Perhaps you are not as impressed as I am with the information content of the CO2 changes with time. Let me try to impress you with the next piece of the pattern that connects this global story with the landcover we are familiar with in our daily lives.

I earlier claimed that the annual climate-driven change in landcover activity is so significant that it can be documented by space observations. Just as we can indirectly observe the biosphere breathing by measuring the atmospheric CO2 concentration, so can we also directly observe the time and space patterns of landcover activity using satellites. All that is necessary is a method of extracting landcover information from satellite observations of the sunlight reflected from the Earth's surface.

To tackle this question we require a measure of greenness of the landcover. This measure can be derived from the contrasting reflection by vegetation of radiation in the visible red and near infrared wavebands.

As discussed in a previous chapter green, (to our eyes) vigorous and actively growing vegetation reflects very little visible red radiation because of the high absorption of this waveband by the green chlorophyll. Conversely these same green (to our eyes) plant canopies show a very high reflection in the near infrared wavebands when the leaves are new and rapidly growing. This spectral contrast was apparent in the images examined earlier: green vegetation was dark in MSS#2 (red) but bright in MSS#6 and MSS#7.

We can devise a very simple index of vegetation vigour or greenness thus: the darker it is in the red wavebands AND the brighter it is the near infrared wavebands, the greener is the landcover. This greenness index can be written very simply as the ratio NIR/RED. That is, a simple ratio of the brightness in the two bands. This ratio is computed and used as an index of greenness and is called (not surprisingly) the Simple Vegetation Index (VI). However, computing ratios of numbers has its dangers, particularly when the denominator becomes small. Therefore this simple vegetation greenness index is "normalised" to make it less non-linear and more robust by rewriting it thus:

(NIR-RED)/(NIR+RED) = NDVI

This index, the Normalised Difference Vegetation Index or NDVI is the most widely used vegetation index in all satellite data analysis.

The NDVI is relevant to this book because it is a very robust index of landcover greenness. This index has been correlated with many vegetation canopy variables, cover and biomass, as well as with measures of processes such as primary production.