GNGTS 2017 - 36° Convegno Nazionale

GNGTS 2017 S essione 2.1 239 Data sources. We first considered the following, public items available at a continental scale: • ������� ������ ���� �� ��� ��������� �� ����� ������� �������� ��� ������ CERESIS (1995) that is the catalogue of South America compiled for GSHAP; • ������� ��� ���������� ������� ���� �� ��� ����������� ���������� ��������� Engdahl and Villaseñor (2002), that is the “Centennial Earthquake Catalog”; • Storchak et al . (2013), that is, the ISC-GEM instrumental earthquake catalogue. CERESIS (1985) has been also considered, to get information on the available macroseismic data points (MDPs). Next, national current catalogues made available by the partners of this project have been considered (Berrocal et al ., 1984; Beauval et al. , 2013; Funvisis, 2014; INPRES, 2015; Leyton et al ., 2009; Observatorio de San Calixto, 2013; Servicio Geologico Colombiano, 2014; Tavera ed., 2001) �� ��������� �� ���� ��������� �� ���������� �������� ���� ���� �� ���� ���� In addition, we have consulted 34 earthquake studies; only part of them were considered by the compilers of the national catalogues. ��� ��� ����� ����������� �� � ����� For the large earthquakes (M ≥ 7.0), the inventory compiled in the frame of the GEM-GEH project (GEH, 2013; Albini et al ., 2014) has been an important source. The entries related to the same event have then been clustered. This operation has been performed in two steps: first, automatically, then manually. The last one has allowed to detect and eliminate several duplications, mainly inside CERESIS (1995), with respect to border earthquakes. Time, location and depth. One entry for each earthquake has been selected as “preferred” with reference to time, location and depth. Priority was given according to the following order: 1) Storchak et al. (2013); 2) Engdahl and Villaseñor (2002); 3) recent earthquake studies; 4) national catalogues; 5) CERESIS catalogue (1995); 6) CERESIS (1985). However, when entries from national catalogues clearly coincided with the one from CERESIS (1985), the last one was selected, as it was the root of them and it gives references. After compiling this material it is possible to say that, in the time-window before 1964, the CERESIS (1995) catalogue contains more entries than the national catalogues which have been submitted to the SARA project (Venezuela, Ecuador, Brazil, Bolivia, Colombia, Chile), or found in websites (Argentina, Peru). It appears that, in many countries, some entries from CERESIS (1995) were not included in the national catalogue because of size threshold. On the other hand, some entries unknown to CERESIS (1995) have been found in some national catalogues. Finally, for each earthquake we compiled a “Task 4” entry adopting time, location and depth from the preferred one. Earthquake size. The data sources considered provide varied type of magnitude (M). CERESIS (1995) entries come with a variety of magnitude types and values; for some entries several M values of varied type are given. We decided to adopt one magnitude value according to the following priority scheme: Mw, Ms, mb, ML, other M. In addition, CERESIS (1995) provides intensity values. Engdahl and Villaseñor (2002), too, provide varied types and values of magnitude, including some of unknown type (UK). Storchak et al . provide Mw. The most updated national catalogues provide varied types of magnitude. Bolivia gives Ms and mb; Brazil gives mainly mb; Colombia and Ecuador gives M of varied types; Chile gives Ms; Peru mostly Ms and some Mw; Venezuela gives Ms. As for the earthquake studies, the modern ones gives Mw of macroseismic origin, mainly assessed with the Bakun and Wentworth method. Magnitudes not assessed in terms of Mw, Ms or mb have been converted to Ms or mb according to data sources. The magnitude of the “Task 4” entries have been compiled selecting the most reliable value available, according to the above mentioned priority scheme: Mw, Ms, mb, (ML), other M. In addition, if we have two or more Mw values, or two or more Ms, from two entries referred to the same earthquake, we selected one of them according to expert judgement. As a general rule we prefer M values the origin of which is known. At this stage, we had Mw values available for 34% of the entries: we had therefore the task of determining Mw for about 66% entries. We considered a number of conversion relationships in literature. We preferred Scordilis relations (2006), which gives values similar to the ones proposed by ISC-GEM, in addition,