Processing Landsat 5-9 to L2 resulted in striping, restricted to mountainous areas. I suspect this is related to the manner of interpolation used by FORCE when reprojecting the DEM for the topographic correction. Such striping is common in case of bilinear interpolation, but is usually solved by cubic interpolation. Does FORCE use the former, or is there any other reason for such behavior?

level2_coreg_L07.prm.txt

The DEM I used is Copernicus 30m DEM. I also tested it with DEM reprojected to the datacube projection. This made the stripes slighty less visible, but they are still there.