The Origin and Mixed-Source Proportion of Natural Gas in the Dixin Area of the Junggar Basin: Geochemical Insights from Molecular and Isotopic Composition

Abstract

The Dixi area of the Junggar Basin has favorable petroleum geological conditions, with the Cretaceous system representing one of the principal hydrocarbon-bearing strata. However, the genetic origin and mixing characteristics of natural gas across different tectonic zones remain insufficiently understood. In this study, a total of 65 natural gas samples were analyzed using molecular composition and stable carbon isotopic data to determine gas origins and quantify the contributions of different source rocks. A novel multivariate mathematical analysis method was developed and applied to convert compositional and isotopic data into quantitative parameters, enabling the accurate estimation of end-member mixing ratios in natural gas. This methodological innovation addresses the challenge of interpreting multi-source gas systems under complex geological conditions. The results show that the Cretaceous natural gas in the Dixi area is derived from three main sources, comprising both oil-type gas from Permian lacustrine source rocks and coal-type gas from Carboniferous coal-measure source rocks. The calculated mixing proportions exhibit significant spatial variation: in the northern Dixi area, coal-type gas dominates (67.8–84.3%), while the southern zone presents a broader mixture (25.6–68.4% coal-type gas). In the Dongdaohaizi Depression, oil-type gas is predominant, accounting for 89.4–97.7%. This study not only clarifies the genetic classification and mixing dynamics of natural gas in the Dixi area but also provides a quantitative framework for evaluating accumulation processes and source contributions in multi-source gas reservoirs. The proposed method offers valuable guidance for assessing resources and optimizing exploration strategies in the Junggar Basin and other similar basins.

1. Introduction

The coexistence of mixed-source natural gases with differing properties is a common phenomenon in natural gas reservoirs. This mainly includes mixtures of gases from different source rocks, as well as mixtures from the same source at different stages of thermal evolution [1,2,3]. Because mixed-source natural gas exhibits integrated characteristics from multiple origins, it is challenging to determine its genesis, source, and the contribution of each component based solely on its geochemical properties. Compared with crude oil, natural gas has a relatively simple composition and lacks diagnostic compounds such as biomarker molecules used for oil–source correlation. This simplicity significantly increases the difficulty of accurately assessing the mixing ratios of natural gas from different sources [4,5]. However, natural gas in discovered fields often exhibits characteristics of multi-source origins and mixed-source reservoir formation across different geological periods [6,7]. Therefore, it is essential to address the specific challenges associated with identifying the origins of multi-source natural gas, particularly in geologically complex basins where gases from different source rocks may be thermally evolved, migrated, or mixed at various stages. These complexities hinder the effectiveness of traditional qualitative assessments. As such, quantitative calculation of source contributions is critical for accurately determining the proportion and influence of each gas source, which in turn supports the reliable reconstruction of gas accumulation processes and source–reservoir relationships.

A wide range of studies have identified thermogenic gas [8,9], biodegraded gas [10], and mixed gases [11] using molecular composition, stable isotope ratios, and thermal maturity modeling [12]. It is now well established that the natural gas in sedimentary basins originates from diverse sources, including thermogenic cracking of kerogen or oil [13], microbial degradation [14], and complex mixing processes [15]. Globally, the formation of large-scale gas accumulations is often associated with oil cracking under high-temperature, high-pressure conditions in deep reservoirs [16]. However, field observations frequently reveal coexisting gas types within a single reservoir, including oil-cracked, coal-derived, and mixed gases [17]. In geologically heterogeneous settings, this genetic diversity further complicates exploration and increases uncertainty [18]. Scholars have attempted to study the geochemical characteristics of mixed-source natural gas and calculate the mixing ratio of natural gas using geochemical indicators such as carbon isotopes and light hydrocarbons. However, on the one hand, these studies are generally aimed at binary natural gas mixed-sources (i.e., two natural gas mixed-sources with different properties). For multi-source natural gas mixed-sources, there are still deficiencies in the simplicity of the research methods and the scope of application [19,20]. On the other hand, the identification diagrams established by these studies are rather complex, making it difficult to apply the theory to the actual identification of the proportion of mixed-sources of natural gas [17].

