A millimeter-scale insight into formation mechanism of lacustrine black shale in tephra deposition background

To reveal the role of tephra in the deposition of black shale during periods of volcanic activity, we performed lithostratigraphic and geochemical analyses on 14 horizontally sliced samples drilled from a 2-cm-thick black shale interval in the lower Ch7 Member of the Upper Triassic Yanchang Formation, southern Ordos Basin. Results indicate that fewer plankton is preserved during tephra deposition than during periods of volcanic quiescence. With the decrease of volcanic activities and tephra deposition, the abundance of redox-sensitive trace elements (RSTEs) and biolimiting elements increases, while terrigenous elements gradually decrease, resulting in the improvement of organic matter (OM) preservation. Paleoenvironmental proxies suggest that the climate during the Late Triassic was generally warm and humid. However, subsequent intense volcanic eruptions may have caused climatic cooling that affected the water column, resulting in enhanced salinity, primary production, water stratification, and bottom water anoxia, leading to enhanced organic carbon production and preservation. Primary productivity and redox conditions controlled the accumulation of organic carbon. Although physical and chemical reactions relating to the deposition of tephra into water are short-lived, climate change induced by volcanic eruptions is the critical cause of black shale formation.

www.nature.com/scientificreports/ of the microenvironment provide us with new insights into the formation mechanism of black shale in the tephra deposition background.

Results
Millimeter-scale mineral composition and lithofacies. Due to the organic matter enrichment, the vertical thin section in Fig. 2 is brown-yellow under the plane-polarized light. No visible fossils or sedimentary structures can be observed. Slice 6 is a set of relatively thick (~ 3 mm) tuff (Fig. 2), mainly composed of tuffaceous and clay minerals. A small amount of feldspar, quartz, and tuff debris are scattered in Slice 6. Locally, the tuff is metasomatized by dolomite, while the intergranular pores are filled with mineral asphalt. Except for slice 6, other slices characterize black shales, although they differ slightly in mineral compositions. Slices 2-5 contain relatively more tuffaceous debris and interstitial filling than slices 7-14. Slices 7-14 contain larger organic debris, more mineral asphalt, and less tuffaceous materials than slices 1-6 ( Fig. 2). As shown in Fig. 2, the TOC content for 14 slices ranges from 4.54 to 11.2%. The TOC content increases with the decreasing burial depth. By comparing, in slices 1-4, TOC is stable at ~ 5%. TOC increases to ~ 6% in slice 5 before it drops back to 4.8% in slice 6 when the tuff band occurs. Subsequently, in slice 7, TOC rapidly increases up to 6.6%. TOC increases steadily in slices 7-9 and stabilizes at ~ 10% in slices 10-14. Overall, TOC in slices 1-5 is approximately half of that in slices 10-14 (Fig. 2). According to the tuff occurrence and TOC content, 14 slices can be divided into three categories: slices 1-5 record the initiation of tephra deposition (phase I: active stage), with a small amount of tephra preserved due to locally weak volcanic activity; slice 6 denotes the deposition during an active regional volcanic eruption (phase II: eruptive stage); and slices 7-14 indicate depositions after the cessation of volcanic activities (phase III: intermittent stage). Slices 7-9 are further classified as phase III A , as they exhibit gradually increasing TOC, whereas slices 10-14 are classified as phase III B because their TOC contents are much higher (~ 10%). Additionally, according to the mineralogical analysis in Fig. 2, the average content of tuffaceous debris is 23% during phase I, and the tuffaceous material is dispersed amongst lacustrine shale. During phase II, the average content of tuffaceous debris is 47%, indicating strong volcanic activity. Entering phase III, the tuffaceous debris content shows a marked decrease (average 4%). The quartz content is relatively stable (~ 12%) throughout 14 samples, exhibiting little change before and after the volcanic activity. The clay mineral content increases significantly after tephra deposition (averaging 43% and 65% in slices 2-5 and 7-14, respectively).
