Rouse et al. (1974) NDVI
Original technical source for the normalized difference vegetation concept.
Learn more: Rouse et al. (1974) NDVISpectral indices guide
NDVI vs NDWI vs NBR: which index should you use?
Choose the index based on the phenomenon you need to observe, not on which acronym is most familiar.
NDVI: vegetation greenness and vigor
NDVI contrasts near-infrared and red reflectance. It is commonly used to observe vegetation presence, relative vigor and change, but should not be treated as a universal measure of crop health.
NDWI: water detection
NDWI variants emphasize water or vegetation water content. Confirm the formula and bands used before comparing results from different sources.
NBR: burn severity and disturbance
NBR contrasts near-infrared and shortwave infrared and is often compared across dates to support burn-severity and disturbance analysis.
Practical rule
- Use NDVI for vegetation patterns.
- Use the appropriate NDWI formulation for open water or moisture questions.
- Use NBR and change in NBR for fire and disturbance analysis.
- Always confirm sensor, date, cloud, resolution and preprocessing consistency.
The real bottleneck
The formula is usually the easy part. Reliable analysis depends on aligned inputs, suitable dates, cloud handling, consistent preprocessing and interpretation in context.
Formula and bands
NDVI = (NIR − Red)/(NIR + Red); it contrasts near-infrared and red reflectance. A common open-water NDWI uses (Green − NIR)/(Green + NIR), while Gao’s vegetation-water formulation uses (NIR − SWIR)/(NIR + SWIR). NBR = (NIR − SWIR2)/(NIR + SWIR2). Confirm the formulation and sensor band mapping before comparing values.
Interpretation controls
- Clouds, shadows and atmospheric correction affect indices
- Acquisition date, seasonality, resolution and sensor consistency matter
- A single date is not change analysis
- Thresholds are project-specific and should not be copied blindly
- Field data and domain expertise are needed to test what a pattern means
A practical choice
Start with the phenomenon, then confirm bands, preprocessing and dates. BGPD-style multitemporal work uses repeated observations and relative behavior rather than treating one index value as an absolute diagnosis.
Sources and further reading
USGS Spectral Indices
Official Landsat spectral-index documentation.
Learn more: USGS Spectral Indices