The nature, type, and origin of diagenetic fluids and their control on the evolving porosity of the Lower Cambrian Xiaoerbulak Formation dolostone, northwestern Tarim Basin, China

The study on Lower Cambrian dolostones in Tarim Basin can improve our understanding of ancient and deeply buried carbonate reservoirs. In this research, diagenetic fluid characteristics and their control on porosity evolution have been revealed by studying the petrography and in situ geochemistry of different dolomites. Three types of diagenetic fluids were identified: (1) Replacive dolomites were deviated from shallow burial dolomitizing fluids, which might probably be concentrated ancient seawater at early stage. (2) Fine-to-medium crystalline, planar-e diamond pore-filling dolomites (Fd1) were likely slowly and sufficiently crystallized from deep-circulating crustal hydrothermal fluids during Devonian. (3) Coarse crystalline, non-planar-a saddle pore-filling dolomites (Fd2) might rapidly and insufficiently crystallize from magmatic hydrothermal fluids during Permian. Early dolomitizing fluids did not increase the porosity, but transformed the primary pores to dissolution pores through dolomitization. Deep-circulating crustal hydrothermal fluids significantly increased porosity in the early stages by dissolving and then slightly decreased the porosity in the late stage due to Fd1 precipitation. Magmatic hydrothermal fluids only precipitated the Fd2 dolomites and slightly decreased the porosity. In summary, Devonian deep-circulating crustal hydrothermal fluids dominated the porosity evolution of the Lower Cambrian dolostone reservoir in the Tarim Basin.


Introduction
Dolostone reservoirs are important components of carbonate hydrocarbon reservoirs in many petroliferous basins worldwide (Ehrenberg et al. 2006;Li et al. 2011a, b;Sonnenberg and Pramudito 2009;Sun 1995;Warren 2000;Zhao et al. 2005), but dolomite genesis and the mechanism for generating dolostone reservoirs remain intensely debated (Hardie 1987; Kirmaci and Akdag 2005;Machel 2004;Morrow 1998;Warren 2000;You et al. 2015). Recent geophysical research found the Lower Cambrian platform margin facies dolostones in the subsurface of the Tabei Uplift (Ni et al. 2015), which made the Lower Cambrian Xiaoerbulak Formation dolostones become potential exploration targets for hydrocarbon reservoirs (Du and Pan 2016;Liu et al. 2017). Limited by scarce well samples, researches on the Lower Cambrian dolostones were mainly conducted in the Sugetbulak outcrop area of the northwestern Tarim Basin, where abundant pores, bitumen, and plentiful pore-filling dolomites were found (Li et al. 2011a(Li et al. , b, 2015Song et al. 2014). Recent researches show that high-quality reservoirs in the Xiaoerbulak Formation are mainly distributed in the platform margin facies dolograinstones (Li et al. 2015;Song et al. 2014) and a few microbial dolostones (Li et al. 2015;Song et al. 2014).
The Lower Cambrian dolostone reservoirs are regarded as one of the most ancient and deeply buried carbonate reservoirs in the world (Li et al. 2016;Pan et al. 2012;Zhang et al. 2014). However, the formation mechanism of the Edited by Jie Hao high-quality dolostone reservoirs remains unknown. Due to the lack of systematically petrographic and geochemical studies, the origin and nature of different diagenetic fluids remain debated. Previous studies concluded that the genesis of the Lower Cambrian dolostones was controlled by one or a mixture of three types of diagenetic fluids: meteoric water, concentrated seawater or heated formation water, and hydrothermal fluids (Cai et al. 2008;Ji et al. 2013;Li et al. 2011a, b;Pan et al. 2012;Zhang et al. 2011Zhang et al. , 2014Zhu et al. 2010). However, the correlation between different types of dolomites and their forming fluids remain unclear, as well as the origin and nature of these different diagenetic fluids (Zhang et al. 2014). Some studies concluded that the saddle dolomite and recrystallized dolomites were precipitated from magmatic hydrothermal fluid (Chen et al. 2009a;Dong et al. 2013;Pan et al. 2009;Zhao et al. 2012;Zhu et al. 2010). Other studies proposed that these dolomites were precipitated from stratigraphic hydrothermal fluids (Pan et al. 2012), such as the heated formation water from the Cambrian dolomite strata (Qian et al. 2012;Zhang et al. 2009Zhang et al. , 2011. Moreover, the influence of the diagenetic fluids on the formation of such a reservoir is also unclear. For example, previous studies (Li et al. 2011a, b;Li et ai. 2016) concluded that primary pores were strongly cemented, and the high-quality dolostone reservoirs were mainly resulted from the corrosion of deep hydrothermal fluids based on the geochemistry property of pore-filling dolomites. However, recent studies have also identified abundant interparticle and intercrystalline pores without any pore-filling dolomites, thus concluding that the primary pores were only slightly cemented and further dissolution occurred during epidiagenesis (Li et al. 2015;Shen et al. 2016).
Based on the basin's sedimentary, tectonic evolution and systematically petrographic and in situ geochemical research on different types of dolomites and reservoir spaces in the dolostones of the Lower Cambrian Xiaoerbulak Formation, this study (1) identified the types, nature, and origin of diagenetic fluids for different dolomites; (2) clarified the influence of different diagenetic fluids on the formation of reservoir spaces; and (3) explained the porosity evolution of the high-quality dolostone reservoir controlled by multiplestage diagenetic fluids. The results of this study can improve our understanding of very ancient and deeply buried carbonate reservoirs.

