Approaches for Modelling Transient Unsaturated-saturated Groundwater Flow during and after Construction

Construction of underground research laboratories and geological disposal facilities has a significant transient impact on groundwater flow, leading to a drawdown in the water table and groundwater pressures, and groundwater inflow into shafts, access ways and tunnels accompanied by desaturation of the surrounding rock. Modelling the impact of underground facilities on groundwater flow is important throughout the construction and operation of the facilities, e.g. estimating grouting and water treatment facility requirements during construction, and estimating the rate of resaturation of the engineered barrier system and the establishment of steady-state groundwater flow after backfilling and closure. Estimating the impact of these effects requires modelling of transient groundwater flow under unsaturated conditions at large scales, and over long timescales. This is a significant challenge for groundwater flow modelling, in particular because of the non-linearity in groundwater flow equations, which can have a marked effect on suitable timesteps for transient calculations. In addition, numerical grids need to be developed at appropriate scales for capturing the transition between saturated and unsaturated regions of the sub-surface, and to represent the features of complex hydrogeological structures such as heterogeneous fractured rock. The Japan Atomic Energy Agency (JAEA) has been developing modelling techniques to overcome these problems as part of the Mizunami Underground Research Laboratory (MIU) Project in the Tono area of Gifu Prefecture, Japan. An integrated geological and hydrogeological modelling, and visualisation system referred to as GEOMASS has been developed, which allows for transient unsaturated groundwater flow modelling in the presence of dynamic underground excavation models. The flow simulator in GEOMASS, FracAffinity, allows for such modelling by the application of sophisticated gridding techniques, allowing for modification of hydraulic conductivity in key zones, and by suitable modification of water retention models (the relationship between saturation and pressure, and saturation and hydraulic conductivity). The approaches that have been developed in GEOMASS have been tested through a series of models of increasing complexity, and the testing has demonstrated that there is no significant impact on estimates of regional groundwater flows or local estimates of flow into underground excavations. The tools and approaches that are described in this paper are of significance in all geological disposal projects, where a key requirement is to estimate and understand the hydrogeological regime and the transient response of groundwater flow to underground construction. Such understanding is important for construction, operation and post-closure phases of facility development. INTRODUCTION Understanding deep groundwater flow relies on a combination of geological knowledge, hydrodynamic testing and monitoring, geochemical characterisation and integration (e.g. Neumann, 2005; Eaton et al., 2007; Follin et al., 2008). Very often, the lack of integration introduces uncertainty in the hydrogeological models that is further transferred to the use of numerical models and to the decision based tools. Thus, the synthesis of research programmes requires software tools that facilitate the integration of the fundamental aspects of groundwater flow in the area of interest. In this respect, the Japan Atomic Energy Agency (JAEA) has been investigating the Tono area (e.g. Mizunami Underground Research Laboratory (MIU) construction site) for many years to develop understanding of the fundamental