Oil-Source Correlation in the Ordovician and Carboniferous Strata from the Southwestern Ordos Basin (China): Implications for a Multivariate Statistical Approach

Abstract

Recently, oil and gas shows have been identified in both the Carboniferous Yanghugou Formation and the Ordovician Wulalike Formation in the southwestern Ordos Basin. However, oil-source correlation in this area is challenging because conventional biomarker parameters are inherently limited, and the two source rock systems exhibit pronounced maturity differences. As a result, traditional biomarker-based approaches often suffer from ambiguity and low reliability. To address these limitations, this study introduces multivariate statistical methods, including hierarchical cluster analysis (HCA) and Q- and R-mode factor analysis (FA). A total of 31 biomarker parameters from 47 source rocks and crude oil samples collected from Wells YT1, YT2, and YT3 were systematically integrated. On this basis, four composite indices were established: the Maturity Index (MI), Organic Matter Origin Index (OMOI), Water Salinity Index (WSI), and Organic Matter Source Index (OMSI). These indices were applied to construct oil-source correlation diagrams, thereby reducing the influence of single-parameter limitations and maturity differences on oil-source identification. The results indicate that the Yanghugou Formation source rocks are characterized by high organic matter abundance (average total organic carbon (TOC) of 4.29%), low maturity, and Type II₂ kerogen deposited in a paralic facies, whereas the Wulalike Formation exhibits lower TOC values (average 0.31%) and represents highly mature Type I marine source rocks. Multivariate statistical analysis shows that the biomarker characteristics of the Yanghugou oil sands are intermediate between those of the two source rock systems. The MI-OMOI diagram further demonstrates that the Yanghugou oil sands have a mixed origin, with contributions from both the Yanghugou and Wulalike Formations, while the Wulalike oils are predominantly self-sourced and self-reservoired. Overall, the composite indices method established in this study effectively improves oil-source identification accuracy under conditions of strong maturity contrast and provides new insights for hydrocarbon exploration in structurally complex areas.

1. Introduction

Oil and gas exploration and development in the central Ordos Basin have reached a near-mature stage. In response to China’s increasing demand for hydrocarbons and the need for sustained long-term production, exploration efforts have progressively shifted toward the basin margins. − In recent years, continued exploration by China’s Changqing Oilfield has led to a series of breakthroughs in the southwestern Ordos Basin. Notably, oil and gas shows have been identified in the Upper Carboniferous Yanghugou Formation and the Middle Ordovician Wulalike Formation, both previously regarded primarily as gas-bearing intervals. Well YT2 encountered oil within the Yanghugou Formation, whereas Well YT3 identified a 0.9-m oil layer and a 3.4-m oil–water transition zone in the Wulalike Formation. These results demonstrate the crude oil accumulation potential of the Yanghugou Formation and confirm the presence of hydrocarbons in the underlying Wulalike Formation. Furthermore, Well YT3 tested daily production rates of 5.3 t of oil and 1013 m³ of gas from the Wulalike Formation, marking the first industrial Paleozoic oil flow discovered in the Ordos Basin. Collectively, these discoveries highlight the significant exploration potential for Paleozoic crude oil in the basin margin (Figure A).

1.

Regional geological map of the Ordos Basin and study area location. Adapted with permission from ref . Copyright 2002 Elsevier Science B.V. (A) Regional overview of the Ordos Basin, China and (B) structural map of the southwestern Ordos Basin.

Nonetheless, hydrocarbon exploration in this area remains challenging because the southwestern basin margin is distant from the basin center and is located within an east–west tectonic convergence zone characterized by frequent tectonic activity, well-developed fractures, and complex structural deformation (Figure B). − Second, the region remains understudied, as evidenced by the unclear organic geochemical characteristics of its crude oil and the absence of commercially viable hydrocarbon flows to date, leaving the petroleum source uncertain. In this study, we focus on the Yanghugou Formation and the Wulalike Formation, systematically evaluating the source rock and crude oil characteristics through a series of geochemical analyses to clarify the origins of crude oil in both formations. Biomarker analysis plays a particularly crucial role in this process. It reflects key characteristics of sedimentary organic matter, including age, source rock type, depositional environment, and thermal maturity. These analyses provide key evidence for identifying reservoir associations, tracing hydrocarbon migration pathways, and determining favorable exploration targets. However, in practical applications, several challenges were encountered.

The central challenge of oil-source discrimination in the study area is the pronounced thermal maturity contrast between the two target stratigraphic intervals, which constrains the application of conventional biomarker compound analysis. In general, biomarker parameters are controlled by multiple factors, and no single parameter can independently and reliably represent a specific geochemical attribute. For example, the Pristane/Phytane (Pr/Ph) ratio, although commonly used to infer depositional redox conditions, is also influenced by thermal maturity, water salinity, and organic matter input. − Similarly, classical maturity indicators such as C₂₇18α­(H)-trisnorhopane/C₂₇17α-trisnorhopane (Ts/Tm), and Ts/(Ts + Tm) are frequently affected by source dependency and clay mineral catalysis. ,, Therefore, a multiparameter integrated assessment is required, providing the rationale for the composite indices developed in this study.

Moreover, many biomarker parameters are valid only within the restricted maturity windows. Their relative abundances generally decrease as thermal evolution advances. For instance, 17α­(H)-hopanes (C₂₉–C₃₃) progressively decline at advanced maturity levels, requiring cautious application in regions characterized by large maturity variations.

Based on our experimental measurements, vitrinite reflectance (R _o) data indicate that Yanghugou source rocks are at a low-maturity stage (0.55–0.68%), whereas Wulalike source rocks range from high- to overmature, with R _o values of 1.05–2.09% (average 1.44%). This pronounced separation in maturity across the hydrocarbon generation window leads to fundamentally different biomarker responses between the two formations such that parameters effective in the low-maturity Yanghugou Formation may become unreliable when applied to the high-maturity Wulalike Formation.

A representative example is the sterane maturity index C₂₉ 20S/(20S + 20R), , which may lose interpretive significance or even exhibit reversal after peak oil generation (R _o ≈ 0.9%) as reaction equilibrium is approached. , Consequently, identical parameter values may correspond to different maturity levels, resulting in nonunique interpretations. Such inconsistencies are also reflected in diagram-based analyses, where sterane ternary plots indicate mixed-source characteristics, whereas n-alkane-isoprenoid distributions suggest a single oil source. These discrepancies indicate that conventional biomarker approaches alone are insufficient for reliable oil-source discrimination.

To improve the reliability of the oil-source correlation, this study applies multivariate statistical analysis of biomarker parameters, whereby parameters representing the same geochemical dimension are evaluated collectively and integrated into comprehensive discrimination schemes. This approach reduces the limitations of individual parameters and minimizes the influence of maturity heterogeneity, allowing for more robust characterization of unknown oil samples.

Although cluster analysis has been widely applied in oil-source studies of the Yanchang Formation, it has commonly been used as a post hoc validation tool and yields relatively coarse results. In this study, cluster analysis is combined with a more refined factor analysis and traditional oil-source correlation methods to investigate the crude oil sources of the Yanghugou and Wulalike formations. This integrated methodology provides a practical framework for resolving ambiguous oil sources in the southwestern Ordos Basin and offers implications for future hydrocarbon exploration along the basin margin.

2. Geological Setting

The Ordos Basin, situated in central China, is the country’s second-largest sedimentary basin, covering 3.7 × 10⁵ km². This multicycle cratonic basin comprises a Meso-Cenozoic inland lacustrine basin superimposed on a Late Paleozoic marine basin. , Structurally, the basin comprises six first-order tectonic units: the Yimeng Uplift, the Weibei Uplift, the Jinxi Fault-Fold Belt, the Yishan Slope, the Tianhuan Depression, and the Western Fold-Thrust Belt.

The western Ordos Basin corresponds to the western part of the Tianhuan Depression. It extends in a north–south direction from Table Mountain in the north to the Pingliang area in the south with an approximate length of 600 km and a width of about 100 km (Figure B). The area is located in the northern segment of the Baikal-East Indian Ocean tectonic belt. Its geological structures and geophysical fields are highly complex due to the combined influence of the surrounding Alxa microcontinental block and the Qinqi tectonic belt. ,, It is the intersection and superposition of tectonic units of different natures, with a low degree of exploration and high research difficulty. , The Paleozoic strata of Cambrian, Ordovician, Carboniferous, and Permian were developed from the bottom up in the West Rim area, in which the Silurian and Devonian periods caused sedimentary discontinuities due to the closure of the argillaceous troughs as a result of the Caledonian orogeny. , The stratigraphic sedimentary facies in the study area exhibit significant vertical and lateral variations. During the Early Paleozoic, marine carbonate strata were predominantly deposited. In the Late Paleozoic, deposition shifted to alternating marine-continental clastic systems, whereas during the Mesozoic the environment further evolved into typical fluvial and lacustrine settings (Figure ).

2.

Sedimentary facies columnar diagram of the Yanghugou and Wulalike Formations in Well YT3. Adapted with permission from ref . Copyright 2025 John Wiley & Sons Ltd.