Geological setting
The Tarim Basin is the largest basin in China with an area of nearly 560,000 km 2 (Wang et al. 2009). The basin is located in northwestern China (Fig. 1a), surrounded by the Tian Shan Mountains and West Kunlun-Altun Mountains in the north and south, respectively (Fig. 1b). The basin has undergone a multiple-stage history of tectonic evolutionary processes, e.g., the Caledonian, Hercynian, Indosinian, and Himalayan cycles (Tang 1997). The six evolution stages during the Phanerozoic can be generally divided into the (1) intra-cratonic extensional basin stage during the Sinian-Early Ordovician, (2) intra-cratonic compressional basin stage during the Middle Ordovician-Middle Devonian, (3) back-arc extensional basin stage during the Late Devonian-Early Permian, (4) retro-arc foreland basin stage during the Late Permian-Triassic, (5) collisional reactivated foreland basin stage during the Jurassic-Paleogene, and (6) Indian-Tibetan collisional successor basin stage during the Neogene-Quaternary (Li 1995;Li et al. 2016;Zhang et al. 2009). Now, the basin is roughly divided into nine principal structural units, including four major uplifts, the Tabei, Tazhong, Bachu, and Southeast, and five major depressions (Fig. 1b), the Kuche, North, Tanggu, Southwest, and Southeast (Wang et al. 1992;Zhang et al. 2014). As a giant petroliferous basin, it contains abundant petroleum and natural gas resources, and the verified hydrocarbon reservoirs include Cambrian and Ordovician marine carbonate reservoirs, Silurian to Carboniferous marine clastic reservoirs, Permian volcanic reservoirs, and Mesozoic and Cenozoic terrestrial fluvial clastic reservoirs (Gu et al. 2002;Pu et al. 2011).
Dolostones are indispensable components of marine carbonate reservoirs in the Tarim Basin (Zheng et al. 2007). Dolostone strata of the basin mainly developed from the Upper Sinian Qigbulak Formation to Lower Ordovician Penglaiba Formation. These dolostones are divided into post-salt and pre-salt petroliferous sequences by regionally distributed Middle Cambrian evaporites interlaying between them (Fig. 2). Numerous wells have been drilled in the post-salt sequences, and Upper Cambrian-Lower Ordovician dolostone reservoirs have been proven excellent (Chen et al. 2009b;Huang et al. 2012). However, the pre-salt sequence is buried underground to depths of 6000-8000 m in the uplifts and 8000-12,000 m in the depressions, overlain by thick Middle Cambrian evaporites (Zhu et al. 2014). As a result, only a few deep wells have been drilled into Lower Cambrian stratum and core samples are rare. Hence, observational and geochemical studies of the Lower Cambrian dolostones are primarily conducted at field sites.
This study was conducted in the Aksu Area of the northwestern Tarim Basin (Fig. 1c), where the Lower Cambrian sequences are well exposed (Fig. 3a). The area lies within the Kalpin Uplift Unit, where a series of folds and thrust belts trend northeast due to the compressional orogeny of the southern Tian Shan Mountains in late Cenozoic (Burchfiel et al. 1999;Li et al. 2015Li et al. , 2016. As a result, the Lower Cambrian stratum is segmented and distributed in several thrust fold belts (Fig. 1c). Preliminary research of the Lower Cambrian dolostones in this area was performed on eight field outcrop sections by our group (Li et al. 2015), which correlated stratigraphy, classified sedimentary facies, measured porosity, and evaluated the reservoirs (Fig. 1c). The Xiaoerbulak Formation is considered the main reservoir interval of the Lower Cambrian (Li et al. 2015;Shen et al. 2016) because of the intensely developed pores (Fig. 3b,c). From the bottom upward, this formation is composed of platform margin facies micritic-microspar dolostones (Fig. 3d), agglutinated microbial reef dolostones (Fig. 3e), dolograinstone (Fig. 3f), and laminated microbial dolostones (Fig. 3g) (Li et al. 2015). Dolograinstone is the dominant high-quality reservoir due to its high average porosity (7%-8%) and the large-scale distribution surrounding the Lower Cambrian platform margin, with thicknesses of ~ 40 m and width of approximately 25 km (Li et al. 2015).

