NDSI: Telling Snow Apart From Cloud and Rock
Snow is bright in visible light but, unlike cloud, nearly black in shortwave-infrared. NDSI turns that one distinguishing trait into the standard test for mapping snow and ice.
How it works
Snow's high green and very low SWIR1 push it well above a ~0.4 threshold, while cloud — bright in both — stays low, which is how NDSI separates the two. It is the basis of the MODIS global snow-cover product.
The math behind it
where Green is green reflectance, SWIR1 is shortwave-infrared 1 reflectance (~1.6 µm).
How to interpret the output
> 0.4 is typically snow / ice (with NIR and thermal checks) · lower values are snow-free.
The bands it needs
| Sentinel-2 | Green B3 · SWIR1 B11 |
| Landsat 8/9 | Green B3 · SWIR1 B6 |
Real-world use
It drives snow-cover mapping for hydrology and water-supply forecasting, glacier monitoring, and winter land-cover work.
Watch out for
NDSI shares its exact band math with MNDWI, so water bodies also score high — you need an NDVI or thermal check to exclude lakes, and thin cloud or shadow can still confuse the 0.4 cut. It also struggles with snow under forest canopy.
Compute NDSI on your own study area
Skip the code. Draw or upload a boundary and Spatial Research Suite runs NDSI on live Sentinel-2 or Landsat imagery — with cloud masking, exports and citations built in.
Run this analysis in GISforus →Frequently asked
What NDSI value indicates snow?
Values above about 0.4 typically indicate snow or ice, usually combined with NIR and thermal checks to exclude water and cloud.
Why does NDSI confuse snow with water?
It uses the same green and SWIR1 bands as MNDWI, so water also scores high; an NDVI or thermal test separates them.