Hyperspectral characteristics and leaf area index monitoring of rice (Oryza sativa L.) under carbon dioxide concentration enrichment

Abstract

Leaf area index is a vital biological parameter, which is widely used to monitor plant growth, evaluate health status, and predict yield. Remote sensing techniques are known to be nondestructive and effective methods for estimating leaf area index of plants, but little attention has been paid to predicting leaf area index under carbon dioxide enrichment. Field experiments were conducted to estimate the rice leaf area index under elevated carbon dioxide using hyperspectral remote sensing. The results showed that various spectral parameters, including the first derivative reflectance at 467 nm, the yellow edge amplitude, the normalized value of the green peak and red valley reflectance, the ratio of red edge and blue edge area, and the normalized value of the red edge and blue edge area, have a significantly correlated with the leaf area index. Further comparing the results of models, the normalized value of the green peak and red valley reflectance exhibited the optimal performance for estimating leaf area index. The best-fitted inversion model was y=6.89x 0.46 with the coefficient of determination was 0.75 and root mean square error was 0.31. This finding is helpful in providing guidance for monitoring rice leaf area index under carbon dioxide enrichment.