Retrieving leaf chlorophyll content using a matrix-based vegetation index combination approach

Abstract

Leaf chlorophyll content (ChlLeaf), which is responsible for light harvesting for photosynthesis, is an important parameter for carbon cycle modeling and agriculture monitoring at regional and global scales. Since the spectral signals of chlorophyll content and leaf area are highly coupled, it is required to remove the effect of the LAI on the retrieval of ChlLeaf from satellite data. In this paper, an approach for the retrieval of ChlLeaf was proposed. A 2-dimensional matrix-based relationship between ChlLeaf and two VIs was established using simulated datasets from the PROSAIL model. The matrix was formed by dividing the two-VI space into m × n cells and assigning the ChlLeaf value to each cell. Based on the matrix, the Chlleaf can be retrieved using the two VIs from observations. Three matrices of different VI pairs for retrieving ChlLeaf were tested using the PROSAIL simulated data. The results show that the matrix formed with two new VIs, RERI[705] and RERI[783], works best. The results from the matrices of two VIs are better than those from individual VIs as well as from VI ratios. The matrices were successfully used to retrieve the ChlLeaf of winter wheat from Sentinel-2 images. The ChlLeaf estimations using the RERI[705]-RERI[783] matrix achieves an accuracy of R2 = 0.70, RMSE = 10.4 μg/cm2, and NRMSE = 11.9%. The estimations using the TCARI-OSAVI and the R740/R705-R865/R665 matrices are also in good agreement with the measured ChlLeaf (R2 > 0.48, RMSE < 13.1 μg/cm2, and NRMSE < 15.1%). The matrix-based VI combination approach has the potential for the operational retrieval of ChlLeaf from multispectral satellite data.