NGR is a useful lithologic parameter because the primeval emitters are at secular equilibrium (i.e., radiation at characteristic energies is constant with time) (Adams and Gaspirini, 1970; Blum, 1997). Radioisotopes with sufficiently long life and that decay to produce an appreciable amount of gamma rays are potassium (40K) with a half-life of 1.3 x 109 yr, thorium (232Th) with a half-life of 1.4 x 1010 yr, and uranium (238U) with a half-life of 4.4 x 109 yr. The total NGR intensity is a function of the combined contributions of K, U, and Th in sediments, matrix density, and matrix lithology. Matrix density mainly results from Compton scattering, and matrix lithology results from photoelectric absorption (Blum, 1997). Clay mineral content is often diluted by other components such as biogenic silica. Because of this dilution, if the NGR signal in the sedimentary sequence is to be used to reconstruct the environmental record, it is important to know which chemical components really relate to the NGR signal.
The NGR intensity of sediments from Cores 186-1150B-22R, 28R, 29R, and 32R and 186-1151A-91R and 105R are plotted with analyzed chemical results in Figures F4, F5, F6, F7, and F8. NGR intensity shows cyclic behavior in these intervals. Cyclic variation corresponding to that of NGR intensity was observed in a series of properties in recovered sediments. Major elements in bulk samples were Na2O, MgO, Al2O3, SiO2, S, K2O, CaO, TiO2, MnO, and Fe2O3 (Table T1). SiO2 is the most abundant element off Sanriku (76.3 wt% on average of all samples) and is more abundant at northern Site 1150 (79.0 wt%) than at Site 1151 (72.2 wt%). In contrast, Al2O3, which is the second most abundant element, is 9.5 wt% on average at Site 1151, compared to 7.4 wt% at Site 1150. MgO, K2O, TiO2, and Fe2O3 have positive correlations with Al2O3 content, which is slightly richer at Site 1151 than at Site 1150, which included siliciclastic minerals. NGR correlates positively with Al2O3 and K2O at Site 1150 (Fig. F9A, F9B). Although this relationship is not clear at Site 1151, NGR intensity and Al2O3 and K2O contents are higher than at Site 1150. The amount of terrigenous minerals including K, Al, and related elements such as clay minerals are the principal sources of the NGR intensity.
Terrigenous components are often diluted by other components such as biogenic components. The sediments consist mainly of biogenic and terrigenous components with small amounts of volcaniclastic particles off Sanriku (Sacks, Suyehiro, Acton, et al., 2000). In turn, SiO2 contents in hemipelagic marine sediments consist of both biogenic and detrital SiO2 with small amounts of volcanogenic silica. Off Sanriku, biogenic silica, especially diatom valves, is a major component of biogenic particles. A plot of SiO2 vs. Al2O3 contents shows wide scattering across the full data range, with a linear negative slope line on the upper limit of their ranges (Fig. F9C). The linear negative slope line means that there is a dilution relationship between biogenic silica and terrigenous aluminum. Other silica on the straight lines, which pass through the origin of the coordinates in Figure F11C, shows terrigenous silica included in siliciclastic minerals. The relation of biogenic and detrital silica is much clearer in Figure F9D. The SiO2/Al2O3 ratio has a clear negative correlation with Al2O3 content (Fig. F9D). The samples with high SiO2/Al2O3 ratios, representing high biogenic silica, correspond to those with high TOC (Fig. F9E). TOC off Sanriku should mainly derive from sea-surface production of siliceous phytoplankton dominated by diatoms. The SiO2/Al2O3 ratio is weakly correlated with the number of diatom valves (Fig. F9F). Because the counting methods for diatom valves under the microscope may not be quantitative, the relationship may not be clear.
Pore volume and/or density may have some control on the NGR signal if downcore variations in NGR activity are low (Blum, 1997). Porosity variations are proportional to the concentration of the matrix in sediments, which may be proportional to the concentration of a radioactive mineral in the sediments. This implies that NGR intensity is a reflection of compositional control of terrigenous minerals including radioactive elements and biogenic silica (Fig. F10A). The high NGR intensity (10-20 cps) shows negative correlation with the SiO2/Al2O3 ratio of weight percent, which ranges between 5 and 15 (Fig. F10A). In addition, there is an excess value of the ratio above the linear negative slope line at ~10 cps of NGR intensity. The NGR intensity has negative correlation with porosity at Site 1150 (Fig. F10B). The correlation at Site 1151 is not as obvious as that at Site 1150 but shows weak negative correlation. Grain density, however, does not vary with the NGR intensity (Fig. F11D). Therefore, the NGR intensity is affected by the compositional changes of diatom-dominant biogenic silica and terrigenous particles dominated by clay minerals. High porosity values corresponding to the high the excess value of the SiO2/Al2O3 ratio correlates to low NGR intensity. The porous structure of diatom valves results in high porosity of bulk sediments.
Consequently, the NGR intensity off Sanriku is controlled mainly by composition of terrigenous minerals with biogenic silica dilution. It was also affected by high porosity related to the amount of diatom valves. Using the analogy that the East Asian monsoon has a great influence on land precipitation and marine productivity in the present day, the NGR signal might to be a proxy of the past intensity of the East Asian monsoon off Sanriku.
