Scientific Reports (Dec 2023)

Reservoir impoundment-triggered seismicity in Brazil: the case of M4.0 Nova Ponte earthquake

  • Haris Raza,
  • Iman Rahimzadeh Kivi,
  • George Sand França,
  • Victor Vilarrasa

DOI
https://doi.org/10.1038/s41598-023-48924-6
Journal volume & issue
Vol. 13, no. 1
pp. 1 – 14

Abstract

Read online

Abstract Reservoir-triggered seismicity commonly occurs as a result of reservoir impoundment. In particular, the Nova Ponte reservoir triggered a series of earthquakes, including the 1998 M4.0 earthquake, which represents the second-largest earthquake triggered by reservoir impoundment in Brazil. The earthquake occurred after prolonged seismic activity following reservoir impoundment starting in 1993. After more than two decades, the mechanisms governing these earthquakes and their relation with the spatiotemporal evolution of the seismic events are still poorly understood. Here, we explain the causal mechanisms of the two largest earthquakes: an initial response M3.5 in 1995 and the delayed M4.0 in 1998. To this end, we numerically simulate the poromechanical subsurface response to reservoir impoundment using a 3D model that includes three geological layers down to 10 km depth. From the proposed potential nodal planes of the 1995 M3.5 earthquake, we show that the earthquake has most likely occurred on a vertical, E–W-oriented strike-slip fault with a reverse-displacement component. Deviatoric stresses generated by the water column loading on the surface, superimposed by undrained pore pressure enhancement in deep low-permeability layers can explain the fault reactivation. We find that for the 1998 M4.0 earthquake to occur, conductive flow pathways with permeability as high as 6.6·10−15 m2 should exist to transmit pore pressure to a deep critically oriented fault. Our analysis raises the importance of accounting for coupled poromechanical mechanisms controlling fault stability, hydromechanical properties of different rock layers, and realistic shape of the reservoir to accurately assess the potential for reservoir-triggered seismicity. We conclude that reliable forecasting models require accurate subsurface characterization before reservoir filling to enable managing the associated reservoir-triggered seismicity.