Modeled and measured fPAR in a boreal forest: Validation and application of a new model

The fraction of absorbed Photosynthetically Active Radiation (fPAR) plays a critical role in carbon balance studies and is one of the Essential Climate Variables (ECV). fPAR can be used to monitor CO2 assimilation by vegetation both seasonally and interannually. Temporal courses of fPAR are difficult to measure in field conditions, and thus, they are most often estimated based on models which quantify the dependency of absorbed radiation on canopy structure. In this study, we demonstrate how a physically-based canopy radiation model was adapted into an fPAR model, and compare modeled and measured fPAR in structurally different forest stands. The model is based on the spectral invariants theory, and uses leaf area index (LAI), canopy gap fractions and spectra of foliage and understory as input data. For validation of the model, measurements of instantaneous fPAR were performed using the TRAC instrument in nine Scots pine, Norway spruce and Silver birch stands in southern Finland. Good agreement was found between modeled and measured fPAR. Next, we applied the model to predict temporal courses of fPAR. For this, continuous data on incoming radiation were obtained from a nearby flux tower. Polynomial functions were fitted to the measured canopy gap fractions to create a hemispherical gap fraction distribution for each stand. These hemispherical gap fraction distributions were used together with sky irradiance models to simulate incoming radiation fields. Field fPAR estimates agreed with modeled ones (RMSE for morning and noon were 0.03 and 0.06, respectively). Application of the model to simulate diurnal and seasonal values of fPAR for the study stands indicated that the ratio of direct-to-total incident radiation and LAI are the key factors behind the magnitude and variation of stand-level fPAR values. (C) 2014 Elsevier B.V. All rights reserved.