A reservoir induced earthquake swarm in the Central Highlands of Sri Lanka

Abstract

An anomalous seismic sequence of five small (MW < 3) felt earthquakes occurred between 29 August 2020 and 05 December 2020 around the Victoria Reservoir in the central highlands of Sri Lanka that clearly exceeded the established national background seismic rate. Using seismic waveform template-matching and a newly developed single-station earthquake location method based on travel-time back-projection, we detected an additional co-located 23 microseismic events, of which 18 occurred within the same period as the felt events. This hitherto undetected seismic swarm defines a seismogenic zone beneath the western flank of the reservoir between 1.5 and 3 km depths. The reservoir-induced peak stresses, resolved on E-W striking faults, predicted from the poroelastic theory that include both drained and undrained crustal responses are ~ 15 kPa in an area overlapping the seismogenic zone, which, together with the physical and spatio-temporal characteristics of the seismic swarm, establish a causal link between reservoir-induced stresses and the earthquake swarm with implications to seismic hazard. This is the first record of induced seismicity in Sri Lanka. The newly developed efficient computational workflows with minimal operational costs described in our study provide a blueprint for monitoring reservoir-induced seismicity in developing countries with severe resource limitations.

Efficient cosine similarity-based image correlation algorithm for object detection and localization

“…So far, the cross correlation has been widely used in object detection. Most recently, P. Herath etc., use template-matching to detect microseismic events [4]. M. kaur etc., use template-based matching algorithm to track target motion on single energy images [5].…”

Cluster Analysis of Seismicity in the Eastern Gulf of Corinth Based on a Waveform Template Matching Catalog

“…This technique has found many applications in case studies in the recent literature as it provides an automated way to drastically increase the volume of a dataset by detecting a large number of smaller events, previously unreported by routine analysis. Herath et al [ 36 ] applied template matching, combined with a single-station travel time back-projection method, to detect an induced seismic swarm at the Victoria Reservoir in Sri Lanka, with events of magnitudes down to M w = −1.3. In another study, Chmiel et al [ 37 ] combined the classic STA/LTA method [ 38 ] with template matching to detect 91 events using 23 templates.…”

A Comprehensive Numerical Model for Reservoir-Induced Earthquake Risk Assessment

“…The impoundment of artificial reservoirs can trigger earthquakes, leading to loss of life and property, and giving rise to safety issues such as chemical spills [1][2][3][4][5]. This issue has garnered significant attention, particularly with the escalating development of water conservation and hydropower projects.…”