Geochemistry and mineralogy of the Upper Ordovician Wufeng formation, Southwestern China: Implications for Paleoclimate construction

Wufeng Formation shale is an important source rock of unconventional hydrocarbons in the Lower Paleozoic shales of Sichuan Basin. However, the study on its provenance and paleoclimate is still relatively limited. In this study, mineralogical and geochemical data of the shales from the Upper Ordovician Wufeng Formation in southwestern China has been used to interpret the provenance and conditions of weathering and paleoclimate. The Wufeng shales have intermediate to high SiO2 (57.72–82.38 wt. %, av. = 68.84 wt. %) and Al2O3 (5.26–16.17 wt. %, av. = 10.62 wt. %), are rich in transition metal elements (i.e. V, Ni, Cu, Co and Cr) and Y as well as moderate depletion in Na2O and Sr, relative to the concentrations of the upper continental crust (UCC). In the chondrite-normalized (CN) rare earth elements (REE) distributions, these rocks display light REE (LREE) enrichment (La/YbCN = 6.69–12.63, av. = 9.28), ﬂat heavy REE (HREE) (Gd/YbCN = 1.35–2.41, av. = 1.70), and clearly negative Eu anomalies (Euan = 0.50–0.66, av. = 0.58), showing similar characteristics with the CN post-Archean Australian Average Shales (PAAS). Wufeng Formation shales are immature composition without evident recycling sediments, and they are originated from an intermediate-felsic igneous source composed of tonalite–trondhjemite–granodiorite (TTG), granitic and andesitic igneous rocks. The chemical weathering conditions of studied shales decreased from moderate to low in the provenance region, suggesting a gradual cooling trend of the climate at Late Ordovician Thus, this article will be helpful to discern the provenance and variations of chemical weathering conditions and paleoclimate of Wufeng Formation shales.

Ordovician glaciations (Sheehan, 2001).The Upper Ordovician Wufeng Formation is mainly composed of black shales which may preserve important information about the palaeoenvironment and paleoclimate of that period (Figure 1(b)).
Previous studies of Wufeng shales mainly focus on its organic petrology (Luo et al., 2017;Luo et al., 2018;Luo et al., 2020), paleogeographic environment (Chen et al., 2004), paleontology (Chen et al., 2000), and petroleum geochemistry (Liang et al., 2008;Chen et al., 2011;Liu et al., 2011;Zou et al., 2012;Guo and Zhang, 2014;Gao et al., 2022), but until now the study on its provenance and the variations of weathering and paleoclimate is still insufficient.Chen et al. (2004) proposed that the palaeoclimate changed from Greenhouse effect to icehouse effect at Ashgillian (Figure 1(b)).Yan et al. (2009) concluded a warmer climate and a colder climate at Katian and Hirnantian, respectively, based on the isotopic compositions of sulfur in iron sulfide (δ 34 S sulfide ) and of carbon in organic matter (δ 13 C org ) in the sediments from the Yangtze Platform, South China.However, the episodic promoted organic matter burial rates during the Early to Late Ordovician, as suggested by the change in high-resolution δ 13 C carb and δ 13 C org records from worldwide sediments, might have contributed to a long-term cooling in climate, which peaked during the Hirnantian glaciation (Brenchley et al., 2003;Yan et al., 2009;Zhang et al., 2010;Zhou et al., 2015;Algeo et al., 2016 and references therein).Cooling in climatic due to prompted organic carbon burial might be an important trigger for the end of Ordovician mass extinction (Sheehan, 2001;Zhang et al., 2010;Zhou et al., 2015;Algeo et al., 2016;Liu et al., 2021;Qiao et al., 2022).Enhanced organic carbon burial associated with paleoclimate changes may have been considered as an important factor of the end-Ordovician mass extinction.
The geochemical features of shales are considered to maintain more information of original geochemical signatures of provenance (McCulloch and Wasserburg, 1978;Bhat and Ghosh, 2001;Qiao et al., 2022).This time, all the samples came from shale rather than sandstone, and we only discussed the shale.The major and trace element concentrations and mineralogy from the Wufeng shales from southwestern China were examined in this study to (1) determine the provenance of the Wufeng shales; (2) evaluate the chemical weathering conditions; and (3) reconstruct the Late Ordovician paleoclimate.

