Geometric, optical and radiative transfer models have potential application in describing pre & post fire affected vegetation. They use data describing the vegetation and the substrate as the input to the models and provide optical spectra as the output. They can be inverted to attempt to characterise changes in vegetation structure, leaf structure, leaf optical properties, the ground layer, the soil type, and more.

Here we present a methodology for ground data collection, in association with a unique method for collecting hand held spectrometric data, as input and output, respectively, to a bottom-up model (BUM) for discriminating fire affects on vegetation.

The outcome is to provide optimal algorithms, regardless of sensor, to discriminate the severity with which fire affects vegetation in the near 2 million square kilometres of tropical savanna woodland in monsoonally effected northern Australia.