Geochemical characterizations. Figure 2 shows the free hydrocarbon (S 1 ) and pyrolyzed hydrocarbon (S 2 ) for 14 samples. In general, evolutions of S 1 and S 2 are similar to that of TOC. Organic petrology. Moreover, kerogen macerals are shown in Fig. 3. Except for slice 6, maceral identification reveals that huminite and liptinite are significantly more abundant in phase III than in phase I, whereas vitrinite and inertinite are less abundant in phase III than in phase I. Therefore, the organic macerals can be classified into two categories. The first includes slices 1-5, which have similar characteristics. The tuff debris is coarse and widely distributed. The mineral bituminite matrix comprises a homogeneous mixture of tuff, clay minerals, and amorphous OM. Liptinite, especially microglobular alginites are relatively enriched in nodular forms. A large amount of inertinite, fusinite, and vitrinite occur ( Table 1). The second category encompasses slices 7-14,  Major terrigenous, biolimiting, and redox-sensitive trace elements. Figure 4 exhibits contents of the terrigenous elements (Al, Si, and Ti), carbon isotope (δ 13 C), and biolimiting elements Fe and P (i.e., elements present as components of DNA, RNA, polypeptides, ATP, phospholipids, and proteins that are involved in photosynthesis) for 14 samples. Overall, from bottom to top, the evolution of the terrigenous element is opposite to that of TOC and biolimiting elements. Aluminum (Al) content ranges from 8.36% (slice 11) to 19.64% (slice 1), with an average of 12.69% (Fig. 4). Silicon (Si) content ranges from 31.91% (slice 13) to 56.24% (slice 2), with an average of 48.14% (Fig. 4). Titanium (Ti) value ranges from 0.347% and 0.792% (average of 0.497%) (Fig. 4). Iron (Fe) values (average of 5.33%) are negatively correlated with Al (r = -0.78), Si (r = − 0.93), and Ti (r = − 0.71) values, but positively correlated with TOC (r = 0.88). Variations in P contents are similar to those in Fe, but P is enriched in slices 9 and 11. Carbon isotope values (δ 13 C) are strongly correlated with TOC contents (r = 0.94). Figure    However, U/Th is negatively correlated with Al (r = − 0.88).

Discussion
Regional correlation between OM accumulation and tephra deposition. Previous studies concerning whether volcanic ash can promote the formation and preservation of organic carbon or not are controversial. Some studies recorded that tephra deposits promote the formation of OM by Fe fertilization 21,35 , oxidant exposure reduction 36 , and reactive oxide complexation 37,38 . However, some other studies suggested that tephra deposition and OM content are not closely related 39 . For example, regarding Fe fertilization in the open ocean from tephra deposition, the average chlorophyll levels return to pre-bloom values within three months of a volcanic event 21,22 . Furthermore, reactive oxide complexation induced by tephra deposition only produces an increase (~ 5%) in the total OM burial rate 40,41 . In the Ordos Basin, tuff layers with thicknesses ranging from a few millimeters to a few centimeters coexist with black shales in the Ch7 Member (Fig. 6) 30 . The region with high TOC does not entirely overlap with the areas of the thickest tuff (Fig. 6). Moreover, black shales with TOC greater than 15% (extremely organic-rich) are distributed throughout the tuff region, regardless of their thick- www.nature.com/scientificreports/ ness (Fig. 6). Studies on the relationship between OM enrichment and tephra occurrence have mostly been qualitative 18,42 . A preliminary conclusion is that tephra is conducive to forming black shales 7 . However, this conclusion is not supported by regional evidence because there is no positive correlation between the cumulative tuff thickness and TOC of shale at the basin scale (Fig. 6). Therefore, the deposition flux of tephra could not be the critical factor in promoting the formation and preservation of organic carbon. In addition, the strong heterogeneity of lacustrine deposits may weaken the genetic relationship between them. Our high-resolution results from a small-scale sample, extrapolated to a regional scale, are probably the best solution to indicate the relationship between the OM accumulation and tephra deposits.