Geological setting
The studied region is located in southeastern Chongqing, close to Sichuan Basin (Figure 2).The structural pattern is partition-style folds in southeast and trough-like folds in northwest (Zhai, 1987).From the Early to Middle Ordovician, the investigated region was submerged by a broad epeiric sea and was isolated by the uplift (Chen et al., 2004;Luo et al., 2016;2017;Mu et al., 2011).During the Late Ordovician, there was a global transgression in Yangtze Block, which led to the development of Wufeng Formation (Mu et al., 2011).The Wufeng Formation mainly comprises black graptolitic shales including Dicellograptus complanatus-Paraorthograptus pacificus graptolite zones (Chen et al., 2000;2005), indicating the transgressive systems tracts (TST) in a typical sea-level sequence.The Guanyinqiao Formation, which includes typical Hirnantia Fauna (Rong et al., 2002) and is primarily composed of argillaceous limestones (Figure 1(b)), lie conformably on the Wufeng Formation, indicating a shelf-margin systems tract (SMST) at the margin of Yangtze Platform.The Guanyinqiao Formation was characterized by a drop in eustatic sea level and cool/cold water (Rong et al., 2002;Yan et al., 2009Yan et al., , 2010)), which is time-equivalent to the Hirnantian period (Sheehan, 2001).Plenty of K-bentonite beds were found in the Ordovician and Silurian sediments on the Yangtze Block.The geochemical data of the K-bentonites indicated a parent magma that was sourced in and within-plate tectonic, volcanic-arc and syn-collisional settings (Su et al., 2006(Su et al., , 2009)).

Sampling and methods
Thirty-seven core shales of the Wufeng Formation were collected from Y1 well, with sampling interval of 10-50 cm (Figures 1(b) and 2 and Table 1).Major elements contents were measured by using the X-ray fluorescence spectrometer (AB-104L, PW2404) at the laboratory of China National Nuclear Corporation (CNNC) Beijing Research Institute of Uranium Geology.The slice melting method was applied for determining concentrations of major elements.The process of measurement was described by Hu et al. (2014) in detail.The standard materials dolomite, andesite, and shale were used as standard samples to check the analysis accuracy with the analytical precision and accuracy of <2% for all major elements.The concentrations of trace element were performed by using the inductively coupled plasmamass spectrometry (ICP-MS) as described by Qi et al. (2000) in detail.The powdered sample of about 0.05 g was combusted in a muffle furnace before put into a PTFE bomb and then adding a mixture solution (HNO 3 : HF = 2:1) (Zhang et al., 2018).Then sealed bomb was put in an electric oven and heated to 185°C for 24 h (Zhang et al., 2018).After cooling, about 500 ng of Rh (standard material), and mixed liquid (water: HNO 3 = 5:2) were added into the bomb, and the sealed bomb was put into an electric oven again at 135°C for heating 5 h to dissolve the residue materials (Zhang et al., 2018).The final dilute factor of trace elements measurement was about 3000 after cooled (Zhang et al., 2018).The analysis accuracy was calibrated by the standard materials of plagiogneiss, andesite and slate (OU-6) (Zhang et al., 2018).The precision of most elemental determination was ±5-10% (relative) (Zhang et al., 2018).The Eu an (Eu anomaly) is calculated as Eu CN /(Sm CN ×Gd CN ) 0.5  (McLennan, 1989), in which CN is chondrite-normalized concentrations (Boynton, 1983).
A Bruker D8 advanced Phaser diffractometer was used to measure X-ray diffraction (XRD), using a 40 mA current and 40 kV voltage, CuKα radiation (k = 1.54 Å), and the detector is LynxEye.Powder samples were randomly oriented, and were taken with 2θ from 3°to 45°(step size of 0.02°).The isolation of clay mineral fraction of < 2 μm was performed by sedimentation based on Stoke's Law after removal of organic matter and carbonate by dissolution with 30% hydrogen peroxide and 0.3 N EDTA.The < 2 μm oriented smear slide was measured three times: (1) after air drying at room temperature (scanning from 2.5°to 15°2θ with a step size of 0.02°); (2) after ethylene-glycol solvation for 8 h (2.5-30°2θ with 0.02°steps); and (3) after heating at 460 °C for 2.5 h (2.5-15°2θ with 0.02°steps).

