GLUE algorithm capability in estimating the van Genuchten soil–water characteristic parameters and their uncertainties

چکیده مقاله

Uncertainty in the model’s structure, parameters, and input data can be viewed as the three primary sources of hydrological models failure; hence, doing an uncertainty analysis on the model can provide valuable insights into the model’s error sources. In this study, the generalized likelihood uncertainty estimation (GLUE) algorithm was used to estimate and examine the uncertainty of parameters and model structure in four soil textural classes using the Monte Carlo approach. Additionally, the generalized sensitivity analysis (GSA) algorithm was utilized to determine the most sensitive parameter in each soil texture. The performance of the GLUE algorithm was compared to that of the particle swarm optimization (PSO) algorithm. The results indicated that GLUE algorithm outperformed PSO in estimating the parameters of the van Genuchten model with RMSE of 0.0031–0.0017 and R2 of 0.99. Additionally, statistical indices demonstrated that the PSO algorithm was more accurate than the RETC and Rosetta models. The posterior distribution of the parameters revealed varying degrees of uncertainty in the model parameters across the studied textural classes, which obviously afects the model’s performance. The results of the GSA algorithm indicated that the sensitivity of the parameters in the textures under study follows a distinct trend. Furthermore, the sensitivity of the parameters is afected not only by the inherent nature of the model parameters, but also by soil heterogeneity and soil management practices. Finally, our fndings indicated that the GLUE method, when applied with the equifnality concept and taking into account the uncertainty of the parameters and the model structure, can perform better than the PSO in predicting the model parameters. The degree of uncertainty in the van Genuchten model parameters and structure is a signifcant factor in the efciency of the methods used to estimate the parameters, which afects the accuracy and efciency of management programs and the output of soil hydraulic and hydrological models.