Parameter uncertainty and sensitivity analysis of the three Gorges Reservoir and Xiangxi River EFDC model

Water quality models are decision support tools for the planning and management of the aquatic environment. However, the application of the model needs an intricate calibration process due to the various ranges of numerous parameters. To help improve the accuracy of model simulation, and reduce the workload of parameter adjustment to similar surface waters, parameter uncertainty and sensitivity analyses of the Environmental Fluid Dynamics Code (EFDC) model were carried out in this study. The EFDC model was first calibrated and simulation results agree well with the measured data. The Generalised Likelihood Uncertainty Estimation (GLUE) and Regional Sensitivity Analysis (RSA) methods are then applied to analyze the uncertainty and sensitivity of the model. Eighteen kinetic parameters related to algae and organic matter are filtered and analyzed. The results show that the sensitivities of the model to eighteen input parameters are significantly different. The modeled algae levels measured as chlorophyll-a (Chl-a) are highly sensitive to the optimal growth rate and the maximum basal metabolism rate of cyanobacteria (BMRc and PMc), with the Sensitivity Indices (SI) at 0.66 and 0.78, respectively. The inorganic nutrient levels (phosphorus, nitrate nitrogen and ammonia nitrogen) are highly sensitive to minimum respiration rates of corresponding dissolved organic matter (KDP and KDN) with SIs at 0.78, 0.56 and 0.88. Dissolved oxygen (DO) is highly sensitive to PMc and KDC, with their SIs at 0.66 and 0.85, respectively. The uncertainty interval is focused on the periods of high algae concentration. The simulated uncertainty in the surface water is higher than that in middle-layer water, and might be related to algal transport processes like settlement and horizontal transport. The results of the uncertainty and sensitivity analyses in this study support a better understand of the modeling mechanisms and provide scientific guidance for calibration in similar waterbodies. © 2022