Problem Set 1: Introduction To Matlab, Featuring The Frequency-Magnitude Relation

Problem Set 1:
Introduction to Matlab, featuring the frequency-magnitude relation

GEOS 626: Applied Seismology, Carl Tape
Assigned: January 17, 2018 — Due: January 29, 2018
Last compiled: January 15, 2018
Overview and instructions

• The purposes of this problem set are to practice using Matlab and to think critically about
histograms. (If you are new to Matlab, let me know.)
• Suggested reading: Stein and Wysession (2003, Section 4.7.1), Gutenberg and Richter
(1944)
• See the handout doc_startup.pdf to get set up on the Linux network. This includes
copying the template file:
cp hw_gr_template.m hw_gr.m
• Possible source of confusion: Following standard mathematical notation, I will use

‘log10 ’ to represent the base-10 logarithm and ‘ln’ to represent the natural logarithm. Note
that in Matlab log10 is log10 and log is ln. (Note that Stein and Wysession (2003) use
‘log’ for the base-10 logarithm.)
• Matlab tips: hw_gr_template.m shows an example of how to plot multiple items on a

log-scaled plot. It also shows how to produce a postscript (PS) or pdf file from a Matlab
file. Make sure you follow the template script before proceeding.
• Latex option: If you want to use Latex for your homeworks, see the templates in the
folder latex (and see the README).
The Gutenberg-Richter frequency-magnitude relation

This problem addresses the famous Gutenberg-Richter frequency-magnitude relation (Gutenberg
and Richter , 1944), which is one of the cornerstone empirical relationships in seismology. It is
given by
log10 N = a − b M, (1)
where
• N is the cumulative number of earthquakes having magnitudes larger than M that occur
in region R within a particular time T
• M is the earthquake magnitude; we take this to be moment magnitude Mw
• b controls the slope of the seismicity distribution in region R within a particular time T ;
b ≈ 1 for most earthquake catalogs. The line will always increase to the left, so Equation (1)
has a negative slope and therefore b > 0 by definition.
• a indicates the seismic activity in region R within a particular time T
1
A subtle point is that in practice the magnitudes are binned, then a line is fit to the histogram.
It is helpful to consider the discrete form of Equation (1):
log10 Ni = a − b Mi , (2)
where the index i refers to the magnitude bin, with Mi being the magnitude at the left boundary
of the bin.
• The cumulative distribution is the cumulative number of events with M ≥ Mi ; the frequency-
magnitude relation is expressed as a cumulative distribution.
• The incremental distribution is the number of events per magnitude bin, Mi ≤ M ≤ Mi+1 .
• A magnitude interval is a range of magnitude, e.g., the magnitude interval [8.7, 9.0].
Problem 1 (10.0). Histograms and earthquake statistics

Make sure that you are very comfortable with the meaning of Equation (1) before
proceeding.
1. (0.2) We consider the Global Centroid Moment Tensor (GCMT) catalog (www.globalcmt.org),
from 01-Jan-1976 to 30-June-2011. Thus T represents the duration of the catalog, and R
represents planet Earth.
Run the Matlab program hw_gr_template.m to generate a global map of the catalog.
(a) (0.0) What is the range of depths of events in the catalog?

(b) (0.1) List three regions of the deepest seismicity.
(c) (0.1) What is the range of magnitudes of events in the catalog?
2. (0.8) As shown in hw_gr_template.m, use the function seis2GR.m to obtain the cumulative
and incremental distributions for the Mw values of the GCMT catalog, using a bin width
of ∆M = 0.1. Examine the output that appears in the command window.
Note: The function seis2GR.m uses the variables names Ncum for the cumulative numbers
(Ni ), N for the incremental numbers, and Medges for the bin edges Mi .
(a) (0.2) What is the maximum value of the incremental distribution?

What is its magnitude interval?
(b) (0.2) What is the maximum value of the cumulative distribution?
(c) (0.2) What is the minimum value of the incremental distribution over the interval
[4.2, 9.1]?
(d) (0.2) What is the minimum value of the cumulative distribution over the interval
[4.2, 9.1]?
2
3. (3.0) Now it’s time to write some lines of code. Using the output from seis2GR.m (as shown
in hw_gr_template.m), plot the cumulative and incremental distributions on the
same plot, similar to the plot in Stein and Wysession (2003, Figure 4.7-2), but note that
your x-axis is Mw , not log10 M0 , and your y-axis is number of earthquakes, not number of
earthquakes per year.
Matlab tip: hw_gr_template.m shows an example of how to plot multiple items with
semilogy axes. Another alternative is to transform N into n = log10 N , then work with n.
(a) (0.0) Which distribution has more scatter?

