Intraplate seismicity in Brazil - Intraplate Earthquakes - page 53

Magnitudes of historical earthquakes were estimated from felt area, as shown in Figure
3.1b . Both total felt area (inside isoseismal II MM) and A IV (area inside isoseismal IV
MM) were used. The relations between felt area and magnitudes are similar to those of the
Central and Eastern United States as reported by Nuttli et al . ( 1979 ) , and shown in Figure
3.1b . Our data are also consistent with the global compilation of Johnston et al .( 1994 ) and
Johnston ( 1996 ) form b > 4.5. An event with magnitude 5 is usually felt up to 200-500 km
away. The new regression shown in Figure 3.1b is not significantly different from the one
used by Berrocal et al . ( 1984 ) . For historical events, m b magnitudes can be estimated with
a standard error of 0.35 units.
Completeness of the catalogue varies considerably according to the region. For the most
populated areas, such as southeastern Brazil, magnitudes above 4.5 (which should be felt
to 100 km distance or more; Figure 3.1 ) are believed to be complete since about 1850. For
uninhabited areas such as the Amazon region, the threshold of 4.5 was only attained in the
late 1960s with the installation of WWSSN stations in South America, and the array station
in Brasilia, central Brazil.

3.3 Seismicity map

Figure 3.2a shows all epicentres of the Brazilian catalogue with magnitudes above 3.0 (a
total of about 800 events) and the main geological provinces in Brazil (Almeida et al .,
2000 ) . Our seismic catalogue is relatively recent: the two earliest well-documented events
with magnitude above 4 occurred in 1807 ( 5m b in northeastern Brazil; Veloso, 2012 )

and 1861 ( 4.4 m b in southeastern Brazil). The maximum observed earthquake, with a

magnitude of 6.2 m b , occurred on 31 January 1955, in the northern part of the Parecis basin
( Figure 3.2a ) . There are sparse historical data on a possibly large earthquake in the central
Amazon basin in 1690, which could have magnitude

7(Veloso 2013 ) , but no reliable

accounts allow a confirmation.
Paleoseismological studies in northeastern Brazil (Bezerra et al ., 2005 , 2011;
Rossetti et al ., 2011 ) have described the collective occurrence of liquefaction structures
in Quaternary gravels, which are consistent with magnitudes as high as 6.0-7.0. These

paleoseismological studies are still patchily distributed in Brazil, but show that magnitudes
upto7m b should be considered in hazard estimates.
The distribution of epicentres in Figure 3.2a is much affected by the population dis-
tribution: the higher population density in southeastern and northeastern Brazil results in
a large number of historical events reported in newspapers, old journals, and topics. To
be able to compare seismicity rates in different parts of the country a more geographi-
cally uniform coverage is necessary. For this reason, we filtered the Brazilian catalogue
( Figure 3.2a ) with time-variable thresholds to produce the “uniform” epicentral map of
Figure 3.2b . We selected events with magnitudes higher than 6 since 1940 (probably
detectable with the worldwide stations reporting to the ISS bulletin); magnitudes higher
than 5.0 since 1962 due to the increased coverage of the WWSS network; magnitudes
higher than 4.5 since 1968 due to the installation of the Brasilia array in 1967, the WWSSN