ABSTRACT
This paper reviews the application of the MX algorithm to earthquakes in the New Zealand local catalogue, reported in Ma Li and Vere -Jones (1997), and compares it to earlier applications of M8 to the Californian local catalogue repotted in Keilis-Borok, Kossobokov and Rinehart (1986), Keilis-Borok, Knopoff, Kossobokov and Rotvain (1990) and Matthews and Switzer (1992)
The New Zealand local catalogue contains many smaller, especially intermediate-depth, events not included in the NEIC catalogue for which M8 was originally designed. Used with local magnitudes ML in place of the M» or m, given in the NEIC catalogue, and applied to the full catalogue of intermediate and deep as well as shallow events, M8 produces 4 TIPS for vents with target magnitude ML≥ 7, one uccessful, one failed, and wo current. Both regions covered by current TIPS have experienced large hallow events since the TIPS were declared, one just under and the other just at ML ≃ 7.
These results are very similar to those concerning the application of M8 to NEIC data for California reported by Keilis-Borok et al/.(1990), and confirmed subsequently in Matthews and Switzer (1992). They suggest that in both cases the algorithm leads to probability gains of about one order of magnitude. In both cases, however, the results appear to rest on some surprising details. It is pointed out in Matthews and Switzer (1992) that the results for California are critically dependent on the inclusion of a group of small events at the Geysers Geothermal area, which may have been partly induced by human activities. In New Zealand, the results are dependent on the inclusion of intermediate-depth earthquakes in the catalogue, and cannot be reproduced if these events are omitted. The possibility cannot be ruled out that the observed effects are due, in part at least, to improved network capability and reporting procedures.
