In this method, groups of signals are summed coherently, prior to detection.
Each coherent sum is square-law detected to form an output waveform. Each
output waveform is mapped adjacent to its predecessors in sequence, to build
the profile shown in the figure. This method is known in the geophysical
community as "coherent stacking".
Most of the clutter data are suppressed by coherent stacking. In other words, the SCR is improved. There are two substantive disadvantages, however. First, the variance of each output waveform is relatively large. Such coherent noise is known in the synthetic aperture radar community as speckle. In this case, speckle is worse because coherent integration reduces the statistical degrees of freedom of the output. Prior to incoherent averaging, there are two degrees of freedom from each iteration of coherent stacking, in contrast to the 64 that incoherent stacking generates in this example. Thus, simple coherent stacking implies relatively poor signal-to-speckle ratio (SSR). The second disadvantage is that much, if not most, of the sounding signal is discarded by coherent stacking. Thus, the SNR of the profile is diminished.