Seismotectonic Clustering and Seismic Gap Identification in the Indo-Eurasian-Burman Triple Junction Region

The convergence of the Indian, Eurasian, and Burma tectonic plates around the Bengal Basin is one of the most seismically complex regions on Earth. However, this zone is typically studied using conventional seismic zoning, which relies on subjective fault definitions and overlooks the structural heterogeneities in the sediment-filled basin. This study applies unsupervised machine learning, specifically Density-Based Spatial Clustering of Applications with Noise (DBSCAN), to analyse the United States Geological Survey (USGS) earthquake catalogue for this area between 1925 and 2025. By transforming the geographic coordinates into a Cartesian coordinate system and a rheologically weighted depth parameter, we identified more than 43 different seismic clusters. Structural orientation was determined using Principal Component Analysis (PCA), and stress regimes were determined using the Gutenberg-Richter b-value. The findings were categorized into high-stress zones (b < 0.8) and creeping segments (b > 1.0). From the event analogue of the recent 2025 Mw 7.7 rupture in the Sangaing fault, we found a similar prolonged seismic sleep near Sreemangal. The observed combination of prolonged quiescence and low b-values is consistent with a locked fault segment, as defined in classical seismic gap theory. The southern end of Cluster 6 on the northern Shillong Plateau is associated with the Oldham fault system, suggesting a potentially locked fault segment. The result stability was confirmed by statistical analysis, which provides new insights into the division of the megathrust and crustal faults, suggesting comparatively elevated seismic hazard potential that warrants careful monitoring.