First, kx will range over .If the x-axis is going to be handled by finite differencing then we will need equation (39):
So if the x-axis is going to be handled by finite differencing then subsequent reference to kx should be replaced by .The finite differencing introduces the free parameter b. Likewise, you could also scale the whole expression by an adjustable parameter near unity. Also, isn't necessarily fixed by the data collection. You could always interpolate the data before processing. A finite-difference method using interpolated data could be mandated by enough lateral velocity variation.