Leaves create a spectral contrast
NDVI compares reflectance in a red band with reflectance in a near-infrared band. Chlorophyll absorbs much of the visible light reaching a leaf, while its internal structure scatters near-infrared light strongly. The index subtracts red reflectance from near-infrared reflectance and divides by their sum. A canopy with a strong contrast therefore receives a high index. This is a dimensionless description of its optical behavior. It is not a measurement of grams of new wood, kilograms of harvested grain or carbon absorbed during the satellite pass.
Sources: ndvi-light
More leaves do not produce equal index increments
The ratio does not increase in a simple, unlimited proportion to vegetation. In dense canopies, NDVI becomes less sensitive to additional foliage, a behavior called saturation. Two forests can consequently have similar NDVI despite differences in their structure or leaf abundance. NASA describes the enhanced vegetation index, EVI, as retaining sensitivity under dense vegetation conditions more effectively. The distinction matters when a map looks uniformly green: a narrow range of index values does not establish that every stand has the same amount of vegetation.
Sources: ndvi-light
Exposed soil contributes to the same pixel
A sparse canopy leaves much of the background visible. Soil brightness then influences the red and near-infrared signals used in NDVI, even though soil is not photosynthesizing. The USGS Soil Adjusted Vegetation Index introduces a correction term to reduce this influence in areas with low vegetation cover. That explains why an index comparison across a patchy field and a closed canopy is harder than comparing two equivalent canopies. The calculation receives reflected light from the observed surface as a whole, including the spaces between plants.
Sources: ndvi-soil
A composite represents a selection of observations
The MODIS vegetation-index product described by NASA combines observations across a 16-day period. Its algorithm excludes poor-quality pixels and then selects a representative observation using high NDVI and viewing-angle criteria. It is therefore not simply a photograph taken on the date printed at the start of the interval, nor an arithmetic average of every daily value. A seasonal curve made from these products tracks canopy greenness through successive observation windows. The compositing rule helps explain why a short event can be less obvious than a sustained seasonal change.
Sources: ndvi-composite
Carbon uptake needs light and conversion efficiency
The MOD17 productivity model illustrates the extra reasoning needed to estimate photosynthesis. It combines absorbed photosynthetically active radiation with an efficiency that depends on vegetation type and environmental stress. Inputs include leaf-area and absorbed-light estimates, sunlight, temperature and atmospheric vapor-pressure deficit. Two similarly green canopies can therefore have different modeled carbon uptake because their available energy or conversion efficiency differs. These estimates depend on a model; attaching carbon units to NDVI cannot produce them. Greenness provides evidence about the canopy; productivity requires its functioning as well.
Sources: ndvi-production
Net growth subtracts the plant’s own carbon costs
Gross primary production describes photosynthetic carbon fixation. Net primary production accounts for carbon used in plant respiration. MOD17 explicitly estimates maintenance respiration in leaves, fine roots and living woody tissue, together with growth respiration, before calculating annual NPP. A rise in gross uptake therefore need not translate into an equal rise in net production. For a claim that vegetation is growing faster, the relevant comparison is a production estimate over a stated period, with its respiration assumptions, rather than a greener color or higher NDVI alone.
Sources: ndvi-production