1) Overview

Purposes of construction of the hydrogeological model and groundwater flow simulation in the 1^st analysis loop are as follows.
understand general hydrology in and around the Shobasama site
estimate the extent of influence on hydrology caused by shaft excavation
clarify the problems to be solved in the subsequent investigations and analyses

(1) Setting of the study area

Study area (ca.4kmca.6km, Fig.4.26) includes the Shobasama site and is surrounded by ridges and streams that are considered as hydraulic boundaries. This area is set to improve the accuracy of groundwater flow simulation. The results of the RHS Project⁴⁵⁾ indicate that mountain ridges generally form hydraulic boundaries. The depth of this study area is set as GL-3,000m, taking depth of the shaft excavated in the MIU Project and the scale of the study area into consideration.

(2) Schedule of shaft excavation

The depth and diameter are set 1,000m (daily excavation rate: 1m) and 6m, respectively. Based on the preparatory examinations carried out until 1998 FY, it is planned to divide the process of excavation into following stages (Fig.4.27).
excavation stage of the upper part (ground surfaceGL-500m)
half-year-long investigation stage
excavation stage of the lower part (GL-500mGL-900m)
three-year-long investigation stage
excavation stage of the bottom part (GL-900mGL-1,000m).
Shaft is expressed as one of the elements in the hydrogeological model because it is thought to give the strongest hydraulic impact on the surrounding rock mass.

(3) Methodology of groundwater flow simulation

For the groundwater flow simulation, a finite element method (henceforth, 'FEM')-based 3D saturated/unsaturated seepage flow simulation program (TAGSAC)⁴⁶⁾ is adopted. TAGSAC is developed for the porous medium. A steady simulation taking the Tono Mine into consideration and a non-steady simulation for predicting the groundwater hydrology affected by the shaft excavation are carried out.

2) Construction of hydrogeological model

(1) Formation of 3D simulation mesh

The study area is modeled using the 3D simulation mesh to carry out the FEM-based 3D saturated/unsaturated seepage flow simulation. The procedures are as follows.

The study area is divided into meshes on a 2D plane so that the Tsukiyoshi Fault, boreholes, shaft excavated in the MIU Project, the exploratory shaft, the second shaft and galleries of the Tono Mine can be shown in the model. The number of element divisions amounts to 875 (EW:25NS:35). Elevations of the individual grid points are calculated from the 20m-interval mesh digital elevation data to express the topography of the study area (Fig.4.28). The ground surface ranges from elevation of 134m to 380m.

Twenty horizontal layers are set between the ground surface and the bottom of the study area, and 3D simulation meshes (FEM model) are prepared. The bottom is set at an elevation of -3,000m, taking the depth of the shaft in the MIU Project into consideration. Also upon this vertical division, the exploratory shaft, the second shaft and galleries of the Tono Mine are taken into consideration. As for the Tsukiyoshi Fault, 1m-wide platy elements are set stepwise to make the element division easy. The number of nodal points and elements in the 3D simulation mesh measures 19,656 and 17,500, respectively. The hydrogeological model is shown in Fig.4.29.

(2) Setting of stratigraphy and physical properties

The 3D simulation mesh is allocated to the geological model (ca.4kmca.6km : See Chapter 4.1.3), which is constructed based on the existing data. However, the size of the 3D simulation mesh doesn't directly corresponds with the thickness of the individual geological units of the geological model. Therefore, geological units at the center of the mesh are allocated to the 3D simulation mesh. As for hydraulic conductivities, figures shown in Tab.4.13 are set for each geological unit on the assumption that each geological formation is homogeneous hydrogeologically. These figures are set based on the results⁴⁷⁾ of groundwater flow simulation carried out in and around the Tono Mine in the RHS Project. Permeability in unsaturated domains are set based on the moisture characteristic curve and relatively hydraulic conductivity curve shown in Fig.4.30^{48), 49), 50)}.

Geology	Hydraulic conductivity (m/sec)
Seto Group	1.010-7
Oidawara Fm.	1.010-9
Akeyo Fm.	1.010-9
Toki Lignite-bearing Fm. (Upper)	5.010-9
Toki Lignite-bearing Fm. (Lower)	1.010-8
Toki granite (Weathered)	1.010-7
"Moderately fracture zone"	1.010-9
Tsukiyoshi Fault	1.010-10

(3) Setting of boundary conditions

Top boundary condition
Recharge rate at the ground surface is employed as the top boundary condition, taking precipitation into consideration. The recharge rate of 0.28mm/day⁵¹⁾ (an average rate calculated from observations between 1990 and 1997 in the vicinity of the Tono Mine) is adopted on the assumption that the recharge rate in the study area is the same as that of the Tono Mine. The ground surface is set as a free seepage face with inflow and outflow of water, taking water inflow into consideration.

