The high chemical activity of carbonate minerals makes carbonates sensitive to postdepositional diagenesis, which can lead to the pore modification events such as primary pore filling and secondary pore formation, and thus endowing carbonates the potential to become high-quality reservoirs^[1]. These pore modification events and corresponding diagenetic environments have been extensively and deeply discussed in the past few decades, and the morphology^[2,3], distribution pattern^[4,5] and geochemical characteristics^[6,7] of cements are the main identification basis. And the diagenesis-porosity evolution geological model established on the basis of above research is the key to identifying the reservoir genesis and evaluating the accumulation effectiveness ^[2]. The marine carbonates in China are old in age, large in burial depth, located in the lower tectonic layer of superimposed basins, and subjected to multi-stages of diagenetic modification^[8], so two problems often emerge when applying traditional methods to study the diagenetic environment-porosity evolution of ancient carbonates in China. One, as the ancient carbonates have small scale and multi-stages of fabrics, thin section observation and whole rock analysis commonly used to study diagenetic environment traditionally are limited in applicability^[9], making it difficult to identify the reservoir genesis; thus using research methods with high spatial resolution has become an inevitable trend. The other, the petrographic observation that traditional porosity evolution study relies on can only qualitatively determine the relative sequence of porosity modification events by the crosscutting relationship of different fabrics, but can’t provide quantitative geochronological information of diagenetic events to identify porosity of the reservoir before hydrocarbon migration, and the identification of porosity before oil and gas migration is one of the critical contents in the evaluation of accumulation effectiveness. The element mapping and carbonate mineral laser U-Pb dating techniques emerged in recent years have the advantages of in-situ, high precision, and high spatial resolution, thus have promising prospects in solving the above two problems.

The dolomite reservoirs of the Upper Sinian Qigebrak Formation in the Tarim Basin develops various kinds of reservoir spaces, including framework (dissolution-enlarged) pores, spongiostromata fenestral pores and karst vugs and caves^[10,11,12]. They are close to the source rock of the Lower Cambrian Yuertusi Formation and potential source rock of Lower Sinian-Nanhua System vertically ^[13,14]. Besides, large amounts of bitumen can be found in the field outcrop, showing promising hydrocarbon prospect. Previous studies proposed that the Qigebrak Formation went through multi-stage diagenetic superimposition ^[15]. However, restricted by the above two issues, its diagenetic environment evolution, reservoir genesis and effective porosity before hydrocarbon migration remain unclear, and this has become one of the essential issues restricting reservoir prediction and accumulation effectiveness evaluation of the Qigebrak Formation.

In this study, taking the microbial dolomite reservoir of the Sinian Qigebrak Formation in the Xiaoerbulakexigou section of the Aksu area, northwestern Tarim Basin as an example, the laser element mapping and carbonate mineral laser U-Pb radiometric dating techniques are used to study diagenetic environment constrained by geochemistry and diagenesis-porosity evolution constrained by geochronology, to figure out the genesis, formation time of different kinds of reservoir spaces and effective porosity before hydrocarbon migration.

The Tarim Basin is the largest hydrocarbon-bearing superimposed basin in western China. It is surrounded by the Mt. Tianshan, Mt. West Kunlun and Mt. Altun, and is in a diamond shape. During the Nanhua Period, the Tarim Basin entered the “tensional rift” evolution stage under the influence of the break-up of Rodinia supercontinent, and the basin basement began to receive deposits^[16]. Since the Sinian Period, the activity of the mantle plume gradually declined, and the basin entered the intracratonic depression stage^[17]. On the basis of filling in the early Sinian, an open-sea carbonate ramp came up in the north of central paleo-uplift of the basin in late Sinian^[18] (Fig. 1a). Previous research suggested that the Upper Sinian Qigebrak Formation in the northwestern Tarim of the Sugetbrak Formation. At the end of the Sinian, the uplift and erosion caused by the Keping movement led to the development of weathered crust at the top of the Qigebrak Formation^[10]. Then at the onset of Cambrian, a rapid transgression occurred, leaving a set of high-quality source rock of Yuertusi Formation above the Qigebrak Formation. The study area has experienced four major regional tectonic events since the Keping Movement, namely, the late Caledonian movement, the late Hercynian movement, the Indosinian movement and the late Yanshanian movement^[19,20]. Among them, the late Caledonian and Indosinian periods were the most active periods of thermal events of the basin^[21]. The above-mentioned tectonic events could cause supergene dissolution, hydrothermal reformation and migration of diagenetic fluids, which had strong impacts on the formation, enrichment and reduction of the reservoir spaces^[8].