Methods
A total of 267 dolostone samples were collected from eight outcrop sections (Fig. 1c), covering all lithofacies described in Sect. 2. Sampling points and serial numbers were marked on corresponding lithological columns (Fig. 4). Thin sections were separately cut from all samples to study the petrographic and reservoir porosity. Dilute hydrochloric acid and alizarin red-S were used to examine the mineralogy in hand specimen and thin sections (Friedman 1959). The scheme of Gregg and Sibley (1984) and Sibley and Gregg (1987) was adopted for the petrographic descriptions and classifications. Descriptions of the reservoir porosity used the classification system of Choquette and Pray (1970). Thirteen representative samples, marked with black stars in Fig. 4, were selected for in situ geochemical experiments. Three types of replacive dolomites (Rd) and two types of pore-filling dolomites (Fd) are classified based on petrography. Due to the sedimentary textures and heterogeneous recrystallization of the samples, three types of replacive dolomites (Rds) commonly have intergrowth relationships. However, their crystals are too small to separate at macroscopic scale and can only be distinguished under microscope. Two types of pore-filling dolomites (Fds) usually develop in pores and small fractures, and they often have an intergrowth relationship , which is occasionally accompanied by late-stage filling calcites (Fig. 5b  with microdrilling due to their mixing and the disturbance of replacive dolomites and filling calcites. Therefore, to avoid mixing between the different generations, the in situ LA-ICP-MS method was adopted in this study. A total of 66 test spots were designed to measure the in situ trace and rare-earth elemental composition of different types of dolomites. The tests were accomplished at the Key Laboratory of Orogenic Belts and Crustal Evolution, Peking University, using Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS). In situ LA-ICP-MS analysis has been widely used on carbonates due to its high sensitivity, excellent spatial resolution (Kamber and Webb 2007;Jochum et al. 2012;Lazartigues et al. 2014;Zhang et al. 2014), and reliable methodology Lazartigues et al. 2014). An Agilent 7500ce ICP-MS equipped with a COMPEX Pro 102 laser ablation system with a 193 nm ArF-excimer laser (Li et al. 2013) using helium as the carrier gas to increase ablated sample transport efficiency was used. The size of the laser-circular spot remained constant at 60 μm in diameter. Carbonate reference materials NIST 610 and 612 were tested as external standards to monitor analytical precision and accuracy. The accuracy was estimated to be < 0.6% for all trace and rareearth elements in NIST 610 and < 5% in NIST 612 (Gao et al. 2002). The detection limits (i.e., background level) varied from 0.003 to 0.30 ppm for most elements except Mn (1.00 ppm), Fe (6.10 ppm), Zn (1.50 ppn), and Ba (0.40 ppm) in this study due to the influence of equipment conditions (Lazartigues et al. 2014).

