GDR 00.1-001 January 2014

Home , Shear vane test, Void ratio

Ground Investigation Report

January 2014 GDR 00.1-001

Femern A/S Vester Søgade 10 1601 København V

Tel.: +45 3341 6300 Fax.: +45 3341 6301 www.femern.com CVR no. 28 98 65 64

Prepared by

Rambøll Arup Joint Venture c/o Rambøll Danmark A/S Hannemanns Allé 53 DK-2300 Copenhagen S Danmark

Phone +45 51611000 Rambøll Arup Joint Venture Danish reg. no: CVR-NR 31749077 Member of FRI

This report is based on the geological/geotechnical knowledge gathered and evaluated by Femern A/S until January 2014.

Prepared JRF, GLH, NLSM, 2014-01-31 CH, UTN Checked PM 2014-01-31

Approved JRF 2014-01-31

Table of Contents

1 Introduction...... 9

2 Executive Summary ...... 12 2.1 Activities ...... 12 2.1.1 Seismic surveys ...... 12 2.1.2 Boring campaigns ...... 12 2.1.3 Geophysical borehole logging ...... 12 2.1.4 Advanced laboratory testing ...... 13 2.1.5 Large scale testing ...... 13 2.1.6 Seismicity ...... 13 2.2 Investigation Results ...... 14 2.2.1 The Quaternary ...... 14 2.2.2 The Palaeogene ...... 15 2.2.3 The Cretaceous ...... 15 2.2.4 Glacial tectonics...... 16 2.2.5 Salt tectonics ...... 16 2.2.6 Geotechnical parameters ...... 16 2.3 Positioning System ...... 17

3 Positioning System ...... 18 3.1 General ...... 18 3.2 Equipment and Procedures ...... 18

4 Offshore Geophysical Investigations ...... 20 4.1 General ...... 20 4.2 Equipment and Procedures ...... 21

5 Onshore Geophysical Investigations ...... 22 5.1 General ...... 22 5.2 Equipment and Procedures ...... 22 5.2.1 Onshore Seismic ...... 22 5.2.2 Continuous Vertical Electrical Sounding (CVES) ...... 22

6 Borings and Cone Penetration Tests (CPTUs) ...... 24 6.1 General ...... 24 6.2 Offshore Type C Borings, Fehmarnbelt ...... 24 6.2.1 General ...... 24 6.2.2 Equipment and Procedures ...... 25 6.3 Onshore Type A and Type B Borings, Lolland and Fehmarn ...... 25 6.3.1 General ...... 25 6.3.2 Equipment and Procedures ...... 25 6.4 Offshore Type A and Type B Borings, Fehmarnbelt ...... 26 6.4.1 General ...... 26 6.4.2 Equipment and Procedures ...... 26

6.5 Sample Handling and Laboratory Classification Works ...... 27 6.5.1 General ...... 27 6.5.2 Equipment and Procedures ...... 27 6.6 Correlation Borings Lillebælt ...... 28 6.7 Correlation Borings Fehmarnsund ...... 28

7 Geophysical Borehole Logging ...... 29 7.1 General ...... 29 7.2 Equipment and Procedures ...... 29

8 Advanced Laboratory Testing ...... 31 8.1 General ...... 31 8.2 Equipment and Procedures ...... 32 8.2.1 General ...... 32 8.2.2 Oedometer Testing...... 32 8.2.3 Triaxial Testing ...... 34 8.2.4 Geological Dating of Samples ...... 35

9 Large Scale Testing ...... 36 9.1 General ...... 36 9.2 Equipment and Procedures ...... 37 9.2.1 Trial Excavation Works and Surveys ...... 37 9.2.2 Instrumentation and Monitoring ...... 37 9.2.3 CPTUs and Block Sampling ...... 38 9.2.4 Plate Load Testing ...... 38 9.2.5 Pile Installation and Tension Load Testing ...... 39 9.2.6 Ground Anchor Installation and Tension Load Testing ...... 40

10 Ground Conditions ...... 42 10.1 Geologic Setting ...... 42 10.1.1 Morphology and Seismic Stratigraphy ...... 42 10.2 Salt Pillow Structures ...... 43 10.3 Seismicity ...... 45 10.4 Ground Water Conditions ...... 46 10.4.1 Lolland ...... 46 10.4.2 Fehmarn ...... 47 10.5 The Series of Layers ...... 48

11 Postglacial and Lateglacial Deposits ...... 49 11.1 Geological Description ...... 49 11.2 Geotechnical Properties ...... 50 11.2.1 General ...... 50 11.2.2 Classification Properties ...... 50 11.2.3 CPTU ...... 51 11.2.4 Stress and Stress History ...... 52 11.2.5 Consolidation Properties ...... 52 11.2.6 Static Shear Strength...... 53 11.2.7 Geophysical Properties ...... 53

12 Glacial Deposits ...... 54 12.1 Geological Description ...... 54 12.2 Geotechnical Properties ...... 55 12.2.1 General ...... 55 12.2.2 Classification Properties ...... 55 12.2.3 CPTU ...... 58 12.2.4 Stress and Stress History ...... 58 12.2.5 Consolidation Properties ...... 59 12.2.6 Static Shear Strength...... 60 12.2.7 Geophysical Properties ...... 62 12.2.8 Small Strain Stiffness and Damping ...... 62 12.2.9 Cyclic Undrained Shear Strength ...... 63 12.3 Recompacted Clay Till Material as Embankment Fill ...... 63 12.3.1 Introduction...... 63 12.3.2 Compaction and Strength Properties of Recompacted Upper Till and Lower Till Deposits ...... 63 12.3.3 Overall Existing Experience ...... 65

13 Clays of Palaeogene Origin ...... 66 13.1 Geological Description ...... 66 13.2 Geotechnical Properties ...... 68 13.2.1 General ...... 68 13.2.2 Classification Properties ...... 69 13.2.3 CPTU ...... 71 13.2.4 Stress and Stress History ...... 74 13.2.5 Consolidation Properties ...... 76 13.2.6 Static Shear Strength...... 78 13.2.7 Geophysical Properties ...... 80 13.2.8 Small Strain Stiffness and Damping ...... 82 13.2.9 Cyclic Undrained Shear Strength ...... 82 13.3 Special Investigations and Evaluations ...... 83 13.4 Large Scale Properties ...... 84 13.4.1 Instrumented Trial Excavation ...... 84 13.4.2 Plate Load Tests ...... 88 13.4.3 Pile and Ground Anchor Tests ...... 90

14 Cretaceous Chalk ...... 98 14.1 Geological Description ...... 98 14.2 Geotechnical Properties ...... 98 14.2.1 General ...... 98 14.2.2 Classification Properties ...... 98 14.2.3 CPTU ...... 99 14.2.4 Stress and Stress History ...... 100 14.2.5 Consolidation Properties ...... 100 14.2.6 Static Shear Strength...... 100 14.2.7 Small Strain Stiffness and Damping ...... 102

15 References...... 103

List of Enclosures

I. Drawing no. 070-02-12: Road Alignment, Combined geotechnical plan and longitudinal section Drawing no. 070-02-13: Rail Alignment, Combined geotechnical plan and longitudinal section

II. List of all Geotechnical Data Reports with Data and Information, Category Number and Comments

List of Appendices (included in a separate binder)

Appendix GDR 00.1-001-A Geological Interpretations of the 1996/2009/2010/2011/2012/2013 borings January 2014

Appendix GDR 00.1-001-B Geotechnical Properties for Postglacial and Lateglacial Deposits January 2014

Appendix GDR 00.1-001-C Geotechnical Properties for Glacial Deposits January 2014

Appendix GDR 00.1-001-D Geotechnical Properties for Clays of Palaeogene Origin January 2014

Appendix GDR 00.1-001-E Geotechnical Properties for Cretaceous Chalk January 2014

Appendix GDR 00.1-001-F Geotechnical Properties for Clays of Palaeogene Origin at Lillebælt, Fehmarnsund and Puttgarden Harbour January 2014

Appendix GDR 00.1-001-G Large Scale Properties, Palaeogene clay January 2014

1 Introduction

This ground investigation report summarizes the geophysical, geological and geotechnical investigations performed in 1995/96 and in the period from April 2008 through January 2014 for the future Fehmarnbelt Fixed Link. The design solution (bridge or immersed tunnel) for the fixed link had not been decided before the different investigations were initiated and the performed investigations were thus directed against both a bridge and an immersed tube tunnel solution.

The investigations for the Fehmarnbelt Fixed Link in 2008 through January 2014 included mainly:  Establishment of a positioning reference system by AXIONET mainly to be used during the construction of the fixed link.  Geophysical offshore investigations performed by Rambøll Arup Joint Venture (RA) comprising marine shallow seismic investigations, marine side scan investigations, marine magnetic measurements and bathymetric measurements. The main survey was performed in 2008. In 2009 an additional bathymetric survey was carried out covering a wide area on both sides of the Rødbyhavn to Puttgarden ferry route; a supplementary survey in a smaller area located west of the 2008-area using similar tools as used for the 2008 survey was also carried out.  Geophysical onshore investigations performed in 2008 and supplemented in 2012 by RA comprising onshore reflection seismic surveys and Continuous Vertical Electrical Sounding (CVES) measurements.  Onshore and offshore geotechnical type A-borings for sampling and geotechnical type B-borings for down-the-hole CPTUs performed by Fugro.  Down-hole geophysical borehole logging performed by RA.  Seabed CPTUs (type C-borings) performed by Fugro.  Onshore Fehmarn CPTU campaign performed by Fugro.  Onshore and offshore geotechnical type A-borings for investigation of the ground conditions in the production site near Rødbyhavn and onshore geotechnical type B- borings with CPTU and sampling for the alignment on Lolland and Fehmarn performed by Aarsleff/GEO. Onshore CPTUs for the production site performed by Ramboll Sweden. Offshore CPTUs for the large scale testing site performed by Aarsleff/GEO.  Offshore CPTUs in the Large Scale Testing Site performed by Aarsleff/GEO  Advanced Laboratory Testing performed by GEO/Deltares/NGI/Fugro on samples recovered during the Boring Campaigns.  Geological (micropalaeontologic) dating of selected samples performed by GEUS.  Large Scale Testing including mainly a phased offshore trial excavation with advanced instrumentation, different kinds of plate load testing, pile installation and tension load testing as well as onshore installation and load testing of ground anchors all in clays of Palaeogene origin performed by Aarsleff/GEO.  Establishment of an overall geological model and detailed description of the ground conditions, focusing on each of the encountered geological deposit. This part of the work also includes an assessment of the groundwater conditions, of the dome structures and of the seismicity in Fehmarnbelt as treated in separate Geotechnical Data Reports, prepared by RA.

It is the objective that the data and test results from the combined geophysical, geological and geotechnical investigations will enable an integrated approach to the interpretation of ground conditions and derivation of geotechnical properties covering the specified con- cepts for the fixed link between Germany and Denmark.

The investigations undertaken from April 2008 to January 2014 have focused on an area approximately 2 km by 27 km situated east of the ferry route between Rødbyhavn in Denmark and Puttgarden in Germany.

This report includes the following:  An executive summary is included in Chapter 2;  The investigations from 2008 through January 2014 are summarised in Chapters 3 through 9; and  The main findings of the investigations are summarized in Chapters 10 through 14.

More detailed information about the main findings can be found in the Appendices GDR 00.1-001-A through GDR 00.1-001-G. Enclosure II contains a list of all the GDRs that have been established and which are available in the Geo Information System, /22/.

In this Ground Investigation Report and in the Geo Information System the term “deposit” is always used when one single geological layer is described, while the term “unit” is used when more than one deposit is included within the same name/description – unless specifically otherwise stated.

The combined geotechnical plan and longitudinal section for the road and rail alignment respectively is presented in Enclosure I. Selected borings and CPTUs appear from these drawings with the boundaries in the soil columns being restructured in accordance with the geological interpretation of the borings as stated in Appendix GDR 00.1-001-A.

The distribution of layers as included in the Geo Information System as digital surfaces represents the geological model for the area considered.

A complete list of the GDRs produced within the Femern/RA Geotechnical Services Agreement and including the report for the geological/geotechnical investigations 1995/96 is attached as Enclosure II. In this enclosure the Data and Information, Category number of each report as being “2” (i.e. limited liability) or “3” (i.e. no liability) has been indicated as well as comments regarding possible old and not updated information (such as coordinate system) included in the actual GDR.

Measured, corrected, correlated and derived values of geotechnical properties are presented in the present report. The derived value is defined as the value of a geotechnical para- meter obtained by theory, correlation or empiricism from test results.

Derived values have been obtained in accordance with ref. /10/ and /11/.

Typical values of the geotechnical properties appear from GDR 01.4-004 /59/.

Femern's overall reporting structure appears from Figure 1-1.

Figure 1-1 Overall reporting structure

Unless otherwise stated, coordinates and elevations/levels are in the Femern Coordinate System (FCS) and the Femern vertical datum FCSVR10 respectively.

The geotechnical background material and data are available within Femern's Geo Infor- mation System /22/.

For abbreviations and definitions not separately defined in the present report, reference is made to Femern's Geo Nomenclature, /12/ and to Femern's Legend and definitions, /43/.

2 Executive Summary 2.1 Activities The investigations have included the following activities:

2.1.1 Seismic surveys The deep seismic survey of 1995 delivered an overall picture of the stratigraphy and tectonic situation in the investigated area; this interpretation remains largely valid. A comprehensive shallower survey was performed in 2008, and a supplementary survey in an area immediately west of the 2008-area was performed in 2010. This confirmed and provided more detail on the conclusions arrived at in 1995. However, observations of disturbances in the layering and specific point-reflections from signals interpreted as cobbles in the Palaeogene layers south of the dome created the wrong interpretation that the upper series of layers south of the dome were dominated by floes of Palaeogene clay inter-layered with clay till. During the 2008/09 boring campaigns it soon became clear that the “boulders” observed on the seismograms were concretions in the Palaeogene clay and that the upper 20-40 m of this clay was folded, but not dislocated as floes.

Two further offshore seismic surveys have been performed as part of the 2008-2010 investigations: One was a bathymetric survey performed by RA and the other was a near- shore survey at both coastlines performed by GEUS.

Different kinds of onshore seismic surveys have been performed in 2008 and 2012.

2.1.2 Boring campaigns The interpretation of the seismic surveys provided the seismic ground model allowing selection of locations for borings in the Boring Campaigns of 2009 and 2010, including 36 boring locations offshore in Fehmarnbelt and 12 boring locations onshore. The information obtained from the high quality deep borings has allowed a very significant increase in the understanding of the stratigraphic, tectonic and geotechnical conditions in the area. The different units identified on the seismic profiles have all been recognised as well defined geologic formations/units. In addition, geotechnical investigations have been performed in 2011, 2012 and 2013 for the production site near Rødbyhavn and along the alignment on Lolland and Fehmarn.

2.1.3 Geophysical borehole logging Geophysical borehole logging has been performed in almost all the boreholes established in 2009 and 2010, but challenges with the stability of un-cased boreholes prevented the use of the optical televiewer, that can only be operated in open holes. The original aim to obtain acoustic “pictures” of the borehole wall to document the condition of the chalk and to illustrate strike and dip of the layers in the folded part of the Palaeogene clay layers has unfortunately not been fulfilled. Moreover, the search for characteristic peaks or patterns, especially in the Palaeogene strata, in the profiles for other log tools has not been successful. However, from the logging important information on details of the series of layers has been gained, helping to correlate the upper layers between the boreholes; the logging has also been an important tool to investigate the quality of the boring works.

2.1.4 Advanced laboratory testing The advanced laboratory testing has included pilot and production testing. The objective of the pilot testing was to clarify procedures to be applied during production testing. The objective of the production testing was to establish soil properties as measured in the laboratory using small scale testing with the identified procedures.