Millimeter-scale evidence of hypoxia influencing OM preservation. The abundance and depletion of different RSTEs, such as V, Ni, Co, Cu, and Mo, are valuable indicators for evaluating redox conditions in modern and ancient sedimentary systems [43][44][45] . The abundance of RSTEs in sediments relates to different reaction pathways under reducing conditions, leading to insoluble sulfides or organic metal complexes. Subsequently, products from the redox reaction are precipitated and adsorbed as oxides and hydroxides on the OM surface 34 . These elements enter lakes in various ways (e.g., through continental runoff) and share the properties of adjacent source rocks. As a result, abundant RSTEs provide sufficient evidence of anoxic conditions. Given that the redox indices might be contradictorily interpreted 46,47 , complementary auxiliary factors were used, such as TOC and RSTE enrichment index, to further validate results suggesting anoxic conditions. The U/Th ratio is used to assess redox conditions further. The oxygen depletion in the process of tephra deposition would decrease the abundance of soluble U in the water, resulting in U depletion in the water column and U enrichment in sediments 43,[48][49][50][51] . Therefore, higher U/Th values might result from anoxic conditions. Additionally, high U/Th values of the slices are associated with organic-rich layers characterized by abundant RSTEs and high values of enrichment index (Fig. 5). Therefore, anoxic conditions are enhanced after the volcanic eruptions, and organic carbon is more effectively preserved. We observe increasing TOC under anoxic conditions in the absence of apparent lithologic change. Temporal variations in the RSTEs indicate fluctuating bottom water conditions. Our data show that the proxies (e.g., RSTEs) for oxygenation of the bottom water fluctuate in phases I, II, and III. Thus, the four intervals (slices 5, 7, 11, and 14) characterized by increasing values of RSTE enrichment index (Fig. 5) and the simultaneous occurrence of thin organic-rich layers indicate recurring episodes of hypoxia during this period.
Reconstructions of temperature, water depth, and salinity. The contents and distributions of some elements are indicators of paleotemperature, paleowater depth, and paleosalinity. For example, Zr is a typical lithophile element (primarily mechanical migration) deposited near its source 52 . Thus, Zr is frequently used as an indicator of the transportation distance from a sediment source. Generally, Zr contents decrease with the increasing distance from the source. Boron can be used for quantitative reconstructions of salinity. Boron concentrations in marine and freshwater environments are generally 80-125 ppm and less than 60 ppm 53 , respectively. The formula proposed by Couch 54 was used in this study, which has good applicability in continental strata, to calculate the paleosalinity of our samples. On average, in Fig. 7, the paleosalinity of the slices is ~ 1.5‰, Figure 7. Zr element, calculated salinity, and Sr/Cu profiles for 14 samples, which could be used to indicate the palaeowater depth, paleosalinity, and paleotemperature, respectively. *The salinity was calculated based on equivalent boron content following the formula proposed by Couch 54 . www.nature.com/scientificreports/ indicating a freshwater environment. Consequently, as the climate during the studied interval becomes drier and colder, the paleowater depth decreases. The paleosalinity gradually increases with a small amount as the evaporation process exceeds the precipitation process. Sr/Cu values in lake sediments are sensitive to changes in climate. Values of 1-10 generally indicate warm and humid conditions, while those greater than 10 indicate hot and dry climates 55 . The Sr/Cu values of the samples (average of 6.59) correspond to a warm and humid environment. The values decrease from 9.26 in slice 1-4.09 in slice 14, indicating a decrease in paleotemperature. Except for an abnormally high value of 10.23 from slice 6, likely caused by elevated temperatures during volcanism, paleotemperatures gradually decrease with time in our samples (Fig. 7). In particular, Sr/Cu values decrease by ~ 50% after the cessation of volcanic activities because tephra can remain in the atmosphere to block the ultraviolet radiation and lower atmospheric temperatures 56 .
