What is the most suitable band or band combination for delineating wetland from upland?
Hi Zhenguo,
Allesandro is right on the money in terms of IR bands, however there are a couple of additional points worth raising:
1. Short wave infra red is better at discriminating open water within wetlands particularly if there are high levels of suspended sediment which tends to push the NIR reflectance up.
2. Depending on what sort of wetlands you're interersted in i.e. lacustrine (lake like) or palustrine (vegetated), you may wish to discriminate wetland vegetation with a higher water content and or more sustained/higher growth rate than surrounding upland vegetation. You can use the methods described in https://profile.usgs.gov/myscience/upload_folder/ci2013Mar1412435621139Drexler%20et%20al.%202004.pdf or http://www.kimberly.uidaho.edu/water/papers/evapotranspiration/Allen_Predicting_ET_demands_for_wetlands_ASCE_1998.pdf to characterise wetland ET, or simply use a sensor that has a thermal band and use a more simple cross plot based approach to identify your wetlands.
3. Lastly, a word of caution, delineating wetlands is a tricky business. Many wetlands fluctuate in extent in response to season, inter-seasonal and inter-annual water and energy availability. I would strongly recommend a multi-temporal approach to wetland delineation, so that you can capture features such as maximum and minimum wetland and/or innundated extent. Using a very deep multi-temporal series may also allow characterisation of wetland connectivity, which may or may not be important in your setting.
Good luck!
thanks a lot! Alessandro. Yes, it does work theoretically. but the vegetation in wetland could limit its application in most cases. so I wonder whether the water content of wetland vegetation was higher than that of upland. were there any field experiments to test it?
Hi Zhenguo,
Allesandro is right on the money in terms of IR bands, however there are a couple of additional points worth raising:
1. Short wave infra red is better at discriminating open water within wetlands particularly if there are high levels of suspended sediment which tends to push the NIR reflectance up.
2. Depending on what sort of wetlands you're interersted in i.e. lacustrine (lake like) or palustrine (vegetated), you may wish to discriminate wetland vegetation with a higher water content and or more sustained/higher growth rate than surrounding upland vegetation. You can use the methods described in https://profile.usgs.gov/myscience/upload_folder/ci2013Mar1412435621139Drexler%20et%20al.%202004.pdf or http://www.kimberly.uidaho.edu/water/papers/evapotranspiration/Allen_Predicting_ET_demands_for_wetlands_ASCE_1998.pdf to characterise wetland ET, or simply use a sensor that has a thermal band and use a more simple cross plot based approach to identify your wetlands.
3. Lastly, a word of caution, delineating wetlands is a tricky business. Many wetlands fluctuate in extent in response to season, inter-seasonal and inter-annual water and energy availability. I would strongly recommend a multi-temporal approach to wetland delineation, so that you can capture features such as maximum and minimum wetland and/or innundated extent. Using a very deep multi-temporal series may also allow characterisation of wetland connectivity, which may or may not be important in your setting.
Good luck!
thanks again, Alenssandro. for your good suggestions. also thank Leo for your good advice. I'd like to try these approaches.
thank Ali, but I don't understand what OIF or IF is.
- Badges
- Science topic
- Similar topics
- Geoscience
- Remote Sensing