科研之友
微信
新浪微博
Facebook
分享链接
摘要
We document and quantify effects of two types of catalogue uncertainties-earthquake location errors and short-term incompleteness-on results of statistical cluster analyses of seismicity in southern California. In the main part of the study we analyse 117 076 events with m >= 2 in southern California during 1981-2013 from the waveform-relocated catalogue of Hauksson et al. We present statistical evidence for three artefacts caused by the absolute and relative location errors: (1) Increased distance between offspring and parents. (2) Underestimated clustering, quantified by the number of offspring per event, the total number of clustered events, and some other statistics. (3) Overestimated background rates. We also find that short-term incompleteness leads to (4) Apparent magnitude dependence and temporal fluctuations of b-values. The reported artefacts are robustly observed in three additional catalogues of southern California: the relocated catalogue of Richards-Dinger & Shearer during 1975-1998, and the two subcatalogues-1961-1981 and 1981-2013-of the Advances National Seismic System catalogue. This implies that the reported artefacts are not specific to a particular (re)location method. The comparative quality of the four examined catalogues is reflected in the magnitude of the artefacts. The location errors in the examined catalogues mostly affect events with m < 3.5, while for larger magnitudes the location error effects are negligible. This is explained by comparing the location error and rupture lengths of events and their parents. Finally, our analysis suggests that selected aggregated cluster statistics (e.g. proportion of singles) are less prone to location artefacts than individual statistics (e.g. the distance to parent or parent-offspring assignment). The results can inform a range of studies focused on small-magnitude seismicity patterns in the presence of catalogue uncertainties.
- 出版日期2015-9
