The impact of forest architecture parameterization on GPP simulations

Abstract

The presence of a forest strongly affects ecosystem fluxes by acting as a source or sink of mass and energy. The objective of this study was to investigate the influence of the vertical forest heterogeneity parameterization on gross primary production (GPP) simulations. To introduce a heterogeneity effect, a new method for the upscaling of the leaf level GPP is proposed. This upscaling method is based on the relationship between the leaf area index (LAI) and the leaf area density (LAD) profiles and the standard sun/shade leaf separation method. The effect of the crown shape and foliage distribution parameterization on the simulated GPP is confirmed in a comparison study between the proposed method and the standard sun/shade upscaling method. The observed values used in the comparison study are assimilated during the vegetation period on three distinguished forest eddy-covariance (EC) measurement sites chosen for the diversity of their morphological characteristics. The obtained results show (a) the sensitivity of the simulated GPP to the leaf area density profile, (b) the capability of the proposed scaling method to calculate the contribution of the different canopy layers to the entire canopy GPP, and (c) a better agreement with the observations of the simulated GPP with the proposed upscaling method compared with the standard sun/shade method.