Geochemistry
Concentrations of major and trace element are shown in Tables 2 and 3.The elemental composition was normalized with upper continental crust (UCC) to determine the geochemical characteristics of the studied samples (Taylor and McLennan, 1985) (Figure 3).The Wufeng shales contain moderate to high SiO 2 (57.72-82.38 wt.%, av.= 68.84wt.%) and Al 2 O 3 (5.26-16.17wt.%, av.= 10.62 wt.%) compared to the UCC.These sediments display slightly higher average SiO 2 and TiO 2 concentrations, and lower average Al 2 O 3 , K 2 O, Fe 2 O 3 , MgO, CaO, MnO, and especially Na 2 O concentrations compared to the concentrations in the UCC (Figure 3).For large ion lithophile elements (LILE; i.e.Rb, Ba, Sr, U, and Th, the average concentrations of Rb, Ba and Th are similar to those in UCC.U are slightly higher than that of UCC in the studied shales, whereas Sr contents are obviously depleted relative to UCC.The transition trace elements, such as Co, Cr, Cu, V, and Ni) are enriched compared to UCC, whereas Sc display slightly depleted.The distributions of high field strength elements (HFSE) are different; Nb, Zr and Hf display slightly depletion, whereas Y is slightly enriched in the UCC-normalized diagram (Figure 3).
Generally, the rare earth elements (REE) concentrations are uniform for the studied samples.Total REE (∑REE) contents fall between 86.67 and 278.94 ppm in these sediments, with an average of 169.80 ppm, while the concentrations of HREE (from Ho to Lu), LREE (from La to Nd), and MREE (from Sm to Dy) range from 4.08 to 11.37 ppm, 72.78 to 240.20 ppm, and 10.00 to 34.15  (Taylor and McLennan, 1985).Note the strong depletion in Na and Sr, and the enrichment in transition metals for most samples.The red dotted line represents the average of the studied shales.ppm, respectively.Our samples display clear LREE (La-Nd) enrichment as indicated by high ratios of La/Yb = 9.92-18.73(av.= 13.77),La/Yb CN = 6.69-12.63(av.= 9.28) and ∑LREE/∑HREE = 14.01-27.28 (av. = 19.82).The obvious negative Eu anomalies (Eu an = 0.50-0.66,av.= 0.58) have been observed in these samples (Figure 4).The HREE display a relatively flat distribution, with Gd/Yb CN ratios falling between 1.35 and 2.41 (av.= 1.70).The PAAS-like shape of the CN REE distributions indicates homogenization of these samples (Figure 4).

Sorting and recycling of sediments
The Index of Compositional Variation (ICV), calculated as equation ( 1), is very useful to determine the clastic sediment composition and maturity (Cox et al., 1995).Oxides are expressed as weight percentages in this formula.Immature minerals, such as detrital ferromagnesian minerals and feldspars, contain ICV values of >1, whereas ICV values are smaller than 1 in the weathering products (muscovite, kaolinite, and illite) (Cox et al., 1995;Cullers and Podkovyrov, 2000).The studied samples display high ICV values, ranging between 0.97 and 1.74 (1.23 on average), suggesting that these samples were compositionally immature.These results indicate that the studied sediments in study area were likely dominated by first cycle sediments and were related to tectonically active settings (Cox et al., 1995;Cullers and Podkovyrov, 2000), according to the geological setting of South China (Su et al., 2006(Su et al., , 2009)).
The Zr/Sc versus Th/Sc diagram was useful tool for the determination of the presence of sorting-related fractionations (McLennan et al., 1993).The data of the samples concerned in this study are varied around the "compositional variations" line and are clustered together (Figure 5(a)), indicating that they are first-cycle deposits without recycling processes, consistent with ICV values.

