We provide a data-driven macro-model-independent stacking technique that mi
grates 2-D prestack multicoverage data into a common-offset (CO) section. W
e call this new process the CO. common-reflection-surface (CRS) stack. It c
an be viewed as the generalization of the zero-offset (ZO) CRS stack, by wh
ich 2D multicoverage data are stacked into a well-simulated ZO section. The
CO CRS stack formula can be tailored to stack P-P, S-S reflections as well
as P-S or S-P converted reflections. We point out some potential applicati
ons of the five kinematic data-derived attributes obtained by the CO CRS st
ack for each stack value. These include (i) the determination of the geomet
rical spreading factor for reflections, which plays an important role in th
e construction of the true-amplitude CO section, and (ii) the separation of
the diffractions from reflection events. As a by-product of formulating th
e CO CRS stack formula, we have also derived a formula to perform a data-dr
iven prestack time migration.