A considerable number of advanced laboratory tests including different kinds of oedometer tests, triaxial tests and direct simple shear tests have been performed for the different deposits.

The major part of the advanced tests has been performed for the clays of Palaeogene origin, being the most challenging soil unit. In addition, a considerable number of advanced tests have been performed on the Glacial tills and on the Cretaceous chalk. Compaction testing and triaxial testing on recompacted Glacial tills have also been performed.

Testing to collect information on how the Palaeogene clay behaves when pumped as cuttings in a water/clay suspension through a long pipe has recently been performed.

2.1.5 Large scale testing The large scale testing has been concentrated in an area with folded Palaeogene Røsnæs clay almost from seabed to 60 m depth below seabed level. The site is located approximately 1 km east of Puttgarden ferry harbour and is characterised by a water depth of 10 m before excavation. The work comprised: o Excavation, carried out in 3 phases, with base area 30 m × 70 m from elevation c. -10 m to elevation -20 m; installation of extenso-piezometers from 3 m to 25 m below excavation base; installation of surface benchmarks; and performance of multibeam surveys. Partial backfilling (3 m) and final backfilling (7 m) of the excavation. o Cone penetration tests around the planned groups of test piles. Cone penetration tests and block sampling from the base of the excavation. Cone penetration testing from the top of (the approximately 3 m high) backfilling in the excavation. o Installation of driven steel tube piles and bored cast-in-situ test piles as well as reaction piles and tension load testing of the piles. o Instrumentation monitoring of the extenso-piezometers (P1-P9) and surface benchmarks within the excavation. o Reinstatement of the site.

In addition installation and tension load testing of ground anchors with fixed lengths in Palaeogene clay has been performed onshore at a test site east of Rødbyhavn.

2.1.6 Seismicity A desk study of the seismicity of the area has been conducted. It is concluded that the fixed link will be constructed in an area classified as a very low seismicity area. This means that the provision of the Eurocode 8 on Earthquake Risk does not need to be observed. It is, however, from the study concluded that the peak horizontal acceleration at the soil surface with a return period of 475 years is estimated to be in the range of 0.014 to 0.036 g.

2.2 Investigation Results From a morphologic point of view, the offshore part of the area can be divided into a central basin area with a seabed elevation below c. -24 m and two gently dipping coastal slope areas (Figure 10-1). As can be seen from the following descriptions, this zoning of the site is important as soft Postglacial and Lateglacial deposits are, with a few exceptions, in the offshore part only present within the central basin. Moving from the surface downwards the deposits detected are:

2.2.1 The Quaternary Postglacial marine sand. The deposit is dominated by sorted sand with shells, normally without any content of organic materials at all, but locally with small amounts of gyttja. The sand is encountered in local accumulations dispersed across the area; at least part of the deposit is mobile.

Postglacial marine gyttja and freshwater peat. The soft organic gyttja deposit has been detected in the central part of the basin, where it, according to the borings, is locally up to 6 m thick. Marine gyttja has also been found in a former bay inside the actual coast line of Lolland. Peat, possibly of Allerød age, has also been detected in a few borings below the marine deposits in a small depression in the till surface on the southern coastal slope and in small parts of the former bay east of Rødbyhavn.

Postglacial/Lateglacial deposits from the different marine and freshwater stages of the Baltic are present in the basin below and around the gyttja deposit. These include bodies of well to poorly sorted sand without any signs of organic activity which are presumed to be Lateglacial meltwater deposits, laminated (“varved”) clay/silt supposed to be a Late- glacial meltwater/freshwater deposit and more massive, layered clay/silt deposits of more uncertain age and depositional environment.

Glacial deposits. During the 1995 to 1996 investigation campaign it became clear that the glacial deposits below the area could be divided into an “upper till” and a “lower till” with rather different appearance and geotechnical properties. The 2008 to 2010 campaign has revealed that both of the “tills” include components of tills from more than one glacial event. As there is still a clear difference between the tills included in the original “Upper till” compared to those included in the original “Lower till” it has been decided to keep the original division, but to rename the original “Upper and Lower till” to “Upper and Lower till Units” and only to use the term “till deposit” for the material left by the glaciers in a single glacial event.

Upper till Unit consists of very hard clay till, often with “normal” clay till composition but also often very silty and/or very sandy. Locally it is described as sand till. A number of observations indicate that the Upper till Unit consists of till deposits from two different glacial events, both with almost identical grain size distribution and both being very hard. The Upper till Unit dominates the northern part of the area, and the surface is generally located at or close to seabed level or immediately below topsoil or Postglacial deposits where encountered onshore on Lolland.

Meltwater deposits dominated by sand but locally including significant silt and clay bodies have been detected within the glacial deposits both in the seismic surveys and in a number of borings. They are mainly concentrated in two separated parts of the investigated area. The meltwater deposits are considered to belong to at least two, but probably three or more different glacial events; one being within the lower part of the Upper till Unit and the other being located between the Upper and Lower till Units.

Lower till Unit. This Unit is dominated by medium plasticity clay till, but it also contains clay till of different composition and floes of high plasticity Palaeogene clay and meltwater deposits. The 2009 and 2010 Boring Campaigns have shown that the unit includes at least three different till deposits. The dominating deposit is the medium plasticity Lower till deposit, generally located above the other deposits within the Lower till Unit. The second deposit is a high plasticity clay till, “lowermost till”, which primarily has been found in the southern part of the area. A third deposit belonging to the Lower till Unit is the “chalk till” with an almost white colour and a CaCO3 content of more than 50 %. This chalk till deposit has only been found in rather few borings, but the borings with chalk till is distributed almost over the entire investigated area. It has been observed that it always appears above, and therefore is younger than, the high plasticity clay till, while its age relation to the medium plasticity clay till is more uncertain.

2.2.2 The Palaeogene Palaeogene clay. The Palaeogene clay is located directly below the Quaternary deposits and locally as floes within the Lower till Unit (albeit those floes are correctly described as Quaternary age deposits of Palaeogene clay). The microfossil analyses of a significant number of samples from the borings have shown that all of the Palaeogene clay formations Æbelø, Holmehus, Ølst, Røsnæs and Lillebælt, and maybe also the Søvind marl, are present in the borings. Moreover, it has been found that the uppermost part of the Røsnæs clay Formation, together with the lower part of the Lillebælt Clay Formation, dominates the floes, while the Røsnæs Clay Formation dominates the very important part of the area immediately north of the German coast where Palaeogene clay is folded up and almost exposed at seabed level.

2.2.3 The Cretaceous Cretaceous chalk would have been deeply buried below younger deposits in the area if it had not been lifted up by a rising salt pillow to its present position, located approximately 16 m below seabed in the culmination point. The Chalk is a typical Danish chalk, very muddy and only slightly indurated. The flint content is typically about 5 %, and the flint is present as nodules; often concentrated in layers but not as plates as known from other (younger) limestone deposits in Denmark.

2.2.4 Glacial tectonics A number of different observations indicate that the strata below the area have been heavily disturbed by ice pressure in one or more glaciations in the Quaternary. The uppermost part of the Palaeogene clay has been pressed up in big floes, and most of these floes have been transported against south and west by the glaciers, finally to be left as isolated floes within the till deposits. Floes of Palaeogene clay are especially common in the south-western part of the investigated area.

The upper part of the Palaeogene clay below the till has been pressed up into a giant fold system, and only below elevation -70 m or deeper has the deposit not been affected by glacier pressure. The folding has clearly weakened the folded clay compared to the intact layers below, with the CPTUs showing the exact level for the interface between the folded and the undisturbed clay.

2.2.5 Salt tectonics Two salt pillows are situated at depth below the general area of the site: One is located immediately north of the midpoint of the investigated area in Fehmarnbelt. The second salt pillow has its culmination point on land north of Rødbyhavn but its southern flank is within the area of interest for the fixed link. Salt pillows are formed by salt deposited in an 800 to 1000 m thick layer and covered by kilometre thick layers of younger deposits. Under high overburden pressure the salt behaves in a plastic fashion and, as its density is lower than the density of the overlying heavily consolidated clay and sand deposits, seeks an upwards path forming local salt pillows (thickenings). Concern has been directed both to the up-doming movement and the associated downwards movement in the area surrounding the dome and to the fact that the process must lead to stretching in the upper layers of chalk which might weaken the chalk and form active faults.

It is concluded that the upwards movement of the central part of the pillow and downwards movement in an area surrounding the culmination point is likely still to be on- going, but only at a maximum average rate of between 0.05 and 0.5 mm/year in the culmination point.

2.2.6 Geotechnical parameters The key to the geotechnical parameters is provide by: 1. The combined plan and longitudinal sections included as Enclosure I (simplified geological profile appears from Figure 2-1). 2. The soil type. 3. The CPTU net cone resistance.

It has been decided that the choice of geotechnical parameters for the deposits relevant for excavation/foundation of structures shall be based on the combined results of CPTU tests performed in the borings, from advanced laboratory tests and from plate load tests conducted as part of the Large Scale Testing.

Figure 2-1 Simplified geological profile across Fehmarnbelt

2.3 Positioning System In both the planning phase and in the construction phase it is important to be able to accurately and unambiguously coordinate and position objects. It is also necessary to have means of obtaining accurate coordinates anywhere in the project area. To fulfil these requirements, a project related positioning reference system, the Fehmarnbelt Coordinate System (FCS) has been established, together with a positioning reference frame (the physical realisation of the reference System).

3 Positioning System 3.1 General For the main part of the field investigations performed during the years 2008 to 2010 the UTM zone 32 N with EUREF89 datum was used as plane coordinate system during survey operations. In addition, geographical coordinates in WGS 84 system were also used. Elevations and levels have been referred to DVR90.

Very early in the planning phase it was realised that the ability to accurately and unambiguously coordinate and position objects is a key requirement; a single ‘coordinate system’ must therefore be adopted for the project. This system could be either the Ger- man or the Danish national systems or it could be a project specific system. Given the very tight requirements for reliability for such a project, the project related system was chosen.

Late in 2010 Femern's own project related coordinate system (FCS) was ready for use (/40/ and /63/). The system included both a Positioning System and a Positioning Frame. It was developed by RA in close collaboration with KMS, BKG and DTU Space and with Axionet as a Contractor for the Positioning Frame.

Programs for conversion between the systems have been developed, and in the Geo In- formation System (/22/) the new coordinates have been added where relevant to the coordinate related information. 3.2 Equipment and Procedures The project specific geodetic infrastructure established for project comprises:  A Reference System (coordinate system and vertical reference definition),  A Reference Frame (the physical realisation of the Reference System) and  A Positioning Service (RTK service) providing high precision positioning and navigation.

The reference system of the FCS is defined as the International Terrestrial Reference System which is also used for GNSS.

The Femern Map Projection (FMP) is defined to be used within the construction area for the fixed link. The height system, FCSVR10, is defined as Mean Sea Level at Rødbyhavn in 2010. The defining stations have been connected to the tide gauge by levelling, using the 1989 hydrostatic levelling between Marienleuchte and Rødbyhavn for the connection across Fehmarnbelt. A geoid model has been developed and fitted to the FCSVR10 by levelled heights of the GNSS stations, and to the ITRF2005 by ellipsoidal heights of the GNSS stations.

Software has been developed for conversion of coordinates between the FCS, the German and Danish realisations of the ETRS89, as well as for conversion of ellipsoidal and ortho- metric heights within the FCS.

Lastly a continuous positioning service has been established. This is an RTK GNSS service based on GPS and GLONASS with the option of including Galileo if this system becomes operational. The service transmits RTK network data by UHF radio. The estimated obtainable position accuracy (RMS) with the FBPS RTK service is 8 mm in the horizontal and 15 mm in the vertical direction for static applications under normal operating conditions. If one (or more) of the FBPS GNSS stations for any reason lose the communication line with the control centre, the station will automatically switch to a single station RTK service. The estimated obtainable accuracy (RMS) with this backup service is 12 mm in the horizontal and 17 mm in the vertical direction for static applications under normal operating conditions.

The combination of the 3D GNSS-based reference frame, the vertical reference frame and the geoid model are referred to as the reference frame. It is realised physically by determination of coordinates of four permanent GNNS stations (2 on Lolland and 2 on Fehmarn) and a number of 3D control points.

All details regarding reference system, map projection, geoid model, coordinate transformation and RTK service can be found in /40/ and /55/.

4 Offshore Geophysical Investigations 4.1 General In the 1995/96 investigations, a comprehensive survey including both shallow and deep reflection seismic tools was performed (/13/). The idea was that major, deep seated structures, would be more easily observed on the deeper profiles. Once the deep seated structures had been identified it would be easier to observe and understand their conti- nuation through the upper layers from the shallow seismic surveys.

As part of the 2008-2010 investigations, surveys in the offshore area have been performed by RA in three separated campaigns. In addition, GEUS, in 2009, performed a near-shore survey along both the German and the Danish coasts.

The first and largest of the seismic surveys was the 2008 survey. This included both an offshore and a near-shore campaign. The offshore survey covered the area with a water depth of 5 m and more in a 2 km wide corridor; the vessel R/V Madoc was used as platform. Data acquisition was performed along lines extending north-south and parallel with the centreline of the investigated area. The original distance between the survey lines was 50 m. In order to accommodate the need for more detailed data for archaeological assessments, fill-in lines with focus on marine magnetometry and shallow seismic were added, thus narrowing the line distance to 25 m.

In the near-shore section the programme was limited to the acquisition of single beam bathymetry with 10 m line spacing. The dinghy ‘Rambunctious’ was used as platform for near-shore bathymetry acquisition.

The investigations are reported in /4/ comprising the bathymetry, magnetic anomalies, seabed classification, selected seismic profiles, and contours of surface as well as thickness of selected seismic units.

In 2009 a bathymetric survey was performed within a much bigger area (c. 836 km2) between Fehmarn and Lolland (/38/). Three vessels were used for the survey (M/V Triad, M/V Ping and M/V Seabeam).

Also in 2009, GEUS/DHI has performed a number of near-shore surveys in water depths between 2 and 6 m along the German and the Danish coasts (/39/). The three purposes for the surveys were archaeological investigations, marine biological mapping and coastal profiling. The survey lines were typically sailed with an interval of 25 m in the direction along the coast, with a little overlap to the area surveyed in 2008 for QA reasons. However, for the coastal profiling the lines were sailed perpendicular to the coast.

Finally, in May 2010 a supplementary survey was performed immediately west of the area covered by the 2008 investigations (/4/).

In addition relevant information on munitions in the Fehmarnbelt has been collated (/42/), and the magnetometric survey maps and the side scan recordings have been carefully studied to detect possible munitions.

4.2 Equipment and Procedures The 2008 survey was performed with the vessel R/V Madoc as platform; the following was carried out:  Multibeam echo sounder bathymetric measurements.  Marine shallow seismic using low and high frequency sources.  Side scan sonar recordings.  Marine magnetometry.

The low frequency shallow seismic profiling was acquired using a Georesources Geo- spark 200 sparker source and a five element hydrophone streamer; the high frequency seismic was recorded using a Benthos Chirp III. The instrument was controlled from a Benthos CL-160 top-site unit.

The magnetometer was a Geometrics G882. A Reson SeaBat 8125 multibeam echo sounder was employed for acquisition of the offshore bathymetry. This instrument emits 512 beams in a beam angle of 60 and thus provides a swathe width of approximately three times the depth below the transducer.

In order to obtain valid depth measurements from the echo sounder, profiles of the sound velocities in the water were measured regularly. For this purpose an FSI CTD was used.

The position of the vessels was obtained using Javad GPS in RTK mode. A base station was installed on top of a silo next to Rødbyhavn.

