Imaging Shallow Cascadia Structure with Ambient Noise Tomography

Robert W. Porritt and Richard M. Allen


Figure 2.40: A - depth slice at 35km below mean sea level. This is plotted in best fit velocity to show crustal to uppermost mantle structure. B - slice at 100km depth plotted in relative velocity to highlight lateral variation.
\epsfig{file=porritt09_fig1.eps, width=8cm}\end{center}\end{figure}

Figure 2.41: A - Longitudinal cross section along 123.2$^{\circ}$W. B - Latitudinal cross section along 44.5$^{\circ}$N. C - same as A, but in best fit velocity. D - same as B, but in best fit velocity. Key aspects of the model are highlighted. Thin red and thin green lines show the estimate of Moho based on greatest vertical gradient within $\pm$ 10km of Crust2.0
\epsfig{file=porritt09_fig2.eps, width=8cm}\end{center}\end{figure}

Along strike variation has been observed along the Cascadia Subduction Zone in studies with a variety of data sets. Body-wave tomography shows a broad zone in the center of the slab with a weak high velocity signal in an atypically quiescent seismic zone (Obrebski and Allen, 2009). Primitive basalts in the arc volcanoes change characteristics along strike, defining four distinct magma sources or plumbing systems (Schmidt et al, 2007). Most striking, however, is the change in recurrence rate of episodic tremor and slip throughout the region (Brudzinski and Allen, 2007). These disparate observations may reflect regional variations in the lithosphere. This study seeks to connect these previous observations by developing a short period surface wave model of structure in the region using ambient seismic noise as the source.

Data Processing

Data for this study comes from the Berkeley Digital Seismic Network (BDSN), Southern California Earthquake Center, the Canadian National Seismic Network, and USArray with a focus on two Flexible Array Experiments. The Flexible Array deployments, FlexArray Along Cascadia For Segmentation (FACES) and Mendocino Broadband, were deployed in 2007 and are nearing the completion of their two year deployments. This is one of the first studies utilizing the approximately one hundred stations in these broadband experiments.

Detailed processing flow for computing group and phase velocity maps can be found in Benson et al, 2007. Group and phase velocity mean values and errors are computed using a jack knife approach from cross correlations of 11 months of continuous data. Dispersion curves over the model space and their corresponding errors are utilized in a Monte Carlo inversion scheme (Shapiro and Ritzwoller, 2002) using PREM as a starting model to compute smooth one dimensional velocity profiles on each tenth of a degree by tenth of a degree node from the surface to 150km depth. The profiles are concatenated together and corrected for topography to build a three dimensional model. To better visualize lateral variations in the mantle, a one-dimensional mean model is extracted and used to compute a three dimensional model of velocity relative to that one-dimensional model.


Figure 2.40 shows two representative slices at constant depth in the Pacific Northwest. In the 35km depth slice, the semi-circular pattern of the Olympic Peninsula (OP), with faster rocks in the center and lower velocities around the outer part, is clearly seen, and the crustal root of the Klamath Mountains (KM) is also clear as a slow mass in the mantle. In the 100km depth section, the overlay of the main arc volcanoes closely matches the edge of the slab. Figure 2.41 shows two cross sections along 123.2$^{\circ}$W and 44.5$^{\circ}$N respectively. Figures 2.41 A and B are shown in velocity relative to a one-dimensional model, while C and D are in best fit velocity from the inversion. Overlain are the slab contours from Audet et al, 2009 in black and black circles showing earthquakes.

The structure is largely consistent with what would be expected; however, unexpected structural variations are apparent. Examples of expected structures include: deep crustal roots of the Klamath mountains, the border between the slab window and the slab following the trend of the Mendocino Fracture Zone, and the high velocity slab dipping the same way as the receiver functions suggest. The offset between the receiver function top of the slab and this image may reflect a layer of oceanic sediments of varying thickness overlying the main basaltic oceanic lithosphere in the subducting slab.

