Abstract:Accurate and nondestructive estimation of cadmium (Cd) status of Cichoriumintybus L. is important for site-specific crop heavy metal stress management. Aiming to rapidly and precision assessment of leaf Cd concentration in Cichoriumintybus L., a hydroponic experiment with three cultivars, i.e., Europea chicory, America chicory and Qianyu no.1 chicory, was conducted in Henan Agricultural University from Dec. 2018 to Mar. 2019. Seven different Cd concentration treatments (0μmol/L, 5μmol/L, 10μmol/L, 25μmol/L, 50μmol/L, 100μmol/L and 200μmol/L) were established with five replications per treatment, and the in situ leaf hyperspectra were taken on at sixleaf and tenleaf stages. Meanwhile, chemical assays of these Cichoriumintybus L. samples were performed in the laboratory. Moreover, data from an independent experiment under 50μmol/L Cd stress condition with nine varieties in Mar. 2019 was also collected to test the transferability of the established optimal monitoring model for leaf Cd concentration prediction. After correlation analysis, stepwise regression (SWR), principal component regression (PCR) and partial least square (PLS) were used to perform the relationship between raw spectral reflectance (R), the first derivative reflectance (FDR) and leaf Cd concentration, respectively. The results showed that leaf Cd concentration in Cichoriumintybus L. was increased with the increase of Cd stress conditions, and the changes in situ leaf spectral reflectance under varied Cd rates were highly significant in the visiblenear infrared region (400~1300nm), with consistent patterns across the different cultivars and growth stages. Using a validation dataset, the best models were calculated with the FDR-PLS method, which yielded the highest coefficient of determination (R2) of 0.92 and the lowest root mean square error (RMSE) and relative percent deviation (RPD) of 181.3mg/kg and 2.96, respectively. The variable importance in projection (VIP) score resulting from PLS regression model was used to determine the effective wavelengths and reduce the dimensionality of the hyperspectral reflectance data. The newly developed FDR-PLS model using the effective wavelengths (659nm, 725nm, 907nm, 1026nm, 1112nm, 1255nm and 1630nm) performed well in leaf Cd concentration prediction with R2 of 0.834 and RPD of 2.41. The validation in the nine varieties’ experiments also indicated an excellent accuracy between the observed and predicted values for leaf Cd concentration (R2 was 0.817, and RPD was 1.77). The overall results demonstrated the applicability and feasibility of the FDR-PLS model for estimating the Cd status of Cichoriumintybus L. using in situ leaf hyperspectral reflectance data.