The study area is located in the Yindongzi region of the southwestern Ordos Basin. It is bounded by the Laochigou area to the north, the Shajingzi area to the east, the Yindong-Guanzhuang area to the south, and the Weizhou-Anguo Fault to the west. Four major north–south faults develop from east to west across this area: the Hui’anbao-Shajingzi Fault, the Qinglongshan-Pingliang Fault, the Weizhou-Anguo Fault, and the Qingtongxia-Guyuan Fault. ,, Influenced by these major faults, multiple thrust and overthrust systems have developed along the southwestern basin margin. Seismic interpretation of Well YT3 indicates that the well is located within an overthrust zone between the Qinglongshan-Pingliang Fault and the Hui’anbao-Shajingzi Fault. The structure is characterized by an overall syncline with locally developed thrust blocks (Figure ).

3.

Seismic profile of the study area. Adapted with permission from ref . Copyright 2025 John Wiley & Sons Ltd.

Overall, stratigraphic thickness in the study area exhibits a systematic pattern of being thicker in the west and north and thinner in the east and south. Core samples from Well YT3 indicate that the Yanghugou Formation is dominated by sandstone and mudstone, with abundant gray-black mudstone; its basal interval consists mainly of dark gray sandstone and ferruginous-aluminous mudstone, with coal seams locally developed. In contrast, the Wulalike Formation is characterized by thick-bedded mudstone and contains abundant graptolite fossils.

3. Samples and Methods

3.1. Samples

This study collected a total of 47 downhole core and crude oil samples from three wells (YT1, YT2, and YT3) in the southwestern Ordos Basin, comprising 14 mudstones, 3 coal samples, and 4 oil sand samples from Yanghugou, alongside 22 mudstone source rock samples and 4 crude oil samples from Wulalike. The 39 source rock samples underwent total organic carbon (TOC) analysis, Rock-Eval pyrolysis, vitrinite/graptolite reflectance measurement, and maceral identification to determine their fundamental geochemical characteristics. For oil-source correlation, 8 source rock samples and 4 oil sand extracts from Yanghugou, plus 2 source rock samples and 4 crude oils from Wulalike were subjected to gas chromatography-mass spectrometry (GC-MS) analysis to characterize their molecular geochemical signatures.

3.2. Geochemical Analytical Methods

3.2.1. TOC and Rock-Eval Pyrolysis

The samples were ground and sieved to pass through a 100-mesh sieve (<150 μm). Subsequently, 10–20 mg of the powdered sample was placed in a porous ceramic crucible preheated at 1000 °C for 4 h. Carbonates were removed by treating the powdered samples with 5% hydrochloric acid (HCl) in a fume hood until all inorganic carbon was eliminated. The crucibles were then rinsed with distilled water to neutrality, dried in an oven, and cooled to room temperature. TOC contents were subsequently determined using a LECO CS-230 carbon–sulfur analyzer. Approximately 1 g of iron flux and 1 g of tungsten flux were added prior to analysis, following the Chinese national standard GB/T 19145-2022 (Determination of Organic Carbon in Sedimentary Rocks).

Subsequently, Rock-Eval pyrolysis was conducted on the prepared samples, with approximately 10 mg analyzed using an OGE-II Rock-Eval instrument. Free oil content (S ₁), source potential (S ₂), CO₂ released during pyrolysis (S ₃), and maximum pyrolysis temperature (T _max) were obtained in accordance with GB/T 18602-2012 (Pyrolysis Analysis of Rock). Source rock parameters such as hydrogen index (HI) and oxygen index (OI) were calculated. , Analytical conditions included a temperature range of 300–650 °C and a heating rate of 25 °C/min.

3.2.2. Vitrinite Reflectance (R _o)

Vitrinite and graptolite reflectance measurements and maceral identification were conducted using a Zeiss Axio Scope.A1/J&M MSP 200 microphotometer. Vitrinite reflectance measurements followed the standard SY/T 5124-2012, and whole-rock maceral identification was performed in accordance with SY/T 6414-2014. Rock samples were prepared as polished blocks, which were subsequently ground and polished for vitrinite and graptolite reflectance determination, maceral identification, and quantitative analysis. The vitrinite reflectance (R _o) value was calculated as the mean of 30–50 measurement points per sample, and a standard reference material was measured after a 2 h interval for calibration control.

Equivalent vitrinite reflectance (*R _o) for Ordovician marine rocks was determined using the graptolite random reflectance (R _graptolite) conversion equation proposed by Luo et al.: *R _o = 1.055 × R _graptolite – 0.053.

3.2.3. Gas Chromatography-Mass Spectrometry (GC-MS)

Following cleaning, air-drying, and grinding to 100 mesh, the source rock and solid bitumen samples underwent Soxhlet extraction with chloroform to obtain Bitumen A. Asphaltenes were then precipitated using n-hexane. The maltene fraction was separated by alumina column chromatography into saturated hydrocarbons, aromatic hydrocarbons, and nonhydrocarbons. The saturated and aromatic fractions were analyzed by gas chromatography-mass spectrometry (GC-MS).

GC-MS analyses were performed using an Agilent 7890B-5977B gas chromatograph-quadrupole mass spectrometer system. Saturated hydrocarbon fractions were analyzed on an HP-5MS fused silica capillary column (30 m × 0.25 mm × 0.25 μm) with an oven program of 50 °C (1 min), ramped to 100 °C at 20 °C/min, then to 320 °C at 3 °C/min, and held for 15.17 min. Aromatic fractions were analyzed on an HP-5MS column (60 m × 0.25 mm × 0.25 μm) with an oven program of 50 °C (1 min), ramped to 315 °C at 3 °C/min, and held for 15.67 min.

For both analyses, the injector temperature was 300 °C; helium was used as the carrier gas (1.04 mL/min for saturated fractions and 1.0 mL/min for aromatic fractions), and the mass spectrometer was operated in full-scan and multiple ion detection (MID) modes with electron ionization at 70 eV and an ion source temperature of 230 °C. Scan ranges were m/z 50–650 for saturated hydrocarbons and m/z 50–550 for aromatic hydrocarbons. All analyses followed GB/T 18606-2017.

3.3. Multivariate Statistical Analysis

Multivariate statistical analyses, including hierarchical cluster analysis (HCA), Q-mode factor analysis (Q-FA), and R-mode factor analysis (R-FA), were applied to evaluate oil-source correlation. HCA groups samples according to similarities in multivariate parameter characteristics, maximizing intracluster similarity while enhancing intercluster separation. Factor analysis (FA) simplifies complex interrelationships among variables by extracting a limited number of latent factors, thereby constructing composite dimensions that capture the dominant geochemical information. ,

In Q-FA, factor loadings (ranging from −1 to +1) reflect the correlation between samples and extracted factors, with absolute values approaching unity indicating strong associations, whereas factor scores represent the integrated contribution of principal factors to individual samples. Although Q-FA focuses on sample classification and R-FA emphasizes variable grouping, the two methods differ only in data matrix transposition and share identical mathematical principles. All factor analyses were conducted using IBM SPSS Statistics, employing a correlation matrix, principal component extraction (eigenvalues > 1), Varimax rotation with Kaiser normalization, and regression-based calculation of factor scores.

4. Results and Discussion

4.1. Basic Geochemical Characteristics of Source Rocks

4.1.1. Organic Matter Abundance

Organic matter abundance, defined as the quantity of organic material per unit rock mass, constitutes the fundamental prerequisite for hydrocarbon generation. Following Chinese industry standard SY/T 5735-2019, this study assessed the source rock organic matter abundance. The TOC content in sediment is expressed as a weight percentage, which can characterize the total amount of organic matter in sediment. Based on TOC analysis, the mudstone source rocks of Yanghugou exhibit TOC contents ranging from 0.34 to 7.45% (average 4.29%) and are therefore evaluated as excellent source rocks. In contrast, Wulalike mudstones show TOC contents of only 0.01–0.87%, averaging 0.31%. Their overall organic matter abundance is relatively low, and they are consequently evaluated as none source rocks (Table and Figure A).

