Sensitivity Analysis of GreenLab Model for Maize

摘要

Sensitivity analysis is a powerful tool which gives important insights regarding parameterization in the process of model design. For an empirical model like the Greenlab model of plant growth relying on parameter estimation from experimental data, such study should open new perspectives for model identification. This study was aimed at analyzing the sensitivity of GreenLab model for maize. When the value of biomass production is considered as the output, the system tends to be linear, the level is above 94% for the SRC (Standardized Regression coefficients) study. Conversion efficiency and characteristic surface are proved to be the most sensitive factors. In Sobol's measure, we excluded the two most sensitive factors in the analysis, then the system linearity is weaker. We obtained the detailed sensitivity indexes for the other uncertain parameters, by which we get the driving forces of maize growth at different stages.