Measurements and simulations of seasonal changes in canopy reflectance of a deciduous forest

Canopy reflectance of forests is well known to show strong seasonal dynamics. Separating this variation into contributions from changes in canopy structure, leaf optical/biophysical properties and understory vegetation remains challenging. For example, the decrease of NDVI during the growing season of deciduous forests is caused by a decrease of near-infrared (NIR) reflectance and often erroneously taken to be equivalent to changes in leaf area. The aim of this study is to identify the main sources for seasonal variability in hyperspectral canopy reflectance of a deciduous forest in central Germany (Hohes Holz), where we measured canopy reflectance and transmittance, leaf reflectance and transmittance, overstory leaf area index, leaf angles, and understory vegetation and soil background spectra every three weeks. Leaf reflectance and transmittance of the four dominant tree species of Hohes Holz (Fagus Sylvatica, Quercus Robur, Carpinus Betulus, Betula Pendula) was additionally measured in an arboretum near Leipzig, Germany, on a weekly basis. Seasonal changes in leaf biophysical parameters were estimated by inversion of the model PROSPECT (Jacquemoud et al., 1990). Measured hyperspectral canopy reflectance and transmittance will be compared to simulated reflectances and transmittances using the radiative transfer formulations of Sinoquet et al. (1990) and Kuusk and Nilson (2014). Mean leaf reflectance initially increased in the NIR from 44.5% to 48.2% from May until beginning of July, while leaf transmittance decreased from 48.4% to 42.4%. This can be explained by an increase from 1.4 to 1.6 of the PROSPECT leaf structure parameter N, an increase in leaf mass per area from 0.005 to 0.007 g/cm2, and a decrease in water content from 0.0097 to 0.0087 g/cm2. From beginning of July end of July NIR reflectance decreased and transmittance increased. In the visible range leaf reflectance increased from 4.6% to 5.7% and transmittance decreased from 9.1% to 6.4% due to an estimated increase of chlorophyll a+b from 29.88 to 36.46 ug/cm2 and of carotenoids by 8.4 to 10.1 ug/cm2. This study will help to understand seasonal changes in canopy structura, canopy reflectance, and spectral indices such as NDVI and the NIRv-Index, which has recently been proposed as a good proxy for plant productivity (Badgley et al., 2017).