Abstract:Sensitivity analysis for parameters of remote sensing physical models is a prerequisite for inversion.The EFAST(Extended Fourier Amplitude Sensitivity Test)as a global sensitivity analysis method,can analyze not only a single parameter's sensitivity but also the coupling effects among parameters.It is usually applied to analyse parameters' sensitivity of the high-dimensional nonlinear models.In this paper,the SAIL model is taken as an example,the EFAST method and the field measured data of winter wheat in Shunyi district in 2001were applied to analyze the model parameters' sensitivity throughout the growing season and in different growth stages respectively.The results are compared with those of the USM(Uncertainty and Sensitivity Matrix)method.The results show that either the EFAST or the USM method for parameters' sensitivity analysis of the SAIL model is feasible;but the EFAST method,which takes into account of the coupling effects among all the parameters and the analysis result is global,compared to the USM method,is more objective and comprehensive.

EFAST Method for Global Sensitivity Analysis of Remote Sensing Model's Parameters