I do not know if there is an absolute method for this. In the book by Charlesworth and Temple there is an image (Figure 5-8) that shows all the arcs for the Compression, Shear and Rayleigh modes and indicates the top and bottom signals for a single planar flaw. This image is a VERY confusing "mess" to look at.
The authors state that "...it may be necessary to make a judgement as to which are the unconverted compression wave signals from the top and bottom of the defect and then calculate the positions and curve-shapes of all possible mode converted signals."