As one of the most important hydrocarbon-bearing basins in China, the Junggar Basin has yielded numerous discoveries of shallow oil and gas accumulations [21,22]. However, deep hydrocarbon exploration in the basin remains in its early stages. In the Dixi area, two sets of humic-type source rocks are present: the Upper Carboniferous Batamayineishan Formation and the Lower Carboniferous Dishuiquan Formation. Previous studies generally suggest that both source rock intervals have contributed to gas accumulation in the region [23]. However, the effectiveness of the Dishuiquan Formation source rocks remains speculative due to limited good penetration in the Dinnan Uplift and the poor imaging quality of available seismic data [24,25]. In addition, the Middle Permian Pingdiquan Formation, one of the basin’s most prolific source rock intervals, is poorly developed in the Dinnan Uplift [26] and has long been overlooked in studies of gas accumulation in the Dixi area. However, the genesis of Cretaceous natural gas and the characteristics of mixed-source differences remain unclear, which limits the understanding of the sources of natural gas and the laws of accumulation and enrichment. Therefore, it is necessary to further clarify the genesis differences of Cretaceous natural gas and determine the proportion of natural gas mixtures.

To address the unresolved issue of gas–source identification in the Dixi area, this study systematically investigates the genetic origins of natural gas and the potential secondary alteration processes affecting it. Based on integrated geochemical analyses, the study delineates the spatial and temporal distribution patterns of gas types, Furthermore, in this study, a multivariate mathematical analysis method is proposed to quantitatively determine the proportions of mixed natural gas sources. This method is particularly valuable for analyzing the proportions of mixed-source gases under complex geological conditions. It provides a scientific foundation for understanding the mechanisms of natural gas accumulation.

2. Geological Setting

The Junggar Basin is located in northwestern China (Figure 1a) at the tectonic junction of the Kazakhstan, Tarim, and Siberian plates. It is composed of a stable cratonic block and surrounding fold belts, covering an area of approximately 134,000 km² [26]. The Dixi area is situated in the western margin of the Dinnan Uplift, within the eastern Junggar Basin of northwestern China (Figure 1b). Structurally, this area lies at the intersection of several second-order tectonic units, including the Dishuiquan Depression to the north and the Dongdaohaizi Depression to the south, which together form a structurally complex yet hydrocarbon-favorable setting (Figure 1c). These depressions act as hydrocarbon-generation centers, while the Dixi area serves as a migration and accumulation zone for natural gas [27,28]. The basin architecture in this region reflects a multi-phase tectonic evolution that has influenced source rock development and hydrocarbon migration. In particular, the Dixi area experienced significant tectonic uplift during the Permian and subsequent subsidence during the Mesozoic, creating favorable conditions for hydrocarbon preservation. The regional stratigraphy is characterized by the presence of multiple source–reservoir–seal assemblages, with Upper Carboniferous and Lower Carboniferous coal-bearing formations [29].

Situated along the western margin of the Dinnan Uplift, the Dixi region hosts abundant Cretaceous hydrocarbon reserves. Regional exploratory drilling has revealed widespread hydrocarbon indicators, highlighting substantial exploration prospects (Figure 1c). This area features multiple gas-prone source rock sequences spanning Carboniferous through Jurassic periods. Composed predominantly of alternating tuffaceous mudstones and volcanic strata (Figure 2), Carboniferous source rocks in the Ludong-Wucaiwan region exhibit TOC concentrations exceeding 1.0%, Type III organic matter, and moderate-to-high thermal evolution (VRo: 1.0–2.0%), demonstrating significant gas-generation capacity [21,22,30,31]. Permian lacustrine mudstones, extensively distributed across the basin, show particularly strong gas potential in the Dongdaohaizi Depression where TOC values surpass 2.5% with mixed Type II–III kerogen and elevated maturity (VRo 1.3–1.5%) [32]. The Lower Carboniferous Dishuiquan Formation presents a transitional marine-terrestrial sequence characterized by conglomerates, sandy conglomerates, siltstones, thick shales, and intermittent coal layers. Similarly to Batamayineishan Formation patterns, these deposits display zonal distribution in the Kelameili foreland, with maximum thicknesses concentrated in the Dishuiquan, Wucaiwan, and Dongdaohaizi depressions. Overlying Permian–Triassic mudstones form effective regional seals, while intra-Carboniferous lithologic variations across depositional cycles create favorable reservoir–caprock configurations. In particular, Permian regional seals effectively preserve hydrocarbons within Carboniferous weathered crust reservoirs [23,24,33].

