I calculated NDVI (Normalised Difference Vegetation Index) and MSAVI (Modified Soil Adjusted Vegetation Index) for a desert environment, where most of the vegetation is shrubs with some trees using RSToolbox in R.

raster_stack = raster_stack("path/file"
    # Find NDVI
    NDVI = spectralIndices(raster_stack, red = 1, nir = 4, indices = "NDVI")
    # Plot
    plot(NDVI, main = "NDVI")
    # Find MSAVI
    MSAVI = spectralIndices(raster_stack, red = 1, nir = 4, indices = "MSAVI")
    plot(MSAVI , main = "MSAVI")

Which of these two indices are more suitable for such a landscape, when the purpose is to analyze vegetation health and quantify vegetation types, and maybe even compare it with LAI and the Modified Chlorophyll Absorption Reflectance Index (MCARI) indices?


enter image description here


enter image description here

RGB enter image description here

  • 1
    Both are close to the line, yet this seems more about the scientific meaning of underlying algorithms, as applied to the Earth, and less about how they can be deployed via GIS.
    – Vince
    Jul 3, 2021 at 2:16
  • 1
    I think that this is a perfectly valid question for SEGIS, and here's why: I've recently started using satellite imagery and NDVI for vegetative biomass change detection in a semi-arid environment. Without this question I would have never heard of MSAVI, which I now intend to look into it as an alternative. So, the post was a great learning opportunity for me, since I don't follow SE Earth Science.
    – Stu Smith
    Jul 3, 2021 at 2:53
  • 1
    @Arthur_Morgan It would help if you edited your post to include a hi-res satellite image, say from Google Earth, of the same area. Also, to my untrained eye, the NDVI example seems to contain a larger range of color, which in turn makes the area's variability pop-out more. However, that might be a reflection of your color ramp choices, so a review of the histograms would be less opinion-based. Finally, your best option would be to get a plant ecologist with local experience to review your results.
    – Stu Smith
    Jul 3, 2021 at 3:20
  • 1
    @Vince I see this as a good question that falls within our realm. It is clearly a spatial question relating to remote sensing and other relevant topics.
    – Aaron
    Jul 3, 2021 at 3:49
  • 3
    @Arthur_Morgan please edit and put that in your question - unless the goal of your work isn't clear then there's no way we can answer "whats the most suitable?" and putting it here will get it lost in the comments.
    – Spacedman
    Jul 3, 2021 at 7:13


Your Answer

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