The bathymetric Large Area survey in 2009 (/38/) was performed with Kongsberg Multi- beam EM3002D instruments as echo sounders.

The Supplementary survey in 2010 (/4/) was performed with the ship R/V Madoc as a platform and a Benthos SIS 1625 was employed for both high frequency seismic investigations and side scan recording. As magnetometer was used a Geometrics G882 instrument, and a Reson SeaBat 8125 was employed for the offshore bathymetric survey.

RA’s bathymetric surveys in 2008- 2010 were performed with reference to IHO Standards for hydrographical surveys, S-44, 5th edition /61/.

The near-shore survey performed in 2009 by GEUS used the ship Føniks Miljø as a platform for the part of the investigations that was related to archaeological purposes. A Geometrics G882 magnetometer was employed, and a Benthos SIS 1626 was used as combined side scan/sub-bottom profiler. For the part of the investigations that was conducted for coastal profiling purposes, the small ship GEUS II was used as a platform and bathymetric data was collected with a Navisound Reson instrument.

5 Onshore Geophysical Investigations 5.1 General The onshore geophysical investigations were performed by RA in 2008 and 2012. The 2008 programme included reflection seismic investigations on Fehmarn and Lolland and Continuous Vertical Electrical Sounding (CVES) on Fehmarn. The 2012 investigation included reflection seismic profiling in an area onshore east of Rødbyhavn where the borings indicated a very uneven surface of the Palaeogene deposits.

The 2008 reflection seismic survey included two lines with a total nominal length of 4.5 km near Rødby and three lines with a length of 5.3 km near Puttgarden. The CVES survey included 11 lines near Puttgarden, with a total nominal length of 9.4 km. The investigations are reported in /3/.

The investigations performed by RA in 2012 included three seismic lines, each recorded in two opposite directions, and one ultra-high resolution seismic line. Total line length of the 2012 survey was 3.6 km. 5.2 Equipment and Procedures

5.2.1 Onshore Seismic The Pulled Array Seismic method is based on a seismic vibrator as energy source and a land streamer with geophones mounted on steel plates as receivers, cf. also Table 5-1.

The investigations were performed using an IVI Minivib T7000 seismic vibrator as energy source. The land streamer is fitted with geophones from Mark Products. The total length of the land streamer is 222.5 m.

Table 5-1 Recording parameters Distance between vibration points 10 m Distance between vibrator and first group 6.25 m Distance between group 2‒49 (first channel is used for 1.25 m correlation of the sweep) Distance between group 50-112 2.5 m Sample distance 0.5 m

5.2.2 Continuous Vertical Electrical Sounding (CVES) CVES is an automatized geoelectrical measuring method which allows rapid collection of large datasets from electrical soundings. The principle is that a current applied to two electrodes generates an electrical potential in the ground which can then be measured using two other electrodes. When the distance between the power electrodes is increased, the penetration into the ground will also be increased. Using varying geometries it is then possible to obtain information of the variation of the resistivity with depth as well as laterally.

The equipment used for the CVES measurement is from the Swedish company ABEM.

All CVES lines are recorded with the 400 m setup, giving an investigation depth of approximately 60m. The interpreted CVES are in /3/ reported as profiles and as contoured maps for the interpreted depth to, and the level of, the Palaeogene Clay surface.

6 Borings and Cone Penetration Tests (CPTUs) 6.1 General Three types of borings/CPTUs have been performed since 2009 and are defined as type A, type B and type C.

All borings and CPTUs performed since 2009 have been numbered with a specific code as defined below where the designation for the type of boring/CPTU is incorporated in the central part of the number as "XY.A.MNO", XY.B.MNO" and "XY.C.MNO". Hence in this number:  "XY" is the last two figures in the year, when the boring/CPTU was performed, e.g. "09".  "A" is used for borings with the main purpose of obtaining high quality samples.  "B" is both used for borings with down-the-hole CPTU alternating with clean out boring with the main purpose of performing deep continuous or nearly continuous CPTU (sampling may also be performed from the clean out boring material) and for onshore single push surface CPTUs performed from a truck mounted CPTU system.  "C" is both used for offshore CPTUs performed from a heavy seabed rig resting on the seabed and for offshore CPTUs performed from a surface rig typically mounted on a jack-up.  "M" is "0" for type A and B borings performed in Fehmarnbelt, "4" for type C borings/CPTUs performed in Fehmarnbelt, "6" for type A and B borings performed onshore on Fehmarn, "7" for type A and B borings performed onshore on Lolland, "8" for type A and B correlation borings performed in Lillebælt and "9" for type A and B correlation borings performed onshore for the Fehmarnsund Brücke.  "NO" is a unique serial number with two figures, e.g. "78".

The locations of the borings and CPTUs appear generally from Enclosure I (Drawing no. 070-02-12 and 070-02-13). 6.2 Offshore Type C Borings, Fehmarnbelt

6.2.1 General In Fehmarnbelt a total of 41 seabed CPTUs have been performed by Fugro from a seabed rig. The target depth was 25 m. The CPTU results appear in /1/.

In addition, for the offshore large scale testing site, CPTUs have been performed by GEO at 3 locations for the pile test sites using a surface rig. The target depth varied between 30 and 40 m. Finally at 5 locations from the base of the trial excavation, seabed CPTUs were performed by GEO from a seabed rig. Target depths varied between 5 and 15 m. In 2013 seabed CPTUs at 3 locations from the surface of the Phase 4 (approximately 3 m high) backfilling of sand in the trial excavation have been performed to a depth of 8 m. The CPTU results appear in /33/.

6.2.2 Equipment and Procedures 2 All CPTUs were performed using a 1000 mm cone measuring cone resistance (qc), the sleeve friction (fs) and the pore pressure (u2).

The seabed rigs and the surface rig had a thrust capacity of 200 kN. The seabed CPTUs were performed according to /17/ with target class 2 accuracy. 6.3 Onshore Type A and Type B Borings, Lolland and Fehmarn 6.3.1 General The onshore borings/CPTUs have included:  Borings type A and type B performed by Fugro in 8 locations on Fehmarn and 4 locations on Lolland (in total 12 type A-borings and 10 type B-borings). Maximum depth of these borings was 100 m. The onshore borings are reported in /2/.  Single push CPTUs (type B) performed by Fugro at 30 test locations on Fehmarn. The single push CPTUs are reported in /24/.  Borings type A performed by GEO in 10 locations to 30 – 35 m depth on Lolland for the production site and shore approach of the Fehmarnbelt Fixed Link. These borings are reported in /44/ and referred to as production site borings.  Single push CPTUs performed by Ramboll Sweden in 55 locations distributed over the production site area on Lolland. The target depth was until refusal at minimum 50 kN penetration thrust or 5 m penetration, whichever came first. These CPTUs are reported in /45/ and referred to as production site CPTUs.  Borings type B (generally supplemented with sampling) performed by GEO adjacent to the alignments in 7 locations on Lolland and 1 location on Fehmarn to depth varying between c. 9 m and 57 m. These borings are reported in /46/ and referred to as alignment borings/CPTUs.

6.3.2 Equipment and Procedures The type A-borings at the 12 locations performed by Fugro were started by use of cable percussion techniques and when coreable layers were met, the drilling work was continued using Geobor-S wireline drilling technique. However, in boring 09.A.601, 09.A.607 and 10.AB.610, Symmetrix destructive drilling techniques were used to overcome cohesionless layers with stones and cobbles. During drilling in non-coreable strata, tube samples were recovered with a push sampler and bag samples were recovered with a bailer sampler or hammer sampler. Coring was performed with the Geobor-S equipped with a triple barrel coring device producing cores with 101 mm diameter.

In the onshore type A-borings (except for 09.A.601) performed by Fugro a 88 mm PVC tube liner was installed and sealed with a bentonite grout enabling geophysical borehole Logging, including VSPs, to be performed (see Chapter 7). Additionally, slotted stand- pipes were installed in two of the borings (09.A.602 and 09.A.607E) for water level observations.

The type B-borings at the 10 onshore locations carried out by Fugro were performed in a distance of c. 5 m from the corresponding A-boring, using a truck mounted CPTU system alternating with another truck mounted drilling system. The procedure included a CPTU push generally followed by drill-outs until the bottom of the borehole was reached. The CPTU truck penetrated CPTU rods to the depth of refusal in each stroke (stroke up to c. 25 m) alternating with clean-out boring and succeeding stroke etc. For some of the onshore CPTUs like 09.B.604A only one single stroke has been performed and reported.

The production site borings were commenced by using dry rotary drilling or percussion drilling within an 8" casing. As soon as practically possible the drilling method was changed to core drilling with Geobor-S triple tube equipment to the bottom of the boring.

The production site CPTUs were performed with CPTU equipment mounted on a crawler Georig 707.

The alignment borings/CPTUs were generally carried out initially using a truck mounted CPTU system for the first push and then followed by clean-out boring by rotary drilling and/or Geobor-S until next CPTU stroke performed with a down-the -hole CPTU equipment to refusal or maximum stroke length of 1.5 m alternating with clean-out boring and succeeding stroke etc. During the clean-out boring bag samples/cores were taken regularly.

The production site borings and alignment borings were sealed with cement/bentonite grout or just bentonite grout until 1 m below the surface. The uppermost 1 meter was filled with excessive material from the drilling works. 6.4 Offshore Type A and Type B Borings, Fehmarnbelt 6.4.1 General The offshore borings/CPTUs have included:  36 type A and 30 type B borings were performed by Fugro in 36 offshore locations. The boring depth was up to 100 m. These offshore borings are reported in /2/ and /23/.  Borings type A performed by GEO in 10 near-shore locations to c. 30 m depth for the production site and shore approach for the Fehmarnbelt Fixed Link to Lolland. These borings are reported in /44/ and referred to as production site borings.

6.4.2 Equipment and Procedures The offshore type A-borings drilled by Fugro, except 09.A.004, were performed from Jack-up drilling platforms “SKATE III” or “Deep Diver”. The borings 09.A/B.004 were performed from the geotechnical drilling vessel “Highland Eagle”. The borings 09.B.007, 09.B.008 and 09.B.017 were performed from the geotechnical drilling vessel “Gargano”. The type B-boring at a location was generally performed within a distance of 5 m of the corresponding type A-boring.

Type A-borings carried out by Fugro were drilled using cable percussion techniques followed by core drilling using Geobor-S.

Type B-borings (down-the-hole CPTU borings) by Fugro were performed with Geobor-S in combination with a non-coring device deployed in the drill bit. The downhole CPTUs employed the WISON XP system on the “SKATE III” and “Deep Diver” and the WISON MkIII system on the “Gargano” and “Highland Eagle”. The CPTUs have been performed with down-the-hole equipment with maximum stroke length of 1.5–3.0 m alternating with clean-out boring and succeeding stroke etc.

The production site borings were commenced by using dry rotary drilling or percussion drilling within an 8" casing. As soon as practically possible the drilling method was changed to core drilling with Geobor S triple tube equipment to the bottom of the boring. These borings were performed from the Jack-up drilling platform "Jack VI".

2 All CPTU tests used a 1000 mm piezocone measuring cone tip resistance (qc), the sleeve friction (fs) and the pore pressure (u2). CPTs were performed according to /17/ with target class 2 accuracy.

After completion the boreholes were sealed with a cement/bentonite grout. 6.5 Sample Handling and Laboratory Classification Works 6.5.1 General The sample handling and laboratory classification works included mainly:  On-site sample handling.  Geotechnical classification and laboratory testing.

The geological descriptions, the sample colour photographs and results of the classification tests and laboratory tests for the borings appear in detail in /2/, /23/, /44/ and /46/.

6.5.2 Equipment and Procedures The samples retrieved from the borings were generally handled as follows:  Hammer samples and push samples taken with Shelby tubes were after measurement of the recovery, cleaned from drill cuttings and sealed with plugs.  Bailer and percussion samples taken with a split spoon were transferred to plastic bags and sealed.  Core samples were, after removal from the barrel, immediately transferred to a split liner, cleaned (by removal of drill fluids and cuttings) and recovery was then determined.

The samples were then transported to the geotechnical laboratories, where the testing / activities for selected samples generally included:  Geological sample description in accordance with Femern's Geo Nomenclature /12/ and DGF Bulletin 1 /54/.  Sample colour photography (for the production site borings and alignment borings / CPTUs only for cores).  Water content determinations in accordance with CEN ISO/TS 17892-1 for soils and ISRM Part 1 for chalk.

 Unit weight determination in accordance with CEN ISO/TS 17892-2 for soils and ISRM Part 1 for chalk (except for the production site borings and the alignment borings/CPTUs).  Labelling and preservation of samples for Femern's Advanced Laboratory Testing and for geological dating (except for the production site borings and the alignment borings/CPTUs).  Particle density test in accordance with CEN ISO/TS 17892-3 (except for the production site borings and the alignment borings/CPTUs).  Particle size analysis in accordance with CEN ISO/TS 17892-4.  Atterberg limits in accordance with CEN ISO/TS 17892-12.  Organic content in accordance with BS1377, Part 3, Clause 4 and BS1377, Part 3, Clause 3 (except for the production site borings and the alignment borings/CPTUs).  Transportation/return of remaining sample material from the laboratories to Femern's sample containers, at that time stored at the farm Lidsø near Rødbyhavn. 6.6 Correlation Borings Lillebælt The purpose of the correlation borings was to analyse the effect the load from the bridge has had on the clay below the foundations. The following investigations were performed:  A 10 m deep boring 10.A.803 was performed by GEO through and beneath the concrete slab inside Pier 3, cf. Appendix GDR 00.1-001-F.  For safety reasons, prior to commencement of the offshore borings a munitions survey with magnetometer and side scan tools was carried out in a minor area to the west of the bridge by GEUS (/42/).  2 offshore sampling type A-borings 10.A.801 and 10.A.802 and two type B-borings 10.B.801 and 10.B.802 were performed from the Jack-up drilling platform “Deep Diver” by Fugro. Boring 10.A.801 and 10.B.801 were performed adjacent to Pier 1 to a depth of 75 m below seabed and boring 10.A.802 and 10.B.802 close to Pier 3 to a depth of 40 m below seabed. The results of the investigation appear in detail in /25/. 6.7 Correlation Borings Fehmarnsund The purpose of the correlation borings was to analyse the effects the load from the c. 22 m high northern embankment of the Fehmarnsund Brücke has had on the clay below. The following investigations were performed:  One type A sampling borehole 10.A.901 and one type B CPTU boring 10.B.901, both to c. 80 m depth performed by Aarsleff/GEO. The results of the investigations are described in detail in /34/.

7 Geophysical Borehole Logging 7.1 General The geophysical borehole logging has been performed by RA in 33 offshore boreholes in Fehmarnbelt, in 7 onshore boreholes on Fehmarn and in 4 onshore boreholes on Lolland.

Vertical Seismic Profiling, VSP, has been performed in 7 onshore boreholes on Fehmarn and in 2 boreholes on Lolland.

The borehole logging has been performed as an integral part of the follow-up of the geological description and classification.

The geophysical borehole logging is reported in /5/. 7.2 Equipment and Procedures The logging programme included a number of different logging probes and the VSPs were performed with a hydrophone array and with an array of two 3D geophones. The choice of log-suite is very dependent on the stability of the borehole wall and on the installation in the borehole. The different logs are to a different degree influenced by the fluid (mud or polymer) in the borehole, by plastic or steel casings and by the borehole diameter.

Based on this fact, the chosen logging strategy has been to try to keep the borehole open without any kind of casing through the chalk and Palaeogene clay, because in an open borehole a wide range of logs can be recorded without any influence from the installation. Only a reduced log programme can be conducted in a plastic liner with or without slots or in a steel casing. However, in most of the borings there were indications for possible stability problems, and in only one borehole was the stability with sufficient certainty adequate for the optical televiewer to be used.