Fine-to-medium crystalline, planar-s dolomites (Rd3)
Rd3 generally display fine-to-medium crystals ranging from 100 to 300 μm with planar subhedral textures. These dolomites usually appear in the zones that have undergone much stronger recrystallization (Fig. 6h). Furthermore, the sedimentary fabrics can be completely replaced by Rd3 through strong recrystallization (Fig. 6i).

Fine-to-medium crystalline, planar-e dolomites (Fd1)
Fd1 are generally planar euhedral crystals ranging from 100 to 500 μm with diamond textures . These dolomites show either clear and homogeneous or with cloudy cores and clear rims , with all dolomites displaying sharp to slightly sweeping extinctions. Fd1 occur as the first generation of filling minerals lining the smooth walls of fractures and dissolution pores, and they generally hackly arrange along the walls and only account for small proportions of the spaces . These dolomites can be followed by Fd2 abruptly, resulting in complete occlusion of pores .

Coarse crystalline, non-planar-a saddle dolomites (Fd2)
Fd2 usually show non-planar anhedral crystals ranging from 0.5 to 3 mm with saddle crystal textures . These dolomites display either cloudy or with thick cloudy cores and fairly thin clear rims, with all dolomites showing curved crystals with cambered cleavages (Fig. 7g-i) and undulating extinctions. Fd2 occur as the second generation of filling minerals growing over Fd1  and are the innermost minerals occluding the remaining pore spaces that have been partially or not occluded by Fd1.

Late-stage filling calcites (Fc)
Filling calcites (Fc) are stained to jacinth by alizarin red-S (Fig. 7c) and are characterized by planar euhedral crystal textures with planar cleavages. These calcites are the latest fracture infills postdating all dolomites, which may have successively precipitated after the Fds; the transition between Fds and Fc is generally abrupt (Fig. 7c). However, Fc are only found in the vugs and fractures of dolostones subjected to weathering and leaching.

Types and distributions
Reservoir spaces are mainly secondary pores due to dolomitization, recrystallization, and dissolution, which have altered the primary pores and generated new dissolution pores (Fig. 8). The reservoir spaces can be divided into 3 types according to genesis: (1) fabric selective dissolution pores, which are generated by altering a distinct sedimentary fabric, and the boundaries of these reservoir spaces do not cut through another fabric, including intraparticle dissolution pores (Fig. 8a,b), interparticle dissolution pores (Fig. 8b,c), intercrystalline dissolution pores (Fig. 8c,d), and bed-parallel dissolution pores (Fig. 8g); (2) fabric nonselective dissolution pores, which are generated by dissolving diverse fabrics, and the boundaries of these reservoir spaces generally cut through at least two kinds of fabrics, mainly enhanced dissolution pores ; (3) fractures, which are generated by structural deformation (Fig. 8c).

Filling situation
Petrography observations indicate that no Fc developed in any reservoir spaces of fresh dolostones. Furthermore, fabric selective dissolution pores are seldom occupied with Fds ( Fig. 8a-g). In contrast, Fabric non-selective dissolution pores are frequently occupied by one or two types of Fds The Fd2 are only found in occlusive enhanced dissolution pores, showing a paragenetic relationship with Fd1 (Fig. 7h,i). Only Fd1 develop in the existing enhanced dissolution pores; they primarily grow around the inner walls and only take up small proportions of these reservoir spaces (Figs. 7d-f, 8f-i). Hence, large proportions of the reservoir spaces can be preserved if Fd2 do not develop. However, the appearance of Fd2 results in the loss of these reservoir spaces because they completely occlude the remaining reservoir spaces that have not been filled by Fd1 (Fig. 7g-i).