1. Geochemical Parameters for Source Rocks of the Yanghugou Formation and Wulalike Formation.

well	stratum	depth/m	lithology	TOC/%	S 1 + S 2/mg/g	HI/mg HC/g TOC	T max/°C	R o/%
YT2	Yanghugou	2245.00	mudstone					0.55
YT2	Yanghugou	2266.00	mudstone					0.57
YT2	Yanghugou	2293.00	mudstone					0.60
YT2	Yanghugou	2303.00	mudstone					0.61
YT2	Yanghugou	2553.00	mudstone	0.92	3.64	381.09	453.00
YT2	Yanghugou	2573.00	mudstone					0.64
YT2	Yanghugou	2577.11	mudstone	0.94	3.34	337.98	455.50
YT2	Yanghugou	2578.00	mudstone					0.65
YT2	Yanghugou	2582.10	mudstone	6.17	24.14	379.33	437.20
YT2	Yanghugou	2584.36	coal	37.73	131.44	319.75	436.00
YT2	Yanghugou	2586.86	mudstone	2.69	10.21	359.48	438.00
YT2	Yanghugou	2587.20	mudstone	5.35	28.34	491.06	429.60
YT2	Yanghugou	2587.95	mudstone	7.45	29.06	374.52	431.40
YT2	Yanghugou	2592.32	mudstone	7.30	25.17	323.15	442.00
YT2	Yanghugou	2598.00	mudstone					0.68
YT3	Yanghugou	2270.28	gray mudstone	6.45	18.74	282.62	441.00
YT3	Yanghugou	2271.41	mudstone	4.43	22.64	493.91	438.00
YT3	Yanghugou	2276.75	coal	37.41	151.05	364.96	437.00
YT3	Yanghugou	2283.08	coal	32.26	137.68	385.91	425.30
YT3	Yanghugou	2284.96	gray mudstone	5.26	18.81	330.02	432.40
YT3	Yanghugou	2285.75	mudstone	0.34	4.11	1123.94	462.50
YT3	Yanghugou	2288.10	gray mudstone	2.37	6.58	264.25	445.00
YT3	Yanghugou	2303.52	mudstone	4.39	20.84	447.38	423.80
YT3	Yanghugou		gray mudstone	5.97	22.63	356.38	434.10
YT1	Wulalike	1204.56	mudstone	0.01	0.13	704.69	463.70
YT1	Wulalike	1204.56	mudstone	0.01	0.15	722.07	453.70
YT1	Wulalike	1210.00	mudstone					1.97
YT1	Wulalike	1496.03	mudstone	0.09	0.16	156.25	456.70
YT1	Wulalike	1497.62	gray mudstone	0.87	0.15	13.38	456.00
YT1	Wulalike	1498.00	mudstone					2.04
YT1	Wulalike	1499.08	mudstone	0.11	0.16	138.89	457.00
YT1	Wulalike	1499.36	gray mudstone	0.11	0.23	162.59	441.80
YT1	Wulalike	1500.14	mudstone	0.24	0.31	56.51	457.20
YT1	Wulalike	1502.48	gray mudstone	0.11	0.26	173.97	449.40	2.09
YT2	Wulalike	3860.00	mudstone					1.05
YT2	Wulalike	3861.40	mudstone	0.29	1.03	263.76	458.10
YT2	Wulalike	3878.00	mudstone					1.10
YT2	Wulalike	3880.00	mudstone	0.68	2.35	199.29	460.20
YT2	Wulalike	3924.27	gray mudstone	0.85	1.31	128.72	451.70
YT2	Wulalike	3924.65	mudstone	0.09	1.69	1430.13	464.20
YT2	Wulalike	3924.65	mudstone	0.30	1.18	280.29	448.60
YT2	Wulalike	3925.00	mudstone					1.27
YT2	Wulalike	3927.80	mudstone	0.34	1.67	360.75	442.30
YT3	Wulalike	4100.08	black mudstone	0.27	1.01	273.87	459.50	1.16
YT3	Wulalike	4091.00	black mudstone	0.16	0.71	343.74	464.90
YT3	Wulalike	4093.00	mudstone					1.08
YT3	Wulalike	4099.90	black mudstone	0.28	0.97	268.79	453.90
YT3	Wulalike	4101.09	gray mudstone	0.59	1.01	124.82	448.30
YT3	Wulalike	4102.35	mudstone	0.29	1.05	286.57	448.60
YT3	Wulalike	4104.48	mudstone	0.18	0.89	400.11	463.90
YT3	Wulalike	4105.00	mudstone					1.18
YT3	Wulalike	4105.62	black mudstone	0.27	1.13	327.63	457.80
YT3	Wulalike	4107.46	mudstone	0.60	1.22	157.86	452.40

^a R _o: equivalent R _o.

4.

Evaluation diagram of organic abundance parameters for source rocks in the study area. (A) Frequency distribution diagram of total organic carbon (TOC) in mudstone source rocks; (B) frequency distribution diagram of hydrocarbon generation potential (S ₁ + S ₂) in mudstone source rocks; and (C) cross-plot diagram of S ₁ + S ₂ vs HI for coal in Yanghugou Formation.

The genetic potential (S ₁ + S ₂), representing the total hydrocarbon yield from source rocks, quantifies the maximum volume of oil and gas generatable by kerogen pyrolysis. , Rock-Eval pyrolysis data delineate a pronounced contrast: Yanghugou mudstones exhibit S ₁ + S ₂ values of 3.34–29.36 mg/g (average 17.02 mg/g), classifying them as good source rocks. Conversely, Wulalike mudstones show markedly lower yields (0.13–2.35 mg/g; avg. 0.85 mg/g), consistent with none source rocks (Table and Figure B).

Coal samples and mudstones were evaluated separately due to differing criteria for organic abundance, as detailed in Table and Figure C. Yanghugou coals exhibit elevated geochemical parameters, with TOC contents ranging from 32.26 to 37.73%, S ₁ + S ₂ values of 131.44–151.05 mg/g, and HI values of 320–386 mg HC/g TOC. Based on these parameters, the unit is classified as marginally effective oil-prone source rock, possessing kerogen-dominated generation potential.

Overall, the organic matter abundance of Yanghugou source rocks (mudstones and coal samples) is higher than that of Wulalike source rocks, indicating a stronger hydrocarbon generation capacity.

4.1.2. Organic Matter Type

Organic matter type critically evaluates source rock quality and hydrocarbon generation-expulsion efficiency. The kerogen-type index (TI) is calculated as TI = (100 × sapropelinite + 50 × liptinite – 75 × vitrinite – 100 × inertinite)/100. Yanghugou source rocks exhibit TI values of −28.30 to 69.20 (average 34.41) They are classified as Type II₂ kerogen with sapropelic/vitrinite-dominant macerals and minor liptinite. In contrast, Wulalike shows elevated TI (66.10–85.30; avg. 80.93), diagnostic of Type I kerogen with sapropelinite-dominated assemblages in ternary diagram lower-right sectors (Table and Figure A), indicating algal and aquatic microbial precursors. , Because these components were not measured for Yanghugou coal samples, coal is not shown in Figure A.

2. Kerogen-Type Classification Table for Source Rocks in the Study Area.

well	stratum	depth/m	lithology	sapropelinite/%	liptinite/%	vitrinite/%	inertinite/%	TI	kerogen type
YT2	Yanghugou	2245.00	mudstone	57.60	12.10	27.50	0.90	42.13	II1
YT2	Yanghugou	2263.00	mudstone	46.30	17.90	35.20	0.60	28.30	II2
YT2	Yanghugou	2266.00	mudstone	72.40	6.70	21.00	3.00	56.97	II1
YT2	Yanghugou	2293.00	mudstone	55.20	15.10	27.50	3.30	38.88	II2
YT2	Yanghugou	2303.00	mudstone	35.50	27.50	29.50	6.70	20.43	III
YT3	Yanghugou	2304.00	mudstone	32.00	2.00	12.00	50.00	–28.30	III
YT3	Yanghugou	2577.11	mudstone	42.70	35.20	20.30	2.20	42.88	II1
YT3	Yanghugou	2582.10	mudstone	59.00	7.00	29.00	2.00	38.10	II2
YT3	Yanghugou	2586.86	mudstone	73.00	2.00	21.00	2.00	56.20	II1
YT3	Yanghugou	2587.20	mudstone	49.00	5.00	31.00	1.00	13.70	II2
YT3	Yanghugou	2592.32	mudstone	79.00	3.00	17.00	0.00	69.20	II1
YT1	Wulalike	1499.08	mudstone	88.00	5.00	2.00	3.00	85.30	I
YT2	Wulalike	3864.00	mudstone	79.50	12.40	0.00	6.20	79.50	II1
YT2	Wulalike	3880.00	mudstone	84.20	9.80	0.00	5.40	83.70	I
YT2	Wulalike	3924.65	mudstone	86.00	0.00	11.00	1.00	76.30	II1
YT2	Wulalike	3925.00	mudstone	82.00	0.00	7.00	10.00	66.10	II1
YT2	Wulalike	3927.80	mudstone	88.00	3.00	5.00	1.00	84.80	I
YT2	Wulalike	4091.00	mudstone	89.70	4.30	0.00	7.10	84.75	I
YT3	Wulalike	4093.00	mudstone	83.50	11.70	0.00	5.20	84.15	I
YT3	Wulalike	4107.46	mudstone	91.00	0.00	5.00	3.00	83.80	I

5.

Kerogen-type assessment map for source rocks in the study area. (A) Ternary diagram of maceral composition. Adapted with permission from ref . Copyright 2013 Elsevier Science B.V.; (B) hydrogen index (HI) vs maximum pyrolysis temperature (T _max) plot. Adapted with permission from ref . Copyright 2020 Elsevier Ltd.

Cross-plots of HI versus T _max provide rapid kerogen typing. The diagrams indicate predominantly Type II₁–II₂ kerogens in Yanghugou (mainly II₂), while Wulalike source rocks exhibit heterogeneous organic matter distributions. Integrated TI assessment classifies Wulalike as a Type I kerogen, demonstrating superior organic quality and enhanced oil-generation potential relative to Yanghugou. The reason for the poor distribution of Wulalike source rocks on the plate may be the high T _max value caused by their high maturity, which in turn leads to inaccurate indication (Figure B).

