13.4 Vegetation Indicator Modeling and Estimation

Learning Guide

Remotely-sensed imagery contains much information on the type, amount, and health of vegetation and other properties of a site. Recall that different surfaces (including different species of plants or plants in different conditions) reflect different amounts of light across the electro-magnetic spectrum.

The different amounts of light reflected by various surfaces across the EMS can be used to predict the amount or type of vegetation.


In image classification, we assign pixels to unique categories based on these spectral differences. But sometimes we want to know about the amount of vegetation, and not just what class a pixel belongs to. In these cases, vegetation indexes or biophysical vegetation models may be useful.

Vegetation Indexes

A vegetation index is a combination of two or more remote sensing image bands designed to highlight or accentuate information about the type or amount of vegetation present. There are many different vegetation indexes.


The most common and well-known vegetation index is the Normalized Difference Vegetation Index (NDVI). The NDVI is calculated as:

where NIR is the near-infrared band, and Red is the red band of a satellite or multispectral aerial image. NDVI takes advantage of the fact that photosynthetically-active plants absorb red light but reflect near-infrared light. Thus, high NDVI values indicate high photosynthetic activity. Low NDVI values can indicate a lack of vegetation, or vegetation that is dead or dormant.


NDVI is related to the amount of photosynthetically active vegetation. In this example of mean NDVI in a portion of the University of Idaho’s Rock Creek Ranch in southern Idaho, you can see higher NDVI values for riparian areas than for surrounding sagebrush uplands. Images from climateengine.org


NDVI is strongly correlated to many different vegetation indicators including cover, density, and biomass. However, it is important to remember that a vegetation index does not directly measure any of these indicators! A vegetation index is nothing more than a ratio of the amount of light reflected in two or more wavelengths!


Because of their correlation with vegetation indicators, vegetation index values are commonly used in monitoring applications and are an input to many other remote sensing techniques. In some cases, NDVI values can be used directly as monitoring indicators.


Darndel et al. (2014) used a time series of NDVI to measure the “green-up” of the Sahel region of Africa over a 30-year time span. Image from Darndel et al. (2014), http://dx.doi.org/10.1016/j.rse.2013.09.011


Biophysical Vegetation Models

In some cases, though, we need or want actual predictions of vegetation indicators, and cannot rely on an indirect indicator like NDVI. In these cases, a statistical model can be constructed the relates vegetation indicator values to remote sensing image data. There are many techniques for doing this, and a discussion of these is beyond the scope of this module. Some examples, however, are below.


Jones et al. (2018) mapped vegetation cover indicators across the western USA from Landsat data using field data collected by the BLM’s Assessment, Inventory, and Monitoring program and the NRCS’s National Resources Inventory. Jones et al. applied their models retrospectively to map vegetation cover by year from 1984 to present.


Homer et al. (2012) used a regression tree method to predict vegetation cover and bare ground over the state of Wyoming using Landsat imagery. Image from http://dx.doi.org/10.1016/j.jag.2011.09.012


Jansen et al. (2017) predicted grassland biomass throughout the growing season using vegetation indexes derived from Landsat TM satellite imagery. Image from Jansen et al. (2017), https://www.mdpi.com/2072-4292/10/7/1057/htm