Introduction

The Grimsel Test Site (GTS) is located south of Guttannen in the upper 'Haslital' in Canton Bern, Switzerland. Geologically, it is situated in the Central Swiss Alps within the Aar Massif. The Aar Massif belongs to the External Crystalline Massifs of the Alps (e.g. von Raumer et al. 1993), forming the basement of the Alpine orogeny. However, its geological record is older than the Alpine orogenesis (e.g. Mercolli et al. 1994). This chapter gives a geological overview of the area surrounding the GTS.

The upper 'Haslital' (Fig. 1.1, Fig. 2.1 and Fig. 2.2) is composed of the following rock units of different geological age: (i) The Guttannen Gneiss Complex, the Ofenhorn-Stampfhorn Gneiss Complex, the Bäregg Gneiss Complex, the Gärstenhorn Gneiss Complex, and the Massa Gneiss Complex form the Pre-Variscan polycyclic metamorphic basement, often referred to as 'Altkristallin' in the German literature (e.g. Abrecht 1994, Labhart 1977, Stalder 1964); (ii) Rocks of the Diechtergletscher Formation and the Trift Formation are Late to post-Variscan volcaniclastic rocks (Berger et al. 2017b); (iii) The Mittagsflue Granite, the Central Aar Granite, the Grimsel Granodiorite, and the Aplitic Boundary Facies are Late to post-Variscan plutonic rocks (e.g. Keusen et al. 1989, Schaltegger and Corfu 1992, Schaltegger 1990). The nomenclature of the rock types mentioned varies in the literature; here we follow the nomenclature proposed by Berger et al. (2017a).

In the following, the geological record of the rock types mentioned will be summarised based on key literature. For further details, the reader is referred to the compilation of Berger et al. (2017b) and the references therein.

Fig. 2.1: Geological overview map of the upper 'Haslital' (modified after Berger et al. 2017a)

Fig. 2.2: Cross-section along the 'Haslital' (modified after Schneeberger et al. 2016) The thrusts and faults are not shown across the entire section as the extrapolation towards depth is increasingly uncertain.

2.1 Pre-Variscan rock units

Volumetrically, the pre-Variscan polycyclic basement represents the dominant rock types within the Aar Massif (e.g. Berger et al. 2017b). They are termed polycyclic due to their complex tectono-metamorphic history.

The pre-Variscan basement rocks are schists and banded gneisses characterised by a strong metamorphic overprint reaching amphibolite facies conditions in the Ofenhorn-Stampfhorn Gneiss Complex (Abrecht 1994, Berger et al. 2017b, Schenker and Abrecht 1987). Locally, migmatic structures occur as relics of the strong overprint.

2.2 Pre-Variscan evolution

The pre-Carboniferous evolution of the area of interest is better recorded in the neighbouring Gotthard Nappe than in the Aar Massif. Relatively little evidence is preserved in the Aar Massif, however what evidence there is indicates that the evolution of the Aar Massif was linked to the evolution of the Gotthard Nappe. This evolution consists of an early high-pressure eclogitic metamorphic event (Oberli et al. 1994) followed by high-temperature metamorphism (upper amphibolite facies to granulite facies) during Middle Ordovician times, leading to anatexis (Berger et al. 2017b). The Ordovician metamorphism was probably caused by subduction and collisional tectonics (Schaltegger 1994). Rapid uplift could be responsible for the observed partial melting (Biino 1994). The Erstfeld Gneiss Complex, outcropping north of the Guttannen Gneiss Complex, bears evidence of amphibolite facies metamorphic overprint, however no relics of an earlier granulite facies metamorphism reported in the Gotthard Nappe are preserved within the Aar Massif (Berger et al. 2017b, Schaltegger 1994). The Ofenhorn-Stampfhorn Gneiss Complex shows evidence of Ordovician amphibolite facies metamorphism with subsequent formation of migmatites (Schenker and Abrecht 1987).

