Modelling canopy radiation budget through multiple scattering approximation: A case study of coniferous forest in Mexico City Valley

摘要

In this paper, we describe some results from a study on hyperspectral analysis of coniferous canopy scattering for the purpose of estimating forest biophysical and structural parameters. Georeferenced airborne hyperspectral measurements were taken from a flying helicopter over a coniferous forest dominated by Pinus hartweguii and Abies religiosa within the Federal District Conservation Land in Mexico City. Hyperspectral data was recorded in the optical range from 350 to 2500 nm at 1nm spectral resolution using the FieldSpec 4 (ASD Inc.). Spectral measurements were also carried out in the ground for vegetation and understory components, including leaf, bark, soil and grass. Measurements were then analyzed through a previously developed multiple scattering approximation (MSA) model, which represents above-canopy spectral reflectance through a non-linear combination of pure spectral components (endmembers), as well as through a set of photon recollision probabilities and inter-ceptance fractions. In this paper we provide an expression for the canopy absorptance as the basis for estimating the components of canopy radiation budget using the MSA model. Furthermore, since MSA does not prescribe a priori the endmembers to incorporate in the model, a multiple endmember selection method (MESMSA) was developed and tested. Photon recollision probabilities and interceptance fractions were estimated by fitting the model to airborne spectral reflectance and selected endmembers where then used to estimate the canopy radiation budget at each measured location.