In situ testing results
Only a single spot was tested in each dolomite grain with homogeneous crystal features. In addition, cloudy cores and clear rims were separately tested for each Fd grain with heterogeneous crystal features. In total, 66 spots in 13 samples were tested, including 12 spots for Rd1 (marked in gray), 12 spots for Rd2 (marked in yellow), 8 spots for Rd3 (marked in orange, including Fd1 cores with geochemical propertied similar to that of Rd3), 25 spots for Fd1 (marked in green), and 9 spots for Fd2 (marked in blue) (Fig. 9). All test results are provided in Table 1. Note that the cloudy cores of heterogeneous Fd1 in SGT-10-6 and SGT-17-1, 4 are marked in orange (Fig. 9d, h) because they show geochemical characteristics similar to Rd3 rather than Fd1. However, the cloudy cores and clear rims of heterogeneous Fd2 in ) and SGT-08 (Fig. 9p) show the same geochemical characteristics to those of homogeneous Fd2. Detailed test results are described as follows in Table 1.

Diagenetic fluids of different dolomites
Based on petrography and geochemical properties of the samples, we interpret the types, nature, and origin of the fluids which formed different types of dolomites.

Origin of replacive dolomites (Rd)
All replacive dolomites may have similar or even the same origin. Rds are distributed in the sedimentary fabrics and recrystallized zones of the Xiaoerbulak dolostones, indicate that these dolomites are likely originated from precursor carbonates, and experienced dolomitization and recrystallization under the influence of dolomitizing fluids. All Rds show nearly the same REE distribution patterns (Fig. 12a-c), and only slight changes in trace elemental composition can be distinguished among them (Fig. 10a-c). Therefore, we propose that the Rd1, Rd2, and Rd3 have similar or even the same origin.
The dolomitizing fluids of Rds may be related to ancient seawater. Rd1, Rd2, and Rd3 show slightly left-leaning REE patterns and slightly positive or no Ce anomalies (Fig. 12a-c) and show inheritance from seawater or precursor carbonates (Alibo and Nozaki 1999;Mclennan 1989;Zhang et al. 2008;Zhao and Jones 2013). Based on the lack of Ce anomalies (Table 1), low ΣREE (Fig. 11), moderate-to-high Fe, and low Mn (Fig. 10a), dolomitizing fluids should be weakly oxidizing (Morford and Emerson 1999) and have non-hydrothermal properties (Chen et al. 2009a;Middleton et al. 1993). Low Mn abundances also imply that meteoric waters and deep fluids did not  influence Rd, because they would have resulted in high-Mn dolomites (Jin et al. 2006). Furthermore, moderateto-high Sr and high Ba contents (Fig. 10b, c) indicate that Rds formed at an early diagenetic stage and only experienced weak recrystallization (Derry 2010;Hecht et al. 1999;Jacobsen and Kaufman 1999;Qing 1998), which is supported by their non-planar-a or planar-s crystal textures and small crystal sizes. Moreover, high Ba contents (Fig. 10c) also imply that these dolomites may have been produced from relatively high-salinity dolomitizing fluids.
From the above, we suggest that the dolomitizing fluids originated from concentrated ancient seawater in the pores under weak oxidizing conditions. Considering the intensely evaporitic environment of the Tarim Basin in the Middle Cambrian, we advocate that formation of the Rds occurred at this time. All values are in ppm. Blank indicates below detection limits. Post-Archean Average Shale (PAAS) compositions used for normalizing calculations are from Mclennan (1989). Ce and Eu anomaly values were calculated using Ce/Ce * = Ce SN /(0.5La SN + 0.5Pr SN ) and Eu/Eu * = Eu SN / (0.67Sm SN + 0.33Tb SN ) (Bau et al. 1996)