In order to construct a continuous sedimentary record and to estimate stratigraphic ages in the logged hole to correlate with ages identified in the recovered cores, it is necessary to integrate the whole-core MST NGR intensity from Holes 1150A and 1150B with the wireline logging NGR intensity in Hole 1150D. As there was a depth shift between depth of recovered core and that of the borehole logs (e.g., Fig. F2), a procedure was necessary to compensate for the depth shift between the holes before core-log integration of NGR intensity. NGR measurements using the HNGS were reported at the most precise depths (Sacks, Suyehiro, Acton, et al., 2000). In this study, the depths of recovered cores were adjusted to the downhole logging depth by graphically correlating obvious peaks and troughs in the NGR profile Site 1150 (Table T4). Variations of the NGR profile in MST and logging data are illustrated in Figure F11. Recovered core depths between the top of the core and 117.7 mbsf, between 700 and 729 mbsf, and below 1146 mbsf shifted linearly -3.3 m, -1.9 m, and -1.1 m, respectively. Logging depths are known to be accurate above 117.7 mbsf because the end of pipe was set at about that depth. NGR was not logged between 700 and 729 mbsf or below 1146 mbsf (Table T4). The depth offset between Holes 1150A and 1150D ranged from +0.9 to -7.6 m (average = -2.4 m) and that between Holes 1150B and 1150D ranged from +1.3 to -7.0 m (average = -1.4 m) (Table T4). Although the logging NGR units (gAPI) are different from the MST NGR units (cps), the trends between core and log NGR can be compared after depth adjustment.
The logging NGR intensity was converted to a time series using a stratigraphic age model. Depths of stratigraphic events in Holes 1150A and 1150B were correlated to Hole 1150D in this procedure. The age model for Site 1150 used in this study is based on biostratigraphy and magnetostratigraphy for Holes 1150A and 1150B (Table T5). Biostratigraphic datums were based on Maruyama and Shiono (this volume), Kamikuri et al. [N1], and Li (this volume). Magnetochronology was based on Leg 186 shipboard stratigraphy (Sacks, Suyehiro, Acton, et al., 2000). In this study, only stratigraphic events that could determine ultimate age were used (Table T5); events that could determine age in some intervals were left out. Middepths of the events were used in the transferring procedure, if stratigraphic events were identified in intervals. As a result, the sedimentary sequence from 1.3 to 9.7 Ma was constructed with a short break at ~5 Ma (Fig. F12).
Based on linear sedimentation rates, which were calculated for the intervals between the stratigraphic events, eight stages of low (L) and high (H) sedimentation rate were identified (Fig. F12A; Table T6). Average sedimentation rates ranged from 3 to 10 cm/k.y. in L stages and from 19 to 73 cm/k.y. in H stages. NGR data were bandpass filtered at 100, 41, and 20 k.y. after resampling the original data in order to equalize sampling intervals to uniform rates (Fig. F12C, F12E, F12G; Table T6). The bandpass filter was centered at a period of 100 k.y. between 83 and 125 ka, 41 k.y. between 35 and 49 ka, and 21 k.y. between 19 and 24 ka. The power spectrum density (PSD) of the maximum entropy method (MEM) for each stage was calculated for the modified time series of the NGR variation between 2 and 100 (1/m.y.) frequency. A numerical series procedure was performed using the software program developed by Paillard et al. (1996). Depending on time control points (average interval = 0.442 m.y.), detailed discussion for values of periodicities was not possible, but the eccentricity, obliquity, and precession band periodicities were detected (Fig. F12I).
The H stages, especially H1, H2, and H4, corresponded to the intervals of low-amplitude precession index and eccentricity calculated by Laskar (1993) (Fig. F12D, F12H), therefore high sedimentation rates in the stages might not be always artificial. During the H stages, the power of the precession band component in the NGR signal was high, particularly in the H4 stage. Obliquity band and eccentricity band components always existed through 10 Ma. The obliquity band periodicities had high power in the L1 stage and moderate power in L4 stage. The power of those became weak in the L stages before the H1, H3, and H4 stages. Intensity changes in the periodic components in each stage through time suggest that the climate system reorganized during the H stages, particularly the H1 and H4 stages. An eccentricity-dominant system with moderate precession and obliquity in the L4 stage seemed to switch an eccentricity-dominant system with low precession and obliquity in the L3 through H4 stages. An eccentricity-dominant system with low precession and obliquity in the L2 stage seemed to switch to an obliquity-dominant system with low precession in the L1 through H1 stages. Transition of the dominant periodicities during the H stages may correspond to major shifts in the climate system. The H1 stage could be correlated with the onset of the East Asian monsoon, which is linked with the uplift of the Himalayas and the Tibetan Plateau. The uplift could have reached sufficient height at that time to produce rainfall in the Asian continental margin, causing increasing river runoff and transporting terrigenous minerals from land into the northwest Pacific. The L2 and L3 stages might indicate the initial stage of the East Asian monsoon intensification through the H2 and H3 stages. The H4 stage could correspond to the strengthening/transitional stage of the monsoon system at ~3.0-2.6 Ma. The timing matches the onset of major Northern Hemisphere glaciation. The L1 stage has been the prevailing stage of the monsoon since 2.7 Ma, relating to Quaternary climatic fluctuations on an glacial-interglacial timescale after the onset of major Northern Hemisphere glaciation.