6.
7 Broad Band Surface Wave Magnitude MS 175
Summary of procedures to calculate Ms
• Calculate the vertical component displacement seismogram (nm), multiply with
the WWSSN relative response function to simulate the WWSSN LP response.
• Read the maximum amplitude and corresponding period within the period range
18–22 s.
• Correct the amplitude reading for the gain for the WWSSN LP curve.
• Make sure reading is not contaminated by microseismic noise.
• Only use stations in the distance range 20–160◦ to calculate Ms . However,
measurements at distance less then 20◦ can be reported for research purposes.
• The hypocentral depth must be less than 60 km.
• When reporting observations, make sure the corresponding phase ID is reported
as IAMs_20.
6.7 Broad Band Surface Wave Magnitude MS
The limitation in the practice of determining the current Ms led IASPEI to
standardize the original surface wave magnitude relation to using the maximum
(A/T)max on a broadband vertical channel. The new magnitude has the IASPEI label
Ms_BB, however we will label it MS for simplicity. It is defined as
MS = log(A/T)max + 1.66 log(Δ) + 3.3 (6.18)
In practice it is not so easy to measure (A/T)max (like for mB ). Since A = VT/2π ,
where V is the velocity amplitude, (6.18) can be written
Ms = log(Vmax /2π ) + 1.66 log(Δ) + 3.3 (6.19)
and Vmax is reported. This formulation takes into account surface-wave maxima at
both lower and higher frequencies than Ms and thus agrees best with the current
IASPEI standard MS calibration formula (see Bormann et al., 2009). The IASPEI
recommendations are:
Measure the maximum velocity amplitude using the vertical component for
surface waves in the period range 2–60 s for earthquakes with hypocentral depth <
60 km and in the distance range 2–160◦ .
The 60 s limit is arbitrary, it could really have been any larger period, however
in practice periods of 60 s are not encountered on velocity broadband records (Peter
Bormann, personal communication). It does mean however, that if the broadband
sensor is not linear at the expected frequency of the maximum amplitude, instru-
ment correction must be made. Figure 6.18 shows an example. It is seen that the 3
traces look almost identical. Ms was determined to 8.1 (amp = 891,000 nm at 22 s)
and MS was 8.2 (365,600 nm/s) so in this case the difference is small. Figure 6.19
shows an example of medium size event. The two top traces are again, as expected,
176 6 Magnitude
Fig. 6.18 Example of determining MS . The top trace shows the original velocity broadband
record, the middle trace the record filtered for MS , the bottom the traces filtered for Ms . The station
is KBS, vertical channel. The location of the reading is indicated by max and the amplitudes are
measured peak to peak divided by 2. The origin time is 2001 0113 17:33, mb (ISC) is 6.3 and Ms
is 7.8 (PDE). Epicentral distance is 79◦ and depth is 60 km
almost identical, while the Ms trace has less high frequency noise due to the fil-
ter in the WWSSN LP response. The magnitudes calculated for this example are
Ms = 4.8 and MS = 5.4 respectively. Thus MS is closer to the average ISC Ms than
the Ms calculated for KONO and there is a substantial difference between the two
magnitudes.
One of the large advantages of using MS instead of Ms is that there are little
limitations of period and distance. This means that MS also can be used for local
Fig. 6.19 Example of determining MS . The top trace shows the original velocity long period
record, the middle trace the record filtered for MS , the bottom the traces filtered for Ms . The station
is KONO, vertical channel. The location of the reading is indicated by max and the amplitudes are
measured peak to peak divided by 2. The origin time is 1996 0623 01:17, mb (ISC) is 5.4 and Ms
is 5.4 (ISC). Epicentral distance is 66◦ and depth is 33 km
6.7 Broad Band Surface Wave Magnitude MS 177
earthquakes and might be an alternative for ML . For a large number of MS mea-
surements from earthquakes in China in a wide period and distance range, Bormann
et al. (2007) observed an excellent average correlation with ML . This suggests that
using MS , in the absence of a well calibrated local/regional ML scale, is a suitable
magnitude proxy at distances between 2◦ and 20◦ . Since MS in the local/regional
range is measured at significantly longer periods (T > 3–15 s) than ML (T < 1 s) it is
much less affected by local/regional differences in crustal structure and attenuation
conditions than ML . This justifies the use of a global average relationship in areas
of unknown attenuation. However, depending on local noise conditions, magnitude
must be above ML 3–4 in order for the surface waves to have significant amplitude
above the noise, see Fig. 6.20.
Fig. 6.20 Calculating MS for a local earthquake. Top trace shows the original vertical broadband
velocity record from station HPK, the UK network. The bottom trace shows the instrument cor-
rected velocity trace using the MS filter (4 pole filter). The event origin time is 2007 0428 07:18,
epicentral distance is 370 km, the depth is 5.3 km and ML = 4.3
This is a local event suitable for MS . Note that the maximum amplitude surface
wave on the original trace appears at higher frequency than the filtered trace due to
the high frequency nature of this local event. The MS calculated was 3.9, similar to
MW of 3.9 (see Chapter 7), but smaller than the ML of 4.3.
The measurement and reporting of Amax , Vmax and T from clearly developed
surface waves of earthquakes at larger depths are strongly encouraged for future
studies on depth-dependent magnitude corrections and attenuation relationships.
Summary of procedures to calculate MS
• Calculate the vertical component velocity seismogram in the period range 3–60 s.
• Read the maximum velocity amplitude and corresponding period.
• Make sure reading is not contaminated by microseismic noise.
• Only use stations in the distance range 2–160◦ to calculate MS .
• The hypocentral depth must less than 60 km.
• When reporting observations, make sure the corresponding phase ID is reported
as IVMs_BB.
