Absolute and relative locations of similar events with application to microearthquakes in southern Iceland

Abstract

S U M M A R YIt is well known that similar earthquakes, i.e. earthquakes having almost identical waveforms, allow extremely accurate relative timing of the seismic arrivals. This has traditionally been used for achieving accurate relative locations of clusters of similar earthquakes. The arrival time differences between similar events depend not only on their relative location but also on the absolute location of the group. Moving a pair of events 200m while retaining their relative locations can cause a 1 ms change in the time difference between the first arrivals of the events at a station 6km distant. A change in time difference of l m s can easily be estimated by cross-correlating the waveforms of the two earthquakes. We use the accurate relative timings to improve absolute locations of groups of similar events, as well as to obtain extremely accurate relative locations. The absolute locations from relative timings are expected to have errors that are independent of the errors associated with locations based on absolute arrival time observations. We analyse data from five earthquake sequences, comprising a total of 96 earthquakes, recorded by a regional network in southern Iceland. One of the clusters is located within the on-land spreading ridge in south-western Iceland, and the other four are within the South Iceland seismic zone, a transform zone between overlapping branches of the spreading ridge. The events vary in magnitude between M , -0.3 and 2.8. After determining the absolute and relative locations of each swarm, we estimate the orientation of a best-fitting plane through the hypocenters. The mean distance of events from a best-fitting plane varies between 4 and 15m for the five swarms. This is comparable to the formal error estimates for the relative locations. Together with (nonunique) fault-plane solutions, the relative locations constrain the fault planes and the type of faulting. Faulting within the nascent transform zone in southern Iceland is predominantly strike slip on near-vertical N-S striking planes, in agreement with the orientation of mapped earthquake fractures in the area. The earthquakes within the spreading zone clearly define a fault plane striking parallel to the ridge and dipping 63". Each group of similar events probably represents repeated slip on the same fault.

Management of Post-mining Large-scale Ground Failures: Blast Swarms Field Experiment for Calibration of Permanent Microseismic Early-warning Systems

“…However, both the input data and the precision of the 3D localisation algorithm remains insufficient for the data recorded with multiplet-type configuration (blasts 9, 10, 11 and 12). It would certainly be useful to use relative localisation methods (Poupinet et al 1984;Slunga et al, 1995;Abdul-Wahed et al, 2006) to gain precision while re-localizing this kind of data. 195750 195560 195560 195504 195456 195449 195268 195129 195126 195122 195115 194914 195717 195526 195508 NL 195409 195409 195245 195022 195034 195022 195022 194799 195778 195526 195514 195491 195426 195426 195245 195092 195098 195092 195092 194864 …”

“…The "multiplet blast" configuration was performed to record "multiplets" a term which refers, in classic seismology, to a group of seismic events showing the following characteristics (Geller and Mueller, 1980;Poupinet et al 1984;Lees 1998;Slunga et al, 1995;Moriya et al 2006;Gibowicz, 2006): similar wave form; different time origin; considered as the result of the relaxation of constrain of a single fracture or fault, similar source mechanisms; close location; different magnitude. The recorded data might indeed be used to test the capabilities of "relative" location algorithms compared to "absolute" and classical seismic location algorithms.…”

Seismotectonic analysis of the Tjörnes Fracture Zone, an active transform fault in north Iceland

“…After locating all subgroups, a single hypocenter is determined for each earthquake by taking a weighted average of all locations for the event. The weights are based on the approximate error estimates produced by the group analysis [Slunga et al, 1995]. Inconsistent estimates are iteratively truncated until a consistent set of absolute locations for each event is achieved.…”

“…Following the method described by Slunga-et al [1995], correlated with waveforms for the 10 to 20 nearest neighbors, and the earthquake is then located together with the 6 to 12 best correlating events of the cluster. Every event will be included in several subgroups and located once for each of those.…”

Eos, Transactions, American Geophysical Union Volume 86, Number 26, 28 June 2005

“…Earthquake locations at Grímsvötn, improved by a double-difference location method [Slunga et al, 1995], and supported by P -wave particle-motion analysis at GRF, reveal a northerly trending event distribution, centered at the southern caldera rim, ~1 km east of the main eruption site ( Figure 1b). Focal depths mostly concentrate above 4 km and decrease by ~2 km near the eruption onset.…”