Bulletin of Earth Sciences of Thailand

Shale pore network and seepage simulation of Huai Hin Lat Formation

Soraya Suninbun — 2026-01-19

Gas shale is a type of petroleum rock associated with physical properties that simplify fluid migration pathways, necessary for the development of petroleum extraction and recovery in the region. Furthermore, it provides a carbon capture and storage (CCS) seal with low permeability and porosity. Shale porosity challenges the measurement of connected micropores. This study aims to investigate pore morphology, pore networks, and rock simulations of gas shale from Dat Fa sub-member of Huai Hin Lat Formation. The geochemical compositions were assessed using XRD and XRF techniques to identify the rock type. Scanning electron microscopy (SEM) analyzed the 2D pore morphology, distribution, and total porosity. X-ray tomographic microscopy generates pore network models that calculate total and effective porosity for seepage flow simulations, enhancing the analysis of gas transport mechanisms in rock formations. Geochemical analysis categorizes the rock types as calcareous shale and dolomitic shale. SEM images of calcareous shale typically reveal parallel flat pores and pinch-out along the laminations. Dolomitic shale exhibited the lamination of dolomite and calcite, with micropores surrounding mineral grains. The total porosity values for calcareous shale and dolomitic shale are 5.54% and 3.04%, respectively. Micro-CT image analysis revealed that the total porosity of calcareous shale ranged from 2.86% to 4.65%, while the effective porosity decreased down to 0.83% and 1.75%. The total porosity of dolomitic shale was estimated to range from 3.51% to 3.70%, with effective porosity ranging from 1.99% to 3.27%. Seepage simulations provide that calcareous shale has more diffusivity in parallel laminations, while dolomitic shale demonstrates greater diffusivity in perpendicular laminations.

Geochemical Characterization of a Fibrous Calcite Vein Using Micro X-ray Fluorescence Imaging: A Case Study

Tindikorn Kanta — 2026-01-19

Micro–X-ray fluorescence (µXRF) imaging is an effective non-destructive method extensively utilized in geosciences for the acquisition of high-resolution spatial elemental data. This study utilizes µXRF to examine a fibrous calcite vein (KK1) from the Permian Khao Khwang Formation in Thailand. The calcite vein, devoid of attached host rock, displays a cone-in-cone texture and a median suture rich in inclusions. µXRF elemental maps indicate a Ca-dominant composition, with median enrichments observed in Fe, Si, Al, K, and Mn, alongside later Sr-rich cross-cutting veinlets. This study illustrates the capability of µXRF to improve comprehension of fluid history, mineral development, and redox conditions within sedimentary systems, thereby assisting geologists in basin analysis and petroleum system assessment.

CropNet: Leveraging SegFormer for Efficient and Scalable Crop Mapping with Sentinel-2 Data

Sathirada Phahurat — 2026-01-19

This research investigates a deep learning-based methodology for crop classification by integrating Sentinel-2 satellite imagery with SegFormer, a state-of-the-art transformer-based semantic segmentation model. The study focuses on five dominant land cover types: rice fields, sugarcane, cassava, para rubber, and pond areas within a part of Khu Mueang District, Buriram Province, Thailand. The main objectives are to develop an efficient classification method using Sentinel-2 satellite data and to evaluate the predictive performance of SegFormer in the agricultural field. Satellite images were acquired via Google Earth Engine (GEE) during the harvest season (Nov 2023–Jan 2024), complemented by ground truth data collected from field surveys and high-resolution drone imagery. Preprocessing steps included cloud filtering, image normalization, and manual pixel-level labeling in QGIS software. The dataset was divided into 512×512 pixel patches, resulting in 780 image–mask pairs allocated for training (480), validation (120), and testing (180). The SegFormer model was trained using Optuna to find the best hyperparameter settings. The model achieved 0.967 pixel-wise accuracy with a validation loss of 0.075 (cross-entropy) on the training and validation datasets, demonstrating strong learning performance during model development. It showed strong classification performance for para rubber and sugarcane. However, it faced challenges in distinguishing cassava, ponds, and bare soil due to class imbalance and spectral similarity.