(b) (2.2)
i. (0.6) Find a best-fitting line, log10 N = a − bM , for the ‘most linear’ section
of the log10 -scaled cumulative distribution. You can simply pick two points and
compute the line, or use a command such as polyfit (and polyval) to apply a
least-squares fit to a set of points. (Do not fit a line to the entire distribution!)
What are your values for a and b? (Show your work.)
ii. (1.0) Plot your best-fitting line over the full range of magnitudes and include this
plot in your write-up.
iii. (0.3) What is the physical (not just geometrical!) meaning of a?
iv. (0.3) What is the physical (not just geometrical!) meaning of b?
(c) (0.6) Assume that the best-fitting distribution (not the GCMT catalog) is ‘reality’.
Based on the idealized cumulative distribution, what is largest earthquake expected
over the duration of the GCMT catalog? (Show your work.)
Hint: Where does N = 1 intersect your best-fitting line?
Note: The expected value does not have to agree with what actually occurred within
the GCMT catalog.
(d) (0.2) The “catalog completeness” (e.g., Wiemer and Wyss, 2000), Mc , represents the
smallest magnitude of above which the frequency-magnitude is true for a particular
seismicity catalog. What is the catalog completeness for GCMT? List your answer
with 0.1 precision. (Provide a brief explanation, but no computation is necessary.)
4. (1.0) Instead of analyzing seismicity, let us now analyze seismicity rate by dividing all
binned values by the duration of the catalog (T , in years).
(a) (0.1) Takeing an average over the entire time interval of the GCMT catalog, how many
earthquakes per year are there?
(b) (0.1) Why is seismicity rate more useful than seismicity?
(c) (0.5) What is the best-fitting line (namely, a and b) for the new distribution? You
can determine this graphically or analytically, from your previous results. (Show your
work.)
(d) (0.1) What magnitude interval averages >100 events per year?
(e) (0.1) How many M ≥ 0 earthquakes are expected on Earth per year?
(f) (0.1) How many earthquakes (of any magnitude) are expected on Earth per year?
3
5. (1.0)
(a) (0.8) Plot the cumulative and incremental distributions for seismicity rate for bin
widths of ∆M = 0.05, 0.10, 0.5, and 1.0.
(b) (0.2) What is the apparent relationship between bin width and the separation between
the cumulative and incremental distributions?
6. (2.0) Define the bin width as
∆M = Mi+1 − Mi (3)
where i increases to the right (the usual convention).

The incremental distribution is given by
−∆Ni = −(Ni+1 − Ni ) = Ni − Ni+1 . (4)
(a) (1.7) Using the discrete frequency-magnitude relation (Eq. 2), show that the incre-
mental distribution can be written as
log10 (−∆Ni ) = a + ∆a − bMi (5)
where ∆a is a function in terms of other variables (but not i).

List the expression for ∆a.
(b) (0.1) What is the relationship between bin width and the shift in the y-intercept?
(c) (0.2) If b = 1 and ∆M = 0.1, what is ∆a?
7. (1.5)
(a) (0.3) Differentiate Equation (1) to obtain dN/dM .

(b) (1.0) Using your expression from (a), as well as the mathematical definition of a
derivative, derive an expression analagous to Equation (5) that is valid for small bin
widths, ∆M ≪ 1. List the expression for ∆a
(c) (0.2) If b = 1 and ∆M = 0.1, what is ∆a?
8. (0.5) Earlier you determined the catalog completeness, Mc .
(a) (0.1) Can the GCMT catalog, M > Mc (see earlier part of this problem for Mc ), be
fit well with a single line?
(b) (0.2) Compute an estimate b for M > 7.5.
(c) (0.1) What does the different b value imply about large events in the catalog?
(d) (0.1) What is a possible reason for this?
Problem
Approximately how much time outside of class and lab time did you spend on this problem set?
Feel free to suggest improvements here.
4
References
Gutenberg, B., and C. F. Richter (1944), Frequency of earthquakes in California, Bull. Seis-
mol. Soc. Am., 34 (4), 185–188.
Stein, S., and M. Wysession (2003), An Introduction to Seismology, Earthquakes, and Earth
Structure, Blackwell, Malden, Mass., USA.
Wiemer, S., and M. Wyss (2000), Minimum magnitude of completeness in earthquake catalogs:
Examples from Alaska, the western United States, and Japan, Bull. Seismol. Soc. Am., 90 (4),
859–869.

Problem Set 1: Introduction To Matlab, Featuring The Frequency-Magnitude Relation

Uploaded by

Copyright:

Available Formats

Problem Set 1: Introduction To Matlab, Featuring The Frequency-Magnitude Relation

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Problem Set 1: Introduction To Matlab, Featuring The Frequency-Magnitude Relation

Uploaded by

Copyright:

Available Formats

Problem Set 1:

Introduction to Matlab, featuring the frequency-magnitude relation

Overview and instructions

• Possible source of confusion: Following standard mathematical notation, I will use

• Matlab tips: hw_gr_template.m shows an example of how to plot multiple items on a

The Gutenberg-Richter frequency-magnitude relation

• M is the earthquake magnitude; we take this to be moment magnitude Mw

• a indicates the seismic activity in region R within a particular time T

Problem 1 (10.0). Histograms and earthquake statistics

(a) (0.0) What is the range of depths of events in the catalog?

(a) (0.2) What is the maximum value of the incremental distribution?

(a) (0.0) Which distribution has more scatter?

6. (2.0) Define the bin width as

where i increases to the right (the usual convention).

−∆Ni = −(Ni+1 − Ni ) = Ni − Ni+1 . (4)

log10 (−∆Ni ) = a + ∆a − bMi (5)

where ∆a is a function in terms of other variables (but not i).

(a) (0.3) Differentiate Equation (1) to obtain dN/dM .

8. (0.5) Earlier you determined the catalog completeness, Mc .

You might also like