Assessing induced seismicity risk for the Lower Yarlung Tsangpo hydropower complex

A large hydropower complex is planned on the Lower Yarlung Tsangpo (YT) with an expected output roughly three times that of the Three Gorges Project. The planned hydropower complex lies in the eastern Himalayan syntaxis, which is characteristic of intricate fault systems, high tectonic strain rates, and strong topographic variations. Reservoir impoundment in such a geologic setting may lead to unintended consequences such as induced seismicity and landslides. A pre-impoundment risk assessment is imperative for the region and the project. With regional faults and stress information, we perform an analysis to identify the fault segments that may be affected by reservoir impoundment and lead to seismicity.
The existing observations from hydraulic fracturing tests indicate that the rotation of S_Hmax orientation shows similarities with the changes in the YT course. To obtain abundant and diverse stress information, we compiled 145 focal mechanisms for the study area covering the period of 2000 - 2023. Moment magnitudes concentrate around 1.5 - 4, and hypocenter depths are in the upper crust (≤ 15 km). Given the complexity of the fault system and the pronounced heterogeneity in the number and distribution of focal mechanisms, we partitioned the study area into four subregions and performed focal mechanism stress inversions separately for each subregion. The inversion results reveal a strike-slip regime in three subregions and a thrust faulting regime in one subregion. The stress ratios for all subregions lie in the range 0.6 - 0.8. The inverted S_Hmax orientations differ markedly between subregions, with a maximum discrepancy of ~58.5°.
To quantify fault destabilization risk, we employ a parameter termed ‘fault instability’ (FI). The FI range is from 0 to 1, ‘0’ for the most stable fault, while ‘1’ for the most unstable fault. It is quantified by fault frictional coefficient μ_f, fault strike and dip, stress field, and pore pressure. To consider the uncertainty in these input parameters, the Monte Carlo sampling is used to constrain the FI. Different fault segments exhibit markedly different FI values. Seismicity over the 23-year period predominantly occur on faults with high FI values, corroborating the qualification of the FI. FI distribution can inform dam siting and tunnel routing. We plan to build a 3D hydro-mechanical model that couples observed and inverted geological data, simulate pore pressure diffusion and water loading effects on Coulomb stress, and assess the resulting changes in FI and induced seismicity risk.