Modeling uncertainties in estimation of canopy LAI from hyperspectral remote sensing data – A Bayesian approach

Abstract

Hyperspectral remote sensing data carry information on the leaf area index (LAI) of forests, and thus in principle, LAI can be estimated based on the data by inverting a forest reflectance model. However, LAI is usually not the only unknown in a reflectance model; especially, the leaf spectral albedo and understory reflectance are also not known. If the uncertainties of these parameters are not accounted for, the inversion of a forest reflectance model can lead to biased estimates for LAI. In this paper, we study the effects of reflectance model uncertainties on LAI estimates, and further, investigate whether the LAI estimates could recover from these uncertainties with the aid of Bayesian inference. In the proposed approach, the unknown leaf albedo and understory reflectance are estimated simultaneously with LAI from hyperspectral remote sensing data. The feasibility of the approach is tested with numerical simulation studies. The results show that in the presence of unknown parameters, the Bayesian LAI estimates which account for the model uncertainties outperform the conventional estimates that are based on biased model parameters. Moreover, the results demonstrate that the Bayesian inference can also provide feasible measures for the uncertainty of the estimated LAI.

Bayesian estimation of seasonal course of canopy leaf area index from hyperspectral satellite data

“…The method is based on Bayesian inversion of a physically-based forest reflectance model. The simulation results in (Varvia et al, 2017) indicated improved estimation accuracy compared to the empirical vegetation index regression approach. Main advantages of the new method are that it also allows simultaneous estimation of other forest reflectance model parameters, such as leaf albedo, and produces uncertainty estimates for the model variables.…”

“…In this section, the Bayesian approach to LAI estimation is shortly summarized. Except for slight adjustments in certain hyperparameters, the methodology is identical to Varvia et al (2017) and the reader is referred there for more detail.…”

“…The wavelength dependent variables ω L and ρ g are approximated using splines following Varvia et al (2017):…”

“…Recently, Varvia et al (2017) proposed a Bayesian method to estimate canopy LAI from hyperspectral satellite measurements. The method is based on Bayesian inversion of a physically-based forest reflectance model.…”

Dynamic Retrieval of Olive Tree Properties Using Bayesian Model and Sentinel-2 Images

“…To overcome such issues, applying Bayesian models to RTM inversion processes provides a straightforward way to quantify the covariance and uncertainty of parameter estimates while integrating different sources of information. RTM inversion has relied on the use of independent prior information to address the otherwise underdetermined challenge of estimating a high number of parameters from a limited number of observations [21], [24], [25]. As these studies neglect the parameter uncertainties or just approximate it, recent research has showed the efficiency of the hierarchical Bayesian technique called the Monte Carlo Markov Chain (MCMC) methods [26], [27].…”