The pre-Variscan polycyclic metamorphic basement also records the Carboniferous and Permian evolution of the Central Alps (Berger et al. 2017b), forming a sequence of SW-NE oriented structures, indicating a north-verging transport direction. This SW-NE trending architecture is proposed to result from Carboniferous nappe tectonics related to the Variscan orogeny (Oberhänsli et al. 1988).

2.3 Late to post-Variscan metasedimentary and volcaniclastic rocks

Late to post-Variscan metasedimentary and volcaniclastic formations are mainly composed of thin, steeply dipping discontinuous bands that strike parallel to the Aar Massif (Fig. 2.1, Berger et al. 2017b). Rocks outcropping within the area of interest belong to the Trift Formation or the Diechtergletscher Formation (Fig. 2.1) and are characterised by rhyolite, epiclastic sediments, tuff and coarse pyroclastic rocks and volcanic breccia (Berger et al. 2017b, Oberhänsli et al. 1988).

2.4 Late to post-Variscan rocks

The outcropping Mittagsflue Granite, Central Aar Granite, Grimsel Granodiorite and the Aplitic Boundary Facies are part of the Haslital Group (Berger et al. 2017b). The Haslital Group belongs to a calc-alkaline to sub-alkaline granitoid suite, with the following differentiation sequence (from most basic to most acidic): Grimsel Granodiorite, Central Aar Granite, Mittagsflue Granite, Aplitic Boundary Facies (Schaltegger 1990, 1994). Different radiometric age datings overlap within error and occurrence of numerous intermingling structures ('Schlieren') indicates a gradual change between the Grimsel Granodiorite (GrGr) and the Central Aar Granite (CAGr) (Schneeberger et al. 2016), corroborating the coeval viscous state. The Haslital plutonic rocks are of Asselian age (299–295 Ma, Schaltegger and Corfu 1992, Schaltegger 1994).

2.5 Variscan evolution

The onset of the Variscan collision orogeny is of Visean age (ca. 346 to 331 Ma, von Raumer et al. 2013), leading to granulite facies metamorphism conditions (Berger et al. 2017b, von Raumer et al. 2013). The granulite facies metamorphism is mostly preserved within the Lauterbrunnen- Innertkirchen Gneiss Complex, which is located north of the study area (north of the map in Fig. 2.1).

During Pennsylvanian times (ca. 323 to ca. 299 Ma), the polycyclic basement was exposed, as rocks of the Diechtergletscher Formation were deposited. Later these rocks were tilted and rapidly transported to 5 – 10 km depth, which corresponds to the intrusion depth of the Central Aar Granite (Berger et al. 2017b). During Asselian (299 to 295 Ma) times, the Haslital Group intruded into the polycyclic metamorphic basement and the volcaniclastic sedimentary rock units. The Haslital Group granitoids are considered to be post-metamorphic and post-tectonic and to be intruded under an extensional tectonic setting (Schaltegger 1990, Schaltegger and Corfu 1992) above a subduction zone at an Andean-type continental margin (Mercolli and Oberhänsli 1988). Variscan metamorphic overprint did not exceed greenschist facies conditions in the area of interest, whereas the southern units of the Aar Massif bear evidence of a metamorphic overprint reaching amphibolite facies conditions (Schaltegger 1994).

2.6 Alpine evolution

2.6.1 Regional tectono-metamorphic evolution

The rock volume of the upper 'Haslital' is part of the inverted European basement of the Alps and thus paleogeographically part of the Helvetic realm (Herwegh et al. 2017, Pfiffner 1993, Schmid et al. 1996). The Alpine orogeny, resulting from the closure of the European Tethys (ocean) followed by continental collision (e.g. Schmid et al. 2004), is recorded as a metamorphic overprint and as a deformational stage within the rock volume of the Grimsel area (e.g. Steck 1968). Fig. 2.3 shows an overview cross-section through the Central Alps illustrating the present-day situation resulting from the Alpine orogeny.