3. Samples and Methods

3.1. Samples

In this study, a total of 65 natural gas samples were collected from 10 wells distributed across the Dongdaohaizi Depression and its surrounding areas (Figure 1c). The sampling locations cover the major gas-bearing structures within the study area in the horizontal dimension, and vertically span five major stratigraphic intervals, including the Carboniferous, Permian, Triassic, Jurassic, and Cretaceous formations (Figure 2). The samples were subjected to gas chromatographic analyses and stable carbon isotope measurements. All the analyses were carried out in the State Key Laboratory of Oil and Gas Resources and Exploration at the China University of Petroleum (Beijing).

3.2. Methods

3.2.1. Gas Chromatographic Analysis of Natural Gas

Natural gas composition was analyzed using a gas chromatograph (7890 GC, Agilent Technologies, Santa Clara, CA, USA; custom-configured by Wasson-ECE Instrumentation, Fort Collins, CO, USA), and quantitative determination of components was performed using the external standard method [34]. The instrument is equipped with one flame ionization detector (FID) and two thermal conductivity detectors (TCDs). High-purity helium (≥99.999%; Air Liquide, Shanghai, China) was used as the carrier gas for the FID and one TCD (for CO analysis), while high-purity nitrogen (≥99.999%; Air Liquide, Shanghai, China) was used for the other TCD (for H₂ and H₂S analysis). The GC oven temperature was programmed as follows: an initial hold at 68 °C for 7 min, followed by a ramp of 10 °C/min to 90 °C with a 1.5 min hold, then a ramp of 15 °C/min to 175 °C, and a final hold at 175 °C for 5 min.

3.2.2. Stable Carbon Isotope Analysis of Natural Gas

After natural gas was enriched using a drainage and gas collection method in supersaturated brine, a 50 μL gas sample was extracted using a gas-tight syringe (Hamilton Company, Reno, NV, USA) and injected into the instrument. The acquisition program was then initiated. Hydrocarbon components were oxidized to CO₂ in the combustion unit and subsequently carried by helium gas into the isotope ratio mass spectrometer (IRMS) for carbon isotope analysis [35]. The instrument used was a MAT 253 stable isotope ratio mass spectrometer (Thermo Fisher Scientific, Bremen, Germany), equipped with an HP-PLOT Q capillary column (30 m × 0.53 mm × 40 μm; Agilent Technologies, Santa Clara, CA, USA). Helium (≥99.999%, Air Liquide, Shanghai, China) was used as the carrier gas, with a column flow rate of 1.3 mL/min. To ensure analytical accuracy, one standard gas was analyzed after every 12 samples to monitor instrument performance.

3.2.3. Multivariate Mathematical Analysis Method

Suppose the absolute content data of Y components of X mixed-source natural gas products are obtained, then A matrix M_x×y, with X rows and y columns can be obtained. Then, the composition of each end-element natural gas should also have the absolute content of Y components. If there are A end-elements (A ≤ X), then the composition of all end-element natural gas also forms a matrix N_A×Y with a rows and Y columns. Then the matrix representing the proportion of mixed-sources must be A matrix with X rows and a columns, denoted as P_X×A. At this point, the following formula holds:

$$M_{X\times Y}=P_{X\times A}\times N_{A\times Y}+e$$

(1)

where M_x×y is a matrix of dimensions X × Y, where each element represents the absolute concentration of each of the Y components in each of the X mixture samples; P_X×A is the proportion matrix (or mixing ratio matrix) of size X × A, representing the proportion (or contribution) of each of the A end-member sources in each of the X mixture samples; N_A×Y is the composition matrix of end-members, where each row represents an end-member source and each element denotes the absolute concentration of the Y components in that source; and e is error matrix of size X × Y, representing the deviation between the modeled and observed component concentrations.