Productivity and water stratification controlled OM preservation. The identification of organic macerals reveals the origin of OM. The OM in our slices is produced by a mixture of in situ and terrigenous photosynthates (Fig. 3). TOC fluctuations reflect environmental characteristics (Fig. 3) and a strong synchronization with peak P and Fe values (Fig. 4). P and Fe are key biolimiting elements for maintaining biomass productivity [57][58][59][60][61] . Our interpretation of the paleoenvironment depends on the known transport pathways of P and Fe to aquatic organisms, but their origins are unclear. During the sequestration of OM (Fig. 4), the decreasing relative abundance of terrigenous elements (Si, Al, and Ti) represents the decreasing river discharge and constrains factors controlling organic carbon production. Thus, rivers are not the primary contributor to the biolimiting elements (P and Fe); instead, they are supplied by other terrigenous sources containing heterogeneous OM (Fig. 4). A possible explanation is that the input of P and Fe into the lake water is relatively stable, but the concentrations of Si and Al are relatively high when the tephra is deposited, resulting in a dilution of P and Fe. Another possible explanation is that hydrothermal activity provides more P and Fe. However, there is insufficient evidence to confirm this scenario. During phase III, the primary productivity increases as the water becomes stratified and the biolimiting element contents (P and Fe) increase. Therefore, the continuous increase in biological oxygen demand in the water exacerbates the oxygen deficiency in the sediments, leading to enhanced OM preservation. Thus, the studied area reflects the influence of regional climate change. After the volcanic activity ceased, the water column is stratified due to the drier and cooler climatic conditions. Increasing paleosalinity in the organic-rich layers supports this inference (Fig. 7). The greater abundance of clay minerals in phase III (Fig. 2) may have provided favorable conditions for OM preservation by adsorption. Even when possible differences in the preservation conditions are considered, the boom of organisms after volcanic activities was significant, showing that nutrient levels were sufficient (Fig. 4). Previous studies have indicated a decrease in productivity with climates getting warmer. For example, the temperature increase affects the supply of nutrients to the euphotic layer, resulting in a decrease in phytoplankton productivity that cannot be offset by an increase in phytoplankton biochemical response rates 62 . For example, chlorophyll concentrations are low in the Pacific Ocean during warm periods 63 . Phytoplankton cell abundance in a cold-water area of the northeastern Atlantic Ocean increased, while that in a warm-water area decreased 64,65 reported a negative correlation between temperature and the volume of phytoplankton cells: the volume decreased by 2.5%/°C, and the biomass of phytoplankton also decreased to some extent 66 . Consequently, the productivity increase shown by our data with decreasing temperature in the background of a warm, humid climate is reasonable.
Black shale formation mechanism in tephra deposition background. Our results demonstrate that volcanic events would enhance the OM preservation and the formation of black shales 38 . Since biolimiting and redox-sensitive elements show strong correlations with TOC in the present study, the combined effects of productivity and hypoxia are the key factors controlling the formation and preservation of organic carbon. Increased RSTE contents, indicating redox conditions, primarily reflect enhanced water stratification. TOC enrichment is associated with relatively high values of paleosalinity (overall freshwater). Water with higher salinity would produce a stronger density gradient, causing more anoxic episodes. The direct impact of tephra deposition was transient and limited. The relatively long-term climate change caused by the volcanic eruptions might explain the enhanced OM accumulation in our samples. The largest tephra flux may have diluted the abundances of Fe and P, which were both sourced from terrigenous input. Intense volcanic eruptions resulted in a clear paleoclimatic cooling when the tephra entered the stratosphere. This would activate hypochlorous anhydride (ClO) through nitrate and sulfate, causing a sudden interdecadal decrease in total stratospheric ozone 56 . The reduction in absorbed heat would cause an interdecadal reduction in stratospheric temperature and geopotential height 56 . Under the warm and humid climate conditions in the Middle and Late Triassic, the Ordos Basin became colder due to tephra eruptions, further strengthening water-column stratification and enhancing TOC preservation. The cooler climate simultaneously enhanced the primary productivity of the phytoplankton in the water. Unlike modern oceans, tephra deposition did not directly promote the accumulation of organic carbon in the ancient lake. The fertilization of tephra has little contribution to the formation of the ancient thick organic-rich shale. Moreover, the extent of tephra deposition is influenced by airflow and topography. Its direct effect (e.g. reducing oxidant exposure) on the lacustrine sediments was probably local (Fig. 6). However, climatic, hydrologic, biological, and other environmental changes caused by volcanic activities were long-lasting and wide-ranging. They are the main contributing factors to organic carbon accumulation. This sedimentary model could provide a new understanding of the formation mechanisms of black shales.

