Amplitudes and Waveform Inversion

   As with any seismic reflection data set, seismic oceanography data contain more information than just the images of ocean structure themselves.  The amplitudes of the seismic reflections contain a potential gold mine of information about the seismic impedance structure of the ocean, and, therefore, about the temperature structure of the ocean.  The first observations of this came in our 2003 Science paper, in which we noted (p. 821) the correlation between seismic reflection amplitudes and the sound-speed changes from a coincident XBT.  This correlation was further demonstrated by our subsequent joint PO/SO work in the Norwegian Sea (Nandi et al., 2003), using many more XBT's (see especially Figs. 3 and 4 from that paper).

   Since those early observations, we've been working on several more quantitative approaches to modeling and inverting the amplitudes of seismic oceanography data, including AVO (amplitude-versus-offset) modeling and waveform inversion.


AVO modeling.

   In 2005, Pedro Páramo and I published in GRL (download here) the first demonstration that the magnitude of temperature contrasts could be successfully deduced from seismic reflection data.  The AVO plot below shows that amplitude information on pre-stack reflection data can be successfully modeled, and that the reflections in this data set are best modeled as the result of temperature, not density, contrasts.  Moreover, the sound-speed (and thus temperature) contrasts gleaned from the modeling match those observed in XCTD data collected simultaneously.

Paramo_Holbrook_fig4

(Click on the image to see a larger version.)


Waveform inversion.

   More recently, Warren Wood and I have been collaborating on taking AVO modeling to the next step, by incorporating formal inversion of the waveforms (wiggle-by-wiggle matching of seismic reflections).  The first results of this work have been published in GRL (download here).  The figure below shows a synthetic result demonstrating that temperature profiles can (given perfect data) be reconstructed with great accuracy from seismic reflection data (but you'd better read the paper to follow the details).  Continuing this work with new and better data is a major focus of our ongoing work.

Wood_et_al_Fig1

(Click on the image to see a larger version.)

To reach me, send me an email.
Or call me at work:  307-766-2427
(But email is better.)