Through the Alternating Least Squares algorithm in multivariate mathematical analysis, continuous estimation, assignment and fitting are carried out. Eventually, if e is minimized, the quantity, composition and proportion of the terminal energy are finally calculated.

4. Results

4.1. Compositional Characteristics of Natural Gas

As shown in analysis of gas samples from the northern and southern parts of the Dixi area and the Dongdaohaizi Depression (

Table 1. Composition of natural gas testing results in the Dixi area, Junggar Basin.

Area	Well	Depth (m)	Formation	Composition of Natural Gas
				C1%	C2%	C3%	C1/(C1–5)	N2%	CO2%
Northern	Dx18	3998.5	T1S	90.46	2.15	0.24	0.98	7.90	0.41
	Dx18	3906	T1S	89.13	3.20	0.26	0.97	8.21	0.14
	Dx10	2192	T1S	89.05	4.29	0.23	0.95	6.22	0.18
Southern	Dx9	2242–2234	K1tg	89.87	5.68	0.24	0.94	4.28	0.19
	Dx9	2109.8	K1tg	89.85	6.05	0.25	0.94	4.10	0.00
	Dx12	1707	K1tg	90.43	3.78	0.26	0.96	5.56	0.07
	Dx12	2205	K1tg	84.25	7.36	0.14	0.92	5.00	0.21
	Dx12	2287	K1tg	89.40	5.45	0.13	0.94	4.47	0.35
	D201	1880	K1tg	80.76	9.20	0.14	0.90	8.11	0.03
	D205	2293.5	K1tg	84.04	8.75	0.26	0.91	6.52	0.04
	D210	1815.5	K1tg	88.40	5.22	0.25	0.94	6.19	0.06
	Dx15	2657	J2t	92.00	3.90	0.42	0.96	3.10	/
	Dx15	2190	K1tg	85.05	2.80	0.24	0.97	11.41	0.34
	Dx15	2451	J2t	85.13	6.67	0.24	0.93	7.02	0.40
	Dx15	2657	J2t	90.56	3.86	0.42	0.96	4.63	0.18
	Dx15	2144	K1tg	85.11	2.2	0.43	0.97	11.58	0.27
Dongdaohaizi	Dn8	3070–3084	C2b	93.35	5.15	0.74	0.96	1.37	0.14
	Dn8	3132	C2b	84.71	10.54	0.68	0.89	2.91	0.12
	Dn8	3650–3665	C2b	90.36	5.2	0.75	0.95	4.34	0.00
	Dn8	3363	P3w	89.42	6.19	0.65	0.94	4.19	0.15
	Dn8	3510	P3w	88.66	7.9	0.75	0.92	3.01	0.11
	Dn1	3345	P3w	87.56	8.76	0.45	0.91	2.51	0.07
	Dn1	3625	P3w	88.4	8.34	0.42	0.91	3.06	0.03
	Dn1	3830	P3w	88.03	9.15	0.41	0.91	2.73	0.02
	C31	3635	J2t	84.41	11.35	0.55	0.88	3.97	0.05
	C341	3502	J2t	85.72	11.21	0.45	0.88	2.38	0.08
	C34	2906	J2t	82.20	9.30	0.41	0.90	6.60	/
	C34	3652	J2t	87.24	7.78	0.54	0.92	3.16	0.06