Origin of Fd1
Fd1 are all distributed in pores and fractures, indicating their direct precipitation from diagenetic fluids rather than replacement of precursor carbonates. Fd1 show planare diamond dolomite crystals, suggesting their slow precipitation and sufficient crystallization, which is supported by relatively high Mn, low Ba, and Sr contents (Fig. 10a-c). Moreover, Fd1 show relatively high Mn contents (Fig. 10a-c) and high ΣREE (Fig. 11), suggesting that the diagenetic fluids of Fd1 were hydrothermal fluids. Fd1 show no Eu anomalies (Table 1); hence, these dolomites probably originated from non-magmatic (crustal) hydrothermal fluids. Note that the Fd1 display roofshaped REE patterns with significant MREE enrichment and depletion in LREE and HREE (Fig. 12d-f), which are common for hydrothermal carbonates generated by low-pH crustal fluids (Hecht et al. 1999). LREE depletion resulted from the hydrothermal recrystallization of dolomites (Kucera et al. 2009), which is supported by the planar-e diamond dolomite crystals and relatively low Sr contents (Fig. 10b). HREEs are bound to less soluble minerals, which implies HREE depletion in the diagenetic fluids (Bau and Moller 1992;Morgan and Wandless 1980). Additionally, Fd1 show relatively moderate-to-high Fe contents (Fig. 10a), indicating weak oxidation fluid conditions (Azomani et al. 2013), which supports the inference of crustal hydrothermal fluids.
We suggest that the diagenetic fluids of Fd1 were most likely deep-circulating crustal hydrothermal fluids that originated from meteoric or marine water carrying crustal features from detrital rocks. Because Fd1 formed prior to Fd2 in petrography, we infer that the deep-circulating crustal hydrothermal fluids should be earlier than Permian magmatic hydrothermal fluids, and they were controlled by regional tectonic and fault activity. Hence, deep-circulating crustal hydrothermal fluids and Fd1 most likely formed during the Late Caledonian-Early Hercynian, i.e., the Devonian, when  (Tang et al. 2012).

Origin of Fd2
Fd2 are all distributed in pores and fractures and show non-planar anhedral crystals with saddle crystal textures and large crystal sizes, suggesting rapid precipitation and insufficient crystallization from diagenetic fluids rather than recrystallization, which is also supported by relatively high Sr contents (Fig. 10b) because recrystallization would result in Sr depletion (Derry 2010;Hecht et al. 1999;Jacobsen and Kaufman 1999;Qing 1998). Moreover, all Fd2 show saddle crystal textures, relatively high Mn contents (Fig. 10a-c), and high ΣREE (Fig. 11), suggesting that the diagenetic fluids of Fd2 should be hydrothermal fluids (Chen et al. 2009a;Huang et al. 2014;Middleton et al. 1993;Sirat et al. 2016). Studies on this kind of pore-filling dolomites showed that the homogenization temperatures could reach 173°-200° (Zhao et al. 2012;Zhu et al. 2010).
Note that the Fd2 display slightly to significantly rightleaning REE patterns with obvious positive Eu anomalies (Fig. 12g,h, Table 1). Positive Eu anomalies may indicate that the hydrothermal fluids should be enriched in Eu 2+ , indicating an acidic and reducing condition (Frimmel 2009), because the hydrothermal fluids would preferentially provide Eu as Eu 2+ under acidic conditions (Bau 1991;Kucera et al. 2009;Morgan et al. 2013), and the Eu 3+ in hydrothermal fluids would be reduced to Eu 2+ under reducing conditions (Bau et al. 1996;Bau and Moller 1992;Hecht et al. 1999). Reducing conditions are also indicated by relatively low Fe contents (Morford and Emerson 1999) of Fd2 (Fig. 10a).
Based on the acidic, reducing, and extremely high-temperature (~ 200 °C) properties, we infer that hydrothermal fluids were most likely magmatic hydrothermal fluids. The abundant Permian acidic magmatic eruptions and intrusions found in the Tarim Basin (Tian et al. 2010;Yang et al. 2007;Yu et al. 2011;Zhang et al. 2008Zhang et al. , 2010Zhang et al. , 2014Zhou et al. 2009) suggest that highly active Permian volcanic-magmatic activities could provide plenty of acidic, reducing, high-temperature, and upward-migrating magmatic hydrothermal fluids enriched in Eu 2+ . This can also be supported by outcrop evidences (Zhang et al. 2014), integrated isotopic geochemistry (C, O, and Sr), and fluid inclusion microthermometry (Dong et al. 2013). Moreover, plentiful magmatic and siliciclastic rocks (Li et al. 2011a, b;Wang et al. 2010;Zhai 2013) rich in Eu 2+ -bearing plagioclase developed in the Precambrian basement of the Tarim Basin; therefore, hightemperature hydrothermal fluids migrating upward through the basement can be easily enriched in Eu 2+ by interacting with these plagioclase-rich rocks and precipitate Fd2 with positive Eu anomalies (Mclennan 1989;Kucera et al. 2009). Therefore, the diagenetic fluids of Fd2 were likely Permian magmatic hydrothermal fluids and the precipitation of Fd2 occurred most probably in the Permian.

