Auxiliary material for
Continental microseismic intensity delineates oceanic upwelling timing along the west coast of
North America
Richard E. Thomson1, Martin Heesemann2, Earl E. Davis3, and Roy A. S. Hourston1
1
Fisheries and Oceans Canada, Institute of Ocean Sciences
P.O. Box 6000, 9860 West Saanich Road, Sidney, BC, V8L 4B2, Canada
2
Ocean Networks Canada, University of Victoria
PO Box 1700 STN CSC, 2300 McKenzie Avenue Victoria, BC V8W 2Y2, Canada
3
Pacific Geosciences Centre, Geological Survey of Canada
P.O. Box 6000, 9860 West Saanich Road, Sidney, BC, V8L 4B2, Canada
Geophysical Research Letters
Introduction
The auxiliary material for this study consists of two Figures and one Table. Figure S1 (fs01.tiff)
presents a comparison of the 15-min time series of the power spectral density for the
microseismic signal recorded during March-April 2012 at the Pacific Geosciences Centre (PGC)
seismological station versus the time series of the 6-hourly running median values of the 15-min
spectral series. Units are given by the left-hand y-axis. The figure illustrates that use of the
median time series removes the spectral contributions from small intermittent earthquakes while
preserving the contributions from the wind-wave induced microseismic signal in the 3-7 s
frequency band. The figure also includes the 6-hourly running median spectral density time
series derived from 15-min spectral data for the CORK bottom pressure recorder located at
~2600 m depth to the west of Vancouver Island (see Figure 1). Units are given by the right-hand
y-axis. The plots show close agreement between the microseismic and bottom pressure variations
in the 3-7 s band.
Figure S2(a) (fs02.tiff) shows the data series used to fill the gap in the PGC microseismic record
for February 1998. The thick line is the daily-mean spectral density calculated from the
microseismic record for the last two weeks of January. This time series was linearly regressed
against the daily-mean significant wave height for the same period recorded at meteorological
buoy 46206 located off the southwest coast of Vancouver Island (see Figure 1 for location). The
least-squares regressional parameters, together with the observed record of significant wave
height for February 1998, were then used to fill the spectral gap for February (thin line). Figure
S2(b) shows the full 6-hourly spectral record for 1998 (including the synthetic time series for
February) and the corresponding cumulative sums based on the three averaging periods. Vertical
lines denote the times of the spring and fall transitions for 1998 derived from the modified
spectral density time series.
Table S1 (ts01.docx) provides the mean and standard deviation of the spring and fall transition
dates (Julian Day) for the central west coast of North America for the period 1993 to 2013.
Times are derived using wind stress from NCAR/NCEP Reanalysis-1 sites along the southern
British Columbia (BC), central Washington (WA) and northern Oregon (OR) coasts and
microseismic data in the 0.24±0.10 Hz (5±2 s) frequency band recorded at the Geological Survey
of Canada seismological station on southern Vancouver Island (see Figure 1). Values in bold
font are considered representative of the transition times for the southern coast of BC. The
upwelling duration in the last column is the difference computed for each year as the difference
between the spring and fall transition dates. For reference purposes, we note that Julian day 94
corresponds to 4 April and JD 103 to April 13 (ignoring leap years). Latitude and longitude have
been rounded to the nearest minute.
1. fs01.tiff Time series of 15-min band-integrated seismic spectral density (thin solid line) and
the corresponding 6-hourly running median value of the 15-min data (thick solid line) for 2012
for the Pacific Geosciences Centre (PGC) seismological site on Vancouver Island. The 6-hourly
median value serves as a form of notch filter that eliminates contributions from relatively intense
but short-duration earthquakes. Seismic spectra are averaged over the 3-7 s band.
2. fs02.tiff The February 1998 data gap. (a) The thick line denotes the daily-mean spectral
density for the last two weeks of January 1998 obtained from the PGC microseismic record. The
thin line denotes the synthetic daily-mean record for the first weeks of February derived through
linear regression of the January microseismic record against the daily mean significant wave
height from meteorological buoy 46206 located off the southwest coast of Vancouver Island. The
regression line was used with the observed record of significant wave height for February to fill
the microseismic gap in February; and (b) the 1998 microseismic record and resulting
cumulative sum (including the synthetic time series for February) for the three averaging periods
of January-May, January-June and January-July. Vertical lines show the times of the spring and
fall transitions.
3. ts01.docx Mean and standard deviation of the spring and fall transition dates (Julian Day) for
the central west coast of North America for the period 1993 to 2013.
3.1 Column “Site: Location”, names of the sites and its latitude and longitude. Wind stress data
are from NCAR/NCEP Reanalysis-1 sites along the southern British Columbia (BC), central
Washington (WA) and northern Oregon (OR) coasts and microseismic data in the 0.24±0.10 Hz
(5±2 s) frequency band recorded at the Geological Survey of Canada seismological station on
southern Vancouver Island. Values in bold font are considered representative of the transition
times for the southern coast of BC.
3.2 Column “Spring Transition”, the mean date (Julian day) of the spring transition.
3.3 Column “Fall Transition”, the mean date (Julian day) of the fall transition.
3.4 Column “Duration (days)”, the mean duration of the upwelling period based on differences in
time (days) between the spring and fall transition dates.