The offshore logging programme included:  Natural Gamma.  Single Induction Conductivity.  Focused Guard Log Resistivity, Deep and Shallow.  Compensated Neutron-Neutron Porosity.  Compensated Gamma-Gamma Density.  Calliper 1-arm or 3-arms.  Sonic P-Wave Velocity.  Acoustic Hardness (only in the open boreholes).  Acoustic Radius (only in the open boreholes).

The onshore boreholes were initially logged through a cemented plastic liner. The total onshore logging programme included:  Natural Gamma.  Dual Induction Conductivity.  Sonic P-Wave Velocity.

 VSP S-wave Velocity.  VSP-P wave Velocity.

The complete log suite for each borehole and the interpreted geological series of layers are together with technical information presented as composite log and VSP sheets in /5/.

8 Advanced Laboratory Testing 8.1 General The Advanced Laboratory Testing comprising pilot and production testing has been performed by GEO, Deltares, NGI and Fugro.

The objective of the pilot testing was to clarify procedures to be applied during production testing. The objective of the production testing was to establish soil properties as measured in the laboratory with the identified procedures. The total number and type of tests for the different soil units reported and considered in this Ground Investigation Report are shown in Table 8-1.

The advanced laboratory tests included:  IL: Oedometer, incremental loading.  IL K0: Oedometer, incremental loading, determination of K0.  CRS: Oedometer, constant rate of strain.  CPR: Oedometer, constant pore pressure ratio.  CADc: Triaxial, anisotropic consolidation, drained failure, compression.  CAUc: Triaxial, anisotropic consolidation, undrained failure, compression.  CAUe: Triaxial, anisotropic consolidation, undrained failure, extension.  CAUcy: Triaxial, anisotropic consolidation, undrained failure, cyclic.  DSSst: Direct simple shear box test, static.  DSScy: Direct simple shear box test, cyclic.  UU: Triaxial, unconsolidated, undrained failure.  RC: Resonant Column.  SP: Standard Proctor.  UCS: Unconfined compression test.  Brazil: Brazil test

The majority of advanced laboratory testing has been performed for the clays of Palaeo- gene origin, being the most challenging soil type, but also a considerable number of advanced tests have been performed for Glacial tills and the Cretaceous chalk.

The specimens have generally been tested for their effective in-situ stress state, but effective stress states reflecting unloading and reloading have also been included.

For Glacial meltwater sand and for the Palaeogene Æbelø formation only limited testing has been performed. Glacial meltwater sand sample material was only recovered in a single borehole (09.A.018) and the Æbelø formation has only been found as thin layers (approximately 1 m thickness) at around 20 m depth or deeper..

Triaxial testing on recompacted clay till (upper and lower till samples) has been performed together with standard proctor testing on the same material to clarify whether this material can be used as fill material in e.g. embankments.

The detailed Advanced Laboratory Testing has been reported in /26/ through /32/ and in /62/, where /26/ includes the details about the applied laboratory procedures. Water used during laboratory testing for saturation and within the triaxial cells is artificially established pore water being chemically almost identical to the in-situ pore water.

X-Ray photos have generally been used for selecting cores for testing. Routine classification tests have been performed for the cores selected for Advanced Laboratory Testing. X-Ray diffraction analyses have been performed for selected cores for determination of clay mineralogy.

A special laboratory investigation of the clays of Palaeogene origin including laboratory mixing, cement stabilisation and unconfined compression testing of this material has been reported in/32/.

Additionally a special laboratory investigation with the purpose to investigate the impact on Palaeogene clay as it is exposed to bentonite slurry and seawater in a static state and during pumping has been performed and reported in /64/. 8.2 Equipment and Procedures 8.2.1 General Details about the equipment and procedures can be found in /26/. The geological dating of selected samples has been performed by GEUS and the results are included in /6/.

8.2.2 Oedometer Testing During oedometer tests it is found that the clays of Palaeogene origin, when mounted in the cell without access to water and loaded to their in-situ vertical effective stress and then flooded, will absorb water and try to swell. This behaviour has been observed for swelling tests and traditional oedometer tests, and it indicates that the clays of Palaeogene origin in nature should expand for their in-situ stress state. This is, however, not believed to be realistic. If the clay expands when flooded at the vertical effective in-situ stress in the laboratory, the water content increases relative to the in-situ state, implying that the properties measured in the test may not be representative.

The observations described above are most likely a consequence of a too low mean stress in the specimen at the time of saturation, meaning that the horizontal effective stresses are not fully regenerated when loading the specimen to its vertical effective in-situ stress. Sa- turation in the oedometer cells is therefore performed at a higher stress level (typically 1.5 to 2.0 times the vertical effective in-situ stress at shallow depth and decreasing with increasing depth).

The methods for estimating the pre-consolidation pressure are described by Akai /18/, Becker /19/, Casagrande /20/ and Janbu /21/. For both the clay till and for the clays of Palaeogene origin it has been a challenge to identify a clear and consistent pre-consolidation pressure.

Table 8-1 Summary of performed advanced laboratory tests

Clay till is a hard to very hard material containing many coarse grains. Trimming of the specimens will therefore usually imply an uneven perimeter and full contact between the soil and the Oedometer ring is therefore not established from the beginning of the test. The stiffness of the material combined with the lack of full contact between the specimen and the ring implies that the measured vertical strain will include an effect of the total specimen volume being squeezed laterally towards the Oedometer ring. A clear break down of the soil skeleton is therefore not seen in all the tests. It has been found that the method of Casagrande will imply a reasonable estimate of the pre-consolidation pressure, ´pc provided that ´pc does not exceed 1.5 to 2.0 MPa (a maximum of 4.8 MPa can be applied in the apparatus).

For clays of Palaeogene origin the rate of secondary compression, C increases with the axial stress and the virgin compression “line” therefore keeps curving in a semi-logarithmic plot. As a consequence of the curving virgin compression line, the methods of Beck- er and Casagrande will lead to pre-consolidation pressures being very dependent of the stress interval used along the virgin compression line. The pre-consolidation pressure as determined by the method of Janbu is located at a stress being slightly lower than where the minimum tangent modulus value is identified and the estimate of the pre-consolidation pressure is therefore not influenced by the stress-strain behaviour for large axial stresses where the creep contribution may be significant.

8.2.3 Triaxial Testing Triaxial testing is generally performed in accordance with Danish practice using test specimens with a height to diameter ratio of unity combined with the use of lubricated ends. The triaxial tests on clay till and Palaeogene specimens have generally been carried out as consolidated undrained tests with pore pressure measurements (measurement on back pressure line).

In the triaxial cell, the specimen is consolidated before shearing is initiated. Two different consolidation approaches were tested during the pilot phase:  Danish experience with triaxial testing on over-consolidated soils is that the specimen should be K0-consolidated to a low estimate of the pre-consolidation pressure ′pc and then unloaded under K0 conditions to the stress state, from which shearing will be initiated.  An alternative procedure is to load the specimen stress controlled directly to the in- situ stress state, from which shearing will be initiated.

The undrained triaxial shear strength in compression was measured using the two procedures above and compared with test results from constant volume Direct Simple Shear test and field measurements using CPTU. Specimens of clay till and Palaeogene clay were used. The findings may be summarised as:  For Upper Clay Till the specimens are loaded stress controlled to a low estimate of the pre-consolidation pressure (K0 = 0.40-0.45) before a stress controlled unloading is initiated, followed by shearing.  For all other soil types the specimen is loaded stress controlled directly to the in-situ stress state from which shearing is initiated after end of primary consolidation has been reached.

 DSS-tests are always consolidated for a low estimate of the pre-consolidation pressure, and then unloaded and consolidated for the axial stress at which shearing is performed. This procedure is adopted in order to obtain representative lateral effective stresses within the specimen

The in-situ stress state was estimated combining the vertical effective in-situ stress (hydrostatic pore pressure distribution) with correlations between the net cone resistance from CPTU and the pre-consolidation pressure, leading to the over-consolidation ratio, OCR. The lateral effective in-situ stress was defined using the coefficient of lateral earth pressure at rest, as measured in the IL K0 tests.

For clay till, a strain rate of 0.3 %/h is used during undrained shearing, whereas 0.05 %/h is used for the clays of Palaeogene origin (undrained shearing). The pressure heads in the triaxial cell have a diameter of 70 mm. For clay till it has been found that a specimen diameter with 70 mm will suffice whereas a specimen diameter of 68 mm for clays of Pala- eogene origin must be used to ensure that the footprint of the specimen will not exceed the area of the pressure head during shearing.

8.2.4 Geological Dating of Samples Geological dating based on microfossils has been performed on 167 samples of clays of Palaeogene origin and 7 samples of chalk. The chosen strategy for the dating has been to start with coccoliths and foraminifera (with CaCO3 shells) and to supplement with the more expensive palynological analyses on samples where the results of the first analyses have not given sufficiently reliable results. Samples without a content of CaCO3 are of course only analysed palynological. Methods and results are further described in /6/.

9 Large Scale Testing 9.1 General The Large Scale Testing has been performed by Per Aarsleff with GEO as subcontractor. NGI has supplied excavation and pile instrumentation and reference measurement device under an appointment from Femern A/S. The works carried out to date (January 2014) include:  Phase 1 excavation (base 30 by 8 m2) to elevation -20 m in 2010.  Piezometer and extensometer installation (extenso-piezometers 3, 9 and 25 m below excavation surface, EP1- EP9) in 2010.  Phase 2 excavation (extension of Phase 1 to form a base 30 by 30 m2) in 2010.  Phase 3 excavation (extension of Phase 1 and 2 to form a base 30 by 70 m2) in 2011.  Phase 4 incremental backfilling of the Phase 1, 2 and 3 excavation with 3 m of sand to elevation c. -17 m in 2011.  Phase 5 final backfilling of the excavation with 7m of sand to elevation c. -10 m in 2013.  Surface benchmark installation (12 surface benchmarks) and 5 additional benchmarks for the multibeam surveys from 2010.  Fourteen multibeam surveys of the 200 by 200 m2 survey area.  Cone penetration tests at 3 locations outside the excavation (10.C.451-453) in 2010, at 5 locations from the base of the excavation (11.C.461-465) in 2011 and at 3 locations from the top of the Phase 4 (3 m) backfilling in the excavation (13.C.466-468), cf. Chapter 6.2.  Block sampling at 6 locations from the base of the Phase 2 and Phase 3 excavation in 2011.  Instrumentation monitoring from 2010; continues until autumn 2013.  Plate load testing (4 vertical load test, 4 horizontal load tests and 1 passive load test) in the Phase 3 excavation in 2011.  Driven pile installation of 5 test piles (DP1-DP5) and 4 reaction piles in 2010.  Bored cast-in-situ pile installation of 5 test piles (BP1-BP5) and 4 reaction piles in 2010.  Driven pile tension load testing of five piles (DP1- DP5), in total 15 tension load tests including repeat testing from 2010.  Bored pile tension load testing of five piles (BP1- BP5), in total 15 tension load tests including repeat testing from 2010.  Installation of 12 vertical ground anchors (G7- G14, G9A, G10A, G17 and G18) and 2 vertical dummy anchors (G15 and G16) in an onshore test area east of Rødbyhavn in 2011. Additional installation of 2 vertical ground anchors (G19 and G20) in the same test area in 2012.  Ground anchor load testing of 12 ground anchors (G7, G8, G9A, G10A, G11- G14, G17 - G20) and 2 dummy anchors (G15 and G16) in 2011- 2012.  Final reinstatement of the test sites.

The works performed by Per Aarsleff and GEO are reported in /33/ through /37/, /47/ and /48/. The results of the works have been elaborated by RA in /49/ through /53/.

9.2 Equipment and Procedures 9.2.1 Trial Excavation Works and Surveys The excavation has been performed by the excavator Komatsu PC3000 mounted on the spud barge Maricavor. Tugboat and split barges have been used for transporting the excavated material to the dump site.

Slopes were excavated with inclination 1 (vertical) to 2 (horizontal).

The 3 excavation phases are illustrated in Figure 9-1 below.

Figure 9-1 Illustration of the 3 excavation phases

Incremental backfilling of the excavation was performed in two phases, Phase 4 and Phase 5, by a hopper dredger pumping the infilling material onto the excavation base.

Multibeam surveys were performed using a Reson Seabat 7125 multibeam echosounder.

9.2.2 Instrumentation and Monitoring An overview of the established monitoring system appears from Figure 9-2.

The six medium depth (9 m) and deep (25 m) extenso-piezometers were installed in partly predrilled boreholes while the three shallow (3 m) extenso-piezometers were pushed in from the base of the Phase 1 excavation. All nine extenso-piezometers were pushed to final depth using GEOs surface rig GEOTop with thrust capacity 20 tons. Automatic logging of data started 2010-09-30.

Four pile piezometers were placed at 14 m (PP1 and PP3) and 19 m (PP2 and PP4) below top of driven pile number DP5.

The reference measuring device (RMD) has been used to measure the relative height difference of the pedestals, extensometer anchors below pedestals and surface benchmarks. After each measurement the height of each pedestal/extensometer anchor/surface benchmark has been calculated relative to pedestal P5, which has been assumed to be a stable reference point. Extension tubes have been added to the surface benchmarks at the base of the trial excavation before the Phase 4 incremental backfilling began. Pedestals were not extended and were instead covered over with a protection cover on a tripod frame.

Data recovery cable & pick-up wire/buoy

-10m Logging station 3m Infill Incl. Incl. -20m Bench Bench -24m -mark -mark -23m PP3 PP1 EP2 EP4 EP8 -29m -29m PP2 PP4 EP3 EP6 EP9

Test pile with piezometers -45m

EP1 EP5 EP7

Figure 9-2 Overview of monitoring system with extenso-piezometers (EP1-EP9) and test pile piezometers (PP1-PP4)

In total as of January 2014, data from the extenso-piezometers has been downloaded 16 times and surveys using the RMD have been carried out 11 times.

9.2.3 CPTUs and Block Sampling The CPTUs performed for the offshore large scale testing site is briefly described in Chapter 6.2.

Block samples at the 6 locations at the base of the excavation have been obtained using a special block sampler tool developed by GEO. The tool has been deployed from the vessel Mira A with a support from the "Jack IV" jack-up platform. The block samples have been taken from base of the excavation at elevation -20 m to the target depth of 1 m below the excavation base. The sample diameter was 300 mm. These samples have been used for advanced laboratory testing, cf. Chapter 8.

9.2.4 Plate Load Testing Three different kinds of plate load tests have been performed:  4 vertical load tests (V1- V4) with 1000 mm diameter 100 mm thick steel plates.  4 horizontal load tests (H1- H4) on fabricated caissons with a footprint of 1.00∙0.50 m2, equipped with 0.10 m skirts.  1 passive load test (P) with 2 parallel steel plates, 6.06 m long by 2.06 m high, separated by 8 hydraulic jacks.

The test set-up for the vertical and horizontal load tests comprised the installation of a guide frame which was used to assist with the positioning of two main frames, one either side of the guide frame. The guide frame measuring 10 m by 10 m was positioned directly on the seabed. The main frames were used as a guide to prepare the test areas, position the test plates and to ensure that the load was applied in the correct orientation during the test. The main frames measuring 8 m by 10 m were positioned directly on the seabed either side of the guide frame.

The passive load test involved pushing the 2 parallel plates in a trench box apart while recording load, displacement and inclination. The trench for the passive test was excavated as part of the Phase 3 excavation.

The overall plate load test arrangement appears from Figure 9-3 below.

Figure 9-3 Overall plate load test arrangement

The set-up for the passive plate load test is illustrated in Figure 9-4.

