Hello, this is more a question about the interaction between microwaves and water surfaces than strictly about S1Tbx. I am trying to create a map of the flooding that occurred in West Germany in July. I went through some papers and posts here about what is the best polarization to detect flooded areas but it’s still not very clear. According to the literature, it seems that HH is the best polarization because water bodies reflect best horizontal polarized radiation and that property should result in lower backscatter over flooded area, because of the specular reflection, so they should appear darker with respect to not flooded areas. On the other hand, Sentinel-1A, provides only VV and VH polarization for the orbit that covers my area of interest. My question is:
Is there any real difference between HH and VV (or VH) polarizations for the task at hand (flood mapping) ?
We had similar discussions on this in these threads
- VV and VH Polarization
- Water threslold
- Difference between VH and VV polarization
- Use of VV & VH polarisations for Water Detection
Generally speaking, VV is more sentitive to vertical structures and HH to horizontal structures. This can make a difference if the water bodies are roughed by wind, have systematic waves or if there is flooded vegetation. But if VV is all you have (with Sentinel-1) you can give it a try.
Good points from @ABraun. I’d also add that your consideration of VV vs. VH might depend on a few factors related to the landcover over which you’re attempting to detect floods. The VV channel backscatter variability tends to be more sensitive to water surface roughness than the VH channel, and often the HH channel. Meaning the backscatter decrease one might expect from flooding may not always be present as rough flood waters may produce higher backscatter than a number of natural and man made bare surfaces for the VV channel. The VV channel backscatter may also be increased by a double bounce effect if emergent wetlands or crops or buildings are being flooded (not as much as HH, but non-trivially at times). In contrast, the VH channel should produce a more consistent decrease in backscatter across different landcover types during flooding as VH tends to produce very low backscatter returns over water while being less sensitive to surface roughness and exhibiting very little double bounce scattering. In summary, it’s well worth giving the VH channel a look if you have the imagery, and perhaps both the VV and VH channels as they may be serve as complements to each other rather than an either-or decision.