Conclusions
At a regional scale, the deposition flux of tephra does not influence the accumulation of organic carbon in the Ordos Basin during the Late Triassic. High-resolution stratigraphical and geochemical characteristics in a 2-cm shale from the lower Ch7 Member in the Ordos Basin were investigated at a millimeter scale. Depth profiles of the minerals, elements and organic components enable us to test possible deposition mechanisms for black shale in the tephra background. As suggested by biolimiting elements, high productivity would cause intense organic carbon deposition. During the dormant period, which is beneficial to OM deposition, the primary productivity is higher than that during the tephra deposition period. Geological factors controlling both booms and the extinctions of aquatic organisms in the lake basin influence OM abundance. Redox-sensitive elements indicate that oxygen levels persistently decreased in the lake bottom water along with the volcanic intermission, which would contribute to the high organic carbon content. In other words, a low oxygen content facilitates OM preservation. The physical and chemical reactions caused by tephra deposition are pretty short-lived. Therefore, the additional nutrients in surface water and anoxia in bottom water triggered by tephra deposition are difficult to record in the geological history. The variations of sedimentary environment recorded in shales resulted from climatic changes caused by tephra. A cooling climate for a considerable time after the eruption probably leads to increased salinity, primary productivity, water stratification, and bottom water anoxia, which promote the input and preservation of organic carbon on long timescales (ca. 1000 yrs). Conventional methods based on XRD analysis on pulverized samples provide average information of sedimentary beddings but may neglect distinct paleoenvironmental information. In this study, a system for lacustrine black shales to recover the information on individual bedding was established. To truly reflect the geological processes, integration of special geological events with the nutrient supply, the hydrologic state of lakes, and the flux of clastic sediments were carried out.

Materials and methods
Ch7 shale samples. The upper Triassic, Yanchang Formation, Ch7 Member Shale was deposited in the southern Ordos Basin, North China 5 . The thickness of the Ch7 Shale varies from 150 to ~ 10 m in Ordos Basin 27 , with ultra-high organic matter abundance 5,6 . The shale contains frequent volcanics thinning vertically from the bottom to top of the Ch7 Member 7 . Well L85 is located at the edge of the Yishan Slope (Fig. 1a). Samples collected from (Fig. 1c) sealed cores in Well L85 are mainly black shales with tuffaceous laminae. To determine the tephra effects on shale formation, it is necessary to investigate the sediments deposited before and after the tephra entered the paleolake water column. Given the deposition duration of the Ch7 Member was at least 5.5 Ma 31-33 , the sedimentation rate is less than 1.8-2.2 cm/ka. A 2-cm-thick shale core (101.6 mm diameter) sample containing a tuff layer (Fig. 2) was selected, which corresponds to an interval of ca. 1000 yrs. In addition, no tuffs exist at least 1 m above the selected sample, avoiding the influence of multiple tephra sets. A thousand years is much longer than the duration of a single volcanic eruption (several months) 21,22 .
