Abstract
The seasonality of Amazon forest productivity and photosynthetic activity has recently been investigated under a new perspective by a series of publications. The debate about possible factors explaining this seasonality is vivid, and the possibility of several hypotheses has been tested, including canopy phenology and leaf demography, and changes in illumination geometry combined with the complex 3D structure of the canopy. A manifold of measurements and techniques have been used to test these hypotheses, including field observations of leaf demography from ground measured litterfall and phenocam, airborne and satellite remote sensing and 3 dimensional radiative transfer modeling. Our study explores the relative influence of leaf demography and illumination geometry on remotely sensed data. To achieve this, we take advantage of the latest advances in the domain of physical modeling at both leaf and canopy scale. The leaf optical properties model PROSPECT-D was used to model leaf optical properties at various growth stages based on field observations and theoretical leaf biochemical composition during its development and senescence. The 3-dimensional radiative transfer model DART was used to simulate various levels of complexity of canopy covers, from a turbid layer to complex canopy derived from airborne LiDAR acquisitions. Data for leaf demography and ontogeny taken from recent publications was used and integrated into canopy simulations corresponding to year-long observations. Data acquisitions were performed in the frame of the HyperTropik project (funded by CNES) and our results focus on analyzing the influence of separated and combined factors on various spectral attributes, including Enhanced vegetation index and hyperspectral metrics.