Exploring the interrelated effects of soil background, canopy structure and sun-observer geometry on canopy photochemical reflectance index

Abstract

Photochemical reflectance index (PRI) is an intriguing avenue for monitoring photosynthetic light use efficiency (LUE), which is a crucial physiological variable and the primary source of uncertainty in gross primary production (GPP) prediction. The correlations between PRI and LUE induced by the energy-dependent xanthophyll cycle are overall convincing at the leaf scale. However, their relationship becomes complicated and less reliable at the canopy scale due to several confounding factors, such as soil background, canopy structure and sun-observer geometry. A fundamental understanding of the interrelated effects of these factors is still missing. In this study, the roles of soil reflectance, canopy structure and sun-observer geometry were examined by investigating radiative transfer processes that determine canopy PRI. Theoretical analysis shows that the deviation between canopy- and leaf-level PRI is mainly attributed to soil background, namely the fraction of sunlit soil in the scene and the discrepancy between leaf and soil reflectance. Canopy structure and sun-observer geometry are not the direct drivers of variations in canopy PRI. They affect canopy PRI by altering the contribution of soil to canopy reflectance. Furthermore, several numerical experiments based on 1D and 3D radiative transfer models, which compare PRI of canopies with black soil (i.e., soil reflectance is zero) and non-black background, were conducted to confirm and validate the decisive role of soil background in explaining the deviation between canopy and leaf PRI. Multi-angular canopy PRI from PROBA-CHRIS data was employed to examine the angular dependency of canopy PRI changing with vegetation coverage. The numerical experiments demonstrate that canpy PRI differs from leaf PRI as large as 150% when canopy leaf area index (LAI) or the fraction of vegetation coverage is small. Canopy PRI is almost identical to soil PRI (i.e., PRI computed from soil reflectance) when the contribution of soil background to canopy reflectance is large. In contrast, when the soil reflectance is zero or LAI is greater than four, canopy PRI is almost identical to leaf PRI and independent of canopy structure and sun-observer geometry. Nonetheless, when leaf and soil reflectance are similar, canopy PRI is always close to leaf PRI regardless canopy structure and sun-observer geometry. The analysis of multi-angular PRI reveals that the angular dependency decreases with vegetation coverage due to weaker soil background effect. This study narrows down the causes of variation in canopy PRI to the fraction of observed soil background and soil reflectance. It suggests that estimating and accounting for the soil contribution to canopy PRI are necessary to mitigate angular and canopy structural effects on canopy PRI and eventually extract leaf physiological information from canopy PRI.