), the methane (C₁) content in the northern Dixi area ranges from 89.05% to 90.46%, and that of ethane (C₂) ranges from 2.15% to 4.29%, and propane (C₃) from 0.13% to 0.75%. In the southern Dixi area, the C₁ content varies between 80.76% and 92.00%, and that of C₂ ranges from 2.20% to 9.20%. For the Dongdaohaizi Depression, the C₁ content ranges from 82.20% to 93.35%, with that of C₂ between 5.15% and 11.35%. Overall, the C₁/(C_1–5) ratios fall within the range of 0.88 to 0.98, the nitrogen (N₂) content ranges from 1.37% to 11.58%, and the carbon dioxide (CO₂) content remains low, ranging from 0% to 0.41%. Across various stratigraphic intervals, the natural gas is dominated by alkane hydrocarbons, characterized by high dryness coefficients and low N₂ and CO₂ contents—features indicative of gases derived from mature source rocks.

4.2. Stable Carbon Isotope Characteristics of Natural Gas

According to the carbon isotope analysis (Figure 3), methane δ¹³C values primarily range from −40‰ to −30‰. Carboniferous gases exhibit the heaviest isotopic signatures, reaching values as high as −32‰, while gases from other stratigraphic intervals show a broader distribution from relatively lighter to heavier values. Ethane δ¹³C values are concentrated between −30‰ and −22‰, with most samples displaying characteristics typical of coal-derived gas. However, gas samples from certain wells in the Cretaceous and Permian intervals show isotopic signatures indicative of oil-derived gas. These results suggest significant variations in the thermal evolution stages of source rocks across different stratigraphic intervals. Such isotopic distinctions provide critical insights into hydrocarbon accumulation mechanisms and the identification of gas origins within the basin.

5. Discussion

5.1. Type of Multi-Origin Natural Gas

Accurate identification of natural gas types is a fundamental aspect of petroleum geology, as it directly influences source determination, reservoir formation model reconstruction, and resource potential evaluation [36,37]. This task becomes particularly challenging in basins with vertically superimposed source rocks and complex accumulation histories, where coal-derived, oil-derived, and biogenic gas may coexist or be mixed. The stable carbon isotopic composition of natural gas is a key indicator for distinguishing genetic types, as it is closely related to the type of source organic matter and its thermal maturity [38,39]. Typically, inorganic gases exhibit a “reversed carbon isotope sequence” (i.e., δ¹³C₁ > δ¹³C₂ > δ¹³C₃), while organically derived alkane gases show a “normal carbon isotope sequence” [38,39]. Carbon isotope analyses of natural gas from the Carboniferous, Cretaceous, Jurassic, and Permian strata in the study area all show normal isotopic ordering (Figure 3), indicating an organic origin. Furthermore, inorganic gases tend to exhibit relatively heavier carbon isotope values. According to Dai Jinxing et al., a δ¹³C₁ value of −30‰ can be used as a threshold to distinguish between organic and inorganic gases: δ¹³C₁ values > −30‰ are characteristic of inorganic gas, while δ¹³C₁ values < −30‰ indicate organic gas. In this study, the maximum δ¹³C₁ values for gases from the four stratigraphic intervals were −31.3‰, −32.6‰, and −33.6‰, respectively, all of which fall below the −30‰ threshold, further confirming the organic origin of the gases in the Dixi area (Figure 4).

Milkov and Etiope proposed one of the most comprehensive classification diagrams for natural gas origins based on geological and geochemical data from 20,621 natural gas samples across 76 countries and six continents. Their diagram refined the distribution fields for thermogenic gas, primary biogenic gas, secondary biogenic gas, and inorganic gas [40]. Building on this work, Milkov further optimized the classification of thermogenic gases by distinguishing coal-derived gas from oil-associated gas using a global geochemical database comprising over 30,600 samples [14]. Given the isotopic differences between coal-derived and oil-derived gases, numerous gas genetic classification diagrams have since been developed by domestic and international researchers [10,38,41]. Among these, the δ¹³C₁–δ¹³C₂–δ¹³C₃ ternary diagram for organically derived alkane gases reveals that four distinct genetic types of natural gas can be identified in the study area [38]. Type I and Type II gases are coal-derived; Type III is a secondary biogenic gas formed by biodegradation of associated oil at shallow burial depths (a mixture of thermogenic and microbial gas); and Type IV is oil-derived gas. Data points from the Carboniferous, Cretaceous, Jurassic, and Permian intervals in the study area are predominantly clustered within the Type I coal-derived gas field (Figure 5). Similarly, δ¹³C₁–δ¹³C₂ values for all four stratigraphic levels closely align with the typical range of humic (coal-derived) gas (Figure 4), indicating that natural gas in the Dixi area is primarily coal-derived in origin.

