Machine learning algorithms are widely used to interpret well logging data. To enhance the algorithms’ robustness, shuffling the well logging data is an unavoidable feature engineering before training models. However, latent information stored between different well logging types and depth is destroyed during the shuffle. To investigate the influence of latent information, this study implements graph convolution networks (GCNs), long-short temporal memory models, recurrent neural networks, temporal convolution networks, and two artificial neural networks to predict the microbial lithology in the fourth member of the Dengying Formation, Moxi gas field, central Sichuan Basin. Results indicate that the GCN model outperforms other models. The accuracy, F1-score, and area under curve of the GCN model are 0.90, 0.90, and 0.95, respectively. Experimental results indicate that the time-series data facilitates lithology prediction and helps determine lithological fluctuations in the vertical direction. All types of logs from the spectral in the GCN model and also facilitates lithology identification. Only on condition combined with latent information, the GCN model reaches excellent microbialite classification resolution at the centimeter scale. Ultimately, the two actual cases show tricks for using GCN models to predict potential microbialite in other formations and areas, proving that the GCN model can be adopted in the industry.

Within carbonate strata across the globe, dolostone reservoirs hold paramount importance in the realm of petroleum development. The genesis of dolostone has long perplexed geologists; however, despite the proliferation of various dolomitization models, the origin of dolostone and its capacity to elucidate the extensive dolomitization processes of the Lower Paleozoic on a platform scale remain subjects to debate. The present study endeavors scrutinizing the dolomitization process of the Lower Cambrian Longwangmiao Formation, with a particular emphasis on the impact of paleogeomorphology on platform- or basin-scale dolomitization. Through meticulous field observations, petrological examinations and geochemical analyses, the correlation between the distribution of dolostones and paleogeomorphological features is elucidated, thereby establishing a dolomitization model that considers paleogeomorphological features. In contrast to prior investigations, this study discerns that the dolomitization process of the Longwangmiao Formation is not only governed by depositional facies environments and sea-level fluctuations but also markedly influenced by paleogeomorphological factors. Specifically, a distinct compositional and distributional gradient of dolomite is observed from the higher terrace to the lower terrace. These findings not only provide novel insights into the dolomitization process within the Sichuan Basin but also serve as a useful reference for analogous processes at the basin or craton scale in other sedimentary basins. The proposed three-terraced evaporated seepage reflux dolomitization model offers a fresh vantage point for describing the extensive dolomitization processes of the Lower Paleozoic on a global platform scale, and underscores the pivotal role of paleogeomorphology in the dolomitization process.

Paleofluid-rock interaction results in the modification of the ultra-deep reservoir quality, but the coupling effects of the paleofluid evolution on reservoir quality modification are still underestimated. Here, the multi-stage dolomite cements have been studied by means of core observations, thin section identification, fluid inclusions, and in situ elements and isotopic analyses in order to reaveal the paleofluid evolution processes and their effects on the reservoir quality. The fibrous dolomite cement (FDC) and bladed dolomite cement (BDC) both have similar geochemical properties and homogenization temperature (Th) to the matrix dolostone. Subsequent early silicification occurs (QZ1). However, silt-to fine-sized crystalline dolomite cement (CD2) has higher concentration of Fe and medium rare earth element, more negative δ¹⁸O than for FDC and BDC, and 103.4–150.2 °C in Th. The first petroleum charge episode solid bitumen 1 (SB1) followed the CD2. The medium to coarse sized crystalline dolomite cement (CD3) is characterized by the higher Mn and lower Sr concentration than CD2, positive δEu anomaly, negative δ¹³C and δ¹⁸O, high ⁸⁷Sr/⁸⁶Sr ratio and 138.9 °C–184.3 °C in Th. The solid bitumen 2 (SB2) occurs after CD3. Subsequently, the saddle dolomite (SD) has higher Fe and Mn concentration and lower Na and Sr than others cement, δEu positive anomaly, negative δ¹³C and δ¹⁸O, high ⁸⁷Sr/⁸⁶Sr ratio and 177.5 °C–243.6 °C in Th. Eventually massive silicification (QZ2) took place. The FDC, BDC, QZ1 and SB1 formed in the early diagenetic stage and have little negative effect on the reservoir quality. But the CD2 and CD3 dramatically decreased the reservoir quality related to a quantity of hydrothermal fluids entrance in the mesodiagenetic stage along the strike-slip faults and the unconformity surface at top of the Deng-2 Member during the late stage of the Caledonian orogeny. The reservoir spaces were retained and enlarged during the late diagenetic stage when the peak petroleum charge occurred, massive silicification and thermochemical sulfate reduction took place with SB2, SD and QZ2 being formed. The research outcome may update our understanding on the retention mechanism and the insights for further hydrocarbon exploration of Precambrian ultra-deep microbial reservoirs.

Recent advances in hydrocarbon exploration have been made in the Member Deng-2 marginal microbial mound-bank complex reservoirs of the Dengying Formation in the western Sichuan Basin, SW China, where the depositional process is regarded confusing. The microfacies, construction types, and depositional model of the Member Deng-2 marginal microbial mound-bank complex have been investigated using unmanned aerial vehicle photography, outcrop section investigation, thin section identification, and seismic reflections in the southwestern Sichuan Basin. The microbialite lithologic textures in this region include thrombolite, dendrolite, stromatolite, fenestral stromatolite, spongiostromata stone, oncolite, aggregated grainstone, and botryoidal grapestone. Based on the comprehensive analysis of “depositional fabrics–lithology–microfacies”, an association between a fore mound, mound framework, and back mound subfacies has been proposed based on water depth, current direction, energy level and lithologic assemblages. The microfacies of the mound base, mound core, mound flank, mound cap, and mound flat could be recognized among the mound framework subfacies. Two construction types of marginal microbial mound-bank complex have been determined based on deposition location, mound scale, migration direction, and sedimentary facies association. Type Jinkouhe microbial mound constructions (TJMMCs) develop along the windward margin owing to their proximity to the seaward subfacies fore mound, with a northeastwardly migrated microbial mound on top of the mud mound, exhibiting the characteristics of large-sized mounds and small-sized banks in the surrounding area. Type E’bian microbial mound constructions (TEMMCs) primarily occur on the leeward margin, resulting from the presence of onshore back mound subfacies, with the smaller southwestward migrated microbial mounds existing on a thicker microbial flat. The platform margin microbial mound depositional model can be correlated with certain lateral comparison profile and seismic reflection structures in the 2D seismic section, which can provide references for future worldwide exploration. Microbial mounds with larger buildups and thicker vertical reservoirs are typically targeted on the windward margin, while small-sized microbial mounds and flats with better lateral connections are typically focused on the leeward margin.