Bayesian estimation of seasonal course of canopy leaf area index from hyperspectral satellite data. (March 2018)

• Record Type: Journal Article
• Authors: Varvia, Petri; Rautiainen, Miina; Seppänen, Aku
• Is Part Of: Journal of quantitative spectroscopy & radiative transfer. Volume 208(2018)
• Page Start: 19
• Page End: 28
• Publication Date: 2018-03
• Subjects: Leaf area index -- Spectral invariants -- Reflectance model -- Uncertainty quantification -- Seasonal dynamics; Spectrum analysis -- Periodicals; Radiation -- Periodicals; Analyse spectrale -- Périodiques; Rayonnement -- Périodiques; Radiation; Spectrum analysis; Periodicals; 543.0858
• DOI: 10.1016/j.jqsrt.2018.01.008
• Languages: English
• ISSNs: 0022-4073

Abstract

Highlights: Bayesian inversion of a forest reflectance model is done using EO-1 Hyperion data. Seasonal dynamics of the estimates and estimate uncertainty are evaluated. Improved LAI estimation accuracy over comparable VI regression was demonstrated. Abstract: In this paper, Bayesian inversion of a physically-based forest reflectance model is investigated to estimate of boreal forest canopy leaf area index (LAI) from EO-1 Hyperion hyperspectral data. The data consist of multiple forest stands with different species compositions and structures, imaged in three phases of the growing season. The Bayesian estimates of canopy LAI are compared to reference estimates based on a spectral vegetation index. The forest reflectance model contains also other unknown variables in addition to LAI, for example leaf single scattering albedo and understory reflectance. In the Bayesian approach, these variables are estimated simultaneously with LAI. The feasibility and seasonal variation of these estimates is also examined. Credible intervals for the estimates are also calculated and evaluated. The results show that the Bayesian inversion approach is significantly better than using a comparable spectral vegetation index regression.