Bottom boundary condition
The bottom boundary condition is set as an no-flow boundary.

Side boundary conditions
The side boundaries are represented by mountain ridges(the northern, eastern and western boundaries) and the Toki River (the southern boundary). The ridges could be considered as watersheds of groundwater, however, some studies point that these ridges are not always coincident with watershed because of local geological structures. In addition, the results⁵²⁾ of groundwater flow simulation encompassing the study area suggest the existence of a southwestward groundwater flow controlled by a regional topography. Accordingly, side boundaries along the ridges are set as permeable, and constant static head are given toward the depth. Constant heads given to the boundaries are determined by the following equation formulating the relationship between an elevations of ground surface and the water levels observed in boreholes around the study area (Fig.4.31).

(Constant head (m)) = 0.86H (elevation (m)) + 18.5

The side boundary along the Toki River is also set as permeable, and constant static head is given toward the depth. An elevation of ground surface is used as constant head given to the boundary.

(4) Galleries in the Tono Mine and shaft in the MIU Project

As for the galleries in the Tono Mine, the pressure head of 0m is adopted as the constant head as the galleries contact with the atmospheric pressure.
As for the shaft in the MIU Project, the corresponding geological units are taken away from the hydrogeological model one by one in response to the progress of shaft excavation. Nodal points representing the shaft wall are set as free seepage points on the assumption that the groundwater seeping out of the shaft wall is all pumped up to the ground surface.

3) Groundwater flow simulation

(1) Steady-state simulation (predictive simulation of groundwater hydrology in the current state)

Prior to a non-steady-state simulation (predictive simulation of groundwater hydrology caused by the shaft excavation), a steady simulation is carried out. The simulation aims at understanding the outline of groundwater hydrology taking the existence of the Tono Mine in the study area into consideration. The validity of the simulation is verified by comparing the simulated results with the pore pressure distribution measured in boreholes in the study area.

i) Comparison between simulated results with pore pressure distribution data

Fig.4.32 shows the comparisons between distributions of the measured hydraulic head and the simulated results.

This simulation model employs coarser meshes than the groundwater flow simulation⁴⁷⁾ carried out around the Tono Mine and gives insufficient consideration to the position of the Tsukiyoshi Fault. Nonetheless, the distributions of groundwater pressure in the AN-1, 68, drilled largely in sedimentary rocks around the Tono Mine, are well coincident with the simulated values in a drop of head caused by the existence of the mine. Also, the general trend of head is well reproduced (Fig.4.32 (a)(d)).

Simulated values for the DH-9 drilled largely in granite (Fig.4.32 (e)) are also coincident with the measured values in the general trend of head. However, as for the head, simulated values are a little higher than the measured values. This discrepancy of ca.30m is thought to be caused by following insufficient reproduction in the hydrogeological model.
permeability contrast generated by the Tsukiyoshi Fault
top boundary condition
position/shape of the Tsukiyoshi Fault

ii) Groundwater hydrology in the current state

This steady simulation shows following results on the distribution of heads and Darcy's velocity vectors (Fig.4.34). Fig.4.33 shows the cross sections of the simulated results.
Heads in the study area generally lower from north to south (Fig.4.34(1)(3)). This indicates that the groundwater flows from north to south as a whole.
The flow of groundwater above an elevation of 0m (henceforth "shallower part") is more affected by the topography of the ground surface than in a part below an elevation of -758m (henceforth "deeper part"). As a result, local distributions of heads and Darcy's velocity vectors are generated (a,c in Fig.4.34(1); d in Fig.4.34(3); e in Fig.4.34(4)). The shallower part is more permeable and susceptible to the top boundary condition than the deeper part, giving rise to larger values of Darcy's velocity vector (Figs.4.34(3) and (4)). The distribution of Darcy's velocity vectors is locally disturbed around the Tono Mine (b in Fig.4.34(1)).
The head distribution in the deeper part gently varies. The Darcy's velocity vectors are nearly horizontal from north to south (Fig.4.34(3)). It suggests that the deeper part is hardly affected by the topography of ground surface.
The groundwater flow simulation based on the current hydrogeological model indicates that the Tsukiyoshi Fault exerts only a small influence on the distribution of heads and Darcy's velocity vectors (Fig.4.34(1)(4)).

