Large 19th Century Earthquakes in Eastern/Central North America: A Comparative Analysis - Historical Seismology: Interdisciplinary Studies of Past and Recent Earthquakes - page 342

biases that are difficult to quantify. For this reason, the comparison between the New
Madrid earthquake magnitudes and the magnitude of the Charleston earthquake is
especially illuminating. That is, while analysis of the New Madrid intensity values
alone might be fraught with uncertainty, a direct comparison with intensity values
for the Charleston earthquake can help constrain the relative sizes of the events.
More recently, Bakun and Hopper (2004) estimated magnitude values for the
New Madrid mainshocks using a new method, one in which intensity versus dis-
tance observations are used together with attenuation relationships developed from
instrumentally recorded earthquakes in central/eastern North America (Bakun et al.,
2003). His preferred estimates are 7.6, 7.5, and 7.8 for NM1, NM2, and NM3, re-
spectively. These estimates are described as M-I, indicating that they are derived
from intensity data. Because the attenuation relationships are derived using Mw
values, it is generally assumed that M-I represents Mw. The approach of Bakun
et al. (2003) does not require isoseismal contours and is thus less subjective than
the method of Johnston (1996). However, Bakun's method reintroduces the problem
that Johnston (1996) attempted to solve with his SCR compilation—namely, the lack
of true calibration events for the largest historical earthquakes. This re-introduces
the need for extrapolation, and its attendant uncertainties. For example, Bakun and
Hopper (2004) consider two different extrapolation techniques, the one that leads
to the preferred values and a second technique that yields values about 0.3 units
smaller. A further potential difficulty is that Bakun et al. (2003) use the 1929 Mw7.3
Grand Banks earthquake (Bent, 1995) to develop their attenuation relationship—the
only Mw

7 earthquake in their dataset. However, this event was located offshore
from Newfoundland, Canada, arguably in a very different tectonic setting than the
New Madrid events. Also, because the event occurred several hundred kilometers
off-shore, its macroseismic effects are not well documented.
Investigation of the Charleston earthquake dates back to the immediate post-
earthquake investigations led by Clarence Dutton, an Army officer detailed to the
U.S. Geological Survey. This effort culminated in the publication of one of the ear-
liest comprehensive, scientific reports of a large earthquake (Dutton, 1889). The so-
called “Dutton Report” includes thorough and consistent compilations of near-field
geological effects of the earthquake and far-field macroseismic effects. Whereas
about 100 or fewer intensity values are available for each of the New Madrid main-
shocks, the Dutton Report provides the basis for assignment of over 1000 intensity
values. In a comprehensive interpretation of these accounts, Bollinger (1977) as-
signed almost 800 intensity values based on the 1337 intensity reports tallied by
Dutton (1889). Bollinger (1977) estimated an m b value of 6.8-7.1 using the same
techniques that Nuttli (1973) used to estimate magnitudes for the New Madrid earth-
quakes.
Whereas an initial review of earlier intensity assignments for the New Madrid
earthquakes suggested immediate biases (and a small handful of outright mistakes)
that led to the reinterpretation by Hough et al. (2000), a initial review of the inten-
sity assignments by Bollinger (1977) reveals the values to have been assigned care-
fully and in keeping with modern conventions. Each of the accounts were evaluated

>