S1 Coherence - Agricultural Zone

Microwave Toolbox
#1

Hi to all of you!

I am working with coherence maps over agricultural zone in Argentina, for crop classification.

I carried out two coherence maps, but one of this have a low coherence. Both are six (6 days) temporal baseline. Considering that spatial and temporal baseline are short, why second coherence maps have low coherence?? Can I help me?

Coh_20200411-20200417
Coh_20200411-202004172438×1672 3.13 MB

Coh_20200312-20200318
Coh_20200312-202003181219×836 909 KB

I see two hypothesis:

  1. MAST and SLV Coregistration problems.

  2. Original SAR Data problems

Have someone similar problem?

Thanks’ a lots.

SBST

#2

some ideas:

  1. have you selected “remove topographic phase” for both maps? I found that it makes a difference for coherence as well.
  2. if there was heavy rain between the first and second image of the second coherence map, crops might have bent which leads to drastic decorrelation. As soon as the main scattering mechanism is altered (e.g. volume scattering to surface scattering) coherence decreases significantly.
  3. please also have a look at these considerations: Unexplained differences between different orbits
  4. harvesting of crops also leads to decorrelation, of couse, but as nearly the entire image has decreased coherence I think it is rather a rain event.
  5. bad coregistration quality also leads to low coherence from the computational side
#3

Hi ABraun! Thank you for your response.

  1. We have removed topographic phase for all maps.
  2. We found low coherence in full image for the map (coh_IW3_VV_23Apr2020_29Apr2020). But the same does not happen with:
  • coh_IW3_VV_17Apr2020_23Apr2020
  • coh_IW3_VV_29Apr2020_5May2020
    Problem occur between 23Apr2020 and 29Apr2020. For example:

coh_IW2_VV_17Apr2020_23Apr2020:
coh_IW3_VV_17Apr2020_23Apr2020_resize

coh_IW2_VV_23Apr2020_29Apr2020:
coh_IW3_VV_23Apr2020_29Apr2020_resize

We consider that when the ground remains bare between both dates, coherence should see high. We investigate the Mission Status (https://sentinel.esa.int/web/sentinel/missions/sentinel-1/mission-status) for this dates, but it is all ok.

Our hypothesis are orbital error or bad focalization in acquisition 29Apr2020. Because the behavior is the same in the three swaths (IW1, IW2 and IW3). For this reason we discard possible heavy rain or weather effects.

Are you agree? Thanks a lots

SBST

#4

have you tried how the coherence looks without the orbit files applied. Sometimes errors are introduced if orbit data is (very rarely) of faulty quality.