Tarbela reservoir, Pakistan: A region of compressional tectonics with reduced seismicity upon initial reservoir filling

Abstract

The Tarbela reservoir is located on the Indus River in the lesser Himalayas of northern Pakistan, a region of considerable natural seismicity. A microseismic survey conducted prior to impounding has indicated that crustal shortening under horizontal north-south compression dominates present-day tectonics, resulting in thrusting and strike-slip motion along numerous faults. The reservoir was first partially filled for a few weeks in 1974 and then completely filled in 1975. In both cases, the microseismic activity decreased slightly during the impounding period and recovered quickly upon commencement of drainage or establishment of a constant reservoir level. In tectonic environments where the maximum principal stress is horizontal and the tectonic system is compressive, a vertical surface load can move the crustal stresses away from a Navier-Coulomb failure criterion, and hence can temporarily decrease seismicity. The observed rapid recovery of seismicity may alternately be explained by: (1) high rates of tectonic strain accumulation which, in turn, increase horizontal stresses at crustal depths; they may quickly overcome the stabilizing effects of the small vertical stress changes and increased friction across faults from reservoir loading and, hence, may return the crustal stress system to one of failure; or (2) high diffusivity of the basement rock results in a pore pressue increase after a few weeks of raised reservoir head; the raised pore pressure equally reduces the effective horizontal and vertical stresses and brings the crustal system back to failure. Since direct monitoring of pore pressure at crustal depths is not available at Trabela, it is presently not possible to distinguish between the two alternatives. Indirect observations, such as long-term monitoring of space-time changes in seismicity, may aid in the resolution of this ambiguity. A few reservoirs, located in regions with preferentially extensional or pure strike-slip tectonics, have triggered or induced seismic activity, whereas Tarbela reservoir, in a region of convergent tectonics, has (to present) only slightly modified the natural seismicity. Here, in a region of active Himalayan tectonics, it is likely that any severe earthquake to occur in the near future would occur sooner or later, regardless of the presence of the reservoir. Within hundreds of years a major earthquake near Tarbela appears tectonically inevitable. Yet the presence of the reservoir may strongly affect, say, the time of occurrence, exact location, and the details of rupture of an earthquake. The time most likely for inducing a tectonic stress release (if any) would be during or shortly after a rapid draw-down of the reservoir, when the decrease in pore pressure from its high level at crustal depths lags behind the instantaneous stress decrease from reservoir unloading.