Fig. 2.3: Cross section through the Central Swiss Alps (modified after Herwegh et al. 2017)

The regional metamorphic overprint reached upper greenschist facies conditions south of the Central Aar Granite, whereas the Guttannen Gneiss Complex shows evidence of lower greenschist facies Alpine metamorphism (Berger et al. 2017b, Frey et al. 1980, Niggli and Niggli 1965). Metamorphic conditions were estimated for the rocks in the Lake Rätrichsboden area to be around 450 °C and 6.5 kbar (Challandes et al. 2008, Goncalves et al. 2012). Bergemann et al. (2017) indicates that the pressure-temperature estimates are maximum estimates and probably overestimate the actual peak metamorphic conditions.

The long-lasting Alpine tectonic evolution of the area has been studied extensively and there is still controversy regarding the number of deformation phases (e.g. Choukroune and Gapais 1983, Herwegh et al. 2017, Rolland et al. 2009, Steck 1968, Wehrens et al. 2016, 2017).

Fig. 2.4: Structural map of the surface area above the GTS (Schneeberger et al. 2017) Structural grouping shown in this figure refers to Tab. 2.1.

Fig. 2.5: Block diagrams showing increasing deformation and strain localisation (modified after Choukroune and Gapais 1983 and Wehrens 2015)

The Alpine solid-state deformation history is best depicted in the post- to Late Variscan plutonites of the Haslital Group, as they are considered to be monometamorphic. Large-scale crustal blocks that are crosscut by kilometre-scale post-collision steep faults form the rock in the 'Haslital' (Fig. 2.4 and Fig. 2.5). The number of faults, the intensity of deformation and the metamorphic grade increase southwards (e.g. Niggli and Niggli 1965, Steck 1968, Wehrens et al. 2017). The ductile deformation resulted in a pervasive foliation and localised high strain shear zones. This results in a pattern of less deformed rock matrix and highly deformed shear zones (Fig. 2.5, Choukroune and Gapais 1983).

The large-scale faults occur as different sets (Fig. 2.4):

• a NE-SW trending fault set (fault set 1, Tab. 2.1), that is a conjugate set of reverse and normal faults with dominant movement south block up.
• a NW-SE and E-W trending predominantly strike-slip fault set (fault set 2).
• the previously mentioned fault sets are crosscut by thrust-related faults that are moderately SE and SW dipping (fault set 3).

Tab. 2.1: Summary of deformation phase models for the Grimsel area

Based on newly acquired structural data and a strict separation between brittle and ductile deformation regimes, a novel grouping of the occurring structures is suggested. The first fault set evolved during peak metamorphic conditions (Fig. 2.6, Herwegh et al. 2017, Wehrens et al. 2017). Radiometric age determination of newly formed sheet silicates suggests an activity between 22–17 Ma (Rolland et al. 2009). Members of this fault set were formed as ductile shear zones, localised high strain zones, within the stability field of biotite (> 400°C). These faults correspond to the Stage 1 faults of Rolland et al. (2009) and the Handegg phase faults of Wehrens et al. (2017). The steeply dipping faults show very little evidence of vertical movement (vertical component of offset typically > 0.5–50 m). However, due to the large number of faults, the cumulative vertical movement adds up to several kilometres of reverse faulting movement for the Grimsel Pass area (Herwegh et al. 2017).

The activity of the second fault set started 14–12 Ma (Fig. 2.6, Rolland et al. 2009) and evolved under lower greenschist facies conditions (Wehrens et al. 2017). The change from the first thrustrelated fault set to the second strike-slip-dominated fault set is gradual (Herwegh et al. 2017, Bergemann et al. 2017), as evidenced by observed oblique stretching lineation. This change from thrust-related faults to strike-slip faults implies a change in the orientation of the principal stress axes.

Fig. 2.6: Alpine Pressure-Temperature-time (PTt) path (modified after Diamond and Tarantola 2015 and Wehrens 2015)

Several kinematic models have been proposed for the Alpine evolution of the 'Haslital' and are part of an ongoing debate. They differ mainly between single-phase (Choukroune and Gapais 1983, Gapais et al. 1987) and multi-phase models (Steck 1968, Herwegh et al. 2017, Rolland et al. 2009, Wehrens et al. 2017). As indicated by Wehrens et al. (2017), the single-phase models are based on observations from the northern part of the CAGr and evidence from the southern GrGr-dominated area was not considered.

