjacent U-channels overlapping each other as shown in (f). The sediment-filled U-channels were carefully extruded from the split cores using a metal spatula. The hole positions of an aluminum sheet (b) coincide with the positions of holes in B-D and rods E, and the holes have slightly large diameter compared with the rods. The insertion of the three rods in D and E into the corresponding holes in aluminum sheet (b) prevented the movement of the sheet during bending and simultaneously allowed the sheet to be bent at a right angle. Rods E were screwed into a male punch A like D, thereby allowing easy removal of the produced U-channel from the punch. The holes in U-channel (e) aided fluid flow in the later impregnation stage.

Fig. 2. Tools for subsampling open core surface. (a–d) Photograph and schematic sketches showing a sheet bender made of aluminum. (e) Schematic sketch showing an aluminum Uchannel. An aluminum sheet (0.2 mm thick) was first cut into 120 mm × 36 mm rectangular pieces with a shearing machine, which in turn was set on a male die A with rods E (b). After pressing a female punch B and plank with rods D downward (c and d), the U-channel (e) was produced. Finally, the sample number and direction were inscribed onto the back of the U-channel using a scriber. The produced aluminum U-channels are pressed into halved cores, with adjacent U-channels overlapping each other as shown in (f). The sediment-filled U-channels were carefully extruded from the split cores using a metal spatula. The hole positions of an aluminum sheet (b) coincide with the positions of holes in B-D and rods E, and the holes have slightly large diameter compared with the rods. The insertion of the three rods in D and E into the corresponding holes in aluminum sheet (b) prevented the movement of the sheet during bending and simultaneously allowed the sheet to be bent at a

right angle. Rods E were screwed into a male punch A like D, thereby allowing easy removal of the produced U-channel from the punch. The holes in U-channel (e) aided fluid

flow in the later impregnation stage.

Fig. 3. Schematic diagrams of resin-embedded sediment sample and its experimental setup. (d) is indicated by red rectangle in Fig. 4a.

Continental Erosion/Weathering Changes in Central Asia Recorded in the Holocene

**3.1 Distribution of elements in Lake Hovsgol sediment** 

**3. Results and discussions** 

(Figs. 5a and c–k).

thin layers and spots (Fig. 5b).

follows:

and f).

presence of carbonates.

**period** 

Sediment from Lake Hovsgol, Northwest Mongolia, by Synchrotron μ-XRF Mapping Analyses 143

Successive maps and profiles of the 11 elements in the upper section of HDP-06 core 1-1 are shown in Fig. 5. From a visual inspection of the distribution and features, the 11 elements were classified into three assemblages–group 1: Ti, Fe, Cu, Zn, As, Rb, Sr, Zr, and Nb; group 2: Br; and group 3: Mn, Fe, and As. Their distributions in the sediment are characterized as

1. Groups 1 and 2 consist of several centimeter-scale layers that alternate with each other

2. The manganese of group 3 is irregularly distributed over the entire section, occurring in

3. A portion of the arsenic from group 3 occurs in thin layers, together with Fe (Figs. 5c

Based on the distributions and assemblages of the elements, group 1 is recognized to be elements composed of rock-forming minerals in the bedrock of the Hovsgol basin (Murakami et al., 2010). The group 1 is therefore terrigenous elements, which were supplied from the drainage basin by erosion and weathering processes. Bromine belonging to group 2 is a biophilic element whose abundance variation in the sediment reflects bioproduction in the lake (Phedorin et al., 2008). In Lake Hovsgol, the record of Br as alternative to diatom and bioSi, especially in the bottom sediment below 5.74 m (Fedotov et al., 2004b), is an important source of information to estimate biogenic production of the lake because the diatom frustules may be dissolved in the sediment high-pH pore water resulting from the

Manganese, Fe, and As of group 3 are elements sensitive to redox condition in the sediment. Existences of these three elements were indentified also in core X106 (Murakami et al. 2010). According to the study, the Mn in the core showed irregular distributions from the last glacial/Holocene transition to the Holocene section. The XRD analyses didn't indicate peaks of minerals containing Mn, thereby suggesting that the Mn in the sediment exists in an amorphous state. On the other hand, the coexistence of Fe and As identified in this study (red arrows in Figs. 5c and f) were identified in the last glacial/Holocene transition section of core X106. Because the section contains pyrite and dolomite as well as arsenic, Murakami

These three elements of group 3 may have responded to redox changes and have subsequently migrated during the diagenetic process in the sediment. We therefore investigate paleoenvironmental implications for variations in the group 1 terrigenous

Comparison of the terrigenous elements of group 1 (Ti, Fe, Cu, Zn, As, Rb, Sr, Zr, and Nb) with the Br of group 2 in the HDP-06 core 1-1 indicates a strong counterphase in the Holocene section (Fig. 5). At the present stage, the age of the section has not yet been determined. However, as shown in Figs. 6b and c, the content variation of Ti in the HDP-06 sediment is well correlative with that of dated X106 core (Murakami et al., 2010). Accordingly, by a correlation between these two profiles, we discuss hereafter nature of the temporal content

**3.2 Continental erosion/weathering changes in central Asia during the Holocene** 

variations of Ti and Br in HDP-06 core, together with paleoproxy records of X106 core.

et al. (2010) suggested the presence of sulfate-reducing bacteria.

elements and group 2 Br records of core HDP-06, as discussed below.

Fig. 4. (a) Schematic sketch showing experimental setup for SR μ-XRF mapping at BL37XU of SPring-8 and XRF spectra in ranges of (b) 0–40 keV and (c) 0–20 keV for upper section of HDP-06 core 1-1 from Lake Hovsgol. Sample of red rectangle in (a) corresponds to Fig. 3d.