Figure 9-4 Illustration of the set-up for the passive plate load test 9.2.5 Pile Installation and Tension Load Testing The steel tube piles DP1-DP4 being OD508 with a wall thickness of 20.0 mm and DP5 being OD508 with 22.2 mm wall thickness were driven to 25 m below seabed. Additi- onally, four reaction piles DR1-DR4 were driven to 22 m below seabed. During pile driving the piles were checked for plugging and plugs were removed before pile driving was resumed.

The bored cast-in-place piles BP1-BP5 were OD610 mm for the upper 10 m below seabed and OD500 mm along the remaining pile length; they were installed to 25 m below seabed using a hydraulic drilling rig mounted on a jack-up. Additionally, four reaction piles BR1-BR4 were installed to 22 m below seabed. The drilling with casing, bucket and auger was continued until the casing was firmly embedded in the clay. Final drilling, in-

stallation of reinforcement cage and concreting by tremmie tube was performed in a “dry” borehole.

All piles were installed through a piling guide frame with a 3 by 3 pile configuration.

All five driven test piles as well as all five bored cast-in-place piles have been tension load tested by means of a large reaction frame placed on top of the four reaction piles. The tension load is applied by a hydraulic jack mounted on the top of the reaction frame.

Displacements of pile heads were measured by displacement transmitters installed on a separate reference girder beam. Supplementary measurements of displacements were performed by means of the RMD.

The set-up for the pile load testing is illustrated in Figure 9-5.

Figure 9-5 Illustration of the set-up for pile load testing

9.2.6 Ground Anchor Installation and Tension Load Testing The anchor bores were constructed using a Klemm 806-4 anchor drilling rig. OD152 mm casing was installed in the bores which were flushed with water until 25.5 m below ground level. Below this level the bores were flushed with air in order not to soften the Palaeogene clay within the bond / fixed length. As a comparative check on construction techniques anchors G17 and G18 were flushed with water over the entire anchor length.

7 post grouted anchors and 7 non- post grouted anchors were installed in the anchor bores with a fixed bond length of 7 m and sheathed free length of 30.5 m. Two dummy anchors were also installed with free length only and therefore had no bond length.

The anchor bars were 36 mm diameter Dywidag WR solid bars of grade 950/1050 MPa steel, with a smooth sheathed free length and a corrugated bond length.

The anchors at the base of the tendons were temporary Stump-Duplex compression type anchors of diameter 82.5 mm, in which load transfer from the anchor to the ground begins at the distal end of the bond length and dissipates towards the proximal end, en- suring that the bond length remains in compression when the anchor is loaded, and the free anchor length therefore extends from ground surface to the base of the bond length.

Three types of static tension load tests were carried out on the installed anchors: 1. Investigation test; anchor incrementally loaded to failure (Method 3, E.4.1 - EN 1537: 1999), in total 8 carried out. 2. Free length friction test; dummy anchor incrementally loaded to failure, in total 2 carried out. 3. Creep test; load reduction and ground anchor displacement recorded under locked off load (Method 2, E.3.4 - EN 1537:1999), in total 5 carried out.

The anchor G8 was tested twice, once in 2011 as a creep test and again in 2012 as an investigation test.

10 Ground Conditions 10.1 Geologic Setting The geologic and tectonic framework for the Fehmarnbelt area is described in /7/ and /8/. In those reports the deeper lying deposits including the thick Permo/Triassic salt layers, the main fault systems and the halokinetic structures (salt pillows) are described. The conclusions from the reports are summarised in Chapters 10.2 and 10.3 below.

The groundwater conditions described in detail in /9/ are summarised in Chapter 10.4 below.

The near surface Postglacial/Lateglacial deposits will be of direct interest to the excavation and foundation works for the fixed link constructions as will, possibly, also be the deposits from the youngest part of the Cretaceous period, from the Palaeogene period and from the rest of the Quaternary period (Neogene deposits have not been, at least not yet, found in the area). The above mentioned deposits have been described in detail in /6/ and the main observations and conclusions are summarised in Chapters 11 through 14.

10.1.1 Morphology and Seismic Stratigraphy The seabed in the investigated area between Fehmarn and Lolland has been divided into a central basin with water depths above c. 24 m and two gently sloping areas from the coasts to the basin with water depths from 0 to c. 24 m (Figure 10-1). In the following sections these areas are termed “the basin” and “the slopes”.

Figure 10-1 Depth profile across Fehmarnbelt. All units in metres

Based on the seismic investigations performed in 1995/96 and in 2008 a number of seismic units were defined, and in the early reports special names for those have been used. However, the Boring Campaigns in 1996, 2009 and 2010 revealed the geologic character of the units mapped from the seismograms, and therefore in this report only the correct geological/sedimentological names and terms have been used.

Table 10-1 Relation between deposits, units and geological characters of the soils in the area Unit Deposit Geological description Postglacial/Lateglacial Marine sand Postglacial marine sand with shells sediments Marine gyttja Marine gyttja (includes also freshwater peat and gyttja) Laminated freshwater Silty clay and silt, often laminated clay and silt Marine silty clay Mostly silty clay. Local sand layers Freshwater sand Sorted sand. Locally with silt and clay layers. Upper till unit Upper till Sandy to very sandy, very hard clay till Meltwater sand unit Meltwater sand Meltwater sand. Mostly medium- coarse. Appears only locally Lower till unit Lower till deposit Mostly medium plasticity clay till Meltwater sand Meltwater sand. Only found in a few borings Chalk till Almost white, very calcareous, hard clay till Lowermost till High to very high plasticity clay till Palaeogene clay unit Søvind Marl Most often very calcareous or (Folded as well as intact) calcareous light grey clay Lillebælt Clay Greenish, very high plasticity clay. Often non-calcareous Røsnæs Clay Often reddish or brownish, very high plasticity clay. Normally calcareous. Ølst Clay Dark or even black, high to very high plasticity clay with pyrite concretions. Numerous ash layers Holmehus Clay Multicoloured (reddish, bluish, greenish) very high plasticity, non- calcareous clay Æbelø Clay (silt) Grey, non-calcareous, often very silty clay or even silt. Cretaceous chalk Maastrichtien and White chalk Campanian deposits

A 3D digital geologic model of the area has been set up, illustrating the distribution of and the shape of the surfaces of the most important layers. This model is available in the Femern Geo Information System, /22/. 10.2 Salt Pillow Structures During the last part of the Permian, Northern Europe including the Danish area was covered by a more or less isolated, wide sea. The climate was warm and dry and the intense evaporation from the sea surface resulted in the water becoming saturated with

salt. The salt precipitated out of solution and collected as deposits on the sea bed. In the deepest part of the sea, salt layers of more than 1000 m thickness were deposited. The Fehmarnbelt area was, however, situated immediately south of the Jylland/Fyn ridge, which during this period was an island in the central part of the sea. The thickness of the salt layer is therefore 200-300 m less below Fehmarnbelt than in the central part of the former sea.

Because the density of the salt is less than the density of the consolidated sedimentary soil/rock covering the salt deposits and because the salt can deform in a viscous manner under pressure, it has a tendency to move upwards towards the surface forming salt pillows, sometimes further developing to domes and even to diapirs. A very large number of salt structures formed by this mechanism are found in the former Zechstein Sea area. Two of them are situated within or so close to the investigated area that their presence may be of some importance for the fixed link project. Both of them are in the pillow stage, in which the lateral movement of salt from surrounding areas flow to a central thickening area resulting in uplift of the overlying layers. The process is presumed to be still on- going.

The presence of one salt pillow below the eastern part of the investigated area for the Fehmarnbelt Fixed Link and another one immediately north of the land connection in Rødbyhavn might influence the project in three different ways:  The vertical movements. It is important to be aware that surface movements above salt structures can be directed both upwards and downwards. This is a result of the salt flowing against the central part of the dome and lifting the surface above. This is flowing in from a “peripheral rim” around the culmination area, making the surface above subsidise. In /8/ information from literature on measured and calculated rates of heave/settlement of the surface above a number of dome structures in Northern Europe are presented. Moreover, information/indications on heave rate of the Fehmarnbelt dome have been collected from the seismic profiles. It is concluded that a heave rate in the range of 0.05-0.5 mm/year is a conservative (upper bound) estimate of the on-going movements of the surface above the Fehmarnbelt dome.

The investigations have revealed that the northernmost part of the fixed link area is situated on the southern flank of the Rødby dome. Based on observations of the thickness and variation in the surface level of the two till units, it has been concluded that the annual heave of this area might be similar to or slightly larger than the heave of the area above the Fehmarnbelt dome.

The deep “hole” in the surface of the Palaeogene clay documented by 09.A.004 might indicate that this boring is situated in a “peripheral rim” area around the Fehmarnbelt dome. However, if this was the case, the "rim" should be clearly visible in the chalk surface too, but the 1996 deeper seismic profiles give no clear evidence for that. De- tails on the 2008/2010 seismic profiles indicate that the subsidence rate of the surface above the area for the possible “rim” might have been significant earlier in the Quaternary period but has been very small or even absent in the last 20,000 years.

 As the salt lifts the overlying strata they may experience tensile fractures (circumfe- rential and radial to the axis of the dome) induced by the anticlinal movement generated by the salt movement at depth.

In the recent boring campaigns attention has been directed to the character of the chalk, but neither in the CPTUs nor the observations during the core descriptions has any sign of karst phenomena or very dense fracturing been observed. However, two “abnormal observations” have been registered: the first is that thin streaks and small bodies of “black matter” have been found in quite a number of the chalk cores. Microfossil analyses performed by GEUS indicate that the black material could be dark clay from the Æbelø Formation, which is the deposit locally situated directly above the chalk. The second observation is that the chalk in some of the samples shows signs of possibly to have undergone slumping after deposition.

It was hoped that the televiewer logs would have delivered important information regarding the character and quality of the chalk. However, in only one of the (shorter) chalk borings was the logging with the optical televiewer successful. The logging profiles for this boring showed no “abnormal” structures/fabric.

 A pattern of normal faults often develop in the layers above and beside a dome. The 1996 seismic investigations and to a lesser extent the 2008 and 2010 seismic investigations have shown a number of smaller faults at the Fehmarnbelt dome. Theoretically, it is a possibility that sudden local movements along faults accompanied by smaller earthquakes may happen. However, the risk for such movements is evaluated to be extremely small.

In this connection it should, moreover, be observed that the area, as described in Chapter 10.3, has been assessed to be a very low seismicity area, and that no significant earthquakes with epicentres close to the Fehmarnbelt dome have ever been registered. 10.3 Seismicity The region including the Fehmarnbelt area is considered to be a stable continental region, meaning that there are no active tectonic plate boundaries close to the area. Tectonic maps show that the regional tectonics almost only contains historic failed rifts and also historic sutures. Nevertheless, the area could contain unidentified zones of weakness where future earthquakes could occur. A number of small scale faults have been identified within the Fehmarnbelt Fixed Link alignment during the ground investigation for the project. However, these small scale faults are not of sufficient dimensions to generate significant earthquakes.

A study of readily available seismic hazard assessments carried out for the region has been performed and reported in /7/. The conclusion is that the Fehmarnbelt location is distant from any active zone and can be considered as located in a very low seismicity area. In Eurocode 8, EN 1998:2004 clause 3.2 it is stated that the provision of EN 1998 does not need to be observed for such areas.

The status as a very low seismicity area has been confirmed by the study of historic earthquakes registered in relevant international and European earthquake catalogues that has

been undertaken and reported in /7/. Only one single event from areas within a distance of 150 km from the fixed link is registered. This was an earthquake that occurred in 1629, but according to the USGS Significant Event register there is uncertainty both with regard to the location and magnitude of this event.

It is estimated that the peak horizontal acceleration on rock in the region with a return period of 475 years is in the range of 0.01 g to 0.02 g. Amplification of the bedrock ground motion would be expected to occur in the overlying soils above bedrock. Eurocode 8 re- commends site amplification factors of 1.35 to 1.80 depending on the ground conditions. The peak horizontal acceleration at the soil surface with a return period of 475 years is therefore estimated to be in the range of 0.014 g to 0.036 g.

It is noted that the Fehmarnbelt Fixed Link does not readily fall within the scope of normal codes such as EN 1998 and that there is neither a Danish National Annex for imple- mentation of Eurocode 8 nor a national seismic hazard map for Denmark. 10.4 Ground Water Conditions It is anticipated to be important for the fixed link project to obtain information on the presence of water bearing sand layers in the upper part of the series of layers in one or both of the coastal zones. From the general knowledge on the Quaternary series of layers it was expected that permeable meltwater sand layers might be present at relevant depths in the investigated area on Fehmarn, while it was evaluated as less probable that such would show up at relevant depths in Lolland (/9/).

10.4.1 Lolland Most of the onshore borings on Lolland in, or adjacent to, the alignment for the fixed link have shown the expected series of layers with a thin, upper Postglacial marine sand layer above a layer of clay till that continues down to the surface of the Palaeogene deposits. An exception to this pattern is boring 09.A.701 which encountered an almost 2.5 m thick meltwater sand deposit between the depths of 15.0 m and 17.5 m. Furthermore in 09.A.702 two very thin meltwater sand layers have been detected in the clay till. To get a better idea if significant water bearing layers are present, boring 10.A.071 was performed in 2010. This boring passed an approximately 2.5 m thick meltwater deposit at 20 m depth, but this was dominated by clayey silt and contained only a 10 cm thick sand layer.

A little to the east of the alignment below the production site area the same, clay/silt dominated meltwater deposit has been detected in several borings, but also meltwater sand deposits have been located in thin, more or less local occurrences below the area (/9/, /45/).

The conclusion is that there are probably rather outspread meltwater deposits in the glacial series of layers in the coastal area of Lolland, but that they only include thin – if any - high permeability water bearing horizons.

Two of the offshore borings, 09.A.018 and 09.A.014 are located rather close to but still more than 1 km from the Lolland coast. Boring 09.A.018 is situated closest to the coast; between elevation -18.5 and -33.0 m a thick deposit that seems to be dominated by meltwater sand has been encountered. Unfortunately, it is uncertain how much meltwater sand

is actually present in the unit, as there has been significant core loss (“no recovery”) at this location. The same unit seems to be present in 09.A.014 between elevations -23.0 m and -35.3 m. Also in the older boring 96.0.006 situated a little further from land than the above mentioned borings, the special meltwater sand dominated unit is present.

In conclusion, no thick, outspread aquifers are present in the coastal area, but smaller groundwater reservoirs might be present in the following geological formations:  In the abutment/portal area a Postglacial marine deposit dominated by sorted, water bearing sand but with significant gyttja layers is present. The layer can have a total thickness of 4.5 m in the area close to the coastline.  Furthermore, a thick and outspread unit dominated by meltwater sand layers but interrupted by thin clay till occurrences seems to be present in elevations typically between elevations -22 and -35 m in the area close to land in the Rødbyhavn area. The layer may be present at the shore as a thin layer of meltwater sand as seen in borings 10.A.071 and 96.0.001.  East of the abutment/portal area, local meltwater sand deposits are more common.

10.4.2 Fehmarn A total of ten borings have been performed onshore on Fehmarn as part of the 2009-2013 Boring Campaigns. Most of them have encountered variable thicknesses of up-thrusted floes of Palaeogene clay within the clay dominated till deposits.

According to the borings, layers of meltwater sand/gravel are only present in the glacial till in small parts of the coastal land area. This conclusion is based on the fact that the deep boring 09.A.603 met no meltwater sand at all, 09.A.604 only passed a very thin sand layer at a depth of 42 m (below the thick floes of Palaeogene clay) and 09.A.605 which was taken down to a depth of 50.5 m without reaching the surface of the Palaeo- gene clay, only met a 0.6 m thick sand layer. On the western side of the railroad, 09.A.606 detected a rather thick unit dominated by meltwater sand but including clay till layers at a depth between 8.0 and 12.5 m, while further inland 10.A.610/610A only encountered three very thin sand layers in the clay dominated till. In contrast to the above locations, quite a different situation is encountered in the area where 09.A.601, 09.A.602 and 09.A.607 and 10.A.607/607A are located. In this area the disturbed series of layers below the upper clay till contains major floes of both meltwater sand/gravel and Palaeo- gene clay. Such conditions have also been found in boring 12.B.651.

