Sensitivity and Uncertainty Analyses of Flux-based Ecosystem Model towards Improvement of Forest GPP Simulation

Abstract

An ecosystem model serves as an important tool to understand the carbon cycle in the forest ecosystem. However, the sensitivities of parameters and uncertainties of the model outputs are not clearly understood. Parameter sensitivity analysis (SA) and uncertainty analysis (UA) play a crucial role in the improvement of forest gross primary productivity GPP simulation. This study presents a global SA based on an extended Fourier amplitude sensitivity test (EFAST) method to quantify the sensitivities of 16 parameters in the Flux-based ecosystem model (FBEM). To systematically evaluate the parameters’ sensitivities, various parameter ranges, different model outputs, temporal variations of parameters sensitivity index (SI) were comprehensively explored via three experiments. Based on the numerical experiments of SA, the UA experiments were designed and performed for parameter estimation based on a Markov chain Monte Carlo (MCMC) method. The ratio of internal CO2 to air CO2 ( f C i ) , canopy quantum efficiency of photon conversion ( α q ) , maximum carboxylation rate at 25 ° C ( V m 25 ) were the most sensitive parameters for the GPP. It was also indicated that α q , E V m and Q 10 were influenced by temperature throughout the entire growth stage. The result of parameter estimation of only using four sensitive parameters (RMSE = 1.657) is very close to that using all the parameters (RMSE = 1.496). The results of SA suggest that sensitive parameters, such as f c i , α q , E V m , V m 25 strongly influence on the forest GPP simulation, and the temporal characteristics of the parameters’ SI on GPP and NEE were changed in different growth. The sensitive parameters were a major source of uncertainty and parameter estimation based on the parameter SA could lead to desirable results without introducing too great uncertainties.

Sensitivity Analysis of Biome-BGC for Gross Primary Production of a Rubber Plantation Ecosystem: A Case Study of Hainan Island, China

“…The difference was not only in the ranking of parameter sensitivity; the latter analysis results did not show differences between carbon flux (GPP, NEE) and LAI ( Figure 2 ). For the settings of the ranges of 10% and 20% perturbation, previous studies have shown that there are certain differences in the analysis results [ 20 ]. The reason for our undifferentiated results may be that the model was revised in this study.…”

“…The commonly used model calibration methods include model parameter optimization based on observation data and data assimilation methods [ 19 ]. Sensitivity analysis methods can quantify the contribution of parameters to the model output, which can be used to explain the functional pattern of the model and its parameter optimization [ 20 ]. The variance-based global sensitivity analysis method can analyze nonlinear models and is widely used in various process models to identify sensitive parameters for different outputs or the coupling effects of different parameters on the outputs [ 21 ].…”

Improving the Simulation Accuracy of the Net Ecosystem Productivity of Subtropical Forests in China: Sensitivity Analysis and Parameter Calibration Based on the BIOME-BGC Model

“…The range of parameter values has a significant effect on the results of sensitivity analyses. A smaller baseline value is more likely to be affected by the parameter range and smaller parameter ranges limit their sensitivity indexes [59]. The wide range of variability in the NEP for the ENF indicates that when generating multiple sets of sample parameters, some combinations contained parameters that were outside the normal range of physiological parameters of the ENF (Figure 4 and Table 4), representing forests experiencing unsustainable or disturbed growth during the years simulated by the model, which led to lower NEP values in the simulations.…”

Sensitivity Analysis of Biome-BGCMuSo for Gross and Net Primary Productivity of Typical Forests in China

“…By correcting related sensitive parameters, the uncertainty of input parameters can be effectively reduced, and the simulation accuracy of the process model can be improved. Previous studies showed that parameters with smaller basic values were more susceptible to the influence of the parameter range, and a narrow range would limit the sensitivity index of a small parameter (Ma et al 2020). Therefore, the range of the parameter is crucial to the result of the sensitivity analysis.…”