Geodynamically, the evolution of the Aar Massif sheds light on the evolution of the Alpine orogeny and is summarised in Fig. 2.7 (Herwegh et al. 2017). Alpine subduction ends with the slab breakoff at about 34–32 Ma, slab-pulling forces are reduced, and the European plate is thus decelerated. The remaining weight of the European mantle lithosphere promotes slowed rollback some 30 Ma ago. The Aar Massif delaminates some 22–17 Ma ago from its lower crust, since tectonics are dominated by sub-vertical extrusion (fault set 1, Fig. 2.6). Severe buoyancy forces due to an over-thickened crust induce this sub-vertical extrusion. Despite predominant delamination and buoyancy processes during post-collision, a compressional component persists and becomes more important after 12 Ma, as documented by the formation of strike-slip-dominated faults in the central part of the Aar Massif (fault set 2, Fig. 2.6) and by thrust-related faults in the northern part of the Aar Massif (fault set 3).

Fig. 2.7: Geodynamic evolution of the Aar Massif since Oligocene times (modified after Herwegh et al. 2017)

Aar: Aar Massif, Got: Gotthard Nappe, AAT: Ausserberg-Avat-Tavestch zone, TALP: Tectonically accreted lower plate, GPSZ: Grimsel Pass Shear Zone

2.6.2 Uplift and exhumation

The extrusion inferred from geodynamic considerations (Herwegh et al. 2017) led to a decrease in temperature and therefore to a gradual change in deformation style. The first fault set evolves under ductile regime deformation in the stability field of biotite (> 400°C). These ductile shear zones are later overprinted by ductile shear zones without biotite, thus formed under slightly lower temperature deformation conditions (Wehrens et al. 2017). With decreasing temperature, the rock mass crosses the brittle-ductile transition zone and brittle deformation reactivates former ductile shear zones (Belgrano et al. 2016, Egli et al. 2018, Hofmann et al. 2004, Kralik et al. 1992, Schneeberger et al. 2016, Wehrens 2015, Wehrens et al. 2016, 2017). Evidence of cyclic embrittlement was reported from the Grimsel area (Wehrens et al. 2016).

The exhumation history of the Aar Massif is reported by diverse authors (Glotzbach et al. 2010, 2011, Michalski and Soom 1990, Pleuger et al. 2012, Reinecker et al. 2008, Valla et al. 2012, 2016, Vernon et al. 2009, Weisenberger et al. 2012). The Miocene to Pliocene exhumation history of the Central Swiss Alps is generally described as exhumation with a constant rate (0.3 to 0.5 mm/yr) since 10 to 14 Ma and a period of increased exhumation rate (ca. 0.7 mm/yr) between 10 to 8 Ma (Glotzbach et al. 2010, Valla et al. 2012, Vernon et al. 2009 and Weisenberger et al. 2012). Michalski and Soom (1990) report slightly higher exhumation rates, however, and used a higher geothermal gradient of ca. 30°C/km, whereas studies since Vernon et al. (2008) use 25°C/km as the local geothermal gradient. Valla et al. (2012) indicate higher exhumation rates for the time period 10–8 Ma with 2–4 mm/yr, whereas Vernon et al. (2009) discuss the possibility of a constant exhumation rate of 0.4 mm/yr for the last 10 Ma based on modelling.

Valla et al. (2016) distinguished three distinct exhumation events since the Late Neogene based on thermal modelling: (i) Late Miocene tectonically driven exhumation, (ii) Pliocene climatically induced regional exhumation pulse, and (iii) Quaternary exhumation event in response to glacial valley carving since ca. 1 Ma. Mey et al. (2016) present an alternative landscape evolution driver with glacial isostatic adjustment in response to the Last Glacial Maximum (LGM) ice cover being responsible for the denudation. For further details on the debate the reader is referred to Mey et al. (2016) and references therein. From GPS data, a modern-day vertical uplift of ca. 2 mm/yr has been suggested (Sanchez et al. 2018). This is substantially higher than the recent vertical crustal movement inferred from precision levelling for the Grimsel area in relation to a reference point in Laufenburg of ca. 1 mm/yr (Nagra 2008). It has to be noted that the two surveys employ different reference levels.

