In the QD1 drilling core, the asphalts are filled in the fractures (Figures 3A,B) or as thin layers in the Jurassic mudstone (Figures 3C–E), and an oil slick occurs when soaked in water (Figure 3F). The fresh asphalt is bright black with visible conchoidal fractures (Figure 3G). Under the microscope, the asphalt contains three obvious groups of macerals—vitrinite, exinite, and inertinite. Overall, the macerals are mainly composed of vitrinite, especially the collinite (Figure 3H) and telinite (Figure 3K), which appear dark gray under oil-immersed reflected light and brownish-black under fluorescence. The exinite components are rich and show grayish black under oil-immersed reflected light (Figures 3H,J,K) and orange to bright yellow under fluorescence. The exinites are dominated by sporophyte (Figure 3I), cutinite (Figure 3M), and resinite (Figure 3L). The inertinite components are less abundant with fusinite and semifusinite (Figure 3K) and show grayish-white under oil-immersed reflected light (Figure 3H) and black under fluorescence (Figure 3I). Some exsudatinites are visible in cracks and residual pore cavities of the microcomponent (Figure 3M), and even fossilized plant leaf lobes (Figure 3J) and flow morphology of macerals are visible (Figures 3J,L). Therefore, the Jurassic asphalt retains the coal macerals and is related to the Jurassic coal.

In the 25 samples (Table 1), the total organic carbon (TOC) of J₁k and J₂y are both greater than 1.0% (1.17–23.00%), making them good to excellent source rocks (Figure 4). The J₁k samples are poor to excellent source rocks with a pyrolysis parameter (S₂) ranging from 2.25 to 71.25 mg HC/g rock. The J₂y samples have higher quality than the J₁k source rock with an S₂ ranging from 3.40 to 33.83 mg HC/g Rock. From the T_max vs. HI crossplot (Figure 5), J₂y source rocks have three kerogen types while J₁k has kerogen of the II and Ⅲ types. Based on the comprehensive pyrolysis parameters, the J₂y source rocks have better quality than J₁k. The testing samples have an Ro from 0.33% to 0.57%, which indicates an immature to low maturity stage. Two Ro values of asphalt are 0.41% and 0.54%, close to the J₁k Ro values (Ro = 0.38%–0.57%, average = 0.44%) and above the J₂y Ro values, which range from 0.33% to 0.41% with an average of 0.37%.

The total rare earth element (REE) concentrations of J₂y mudstone are the highest, from 275.59 to 317.45 μg/g (Table 2), followed by those of the J₁k mudstone at 139.20–143.77 μg/g. J₁k coal has an REE concentration of 68.71 μg/g, which is slightly higher than the asphalt (35.88–41.92 μg/g). Light rare earth elements (LREEs) are more enriched in all samples than heavy rare earth elements (HREE), and the LREE/HREE ratios of J₂y mudstone are from 9.27 to 9.70, which is higher than the 6.37–6.95 of J₁k mudstone, higher than the 4.73 of J₁k coal, and higher than the 1.45–2.04 of the asphalt. The chondrite normalization values are after Boynton (1984), and all of the chondrite-normalized REE curves show the right-sloping characteristics with LREE enrichment. However, the curves gradually flatten from the J₂y mudstone to J₁k mudstone to J₁k coal and then to asphalt (Figure 6). All the samples show pronounced negative Eu anomalies. The δEu values of asphalt value vary from 0.84 to 0.92, close to those of J₁k mudstone and coal, ranging from 0.83 to 0.89, much higher than J₂y values of 0.70–0.80. The Ce anomalies are weak and relatively close, with slightly lower δCe values of asphalt, ranging from 0.90 to 0.98.

The J₂y/Ho ratios vary from 23.80 to 27.50, which is lower than those of J₁k mudstones (30.67–31.50) and asphalt Y/Ho ratios of 30.04–44.69, which are close to the 36.3 of J₁k coal (Table 2). The (Ce/Yb)_N, (La/Yb)_N, (La/Sm)_N, and (Gd/Yb)_N ratios of J₂y mudstone are 8.84–9.11, 11.39–11.95, 3.32–3.47, and 2.22–2.39, much higher than those of J₁k samples, which are 3.18–5.96, 4.49–7.63, 2.90–3.57, and 1.00–1.78, respectively. The corresponding ratios of asphalt samples were 0.75–0.95, 1.15–1.17, 1.08–1.15, and 0.77–1.07, respectively.