Basin was deposited in inner-middle ramp facies^[12], and was in parallel unconformable contact with the underlying clastic rocks

The Xiaoerbulakexigou section is a typical section of the Upper Sinian in the study area^[15]. Based on measurement results of the section, the Qigebrak Formation is divided into 4 lithological members mainly dominated by microbial dolostones (Fig. 1b): 1^st member, 22.55 m thick, interbedding of thin layered microbial dolostone and clastic rock; 2^nd member, 47.35 m thick, dominated by thin layered horizontal-microwave stromatolite dolostone, which forms several cycles with grain dolostone and thrombolite dolostone; 3^rd member, 43.00 m thick, thick-massive layered spongiostromata dolostone and thrombolite dolostone constitute several cycles, and the development degree of stromatolite decreases sharply; 4^th member, 60.60 m thick, the top 10.5 m is karst breccia, and the lower part is the same as that of the 3^rd member in lithology, but has a large number of horizontal dissolved fissures from several centimeters to several meters in length, with multi-phases of cements (Fig. 2a). They are distributed in the range of 10-50 m below the unconformity, and are deemed the product of karst phreatic zone^[10]. The dissolved fissures gradually decrease in scale downward, which is related to the gradual weakening of karstification strength. The dissolved fissures have macroscopic and microscopic characteristics similar to "grape lace structure" in the Dengying Formation, thus it is named "the lace structure (LS)" in this paper.

There are two types of reservoirs in the Qigebrak Formation (Fig. 1b). The middle reservoir section develops microbial dolostone reservoirs with framework (dissolution-enlarged)) pores and spongiostromata fenestral pores as main reservoir space that are controlled by sedimentary microfacies and high-frequency exposure ^[11,12]. The upper reservoir section develops karst reservoir with spongiostromata fenestral pores and karstic vugs and caves as main reservoir space that is controlled by sedimentary microfacies and the uplift at the end of Sinian.

Cement is the critical carrier for the study of carbonates diagenetic environment^[2]. Microscopic petrographic investigations revealed that 5 phases of dolomite cements were developed in various reservoir spaces of the Qigebrak Formation according to the shape, size and development sequence.

(1) Fibrous isopachous dolomite (FID): it is extensively developed in the reservoir spaces in the 4^th member. It is distributed in the form of thin rim at the edge of pores, dissolved vugs and fissures (Fig. 2b). The thin rim cement is composed of tightly spaced fibrous crystals perpendicularly to the edge of the host rocks (HR), and it contains multiple layers in alternate light and dark colors. Though this dolomite cement is quite common, it does not occupy a large proportion of the reservoir space.

(2) Bladed dolomite (BD): it is the second-stage cement developing in the dissolved fissures after the fibrous isopachous dolomite. The crystals are obviously larger than the fibrous dolomite and take on blade shape. Its aggregation also grows in a rim around the edge of dissolved fissures toward the center (Fig. 2c). However, unlike the extensively developed fibrous isopachous dolomite, this type of dolomite only develops in large-scale dissolved fissures, and is rarely seen in smaller pores and vugs.

(3) Powder-fine crystalline dolomite (PFCD): it is the most common cement and is widely distributed in the reservoir spaces. The crystals are clean and bright, which are significantly different from the darker fibrous isopachous and bladed dolomites. The crystals are usually less than 0.2 mm, and in anhedral and subhedral forms (Fig. 2d).

(4) Medium crystalline dolomite (MCD): it can only be found in large-scale vugs and dissolved fissures, indicating that the crystal size is proportional to the size of growth space. The crystals are also clean and bright, 0.25-0.5 mm in diameter, and in subhedral-euhedral form (Fig. 2e).