5.2. Origin of Multi-Origin Natural Gas

The high to over-mature thermal evolution stage provides a critical geochemical basis for identifying the genetic type and source of natural gas [42,43]. Within the thermal evolution sequence of source rocks, when vitrinite reflectance (Ro) reaches 1.3–2.0% (high maturity stage), gas generation from kerogen degradation enters a peak phase, and oil-cracking gas begins to dominate. When Ro exceeds 2.0% (over-mature stage), liquid hydrocarbons are completely cracked into dry gas, and the carbon isotope of methane becomes significantly heavier (δ¹³C₁ > −30‰). At this stage, the molecular composition, isotopic fractionation effects, and light hydrocarbon indicators (e.g., benzene/n-hexane ratio) can effectively distinguish oil-derived gas from coal-derived gas while excluding interference from biogenic gases. In addition, residual aromatic compounds and diamondoids preserved under over-mature conditions provide important geochemical markers for tracing the type of source material and the cracking pathway, thereby offering key indicators for evaluating deep to ultra-deep natural gas resources [44]. As shown in the δ¹³C₂ versus (δ¹³C₂–δ¹³C₁) cross-plot (Figure 6), the natural gas samples in the study area are primarily identified as middle to late-stage humic-derived gas, indicating that the associated source rocks have undergone high to over-mature thermal evolution.

Based on the δ¹³C₁ and δ¹³C₂ isotopic characteristics of natural gas, the distribution of samples can be compared with different types of kerogen (e.g., Type II and Type III). The diagram (Figure 7) includes the δ¹³C₁–δ¹³C₂ evolutionary fields associated with various kerogen types from typical basins worldwide, such as the Delaware/Val Verde Basin, Sacramento Basin, and Niger Delta [45]. Most of the natural gas data points from the study area—including formations such as the Carboniferous (C), Permian–Triassic (P–T), Jurassic (J), and Cretaceous (K)—fall within the field corresponding to Type III kerogen (Figure 7), indicating that the gas is predominantly derived from coal-bearing or humic organic matter. This finding is consistent with previous research on the types and thermal maturity of organic matter in the region, further confirming that the natural gas in the study area exhibits typical characteristics of coal-derived (humic) gas.

In addition to carbon isotopic signatures, the compositional characteristics of natural gas and their corresponding ratios also serve as critical indicators for determining gas origin [10,46]. The distribution and abundance of light hydrocarbons—such as methane, ethane, and propane—reflect their genetic signatures. For instance, high-maturity thermogenic gas typically exhibits heavier carbon isotopic values (e.g., more negative δ¹³C₁ values). In contrast, low-maturity organic gases may show a typical isotopic sequence (δ¹³C₁ < δ¹³C₂ < δ¹³C₃) [46]. Moreover, coal-derived gas generally features more negative δ¹³C₁ values, whereas oil-derived gas tends to have lighter methane isotopes, underscoring the isotopic distinction between these genetic types [10]. In this study, most of the natural gas samples from the four stratigraphic intervals are plotted within the kerogen-cracking gas field (Figure 8), indicating that these gases primarily originated from the thermal degradation of kerogen. Additionally, the relationship between the rate of change in ln(C₂/C₃) versus ln(C₁/C₂) helps differentiate primary versus secondary cracking gases from kerogen [47]. When the rate of change in ln(C₁/C₂) exceeds that of ln(C₂/C₃), the gas is interpreted as primarily derived from initial kerogen cracking; the reverse suggests secondary cracking. According to the results of this study, the ln(C₁/C₂) variation is significantly higher than that of ln(C₂/C₃) across the four stratigraphic intervals, suggesting that the natural gases in the study area are predominantly products of primary kerogen cracking.

