Paleoclimatic and provenance implications of magnetic parameters from the Miocene sediments in the Subei Basin

Thick sediments from foreland basins usually provide valuable information for understanding the relationships between mountain building, rock denudation, and sediment deposition. In this paper, we report environmental magnetic measurements performed on the Miocene sediments in the Subei Basin, northeastern Tibetan Plateau. Our results show two different patterns. First, the bulk susceptibility and SIRM, ARM, and HIRM mainly reflect the absolute‐concentration of magnetic minerals; all have increased remarkably since 13.7 Ma, related to provenance change rather than climate change. Second, the ratios of IRM100mT/SIRM, IRM100mT/IRM30mT, and IRM100mT/IRM60mT, together with the redness and S ratio, reflect the relative‐concentration of hematite, being climate‐dependent. Their vertical changes correlate in general with the long‐term Miocene climatic records of marine oxygen isotope variations, marked by the existence of higher ratios between 17 and 14 Ma. This may imply that global climate change, rather than uplift of the Tibetan Plateau, played a dominant role in the long‐term climatic evolution of the Subei area from the early to middle Miocene.


Introduction
The Cenozoic Era is characterized primarily by a cooling trend, revealed by comprehensive marine oxygen isotope records (Zachos et al., 2001(Zachos et al., , 2008. The most significant cooling event occurred at 34 Ma and marked by the formation of the Antarctic Ice Sheet (Zachos et al., 2001). The Mid-Miocene Climatic Optimum (MMCO), with temperatures of ~6 °C higher than at present, was the warmest period since 34 Ma (Miller et al., 1991;Wright et al., 1992;Flower and Kennett, 1994;Zachos et al., 2001). Furthermore, global cooling after the MMCO, since 14 Ma, is believed to have been controlled by the Antarctic cryosphere evolution and/or uplift of the Tibetan Plateau (Molnar et al., 1993;Flower and Kennett, 1994;Miao YF et al., 2012), but which factor played the dominant role is still under debate (Raymo and Ruddiman, 1992;Flower and Kennett, 1994). Understanding regional climate responses to global change and mountain uplift will help address the controversy. In the Tibetan Plateau, sediments in foreland basins offer valuable information on the relationships between mountain up-lift, rock denudation and sediment deposition. If we can recover regional climate history, it is possible to examine local responses to tectonic events and global climate change. However, compared to abundant marine records (Flower and Kennett, 1994;Abreu and Haddad, 1998;Flower, 1999;Turco et al., 2001;Zachos et al., 2001;Larsson et al., 2011;Holbourn et al., 2014), long-term terrestrial paleoclimate reconstructions are still limited.
The objectives of this study are (1) to reconstruct the climate history from the early to late Miocene in the northeastern Tibetan Plateau, and (2) to examine its relationships with the global climate change and uplift of the Tibetan Plateau.

Geological Setting
The Subei Basin is located at the junction area of the Altyn Tagh Mountains and the western end of the Danghe Nan Shan (Figure 1). The elevation ranges from 2500 to 3000 m. The Altyn Tagh Fault (ATF) is the first-order feature of the Tibetan Plateau and marks its northern edge ( Figure 1); it turns sharply to become a series of thrusts and thus plays a critical role in controlling the development of the thrust-fold belts around the Danghe Nan Shan.
The Subei Basin is cut by three main river channels: Yandantugou, Tiejianggou, and Xishuigou from west to east (Figure 2), respectively. Comparison of magnetostratigraphic results for three sections indicates that the Tiejianggou section holds the most complete set of strata from Oligocene to Miocene (Gilder et al., 2001;Yin A et al., 2002;Sun JM et al., 2005). Thus, our studied section fo-cused on the Tiejianggou section ( Figure 2). Magnetostratigraphic results combined with biochronology constrain our sequences from 22.8 to 9.0 Ma (Wang XM et al., 2003;Sun JM et al., 2005). The Tiejianggou Section includes three formations from south to north: the Oligocene Paoniuquan Formation, the Miocene Tiejianggou Formation, and an unnamed formation in late Miocene to Pliocene. The Tiejianggou Formation is limited by two faults (F0 and F1), as its upper and lower boundary, respectively. It is composed of three sedimentary units.
Based on the paleomagnetic results, we calculated the sedimentary rates by using linear interpolations of two neighboring ages ( Figure 3). The studied strata can be subdivided into three parts: (1) The lowest unit (22.8−18.8 Ma) mainly contains reddish finegrained mudstones with a thickness of 659 m, representing a shallow lake environment, with a low sedimentary rate of 164.75 mm/ka; (2) The middle unit (18.8−13.7 Ma) is a mix of brownish mudstones or siltstones and growing interbedded grey sandstones with a thickness of 908 m, implying fluvial and lacustrine environments with a sedimentary rate of 178.04 mm/ka; (3) the uppermost unit (13.7−9.0 Ma) is predominated by conglomerates and thin siltstone intercalations representing high energy floods with sedimentation rate of 250 mm/ka in its lower part, increasing to 450 mm/ka after 11 Ma.