In the offshore borings located closest to the Fehmarn coast the only sand detected is Postglacial marine sand in the surface layers of 09.A.001, 09.A.002, 09.A.003, 09.A.004 10.A.051, 10.A.052 and possibly as a very thin layer in 09.A.005. In the older part of the series of layers no sand seems to be present in the near-shore coastal area. Also boring 10.A.072, which was performed specifically to look for possible water bearing sand layers, only found clay deposits below the upper Postglacial marine sand.

There is no indication of sand layers in the geophysical CVES/Mep lines except for in the north-western part of the mapped area where the surface of the glacially disturbed Palaeo- gene clay deposits is estimated to be at a deep level. This sand layer has been verified by boring 09.A.606 in which sand was found between elevation c. -8 m and -12 m.

In the area where the borings 09.A.601, 09.A.602 and 10.A.607/607A are located, rather thick floes of meltwater sand/gravel seem to be present at depths greater than 10 to 12 m. The horizontal extent of these water bearing zones is not known, albeit sand layers of several hundreds of meters maybe encountered in the area. However, given that the glacial layers are considered to be disturbed, folded and faulted, it is not likely that sand layers of much larger extent may be met. 10.5 The Series of Layers The series of layers are described in detail in the succeeding Chapters 11 through 14 and the corresponding appendices.

11 Postglacial and Lateglacial Deposits 11.1 Geological Description An overview of the Postglacial/Lateglacial history of the Baltic and the proposed rela- tions between the historic episodes and the deposits in the area is illustrated in Table 11- 1.

It should be noted that, except for the Littorina Sea/present day sea stage in Table 11-1, the described episodes and related deposits are only relevant to the areas within the “basin”. The reason for this is that the relative water level was so low in the Lateglacial and early Postglacial time that the basin slopes were dry land at that time. No deposits were formed there except in local lakes/bogs where freshwater gyttja and peat were deposited. Such a bog has been detected below the present day seabed in the southern slope area. Finally, in the period for the Ancylus Lake the water level was still rising, and the slope areas nearest to the basin were incorporated in the lake.

Table 11-1 Lateglacial/Postglacial development of the Baltic, including the actual area Stage Period (before Typical deposit present) Littorina Sea/Present sea 7,800–now Sand in exposed areas. Gyttja locally in the deepest, low oxygen parts and in former bay on Lolland Ancylus Lake 9,000–7,800 Mostly silty clay. Sometimes layered/laminated. Yoldia Sea, higher salinity 9,500–9,000 Medium plasticity clays, often without visible layering. A few shells. Maybe near-shore sand deposits. Yoldia Sea, low salinity 10,250–9,500 Mostly silty clay and silt without a distinct layering. Local sand deposits in areas where rivers entered the lake. No shells. Baltic Ice Lake 13,000-10,250 Distinctly layered/laminated (“varved”) clay/silt. Maybe sand deposits in areas for inflow to the sea. Inside the shoreline, Allerød peat and freshwater gyttja. Meltwater stage 15,000-13,000 Meltwater sand and gravel. End of glaciations in area 15,000 Clay till deposition stops.

11.2 Geotechnical Properties 11.2.1 General The geotechnical properties in these soils have been investigated through:  Classification testing by Fugro of samples from the type A-borings (/2/, /23/).  In-situ testing (CPTU) by Fugro in the type B-borings (/2/, /23/).  In-situ testing (CPTU) by Fugro in the type C-borings (seabed CPTUs) (/1/).  Classification testing by GEO of selected samples from the type A-offshore borings  Advanced Laboratory Testing (/27/).  In situ testing (CPTU) by Ramboll Sweden in the type B borings for the production site, cf. /45/.  Classification testing by GEO of selected samples extracted from the type A borings for the production site, cf. /44/.  Advanced geotechnical testing by GEO of selected samples from the type A-borings (/27/).

Details of the geotechnical properties for Postglacial and Lateglacial deposits can be found in Appendix GDR 00.1-001-B.

11.2.2 Classification Properties An overview of the geotechnical classification properties for the Postglacial and Late- glacial deposits appears from Table 11-2 and Table 11-3.

Table 11-2 Basic geotechnical classification properties for Postglacial and Lateglacial deposits

Soil type w  wL wP IP Postglacial Arithmetical mean 25.9% 18.3 kN/m3 - - - marine Standard deviation 12.3% 1.6 kN/m3 sand/gravel Number of tests 75 19 Postglacial Arithmetical mean 104.2% 13.7 kN/m3 96.3% 47.1% 49.2% marine Standard deviation 68.0% 2.0 kN/m3 60.0% 35.9% 28.5% /freshwater Number of tests 53 30 21 21 21 gyttja and peat Postglacial/Late Arithmetical mean 33.6% 18.6 kN/m3 42.4% 18.3% 24.1% glacial marine Standard deviation 14.1% 1.8 kN/m3 14.2% 4.9% 10.5% /freshwater Number of tests 205 145 59 59 59 clay/silt Postglacial/Late Arithmetical mean 20.2% 20.5 kN/m3 - - - glacial marine Standard deviation 4.8% 2.5 kN/m3 /freshwater Number of tests 35 13 sand/gravel

Table 11-3 Additional geotechnical classification properties for Postglacial and Lateglacial deposits

Soil type ds e Clay T CaCO3 gl Postglacial Arithmetical mean 2.62 0.76 8.2% ‒ 1.1% marine Standard deviation 0.03 0.5 8.5% 1.4% sand/gravel Number of tests 11 8 31 15 Postglacial Arithmetical mean 2.45 3.41 13.9% 9.8% 12.6% marine Standard deviation 0.31 1.37 8.4% 13.4% 17.5% /freshwater Number of tests 11 15 36 4 17 gyttja and peat Postglacial/Late Arithmetical mean 2.67 0.95 30.0% 17.2% 2.5% glacial marine Standard deviation 0.04 0.43 15.5% 4.6% 1.2% /freshwater Number of tests 30 72 122 7 42 clay/silt Postglacial/Late Arithmetical mean 2.67 0.68 7.8% ‒ 0.7% glacial marine Standard deviation 0.02 0.20 8.0% 0.7% /freshwater Number of tests 8 8 22 9 sand/gravel

The plasticity chart for the Postglacial marine/freshwater gyttja and marine/freshwater clay/silt is included as Figure 11-1.

Figure 11-1 Plasticity chart for Postglacial/Lateglacial deposits

11.2.3 CPTU A considerable number of the type B-borings with CPTU and the seabed CPTUs (type C- borings) have penetrated the Postglacial and Postglacial/Lateglacial deposits.

The data compilation and data processing for the CPTUs have been described in /60/.

It appears that the corrected and correlated qnet generally varies:  Between 1 and 40 MPa and generally increasing with depth for the Postglacial / Late- glacial sand.  Between 20 and 300 kPa and generally increasing slightly with depth for the Post- glacial gyttja.  Between 100 and 1000 kPa and generally increasing slightly with depth for the Post- glacial/Lateglacial silt/clay.

It shall be noted that the CPTU data have been sorted and filtered as described in /60/.

11.2.4 Stress and Stress History Based on the oedometer tests the pre-consolidation stresses may be assessed. Although these stresses correspond to over-consolidation ratios (OCRs) equal to one or above, these deposits should generally be anticipated to be normally consolidated corresponding to OCR = 1.

The coefficient of earth pressure at rest K0 has for the Postglacial and Lateglacial deposits been assumed to:  0.7 for gyttja, peat and clay/silt,  0.5 for sand/gravel.

11.2.5 Consolidation Properties

The laboratory coefficient of consolidation ck has for the deposits been determined between c. 10-8 and 10-5 m2/sec.

The compression ratios vary and the results of the tests are included in Table 11-4.

Table 11-4 Compression ratios for Postglacial and Lateglacial deposits Soil type Q Postglacial Arithmetical mean 32.7% marine/freshwater gyttja Standard deviation 15.0% and peat Number of tests 5 Postglacial/Lateglacial Arithmetical mean 9.2% marine/freshwater Standard deviation 3.7% clay/silt Number of tests 6

An initial very rough prediction of the compression ratio Q [%] may be obtained through the correlation qnet/7 for Postglacial marine/freshwater gyttja and peat and qnet/76 for Postglacial/Lateglacial marine/freshwater clay/silt where qnet has units of kPa.

The rate of secondary consolidation Cα (or εs) must generally be expected to be considerable. For a peat specimen Cα has been determined to vary between 2.0 and 3.8 %/lct [log10 cycle of time], increasing with increasing vertical effective stress level.

The laboratory coefficient of permeability k has been determined at the vertical effective in-situ stress level σ′v0 to be:  Between c. 10-9 and 6·10-9 m/sec for the Postglacial marine gyttja and peat.

 Between c. 0.1·10-9 and 2·10-9 m/sec for the Postglacial/Lateglacial marine/freshwater clay/silt.

11.2.6 Static Shear Strength The undrained shear strengths as determined by advanced laboratory testing are summarised in Table 11-5.

Table 11-5 Undrained shear strength for Postglacial and Lateglacial deposits C E DSS Soil type cu cu cu Postglacial Arithmetical mean 32 kPa - - marine/freshwater gyttja Standard deviation 25 kPa and peat Number of tests 3 Postglacial/Lateglacial Arithmetical mean 33 kPa 13 kPa 13 kPa marine/freshwater Standard deviation 12 kPa clay/silt Number of tests 5 1 1

With due regards to the rather few tests and the considerable variation of the cone factors, C a cone factor (related to cu ) of 15 for Postglacial marine/freshwater gyttja and peat and for Postglacial/Lateglacial marine/freshwater clay/silt is suggested to be applied for pre- C diction of cu based on qnet.

The lower bound value of the effective strength parameters for Postglacial marine/freshwater gyttja and peat as well as for Postglacial/Lateglacial marine/freshwater clay can be estimated to φ′ = 20.4° and c′ = 8 kPa. However for lower effective stress levels the effective strength parameters for Postglacial marine/freshwater gyttja and peat as well as for Postglacial/Lateglacial marine/freshwater clay are assessed to be: φ' = 22°, c' = 5 kPa

Based on the sorted cone resistance in the B-borings with an adjacent A-boring in Post- glacial and Lateglacial sand, typical triaxial peak friction angles φ' = 30°- 40° have been assessed. The lower part of these friction angles may be due to the presence of silt or clay in the strata.

Anisotropy factors have been addressed for a single series of tests of Postglacial / E DSS C Lateglacial silt/clay and it appears that cu ≈ cu ≈ 0.5 cu . 11.2.7 Geophysical Properties Due to concern for the stability of the borehole, neither the Postglacial sand nor the basin deposits have been characterised by downhole geophysical log data.

12 Glacial Deposits 12.1 Geological Description The series of glacial layers are divided into six different deposits. These deposits, listed in the geological sequence typically encountered, are:  Upper till deposit(s)  Meltwater Silt/Clay deposit  Meltwater Sand deposit  Chalk till deposit  Lower till deposit  Lowermost till deposit

During the initial phases of the project, focus was given to the deposits. As the project has evolved, these deposits have been grouped into Glacial units, defined by:  Upper Till Unit: Upper till deposit(s) and Meltwater silt/clay deposit  Lower Till Unit: Lower till deposit, Lowermost till deposit and Chalk till deposit  Meltwater Sand Unit: Meltwater sand deposit as encountered in the borings 09.A.018 and 10.A.065

The upper till deposit is typically a hard to very hard, sandy to very sandy clay till. Lo- cally it includes meltwater sand layers/floes, and also floes of the lower till deposit seem to be present locally within it too. A number of observations indicate that the upper till deposit includes two till deposits with remnants of a separating meltwater deposit present locally.

It has been observed that a system of closely spaced, vertical fractures at least in parts of the area intersects the upper till(s).

Deposits of meltwater sand and meltwater silt/clay separate the upper till deposit from the lower till unit in parts of the area. The sand varies widely both in grain size and degree of sorting over the area and both coarse, very gravelly sand and fine grained, very silty sand is included in this deposit. Because of poor recovery, the more precise composition of the layer is not known.

The lower till deposit typically consists of medium plasticity clay till.

The lowermost till deposit typically consists of high plasticity clay till.

The chalk till deposit consists of very calcareous clay till.

Floes of meltwater sand and of clay of Palaeogene origin are present within the lower till unit.

The lowermost till deposit of high plasticity clay till is almost consistently situated below the lower till deposit of medium plasticity clay till with a well-defined boundary. This indicates that the lowermost till deposit probably is deposited during a separate, older glacial event.

The upper, folded part of the underlying Palaeogene clay could also be regarded as a part of the lower till unit, but the authors have chosen to use the term "Clays of Palaeogene origin" to underline that they originally were deposited during the Palaeogene period, but that later geological events may have changed the character of the clay.

It is noted that cobbles and boulders are rare, however it is important to be aware that these are certainly present in the till deposit. The occurrence of cobbles and boulders in the borings has been analysed in /60/ and in /41/. 12.2 Geotechnical Properties 12.2.1 General

The geotechnical properties of the Glacial deposits have been investigated through:  Classification testing by Fugro of samples from the type A-borings /2/ and /23/.  In-situ testing (CPTU) by Fugro in the type B-borings /2/ and /23/.  In-situ testing (CPTU) by Fugro in the type C-borings (seabed CPTUs) /1/.  Classification testing by GEO of selected samples extracted from the type A-borings for Advanced Laboratory Testing /28/.  Advanced geotechnical testing by GEO of selected samples extracted from the type A-borings /28/.  In-situ testing (CPTU) by Ramboll Sweden in the type B borings for the Production site, cf. /45/.  Classification testing by GEO of selected samples extracted from the type A borings for the Production site, cf. /44/.  In-situ testing (CPTU) by GEO in the type B borings for the Alignment Lolland and Fehmarn, cf. /46/.  Classification testing by GEO of selected samples extracted from the type B borings for the Alignment Lolland and Fehmarn, cf. /46/.

Details of the geotechnical properties for Glacial deposits can be found in Appendix GDR 00.1-001-C.

It must be noted that the geotechnical properties for the floes of clays of Palaeogene origin have been described under the geotechnical properties for clays of Palaeogene origin.

12.2.2 Classification Properties Review of the results of the laboratory classification testing, undertaken on samples of the Glacial deposits, confirms the presence of individual Glacial deposits. These deposits are particularly well defined from a review of the plasticity chart, plasticity index and clay content test results, where the deposits fall into groupings or are evident through stepped variations of geotechnical parameters.

Assessment of the plasticity test results reveals a significant variance of plasticity between the Glacial deposits. The test results indicate that the Upper and Chalk till deposits are typically of low plasticity, whilst the Meltwater silt/clay and the Lower till deposits are of low to medium plasticity. The plasticity properties of the Lowermost till deposit are noted to be highly variable, and range from medium to very high plasticity.

The plastic and liquid limit results within the Upper till, Chalk till and Lower till deposits are shown to be relatively consistent within each of the deposits; with the exception of occasional elevated liquid limit results. Few tests were undertaken on samples of the Meltwater silt/clay deposit; with the results indicating irregular ranges of plastic and liquid limits. The results of Atterberg Limits tests undertaken on samples of the Lower- most till deposit reveal a significant range between the liquid and plastic limits which appear variable with depth.

