NESZ Subtraction for Extra wide mode data of Sentinnel-1

Hey all,
I am currently using Sentinel-1 EW (Extra wide) mode data for my studies on polar continents,.
Now, this data has white stripes in cross polarised modes (HV) especially. While I apply S-1 TNR - it does not show produce effective results. Is it something else I can do to overcome these white stripes. Also, I come across few research papers that state this issue and have come up with a solution say on NESZ subtraction, something that says noise vectors to be subtracted for NESZ - Noise equivalent sigma nought subtraction.
I do not find any literature or information on this, can someone please help on how I can perform this on an EW mode s-1 data