(5) Saddle-shape dolomite (SSD): mainly distributed in large-scale dissolved fissures and between the karst breccia at the top, and is the latest diagenetic product filling in the dissolved fissures. It has the characteristics of coarse crystal, saddle shape, undulatory extinction under cross-polarized light and accompanying quartz (Fig. 2f), indicating its hydrothermal origin.

The above cements are widely distributed in the Qigebrak Formation. In the middle reservoir section, due to the low influence of the Keping movement, only powder-fine crystalline and medium crystalline dolomite are developed, but their characteristics are totally consistent with that in the upper reservoir section. However, in the upper reservoir section, all the 5 phases of cements can be found in the dissolved fissures and has relative complete cement sequence. Thus, we select samples containing the lace structure in the 4th member, which can completely represent the diagenetic sequence and are the ideal material reflecting the diagenesis-porosity evolution history of the Qigebrak Formation. At the same time, powder-fine and medium crystalline dolomite, which fill in the reservoir space in large quantities, are the main cause for the porosity loss. Thus clarifying the diagenetic environment and geological time of multi-phases of dolomite cements not only provides important information for the establishment of diagenesis-porosity evolution history, but is also of great significance to the identification of porosity before hydrocarbon migration and the effectiveness of accumulation.

All samples in this study were collected from the Upper Sinian Qigebrak Formation of the Xiaoerbulakexigou section in Aksu area, northwestern Tarim Basin. The sampling location and characteristics of the samples are shown in Figs. 1b and 3. According to the preparation process introduced by Shen et al.^[22], the hand specimens were fixed in 2.5 cm diameter epoxy resin blocks or made into thin sections 100 μm thick. Then the samples were pre-treated in the ultra-clean laboratory to eliminate the surface contamination.

3.1.1. Laser element mapping

With advances in analytical techniques, the elemental content analysis has evolved from whole rock analysis to in-situ spot analysis ^[9]. The multi-stage cements in ancient carbonates are often different greatly in composition, thus visually displaying the plane changes of element content within millimeter-centimeter zones by element mapping technique is critical to understand the formation process of the cements^[23,24]. A variety of micro-area analysis instruments have been applied to quantitatively mapping the sample element content at home and abroad, but they have advantages and disadvantages in terms of detection limit, resolution, analysis procedure and cost^[24]. The recently emerging LA-ICP-MS element mapping has the advantages of short analysis time, low cost, simple sample preparation process, low detection limit and simultaneous multi-element analysis, which expands the application of mapping technique in earth science. Element mapping was mainly used to study the formation process and environment of metal minerals ^[25], magmatic minerals^[26] and speleothem^[27,28]. In this work, taking the Qigebrak Formation as an example, we applied the technique to the diagenesis process and environment study of ancient carbonate rocks (Fig. 4).

3.1.2. Carbonate mineral laser U-Pb radiometric dating

Carbonates are widely distributed in various geological environments, and carbonate mineral dating has broad application prospects in determining the age of geological events^[29], but the Rb-Sr, K-Ar, Re-Os and Nd-Sm radiometric dating methods cannot obtain stable and reliable ages of carbonate minerals^[30]. Moorbath et al.^[31] first confirmed the feasibility of U series dating in carbonates, then Smith and Farquhar^[32] successively reported examples of U-Pb radiometric dating of low-uranium carbonates (uranium content of (100-500)×10^-9), making U-Pb radiometric dating method the only absolute geochronometer currently applicable to carbonates^[33]. This method has been widely used to determine the time of mineral transformation of aragonite^[34], fault slip^[35], dolomitization^[36], regional tectonic and diagenetic fluid activity^{[22, 33]} by direct dating speleothem^[37], calcareous concretion^[38], calcareous fossils^[34], calcite veins^[35], seepage reflux dolostone^[36] and vug cements^{[22, 33]}, etc. Especially for diagenetic cements, determining the absolute age of the events that led to their precipitation can deepen the understanding on the overall evolution of reservoir and provide support for the study of thermal and accumulation history.