Provenance of Wufeng shales
Before using immobile elements to discriminate the provenance, element mobility in the samples concerned in this study during the processes, such as chemical weathering, sorting, and/or others, should be evaluated (Singh, 2009).Chemically mobile elements display a wider scatter, while chemical immobile elements will keep constant and show a linear array extending from the origin along radians.(Fralick and Kronberg, 1997).A linear arrangement of samples illustrates the immobility of the elements in the studied samples (Fralick and Kronberg, 1997).In the plots of SiO 2 -Al 2 O 3 , -Fe 2 O 3 , and -TiO 2 (Figure 5(b)-(d)), the linear arrays of points along lines extending toward 100% SiO 2 indicate that these elements were immobile in chemistry.These immobile elements can be used to discriminate the source rocks because TiO 2 /Al 2 O 3 ratios vary in different igneous rocks (acidic > intermediate > mafic > ultramafic) and Ti and Fe are enriched in mafic minerals (Nesbitt and Wilson, 1992;Sugitani et al., 1996).In the plots of TiO 2 /Al 2 O 3 versus TiO 2 /Fe 2 O 3 and TiO 2 /Al 2 O 3 versus Fe 2 O 3 /Al 2 O 3 , the elements from the most studied samples distribute close to the average compositions of andesites (Figure 6), indicating that the mainly source was andesites or rocks with similar composition.
The Eu anomalies and the CN REE distributions can be used to discriminate the provenances of sediments (McLennan et al., 1993;Fedo et al., 1996).Felsic rocks generally show negative Eu anomalies and higher ratios of LREE/HREE, however, basic rocks usually display no Eu anomalies with low ratios of LREE/HREE (Cullers and Graf, 1983;Taylor and McLennan, 1985;Luo et al., 2015;Qiao et al., 2021;Radwan, 2022).The consistent CN REE distributions illustrate that they were likely to be mainly sourced from felsic source rocks, as suggested by negative Eu anomalies, enriched LREE, and relatively flat distributions of HREE (Figure 4).Hayashi et al. (1997) proposed that the relationship between TiO 2 and Zr could be used to discriminate sediment provenance.The strongly positive correlation between TiO 2 and Zr suggests a homogeneous source, and the samples are plotted in the felsic igneous rock field (Figure 7(a)), showing a predominantly felsic source for the studied shales.The studied samples contain low and constant ratios of La/Th, ranging from 2.21 to 4.32.The Hf contents span between 1.95 and 5.98 ppm (3.87 ppm on average; Table 3).Ti/Zr and Co/Y ratios concerned in this study are 0.18-1.23 and 13.95-36.01,respectively.In the plots of La/Th versus Hf (Floyd and Leveridge, 1987)  decentralized La/Sc ratios, also indicating an intermediate to felsic source components (Figure 7(d)).In the diagrams of Sc-La-Th and Co-Th-Hf, the studied samples concerned in this study distribute around the average composition of TTG and granites (Figure 8).
In conclusion, the source rocks of the Wufeng shales were dominated by TTG-like, andesitic and granitic igneous rocks, which were only present in the older Proterozoic basement in this area (Chen et al., 1995;Yan et al., 2008;He et al., 2010), indicating that the Wufeng shales were derived from the Proterozoic basement.Hayashi et al., 1997), note that the studied samples mainly distribute in the felsic igneous field; (b) La/Th versus Hf (after Floyd and Leveridge, 1987), note that all samples distribute close to the TTG and felsic volcanic rocks; (c) Co/Y-Ti/Zr, note that the samples mainly plot around TTG and andesites; (d) La/Sc versus Co/Th (Gu et al., 2002), note that these samples distribute around the TTG and granites.The average compositions of Phanerozoic granites, TTG, andesites and basalts are cited from Condie (1993).
