ALOS is L-band, Sentinel-1 is C-band. Both sensors have different interactions with the surfaces presenti in your study area. None of them is correct, nor false. But generally, L-band is more resistent to temporal decorrelation. As you see in your screenshots, coherence is not ideal in your image. The higher the coherence (and at best over large areas) the btter is the quality of the phase information to be expected. But there can still be atmospheric artifacts in the data.
The elevation you show in the screenshots (bottom right) is probably the one SNAP outputs from the Terrain Correction step. It is important to distinquish the one created as result of the InSAR DEM process and the one automatically downloaded as ancillary height information. Please have a look here: Using VV and VH to make DEM - #2 by ABraun