 |
| Titel |
Theoretical estimates of magnitudes of earthquakes induced by pore-pressure perturbations with large aspect ratios |
| VerfasserIn |
Martin Galis, Jean-Paul Ampuero, P. Martin Mai, Frédéric Cappa |
| Konferenz |
EGU General Assembly 2017
|
| Medientyp |
Artikel
|
| Sprache |
en
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 19 (2017) |
| Datensatznummer |
250142482
|
| Publikation (Nr.) |
 EGU/EGU2017-6111.pdf |
|
|
|
|
|
| Zusammenfassung |
| Being able to reliably and accurately estimate the possible maximum magnitude of
fluid-injection-induced earthquakes is of critical importance to quantify the associated
seismic hazard and to define operational constraints for geo-reservoirs. In previous studies,
we developed theoretical estimates of the magnitude of fluid-injection-induced earthquakes
based fracture mechanics, assuming circular pressure perturbations. However, natural
reservoirs are typically much wider than thicker. Therefore, here we discuss the application
of our model to horizontally elongated pressurized regions with realistic aspect
ratios.
Assuming circular pressure perturbations, we derived a physical model estimating how
large a rupture will grow on a given fault and for a given pore-pressure perturbation. We used
two approaches. The first, semi-analytical approach is based on pore pressure evolution
obtained by solving the diffusion equation for a cylindrical reservoir with no-flow boundaries.
The second approach is an approximation to the first one, based on a point-load
approximation of the pres-sure perturbation on the fault, allowing derivation of a complete
analytical formula relating the magnitude of the largest arrested rupture, Mmax−arr, to
injection and slip-weakening friction parameters. We found that the Mmax−arr scales with
cumulative injected fluid volume as a power law with exponent of 3/2. In contrast, the
Mmax relation by McGarr (2014) is a linear scaling (exponent of 1). While for the
dataset used by McGarr (2014) the difference between our and McGarr’s models is
relatively small, inclusion of datasets with broad range of injected fluid volumes
(from 10-10m3 to 1010m3) suggests better agreement with our model. However,
inclusion of extended pressure perturbations into our two models, while maintaining the
(semi-)analytical character, is not viable. Therefore, we perform numerical dynamic-rupture
simulations to investigate rupture nucleation and arrest for pressure perturbation with
large aspect ratios. We then interpret our findings and discuss implications for the
two previously derived (semi-)analytical models. In particular, we find that our
circular perturbation equations are also an adequate predictor of the final size under
elongated perturbations for a wide range of aspect ratios and background stress
values. |
|
|
|
|
|
|