Deconvolution of pigment and physiologically related photochemical reflectance index variability at the canopy scale over an entire growing season

The sensitivity of the photochemical reflectance index (PRI) to leaf pigmentation and its impacts on its potential as a proxy for light‐use efficiency (LUE) have recently been shown to be problematic at the leaf scale. Most leaf‐to‐leaf and seasonal variability can be explained by such a confounding effect. This study relies on the analysis of PRI light curves that were generated at the canopy scale under natural conditions to derive a precise deconvolution of pigment‐related and physiologically related variability in the PRI. These sources of variability were explained by measured or estimated physiologically relevant variables, such as soil water content, that can be used as indicators of water availability and canopy chlorophyll content. The PRI mainly reflected the variability in the pigment content of the canopy. However, the corrected PRI, which was obtained by subtracting the pigment‐related seasonal variability from the PRI measurement, was highly correlated with the upscaled LUE measurements. Moreover, the sensitivity of the PRI to the leaf pigment content may mask the PRI versus LUE relationship or result in an artificial relationship that reflects the relationship of chlorophyll versus LUE, depending on the species phenology. In this study, a new method to analyze PRI variability at canopy scale is shown. Two components of PRI variability were successfully identified, and explained. While the first component (explaining most of PRI daily variability) allowed the retrieval of canopy‐scale light use efficiency, the second component (explaining most of PRI seasonal variability) was shown to interfere with the use of the PRI as a proxy of LUE. The relevance of these findings to the direct use of PRI as a proxy of the LUE in remote sensing is discussed in details, and potential solutions are proposed.