The 3D plant canopy radiative transfer analysis in an Alaskan black spruce forest: the characteristics of fraction of absorbed photosynthetically active radiation in the heterogeneous landscape

Over the last couple of decades, the three dimensional plant canopy radiative transfer models have been developed, improved and used for the retrievals of biophysical variables of vegetative surface. Fraction of absorbed photosynthetically active radiation (FAPAR) by plant canopy, a similar variable to heating rate in the atmosphere, is one of the important biophysical variables to infer the terrestrial plant canopy photosynthesis. FAPAR can be estimated by the radiative transfer model inversion or the empirical relationships between FAPAR and vegetation indices such as normalized difference vegetation index (NDVI). To date, some global FAPAR products are publicly available. These products are estimated from the moderate resolution satellites such as MODIS and SPOT-VEGETATION. One may apply the similar FAPAR algorithms to higher spatial resolution satellites if the ecosystem structures are horizontally homogeneous, which means that the adjacent satellite pixels have a similar spectral properties. If the vegetation surface is highly heterogeneous, "domain average FAPAR", which assumes no net horizontal radiation fluxes, can be unrealistically high (more than 1). In this presentation, we analyzed the characteristics of FAPAR in a heterogeneous landscape. As a case study, we selected our study site in a sparse black spruce forest in Alaska. We conducted the field campaigns to measure forest structural and optical properties that are used in the radiative transfer simulation. We used a 3D radiative transfer, FLiES (Kobayashi, H. and H. Iwabuchi (2008), A coupled 1-D atmosphere and 3-D canopy radiative transfer model for canopy reflectance, light environment, and photosynthesis simulation in a heterogeneous landscape, Remote Sensing of Environment, 112, 173-185) to create a high resolution simulated spectral reflectance and FAPAR images over the course of the growing season. From the analysis, we show (1) FAPAR with no net horizontal fluxes assumption can be higher than 1 when the spatial resolution is higher than 5m, and (2) the relationship between FAPAR and a vegetation index (NDVI) becomes unclear when the spatial resolution is higher than 1. Our results show the complexity in estimating FAPAR when using high-resolution satellite data.