Normalising Sentinel-1 Ascending and Descending Pass Values

Hi Team,
I’m trying to carry out a time-series analysis on the basis of Beta0 coefficient in Sentinel-1.

My area of interest has a revisit frequency of 12 days (Only S1A is available) and hence I decided to use both Ascending and Descending passes of S1A. But the values I get from these passes keep varying abruptly and hence I cannot draw any conclusion out of the combined time-series.

Is there any way through which I can get Beta0 values by making use of both Ascending and Descending passes?

Here are the preprocessing steps that I have followed and parameters that I have used in SNAP:

  1. Apply Orbit File
  2. Thermal Noise Removal
  3. Calibration => Create beta0 virtual band
  4. Speckle Filter => Median (3x3 Window)
  5. Terrain Correction =>
  6. DEM => SRTM 3sec
  7. DEM Resampling Method => Nearest Neighbour
  8. Image Resampling Method => Nearest Neighbour

Please let us know if these are the correct steps to get the Beta0 values.

Images Used (All images are S1A_IW_GRDH_1SDV):

  1. subset_20210601_Orb_NR_Cal_Spk_TC - Ascending Pass
  2. subset_20210609_Orb_NR_Cal_Spk_TC - Descending Pass
  3. subset_20210613_Orb_NR_Cal_Spk_TC - Ascending Pass
  4. subset_20210621_Orb_NR_Cal_Spk_TC - Descending Pass
  5. subset_20210625_Orb_NR_Cal_Spk_TC - Ascending Pass


Are you using any time series algorithm such as StaMPS?

I have simply used both passes and followed the preprocessing steps mentioned in the original post.
Any random pixel is giving me a zig-zag kind of pattern indicating that I can relate one Ascending date with other Ascending dates but not with the Descending dates and vice-versa.

If you want to combine ascending and descending pass data to reduce radiometric effects caused by topography you should test the new function: Spatial reference lost after applying multitemporal compositing - #12 by ABraun

Regardless of that, I recommend the 1Sec SRTM and bilinear resampling.

Radiometric terrain flattening is the step that minimises the effect of the viewing geometry.