Uplift and exhumation related to glacial activity (as discussed below) resulted in the formation of so-called exfoliation joints, which occur within the uppermost 200 m of rock volume (Ziegler et al. 2013). Three distinct exfoliation joint types are reported for the Grimsel area: (i) closely spaced (< 1 m) joints that are parallel to today's topography, (ii) moderately spaced (0.6–2 m) exfoliation joints subparallel to today's topography and (iii) widely spaced (>> 2 m) joints that are not parallel to the ground surface but could be controlled by former topographies. Joint spacing inferred from mapping was confirmed on a larger scale based on photogrammetric models (Ferrari et al. 2018). Four exfoliation joint generations were identified: Early Pleistocene (ca. 1.5–1 Ma), Middle Pleistocene (ca. 0.7–0.4 Ma), Late Pleistocene (0.1–0.02 Ma) and Late Glacial/Holocene (< 0.02 Ma).

The latest step of the landscape evolution in the Grimsel area is related to glacial activity (Kelly et al. 2006). Multiple glacial stages throughout the Pleistocene shaped the valley morphology, which is likely to be pre-conditioned by the tectonic setting. Ice coverage was a least 600 m thick on the Grimsel Pass, as evidenced by trimlines and bedrock carving (Florineth and Schlüchter 1998). Recent modelling suggests even greater glacier thicknesses (Seguimot et al. 2018). The highest ice surface during the LGM was attained at 23.0 ± 0.8 ka in the upper 'Haslital' (Wirsig et al. 2016). It was followed by significant retreat not later than 17.7 ± 0.8 ka (Wirsig et al. 2016). Post-glacial tectonic faulting was observed SE of the Grimsel Pass (Ustaszewski and Pfiffner 2008). The fault is characterised by a 7 m crest displacement resulting in a tectonic slip rate of about 0.4 mm/yr since the LGM.

2.7 Regional hydrological setting

The regional hydrological setting has been characterised based on hydraulic modelling validated against measured hydraulic heads (Herzog 1989, Voborny et al. 1991). Generally, principal infiltration areas for hydraulic heads measured at the GTS are the Alplistock and the Juchlistock with discharge in the Aare valley (Lake Grimsel or Lake Rätrichsboden; Fig. 2.8). These "undisturbed" conditions are altered by the presence of the underground facilities (GTS and main access tunnel of KWO, Voborny et al. 1991). Also, at the surface the water has no natural flow due to the numerous engineered systems of KWO (e.g. dams and water tunnels). The GTS and the main KWO access tunnel act as sinks, but with little water inflow due to the low hydraulic conductivity. The occurrence of these sinks leads to high hydraulic gradients in the western part of the GTS (Fig. 2.9). The labelled black lines in Fig. 2.9 represent equipotential lines where their spacing indicates the hydraulic head gradient. The Bächli zone that forms the Bächli valley (Bächlital, Fig. 1.1) is of regional importance and drains part of the Juchlistock/Alplistock catchment into Lake Rätrichsboden (Voborny et al. 1991).

Fig. 2.8: Map showing the equipotential lines for hydraulic heads with regional brittle
discontinuity network (black lines) calculated on a horizontal section at the elevation
of the GTS (modified after Voborny et al. 1991)

The map is based on model calculations containing the local topography, measured
transmissivity values for rock matrix and faults, and measured hydraulic heads at the GTS.

Fig. 2.9: Cross-section through the GTS showing the regional distribution of hydraulic
pressure in metres above the GTS level (modified after Frick et al. 1992)

The cross-section is based on model calculations containing the topography, faults and
measured hydraulic heads at the GTS.