4.1.3. Organic Matter Maturity

Assessment of organic matter maturity critically determines hydrocarbon generative capacity and resource potential, with vitrinite reflectance (R _o) serving as the primary indicator. Yanghugou source rocks exhibit R _o values of 0.55–0.68% (avg. 0.61%), indicating low maturity. Conversely, Wulalike shows significantly higher thermal evolution (R _o = 1.05–2.09%; avg. 1.44%), confirming mature-stage conditions with samples reaching high overmaturity (Table ).

Additionally, T _max data is an important indicator reflecting the maturity of organic matter. T _max values for Yanghugou source rocks predominantly range from 425 to 440 °C, consistent with low maturity. In contrast, Wulalike samples show T _max signatures characteristic of mature to high maturity (Figure ), confirming significantly more advanced thermal evolution in the Wulalike Formation compared with the Yanghugou Formation.

6.

Frequency distribution plot of T _max for source rocks in the Yanghugou Formation and Wulalike Formation study area.

4.2. Molecular Geochemical Characteristics of Source Rocks, Oil Sands, and Crude Oils

4.2.1. N-Alkanes

N-Alkanes, primarily derived from fatty acids and hydrocarbons in organisms, exhibit distribution patterns reflecting depositional environments and organic matter sources. High Carbon Preference Index (CPI) values indicate low maturity with terrestrial input, whereas CPI ≈ 1 characterizes marine input and/or high thermal maturity. Yanghugou source rocks exhibit CPI values of 0.49–1.40 (avg. 1.15) and an OEP value of 1.04–1.26 (avg. 1.13), demonstrating odd-over-even predominance (OEP). These values indicate significant terrestrial input and a low maturity. Conversely, both Yanghugou oil sands and Wulalike samples show CPI and the OEP approximately equal to 1, reflecting no distinct odd–even predominance, a characteristic indicative of marine input and/or high maturity.

The light-to-heavy alkane ratio (nC₂₁ ^–/nC₂₂ ⁺) of Yanghugou source rocks ranges from 0.51 to 1.72 (avg. 1.00), while Yanghugou oil sands show ratios of 0.45–1.03 (avg. 0.67). Both show relatively high abundances of long-chain n-alkanes (nC₂₂ ⁺), with some samples exhibiting a bimodal distribution. Low-molecular-weight n-alkanes originate from microalgae, whereas abundant long-chain n-alkanes (C₂₂ ⁺) are primarily derived from higher-plant waxes. , This indicates significantly higher plant input into Yanghugou source rocks and oil sands. In contrast, Wulalike source rocks and crude oils exhibit higher nC₂₁ ^–/nC₂₂ ⁺ ratios with front-peak distributions, suggesting predominantly algal-derived organic matter. The characteristic unimodal distribution dominated by low-molecular-weight homologues reflects high thermal maturity, further confirming Wulalike’s advanced maturation state.

Pristane (Pr) and phytane (Ph) serve as key indicators of depositional environments. The Pr/Ph ratio reflects organic matter type and redox conditions during early diagenesis, , while Pr/nC₁₇ and Ph/nC₁₈ ratios provide insights into source rock type, depositional setting, and thermal maturity. Yanghugou source rocks exhibit Pr/Ph = 1.24–3.84 (avg. 2.71), indicating terrigenous input under oxic conditions. Yanghugou oil sands show Pr/Ph = 0.89–1.23 (avg. 1.09), suggesting a potential hypersaline environment. Conversely, Wulalike samples display Pr/Ph = 1.50–2.00, signifying deposition in a moderately reducing environment.

The Pr/nC₁₇ vs Ph/nC₁₈ cross-plot (Figure A) shows Yanghugou source rocks plotting within the mixed organic matter field with affinity to terrigenous Type III organic matter, while Yanghugou oil sands plot toward marine and saline lacustrine facies. Wulalike samples cluster in the lower-left quadrant near mixed/marine fields, a distribution attributed to their advanced thermal evolution that depleted pristane and phytane concentrations. For this reason, this diagram has a limited ability to indicate oil-source correlations and requires additional information for a reliable comparison.

7.

Biomarker parameter diagrams for source rocks, oil sands, and seep oils from the Yanghugou Formation and Wulalike Formation. (A) Cross-plot of Pr/nC₁₇ versus Ph/nC₁₈ for samples from the Yanghugou Formation and Wulalike Formation in the study area. Adapted with permission from ref . Copyright 1980 Elsevier Ltd.; (B) ternary diagram of regular steranes (C₂₇–C₂₉). Adapted with permission from ref . Copyright 1979 Elsevier Ltd.; (C) sterane C₂₉ββ/(ββ + αα) vs C₂₉ 20S/(20S + 20R); and (D) distribution map of hopanes/steranes and tricyclic terpanes/hopanes in samples of Well YT2 in the study area.

4.2.2. Steranes

Relative abundances of C₂₇, C₂₈, and C₂₉ regular steranes serve as key indicators of organic matter sources and depositional environments. − C₂₇ steranes derive from algae and lower aquatic organisms, , C₂₈ steranes relate to phytoplankton input, and C₂₉ steranes originate primarily from higher plants. , All samples plot predominantly within the mixed planktonic–terrigenous source field (Figure B).

The C₂₉ sterane ββ/(ββ + αα) vs C₂₉20S/(20S + 20R) cross-plot assesses thermal maturity of source rocks and crude oils. , Both ratios increase with maturity, reaching equilibrium values of 0.52–0.55 and 0.67–0.71, respectively, at peak oil generation. Wulalike samples consistently plot within the mature zone. However, some Yanghugou oil sands exhibit higher maturity than Yanghugou source rocks (Figure C), requiring verification with supplementary maturity parameters.

4.2.3. Terpanes

Elevated C₁₉–C₂₁ tricyclic terpane (TT) abundances indicate significant terrigenous higher-plant input, whereas C₂₃TT dominance suggests aquatic origins. Yanghugou source rocks show a mean C₁₉TT/C₂₃TT ratio of 2.38. Conversely, both Yanghugou oil sands and Wulalike samples exhibit substantially lower ratios (avg. < 0.5), demonstrating substantial terrigenous plant contribution to Yanghugou source rocks versus predominantly aquatic-derived characteristics in Yanghugou oil sands and Wulalike samples.

A high steranes/hopanes ratio (St/H) indicates lower aquatic algal derivation, reflecting marine organic matter, while low ratios signify terrestrial input. Zhu classified marine versus terrestrial oils in the Tarim Basin using H/St versus ∑C_19–29TT/hopanes correlations. As Figure D shows, Yanghugou source rocks exhibit terrestrial to marine–terrestrial transitional characteristics; Wulalike samples display marine-origin features; and Yanghugou oil sands cluster near the marine–terrestrial boundary, suggesting mixed origins.

Ts and Tm are C₂₇ tricyclic terpanes derived from hopanoids by the loss of three methyl groups. They are a pair of isomers. Owing to the greater thermal stability of Ts relative to Tm, the Ts/(Ts + Tm) ratio increases progressively with advancing thermal maturity. This ratio is applicable across a wide maturity range, from immature to highly mature. Yanghugou source rocks exhibit significantly lower Ts/(Ts + Tm) ratios than the Yanghugou oil sands, indicating a relatively higher maturity in the oil sands. In contrast, Wulalike source rocks display higher Ts/(Ts + Tm) ratios, similar to those of their genetically related crude oils.

Yanghugou oil sands exhibit gammacerane/C₃₀ hopane (Ga/C₃₀αβ) ratios of 0.08–0.38 (avg. 0.18). According to salinity assessment criteria for depositional water bodies, these values suggest deposition in a moderately saline aquatic environment. In contrast, Yanghugou source rocks show lower Ga/C₃₀αβ ratios (0.02–0.08), consistent with a freshwater-dominated depositional setting.

Based on the above results, Yanghugou crude oil sands exhibit complex origins. These oil sands display higher thermal maturity than their source rocks, whereas conventional biomarker cross-plots (e.g., Figure A,B) yield ambiguous results. This ambiguity arises from (1) thermal maturity differences compromising key biomarker parameters and (2) inherent limitations of traditional methods in resolving multivariate data. To address this, we employ a multivariate statistical analysis of biomarker parameters to mitigate maturity effects and multidimensional constraints. This approach enables precise characterization of Yanghugou crude oil signatures and establishes robust oil-source correlations.