5.3. Quantitative Calculation of Mixed-Sources of Natural Gas

The conventional research methods for mixed-source oil and gas are all the forward thinking mode from end-element oil and gas to mixed-source oil and gas [17]. For the oil and gas reservoirs with common mixed-source conditions, end-element oil and gas cannot be found [35]. Therefore, it is necessary to explore a reverse thinking mode from mixed-source oil and gas to end-element oil and gas.

In the multivariate mathematical analysis of mixed-source oil, the most fundamental requirement is that the composition data of crude oil must be quantitative data (such as the absolute concentration of biomarker compounds in crude oil). Peters conducted a comparative study, suggesting that if the data is not quantitative, the final calculated results will differ significantly from the actual results [48]. When applying this method to the research on mixed-source natural gas, a crucial issue is that the analytical data from natural gas is not quantitative data. For example, the composition of natural gas is generally the percentage content of certain carbonaceous components. Taking methane as an example, the methane composition data of natural gas is the total proportion of ¹²C₁H₄ and ¹³C₁H₄ in the entire natural gas (assuming no consideration of hydrogen isotopes). If the methane isotopes change, that is, the relative proportion of ¹²C₁H₄ and ¹³C₁H₄ changes, it is very likely that the percentage content of the entire methane component does not change. In this case, the isotopic composition of natural gas has actually changed, but it is not reflected in the composition ratio of the natural gas components. This will lead to a misjudgment of the mixed-source ratio. In addition, the isotopic values of natural gas are a ratio relative to the standard sample. Therefore, such data on the composition and isotopes of natural gas cannot be directly applied in the multivariate mathematical analysis of the proportion of mixed-sources.

In response to this key issue, in this study, a mathematical transformation approach was adopted to convert common natural gas composition and isotope data into quantitative data, so as to conduct subsequent research using multivariate mathematical analysis methods. It is known that the calculation formula for carbon isotopes is:

$${\delta }^{13}C=\left[\left({\displaystyle \frac{{}^{13}C/{}^{12}C_{\mathrm{sanple}}}{{}^{13}C/{}^{12}C_{\mathrm{standard}}}}\right)-1\right]\times 1000$$

(2)

where δ¹³C (in ‰, per mil) represents the relative difference in the ratio of ¹³C to ¹²C between the sample and an international reference standard; (¹³C/¹²C)_sample is the carbon isotopic ratio measured in the sample; and (¹³C/¹²C)_standard is the carbon isotopic ratio of the reference material, typically the Vienna Pee Dee Belemnite (VPDB) standard, with a known value of 11,237.2 × 10⁻⁶.

Therefore, it is not difficult to calculate (¹³C/¹²C)_sample. For methane (an alkane), there are only two types of carbon atoms, namely ¹³C₁ and ¹²C₁, so the sum of their percentage contents is 1. We can establish a system of linear equations in two variables:

$$\left\{\begin{array}{c}{}^{13}C_1÷{}^{12}C_1=k\\ {}^{13}C_1+{}^{12}C_1=1\end{array}\right.$$

(3)

where k represents the (¹³C/¹²C)_sample ratio. In the case of methane, the sample contains only one carbon atom. For clarity, we use symbolic notations such as ¹³C₁ and ¹²C₁ to denote methane molecules containing a single ¹³C or ¹²C atom, respectively.

The proportions of both ¹³C₁ and ¹²C₁ contained in the methane component of the natural gas can be calculated. By multiplying the proportion of the two by the proportion of the methane component in the total natural gas, the absolute content of ¹³C₁ or ¹²C₁ in the total natural gas can be obtained. Through incorporating the proportion of mixed methane, ethane, and propane gases using Equations (2) and (3), the ratio of the time the experiment of the absolute content of vigor and the ratio of gas components can be determined.