Sample processing. The 2-cm core sample was cut into two parts vertical to bedding (approximately 1:4 volume ratio). The small part was prepared as a large thin section vertical to bedding over the entire sampling range (leftmost image in Fig. 2). The larger part was slabbed and then sliced parallel to bedding at ~ 1 mm resolution, using a diamond wire saw. The 2-cm-thick sample was cut into 14 continuous slices (i.e., L85N-1 to L85N-14 from bottom to top, referred to as slices [1][2][3][4][5][6][7][8][9][10][11][12][13][14]. The thickness of each slice ranges from 0.6 to 1.1 mm. Noticeably, slice 6 is a 3-mm-thick tuff. 4 cm 2 segment of each slice was saved (in the same position) to prepare the transmission light and reflected light thin sections (horizontal thin section, slice 9 is absent due to damage). The rest part of each slice was respectively dried and powdered to 200 mesh, and subsamples were taken for individual analyses, including XRD, TOC, Rock-Eval, major and trace elements, and carbonate C-isotope.
Facies identification via optical microscopy. Sample preparation followed ASTM D2797 67 wherein the vertical slice was mechanically ground and polished by paper (120-1200 grit) followed by final polish on matte clothes with 1.0 μm alumina to achieve smooth surfaces for microscopic analysis. We took images using a transmitted light microscope (Olympus BX-53 manufactured by the Olympus Corporation, Tokyo, Japan) at the Central Laboratory of Geological Sciences (CLGS), PetroChina Research Institute of Petroleum Exploration and Development (Beijing, China). From the vertical thin section images, minerals and tuff layers are easy to be observed.
Organic petrography analysis via optical microscopy. The 13 slices (except slice 9) were carried out organic petrography analysis by using maceral identification in epifluorescence followed ASTM D7708 68 . Maceral identification employed an Olympus BX53M microscope with LED illumination with the computer program PRECiV™ by the China University of Geoscience.
Bulk characterization of shale. Total organic carbon (TOC) analysis and Rock-Eval were carried out at the CLGS following methods previously described in Barker 69 and Peters et al. 70  www.nature.com/scientificreports/ Mineral composition determination. XRD measurements to determine mineralogy have been performed on Rigaku D/Max 2500 v/PC system using Cu Kα radiation (Rigaku Corporation, Tokyo, Japan) at the CLGS. Quantitative phase analysis is performed by using MDI Jade software. The precision of these measurements is better than 0.1wt% for phases in which the content is above 2%. Mineral compositions relate to the crystalline content of the analyzed samples.
Major and trace element chemistry. Major and selected trace elements in samples rock powders were determined using an Axiom AX X-ray fluorescence spectrometer (Axiom Electronics LLC, Hillsboro, USA) and an ELEMENT XR inductively coupled plasma-mass spectrometer (Thermo Electron Corporation, Waltham, USA) at the Beijing Research Institute of Uranium Geology, China (BRUG). Internal precision and external reproducibility are typically better than 1% and 3%, respectively.
Isotope ratio measurements. All samples were selected for isotope ratio measurements. The stable carbon isotope ratio of carbonate (δ 13 C carb ) was measured on duplicate subsamples of 500 μg using an isotope ratiomass spectrometer (Finnigan MAT-253, Thermo Electron Corporation, Waltham, USA) at the BRUG (China).
The δ 13 C results are reported in delta notation concerning PeeDee Belemnite (PDB). The typical values were better than 0.06%.
Statistical and multivariate analysis. The relative abundance for each element was estimated by the average value of the background level 34 based on the following equation: where [Element A ] represents the concentration of each sample and [Element A ] background is the average value of the corresponding element for all the samples. EI means the enrichment index. The correlation coefficient (r) of any two proxies is calculated by the Pearson's equation 71 : where x and y are the corresponding sample's average. The results of this equation go from a perfect negative correlation (− 1), no correlation (0) to a perfect positive correlation (1).

Data availability
The raw datasets of this study are provided in the Supplementary Information, including five tables (supplementary Tables S1-S5). (1)