4.3. Applying Multivariate Statistics to Oil-Source Correlation

4.3.1. Hierarchical Cluster Analysis

In this study, 6 source rocks and 6 oil sands from the Yanghugou Formation and 4 crude oil samples and 2 source rock samples from the Wulalike Formation were used. In addition, oil-bearing sandstone samples from L33 and B45 wells of Chang8 oil formation in the Huanxian block, which is adjacent to the work area, were imported to the work area in consideration of the potential connection between the target layer and the crude oil of Yanchang Formation in the work area. In summary, a total of 20 source rock, crude oil, and oil-bearing sandstone samples were used for this cluster analysis. In terms of variables, this paper collects 31 n-alkanes and terpenes as variables in this multivariate statistical analysis mainly through the book “The Biomarker Guide: Volume 2, Biomarkers and Isotopes in Petroleum Systems and Earth History” written by famous scholars Peters et al. and the research cases of many scholars. The biomarker parameters for 31 n-alkanes, terpenes, and sterols were collected as variables in this multivariate statistical analysis study, as shown in Table . These biomarker parameters are representative of most of the parameters used in conventional analytical studies.

3. Biomarker Parameter Compilation Table .

well	stratum	depth (m)	lithology	A	B	C	D	E	F	G	H	I	J	K	L	M	N	O
YT2	Yanghugou	2584.18	coal	1.19	1.05	2.41	0.56	0.62	0.38	0.43	0.17	0.40	0.43	0.02	0.15	0.05	3.95	0.05
YT2	Yanghugou	2584.36	coal	0.49	1.03	3.67	2.33	0.86	0.18	0.37	0.22	0.40	0.55	0.26	0.06	0.03	2.29	0.10
YT2	Yanghugou	2586.86	mudstone	1.19	1.11	3.24	1.19	0.51	0.47	0.34	0.25	0.41	0.61	0.39	0.18	0.03	4.87	0.12
YT2	Yanghugou	2592.02	mudstone	1.40	1.14	3.84	1.29	0.48	0.23	0.42	0.20	0.39	0.51	0.06	0.21	0.07	4.17	0.10
YT2	Yanghugou	2592.32	mudstone	1.28	1.26	3.04	1.19	0.53	0.20	0.19	0.34	0.47	0.73	0.32	0.17	0.08	2.49	0.09
YT3	Yanghugou	2271.41	mudstone	1.31	1.15	2.36	0.48	0.26	0.17	0.36	0.26	0.38	0.69	0.37	0.19	0.07	2.43	0.12
YT3	Yanghugou	2276.75	coal	1.12	1.04	1.85	0.49	0.37	0.33	0.37	0.30	0.33	0.91	0.44	0.17	0.04	2.59	0.09
YT3	Yanghugou	2288.10	mudstone	1.18	1.12	1.24	0.83	0.61	0.56	0.48	0.23	0.29	0.78	0.58	0.39	0.08	2.83	0.13
YT2	Yanghugou	2573.45	oil sand	1.08	1.03	1.21	0.60	0.42	0.33	0.31	0.27	0.42	0.65	0.18	0.90	0.10	0.46	0.09
YT2	Yanghugou	2598.73	oil sand	1.07	1.02	0.89	0.28	0.52	0.77	0.42	0.21	0.37	0.56	0.38	3.26	0.32	0.75	0.15
YT3	Yanghugou	2286.85	oil sand	1.08	1.00	1.23	0.51	0.50	0.56	0.36	0.26	0.38	0.67	0.40	0.95	0.05	0.77	0.11
YT3	Yanghugou		oil sand	1.08	1.07	1.01	0.44	0.46	0.40	0.38	0.26	0.35	0.74	0.36	0.79	0.04	0.73	0.17
YT2	Wulalike	3867.40	mudstone	1.07	1.02	1.23	0.14	0.11	0.66	0.41	0.28	0.31	0.93	0.35	2.41	0.05	0.63	0.15
YT2	Wulalike		mudstone	1.04	1.00	1.88	0.10	0.05	0.62	0.38	0.25	0.37	0.67	0.46	1.13	0.11	1.67	0.13
YT2	Wulalike		oil	1.05	1.01	2.00	0.10	0.06	1.12	0.43	0.28	0.30	0.94	0.61	2.01	0.11	1.16	0.31
YT2	Wulalike		oil	1.07	1.01	1.89	0.11	0.07	1.13	0.44	0.27	0.29	0.95	0.57	0.71	0.11	1.17	0.23
YT3	Wulalike		oil	1.04	1.02	1.74	0.12	0.08	1.06	0.42	0.27	0.31	0.89	0.47	2.26	0.06	1.36	0.22
YT3	Wulalike		oil	1.04	1.01	1.49	0.12	0.07	1.26	0.40	0.24	0.35	0.69	0.60	1.16	0.09	1.52	0.26
B45	Yanchang	2812.60	oil sand	1.07	1.02	0.78	0.26	0.26	0.69	0.35	0.28	0.37	0.74	0.51	5.51	0.23	2.94	2.50
L33	Yanchang	2821.40	oil sand	1.10	1.05	0.83	0.28	0.29	1.38	0.29	0.28	0.43	0.65	0.32	5.70	0.28	2.32	1.73

well	stratum	depth (m)	lithology	P	Q	R	S	T	U	V	W	X	Y	Z	AA	AB	AC	AD	AE
YT2	Yanghugou	2584.18	coal	0.08	0.08	0.37	0.47	0.17	1.00	0.60	0.15	0.24	0.03	0.02	0.01	1.23	0.01	0.94	0.02
YT2	Yanghugou	2584.36	coal	0.02	0.02	0.42	0.42	0.30	0.60	0.57	0.20	0.30	0.03	0.01	0.05	1.21	0.01	0.91	0.01
YT2	Yanghugou	2586.86	mudstone	0.03	0.04	0.41	0.39	0.49	0.66	0.60	0.42	0.33	0.06	0.04	0.02	0.16	0.02	0.97	0.07
YT2	Yanghugou	2592.02	mudstone	0.08	0.12	0.33	0.36	0.23	0.52	0.60	0.42	0.42	0.03	0.03	0.02	0.96	0.02	1.02	0.04
YT2	Yanghugou	2592.32	mudstone	0.06	0.07	0.41	0.41	0.55	0.66	0.59	0.52	0.38	0.04	0.04	0.03	0.21	0.01	1.03	0.06
YT3	Yanghugou	2271.41	mudstone	0.05	0.07	0.34	0.21	0.58	0.57	0.57	0.47	0.38	0.10	0.08	0.06	0.39	0.02	0.75	0.07
YT3	Yanghugou	2276.75	coal	0.04	0.07	0.38	0.34	0.51	0.64	0.61	0.27	0.31	0.06	0.03	0.05	0.12	0.04	0.64	0.06
YT3	Yanghugou	2288.10	mudstone	0.05	0.07	0.29	0.24	0.57	0.61	0.57	0.34	0.26	0.07	0.05	0.07	0.57	0.06	0.60	0.12
YT2	Yanghugou	2573.45	oil sand	0.38	0.55	0.31	0.28	0.42	0.82	0.54	0.20	0.18	0.40	0.09	0.26	0.63	0.21	0.53	0.23
YT2	Yanghugou	2598.73	oil sand	0.13	0.21	0.55	0.41	0.54	0.83	0.47	0.20	0.14	0.54	0.23	0.23	0.55	0.81	0.66	0.52
YT3	Yanghugou	2286.85	oil sand	0.08	0.12	0.49	0.49	0.53	0.29	0.48	0.24	0.15	0.49	0.18	0.20	1.56	0.21	0.58	0.28
YT3	Yanghugou		oil sand	0.11	0.14	0.50	0.44	0.54	0.74	0.60	0.23	0.17	0.41	0.14	0.30	0.51	0.17	0.57	0.18
YT2	Wulalike	3867.40	mudstone	0.20	0.34	0.51	0.39	0.19	0.43	0.60	0.19	0.16	0.61	0.33	0.28	1.69	0.52	0.62	0.45
YT2	Wulalike		mudstone	0.07	0.11	0.37	0.37	0.49	0.61	0.53	0.35	0.25	0.32	0.13	0.13	0.65	0.39	0.56	0.25
YT2	Wulalike		oil	0.06	0.12	0.39	0.49	0.54	0.72	0.55	0.00	0.24	0.63	0.38	0.31	1.18	0.35	0.58	0.39
YT2	Wulalike		oil	0.09	0.34	0.34	0.45	0.57	0.71	0.58	0.13	0.06	0.72	0.41	0.29	1.36	0.23	0.49	0.19
YT3	Wulalike		oil	0.07	0.10	0.48	0.45	0.61	0.63	0.59	0.25	0.15	0.54	0.34	0.22	0.89	0.31	0.59	0.40
YT3	Wulalike		oil	0.16	0.33	0.42	0.40	0.52	0.70	0.56	0.17	0.22	0.68	0.37	0.26	0.79	0.36	0.47	0.25
B45	Yanchang	2812.60	oil sand	0.28	0.31	0.59	0.59	0.53	0.79	0.76	0.25	0.32	0.83	1.65	0.64	0.62	1.08	0.48	0.41
L33	Yanchang	2821.40	oil sand	0.22	0.25	0.53	0.45	0.52	0.78	0.64	0.27	0.17	0.89	1.53	0.60	0.67	1.17	0.50	0.46

