GNGTS 2017 - 36° Convegno Nazionale

240 GNGTS 2017 S essione 2.1 it provides uncertainty. We also considered Contreras Luarte (2009) for Chile, but its range of definition is very limited, and Assumpção et al . (2014) ��� ������� ���� ���� ��� ����� � for Brazil. Only this one shows a different behaviour; therefore we adopted it for Brazil, only. Determining Mw from macroseismic data�. At this stage we still had hundreds of entries without Mw, half of them from Peru. For all the entries, we had Io given by CERESIS (1995) catalogue. The best would be to determine Mw from the macroseismic data points (MDPs), making use of repeatable procedures such as the models proposed Bakun andWentworth (1997) or Gasperini et al . (2010), as it has already been done for some earthquakes in Venezuela, Colombia and Ecuador. However, this requires MDPs, which are not always available, and the determination of the regional coefficients of the models: • to Colombia through a calibration process; this process was developed by the present study using the method of Bakun and Wentworth calibrated regionally for some events of 20-21 st century that have a sufficient number MDPs from Colombian Geological Survey (2013) and it was applied to 29 historical earthquakes of Colombian territory; • to Venezuela and Ecuador, the strategy adopted was to use Mw/I empirical relationships available from literature (Beauval et al. , 2010; Palme, 2005); • to Peru-Chile, Colombia, Bolivia and Argentina from Mw/I linear empirical relationships determined in the present study. For sake of homogeneity we have assessed the final Mw uncertainty as equal to 0.60 unit, that correspond to the mean of 95% confidence level to one intensity data point following the Bakun and Wentworth (1997) method. Results and conclusion. We have established the lower Mw threshold at 5 (-0.2) for the areas of the Andean region; for Brazil, no lower threshold has been established. In such a way we got 2556 events, the distribution of which by data sources is presented in Fig. 1. Having now determined Mw for all these earthquakes, for the first time we can see the seismicity plotted in terms of Mw (Fig. 2). We can also show the earthquake history before 1964 (Fig. 3). We have still 1766 events with size below the adopted threshold, for most of which the Mw to other parameters regressions cannot be applied, because they are out of reliability range. In addition, we have 227 without any size. Fig. 3 - Earthquake history before 1964 (Mw ≥ 5.0) of South America.