Influence of diagenetic fluids on the formation of diverse reservoir spaces
Three distinct types of diagenetic fluids are interpreted from 5 types of dolomites as described previously; these diagenetic fluids have played an important role in the formation of diverse reservoir spaces. Here, we interpret the influence of different diagenetic fluids on the formation of differing reservoir spaces based on petrography observation.

Influence of dolomitizing fluids related to ancient seawater
Fabric selective dissolution pores were all accompanied by Rds, and no Fds were found in these reservoir spaces . This observation suggests that deep-circulating crustal hydrothermal fluids and magmatic hydrothermal fluids had no clear influence on the formation of fabric selective dissolution pores. Therefore, the fabric selective dissolution pores may have been altered from primary pores rather than newly generated under the influence of dolomitizing fluids related to ancient seawater (Fig. 6e, f). Most intraparticle dissolution pores show wave-like irregular boundaries. Furthermore, Rds surrounding the edges of these pores usually display outward-convex crystal morphology (Fig. 8a, b), indicating that no carbonate minerals dissolved during dolomitization. The dolograinstone cements surround the grains in a teeth-like shape (Figs. 7e,8c,d), suggesting that the interparticle dissolution pores were inherited and altered from primary interparticle pores during dolomitization. Moreover, the bed-parallel dissolution pores are accompanied by enhanced dissolution pores and Fd1 (Fig. 8g), suggesting the occurrence of precursor primary bed-parallel pores because they are expected to provide the necessary channels for diagenetic fluids to generate these pores and precipitate Fd1. Moreover, the intercrystalline dissolution pores are only found in the strongly recrystallized zones (Figs. 7d, 8d, e).
Therefore, we propose that these fabric selective dissolution pores were inherited and altered from precursor primary pores through dolomitization and recrystallization, and the dolomitizing fluids related to ancient seawater rarely generated new reservoir spaces.

Influence of deep-circulating crustal hydrothermal fluids
Fd1 are frequently found growing around the inner walls and take small proportions of fabric non-selective dissolution pores (enhanced dissolution pores) h,i,. The boundaries of fabric non-selective dissolution pores show outward-convex morphology with regular and smooth profiles, regardless of Fd1 growth around the inner walls h,i,8h,i). Notably, the edges of Rd3 were dissolved as denoted by embayed shapes (Fig. 8e). Therefore, we propose that these fabric non-selective dissolution pores were generated by the dissolution of Rds. The cloudy cores of Fd1 show similar geochemical properties with Rd3 (Fig. 9d, h, Table 1), which indicates that the deep-circulating crustal hydrothermal fluids would have interacted with the Rds. Based on these arguments, we conclude that the deepcirculating crustal hydrothermal fluids generated the fabric non-selective dissolution pores. These hydrothermal fluids dissolved Rds in the early-middle stages and precipitated Fd1 at a late stage by using Rds as crystallization centers. However, further study is required to interpret the mechanism by which the deep-circulating hydrothermal fluids dissolved Rds.

Influence of magmatic hydrothermal fluids
No Fd2 individually filled the fabric non-selective porosity although they did fill fractures alone, as described before, Fd1 occurred as the first generation of infills of the fabric non-selective dissolution pores, while Fd2 are second-generation infills growing over Fd1. However, Fd1 show straight and smooth crystal edges without any dissolution trails (Fig. 7h, i). Therefore, we infer that magmatic hydrothermal fluids have no dissolution properties and only precipitated Fd2 as the innermost infills, occluding the remaining reservoir spaces that were not occluded by Fd1.