Magnetic Experiments
Two hundred ten samples were obtained, pulverized and then Ku nl un Fa ul t Q a i d a m Q im a n T a g h N o rt h A T F A lt y n Ta g h M ts . So ut h AT F D a n g h e N a n S h a n Q ili an M ts . placed into pre-weighed 2cm × 2cm × 2cm plastic boxes and weighed. measurements of low-frequency (976 Hz) and high-frequency (15616 Hz) magnetic susceptibility were made with a Kappabridge MFK1-FA. Anhysteretic remanent magnetization (ARM) was imparted in a 100 mT peak alternating field with a 0.05 mT direct current field superimposed, using a 2G-760 model Uchannel superconducting magnetometer. Saturation isothermal remanent magnetization (SIRM) was attained in a 1 T magnetic field using the 2G-660 Pulse Magnetizer. Followed by three AF demagnetizations of 30 mT, 60 mT, and 100 mT, respectively, the corresponding IRM 30mT , IRM 60mT , IRM 100mT were produced. After demagnetization, samples were subjected to a 1 T field and then a 0.3 T bias field for measurement of IRM 30mT . The parameters used in the context were given according to Liu QS et al. (2012):

Color Measurements
Additionally, we performed color measurements. Samples were first dried at 40 °C for 24 h and were then crushed and measured using a Minolta-CM2002 spectrophotometer. For all samples, lightness (L*), redness (a*), and yellowness (b * ) were attained. All the above experiments were carried out in the Institute of Geology and Geophysics, Chinese Academy of Sciences.

Magnetic Susceptibility (χ) and Frequency-Dependent Susceptibility (χ fd )
The values of χ range from 1.4 to 13.0 × 10 −7 m 3 /kg with an average value of 4.18 × 10 −7 m 3 /kg. Two stages can be identified in the whole section ( The value of χ fd fluctuates from −0.2 to 0.2 × 10 −7 m 3 /kg with an average value of 0.02 × 10 −7 m 3 /kg.

Interparametric Ratios
We investigated three interparametric ratios, IRM 100mT /IRM 30mT , IRM 100mT /IRM 60mT , and IRM 100mT /SIRM. They show a similar trend but reveal a different pattern from the bulk magnetic susceptibility and remanent magnetization. We separated the whole section into three stages (

Color Records
We used lightness L * , redness a * , and yellowness b * of the spherical L * a * b * color space. Color parameters of a * and L * records generally show a trend similar to that of the S ratio ( Figure 6). In Stage I

Mountain Uplift Since 13.7 Ma
There is an obvious increase at 13.7 Ma for all four magnetic parameters (Figure 4). Magnetic susceptibility reflects the contributions of all magnetic materials; remanent magnetization refines the picture by reducing the effects of superparamagnetic (SP) particles. ARM is affected by grain size changes and is very sensitive to the concentration of SD particles, while SIRM is responsive to the general magnetic mineral concentration as long as the grain size and mineralogy remain relatively constant. HIRM indicates the mass concentration of high-coercivity minerals, e.g., hematite and goethite. χ fd is used to indicate the presence of SP particles (Zhou LP et al., 1990). The increase at 13.7 Ma implies a higher mass concentration of both ferrimagnetic materials and high-coercivity materials. Both tectonic events and climate change could account for such a change (Zhou LP et al., 1990;Sun JM et al., 2005). Several factors are known or hypothesized to be related to the magnetic properties, including input of aeolian dust, pedogenic processes, low-temperature oxidation, and increased source materials due to active tectonics.
Aeolian dust flux was stable during 22−7 Ma according to Guo ZT et al. (2002). Additionally, the particle size of dust is usually fine, and this is inconsistent with the up-section coarsening trend in our study.
The value of χ fd fluctuates greatly in the lower part of the section; its value is higher after 13.7 Ma. It could be logically speculated that the presence of diamagnetic materials such as quartz and plagioclase could account for the lower and fluctuating values before 13.7 Ma. Moreover, the value of χ fd % is generally less than 0.01. Therefore, we believe that the contribution of pedogenic processes can be neglected.
According to Ritts et al. (2004), the major paleocurrent in the Subei area changed from southeast to southwest in the middle late Miocene. Detrital apatite fission track and detrital zircon dating also indicate a sediment source change around 14 Ma (Li JF et  Sun JM et al. (2005). Corresponding sediments depositional rates are calculated and plotted. doi: 10.26464/epp2020030 311 al., 2014;Lin X et al., 2015). Based on sedimentary facies and sediment accumulation rate analysis, Sun JM et al. (2005) suggested that the source material changed at around 13.7 Ma, and that the main magnetic material changed from hematite to magnetite and hematite. We also suggest that the increase of susceptibility and magnetic remanences reflect the change of source materials.