With the exception of the Lowermost till deposit, the typical values of activity within the Glacial deposits are less than 1.25, and thus, the soil can be classified as ‘in-active’ or ‘normal’. However, the range of activity values within the Lowermost till deposit typically range between 0.90 and 2.00 indicating proportions of the unit are ‘active’. Given this, the Lowermost till deposit may be prone to swelling and/or shrinkage with changes in water content.

An overview of the geotechnical classification properties for the Glacial deposits is provided in Table 12-1 and Table 12-2.

It should be noted that the Meltwater sand deposit according to the soil descriptions includes thin layers of clay till and very silty meltwater sand.

Table 12-1 Basic geotechnical classification properties of Glacial deposits

Glacial deposit w wL wP IP IL  Arithmetical mean 10.3% 20.0% 11.2% 8.8% -0.13 23.0 kN/m3 Upper till Standard deviation 2.6% 2.8% 1.4% 2.4% 0.25 1.1 kN/m3 Number of tests 1100 316 315 315 218 421 Arithmetical mean 20.4% 33.3% 17.3% 16.1% 0.38 20.7 kN/m3 Meltwater 3 Standard deviation 7.1% 15.7% 6.8% 10.2% 0.55 1.3 kN/m Silt/Clay Number of tests 67 20 20 20 7 25 Arithmetical mean 16.9% - - - - 21.1 kN/m3 Meltwater 3 Standard deviation 5.9% 1.8 kN/m Sand Number of tests 98 44 Arithmetical mean 12.1% 22.7% 14.0% 8.7% -0.30 22.4 kN/m3 Chalk till Standard deviation 4.2% 4.5% 1.9% 4.3% 0.21 1.2 kN/m3 Number of tests 94 42 42 42 20 63 Arithmetical mean 11.6% 28.0% 12.3% 15.7% -0.05 22.6 kN/m3 Lower till Standard deviation 3.1% 9.2% 2.4% 7.7% 0.24 1.1 kN/m3 Number of tests 736 198 198 198 55 613 Arithmetical mean 17.7% 58.8% 18.9% 40.0% -0.02 21.4 kN/m3 Lowermost 3 Standard deviation 5.8% 23.7% 5.5% 19.5% 0.15 1.4 kN/m till Number of tests 273 96 96 96 20 209

Table 12-2 Additional geotechnical classification properties of Glacial deposits

Glacial deposit e ds Clay T Activity CaCO3 Arithmetical mean 0.27 2.65 14.4% 0.61 25.5% Upper till Standard deviation 0.10 0.04 7.1% 0.27 8.0% Number of tests 117 75 260 79 10 Arithmetical mean 0.58 2.68 20.7% 0.92 - Meltwater Standard deviation 0.11 0.04 11.9% 0.42 Silt/Clay Number of tests 12 7 32 9 Arithmetical mean 0.49 2.64 11.1% - 30.8% Meltwater Standard deviation 0.17 0.04 12.4% - deposits Number of tests 23 16 61 1 Arithmetical mean 0.31 2.68 27.5% 0.44 59.7% Chalk till Standard deviation 0.04 0.04 7.7% 0.38 11.0% Number of tests 26 17 37 21 8 Arithmetical mean 0.30 2.66 20.1% 0.76 23.7% Lower till Standard deviation 0.07 0.04 7.5% 0.29 11.8% Number of tests 236 98 303 124 31 Arithmetical mean 0.47 2.66 29.2% 1.38 21.7% Lowermost Standard deviation 0.20 0.04 11.1% 0.67 6.6% till Number of tests 66 39 161 62 6

The plasticity chart for the Glacial deposits is included as Figure 12-1.

Figure 12-1 Plasticity chart for Glacial deposits

12.2.3 CPTU In order to evaluate the stratification and to allow comparisons between laboratory test results with cone resistance profiles, a number of CPTUs were undertaken adjacent to A- borings. Consequently, the following typical ranges of net cone resistance, qnet, have been determined for each of the following Glacial deposits:  Upper till deposit– typically between 10 and 35 MPa (see following text)  Meltwater silt/clay deposit– typically between 1 and 40 MPa  Meltwater sand deposit– typically between 20 and 50 MPa  Lower till deposit– typically between 4 and 20 MPa  Lowermost till deposit– typically between 2 and 6 MPa at depths less than 25 m and typically between 4 and 15 MPa at depths greater than 25 m

It should be noted that a significant number of CPTUs undertaken within the Upper till deposit met refusal before the test push could be completed. Review of cone resistance profiles suggest that approximately 40 % of CPTU pushes typically met refusal in the Upper till deposit. Given that the area where the piezocone met refusal was drilled out, it is unknown whether a part of these refusals were a result of the piezocone encountering cobbles / boulder obstructions or whether the soil strength in all the actual tests exceeds the limit of the test equipment. As a result, the reported upper limit of the typical range of the Upper till deposit should be adopted with significant caution.

An accurate profile of cone resistance could not be determined within the Chalk till deposit as the piezocone repeatedly met refusal. As a result, a limited number of net cone resistance values were obtained within the Chalk till deposit and are not considered representative of the strata strength.

The data compilation and data processing for the CPTUs have been described in /60/.

12.2.4 Stress and Stress History In order to identify the stress history, a number of oedometer tests were carried out on samples of the till deposits.

It has been attempted to correlate the net cone resistance and the estimated pre-consolidation pressure for the tills. Besides the challenges already pointed out in Chapter 12.2.3, the correlations have also been influenced by the net cone resistance that carries a large scatter with jumps, reflecting e.g. gravel grains and variability within the soil formation. Because of these variations no correlation could be established between pre-consolidation pressure and the net cone resistance for the Upper till deposit. However, for the Lower till deposit, an estimate of the pre-consolidation pressure may be found using the correlation: σ′pc = 0.3qnet. It is suggested that this correlation may be used for all the clay till formations applying the rule qnet ≤ 8000 kPa.

Following the correlation between qnet and 'pc, OCRs were found to range between 5 and 25 for the Upper till deposit and between 2.5 and 13 for the Lower till deposit.

To establish a coefficient of earth pressure at rest, the results of incremental loading oedometer tests with measurement of horizontal stress were reviewed. Although this review concluded no apparent trends, GEO have determined equations to fit the upper and lower bound test results for both the Lower till and the Chalk till deposits. For the 0.40 upper bound results, K0 is approximately 0.42OCR and for the lower bound results, 0.28 K0 is approximately 0.42OCR . 12.2.5 Consolidation Properties A number of incremental loading oedometer tests have been undertaken on samples of the Glacial deposits in order to determine the constrained secant oedometer modulus,

Eoed,sec, the coefficient of consolidation, ck, the compression ratio, Q, and swelling ratio, Qun, the rate of secondary consolidation Cα (= εs) and the rate of secondary swell, Csw.

It is apparent that the derived values of Eoed,sec are markedly different between Glacial deposits and have been shown to increase with depth. The minimum derived values of

Eoed,sec, for a stress increase not exceeding 500 kPa from the in-situ vertical effective stress, have been shown to be:  600 MPa for Upper till deposit;  1114 MPa for Chalk till deposit;  83 MPa for Lower till deposit; and  31 MPa for Lowermost till deposit.

It should be noted that /28/ has devised a series of trend lines to predict lower bound values of Eoed,sec for each Glacial deposit. The adjusted trend lines predict Eoed,sec values based on the plasticity index of the soil, provided that the unloading stress decrement

(σ′unl) is between 120 kPa and 500 kPa, cf. Table 12-3.

Table 12-3 Trend lines for lower bound values of oedometric secant modulus, where Eoed,sec = A + B·σ′unl (kPa) and 120 kPa < σ′unl < 500 kPa

Glacial deposit Ip [%] A [kPa] B Upper till < 10 200·103 1000 Chalk till - 200·103 500 < 10 40·103 750 10–14 20·103 750 Lower till 14–18 0 750 > 18 0 500 Lowermost till - 0 250

The compression ratio, Q of each of the Glacial deposit were found to be highly variable; with values in the Upper till deposit ranging between 1.9 and 6.0 % and between 3.1 and 7.1 % in the Lower till deposit. However, GEO have established an approximate relationship between Q and plasticity index, Ip. For Glacial till with an Ip less than 9 %, Q varies between 2 and 4 %, and, with an Ip between 10 and 16 %, Q varies between 4 and 7 %. Similarly, samples of till with an Ip of approximately 38 % were found to have a Q value in excess of 8 %.

Average measured Qun values are 0.3 % for Upper till deposit, 0.7 % for Lower till deposit and 2.3 % for Lowermost till deposit.

Review of the creep properties noted a maximum value of Cα along the initial loading curve of 0.3 %/lct and a maximum value of Cα along the reloading curve of 0.1 %/lct. In both the loading and reloading testing, the rate of secondary consolidation was found to increase with increasing effective stress.

/28/ indicates that Glacial deposits with an Ip less than 20 % have no swell potential.

The coefficient of permeability, k (hydraulic conductivity), has been determined during CRS and IL testing at the corresponding vertical effective in-situ stress level, σ′vo, at reloading. These tests have concluded average k values in the laboratory ranging from 3 to 50·10-12 m/s for the Upper Till, 2 to 2010-12 m/s for the Chalk till, 5 to 5010-12 m/s for the Lower till and 2 to 410-12 m/s for the Lowermost till. Such values would indicate that these Glacial deposits have very low permeability.

Danish experience with the till deposits reveals a field value of the coefficient of permeability of 10-8 to 10-7 m/s. This value is higher than that measured in the laboratory; part of this difference may be that water will flow through cracks in the soil providing a higher mass permeability. The coefficient of consolidation should be based on this field value together with the constrained oedometer moduli given above.

12.2.6 Static Shear Strength

The uniaxial compressive strength (σc) of Upper till deposit is generally in the range 0.4 to 1.35 MPa with a typical value of 0.8 MPa. The measured Brazil tensile strength (σt) is typically in the range 0.1 to 0.25 MPa with a typical (σc/σt) ratio of 5.9.

C The undrained shear strength of the Upper till deposit in triaxial compression (cu ) is measured to be in the range 650 kPa to 1200 kPa, which is approximately twice the undrained shear strength indicated by the UCS testing assuming cu = σc /2. It has been observed that the Upper till deposit is intersected by a closely spaced net of vertical fractures and it is probable that the low strengths obtained from UCS testing are caused by the vertical fractures in the Upper till deposit.

In the triaxial compression and extension tests, and in direct simple shear tests, the undrained shear strength of Lower till deposit and Lowermost till deposit has been determined to be in the range 130 kPa to 670 kPa.

The Chalk till deposit shows strong dilatants behaviour when tested in triaxial compression, with correspondingly high undrained shear strengths that increase with increasing strain. The undrained shear strength in triaxial compression is in the range 1200 kPa to 2300 kPa.

A number of laboratory tests were undertaken on samples of Glacial deposits that were recovered from borings drilled within a distance of 5 m from CPTUs. The laboratory strength testing results was thus correlated with the CPTU results.

The net cone resistance qnet is often considered directly proportional to the undrained shear strength (cf. e.g. /16/). The undrained shear strength is, however, not unambiguous- C C ly defined. Compared to cu (CAUc-tests), /14/ indicates a cone factor Nkt (= qnet/cu ) ~ 10 for clay till. In ref. /15/ an average cone factor Nk (~ qc/cv) ~ 10 has been found for clay till of the Storebælt till type 0-1.

Site correlations show a high variability in the cone factors calculated for the Glacial tills and although there are a limited number of correlated tests, the average Nkt values are generally in agreement with past experience. Because of the high variability in the cone fac- C tors it is proposed that a common Nkt value of 12.0 is adopted when predicting cu on basis of qnet for all Glacial tills. Based on the CPTU profiles and this cone factor both lower and higher undrained shear strengths than the range stated above for the advanced laboratory tests must be expected to be present in the Glacial tills

Effective shear strength parameters have been assessed using undrained triaxial compression and extension tests with pore water pressure measurements together with drained triaxial compression tests. The estimates are included in Table 12-4 using linear regression on test data.

Table 12-4 Estimates of effective shear strength properties Glacial deposit φ′ [] c′ [kPa] No. tests Upper till 33.4 54 5 Meltwater Sand 37.6 44 3 Chalk till 36.2 99 5 Lower till 36.2 0 8 Lowermost till 31.3 0 1

Danish experience notes that the undrained shear strength of clay till can be related to the pre-consolidation pressure and in-situ vertical effective stress using the SHANSEP procedure. The Upper till and Chalk till deposits have higher undrained shear strengths than the tills to which the SHANSEP procedure has previously been applied. For the Lower and the Lowermost till deposits, the undrained shear strength has been measured lower than what would be predicted by the SHANSEP approach using ´pc = 0.3qnet.

Compression, extension and direct simple shear measurements of the undrained shear strength at close proximity to one another demonstrate that the tills exhibit anisotropic behaviour. The limited number of tests carried out show large variability in the ratios between those values obtained, and it is not possible to provide reliable predictions of anisotropic factors for individual Glacial deposits. It is proposed that the following correlations be adopted for all tills:

DSS C  cu /cu = 0.75. E C  cu / cu = 0.65.

12.2.7 Geophysical Properties The P-wave velocities measured in the glacial deposits are in Table 12-5 separated into three parts: a part that is mainly floes of Palaeogene clay and upper and lower parts that are mostly till and sand. The change in measured velocities from the VSP does not always match the interpreted lithology. This could be due to the limited precision of the VSP method and the difficulties with interpreting velocity intervals in grading lithologies.

Table 12-5 Derived Sonic and VSP interval velocities in the Glacial deposits

Geology Sonic P-VSP S-VSP Int. Vel. [m/s] Int. Vel. [m/s] Int. Vel. [m/s] Range Average Range Average Range Average Upper till - - 1300-1800 1400 275-525 375 Lower till 1770- Glacial 1900 1875-2075 2000 300-700 450 2140 deposits Floes of 1570- Palaeogene 1585 1675-1725 1700 175-400 300 clay 1600

The shear modulus (G0), Poisson’s ratio (ν) and Young’s modulus (E) have been calculated as appearing from Table 12-6 using the results from the VSP measurements and the density log. Stiffness values are small strain values, reflecting an undrained response.

Table 12-6 Calculated parameters in the different glacial deposits from VSP

Geology G0 Poisson’s ratio Young’s module [MPa] [-] [MPa] Range Average Range Average Range Average Upper till 70-680 350 0.419-0.490 0.453 205-1930 1005

Glacial Lower till 85-1090 400 0.433-0.492 0.473 250-3120 1180 deposits Floes of Palaeogene 55-345 150 0.470-0.494 0.486 170-1010 435 clay

The range in velocities reflects the heterogeneity of the deposits which also appears from the full wave form data showing a very inconsistent pattern.

12.2.8 Small Strain Stiffness and Damping Small strain testing of the Glacial deposits proved difficult to achieve due of the inherent high strength and stiffness of these materials. Because of these constraints a limited number of laboratory tests has been completed to provide small strain stiffness data for the Glacial deposits and to define how soil stiffness and damping vary with strain. Testing has been limited to specimens of Lower till deposit and Chalk till deposit; no small strain testing has been undertaken for the Upper till deposit, the Lowermost till deposit or the meltwater deposits.

The test results indicate that small strain stiffness (G0) derived from VSP logging, and laboratory determined small strain stiffness (Gmax) derived from bender element and resonant column testing, generally vary between approximately 200 and 600 MPa. The lower values apply to the Lowermost till deposit and Lower till deposit at depths less than about 15 m, while the higher values were measured in Upper till deposit at depths less than about 15 m. Below 20 m depth the G0 values generally converge to values between 300 and 400 MPa.

Laboratory testing, downhole logging and CPTU results indicate Gmax/qnet and G0/qnet ratios in the range 20 to 100 with no distinct variation between the Glacial deposits.