By using the multivariate mathematical statistics analysis method, the data in Table 1 except for the mixed ratio were calculated. The end-element contribution ratios obtained through the Alternating Least Squares algorithm are shown in

Table 2. The mixed proportion calculated through the multivariate data analysis method.

Area	Formation	δ13C1/%	δ13C2/%	δ13C3/%	13C1	12C1	13C2	12C2	13C3	12C3	Coal Type Gas /%	Oil Type Gas/%
Northern	P	−31.61	−29.56	−28.79	8629	821,441	471	19,068	210	19,298	62.2	37.8
	P	−31.14	−29.54	−28.45	8603	819,653	290	16,551	182	16,782	84.3	15.7
	T	−32.94	−31	−30.11	8657	824,569	414	18,427	203	18,659	67.8	32.2
Southern	J1b	−37.98	−27.88	−24.77	8724	830,938	372	17,842	196	18,075	68.4	31.6
	J2x	−39.03	−27.97	−25.94	8647	822,918	532	19,864	219	20,098	45.2	54.8
	J1s	−39.04	−27.68	−25.3	8567	815,091	650	21,525	237	21,760	25.6	74.4
Dongdaohaizi	J1b	−43.61	−25.1	−18.89	8613	819,303	694	22,033	243	22,269	10.6	89.4
	K1s	−44.93	−25.67	−22.9	8584	816,329	749	22,892	253	23,129	2.3	97.7

. The results show that the proportion of mixed-sources varies in different structural areas of the Di Nan Uplift. In the northern part of the Di Xi area, the proportion of coal-type gas mixed in the natural gas is the largest, accounting for approximately 67.8% to 84.3%. In the southern part of the Dixi area, most of the natural gas is a mixture, and the proportion of coal-type gas mixed in is approximately 25.6% to 68.4%. The natural gas in the Dishui Spring area is mainly oil-based gas, accounting for approximately 89.4% to 97.7%.

During the depositional period of the Carboniferous, the eastern part of the Junggar Basin was interpreted to have been situated within a foreland basin setting, while the western part was associated with a trench–arc–basin system. Under this tectonic regime, a thick succession of marine marine–continental transitional deposits was developed. Based on preliminary reconstructions, the extent of the Early Carboniferous marine basin in northern Xinjiang was estimated to exceed 40 × 10⁴ km², significantly surpassing the current geographical limits of the basin. Thick marine–transitional source rocks within the Carboniferous have been observed in surface outcrops and well data in the eastern basin [30]. These source rocks are considered more laterally extensive, more continuous, and of greater thickness, suggesting a higher potential for hydrocarbon generation.

These results not only improve our understanding of the gas charging history and accumulation pathways, but also provide a scientific basis for evaluating accumulation effectiveness and predicting favorable zones. Therefore, this work offers practical guidance for assessing natural gas resources and optimizing exploration strategies in the Junggar Basin and similar multi-source basins.

6. Conclusions

This study clarifies the genetic types and origins of natural gas in the Dixi area of the Junggar Basin through analysis of gas composition and stable carbon isotopes.

(1) The Cretaceous natural gas in the Dixi area originates from three distinct sources. It consists of both oil-type gas and coal-type gas. The oil-type gas is mainly derived from Permian lacustrine hydrocarbon source rocks. In contrast, the coal-type gas primarily originates from Carboniferous coal-measure source rocks.

(2) The proportion of mixed-sources varies in different structural areas of the Dinan Uplift. In the northern part of the Dixi area, the proportion of coal-type gas mixed in the natural gas is the largest, accounting for approximately 67.8% to 84.3%. In the southern part of the Dixi area, most of the natural gas is a mixture, and the proportion of coal-type gas mixed in is approximately 25.6% to 68.4%. The natural gas in the Dishui Spring area is mainly oil-based gas, accounting for approximately 89.4% to 97.7%.