There is a reduction in the velocity of the subducting Juan de Fuca slab between 45$^{\circ}$N and 47$^{\circ}$N in a similar location as the weakening of the high velocity slab shown in Obrebski and Allen, 2009. There is also a high velocity lower crustal layer between 43$^{\circ}$N and 47$^{\circ}$N which correlates with the Siletzia terrain (Brudzinksi and Allen, 2007). The Siletzia terrain has the longest recurrence interval of episodic tremor and slip (ETS) events throughout the subduction zone. Because this is a high shear wave velocity zone and ETS is less active in this zone, it is likely this is a region of lower fluid content than the rest of the subduction zone. Further analysis incorporating receiver functions and estimates of Vp/Vs ratio could confirm this finding.


We would like to acknowledge our co-PI's and collaborators on the flexible array experiments. The Mendocino Broadband experiment was made possible through NSF grants EAR0643392 and EAR0745934, with help from Gene Humphreys, Leland O'Driscoll, Alan Levander, and Yongbo Zhao for fieldwork and discussions. The FlexArray Along Cascadia was funded through NSF grant EAR0643007 with co-PI Mike Brudzinksi and his students Devin Boyarko and Stefany Sit.

We thank Morgan Moschetti, Michael Ritzwoller, Yingji-Yang, Nikolai Shapiro, and Fan-Chi Lin for help with the ambient noise tomography processing flow to create the model. Data from this study came from the Earthscope USArray/Transportable Array, the Canadian National Seismic Network through the AutoDRM system, the Berkeley Digital Seismic Network, and the Southern California Earthquake Center.

Field assistance has been provided by the following: Derry Webb, Marcos Alvarez, Lloyd Carothers, Eliana Arias Dotson, Pat Ryan, Lisa Linville, Pallavi Chethan, Kevin Jensen, Chris McMillan, Stefany Sit, Andrew Tran, Summer Ohlendorf, Dan'L Martinez, Valerie Zimmer, Will Levandowski, Amanda Thomas, Heidi Reeg, Nickles Badger, Tom Owens, Eileen Evans, Holly Brown, Joanne Emerson, Ajay Limaye, Rick Lellinger, Mathias Obrebski, Samanta Cubias, Caroline Eakin, Liang Zhao, and David Belt.

This work has been made possible with the resources available through the PASSCAL instrument center at New Mexico Tech.


Audet, P., Bostock, M. G., Christensen, N. I., and Peacock, S. M., Seismic evidence for overpressured subducted oceanic crust and megathrust fault sealing, Nature, 457, 76-78, 2009.

Benson, G.D., Ritzwoller, M.H., Barmin, M.P., Levshin, A.L., Lin, F., Moschetti, M. P., Shapiro, N.M., Yang, Y., Processing seismic ambient noise data to obtain reliable broad-band surface wave dispersion measurments, Geophysical Journal International, 169, 1239-1260, 2007.

Bostock, M.G., Hyndman, R.S., Rondenay, S., and Peacock, S. M., An inverted continental Moho and serpentinization of the forearc mantle, Nature, 417, 53-539, 2002.

Brudzinksi, M. and Allen, R. M., Segmentation in Episodic Tremor and Slip All Along Cascadia, Geology. 35 (10) 907-910, 2007.

Obrebski, M. J., and Allen, R. M., Plume Vs. Plate: Convection beneath the Pacific Northwest, Berkeley Seismological Laboratory Annual Report, 2009.

Schmidt, M. E., Grunder, A.L., and Rowe, M., Segmentation of the Cascades Arc as indicated by Sr and Nd isotopic variation among primitve basalts, Earth and Planetary Science Letters, 266, 166-181, 2007.

Shapiro, N. M., and Ritzwoller, M. H., Monte-Carlo inversion for a global shear velocity model of the crust and upper mantle, Geophysical Journal International, 151, 88-105, 2002.

Xue, M., and Allen, R. M., The Fate of the Juan de Fuca Plate: Implications for a Yellowstone Plume Head, Earth and Planetary Science Letters, 264, 266-276, 2007.

Berkeley Seismological Laboratory
215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760
Questions or comments? Send e-mail:
© 2007, The Regents of the University of California