Inferring fluid volume during earthquake swarms using seismic catalogues

Abstract

Summary Many studies have pointed out a correlation between either the cumulative or the maximum seismic moment and the injected fluid volume when analyzing global datasets of fluid injection-induced earthquake sequences. However, those correlations become quite uncertain when looking at individual episodes, mainly because of the large aseismic component of the induced deformation. If natural swarms are thought to result from the same physical processes as sequences from anthropogenic origin, little is still known about them as observations are limited by the depth of the active zone and the moderate deformations. In this work, we make profit of the similarity between both natural and injection-induced swarms. To this aim, we develop new relations between seismic observables and hydraulic attributes by using a global compilation of injection-induced earthquake catalogs, leading to two methods to estimate the injected fluid volume based solely on earthquake catalogs. Once the precision of our approaches is validated, we estimate the volume and flowrate of fluids circulating in diverse natural swarms, shedding a new light on the fluid dynamics that trigger them.

Prevalence of Aseismic Slip Linking Fluid Injection to Natural and Anthropogenic Seismic Swarms

“…Our approach with D max might therefore be challenged for sequences with a high b ‐value, for which average slip might be better estimated by taking into account the cumulative seismic slip, computed on a spatial grid for instance, rather than by only considering the biggest event. However, as a first approximation, the value obtained with D max seems to be a reasonable estimate of total moment released, as validated in another work (Danré et al., 2022).…”

“…For instance, $${\Delta}{\sigma }_{\mathrm{e}}$$ ∼ 1 kPa for the Soultz‐sous‐Forêt stimulations (1993, 1995, and 1996) suggests an important aseismic moment release, while $${\Delta}{\sigma }_{\mathrm{e}}$$ = 1 MPa for the Basel injection might mean that the contribution of aseismic slip is relatively less important in this case. Therefore, variability in effective stress drop values within the studied sequences might indicate variability in the importance of aseismic slip release among swarms, and $${\Delta}{\sigma }_{\mathrm{e}}$$ could be considered as a valuable proxy to the seismic‐to‐total moment ratio (Danré et al., 2022).…”

Control of Seismicity Migration in Earthquake Swarms by Injected Fluid Volume and Aseismic Crack Propagation

“…We propose here a way to combine this approach with the observation made of V inj ∝ R 2 in order to fully constrain the injected fluid volume. The squared seismicity front distance evolution gives the injected volume V inj (t) shape, while the final value of injected volume constrained through the statistical analysis of Danré et al (2023) gives its amplitude.…”

“…respectively. Despite the inherent uncertainties in the methods employed byDanré et al (2023), which may lead to deviations of up to a factor of four in the final volume estimates, our primary focus remains the examination of temporal variations in volume. For Corinth 2015, the flow rate seems quite constant as volume increases linearly, while for the Cahuilla swarm the injection seems to accelerate with time.…”

“…For Corinth 2015, the flow rate seems quite constant as volume increases linearly, while for the Cahuilla swarm the injection seems to accelerate with time. Taking distances of R = 600 m and R = 3,500 m for Corinth 2015 and Cahuilla, respectively(Danré et al, 2022), and estimated volume values of 5,100 m 3 and 13.10 6 m 3 as obtained fromDanré et al (2023) leads to s d estimates of s d = 70 m −1 and s d = 9 m −1 for the two sequences, respectively. Therefore, it seems possible to combine our observations of V inj ∝ R 2 and already existing methods of final volume estimation to better constrain the temporal injection dynamics occurring during fluid-induced natural earthquake swarms.…”