Evaluation of the influence of two operational fraction of absorbed photosynthetically active radiation (FAPAR) products on terrestrial ecosystem productivity modelling

Abstract

The fraction of photosynthetically active radiation absorbed by vegetation (FAPAR) represents the available light energy for plant productivity and is the key variable influencing photosynthesis, transpiration, and energy balance in most terrestrial vegetation productivity models. With availability of earth observation data from different satellite sensors increasing, a number of FAPAR products are being generated. Several studies have investigated the differences between these products. However, very few studies have investigated how the differences between these products influence the output from ecosystem productivity models that utilise them. This study evaluated the influence of two operational FAPAR products (i.e. the MODIS and CYCLOPES FAPAR products) on the terrestrial vegetation primary productivity predicted by the Carnegie–CASA model across various biomes in the USA. The GPP predicted by the Carnegie–CASA model was compared to GPP measurements from various flux tower sites representing five biomes (i.e. croplands, broadleaf deciduous forests, grassland, needle-leaf evergreen forests, and savanna woodland). With the exception of cropland sites, the two FAPAR products resulted in GPP predictions which were higher than the in situ GPP measurements for the evaluated biomes. However, the CYCLOPES FAPAR product resulted in GPP outputs which were closer (lower RMSE values) to the in situ measurements than the MODIS FAPAR product. The two FAPAR products do not account for the FAPAR absorbed by non-photosynthetic elements of the canopy, which may lead to overestimation of the FAPAR that is actually used in photosynthesis. This could explain the higher GPP values derived using these products when compared to the in situ GPP measurements.

Acknowledgements

The authors would like to thank the University of Southampton’s Geography and Environment Academic Unit and the Overseas Research Students Awards Scheme (ORSAS) for providing funding to facilitate this study. The authors would also like to thank Dr Guido van der Werf, VU University in Amsterdam, for supplying the Carnegie–CASA model source code. We would also like to thank NASA for providing the MODIS data sets, POSTEL for the CYCLOPES data sets, and the many researchers within the AmeriFlux consortium for availing the flux tower data for use in the advancement of science. Finally, we would like to thank the two anonymous reviewers, whose input helped improve this manuscript.