@@ -1084,29 +1084,37 @@ covariance functions are all practically indistinguishable.
%% TRANSIENT STRAIN RATES
\subsection{Transient Strain Rates}\label{sec:Results}
% Stating how we calculate strain and refering to the figures
Having established a noise model and a prior for transient
displacements, we use the GNSS dataset to calculate transient strain
rates in the Puget Sound region. We calculate transient strain rates
for each day from January 1, 2010 to May 15, 2017. The strain rates
are estimates at a grid of points spanning the study area. In Figure
\ref{fig:StrainMap} we show the transient strain rates on January 1,
2016, which coincides with the height of the winter 2015-2016 SSE. We
have included an animation showing the map view of strain rates
through time as supplementary material. The strain rates shown in
Figure \ref{fig:StrainMap} are generally similar to the strain rates
during the other six SSEs considered in this study. The SSEs cause
trench perpendicular compression in the Olympic Peninsula and
extension east of Puget Sound. The strain transitions from compression
to extension around the southern tip of Vancouver Island, which
coincides with the location of thrust slip for SSEs in the Puget Sound
region \citep[e.g.,][]{Dragert2001,Wech2009,Schmidt2010}. Thus, this
pattern of strain is to be expected. During the period in between
SSEs, secular strain rates indicate trench perpendicular compression
throughout this study region \citep{Murray2000, McCaffrey2007,
McCaffrey2013}. When comparing inferred strain rates from SSEs to the
secular strain rates, we see that SSEs are concentrating tectonically
accumulated strain energy towards the trench, and presumably pushing
the subduction zone closer to failure.
displacements, we can now use the GNSS data to calculate transient
strain rates, $\strain$, in the Puget Sound region. We evaluate
$\strain$ at a grid of points spanning the study area for each day
from January 1, 2010 to May 15, 2017. In Figure \ref{fig:StrainMap},
we show a map view of $\strain$ on January 1, 2016, which coincides
with the height of the winter 2015-2016 SSE. In Figure
\ref{fig:StrainTs}, we show a time series of $\strain$ at a position
on the Olympic Peninsula, which tends to be where $\strain$ is the
largest during SSEs. We also include a supplementary animation showing
a map view of $\strain$ over time.
% Describe the strain during SSEs and tectonic implications
The strain rates shown in Figure \ref{fig:StrainMap} are generally
similar to the strain rates during the other six SSEs considered in
this study. The SSEs cause trench perpendicular compression in the
Olympic Peninsula and extension east of Puget Sound. The strain
transitions from compression to extension around the southern tip of
Vancouver Island, which coincides with the location of thrust slip for
SSEs in the Puget Sound region \citep[e.g.,][]{Dragert2001, Wech2009,
Schmidt2010}. Thus, this pattern of strain is to be expected. During
the period in between SSEs, secular strain rates indicate trench
perpendicular compression throughout this study region
\citep{Murray2000, McCaffrey2007, McCaffrey2013}. When comparing
inferred strain rates from SSEs to the secular strain rates, we see
that SSEs are concentrating tectonically accumulated strain energy
towards the trench, and presumably pushing the subduction zone closer
to failure.
%A key difference between the strain inferred here and strain that can
%be derived from fault slip models is that our estimated strain rates
@@ -1122,21 +1130,23 @@ the subduction zone closer to failure.
\begin{figure*}
\includegraphics{figures/strain_map/strain-map.pdf}
\caption{
Estimated transient strain rates during the Winter 2015-2016 SSE.
Strain glyphs show the normal strain rate along each azimuth, where
orange indicates compression and blue indicates extension. The shaded
regions indicate one standard deviation uncertainties in the normal
strain rates.
Transient strain rates during the Winter 2015-2016 SSE. The glyphs
show the normal strain rates as a function of azimuth, where orange
indicates compression and blue indicates extension. The shaded regions
indicate the one standard deviation uncertainties for the normal
strain rates. It is common to depict strain rates by showing just the
principal normal strain rates; however, we deviate from this
convention because there is no straight-forward way of showing the
uncertainties for the principal normal strain rates.
}
\label{fig:StrainMap}
\end{figure*}
In Figure \ref{fig:StrainTs} we show the time dependence of estimated
transient strain rates at a position on the Olympic Peninsula, where
transient strain rates from SSEs are largest. To verify that the
estimated transient strain rates are accurately identifying
geophysical signal, we compare the signal-to-noise ratio from eq.
(\ref{eq:SNR}) to the frequency of seismic tremor (Figure
% compare seismic tremor to SNR
To verify that the estimated transient strain rates are accurately
identifying geophysical signal, we compare the signal-to-noise ratio
from eq. (\ref{eq:SNR}) to the frequency of seismic tremor (Figure
\ref{fig:StrainMag}). A signal-to-noise ratio greater than ${\sim}3$
can be interpreted as a detected geophysical signal. We detect nine
distinct events, which each correspond to peaks in seismic tremor. The