Travel-time residuals. Rather than deluge you with individual plots, I have combined the raw travel time residuals into Flash files that allow you to rapidly change from azimuth to azimuth and more easily see variations reflecting the position and depth of source anomalies. I hope to add a similar view of the post-inversion anomalies shortly. Although you can download the Flash file for the demeaned IASPEI residuals and the residuals corrected for Great Valley sediments, you might want to first check out the html version of either the demeaned 1-D residuals or the 1-d residuals corrected for Great Valley sediments in order to get some text guidance on how to view these. (Post-inversion residuals are also animated)
Below are images from two recent versions of tomography and Heidi's final thesis tomography (corrected for a 1 s bug that didn't cause too much difference). The tomography Heidi did is *wrong*: the SPE dataset, which was acquired on the east side of the Isabella anomaly in 1997, was accidentlly replaced with vanilla IASPEI times. This created a bunch of odd divots in the inversion.
The other two inversions simply differ in starting model. The "1-D" one starts from a uniform 1-D starting model; there are no station corrections. "GV" has had station corrections applied to the 8 stations in the valley or a partial correction for 2 stations on thrust sheets on the west side. These corrections are based on a simple velocity-depth curve used in the oil industry as a guide to migrate their sections. Both inversions fit the data equally well (there is a tiny numerical edge to the GV inversion), with a 90%+ variance reduction the way Roecker calculates variance and a ~75% reduction when I do the calculations (I will figure out the difference before too long). Final RMS values are about 0.2s, suggesting there is still some signal in the data from structure either too rough or too fine or too deep for this inversion.
The main differences are generally 120 km and above and south of about 38N, though there are some changes in the northern Great Valley from adding sediment delays. I have not included the 0 km slice as that is always a bit iffy, but it usually looks like the 20 km slice but with lower amplitude.
Click on an image for a bigger view.
Alternative is to view these as a single map with buttons to choose depth and model: as a freestanding Flash file (you can resize and save to disk) or within html frame. Also note there is a Flash file for the detail of models covering the area of Josh's ANT (also within an html frame). Animaition with maps of the original Moschetti and Stachnik ANT/surface wave tomos and the inversions derived from those starting models is also available as freestanding Flash or in html frame.
| Depth | GV | 1-D | Reeg MS |
|---|---|---|---|
| 20 km | |
|
|
| 40 km | |
|
|
| 70 km | |
|
|
| 120 km | |
|
|
| 170 km | |
|
|
| 220 km | |
|
|
| 270 km | |
|
|
| 320 km | |
|
|
Animations:
Links to animations of sections through the models:
Great Valley station corrections: WSW-ENE oriented, S to N movie, NNW-SSE west to east movie, E-W from south to north through slab at north
1-d start, no station corrections: WSW-ENE oriented, S to N movie, NNW-SSE west to east movie, E-W from south to north through slab at north
Below is a comparison of Josh Stachnik's ANT (the no Moho version) redone to match the P-wave parameterization (including a conversion from S to P using relations described by Brocher, 1995). Map is the area that Josh's inversion covers, which is a subset of the P-tomo area shown above. Righthand column is inversion of the teleseismic P times from a starting model of the Stachnik ANT for depths of 0-40 km and 1-D IASPEI below that. This inversion produces misfits about 5% higher (in variance) than the Great Valley P tomo or the starting 1-D model shown above.
Also note there is a Flash file for the detail of models covering the area of Josh's ANT (also within an html frame). Click on images below for large version.
| Great Valley P tomo | Stachnik ANT as P | P tomo start Stachnik | |
|---|---|---|---|
| 0 km |  |
 |
 |
| 20 km |  |
 |
 |
| 40 km |  |
 |
 |
| 70 km |  |
 |
 |
| 120 km |  |
 |
 |
There are a number of things to note comparing the Great Valley inversion to the converted ANT. First, as the Great Valley sediments were removed via station corrections on the P model (left column), the absence of the big red blob in the 0 km slice is expected. However, the much lower intensity of the foothills anomaly in the P tomo at that depth probably is reflecting some vertical resolution issue. In the 20 km slice, the big difference is the much smoother and large low wavespeed body under and east of the Sierra. With the exception of the high wavespeeds SE of Walker Lake, though, the boundary between fast and slow is much the same in both models. Whether the high wavespeed body in the P tomo SE of Walker Lake can move down if the top is held constant remains to be seen; this seems a rather fundamental difference. At 40 km,a lot of the difference might come down to the mean wavespeed--the P tomo has a larger footprint. Boundary between fast and slow is again almost identical, and lateral variation of about 8% is quite similar. The high wavespeed thing SE of Walker Lake again differs between the two models; there are additional differences in the north where P tomo suggests continued low wavespeeds and ANT ends low wavespeeds. The P tomo is picking up a larger anomaly in 20 and 40 km depths near Isabella Anomaly than ANT; whether this is a coverage issue with the ANT or a real difference is unclear.
Comparing the inversion that started from the ANT model to the other structures yields some oddities and indicates some artifacts. Because the ANT did not extend somewhat farther south, the low wavespeeds of the southern San Joaquin Valley were left out and, absent any station corrections, arrivals to stations like V05C (on deep sediment just east of 120W and just south of 36N) appeared to be quite late, probably producing the red corner seen at many depths near 120W, 36N and tending to revert the shape of the Isabella anomaly at 120 km to that of the original 1-D solution. The P tomography seems to be compatible with the much larger gradients in the 0 and 20 km depth range, increasing the amplitude of this contrast at 20 and 40 km, but the tele P waves seem to demand high wavespeeds somewhere SE of Walker Lakeabove 70 km but lowering them back down below 70 km. The increased wavespeeds in the Great Valley at 20, 40 and 70 km depth indicates that the anomaly determined from the ANT from 55 km on up is insufficient to fit the P delay times; this quite plausibly indicates that a large anomaly exists near 70 km depth (such a test requires fixing the wavespeeds in the crust, which isn't a great idea for this inversion given the broader aperture, but there is an anomaly in the ANT/surface waves here). The utter collapse of commonality at 120 km depth, save for the Isabella Anomaly, is somewhat baffling. Whether this could be reconciled by starting with the ANT all the way down to 120 km is unclear.
SNEP Main Page | C. H. Jones | CIRES | Dept. of Geological Sciences | Univ. of Colorado at Boulder