In the past, carbonates U-Pb dating relied on the isotope dilution method, but pretreatment procedures such as microdrilling and chemical purification are extremely tedious and time-consuming, and have the potential risk of contaminating samples, which limit the promotion of this method. With the discovery of carbonate laser U-Pb dating reference material such as ASH-15^[37], WC-1^[39] and AHX-1^[22], the laser ablation microsampling method has been applied to carbonate U-Pb dating successfully. This technique uses a 30 μm-250 μm laser spot to ablate 30-60 times within monogeneration cement, and the obtained U/Pb ratio with a wide variation range can be used to fit ideal isochrons. Compared with the isotope dilution method, the laser ablation method has the advantages of high resolution (> 5 μm), high accuracy, high success rate, and high analysis speed, which provides support for micro-area research^[33]. In this work, taking the Qigebrak Formation as an example, we applied the technique to the study of diagenesis-porosity evolution and identification of porosity before hydrocarbon migration of ancient carbonate formation (Fig. 5).

Laser element mapping was done at the Radiogenic Isotope Facility (RIF) of the University of Queensland using a ASI RESOlution SE laser ablation system connected to a Thermo iCap-RQ ICP-MS. The ablation used a 50 μm-square spot at a laser repetition rate of 20 HZ, the laser energy of 3 J/cm², and a spot movement rate of 0.05 mm/s. NIST 614 was used as the external calibration standard^[40], and 2D element maps were built with Iolite 3.6 after raw data processing^{[26, 41]}.

Carbonate mineral laser U-Pb radiometric dating was also done at RIF of the University of Queensland using an ASI RESOlution SE laser ablation system connected to a Thermo iCap-RQ ICP-MS or Nu Plasma II MC-ICPMS. The ablation used a 100 μm-round spot at a laser repetition rate of 10 HZ, laser energy of 3J/cm², and an ablation time of 15-25 s. NIST614^[40] and WC-1^[39] were used as reference material to normalize element concentration and isotope ratio, and in-house reference material AHX-1 was used to monitor the regressed ages^[22]. The sample ages were regressed on Tera-Wasserburg plots using Isoplot 3.0^[42] after raw data were processed by Iolite 3.6^[41].

The tests of laser carbon and oxygen isotopic composition, strontium isotopic composition and Cl microscopy were completed at the Key Laboratory of Carbonate Reservoir (KLCR) of CNPC. For detail testing methods, see Pan et al.^[43].

4.1.1. Fibrous isopachous dolomite

The fibrous isopachous dolomite is generally considered to be the typical product at sea bottom, which is formed by metasomatic alteration of aragonite or Mg calcite precursors^[44,45]. In the element mapping images, the concentrations of Li, Mn, Fe, Sr, Ba, Th and U of fibrous isopachous dolomite are consistent with those of the host rock, reflecting they were formed in similar formation environments. The fibrous isopachous dolomite yielded δ¹⁸O values ranging from -6.656‰ to -4.810‰ vPDB, which is slightly less positive than that of host rock, and it yielded δ¹³C values ranging from 2.033 ‰ to 2.158‰ vPDB, which is consistent with that of host rock. The ⁸⁷Sr/⁸⁶Sr value is 0.708 840, which are within the range of contemporary seawater composition ^[46]. Cathodoluminescence is generally dull luminescent, which is also quite consistent with the host rock, but bright luminescent bands can be found in this cement. Therefore, it is inferred that the fibrous isopachous cement is the product of seawater cementation.

