The left panel of Figure shows the synthetic model with 7 reflectors overlapped. The challenge for this model is to determine the low velocity zone under the top of the anticline and to image the bottom of the model. The initial velocity I use is a linear v(z) model, as shown in the right panel of Figure . Figure shows the migration result using the initial velocity model. We can see the upper part of the model is well imaged, but the lower part is severely mispositioned. The left three panels of Figure shows three angle domain common image gathers (ADCIG) computed using the initial velocity model. The X positions are 8.0, 10 and 12.0 km respectively. Notice the smiling events in the ADCIGs indicating the initial velocity value is too slow at these locations.