Groundwater simulations are generally intended to characterise the hydrogeological/hydrological flow in the subsurface or to constrain the physical properties of the medium itself.
As for all modelling situations, external input parameters are required to initiate the modelling process. These parameters can be roughly divided in the three groups, namely the physical parameters such as permeability or, hydric parameters such as recharge characteristics and geometrical parameters such as aquitard/aquifer extent and location. The presented work will focus on that last aspect and more importantly on the impact of the quality and reliability of the prior information used to derive aforementioned geometries in complex geological settings.
When the geological structures are too complex to be appropriately modelled using the layered earth hypothesis, hydrogeologists rely on structural measurements, drillholes, maps and geophysical data to build geological models meant to represent the subsurface as accurately as possible. Therefore, hydrogeological simulations are heavily reliant on the quality of geological information and the reliability of the modelling engine selected. That is, the global uncertainty of the geological modelling process propagates and compounds with that of the hydrogeological modelling process.
Implicit geological modelling engines such as GeoModeller are considered fairly reliable and their behaviour under uncertainty has been extensively studied in the recent year by the geological modelling community. An important concept which developed from these studies is that what appears like minor changes in the structure of a geological model may alter its lithological topology significantly. This, in turn, massively impacts the end results of a groundwater simulation because of unforeseen topology shifts resulting, for example, in aquitard breaches. The relationship between geological uncertainty and hydrogeological uncertainty is non-linear.
The presented work explores the consequences of accurate geological uncertainty propagation on the quality and reliability of hydrogeological simulations using the Saxony geological model developed jointly by Intrepid Geophysics and DHI as a case study.