The Himalaya-Tibet orogen, characterized by active deformation, significant strain accumulation, and dynamic geomorphic processes, presents an ideal environment to investigate low-frequency seismic events and their underlying geodynamic mechanisms. These events are crucial for revealing slow, long-duration processes often overlooked by conventional high-frequency analyses, yet they can be robustly detected using the semblance technique to measure waveform coherence across seismic arrays. In this study, we extend conventional centroid single force and moment tensor inversion techniques by incorporating a Bayesian bootstrap resampling framework to systematically quantify uncertainties in source parameters. This advanced approach enables the resolution of both shear and non-shear components, effectively discriminating between tectonic earthquakes and non-tectonic surface processes—such as landslides and glacier dynamics—prevalent across both the Himalayan front and the Tibetan Plateau. Consequently, our work contributes to the creation of a comprehensive low-frequency seismic catalog for the entire region while deepening our understanding of the diverse geodynamic processes governing this continental collision zone.