Sensitivity analysis of the row model#039;s input parameters

摘要

In this study, we use the variance based sensitivity method to analyze the sensitivities of the row crop model's input parameters. This method consists of three steps: sample generation, model execution and the calculation of sensitivity indices. The results of sensitivity analysis of row crop model's input parameters indicate that the sensitivities of the row model's parameters are different under different viewing angles. We also find that the row structure of canopy can affect the sensitivities of model's inputs parameters. LAI is found to be generally the most sensitive parameter in three typical spectral bands. a and h are also identified as sensitive parameters in the along-row or near-along row directions in the VNIR bands. LIDF.a is relatively sensitive at some viewing angles and LIDF.b is insensitive in three typical spectral bands. k has marginal influence on the model outputs in all of the viewing angles and spectral bands, except around hotspot directions in the red band. The main contribution of this work is that the technique of global sensitivity analysis is applied to the row model and the results are informative for the retrieval of the parameters during the inversion.