Specially, the concentrations of Mn, U changes significantly between the light and dark layers of the fibrous isopachous dolomite. The dark layers with high Mn concentration (about 500×10^-6) appear as bright red-yellow bands in the element mapping images, but the light layers with low Mn concentration (smaller than 100×10^-6) appears as a blue bands (Fig. 4g). Previous studies have shown that ferruginous sea-water prevailed globally at the end of Sinian, which was characterized by suboxic surface seawater and anoxia reduced deep water^[47,48]. Such paleo-oceanic environment led to high Mn and Fe concentrations of the Sinian dolostones around the world, including the Qigebrak Formation^[48,49]. The frequent fluctuation of sea-level and the paleo-oceanic properties of Precambrian when the fibrous isopachous dolomite was formed may be the causes that the fibrous isopachous dolomite exhibit special element distribution pattern. The Precambrian seawater redox model proposed by Hood and Wallace^[48] can be directly used to explain this phenomenon. During the fall of sea-level, the oxic-suboxic surface seawater led to the depletion of Mn in the form of poorly crystalline Mn oxide phases^[50], resulting in a lower Mn concentration in the cement. While during the rise of sea-level, the upwelling of more reduced seawater caused the rapid dissolution of Mn oxide, and thus leading to the abrupt increase of Mn concentration in the precipitating water^{[48, 50]}. In oxic environment, U was absorbed strongly onto Mn-oxides^[51]. The dissolution of Mn oxide during transgression was also accompanied by the release of U, thus the U concentration co-varies with Mn concentration in the light and dark layers. But the Fe oxyhydroxides are more stable at lower Eh values, thus the distribution of Fe concentration is more uniform^[48]. Meanwhile, the cathodoluminescence of carbonates is strongly controlled by Fe²⁺ (quencher) and Mn²⁺ (activator)^[48], which leads to the periodic change of the cathodoluminescence intensity of the light and dark layers of fibrous isopachous cement.

4.1.2. Bladed dolomite

The element mapping images show that the bladed dolomite has uniform element concentration and does not develop ring structure, indicating it was formed in a relatively stable environment. The dolomite in sample Q-58-1-2 shows blade shape under plane and cross polarized light (Fig. 4a, 4b), and it has much lower Mn and Fe concentrations in the element mapping images (about 100×10^-6). Research on modern seawater and freshwater shows that the Mn and Fe concentrations in freshwater are about 0.4×10^-9 and 20×10^-6, which are 50 and 197 times of the seawater concentration respectively ^[49]. However, due to the extremely high concentrations of Mn and Fe in Sinian seawater ^[48], the sudden drop of the concentrations of these two elements reflects the weakening of the influence of seawater then. The Sr concentration is lower than that of host rock and fibrous isopachous dolomite, indicating that the salinity of the diagenetic fluid has decreased. The bladed dolomite has δ¹⁸O values ranging from -6.65‰ to -6.61‰, suggesting that its diagenetic environment increased in reducibility. It has δ¹³C values from 1.001‰ to 1.598‰, which is somewhat lower than that of the fibrous dolomite, while it has slightly higher ⁸⁷Sr/⁸⁶Sr values of 0.708 85, indicating atmospheric freshwater mixed into the water system during its formation^[49]. In addition, the bladed dolomite exhibits dull luminescence in CL test. The above evidences shows that the bladed dolomite was affected by atmospheric freshwater, but it still has many geochemical features inheriting from the fibrous isopachous dolomite, thus it is concluded that the bladed dolomite was formed in confined seawater affected by freshwater.

4.1.3. Crystalline dolomite

Crystalline dolomite is the most common cement in the Qigebrak Formation, which is widely distributed in the middle and upper reservoir sections. According to the difference of crystal size, the crystalline dolomite is divided into powder-fine and medium crystalline dolomites. The former partially or fully fills microbial framework (dissolved) pores, spongiostromata fenestral pores and small-scale lace structure (Figs. 2d-e, 3g), and thus is the main cause of porosity loss. The latter is only found in large-scale vugs and lace structures, occupying the space left by the powder-fine crystalline dolomite (Figs. 2e and 4a). According to Moore^[2], Sr has a very low partition coefficient in calcite and dolomite in burial environment, thus low Sr concentration is a sign of buried cement; in contrast, U is in a soluble oxidation state in oxic water and it would be reduced to insoluble state in the anoxic environment^[51]. Therefore, with the gradual enhancement of the reducibility in the diagenetic environment, U would be strongly depleted and drop sharply in concentration. The low concentrations of Sr and U of crystalline dolomite in the element mapping images undoubtedly indicate the burial diagenetic environment, and the crystal size is proportional to the growth space. The two types of crystalline dolomite did not change significantly in δ¹³C values, while have δ¹⁸O values turning more negative.

