NDVI (Normalized Difference Vegetation Index), introduced by Rouse and colleagues for NASA in 1973, is the most established and most widely used vegetation index for measuring plant cover from remote sensing. It is the standard tool in agriculture, forestry, ecology and landscape monitoring.

Formula and physical basis

NDVI = (NIR − Red) / (NIR + Red) — the value range is between −1 and +1.

Chlorophyll strongly absorbs visible red light (~660 nm); the spongy mesophyll inside the leaf strongly reflects near-infrared (~840 nm). The reflectance gap between these two bands is proportional to plant health. In a healthy leaf the gap widens and NDVI rises. When a plant suffers water stress, disease or chlorophyll loss, it absorbs less red and reflects less NIR; NDVI falls.

Data sources

For NDVI computation, the most common open source is the European Space Agency's (ESA) Copernicus programme satellites Sentinel-2A (2015) and Sentinel-2B (2017). The pair offers a 5-day revisit, 10 m spatial resolution and 13 spectral bands. NDVI uses Sentinel-2's B4 (Red, 665 nm) and B8 (NIR, 842 nm) bands.

Other common sources: Landsat-8/9 (30 m, 16 days), MODIS (250 m, daily), Planet (3 m, daily, commercial), and unmanned-aerial-vehicle (UAV) multispectral cameras (sub-metre, on demand). Resolution, revisit cadence and cost are chosen to match the application.

Value ranges and interpretation

  • −1 – 0 — water surface, snow, cloud.
  • 0 – 0.1 — bare soil, asphalt, concrete, dry rock.
  • 0.1 – 0.3 — sparse herbaceous cover, stressed or drying turf, stubble.
  • 0.3 – 0.6 — moderately dense green; park turf, young crops, shrub.
  • 0.6 – 0.9 — dense, healthy cover; well-maintained turf, mature deciduous canopy.
  • Above 0.9 — rare in practice; usually a sign of calibration error or sensor saturation.

Time-series analysis

The real value of NDVI emerges not from a single image but as a time series. A typical seasonal curve for healthy perennial cover: rapid rise in spring, plateau in early summer, slight decline in late summer, gentle decrease in autumn. A meaningful departure from this curve — an unexpected drop, a lost plateau or delayed green-up — is a stress signal. Thresholding a moving average (e.g. three-week) is a common way to issue an alert before visible symptoms appear.

Early-warning value: in field studies the NDVI decline typically begins 7–10 days before visible stress symptoms. A drop in NDVI alone, however, does not name the cause: water stress, fungal/insect disease, fertiliser deficiency or physiological disorder can all produce a similar signal. Diagnosis is usually confirmed by cross-referencing ground measurements (soil moisture, leaf temperature, visual inspection).

Limitations and alternative indices

Cloud and haze obscure NDVI; the proportion of cloud-free imagery varies between 30–70% depending on region and season. Snow cover suppresses the signal in winter. Atmospheric distortions (aerosol, water vapour, cirrus) require atmospheric correction.

In young or sparse cover, the underlying soil makes NDVI read lower than it should; the soil-corrected SAVI (Soil-Adjusted Vegetation Index) is preferred. In dense, saturated canopies NDVI saturates; EVI (Enhanced Vegetation Index) is more sensitive in that regime. For analyses focused on plant water content, NDWI (Normalized Difference Water Index) is used. For direct chlorophyll-density measurement, red-edge indices (e.g. CIred-edge using Sentinel-2's B5–B7 bands) are common.