The value of Modified Normalized Difference Water Index (MNDWI) often varies between -1 to 1. It uses green and SWIR (Short Wave Infra-Red) bands for the enhancement of open water features. Thus could help to diminish built-up area features that are often correlated with open water in other indices.

Hi, first of all you need to pass the values od DN to values of surface reflectance.

For do that you need some formulas and that formulas depends on the landsat satellite you are using. When you have surface reflectance values you applied the formula of Modified Normalized Difference Water Index (MNDWI). It's something like Green-MIR/Green+MIR (MIR = Middle Infrared) Green and MIR refer to some bands of your satellite image, and it depends on which satellite landsat you are using.

Good luck.

Dear dulan. My suggestion to you is to again calculate your index. Actually it is not about the method. Its about your calculation. Even if you just use raw DN values, and apply a correct calculation method, you should get values between -1 and +1. My suggestion for you is to use raster calculator in ArcGis and use parenthesis. For example (band 5 - band 4)/(band 5 + band 4). It would be totally wrong if you want to rearrange and reclassify your values.

Nima Karimi

Dear Nima, I used surface reflectance data (Collection-1 Level-2) of Landsat 5. So i applied MNDWI algorithm using raster calculation which is Float("Green"-"SWIR2")/Float("Green"+"SWIR2"). But DN value range changed from large negative value to large positive value. Ex: (-217 to +223), (-111 to +67). But i used Collection-1 Level-1 ( Landsat 5) data, and applied above algorithm got value from -1 to +1. That's fine, but i can't used collection-1 level-1 data because of it has surface reflectance + atmospheric reflectance.

Thanks for your response.

Adrian Jacinto

Dear Adrian, I used Collection-1 Level-2 data which is mention surface reflectance Landsat data. Is collection-1 level-2 data sets need to convert surface reflectance again?

Thanks for your response.