I have composed a mosaics of 3 images of Landsat 8 Surface Reflectance Level 2 for the bands NIR and RED. According to the Landsat Surface Reflectance Product Guide each band is supposed to be multiplied pixel wise times a scaling factor of 0.0001. As I have understood, it might occur that after such multiplication the obtained surface reflectance values may slightly surpass/exceed the commonly known upper and lower bound of the reflectance range going from 0 to 1. Then, I calculated NDVI and the values do not strictly go from -1 to 1 but from -0.5 to 0.98.

Could this have to do with the fact that the input surface reflectance bands (NIR and RED) slightly surpass the 0 to 1 values expected for reflectance?

Does NDVI always range from -1 to 1 or it can also range within the values -1 to 1?


1 Answer 1


If you first multiply every layer by 0.0001 and then calculate NDVI the surpassing shouldn't happen, as you are dividing a subtraction of two positive numbers (band values) by the addition of the same two numbers; this is because it is a normalized index, which mathematically means that it will range from -1 to 1 in this case; as you state, your values range from -0.5 to 0.98, this lays withing the range; surpassing the range would be, e.g. -1.02 to 1.04


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.