Generally, as the burial depth and the reducibility increase, low-valence Mn and Fe ions would replace Ca and Mg ions in the crystal lattice of dolomite, resulting in increase of the Mn and Fe concentrations^[2]. The powder-fine crystalline dolomite of sample Q-58-1-2 shows bright red to yellow bands with high Mn and Fe concentrations in the element mapping images. In contrast, the medium crystalline dolomite had lower Mn and Fe concentrations (Fig. 4g-4h), indicating that it might be affected by extraneous fluid during its slow crystallization process. However, the concentrations of Sr and U and oxygen isotopic composition suggest that the medium crystalline dolomite was still formed in a highly reduced environment, thus it is speculated that the medium crystalline dolomite may be affected by deep-circulation groundwater^[52].

4.1.4. Saddle-shape dolomite

The saddle-shape dolomite is mainly distributed in large- scale lace structure and between karst breccia, and is the latest diagenetic product filling the residual reservoir space. It comes in coarse saddle-shape crystals, shows undulatory extinction, and is associated with quartz (Fig. 2f), undoubtedly indicating its hydrothermal fluid origin ^[49]. Both its δ¹⁸O and δ¹³C become more negative, reaching -3.53‰ to -3.51‰ and -10.39‰ to -9.50‰ respectively, which also suggests the strong influence of hydrothermal fluid^[49]. In addition, the saddle-shaped dolomite has ⁸⁷Sr/⁸⁶Sr values significantly higher than other dolomites, indicating that its diagentic fluid is affected by crust-derived strontium with a high ⁸⁷Sr/⁸⁶Sr value.

Combining the above analysis with the age data in Table 1, it is inferred that the Qigebrak Formation reservoir in the Xiaoerbulakexigou section has experienced the following diagenetic environment and process: 1. Ediacaran sedimentary environment (constraining age (576±16) Ma), multiple types of microbial dolostones, mainly horizontal-microwave stromatolite dolostone and spongiostromata dolostone were formed. 2. Freshwater diagenetic environment (constraining age (560±26) Ma): the Keping movement caused the exposure of the Qigebrak Formation, and then the formation suffered atmospheric freshwater leaching, giving rise to non-fabric selective dissolved fissures and caves. 3. Marine diagenetic environment (constraining age (556±17) Ma): transgression again led to the precipitation of fibrous isopachous dolomite cement at the edge of the dissolved vugs and fissures. 4. Extremely shallow buried diagenetic environment (constraining age (542.7±8) Ma): bladed dolomite was formed in sealed seawater affected by atmospheric freshwater. 5. Buried diagenetic environment (constraining age (486.3±6.8) Ma): buried diagenetic fluid dominated the precipitation of crystalline dolomite, but the activities of deep-seated faults might make the crystalline dolomite affected by atmospheric freshwater intermittently. 6. Hydrothermal diagenetic environment (constraining age (215±30) Ma): tectonic thermal events in the early Indosinian led to the precipitation of saddle-shape dolomite.

The diagenetic environment evolution of the Qigebrak Formation was finely constrained by combining element mapping with other geochemical tracers, to provide a reliable basis for the identification of reservoir genesis. The reservoir spaces in the microbial dolostone reservoir of Qigebrak Formation were mainly formed in the sedimentary environment and freshwater diagenetic environment before burial, while the dolomite cements formed in marine, burial and hydrothermal environments caused gradual filling of the reservoir spaces.

4.2.1. Ages of host rock and dolomite cements

Ages of host rock: The dolomite host rock samples from the middle part (sample Q-58-1-2) and top (sample Q-76-1) of the Qigebrak Formation were dated at ages of (576±16) Ma and (560±26) Ma, respectively (Fig. 5a, 5b). Several chronological studies of Sinian System in the study area were conducted: Xu et al.^[17] dated the zircon in the basalt samples of the Lower Sinian Sugetbrak Formation at (615.2±4.8) Ma and (614.4±9.1) Ma. Li et al.^[53] dated the detrital zircon in the fine sandstone sample from the upper part of the Sugetbrak Formation of the Wushi section at (602±23) Ma. The ages of the dolomite host rock samples in this study are consistent with the age of the Ediacaran Period (635-541 Ma) and the regional zircon chronological framework, and represent the sedimentary age of the stratum. Besides, Wang et al.^[54] proposed that the matrix dolomite in the Qigebrak Formation originated from microbially mediated primary precipitation, thus the ages may also represent the age of early dolomitization.