The Jurassic alkanes are primarily distributed from n-C₁₂ to n-C₃₅, with n-C₂₅ as the single dominant n-alkane on the m/z 85 fragmentation diagram (Figure 7). The n-alkane composition has a distinct odd-even dominance with the odd n-alkane abundance higher than the even ones. The J₁k samples have an OEP (2.22–2.47) and CPI (1.38–1.75) similar to those of asphalt samples (OEP: 2.78–4.09; CPI: 1.84–2.30) and significantly lower than those of J₂y samples, which have OEP values of 5.02 and 5.19 and CPI values of 3.04 and 3.10, respectively (Table 3). The Pr/Ph ratios of Jurassic samples are all higher than 2.5, and the Pr/Ph values of the two J₂y mudstones are 3.01 and 3.62, slightly higher than that of the J₁k mudstone (2.85). However, the Jurassic mudstone Pr/Ph values are significantly lower than those of J₁k coal (5.12) and asphalt (5.09–5.85).

Quantified using m/z 191, the relative abundances of tricyclic and tetracyclic terpenoids are generally low, while the pentacyclic triterpenes are abundant, especially the C₂₉ to C₃₂ hopanes (Figure 7). The Ga/C₃₀H ratios range from 0.03 to 0.09 for all Jurassic samples (Table 3). The Ts/(Ts+Tm) values of the J₁k coal and mudstone are 0.78 and 0.90, respectively, significantly higher than those of the two J₂y mudstones (0.44 and 0.56). The asphalt extracts also have high Ts/(Ts+Tm) ratios, 0.52 and 0.56 for the two samples, respectively.

From the m/z 217 sterane compounds (Figure 7; Table 3), C₂₇, C₂₈, and C₂₉ conventional steranes are distributed in a “V” shape with the abundance of C₂₉αααR sterane higher than C₂₇αααR and C₂₈αααR steranes. The asphalt extracts have a slightly higher proportion of C₂₉ steranes and a lower proportion of C₂₈ steranes, distinguished from the similar sterane compositions of the J₁k and J₂y. Meanwhile, the Jurassic extract values from the two maturity indexes of C₂₇ (DSt/St) and C₂₉ 20s/(20s+20r) are close. The J₁k C₂₉ ββ/(αα+ββ) values are 0.51 and 0.46, higher than two J₂y extracts (0.27 and 0.38) and close to two asphalt values of 0.52 and 0.50.

The Jurassic asphalts have low Re and Os contents (Table 4). The Re content is between 0.82 and 11.36 ppb, with Os content varying from 26.82 to 131.85 ppt (Table 4), which is different from the Re and Os composition of the continental crust (Re content less than 1 ppb, and Os content less than 50 ppt). The ¹⁸⁷Re/¹⁸⁸Os values of asphalt vary from 58.7 to 576.1, while the ¹⁸⁷Os/¹⁸⁸Os ratio is in the range of 1.56–3.09 (Table 4). The linear fitting between ¹⁸⁷Re/¹⁸⁸Os and ¹⁸⁷Os/¹⁸⁸Os is relatively high with an R ² of 91.8%, and the asphalt forming age is calculated to be 192 +30/−45 Ma with the isochrone slopes (Figure 8).

The rare earth elements have very similar and highly stable chemical properties that are not easily affected by metamorphism and can be used to determine the depositional environment, type, and origin of source rock (Condie, 1991; Tribovillard et al., 2006). The J₂y mudstones have significantly higher REE content and (La/Sm)_N and (Gd/Yb)_N ratios than the J₁k, indicating that the J₂y deposited with higher LREE and HREE fractionation under a warmer and wetter environment (Fleet, 1984). The J₁k coal and asphalt have similar REE content, reflecting a similar sedimentary environment. The J₂y REE curves appear to have steeper right-sloping characteristics than the J₁k ones (Figure 6) and have higher ratios of (La/Yb)_N and (Ce/Yb)_N, with the LREE higher than HREE. Compared to braided-river delta and shallow coastal lakes of J₁k sedimentary environment close to the sediment source (Cheng et al., 2008), the J₂y distal sediments have fine particles and strong adsorption with the REE characteristics mentioned above. The δCe values of all samples are relatively close, with similar oxidizing properties of depositional environments (Elderfield and Pagett, 1986). However, the δEu values of asphalt and J₁k samples are significantly higher than the J₂y ones, reflecting the differences in sedimentary waters (Hass et al., 1995). The rare earth elements Y and Ho have very similar geochemical properties, but Y has an unusually low affinity for Fe oxides compared to Ho (Gong et al., 2021). It is clear that the J₂y Y/Ho values are higher than those of the asphalt and J₁k samples, indicating the J₂y depositional environment was more reduced. Therefore, combined with the REE partition curves and ratio relationships (Figure 6; Table 2), the asphalts have a closer relationship with the J₁k, especially the J₁k coal, suggesting that they are more closely related.

