Normalized Difference Water Index

I have satellite image (Landsat 8), containing all the bands data.

Using green & NIR band I have calculated the NDWI.

Now I want to classify the Water Index, but don't know the theory behind classification of the water index.

Can anyone help me? What should the classes    be based on?

Taxonomy

• GIS & Remote Sensing Technology
• GIS
• Remote Sensing & Data Analysis
• GIS for network
• GIS & Remote Sensing
• Satellite
• r programming

3 Answers

1. I have used NDWI indexes to quickly identify water bodies quickly and to calculate differences in aerial 

   extents of water. One thing to keep in mind with the published equations is that they do not work in all instances. For areas subject to the introduction of water high in sediment concentration, the Equations may not work. 

2. Please refer this link

    

   https://www.researchgate.net/publication/224184469_Automated_Water_Spread_Mapping_Using_ResourceSat-1_AWiFS_Data_for_Water_Bodies_Information_System

3. Hi,

   depends what you want to do. Generally, NDWI is used to identify Water bodies or vegetation cover.

   Please find here an article with some methodology which could help you :

   https://repository.library.fresnostate.edu/bitstream/handle/10211.3/191777/Using%20the%20Normalized%20Difference%20Water%20Index.pdf?sequence=1

   It also addresses some boundary problems (depending on your workscale, you might have or have not that problem).

   Regards

   Note : using a quick google search might be helpful how to analyze this kind of parameters.

   2 Comments

   1. Maybe that there are no waterbodies in your analysed area ?

      I don't know yout project nor what area you are analysing. Maybe a calculation mistake ?

      Compare the results to the actual satellite image and check how it compares.

      I can't help much more then that from here.

      1 Comment reply

      1. I'm using this formula: NDWI = ((Green-NIR)/(Green+NIR)). 

          

   2. In this paper they are using the assertion based on McFeeters paper that, NDWI values greater than zero are identified as water surfaces. But, what if I have NDWI values in the range of negative values only, added that it contains water bodies. I have Landsat image of Jaipur for 15th May, and calculated NDWI using Green & NIR band. I'm getting the NDWI value ranging from -0.268 to -0.134. What should I interpret from this result then?

      1 Comment reply

      1. Gajanan Kothawade visualize your satellite imagery in FCC. Check if there are any waterbodies visually. Jaipur + Summer could mean relatively less water bodies. Right?