(2) Non-steady-state simulation (Predictive simulation of groundwater hydrology affected by the shaft excavation)

The non-steady-state simulation is carried out on the assumption that shaft is excavated as shown in Fig.4.27.

i) Changes in heads and Darcy's velocities caused by the shaft excavation

The results of predictive simulation of groundwater hydrology caused by the shaft excavation are summarized as follows (Fig.4.35). The cross sections shown in Fig.4.35 are cut at the same planes as Fig.4.33.
Heads are lowered around the shaft. As a result, the groundwater around the shaft not only flows but also increases in Darcy's velocity toward the shaft
The groundwater flows toward the shaft at a depth exceeding 1,000m (Fig.4.35 (2) and (3)).

ii) Changes in groundwater hydrology caused by the shaft excavation

Changes in groundwater hydrology caused by the shaft excavation are as follows (Fig.4.36). The cross sections shown in Fig.4.36 are cut at the same planes as Fig.4.33.
The distribution of head drops (drawdown) on a horizontal plane shows a nearly concentric pattern with the shaft as its center (Fig.4.36(1) and (2)). It indicates that the Tsukiyoshi Fault exerts a small influence on the groundwater hydrology in the groundwater flow simulation using the hydrogeological model.
The distribution of drawdown in a vertical section across the shaft shows an onion-shaped pattern (Fig.4.36(1) and (2)). It suggests that effects of the top boundary condition are remarkable enough to prevent the heads near the surface from greatly falling.
Drawdown caused by the shaft excavation are within the boundary of the ca.4kmca.6km study area (Fig.4.36(1) and (2)).

4) Examinations on boundary conditions

Examinations on the results of the 1^st analysis loop lead to the following problems to be solved during the 2^nd analysis loop.
As for the steady-state-simulation, the results approximately reproduce the current groundwater hydrology. It suggests that it is appropriate to have the side boundaries coincide mainly with mountain ridges and a river and to set them permeable boundaries. However, simulated values are generally a little larger than measured values. This is thought to be caused by the setting of relationship between groundwater level and elevation, which the hydrostatic pressure (side boundaries : Fig.4.29) is based on. Also, overestimation of the recharge rate is thought to be the cause. Thus, they should be reexamined.
The non-steady-state simulation results shows a small drawdown in the shallow part despite the assumption of the shaft excavation as is shown in Fig.4.36. It suggests that the recharge rate exerts a great effect on groundwater hydrology in the shallow part. Therefore, it is necessary to give different recharge rates to individual areas and geological formations instead of giving fixed value to the whole study area as in this simulation.
Effects of the shaft excavation do not extend to the bottom boundary of the study area. It suggests that it would be allowable to postulate the bottom boundary to be no-flow boundary at the same depth as this simulation.

5) Summary

Simulation results are as follows.
The groundwater flows from north to south as a whole.
Drawdown caused by the shaft excavation are within the boundary of the ca.4kmca.6km. It suggests that the size of the study area is appropriate for understanding the extent of area affected by the shaft excavation.
In the groundwater flow simulation using the hydrogeological model, the horizontal distribution of the drawdown shows a concentric pattern with the shaft as its center. It does not indicate that the Tsukiyoshi Fault forms an impervious zone.
The results of the groundwater flow simulation (2^nd analysis loop) also shows the appropriateness of the size of the study area and setting of the boundary conditions.

6) Future tasks

While the steady-state and non-steady-state simulations allow understanding the overall groundwater hydrology in the study area, none of the investigation results obtained in the area is used for these simulations. The following are extracted as tasks to be dealt with in the future.
Method of setting the geological units for geological model
extraction of geological units taking the heterogeneity of granite into consideration
Varieties, quality and quantity of information needed for modeling and groundwater flow simulation
understanding of hydrogeological properties of individual geological units and prediction of 3D distribution of hydraulic conductivity
Methodology for construction of the hydrogeological model
method for estimating 3D distribution of hydraulic conductivity
Methodology of groundwater flow simulation
method for saturated/unsaturated simulation and the applicability of finite element method and finite difference method
Method of setting hydraulic boundary conditions
re-examination of setting the top and side boundary conditions on the basis of the accumulated data
Assessment of the uncertainty contained in obtained data, models and the results of the groundwater flow simulation
factors which contribute to the reduction of the uncertainty, and the priority of the data acquisition.