NDVI Explained: Formula, Values and Bands for Landsat 8 and Sentinel-2
In brief
- NDVI (Normalised Difference Vegetation Index) measures how green and healthy vegetation is from satellite or drone images.
- Formula: NDVI = (NIR − Red) / (NIR + Red). Values run from −1 to +1.
- Rough guide: water below 0, bare soil and rock about 0–0.2, sparse vegetation 0.2–0.5, dense healthy vegetation 0.6 and above.
- Landsat 8 and 9: Red = Band 4, NIR = Band 5. Sentinel-2: Red = Band 4, NIR = Band 8.
NDVI is the most widely used index in remote sensing. Farmers use it to spot stressed crops, ecologists to map forests and drought, and climate researchers like me to track how landscapes change over years. This guide explains what NDVI stands for, the formula, how to read the values, and how to calculate it from free Landsat and Sentinel-2 imagery.
What does NDVI stand for?
NDVI stands for Normalised Difference Vegetation Index. It works because healthy leaves absorb most red light (for photosynthesis) and strongly reflect near-infrared (NIR) light. The bigger the difference between NIR and red reflectance, the more dense and active the vegetation.
The NDVI formula
NDVI = (NIR − Red) / (NIR + Red). Dividing by the sum “normalises” the result, so it always falls between −1 and +1 regardless of overall brightness. That makes NDVI comparable between dates and places, as long as the images are properly corrected for the atmosphere.
| NDVI range | Typical surface |
|---|---|
| Below 0 | Water, snow, clouds |
| 0 to 0.2 | Bare soil, rock, sand, built-up areas |
| 0.2 to 0.5 | Sparse vegetation, grassland, crops early in the season, stressed plants |
| 0.6 and above | Dense, healthy vegetation such as forests and crops at peak growth |
These ranges are guides, not fixed rules: soil colour, season, sensor and atmospheric correction all shift the numbers. Always compare like with like, for example the same field in the same month across several years.
NDVI formula for Landsat 8 and Sentinel-2
| Satellite | Red band | NIR band | NDVI formula |
|---|---|---|---|
| Landsat 8 / 9 (OLI) | Band 4 | Band 5 | (B5 − B4) / (B5 + B4) |
| Sentinel-2 (MSI) | Band 4 | Band 8 | (B8 − B4) / (B8 + B4) |
| Landsat 5 / 7 | Band 3 | Band 4 | (B4 − B3) / (B4 + B3) |
Use surface reflectance products (Landsat Collection 2 Level-2, Sentinel-2 Level-2A) rather than raw digital numbers, and mask clouds and shadows first. In QGIS, load the two bands and use the Raster Calculator with the formula above. In Google Earth Engine, the normalizedDifference function does it in one line.
What NDVI is used for
Crop monitoring and yield estimates; drought and vegetation-stress mapping; tracking deforestation and regrowth; land cover classification; and climate studies of greening and browning over decades. In mountain regions, NDVI time series help show how vegetation responds to warming and to disasters such as floods. Similar satellite methods are used to monitor glacial lakes; see a sample GLOF exposure brief.
Limits of NDVI
NDVI saturates over very dense forest, so it stops rising even when biomass keeps increasing. Bright soils in sparse vegetation can distort it, which is why indices such as SAVI (soil-adjusted) and EVI (enhanced vegetation index) exist. And NDVI shows greenness, not cause: a drop could mean drought, pests, harvest or clearing, so field knowledge still matters. If you need an NDVI or land-cover analysis for research or a project, see GIS and remote sensing services.
NDVI vs EVI vs SAVI
| Index | Formula idea | Best for |
|---|---|---|
| NDVI | (NIR − Red) / (NIR + Red) | General greenness, crops, drought, change over time |
| EVI | Adds the blue band and correction terms | Dense forests where NDVI saturates; reduces atmospheric effects |
| SAVI | Adds a soil-brightness factor (L, often 0.5) | Sparse vegetation and drylands where bare soil shows through |
| NDWI | Uses green and NIR (or NIR and SWIR) | Open water and vegetation water content |
In practice, many projects calculate NDVI first because it is simple and comparable across decades of Landsat data, then add EVI or SAVI where dense canopy or bright soils make NDVI less reliable.
A simple NDVI workflow
1. Choose your area and dates, ideally the same season in each year you compare. 2. Download surface reflectance images from USGS EarthExplorer (Landsat) or the Copernicus Data Space (Sentinel-2), or work in Google Earth Engine. 3. Mask clouds, cloud shadows and snow using the quality bands. 4. Calculate NDVI with the formula above. 5. Summarise it by field, district or land-cover class, or map the change between years. 6. Check a sample of results against field photos or high-resolution imagery before drawing conclusions.
Frequently asked questions
What is the NDVI formula?
NDVI = (NIR − Red) / (NIR + Red), using near-infrared and red reflectance.
What is a good NDVI value?
Values above about 0.6 usually indicate dense, healthy vegetation; 0.2–0.5 sparse or stressed vegetation; below 0.2 bare ground; below 0 water or snow.
Which bands are used for NDVI in Landsat 8?
Band 5 (near-infrared) and Band 4 (red): NDVI = (B5 − B4) / (B5 + B4).
Which bands are used for NDVI in Sentinel-2?
Band 8 (near-infrared) and Band 4 (red): NDVI = (B8 − B4) / (B8 + B4).
What is the range of NDVI?
From −1 to +1.
Get the climate stories that matter, explained, every Sunday.
One email a week: new explainers, data and free tools. No spam, unsubscribe any time.
More from climaRevolt
Climate Tipping Points: Which Have We Already Crossed?
What Is a GLOF? Glacial Lake Outburst Floods, Explained
The Fight for the High Seas: Inside the Treaty and the Deep-Sea Mining Showdown