The Jurassic biomarker compounds show the characteristics of high Pr/Ph, high C₂₉ steroid, and low abundance of tricyclic terpene alkanes, highly consistent with the Jurassic source rocks studied by previous researchers (Hendrix et al., 1995). The asphalt extracts have similar Pr/Ph ratios with J₁k coal, which suggests they have similar depositional environment oxidizing properties and closer affinities. The peak alkane is n-C₂₅ for all samples (Figure 7; Table 3), reflecting that the regional Jurassic organic matter sources are highly related to terrestrial sources. In terms of the m/z 217 sterane compounds (Table 3), the J₂y extracts have similar C₂₇ regular sterane content with the J₁k and slightly higher C₂₇/C₂₉ regular sterane ratio and slightly lower C₂₈/C₂₉ regular sterane ratio, indicating that the origins of J₂y organic matter have more algae (Volkman, 1986). The CPI and OEP of the Jurassic source rocks and asphalt are all greater than 1.0, suggesting that their thermal maturity is low. In contrast, the J₁k CPI and OEP values are lower than those of the J₂y (Table 3) and close to those of asphalt, indicating that the asphalts have similar thermal maturity to the J₁k. By the maturity-related biomarker parameters of Ts/(Ts+Tm), C₃₁22S/(22S+22R), and C₃₂22S/(22S+22R), the J₁k and asphalt have proximate maturity with biomarker parameters higher than those of the J₂y. By the C₂₉ ββ/(αα+ββ) ratio, the J₁k has similar thermal maturity with asphalt but slightly higher than the J₂y extracts. Therefore, the Jurassic asphalts are more closely related to the J₁k.

Even if the source rock underwent hydrocarbon generation and expulsion processes, the Re-Os system can still remain closed and record its sedimentary age (Creaser et al., 2002). However, the isochronous age of asphalt Re-Os isotopes may reflect the sedimentary age of source rock or hydrocarbon migration/accumulation age, which mainly depends on the closure of the Re-Os isotopic system. If the Re-Os isotopic system is not closed with different Re and Os isotopes evenly mixed, then the asphalt Re-Os isochronous age represents the migration/accumulation age (Selby and Creaser, 2005). Using the plots of ¹⁸⁷Os/¹⁸⁸Os versus ¹⁸⁷Re/¹⁸⁸Os, the age of the Jurassic asphalt is 192+30/−45 Ma, similar to the J₁k deposition time, suggesting that the asphalt is closely related to J₁k strata. It is obvious that the J₂y asphalt retains the Re-Os characteristics of J₁k coal, and the Re-Os isotope system is closed during the entire hydrocarbon generation and migration process.

More importantly, from the macerals, the Jurassic asphalt is mainly precipitated from exinite-rich components such as sporophytes, alginate, and resinite during the coalification. The asphalt precursor was mixed with vitrinite, inertinite, and other coal macerals and sometimes, flow traces are visible under the microscope (Figures 3I,M). They were formed during the early coalification equivalent to the oil window (Stach et al., 1982) and then filled in Jurassic pores and fractures.

Coal is not only an important gas-source rock but also an oil-source rock, which has been confirmed by the petroleum exploration practices in western Chinese basins, such as the Tuha Basin (Zhao and Cheng, 1998). Generally, these coals were deposited in a deltaic inter-bay swamp (Shao et al., 2003). The J₁k sedimentary facies were braided-river delta and shallow coastal lakes, where ferns and other herbaceous plants flourished (Vakhrameev, 1991). Undecomposed hydrogen-rich components carried by the river accumulated here. The enrichment of detrital exinite, coupled with the strong micro-biolysis of peat swamps, resulted in coals with high hydrogen content, which laid a good environmental foundation for coal-derived hydrocarbons.

The Jurassic has an unconformable contact relationship with the Paleozoic basement and steadily developed in the Early and Middle Jurassic. Overall uplifting occurred in the Late Jurassic, which resulted in the absence of the Middle and Upper Jurassic in the Ruoqiang Sag. In the Jurassic, the paleo-temperature gradient was not high, 2.8°C/100 m (Li et al., 2005), and the J₁k source rock does not reach the Ro 0.5% with a maximum burial depth of only 1840 m in the QD1 well (Figure 9).

The Cretaceous in the Ruoqiang Sag is only a localized remnant even though it is 100 m thick in the Tadong and Tangguzibas Depression due to the uplifting in the study area and strong denudation of the pre-deposited strata in the late Cretaceous (Yang et al., 2001; Arnaud et al., 2003). During the Cretaceous, the J₁k source rock underwent continued subsidence and uplifting. Its maximum burial depth of 2029 m did not meet the hydrocarbon threshold. The J₁k source rock did not mature with a paleo-temperature gradient of 2.7°C/100 m (Li et al., 2005).