^aA, CPI; B, OEP; C, Pr/Ph; D, Pr/nC₁₇; E, Ph/nC₁₈; F, St/H; G, C₂₇/C_27–29: C₂₇αααR/C_27–29αααR; H, C₂₈/C_27–29: C₂₈αααR/C_27–29αααR; I, C₂₉/C_27–29: C₂₉αααR/C_27–29αααR; J, C₂₈/C₂₉: C₂₈αααR/C₂₉αααR; K, rearranged steranes/steranes; L, ∑C_19–26TT/C₃₀αβ; M, C₂₄TET/C₃₀αβ: C₂₄ tetracyclic terpane/17α­(H), 21β­(H)-C₃₀hopane; N, C₃₀*/C₂₉Ts: C₃₀ diahopane/C₂₉Ts; O, C₃₀*/C₃₀αβ; P, Ga/C₃₀αβ; Q, Ga/C₃₁αβ: Ga/17α­(H), 21β­(H)-C₃₁hopane­(22S + 22R); R, C₂₉ββ/(αα + ββ); S, C₂₉20S/(20R + 20S); T, C₂₇ rearranged sterane 20S/(20S + 20R); U, C₂₉ rearranged sterane 20S/(20S + 20R); V, C₃₁αβ22S/(22S + 22R); W, C₂₉βα/C₂₉αβ: 17β­(H), 21α­(H)-C₂₉ norhopane/17α­(H), 21β­(H)-C₂₉ norhopane; X, C₃₀βα/C₃₀αβ: 17β­(H),21α­(H)-C₃₀ moretane/C₃₀αβ; Y, Ts/(Ts + Tm); Z, Ts/C₃₀αβ; AA, C₂₉Ts/(C₂₉Ts + C₂₉αβ); AB, C₂₇/C₂₉; AC, C₂₃TT/C₃₀αβ; AD, C₂₉αβ/C₃₀αβ; AE, ∑C_19–29TT/(∑C_19–29TT + hopanes).

The biomarker parameters of 31 n-alkanes, terpanes, and steranes from each sample were analyzed by clustering. The results define two distinct clusters (Figure ): Cluster I contains all Wulalike and Yanghugou samples; Cluster II comprises Yanchang oil sands. Yanghugou oil sands show genetic affinity with Wulalike samples and occupy an intermediate position between the two source rock end-members, implying mixed sourcing. Yanchang samples exhibit no significant correlation with other samples, indicating genetic isolation from the local reservoirs.

8.

Dendrogram of hierarchical clustering for samples in the study area.

4.3.2. Q-Mode Factor Analysis

After the initial clarification of the oil source of the oil sands in Yanghugou Formation by cluster analysis, the molecular geochemical data were further analyzed by factor analysis. However, before proceeding to the discussion of the results of the Q-FA, it is first necessary to perform the Kaiser–Meyer–Olkin (KMO) test and the Bartlett’s sphericity test on the data; the KMO value for the present factor analysis was 0.805, which meets the lower condition of the lower limit for performing factor analysis (KMO ≥ 0.5). The observed values of Bartlett’s spherical test statistic were 1115.514, and the corresponding probability p of 0.000 was less than the significance level of 0.05, indicating that the molecular geochemical data of the present study are suitable for Q-FA and the results are reliable.

The results of Q-FA are divided into three categories. The first category is Yanghugou source rocks, and the second category is Yanghugou oil sands, Wulalike source rocks, and crude oil, which indicates that Wulalike crude oil originates from its own source rocks and also reflects to a certain extent that Yanghugou oil sands and Wulalike source rocks have a certain oil-source correlation, and it is speculated that it may be a mixed source. The third category is Yanchang oil sands and one Yanghugou oil sand, but this oil sand sample has a high factor loading of about 0.5 in F2, which indicates that there is a correlation with the second category, while the two Yanchang oil sands have a very low factor loading in F2, which is poorly correlated with the samples in F1 and F2 (Table ).

4. Rotated Factor Loadings Matrix (Q-Mode FA).

well	stratum	depth/m	lithology	F1 (I)	F2 (II)	F3 (III)
YT2	Yanghugou	2586.86	mudstone	0.980	0.117	0.140
YT2	Yanghugou	2592.02	mudstone	0.943	0.263	0.072
YT3	Yanghugou	2276.75	coal	0.940	0.252	0.099
YT2	Yanghugou	2584.18	coal	0.922	0.242	0.119
YT3	Yanghugou	2271.41	mudstone	0.919	0.346	0.055
YT2	Yanghugou	2592.32	mudstone	0.911	0.332	–0.019
YT3	Yanghugou	2288.10	mudstone	0.902	0.216	0.172
YT2	Yanghugou	2584.36	coal	0.802	0.352	–0.122
YT2	Wulalike		mudstone	0.694	0.568	0.381
YT3	Yanghugou	2286.85	oil sand	0.340	0.846	0.156
YT2	Yanghugou	2573.45	oil sand	0.340	0.817	0.113
YT2	Wulalike		oil	0.484	0.768	0.153
YT2	Wulalike	3867.40	mudstone	0.012	0.766	0.581
YT3	Yanghugou		oil sand	0.472	0.754	0.184
YT2	Wulalike		oil	0.331	0.737	0.536
YT3	Wulalike		oil	0.352	0.658	0.642
YT3	Wulalike		oil	0.539	0.611	0.446
L33	Yanchang	2821.40	oil sand	0.019	0.180	0.974
B45	Yanchang	2812.60	oil sand	0.104	0.073	0.968
YT2	Yanghugou	2598.73	oil sand	–0.032	0.498	0.800

4.3.3. R-Mode Factor Analysis

Given their transposed data matrix relationship, R-FA used identical samples and parameters. R-FA of 31 biomarker parameters from 20 samples extracted seven principal factors, representing seven latent geochemical dimensions (Table ). Parameters grouped within the same factor exhibit strong correlations and, therefore, share common geochemical implications. Accordingly, the geochemical significance of each factor can be inferred from the diagnostic parameters with high loadings and their collective behavior.

5. Rotated Factor Loadings Matrix (R-Mode FA).

biomarker parameters	F1	F2	F3	F4	F5	F6	F7
∑C19–26TT/C30αβ	0.944	0.101	0.207	0.000	0.107	–0.002	0.063
C23TT/C30αβ	0.938	0.089	0.250	–0.024	0.006	–0.002	0.096
Ts/C30αβ	0.858	0.154	0.197	0.057	0.385	–0.030	0.090
C29ββ/(αα + ββ)	0.837	–0.008	–0.053	0.132	–0.064	–0.235	–0.185
C29Ts/(C29Ts + C29αβ)	0.807	0.271	0.425	–0.006	0.168	–0.151	0.046
C24TET/C30αβ	0.804	–0.040	0.135	0.016	–0.189	0.128	0.385
C30*/C30αβ	0.794	0.070	0.123	0.138	0.496	–0.028	0.168
∑C19–29TT/(∑C19–29TT + hopanes)	0.787	0.281	0.297	–0.166	–0.292	–0.015	–0.129
Ts/(Ts + Tm)	0.745	0.386	0.431	–0.184	–0.052	–0.139	–0.093
C2920S/(20R + 20S)	0.638	–0.028	–0.096	–0.129	0.251	–0.424	–0.208
St/H	0.581	0.466	0.147	–0.388	–0.074	–0.062	0.045
Pr/Ph	–0.528	–0.397	–0.477	0.127	0.280	0.006	–0.174
rearranged steranes/steranes	0.183	0.891	–0.146	–0.115	–0.068	–0.119	0.091
C28/C29	–0.029	0.886	0.154	–0.054	0.121	–0.068	–0.273
C27 rearranged sterane 20S/(20S + 20R)	0.198	0.740	–0.188	0.284	–0.256	0.057	0.361
Ph/nC18	–0.251	–0.658	–0.318	0.255	–0.063	–0.266	0.217
C29αβ/C30αβ	–0.395	–0.621	–0.488	0.237	0.168	0.150	–0.115
Pr/nC17	–0.397	–0.541	–0.396	0.354	0.167	–0.320	–0.008
Ga/C31αβ	0.236	0.074	0.918	–0.043	–0.015	–0.019	0.030
Ga/C30αβ	0.398	–0.107	0.861	0.134	0.086	0.010	0.110
C27/C27–29	–0.125	0.126	–0.107	–0.932	–0.068	–0.003	0.012
C29/C27–29	0.091	–0.642	0.001	0.679	0.000	0.067	0.124
C28/C27–29	0.101	0.635	0.174	0.657	0.133	–0.011	–0.212
C29βα/C29αβ	–0.192	–0.140	–0.359	0.609	0.068	0.566	0.000
OEP	–0.324	–0.166	–0.314	0.571	0.170	0.483	0.058
C31αβ22S/(22S + 22R)	0.308	0.046	0.047	0.130	0.904	0.044	0.077
C30*/C29Ts	–0.137	–0.398	–0.501	0.020	0.587	0.319	0.208
C30βα/C30αβ	–0.279	–0.270	–0.421	0.385	0.460	0.333	0.057
CPI	–0.076	0.007	0.024	0.069	0.088	0.924	0.009
C29 rearranged sterane 20S/(20S + 20R)	0.168	–0.205	0.237	–0.147	0.192	–0.077	0.764
C27/C29	0.088	–0.146	0.205	–0.542	–0.066	–0.321	–0.649

To clarify the geological implications of the extracted factors, the known geochemical significance of the biomarker parameters associated with factors F1, F2, F3, and F4 was first summarized (Table ). Although individual biomarker parameters may carry multiple interpretive meanings, clear clustering patterns emerge at the factor level. In particular, parameters loading strongly on F1 predominantly reflect thermal maturity, whereas those associated with F2 are mainly related to the organic matter type. On this basis, the indicative significance of factors F1–F4 is interpreted sequentially below, providing a coherent framework for constructing a comprehensive oil-source discrimination scheme.

