Research Associate University of Illinois at Urbana-Champaign
Disclosure: Disclosure information not submitted.
Exploring natural variation in photosynthesis provides a promising avenue to increase crop yield to feed a growing world population, yet techniques are still lacking to screen genetic variation in a high-throughput manner. Remote sensing-inverted solar-induced fluorescence (SIF) has been widely used to quantify photosynthesis at broad-scale, yet there is a need to further evaluate whether SIF-associated parameters (SIF or SIF yield) can be leveraged to relieve the bottleneck in phenotyping of photosynthesis. In this study, we used time-synchronized hyperspectral images (with the spectral region from 400 to 900 nm) and irradiance measurements of sunlight under clear-sky conditions to estimate SIF and SIF yield. Ground truth photosynthetic variables were collected over 10 tobacco cultivars in 2017 (July 6, 7, 12, 18, and 31) and 2018 (July 24 and 25) using a portable leaf gas exchange system. Our results showed that when measurements of all days were considered, both SIF and SIF yield had no statistical relationship with photosynthetic variables. When measurements were separated by groups of days (i.e., group 1: July 6, 7, and 12, 2017, group 2: July 18 and 31, 2017, and group 3: July 24 and 25, 2018), however, SIF yield exhibited a statistically significant negative relationship with photosynthetic variables. More specifically, cross-validation regression results revealed that the performance (i.e., coefficient of determination, R2) of SIF yield to predict Vcmax (Jmax) ranged from 0.71 to 0.87 (0.62 to 0.92). Compared to previous studies relating reflectance spectra to photosynthetic capacities, our study has the potential to further avoid costly hyperspectral computations associated with inversions of leaf/plot-level physiological traits and/or optimization of spectral indices for characterizing photosynthetic variation. The presented approach also has merits to overcome future needs for time laborious ground-truth data to build models that predict photosynthetic performance.