Integrated 3D Petographic and Petrophysical Insights into the Permian Amb Formation, Pakistan.

This study utilized advanced methods to train well log data based on outcrop facies and then applied these models to other available well logs across the basin. The Amb Formation spans over 60,000 km² in the Trans- and CisIndus Basins, with its maximum thickness in the western Salt Range. Using machine learning and 3D modeling techniques, a clearer picture of the basin’s structure was created. The formation’s thickness increases in the western basin, with shallow depth in the southeast compared to the northwest. It mainly consists of quartz-rich siliciclastic or pure carbonate facies, with siliciclastic content increasing in the northwest. Reservoir properties show higher porosity and permeability in the northeast, while the southwest has lower values. The Potwar Basin has higher shale volume, while the Punjab plain has lower values. Macro porosity analysis indicates low porosity (<7%) except in certain facies like AM-3, which shows good porosity in the southeast. The Amb Formation’s facies and reservoir properties resemble those of the Permian succession in the Persian and Arabian platforms, but its lack of source rock may explain its non-productivity.

Three-dimensional modelling of stratigraphic facies and reservoir properties plays an important role in exploration and research. The Upper Permian Amb Formation of the Upper Indus Basin, Pakistan, was the subject of this experiment, which used multiple datasets from the field, the laboratory and machine learning techniques (ML) to model formation properties in three dimensions. The borehole data were trained with petrographic data to predict the facies and reservoir properties in the basin using an artificial neural network algorithm (ANN). Using field, borehole samples and borehole logs, four main microfacies were identified for the Amb Formation, comprising bioclastic packstone (AM-1), siliciclastic-rich bioclastic wackestone (AM-2), bioclastic wackestone (AM-3) and siliciclastic wackestone microfacies (AM-4). The intelligently predicted lithologies across the basin are grainstone, mudstone, sandstone, quartz wackestone, dolomite, sandy limestone, calcareous sandstone, bioclastic packstone, bioclastic wackestone and siliciclastic wackestone). The rock types predicted by machine learning match the petrographically determined rock types by 80%. The data generated by ANN are modelled three-dimensionally using Sequential Gaussian Simulation (SGSIM). The spatial distribution of facies shows that the bioclastic packstone facies occurs mainly in the southern and northeastern parts of the basin. In contrast, the siliciclastic-rich bioclastic wackestone facies is mainly developed in the northern parts of the basin. The bioclastic and siliciclastic wackestone facies are evenly distributed across the basin. In terms of porosity, the siliciclastic wackestone (AM-4) has a good porosity value (> 7%), while the other facies have a low porosity value. The petrophysical results show an increase in the average (ΦA) and effective porosity (ΦE) of the Amb Formation towards the northeast and an increase in permeability (k) towards the west, i.e., the Trans-Indus Basin (TIB), compared to the eastsoutheast. The volume of shale (Vsh) increased towards the Potwar Basin compared to the Punjab Plain. Regarding the deposition, the potential microfacies AM-1 are present throughout the Upper Indus Basin, specifically concentrated in the south.