Earth and Planetary Physics
Here is a possible scenario. The magnetite (source material) had undergone prolonged chemical weathering processes before it was transferred to the foreland basins by low energy rivers in stable tectonic periods; most of the minerals were preserved in the form of hematite. However, during periods of active tectonics, new bedrocks were uplifted and new minerals derived from the bedrocks were only partially oxidized because of intense erosion and high energy river floods. These resulted in the accumulation of molasse deposits in the foreland basins, with the minerals being a mixture of magnetite and hematite. It is worth noting that the HIRM results show that the absolute-concentration of hematite increased after 13.7 Ma. However, just as in the proposed scenario above, the absolute-concentration of magnetite also in-creased considerably, as evidenced by the sharp-increase of bulk magnetic susceptibility, due to the fact that the bulk susceptibility of magnetite is 1000 times greater than that of hematite (Collinson, 1983). Therefore, the relative-concentration of hematite, which is climate-dependent, did not increase after 13.7 Ma, a conclusion that can be further supported by the magnetic ratios and redness as indicated by Figures 5 and 6. Conclusively, the activation of tectonics in the Subei area since 13.7 Ma is the fundamental cause of the change of magnetic properties. This is in accordance with other reports concerning the uplift history of the northern Tibetan Plateau (Li JJ et al., 1997;Sun JM et al., 2005;Bovet et al., 2009;Li JF et al., 2014;Wang CS et al., 2014;Zhuang GS et al., 2014;Lin X et al., 2015;Lin XB et al., 2016;He PJ et al., 2018).

MMCO During 17-14 Ma
The interparametric ratios, IRM 100mT /IRM 30mT , IRM 100mT /IRM 60mT , and IRM 100mT /SIRM, can be used as indicators for the relative contributions of hematite (Deng CL et al., 2006). The S ratio has been widely used as a climatic proxy to measure the relative abundance of high-coercivity minerals in a mixture with ferrimagnetic minerals (e.g., magnetite, maghemite) (Yamazaki and Ioka, 1997;Rousse et al., 2006;Liu QS et al., 2007Ao H et al., 2010;Fang XM et al., 2015;Zan JB et al., 2018;Guan C et al., 2019). Approaching unity points to the domination of ferrimagnetic minerals, and the S ratio will decrease with an increase of hematite. Redness is an intuitive proxy to reflect the concentrations of iron-bearing components, mainly hematite, and is thus widely used (Nagao and Nakashima, 1992;Balsam et al., 1999;Helmke et al., 2002;Abdul Aziz et al., 2003;Aziz et al., 2004;Jiang HC et al., 2008;Sayem et al., 2018). Synthetically, these proxies suggest a maximal relative abundance of hematite from 17 to 14 Ma and a decreasing trend of hematite concentration after 14 Ma.
As we have discussed above, tectonics had a profound influence on the abundance of hematite and magnetite. However, this tectonic influence can be eliminated by using the above ratios. In such a case, all these ratios reflect the relative-concentration of hematite to the total magnetic minerals rather than the absoluteconcentration of hematite or magnetite. This can be evidenced by the fact that all the magnetic ratios show a trend similar to that of the redness (Figures 5 and 6), which reflects mainly the climatecontrol degree of chemical weathering.  Figure 5), indicating high values of the relative concentrations of hematite, thus the climate condition in this stage must be the warmest and the most humid of the Miocene. This warm phase corresponds to the MMCO (Miller et al., 1991;Wright et al., 1992;Flower and Kennett, 1994;Zachos et al., 2001;Sun JM and Zhang ZQ, 2008;Zan JB et al., 2015;Lin XB et al., 2016;Guan C et al., 2019). Stage III (14−9.7 Ma): decreasing ratios of IRM 100mT /IRM 30mT , IRM 100mT /IRM 60mT , and IRM 100mT /SIRM and increasing values of the S ratio are consistent with lower relativeconcentrations of hematite and thus suggestive of a weakening weathering process under a cooling climate.

Conclusion
Based on detailed analyses of the magnetic and color parameters of the Miocene strata in the Subei Basin, we can draw the following conclusion: (1) The abrupt increase of bulk magnetic susceptibility, SIRM, ARM, and HIRM since 13.7 Ma mainly reflects the change of the absolute-concentration of magnetic minerals, which is related to changes in source materials in response to the intensive rock denudations related to tectonic uplift of the adjoining mountains since the late Miocene. (2) Different from the bulk magnetic susceptibility, SIRM, ARM, and HIRM, the S ratio, and the other ratios of IRM 100mT /SIRM, IRM 100mT /IRM 30mT , IRM 100mT /IRM 60mT , as well as the redness, all reflect changes in the relative-concentration of hematite, which were thus climate-dependent. The change pattern indicate three stepwise climatic changes: Stage I (22.8−17 Ma) is a temperate period with a slight warming trend; Stage II (17−14 Ma) is characterized by a warm phase corresponding to the MMCO; Stage III (14−9.7 Ma) reveals a weakening weathering process related to global cooling.