The data of Wufeng Formation is plotted in the A(Al 2 O 3 )-CN (CaO* + Na 2 O)-K (K 2 O) diagram (Figure 9(a)).This ternary can be used to evaluate the degree of K-metasomatic effects and chemical weathering conditions (Nesbitt and Young, 1984;1989;Fedo et al., 1995), indicating that the removal of Na and Ca was intermediate extent because of plagioclase destruction with the data plotting intermediate between plagioclase-K feldspar and A-K lines.The plots do not deviate from the predicted weathering trend line and exhibit no trends toward the K apex, indicating that K-metasomatism did not happen in these sediments during diagenesis.This trend may be the nature of non-steady state weathering related to the different conditions and different provenance of balance between physical (such as erosion, tectonism, uplift) and chemical processes (Nesbitt et al., 1997).The vertical dimension (percent Al 2 O 3 ) in A-CN-K diagram is in agreement with CIA values, also suggestive of low to moderate chemical weathering conditions.Harnois (1988) proposed chemical index of weathering (CIW) which can be calculated as equation (3).Like CIA, CIW also can be used to monitor the degree of conversion from feldspars to clay minerals (Fedo et al., 1995;Maynard et al., 1995).This parameter is similar to CIA, even though it eliminates K 2 O. CaO*, in the formula of CIW and PIA, is it equivalent to that in the formula of CIA.    3) also indicate low to moderate chemical weathering conditions of the provenance area (Nesbitt and Wilson, 1992;Fedo et al., 1996).The dominant clay minerals are illites and mixed-layer clays in the studied sediments, which also supports the low to moderate chemical weathering conditions (Nesbitt and Young, 1982).The intensity of chemical weathering conditions is primarily determined by precipitation and temperature, and a warm and wet climate condition is stimulative for chemical weathering condition (Nesbitt and Young, 1982;Riebe et al., 2004).It is worth noting that CIA, CIW, and PIA gradually decrease from the bottom toward the top part of the Wufeng sediments (Figure 10).On the other hand, the intense paleoweathering conditions are characterized by the abundance of kaolinite and chlorite, and the paucity of particularly plagioclase, detrital feldspars (Nesbitt and Young, 1982;Schoenborn and Fedo, 2011;Jian et al., 2013).For the studied samples, the kaolinites and chlorites display decreasing trends with decreasing depth, similar to the stratigraphic variations of CIA and PIA, however, the plagioclase contents display the reverse trend (Figure 10), suggesting that the chemical weathering is gradually decreasing at the Late Ordovician and humidity seems to have decreased from the basal part of the Wufeng Formation.(Nesbitt and Young, 1984).Note that all studied samples distribute along the line parallel to the line A-CN; (b) Ternary plot of (Al 2 O 3 -K 2 O)-CaO*-Na 2 O ((A-K)-C-N) (Fedo et al., 1995).Note that the low to moderate values of CIA and PIA, indicating low to moderate chemical weathering of the source area.
The intensity of the chemical weathering controlled the mobility of elements in the provenance region.Compared to immobile element (Rb, Cu, and Al), sequestered in the residual phases K, Na, Ca, Mg, and Sr are regarded as mobile elements during the process of weathering (Nesbitt andYoung, 1984, 1989) 10).These results also indicated that the weathering intensity in the provenance region decreased gradually.The stratigraphic variations of CIA, PIA, elements ratios and mineralogy indicate that the chemical weathering degree decreased from moderate to low in the provenance region at Katian, suggesting a gradual cooling trend of the climate at Late Ordovician.

