MichSeis seismograph files each contain one hour of data.
The analysis program reads in the hour-long files, uses
a moving window, duration about one minute, to calculate the
spectral amplitude for periods that range from 4 to 12 sec.
The maximum amplitude value for each period within each hour
is kept and displayed. As seen in the graph, I analyzed each
and every hour around the storm peak, but only analyzed a few
hours per day in the quiet time before and after the passage of
"Bonnie". Since the "Bonnie" microseisms are dominated by a period 
of about 5 sec, the graph just plots the microseism 
amplitude for a 5 sec period.
But for the "Georges" microseisms, there were significant
amplitude peaks for periods of 5 sec and 8 sec.

A few earthquakes occurred around the world during the time
intervals displayed, I stepped over the hours that included
these events.

To interpret the amplitude values of the graph:
a value of, say, 1 micron/sec for a particular hour means that
somewhere in that hour-long seismogram, there was a burst
of 5-sec microseism waves that had an average amplitude
(over a one-minute time interval) of 1 micron/sec for the
vertical ground velocity. 

<LJR>