Ages of dolomite cements: Based on the obtained U-Pb ages, it is concluded that the cements filling in the reservoir space of the Qigebrak Formation were formed in 4 stages: fibrous isopachous dolomite, bladed dolomite, crystalline dolomite, and saddle-shape dolomite stages. The cement sequence in the large dissolved vugs is relatively complete, while that in smaller pores (such as spongiostromata fenestral pores) is mostly incomplete.

(1) Fibrous isopachous dolomite: yielded 2 isochrons at (553±20) Ma and (556±17) Ma, respectively (Fig. 5c, 5d). The ages are quite close to the stratigraphic age of the top of the Qigebrak Formation, indicating a rapid sea-level fall soon after the syndeposition, and then the Keping movement led to the exposure and the formation of dissolved vugs and caves. When the sea-level rose again, the fibrous isopachous dolomite precipitated as the first-stage cement in the dissolved fissures.

(2) Bladed dolomite: the bladed dolomite yielded two almost isochronous ages of (542.7±8) Ma (Fig. 5e) and (542±26) Ma. Geochemical evidence shows that bladed dolomite was formed in the sealed seawater affected by freshwater, indicating that the Qigebrak Formation was still affected by supergene karstification and was kept in very shallow burial diagenetic environment then. Both the fibrous isopachous and bladed dolomite represented the early cementation. The U-Pb ages of them are equivalent to or slightly later than that of the host rock, which strongly suggests that the reservoir space was mainly formed in the sedimentary environment and supergene dissolution before burial, but not the product of burial dissolution.

(3) Crystalline dolomite: The U-Pb age of powder-fine crystalline dolomite in this study is (486.3±6.8) Ma (Fig. 5f, 5g), and the age of medium crystalline dolomite is (472.3±7.7) Ma (Fig. 5g), which are consistent with the diagenetic sequence identified by microscopic petrographic investigation. Although the ages of the two types of crystalline dolomite have a difference of 14 Ma, they were both formed in the early Caledonian. In this period, the subduction of the North Kunlun Ocean to the west-central Kunlun Island arc led to the formation of multiple uplifts such as Tazhong, Hetian and Tabei^[20]. The extensive uplift movement at this time led to the shift between high and low potential energy regions, resulting in the precipitation of buried cement^[8]. The associated faulting could significantly enhance the influence of atmospheric freshwater on the underlying strata, which is consistent with the speculation that the medium crystalline dolomite was affected by atmospheric freshwater.

(4) Saddle-shape dolomite: yielded an isochron at (215±30) Ma (Fig. 5h). The U-Pb age shows that it was formed in the early of Indosinian, which may be related to the strong uplift and fold movements caused by the collision between the Qiangtang terrane and the Tarim craton ^[20]. The active tectonic movement also made the Permian-Triassic become one of the most active periods of deep thermal events in the Tarim Basin^[21].

4.2.2. Diagenesis-porosity evolution history

Based on the diagenetic environments revealed by geochemistry and U-Pb radiometric ages, the diagenetic environment evolution history and diagenesis-porosity evolution curves of both the middle and upper reservoir sections were established (Fig. 6), and the effective porosity before hydrocarbon migration was identified.

According to the estimates based on the initial porosity of modern microbialite^[55], plunger porosity and the microscopic distribution of cement and residual pore^[2], the initial average porosity of the middle reservoir section with mainly stromatolitic framework pores and spongiostromata fenestral pores was estimated at 20%. The Keping movement at the end of the Sinian has almost no effect on the middle reservoir section. During the early Caledonian, crystalline dolomite precipitated in the stromatolitic framework (dissolved) pores and spongiostromata fenestral pores, making the porosity drop to 6%, which has been maintained to the present.

