The Xiaoerbulake Formation in the Tarim Basin is considered one of the most important deep to ultradeep hydrocarbon reservoirs in the world. The objective of the present study is to analyze the facies, depositional environment and reservoir quality of the Early Cambrian Xiaoerbulake Formation in the Tarim Basin through integrated analysis of 120 m of cores, 3240 m of well cuttings, wireline logs and over 1100 thin sections from 17 exploration wells. Early Cambrian Xiaoerbulake Formation deposits in the Tarim Basin were deposited on a ramp setting. The ramp was occupied by seven facies associations and ten facies types ranging from the inner ramp to the outer ramp. These facies associations include tidal flat, lagoon, high-energy shoal, sabkha, inter-shoal, low-energy shoal and open shelf. Their distribution was controlled by paleogeographic patterns, sea level changes and the paleoclimate. Reservoir quality is considerably affected by facies together with diagenetic processes, including dolomitization and anhydrite cementation. High-quality reservoirs are found in the dolograinstone of high-energy shoal environments, which are favorable exploration facies in the carbonate ramp. Their pore space types consist of interparticle pores, intraparticle pores, intercrystalline pores, vuggy pores and moldic pores, with a porosity of 1.4%–7.5%. This study will help with our understanding of the stratigraphic framework, sedimentary-facies evolution and high-quality reservoir distribution of the Early Cambrian carbonate ramp in the Tarim Basin, facilitating exploration and the production of hydrocarbons from the Xiaoerbulake Formation.

Accurate prediction of air pollutant concentrations, particularly fine particulate matter (PM_2.5), is essential for controlling and preventing heavy pollution incidents by providing early warnings of harmful substances in the atmosphere. This study proposes a novel spatiotemporal model for PM_2.5 concentration prediction based on a Conditional Wasserstein Generative Adversarial Network with Gradient Penalty (CWGAN-GP). The framework incorporates three key design components: First, the generator employs an Inception-style Convolutional Long Short-Term Memory (ConvLSTM) network, integrating parallel multi-scale convolutions and hierarchical normalization. This design enhances multi-scale spatiotemporal feature extraction while effectively suppressing boundary artifacts via a map-masking layer. Second, the discriminator adopts an architecturally enhanced U-Net, incorporating spectral normalization and shallow instance normalization. Feature-guided masked skip connections are introduced, and the output is designed as a raw score map to mitigate premature saturation during training. Third, a composite loss function is utilized, combining adversarial loss, feature-matching loss, and inter-frame spatiotemporal smoothness. A sliding-window conditioning mechanism is also implemented, leveraging multi-level features from the discriminator for joint spatiotemporal optimization. Experiments conducted on multi-source gridded data from Dongguan demonstrate that the model achieves a 12 h prediction performance with a Root Mean Square Error (RMSE) of 4.61 μg/m³, a Mean Absolute Error (MAE) of 6.42 μg/m³, and a Coefficient of Determination (R²) of 0.80. The model significantly alleviates performance degradation in long-term predictions when the forecast horizon is extended from 3 to 12 h, the RMSE increases by only 1.84 μg/m³, and regional deviations remain within ±3 μg/m³. These results indicate strong capabilities in spatial topology reconstruction and robustness against concentration anomalies, highlighting the model’s potential for hyperlocal air quality early warning. It should be noted that the empirical validation is limited to the specific environmental conditions of Dongguan, and the model’s generalizability to other geographical and climatic settings requires further investigation.