Porosity evolution controlled by multiple-stage diagenetic fluids
Based on our interpretation of the diagenetic fluids and their influence on the formation of reservoir spaces, this potential sequence of diagenetic fluids can be summarized: (1) dolomitizing fluids related to concentrated ancient seawater in shallow burial environment during the Middle Cambrian; (2) deep-circulating crustal hydrothermal fluids during the Devonian; and (3) magmatic hydrothermal fluids during the Permian.
The average porosity of the samples is 7%-8% (reported by our research before (Li et al. 2015)). More than 260 viewsheds under microscope (> 20 for each sample) were calculated and the volume percent of the Fd1 (2%), Fd2 (1%), fabric selective pores (4%-5%), and fabric non-selective pores (6%) can be measured, so we can propose the porosity evolution of high-quality dolostone reservoir (i.e., dolograinstone reservoir) in the Xiaoerbulak Formation as  Fig. 13 The proposed porosity evolution of the high-quality dolostone reservoir controlled by multiple-stage diagenetic fluids in the Lower Cambrian Xiaoerbulak Formation. (The initial porosity is estimated with the dolograinstone regardless of matrix and cements) follows ( Fig. 13): (1) Initial carbonate grains continually precipitated from Early Cambrian seawater and accumulated overlying the seafloor; sediments at this stage were loose and had very high primary porosity (> 20%).
(2) Primary sediments were gradually buried, compaction, and cementation strengthened with the increase in burial depth, and the average porosity of newly formed grainstones declined. In the Middle Cambrian, dolomitizing fluids participated in the diagenetic process of grainstones, transformed grainstones into dolograinstone, and altered the interparticle pores to interparticle and intercrystalline dissolution pores, and the average porosity turned to 4%-5% (estimated average porosity from dolograinstone with only fabric selective pores).
(3) Deep-circulating crustal hydrothermal fluids participated in the diagenetic process during the Devonian; the fluids dissolved considerable amounts of Rds in the early-middle stages and then precipitated Fd1 in the late stage; the average porosity initially increased to 10%-11% (the porosity of all fabric selective and non-selective pores, regardless of filling minerals in them) and then decreased to 8%-9% (due to filling by Fd1). (4) Magmatic hydrothermal fluids joined in the diagenetic process during the Permian and precipitated Fd2, and the average porosity decreased to 7%-8% (the average porosity of dolograinstone reported by our research before (Li et al. 2015)).

Conclusions
Based on the systematically petrographic and in situ geochemical study on different types of dolomites and reservoir spaces in the dolostones of the Lower Cambrian Xiaoerbulak Formation, this research analyzed the types, nature, and origin of the diagenetic fluids of dolomites and their influence on the porosity evolution of this high-quality dolostone reservoir. Three types of replacive dolomites (Rd) and 2 types of pore-filling dolomites (Fd) are classified based on petrography. And the diagenetic fluids can be divided into three types according to in situ geochemical properties: (1) Rds with slightly left-leaning REE patterns, low ∑REE, low Mn, moderate-to-high Fe and Sr, and high Ba contents were deviated from shallow burial dolomitizing fluids, in an early diagenetic stage, which might probably be related to concentrated ancient seawater in pores; (2) Fd1 with roof-shaped REE patterns, high ∑REE, low Ba and Sr, moderate-to-high Fe, and high Mn contents were likely slowly and sufficiently crystallized from deep-circulating crustal hydrothermal fluids during Devonian; and (3) Fd2 with slightly to significantly right-leaning REE patterns, obvious positive Eu anomaly, high ∑REE, low Fe and Ba, and high Mn and Sr contents were might rapidly and insufficiently precipitated from magmatic hydrothermal fluids during the Permian.
The porosity evolution under the control of multiple-stage diagenetic fluids can be interpreted: (1) Early dolomitizing fluids (concentrated ancient seawater) altered primary pores to fabric selective dissolution pores through dolomitization during the Middle Cambrian; the average porosity was about 4%-5% after compaction, cementation, and dolomitization.
(2) During the Devonian, deep-circulating crustal hydrothermal fluids significantly increased porosity in the early stages by dissolving and then slightly decreased the porosity in late stage due to Fd1 precipitation. The average porosity increased to 10%-11% after sufficient dissolution and then slightly decreased to 8%-9% after the Fd1 precipitation. (3) Magmatic hydrothermal fluids only precipitated the Fd2 and slightly decreased the porosity during the Permian. The average porosity slightly decreased to 7%-8%.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.