Siegfried R.C. Rohdewald Intelligent Resources Inc. Vancouver, British Columbia Canada

1. SAGEEP 2014 Refraction/Reflection SessionOptimized interpretation of SAGEEP 2011 blind refraction data with Fresnel Volume Tomography and Plus-Minus refraction Siegfried R.C. Rohdewald Intelligent Resources Inc. Vancouver, British Columbia Canada

2. Smooth Inversion = 1D gradient initial model +2D WET Wavepath Eikonal Traveltime tomography Get minimum-structure 1D gradient initial model : Top : pseudo-2D DeltatV display • 1D DeltatV velocity-depth profile below each station • 1D Newton search for each layer • velocity too low below anticlines • velocity too high below synclines • based on synthetic times for Broad Epikarst model (Sheehan, 2005a, Fig. 1). Bottom : 1D-gradient initial model • generated from top by lateral averaging of velocities • minimum-structure initial model • DeltatV artefacts are completely removed

3. 2D WET Wavepath Eikonal Traveltime inversion • rays that arrive within half period of fastest ray : tSP + tPR – tSR <= 1 / 2f (Sheehan, 2005a, Fig. 2) • nonlinear 2D optimization with steepest descent, to determine model update for one wavepath • SIRT back-projection step, along wave paths instead of rays • natural WET smoothing with wave paths (Schuster 1993, Watanabe 1999) • partial modeling of finite frequency wave propagation • partial modeling of diffraction, around low-velocity areas • WET parameters sometimes need to be adjusted, to avoid artefacts • see RAYFRACT.HLP help file Fresnel volume or wave path approach :

5. Interpretation with Plus-Minus refraction method (PMR) • Assign traces to 3 layers : weathering layer, overburden layer, basement. See next slide. • Velocity and thickness of weathering layer determined with slope-intercept, between adjacent shot points. Model can vary laterally. • Velocity of overburden layer also determined with slope-intercept method • Bottom of overburden layer (top of basement) determined with Plus-Minus refraction (PMR) • Velocity of basement from PMR

6. Semi-automatic mapping of traces to refractors • yellow is weathering layer, red is overburden layer, green is basement • mapping is not required for Smooth inversion. Station spacing is 3m. • specify 1D velocity model : upper velocity limits for weathering, 1st refractor • specify lateral and vertical smoothing of CMP-sorted traveltime field • map traces to refractors by matching apparent velocity to 1D velocity model • lateral smoothing of crossover distance, after mapping to refractors

7. Plus-Minus refraction interpretation • basement velocity dips to below 2,000 m/s at station no. 63 • this hints at a basement fault zone, dip of fault is not visible • lateral smoothing of refractors, for Plus-Minus method (Hagedoorn, 1959) • overburden refractor colored blue, basement refractor colored black

8. 1D initial model : smooth DeltatV inversion

9. WET with Ricker wavelet weighting Wavepath width 30%, 100 Iterations Wavepath width 10%, 100 Iterations Wavepath width 5%, 100 Iterations Wavepath width 3%, 100 Iterations e) 7th run, 6th run as starting model, wavepath width 3%, 100 WET iterations, RMS error 0.6% Figure 1: WET with wavepath velocity update weighted with a Ricker wavelet (a) 1D-gradient starting model, (b) velocity tomogram obtained with wavepath width 30%, (c) 10%, (d) 5%, (e) 3%. Color scale is velocity in m/s. Contour interval is 250 m/s. Horizontal axis is offset from first profile receiver, in m. Vertical axis is elevation in m. Overburden Plus-Minus refractor is colored cyan, basement refractor colored orange. These two refractors are the same in (b) through (e). See Figure 4.

10. WET with Gaussian weighting Wavepath width 30%, 100 Iterations Wavepath width 10%, 100 Iterations Wavepath width 5%, 100 Iterations Wavepath width 3%, 100 Iterations

11. Wavepath coverage for Ricker weighting Wavepath width 30%, 100 Iterations Wavepath width 10%, 100 Iterations Wavepath width 5%, 100 Iterations Wavepath width 3%, 100 Iterations

12. Discussion • Wide wavepaths (low frequency) make WET inversion less dependent on the starting model and more robust but produce a smooth tomogram • Narrow wavepaths can give a sharper tomogram, but WET becomes more dependent on the starting model (previous run) and less robust • WET images the dipping low-velocity fault zone (Zelt et al., 2013) more realistically with iteratively decreasing wavepath width • Contours for velocity 2,500 m/s and higher velocities become more parallel to the fault zone, which dips down to the right (towards offset 250m at elevation of -80m). • Thin wavepaths make WET tomography more prone to generating artefacts, especially with bad or noisy first break picks and strong refractor curvature.

13. Conclusions • Plotting 1.5D Plus-Minus refractors on the WET tomogram allows interactive adaptation of parameters, until the layered analysis matches the 2D velocity tomogram • Layered refraction modeling is non-unique and subjective due to mapping of traces to assumed refractors and lateral smoothing, necessary for refractor velocity estimation and time-to-depth conversion • WET interpretation depends on the maximum allowed basement velocity, which may not be well-constrained by the first break picks • Gaussian weighting can produce more focused tomograms for wide wavepaths than Ricker weighting • Tomograms obtained with an iterative approach of wavepath adjustment show improvement compared to the standard ray-based approach.