6. Sorting and Summarizing the Geochemical Significance of the Parameters Contained in Factors F1 and F2.

factor	biomarker parameters	established geochemical significance
F1	∑C19–26TT/C30αβ	maturity, water salinity ,
	C23TT/C30αβ	organic matter source
	Ts/C30αβ	maturity, sedimentary environment
	C29ββ/(αα + ββ)	maturity
	C29Ts/(C29Ts + C29αβ)	maturity ,
	C24TET/C30αβ	maturity, organic matter origin, water salinity
	C30*/C30αβ	maturity, , sedimentary environment ,,
	∑C19–29TT/(∑C19–29TT + hopanes)	maturity, water salinity ,,
	Ts/(Ts + Tm)	maturity, , sedimentary environment
	C2920S/(20R + 20S)	maturity
between F1 and F2	St/H	organic matter origin ,
	Pr/Ph	maturity, organic matter origin, sedimentary environment, water salinity
F2	rearranged steranes/steranes	organic matter source, maturity, sedimentary environment ,
	C28/C29	organic matter origin
	C27 rearranged sterane 20S/(20S + 20R)	uncertain
	Ph/nC18	sedimentary environment, , maturity
	C29αβ/C30αβ	organic matter origin, mineralogical characteristics
between F2 and F3	Pr/nC17	organic matter origin, , maturity, water salinity
F3	Ga/C31αβ	water stratification environment, water salinity ,
	Ga/C30αβ	water stratification environment, water salinity ,
F4	C27/C27–29	organic matter source ,
	C29/C27–29	organic matter source ,
	C28/C27–29	organic matter source ,
	C29βα/C29αβ	uncertain

Factor F1 contains 12 biomarker parameters. Considering that the absolute factor loadings of St/H and Pr/Ph in F1 are lower than 0.6 and both parameters show relatively high loadings in F2, these two parameters were excluded from F1. Similarly, Pr/nC₁₇ was excluded from F2 due to its shared loading characteristics. Then, among the ten biomarker parameters included in F1, nine parameters ∑C_19–26TT/C₃₀αβ, , Ts/C₃₀αβ, C₂₉ββ/(αα + ββ), C₂₉Ts/(C₂₉Ts + C₂₉αβ), , C₂₄TET/C₃₀αβ, C₃₀*/C₃₀αβ, , ∑C_19–29TT/(∑C_19–29TT + hopanes), ,, Ts/(Ts + Tm) , and C₂₉20S/(R + S) are all well correlated with thermal maturity and generally increase with increasing maturity. Therefore, F1 is interpreted to represent thermal maturity, with higher factor scores indicating higher maturity levels.

While F1 primarily reflects maturity, some parameters located near the F1–F2 boundary provide insight into the transition between maturity and organic matter characteristics. For example, the St/H ratio shows variation with increasing maturity, potentially due to the lower thermal stability of hopanoid compounds, yet it also carries information related to organic matter input. , Similarly, Pr/Ph correlates with maturity but is widely used as an indicator of organic matter origin, with values greater than 3 typically reflecting terrestrial input. , The significant loadings of these parameters in F2 suggest that beyond maturity control, organic matter origin exerts an important influence, indicating that Factor F2 is dominantly associated with organic matter type.

This interpretation is further supported by parameters that load strongly on Factor F2 including C₂₈/C₂₉ (high values indicative of marine origin) and C₂₉αβ/C₃₀αβ (high values indicative of terrestrial input). In addition, rearranged steranes/steranes ratios and the Pr/nC₁₇ ratio, which lies at the boundary between F2 and F3, are also closely related to organic matter origin. ,, Integrating these observations, F2 is interpreted as reflecting the organic matter origin. According to the signs of the factor loadings (Table ), higher F2 scores indicate a stronger marine organic matter contribution, whereas lower scores suggest increased terrestrial input.

Having established maturity (F1) and organic matter origin (F2), the interpretation of Factor F3 focuses on depositional environmental conditions. F3 is dominated by Ga/C₃₀αβ and Ga/C₃₁αβ. Gammacerane is commonly present in source rocks and crude oils, and elevated gammacerane contents are widely regarded as indicators of water-column stratification during deposition. Consequently, higher Ga/C₃₀αβ reflect higher paleosalinity and more strongly reducing depositional environments. , Therefore, F3 is interpreted to represent water salinity with higher factor scores corresponding to higher salinity conditions.

Finally, Factor F4 is composed of three sterane distribution ratios (C₂₇/∑C_27–29, C₂₈/∑C_27–29, and C₂₉/∑C_27–29), which are routinely used to infer organic matter sources. , Based on the positive and negative factor loadings (Table ), higher F4 scores indicate a greater contribution from higher-plant-derived organic matter, whereas lower scores reflect dominance of algal and lower aquatic organic matter sources.

Consequently, based on the factor score calculations, four composite indices are established to evaluate key petroleum geochemical characteristics: the Maturity Index (MI), Organic Matter Origin Index (OMOI), Water Salinity Index (WSI), and Organic Matter Source Index (OMSI), which are defined as follows

$$\mathrm{Maturity\; Index}(\mathrm{MI})=\sum _{j=1}^n{\mathit{\alpha }}_j{x}_j$$ 1

$$\mathrm{Organic\; Matter\; Origin\; Index}(\mathrm{OMOI})=\sum _{j=1}^n{\beta }_j{x}_j$$ 2

$$\mathrm{Water\; Salinity\; Index}(\mathrm{WSI})=\sum _{j=1}^n{\gamma }_j{x}_j$$ 3

$$\mathrm{Organic\; Matter\; Source\; Index}(\mathrm{OMSI})=\sum _{j=1}^n{\delta }_j{x}_j$$ 4

Here, j denotes the sequence number of a biomarker parameter for a given sample; n represents the total number of parameters; α, β, γ, and δ are the corresponding score coefficients (Table ); and x is the measured value of each parameter. Using the above formula, the biomarker parameters listed in Table were calculated, and four composite indices were constructed. These indices were then plotted as scatter diagrams (MI vs OMOI and WSI vs OMSI). On this basis, a comprehensive geochemical diagram was developed as a diagnostic tool to rapidly evaluate thermal maturity, organic matter origin, water salinity, organic matter source, and oil-source correlations for samples from the Yanghugou and Wulalike Formations (Figure ).

7. Factor Score Coefficient Matrix.

no. (j)	biomarker parameters	F1 (α)	F2 (β)	F3 (γ)	F4 (δ)	F5	F6	F7
1	CPI	0.045	–0.029	0.043	–0.094	–0.014	0.505	–0.111
2	OEP	0.003	0.010	–0.030	0.111	0.005	0.164	–0.032
3	Pr/Ph	–0.026	–0.032	–0.079	–0.006	0.116	–0.038	–0.101
4	St/H	–0.031	–0.055	–0.058	0.114	0.049	–0.247	0.022
5	C27/C27–29	0.006	–0.109	–0.067	0.077	–0.082	–0.199	0.147
6	C29/C27–29	0.074	0.058	–0.075	–0.106	–0.016	0.039	0.033
7	C28/C27–29	–0.005	0.001	–0.089	–0.330	0.041	0.099	0.061
8	C28/C29	–0.054	0.179	0.080	0.273	0.073	–0.090	–0.133
9	rearranged steranes/steranes	0.055	–0.141	0.051	0.202	–0.105	–0.024	0.002
10	∑C19–26TT/C30αβ	–0.089	0.222	0.041	0.039	0.148	–0.049	–0.113
11	C24TET/C30αβ	–0.009	0.228	–0.145	0.004	0.012	–0.090	0.115
12	C30*/C29Ts	0.154	–0.039	–0.053	–0.004	–0.007	0.058	–0.028
13	C30*/C30αβ	0.157	–0.070	–0.080	–0.014	–0.174	0.103	0.167
14	Ga/C30αβ	0.033	–0.036	–0.127	–0.125	0.221	0.115	0.079
15	Ga/C31αβ	0.088	0.007	–0.023	0.013	0.191	–0.026	0.048
16	C29ββ/(αα + ββ)	–0.058	–0.076	0.338	0.074	0.066	0.015	0.029
17	C2920S/(20R + 20S)	–0.101	–0.037	0.362	0.037	0.061	0.016	0.011
18	C27 rearranged sterane 20S/(20S + 20R)	0.194	–0.057	–0.157	0.077	–0.109	–0.053	–0.161
19	C29 rearranged sterane 20S/(20S + 20R)	0.118	–0.029	–0.132	–0.016	0.099	–0.163	–0.127
20	C31αβ 22S/(22S + 22R)	0.018	0.201	–0.153	0.120	–0.142	–0.053	0.242
21	C29βα/C29αβ	–0.069	–0.022	0.097	–0.085	0.106	–0.113	0.486
22	C30βα/C30αβ	–0.030	0.058	0.058	–0.024	0.435	–0.041	0.016
23	Ts/(Ts + Tm)	0.059	–0.003	–0.078	0.121	–0.071	0.231	–0.091
24	Ts/C30αβ	0.001	0.003	–0.066	0.031	0.158	0.089	–0.017
25	C29Ts/(C29Ts + C29αβ)	0.072	0.019	0.040	–0.009	–0.009	0.006	–0.071
26	Pr/nC17	0.102	0.006	–0.019	0.002	0.146	0.002	0.001
27	Ph/nC18	0.060	0.016	0.062	0.030	0.079	–0.046	0.002
28	C27/C29	0.034	–0.104	0.044	–0.122	0.027	–0.010	–0.373
29	C23TT/C30αβ	0.154	–0.049	–0.043	–0.005	–0.054	0.064	–0.005
30	C29αβ/C30αβ	0.038	–0.106	–0.098	0.010	0.011	0.052	–0.109
31	∑C19–29TT/(∑C19–29TT + hopanes)	0.145	–0.029	–0.041	–0.012	–0.166	0.103	–0.120