According to the microscopic pore and cement distribution, it is speculated that the spongiostromata dolostone had a very high initial porosity (Fig. 3g). Thus the initial porosity of the upper reservoir section, whose reservoir rock is dominated by spongiostromata dolostone, is estimated at 30%. The dissolution pores formed at the end of Sinian made the average porosity of the reservoir increase to 35%. Subsequently, the fibrous isopachous and bladed dolomites partially filled the reservoir spaces, making the average porosity reduce to 28%. In the early Caledonian, the precipitation of powder-fine crystalline dolomite caused the porosity to drop sharply to 15%. Then the medium crystalline dolomite mainly filled larger residual pores, making the average porosity decrease further to 10%. The hydrothermal fluid activities in the early Indosinian caused filling of the reservoir, but the reservoir didn’t change much in porosity, with the average porosity reducing to 8%, and this porosity has been kept to now.

Based on the diagenesis-porosity evolution history of the microbial dolostone reservoirs in the Qigebrak Formation, combined with the burial history of the Qigebrak Formation and the thermal history of the high-quality source rock in the Yuertusi Formation, the time of hydrocarbon migration, effective porosity before oil migration and accumulation stages were evaluated. The upper reservoir section of the Qigebrak Formation is stably distributed in Keping-Tabei uplift. In the Early Ordovician, the source rock of the Yuertusi Formation in Keping-Tabei area entered the hydrocarbon generation threshold, and reached mature-high mature stage in scattered areas like southeast Awati Sag, so this period is the earliest favorable hydrocarbon accumulation period for the Precambrian-Lower Paleozoic reservoir in this area^[56]. In the Late Caledonian stage, the Yuertusi Formation deposited in the Manjar Sag and Awati Sag experienced rapid thermal evolution and entered the massive liquid hydrocarbon generation period. The generated hydrocarbons migrated to the northern uplift, and at this period the dolomite reservoir of the Qigebrak Formation still maintained at a porosity of 8% to 11% after the precipitation of crystalline dolomite, which is the main period of accumulation^[56]. Affected by tectonic movements, the Yuertusi source rock had secondary hydrocarbon generation and liquid hydrocarbon cracking during the Hercynian. The gaseous hydrocarbon generated in the slope area of the Manjar Sag and Awati Sag migrated to the adjacent Keping-Taibei ancient uplift, and the reservoir of the Qigebrak Formation maintained at a porosity of 6% to 10% at this point. Since the Middle Permian, the Yuertusi Formation in the Manjar Sag and Awati Sag generally has entered the overmature stage and no longer has the ability of generating hydrocarbon, and the late Himalayan stage was the adjustment period of paleo-hydrocarbon reservoirs. The above analysis shows that the microbial dolomite reservoir of the Qigebrak Formation still maintained desirable reservoir properties during the peak hydrocarbon generation period of the Yuertusi Formation and had the potential to accumulate effectively.

Supported by petrographic investigations, the element mapping, stable isotopic composition, strontium isotopic composition and cathodoluminescene analysis were performed on the various phases of dolomite cements precipitated in the vugs. It is proposed that the microbial dolomite reservoirs of the Qigebrak Formation successively experienced freshwater, marine, extremely shallow burial, buried and hydrothermal diagenetic environments after synsedimentary dolomitization. It is revealed that the reservoir space was mainly formed in the sedimentary environment (primary pores) and freshwater diagenetic environment (supergene dissolution pores and vugs) before burial. The seawater, burial and hydrothermal diagenetic environment caused the gradual filling of dolomite cements.

Based on the understanding of diagenetic environment and reservoir genesis, the U-Pb dating was carried out on the dolomite cements in the vugs and dissolved channels, and the diagenesis-porosity evolution curves of the microbial dolomite reservoir of the Qigebrak Formation were established. It is revealed that the cement precipitation and porosity loss mainly occurred in the Early Caledonian. During the peak hydrocarbon generation periods of the Yuertusi Formation, which were the Early Caledonian, Late Caledonian and Late Hercynian, the porosity of the Qigebrak Formation reservoir still maintained 6%-10%, indicating the potential of effective accumulation and hydrocarbon exploration.

The application of element mapping and carbonates U-Pb dating techniques in the microbial dolomite reservoirs of the Qigebrak Formation not only provides an approach to precisely constrain the diagenetic environment changes and reconstruct diagenesis-porosity evolution history constrained by geochronology, but also sheds lights on the study of reservoir genesis of ancient carbonates, the identification of effective porosity before hydrocarbon migration and the analysis of effective accumulation assemblages.