The regional Paleogene exhibits weakly extensional sedimentation of sandstone and mudstone (Li, 1998), while the Himalayan movement that occurred at the end of the Paleogene resulted in a large-scale withdrawal of the lake water in the Ruoqiang Sag (Li et al., 2002), and the Paleogene near the QD1 well was also completely eroded. In the Miocene, affected by large-scale thrusting activities in the West Kunlun tectonic belt, the Ruoqiang Sag underwent flexural subsidence and deposited a thick Neogene (Tian et al., 2023). With a low paleo-temperature gradient of 2.1°C/100 m (Li et al., 2005), the J₁k burial depth met the hydrocarbon threshold with a maximum burial depth of 2500 m and Ro 0.5%, and J₁k coal began early hydrocarbon generation evolution (Figure 9). The coal macerals began the asphaltenization stage of coalification, and the exsudatinites filled in the fractures and pores, or were preserved in situ, or were transported to the overlying J₂y mudstone under the presence of faults, where they were accumulated, were preserved, and formed an asphalt-like appearance. After that, the Altyn Tagh Fault underwent a large-scale strike-slip activity, and the deposited Neoproterozoic was completely denuded. With the rapid rise of the Tibetan Plateau after the Pleistocene, the Quaternary was deposited only several meters thick near the QD1 well.

After the J₁k hydrocarbon evolution, the asphalt was formed from the Miocene and maintained a similar REE composition as the J₁k coal. The rhenium–osmium isotopes in the asphalt had not been sufficiently fractionated, so the ages of asphalt were consistent with the J₁k. Meanwhile, the rhenium and osmium contents indicated that they came from a source rich in organic matter rather than crustal mixing during the transport process (Sun et al., 2003; Selby et al., 2007). Due to the late formation of asphalt, the 25-norbornanes were rarely detected or not detected in the m/z 191 composition (Figure 10). The 25-norhopanes are believed to originate from the loss of a methyl group from C-10 in hopanes and have a high abundance of biodegradation (Volkman et al., 1983). C₂₉ 25-nor-17α-hopane (D29) from the C₃₀ hopane has very low content of both m/z 191 and m/z 177 ions, and the ion content values of D30 and D31 from C₃₁–C₃₂ 17α-hopane (22S + 22R) are only slightly higher (Figure 10). The asphalt did not undergo many secondary changes, such as oxidization and biodegradation, during its formation process.

Relying on samples of the QD1 well and outcrops in the Ruoqiang Sag, the Jurassic asphalt is confirmed to be derived from the J₁k coal and serves as the product of hydrocarbon generation and evolution in the early maturity of J₁k coal. The current understanding diverges considerably from the previous understanding of regional oil sources (Xu et al., 1994; Chen and Hu, 1996; Huang et al., 2013), but the current evidence is much stronger and more robust. Meanwhile, in the Tula Basin (Figure 1B), which is faulted by the Altyn Tagh Fault with Ruoqiang Sag, oil sandstones more than 100 m thick with a large amount of asphalt veins were found in the Upper Jurassic (Guo et al., 1998; Robinson et al., 2003) and were considered to be similar to the Jurassic oils in the Qaidam Basin by previous researchers, whose source rocks are also from the Middle and Lower Jurassic (Zeng et al., 2022). This undoubtedly increases the confidence in the Jurassic oil and gas exploration in the Ruoqiang Sag.

Although the understanding of the low maturity evolution degree of J₁k source rock is unfavorable for the regional Jurassic petroleum exploration, it confirms that the J₁k coal-bearing source rock has a very good hydrocarbon generation potential. Furthermore, the sampling study area is located in the premontane area at the southern margin of the Ruoqiang Sag (Figure 1). The current morphology of the Ruoqiang Sag is generally characterized by a northeast-oriented belt that is higher in the south and lower in the north (Figure 11). The Jurassic has a greater burial depth away from the premontane area and develops with a large area of J₁k and J₂y lacustrine deposits (Cheng et al., 2008; Jiang et al., 2014). From the seismic section (Figure 11), the Lower Jurassic strata are buried more than 3000 m deep and are hundreds of meters thick. Therefore, the J₁k coal-bearing source rock has matured and could provide hydrocarbons for the Ruoqiang Sag. The J₁k and J₂y have developed sedimentary facies such as deltas and shallow coastal lakes, which are favorable for the sandstone reservoirs, while the J₂y mudstone has pure lithology and large thickness and could be a very good caprock. From the analysis of the Jurassic lithological assemblage, the Ruoqiang Sag has developed a good reservoir-seal assemblage. Therefore, fault-related tectonic and lithologic traps near the hydrocarbon stove are targets for regional Jurassic petroleum exploration.