Conclusions
The Wufeng shales contain intermediate to high SiO 2 and Al 2 O 3 , show low to moderate depleted in Sr, Na 2 O, and, CaO, and are rich in transition metal elements and Y compared to their concentrations in the UCC.The Wufeng shales might be the product of first-cycle deposits and are related to the tectonically active settings, as indicated by ICV values and the diagram of Zr/Sc versus Th/Sc.The Wufeng shales were derived from a mixture of felsic-intermediate provenances including TTG-like, granitic and andesitic igneous rocks.The source area experienced low to moderate chemical weathering.The paleoclimate gradually cooled during Late Ordovician.

Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Figure 2 .
Figure 2. Location and paleogeographic map of the study area in South China.

Figure 3 .
Figure 3. UCC-normalized major and trace elements for the Wufeng shales(Taylor and McLennan, 1985).Note the strong depletion in Na and Sr, and the enrichment in transition metals for most samples.The red dotted line represents the average of the studied shales.

Figure 4 .
Figure 4. CN REE distributions of the wufeng sediments.PAAS values (after Taylor and McLennan 1985) are shown as a red line.Values of.Note negative Eu anomaly, fractionated LREEs, and flat HREEs, similar to the CN (chondrite data after Boynton, 1983) PAAS.

Figure 5 .
Figure 5. (a) Th/Sc versus Zr/Sc plot for the Wufeng sediments (after McLennan et al. 1993).Note that the samples plot along the "compositional variations" line and cluster together, suggesting no obvious influence of recycling and heavy mineral sorting.(b-d) Major oxides versus SiO 2 for the Wufeng shales.Note the strong negative correlations for SiO 2 versus Al 2 O 3 , Fe 2 O 3 , and TiO 2 .

Figure 7 .
Figure 7. (a) TiO 2 versus Zr (afterHayashi et al., 1997), note that the studied samples mainly distribute in the felsic igneous field; (b) La/Th versus Hf (afterFloyd and Leveridge, 1987), note that all samples distribute close to the TTG and felsic volcanic rocks; (c) Co/Y-Ti/Zr, note that the samples mainly plot around TTG and andesites; (d) La/Sc versus Co/Th(Gu et al., 2002), note that these samples distribute around the TTG and granites.The average compositions of Phanerozoic granites, TTG, andesites and basalts are cited fromCondie (1993).

Figure 8 .
Figure 8. La-Th-Sc and Th-Hf-Co ternary plots.Note the studied samples plot around to TTG and granites.The average compositions of Phanerozoic granites, TTG, andesites and basalts are cited from Condie (1993).

Figure 9 .
Figure 9. (a) Ternary plot of Al 2 O 3 -(CaO*+Na 2 O)-K 2 O (A-CN-K)(Nesbitt and Young, 1984).Note that all studied samples distribute along the line parallel to the line A-CN; (b) Ternary plot of (Al 2 O 3 -K 2 O)-CaO*-Na 2 O ((A-K)-C-N)(Fedo et al., 1995).Note that the low to moderate values of CIA and PIA, indicating low to moderate chemical weathering of the source area.
. Thus, Rb/Sr, Sr/Cu, Na 2 O/Al 2 O 3 , and CaO/Al 2 O 3 are useful for the evaluation of the variations of chemical weathering.The higher Rb/Sr, and lower Sr/Cu, Na 2 O/Al 2 O 3 , and CaO/Al 2 O 3 values generally imply a stronger chemical weathering.In the studied samples, Rb/ Sr values show a decreasing trend with decreasing depth, whereas the Sr/Cu, Na 2 O/Al 2 O 3 , and CaO/Al 2 O 3 ratios display the opposite trends (Figure
b "\" represents that the minerals are absent in the samples.

Table 3 .
Trace elements (in ppm) of the Wufeng shales.
Table3 and Figure 9(b)), indicating chemical weathering degree was low to moderate in the provenance region.Such conclusion is consistent with the results of CIA and CIW values.CIA values are positively related to PIA and CIW values (r 2 = 1.00 and 0.99, respectively).In the UCC normalized multielement diagram, the lowly to moderately depleted CaO, Na 2 O and Sr and slightly enriched Ba and Rb (Figure