Tomographic algorithms usually have trouble resolving rapid variations in the near surface velocity field. This is why most software packages allow users to include some sort of high-frequency fudge factor when computing statics.
This walk through will show how to update the velocity field using shot/station stacks, thus mitigating or eliminating the need for the high frequency terms.
Here are the basic steps:
- Generate a stack. This creates one stacked trace for each live shot and station. Note that only the travel times through the tomography model are applied, not the high frequency fudge factors.
- Pick the stack. There are some automated tools to do this (threshhold, nearest time, etc) or the user may pick the stacks.
- Use the picks to update the velocity field.
The above steps will likely be done several times until the stacks are flat.
Open a tomography model that has already been updated.

Create a stack
Select the “Shot/station stacks” tab and then click on the green “+” in the upper-center of the display. This will open the new stack creation dialog:

Give the new stack a name. Don’t worry too much about the options selected for the first set of stacks, they will certainly be modified during later stack runs. In particular, selecting the best offset range is something of an art. It may be important to apply RMS trace balancing. For very large surveys it may be wise to decimate the trace count. Click “OK” to start the stack job.
The stack will run in a separate process. A dialog will pop open informing you the job is done.
Pick the stacks
Be sure the newly created stack is selected in the “Stack selection” combo. There is only one Branch for tomography.
Aside – if you’re unhappy with the current stacks you should experiment with different stacking options. The current stacks will not be overridden as long as you be sure to change the stack name.
You may need to experiment with different grouping options on the left. By default the shots/stations are grouped by column, but for some surveys it may be more appropriate to group by count.
Go ahead and pick the stacks, it’s just like picking first arrivals.
You can flip polarity if needed.
It’s worth experimenting with the various auto picking options. They can work well on relatively clean data.
It’s highly recommended to color the basemap on the right using the picks. This can help avoid leg jumping and other inconsistencies.
The following screenshot shows a receiver line. The very large station to station stack variation is caused by the shallowest layers not being correctly described by the tomography inversion.

Update the velocity field
Once both shots and stations have been picked it’s time to update the velocity field.
Select the up-arrow in the upper-left tab display. Note that Phoenix forces you to create a new tomography model, it will not permit the current velocity to be overridden.

The amount of update is controlled by the “Update percent” option. It will vary from a large amount at the surface tapering down to close to zero at the maximum depth.
Be sure to select the correct pick version! For example, if you’re using threshhold auto picks be sure to select that option.
After selecting the options click “Update velocity – create new tomography model”. Then follow the directions in the dialog and open the new model.
Comparing the two models
At this point you should have both models open. In the “Displays” button select “Profile comparison display” (NOT “Horizontal profile display”!). Left-click and drag a profile line on the basemap, then select the correct models. You should see something like this:

Create new stacks using the new tomography model
Repeat the above steps until the stacks are relatively flat.
Keep in mind that Phoenix creates a new tomography model each time you go through the above procedure. This make it easy to go back to an earlier model and try different options, such as stacking offset ranges.
Comparing statics
It’s important to understand how much the above procedure changes the velocity field. One simple way to do this is to compare statics fields computed WITHOUT the fudge factors.
Compute statics for both models using the same options. Be sure that “Apply residual term” is unchecked.

Next select the “Column difference” tab in one of the model windows. Select the appropriate options on the left.
The plot show the difference in vertical travel times through the models.