^aBiomarker parameter: Excluded parameters do not participate in the calculation.

8. Computation Results of Key Sample MI and PMI Indices.

well	stratum	depth	lithology	MI	OMOI	WSI	OMSI
YT2	Yanghugou	2584.18	coal	0.172	–0.033	0.057	0.002
YT2	Yanghugou	2584.36	coal	0.163	0.051	0.012	0.043
YT2	Yanghugou	2586.86	mudstone	0.191	0.166	0.025	0.093
YT2	Yanghugou	2592.02	mudstone	0.175	0.014	0.070	0.046
YT2	Yanghugou	2592.32	mudstone	0.193	0.179	0.044	0.188
YT3	Yanghugou	2271.41	mudstone	0.173	0.248	0.041	0.085
YT3	Yanghugou	2276.75	coal	0.179	0.297	0.039	0.059
YT3	Yanghugou	2288.10	mudstone	0.210	0.291	0.040	0.003
YT2	Yanghugou	2573.45	oil sand	0.374	0.168	0.330	0.083
YT2	Yanghugou	2598.73	oil sand	0.996	0.195	0.119	0.021
YT3	Yanghugou	2286.85	oil sand	0.456	0.230	0.073	0.057
YT3	Yanghugou		oil sand	0.406	0.244	0.089	0.044
YT2	Wulalike	3867.40	mudstone	0.776	0.246	0.192	0.027
YT2	Wulalike		mudstone	0.459	0.288	0.063	0.061
YT2	Wulalike		oil	0.700	0.389	0.066	–0.005
YT2	Wulalike		oil	0.440	0.396	0.155	0.004
YT3	Wulalike		oil	0.715	0.357	0.060	0.030
YT3	Wulalike		oil	0.539	0.338	0.174	0.026
B45	Yanchang	2812.60	oil sand	1.783	0.306	0.207	0.067
L33	Yanchang	2821.40	oil sand	1.733	0.237	0.164	0.100

9.

Oil-source correlation diagram based on composite geochemical indices for the Yanghugou and Wulalike Formations in the southwestern Ordos Basin. (A) MI-OMOI composite index diagram for oil-source correlation and (B) WSI-OMSI composite index diagram for oil-source correlation.

Considering that the Yanchang oil-bearing sandstones show little correlation with other samples in the cluster analysis, and given that this study focuses on the oil-source correlation between the Yanghugou and Wulalike Formations, the two Yanchang oil sands were excluded from the scatterplot projection. A total of 18 samples were therefore projected in Figure , where the Yanghugou and Wulalike samples exhibit clear and systematic distribution patterns.

In Figure A, the MI-OMOI diagram defines a geochemical template in which the left portion represents lower thermal maturity and the right portion higher maturity, whereas the upper domain indicates a predominantly marine organic matter origin and the lower domain a terrestrial-dominated origin. Yanghugou source rocks cluster in the lower-left area, reflecting relatively low maturity and mixed organic matter input, consistent with a paralic setting. In contrast, Wulalike source rocks and crude oils plot mainly in the upper-right area, indicating higher maturity and marine organic matter origin. The Yanghugou oil sands span the distribution fields of both source rock systems, suggesting mixed-source contributions, whereas Wulalike oils closely overlap with their own source rocks, supporting a self-sourced petroleum system.

In Figure B, based on the geochemical significance of the WSI and OMSI indices, the left side of the diagram represents relatively low water salinity (freshwater conditions), whereas the right side reflects higher salinity (saline conditions). Vertically, the upper part of the diagram indicates a greater contribution from higher-plant-derived organic matter, while the lower part is dominated by aquatic algae and microorganisms.

Compared with that in Figure A, oil-source discrimination in Figure B is less distinct, which is likely due to the smaller number of biomarker parameters incorporated in factors F3 and F4. Nevertheless, the diagram still provides useful geological and geochemical information. Yanghugou source rocks are mainly concentrated on the left side of the diagram, suggesting deposition under relatively freshwater conditions. In contrast, both Wulalike source rocks and crude oil samples plot to the right of the Yanghugou samples, indicating higher depositional water salinity consistent with a marine environment. Vertically, Yanghugou samples show broader variability with locally higher-plant input, while Wulalike samples are dominated by algal and microbial organic matter. Once again, the Yanghugou oil sands plot between the two source rock fields, reinforcing a mixed-source origin.

By integration of the two geochemical templates for oil-source correlation, it can be concluded that the Yanghugou source rocks are characterized by low thermal maturity and paralic facies deposition, with mixed organic matter input. In contrast, the Wulalike source rocks exhibit higher maturity and clear marine depositional characteristics dominated by algal organic matter. With respect to oil-source correlations, the Yanghugou oil sands reflect mixed contributions from both the Yanghugou and Wulalike source rocks, whereas the Wulalike crude oils are interpreted as self-sourced and self-reservoired within the same formation.

Overall, consistent sample distributions observed in both diagrams demonstrate that the composite index approach effectively overcomes the limitations of single-parameter biomarker methods and minimizes the influence of maturity differences. This integrated interpretation further confirms that Wulalike oils are self-sourced, whereas the Yanghugou oil sands record mixed contributions from both the Yanghugou and the Wulalike source rocks.

5. Conclusion

Regarding the fundamental geochemical characteristics of source rocks, the Yanghugou source rocks (mudstones and coals) exhibit high organic matter abundance and maturity, indicating favorable hydrocarbon generation potential. The Wulalike source rocks possess moderate organic matter abundance with favorable organic matter types and are predominantly within the oil-generation window. Locally (Well YT1), the degree of evolution is elevated, reaching the wet gas generation window. These findings provide crucial geochemical evidence for the regional hydrocarbon resource evaluation.

Comprehensive biomarker analysis indicates that organic matter in Yanghugou source rocks primarily originates from terrestrial higher plants and was deposited under oxidizing conditions with relatively low maturity. In contrast, Yanghugou oil sands exhibit mixed genesis, incorporating both marine/saline lake and terrestrial organic input. They were deposited in a highly saline environment and underwent greater thermal evolution, exhibiting maturity significantly higher than that of the associated source rocks. Wulalike source rocks and crude oil samples display typical marine organic sources, were deposited in relatively reducing environments, and exhibit the highest maturity among the samples examined in this study.

Multivariate statistical analysis confirms that Yanghugou oil sands exhibit not only affinity with their own source rocks but also strong affinity with Wulalike source rocks. Their biomarkers exhibit characteristics intermediate between those of Yanghugou source rocks (low maturity, dominated by terrestrial organic matter) and the Wulalike source rocks (high maturity, marine origin), confirming that they represent products of mixed provenance. Wulalike crude oil shows a high degree of consistency with its own source rock characteristics, consistent with a self-sourced and self-reservoired model.

The comprehensive diagram constructed through composite indices (MI, OMOI) further confirms these relationships: Yanghugou oil sands plot across fields representative of both Yanghugou (paralic facies) and Wulalike (marine) sources, demonstrating dual-source contributions. Conversely, the Wulalike samples are concentrated in the high-maturity marine-origin zone, indicating a clear and unambiguous oil-source correlation.

All data supporting the findings of this study are included in the manuscript.

J.S.: Writingoriginal draft, methodology, investigation, data curation; K.S.: writingreview and editing, investigation, funding acquisition, conceptualization; J.H.: methodology, investigation; Y.S.: resources; Y.X.: resources, visualization, formal analysis, supervision; J.L.: visualization, investigation, data curation.

The authors declare no competing financial interest.