Biophysical parameter retrieval is an approach in remote sensing that aims to estimate parameters which have physical meaning related to properties of living organisms. The goal is to provide quantitative results directly relating to the biophysical state, but independent of acquisition conditions and technology. Assessment of vegetation status is a key motivation for this, because through plant respiration and photosynthesis, vegetation is critical for modelling terrestrial ecosystems and energy cycles in environmental studies.
Important parameters describing canopy structure include leaf area index (LAI), green cover fraction (fCover), fraction of absorbed photosynthetically active radiation (fAPAR), plant height, biomass and leaf angle distribution. At leaf biochemical level, leaf chlorophyll/water, fuel moisture and leaf pigmentation content are used.
Visual inspection can provide a first assessment of plant status. For detailed measurements of biophysical parameters, mostly destructive methods have been used. Chemical measurement techniques on leaf samples can measure pigment concentrations very accurately, but are time consuming and only use very limited samples.
Much more extensive data can be collected using earth observation imagery. These range from large scale spaceborne observations with high frequency at coarse resolution to dedicated UAV flights which can offer spectral information of individual plants. Radar and LiDAR acquisitions, which are insensitive to weather conditions, now complement optical observations.
Methods to retrieve the parameters from remote sensing data fall into two main categories. Statistical models empirically match data to a biophysical variable. Univariate techniques use a single quantity derived from the data, usually a vegetation index whereas multivariate techniques link a combination of measurements at different wavelengths to one or more biophysical parameters.
Physically-based modeling is an alternative approach which uses advanced radiative transfer models to describe the transfer and interaction of radiation inside a leaf or canopy based on robust physical, chemical, and biological processes. They compute the interaction between solar radiation and plants and provide as such a better understanding between biophysical variables and reflectance characteristics. Good examples are Leaf optical models such as PROSPECT and LIBERTY which simulate leaf optical properties by absorption and scattering coefficients. Canopy reflectance models simulate canopy reflectance as a function of a complex description of plant structural and radiometric attributes to develop a quantitative understanding of remote sensing information.