Application of Sobol GSA to the numerical model of seawater intrusion SEAWAT – A case study in Longkou City

Northwestern coastal plain of Longkou city, Shandong province, China Parameter uncertainty is one of the main uncertainty sources in groundwater numerical modelling. Owing to the strong nonlinearity and significant parameter interactions inherent in seawater intrusion (SI) models, Global Sensitivity Analysis (GSA) is essential for deciphering the complex dependencies between inputs and outputs. Therefore, the present study investigates the spatiotemporal sensitivity of seven parameters in an SI model by integrating the Sobol' GSA method directly into the SEAWAT_V4 source code. (1) This study provides, a spatiotemporally explicit sensitivity map for a complex multi-layer coastal aquifer system, revealing that the dominance of parameters ( PR , DP , HK ) shifts not only across space (e.g., from advection-dominated near coast to dispersion-dominated in mixing zones) but also over time, reflecting dynamic hydrogeological processes. (2) Crucially, the analysis quantifies significant interactions among key parameters ( HK , PR , DP ), demonstrating that their combined effects can outweigh individual impacts in critical zones, an aspect often overlooked in management strategies. • The global sensitivity of SEAWAT model in Longkou area was analyzed. • The sensitivity ordering of the parameters varied spatially and temporally. • The parameters interactions had a significant impact on the simulated Cl - concentration.