Due to the size of the data set firm conclusions on how stiffness and damping vary with shear strain have not been determined. Normalised plots of G/Gmax follow the expected ‘S’ shape profile when a stiffness degradation profile is plotted against the log of cyclic shear strain, but the damping ratio values derived from cyclic triaxial tests appear to be high and do not show the expected trend of increasing damping ratio with increasing cyclic shear strain.

12.2.9 Cyclic Undrained Shear Strength To determine the number of cycles required to fail the Lower till deposit with a combination of average and cyclic shear stresses, i.e. reach 15 % shear strain, six cyclic undrained direct simple shear tests were undertaken.

Simplified diagrams have been used to assess the undrained cyclic shear strength. Such methods have indicated that the undrained cyclic shear strength equals approximately 70 % of the undrained static shear strength when number of cycles to failure N = 10. 12.3 Recompacted Clay Till Material as Embankment Fill 12.3.1 Introduction In Denmark there is a long tradition for using excavated clay till as earthworks material in e.g. railway and road embankments. Selected information about the existing knowledge of clay till as earthworks material in e.g. embankments has in /41/ been collected and compared with the results of a limited number of compaction and strength tests performed on selected samples of recompacted Upper and Lower till deposits from the fixed link area. These compaction and strength tests are summarised in Chapter 12.3.2 and the overall existing experience with earthworks using clay till is summarised in Chapter 12.3.3.

12.3.2 Compaction and Strength Properties of Recompacted Upper Till and Lower Till Deposits A total of 4 Standard Proctor tests and 4 anisotropic undrained triaxial compression tests have been performed.

Two of the Standard Proctor tests have been performed on mixed material of Upper till deposit from the geotechnical borehole 09.A.019 and the remaining two Standard Proctor tests have been performed on mixed material of Lower till deposit from the geotechnical borehole 09.A.006.

The main results of the 4 Standard Proctor tests are summarised in Table 12-7.

Table 12-7 Results of Standard Proctor tests Natural water Optimum water Maximum Material content content dry density [%] [%] [Mg/m3] Upper Till deposit 9.9 9.3 2.04 (Proctor 03) Upper Till deposit 9.9 8.0 2.07 (Proctor 04) Lower Till deposit 13.0 10.9 1.95 (Proctor 01) Lower Till deposit 13.0 11.2 1.94 (Proctor 02)

It appears from the Standard Proctor curves that the achievable degree of compaction, as expected, is very dependent of the water content at time of compaction for both the Upper and Lower till deposits. It furthermore appears that the optimum water content for compaction is approximately 1–2 % lower than the natural water content for the Upper till deposit and approximately 2 % lower than the natural water content for the Lower till deposit.

The material for the triaxial tests was taken from the same batch as the material used for the corresponding Standard Proctor tests and the triaxial samples have been prepared in the proctor mould. Two triaxial samples were prepared for each of the Upper and Lower till deposits. Trimming of the triaxial specimens have been performed in the central part of the proctor sample. One of the specimens for each deposit was established using intended 92 % SP (wet side) and the second specimen was established using intended 96 % SP (wet side). The triaxial specimens have been saturated and consolidated for a vertical effective stress ≈ 100 kPa (horizontal effective stress ≈ 57 kPa) before shearing.

Table 12-8 Main results of triaxial tests Natural water Corresponding Undrained content for degree of com- shear Material specimen paction strength [%] [% SP] [kN/m2] Upper Till deposit 12.4 94 60 (Proctor 03) Upper Till deposit 12.3 91 29 (Proctor 04) Lower Till deposit 13.2 96 47 (Proctor 01) Lower Till deposit 16.4 91 39 (Proctor 02)

The main results of the 4 triaxial tests have been included in Table 12-8. The table shows C that the ratio cu /′v varies between 0.3 and 0.6. 12.3.3 Overall Existing Experience Besides compaction, the geotechnical strength and deformation properties for the earthworks clay till material depend typically on:  The geotechnical properties such as the grain size curve and plasticity for the material.  The water content for the earthworks material at the time of compaction.

The water content in the remoulded clay till depends very much on the methods used for excavation, transportation and storing.

For clay till where excavation, transport, storing and placing of the material in large lumps can be expected, the conditions in relation to compaction, drainage, strength and deformation properties can be compared directly to work with clay till onshore. The clay till has to be considered as disturbed or remoulded with a change in water content and degree of saturation and this will influence the strength and deformation properties. The main challenge is the water content in the earthworks material prior to compaction. This water content will typically be increased due to precipitation and other exposures to water, and the problem with a too high water content is especially common for earthworks conducted in the time of the year where the precipitation exceeds the evaporation, which in Denmark in a normal year is between mid-October and early May.

13 Clays of Palaeogene Origin 13.1 Geological Description All the Palaeogene clay formations from the youngest Søvind Marl Formation overlying the Lillebælt Formation, the Røsnæs Formation, the Ølst Formation, the Holmehus For- mation, the Æbelø Formation and maybe also (traces of) the Kerteminde Marl Formation are present below the Belt. Dating of 163 samples of Palaeogene clays, /6/, has formed the basis for the classification of the clay into Formations as shown on the longitudinal profile in Enclosure I (Drawings no. 070-02-12 and 070-02-13).

As part of the work to build up a reliable geological model for the area, it has been of significant importance to get detailed information on the exact stratigraphic position of the Palaeogene clays from boring to boring. As this cannot be gained solely from visual description, it was decided to perform micropalaeontologic investigation of a significant number of samples to support the geological descriptions.

In several borings it was observed that ash layers are (steeply) dipping. Furthermore, layers from some of the Palaeogene formations are apparently extremely thick compared to observations at other locations with Palaeogene deposits in Denmark. As an example, boring 09.A.002 passed through more than 80 m of Røsnæs clay even though 30 m is the maximum thickness previous encountered in Denmark. This observation indicates that the boring has not passed through the deposit in a direction perpendicular to the bedding.

Moreover, there are situations where two neighbouring borings from top to bottom of the Palaeogene clay layer have been drilled in different formations (boring 09.A.003 Røsnæs Formation and 09.A.010 Ølst Formation, the borings being 600 m apart), and in boring 09.A.015 and 09.A.015A older layers are situated above younger layers. All those observations indicate that the uppermost part of the Palaeogene clay has been squeezed into giant, sometimes over-lapped anticlines by glacier pressure in the Quaternary period.

In almost all the borings that have been drilled deep into the Palaeogene clay, a marked increase in the CPTU cone resistance is seen at the approximate elevation -70 to -80 m. Dating of selected samples has shown that the level of the sudden increase in cone resistance is not related to a specific stratigraphic horizon/boundary but instead is located in different stratigraphic levels from boring to boring. Based on these observations, the boundary is interpreted as the sliding surface at the base for the glacial deformation of the upper part of the Palaeogene layers; i.e. the boundary between folded and intact clay.

All of the Palaeogene deposits encountered, irrespective of the formation assigned to them, are very high plasticity clays. The clay content is typically 70-80 %, but may be less than 60 % in the Ølst Formation. The plasticity index is typically 50-140 % but slightly higher for the Ølst Formation, c. 90-150 %. Fissures abound, commonly associated with slickensides.

It is noted that the term “Tarras” often used in older German geological literature is assumed to be synonymous with “Palaeogene clay” and that it further is assumed that “Grüner Tarras” might be synonymous with Lillebælt clay and “Roter Tarras” with Røs- næs clay.

The distribution of the Palaeogene deposits below the area appears from the longitudinal profile in Enclosure I (Drawings no. 070-02-012 and 070-02-13). The result of the micropalaeontologic dating is described in /6/.

The Lillebælt clay is most often greenish or greenish grey in colour but can include layers with a more reddish tint. It is in the Fehmarnbelt area normally non-calcareous or sometimes slightly calcareous. It contains a few thin ash layers and also concretions orientated in bands parallel with the bedding planes are seen. The known thickness of the layer is 40 m in central Jylland (Viborg and Ølst), 70 m at the new Lillebælt Bridge and 100 m in the southernmost part of Denmark.

In Røsnæs clay red and olive brown colours dominate but other shades of brown, olive grey and patches of grey also appear. Due to a significant content of micro-shells, the clay almost always has some calcareous content (typical contents of CaCO3 is 2-4 %). Most of the layer has no organic content at all, but two thin bands with a characteristic black colour and rich in organic matter are present in the uppermost part of the unit. The clay contains abundant nodules or concretions (siderite and others), and silty lilac or dark green ash layers appear. The Røsnæs clay is in a boring located at Odder in Jylland 28 m thick, which is considered to be the largest thickness measured in Denmark.

Clay from the Ølst Formation is almost black, dark grey or very dark grey/olive grey, very high plasticity clay. It is almost always non-calcareous; layers of volcanic ash are common. Burrows are observed and concretions of carbonate and pyrite occur. In boring 96.0.009 a 0.3 m thick carbonate cemented bed is observed at level -91 m. The thickness of the Ølst clay layer is typically 10-15 m in the Danish area.

Holmehus clay is very high plasticity clay, normally with clear bluish, reddish or, most commonly, greenish colour. It is non-calcareous and contains no ash layers. It is often heavily bioturbated. The thickness of the unit varies from approximately 3 m at the location of Ølst to 40 m in the Odder-area south of Aarhus.

The Æbelø Formation consists of light grey to grey, non-calcareous, silty to very silty clay or even silt. It contains many silicified layers. The thickness varies between 15 and 60 m, and it is often covered by rather thick transition layers to the overlying Holmehus Formation.

It must be noted that the floes of Palaeogene clay within the Glacial deposits have been described under the geological description for Glacial deposits. The geotechnical properties of these floes are included in the following Chapters.

13.2 Geotechnical Properties 13.2.1 General The geotechnical properties of these deposits have been investigated through:  Classification testing on samples from the A-boreholes as identified in Chapter 6.  In-situ testing (CPTU) in the B-boreholes as identified in Chapter 6.  Advanced geotechnical laboratory testing by GEO of selected samples from the A- boreholes (/29/).  DSS testing by NGI to study secondary consolidation for a combined stress state with axial and shear stresses, /62/

The overall behaviour of the clays of Palaeogene origin may be characterised by the following:  The content of clay minerals from the smectite group in the clays of Palaeogene origin implies that the soil expands when unloaded and tends to absorb water.  As the soil volume increases during unloading (caused by e.g. melting ice), the void ratio increases and the apparent pre-consolidation pressure is reduced. This is seen on reloading stress paths during oedometer testing where ´pc is identified for an axial stress significantly smaller than the maximum axial stress from a former load step within the same test.  The soil therefore has a moderate memory but its behaviour in unloading shows a high degree of knowledge as repeated unloading-reloading loops are more or less identical in a stress strain plot. This behaviour is also described in e.g. /58/.  Following these tendencies, the in-situ state of the Palaeogene clay reflects an unloaded soil, that, when reloaded, behaves as a moderately over-consolidated clay with an apparent over-consolidation ratio of approximately two to five. When the soil is unloaded in 1D from the in-situ state, it will reach a new state when fully swelled.

Figure 13-1 Plasticity chart for clays of Palaeogene origin

Main findings documenting the statements above are given in the following subsections comprising field testing, laboratory testing and large scale testing. More details of the geotechnical properties for clays of Palaeogene origin can be found in Appendix GDR 00.1-001-D and Appendix GDR 00.1-001-G.

13.2.2 Classification Properties The plasticity chart for the clays of Palaeogene origin is shown in Figure 13-1 from which it is seen that the major part of the clays belongs to clay of very high plasticity. An overview of the geotechnical classification properties for clays of Palaeogene origin appears from the Tables 13-1 and 13-2 for the intact clays and the folded/floe clay, respectively.

Table 13-1 Classification properties for intact Palaeogene Clay

Røsnæs Ølst Holmehus

 32.5 % 41.6 % 32.5 % Water content, w  2.7 % 6.9 % 0.9 % N 254 72 4  18.9 kN/m3 18.2 kN/m3 19.7 kN/m3 Total saturated unit weight,   0.9 kN/m3 0.9 kN/m3 2.4 kN/m3 N 277 69 4  31.7 % 40.6 % 30.5 %

Plastic limit, wP  3.3 % 8.0 % 0.7 % N 72 18 2  114 % 154 % 117 %

Liquid limit, wL  22.2 % 20.4 % 0.7 % N 72 18 2  82.0 % 113 % 86.0 %

Plasticity index, IP  21.0 % 24.2 % 1.4 % N 72 18 2  Arithmetical mean value  Standard deviation N: Number of data

Table 13-2 Classification properties for the Palaeogene clays from the different formations when they appear as folded strata or in dislocated floes

Røsnæs Ølst Holmehus

 36.5 % 43.7 % 37.4*/39.8 %

Water contents, w  4.6 % 8.6 % 5.4*/6.8 % N 1151 153 107*/89

 18.5 kN/m3 18.2 kN/m3 18.6*/18.1 kN/m3

Total saturated unit weight,   0.8 kN/m3 0.9 kN/m3 0.9*/1.1 kN/m3 N 1167 130 86*/73

 31.7 % 39.5 % 35.3*/37.7 %

Plastic limit, wP  4.9 % 6.9 % 5.0*/8.4 % N 335 53 22*/26

 136 % 156 % 151*/148 %

Liquid limit, wL  32.0 % 28.3 % 21.8*/34.6 % N 335 53 22*/26

 104 % 116 % 116*/110 %

Plasticity index, IP  31.1 % 28.6 % 21.6*/31.6 % N 335 53 22*/26 and N: See footnote for Table 13-1 *: Numbers valid for borehole 10.A.058 only. The main relevant occurrence of Holmehus Clay is found within borehole 10.A.058, and classification properties from this borehole are thus presented separately.

13.2.3 CPTU A considerable number of the B-borings have penetrated the clays of Palaeogene origin. Considering the net cone resistance, the following tendencies are found:  The net cone resistance increases with depth for the folded Røsnæs deposit, starting at 0.3-0.5 MPa at seabed level and increases to a value not exceeding 4.0 MPa at approximately 30 m depth.  A similar trend is seen for the folded Ølst deposit, but the ash layers imply a scatter in the net cone resistance.  For the Holmehus Formation as found in borehole 10.B.058 the scatter in cone resistance is also seen but the net cone resistance is generally 20-30 % higher than found in the corresponding depth of the Røsnæs Clay Formation.

The net cone resistance has been used together with the undrained shear strength in com- C pression from triaxial testing to establish the cone factor Nkt = qnet/cu , see /16/.

The normalised cone resistance is defined by Qt = qnet/′v0 where ′v0 is the estimated vertical effective in-situ stress. Assuming that the cone factor, Nkt is constant within one clay unit, e.g. folded Røsnæs clay, the normalised cone resistance may be expressed by Qt C = Nktcu /′v0 and Qt therefore links directly to the SHANSEP relation through the equation:

B C Qt = NktAOCR = Nktcu /′v0 where “A” and “B” are constants for a site-specific soil type.

The following tendencies are found for the folded and the intact Røsnæs formation and are illustrated in Figure 13-2:  For the folded Røsnæs clay Qt decreases with depth, provided that the formation is NOT overlain by a different deposit, e.g. clay till.  If the folded Røsnæs formation is overlain by a different deposit (with a thickness exceeding say 5 to 10 m), Qt is more or less constant with depth, meaning that OCR is approximately constant.  For the intact clay, Qt is generally constant but the value may vary slightly from borehole to borehole.

Figure 13-2 Normalised cone resistance, Qt, versus depth for boreholes 10.B.051 and 10.B.054 where: 10.B.051: Folded Røsnæs clay from seabed and 10.B.054: Glacial and younger deposits located between the seabed at elevation -18 m and elevation -47 m

CPTUs were performed within the large scale trial excavation area (folded Røsnæs clay) with the normalised cone resistance being depicted in Figure 13-3:  The base of the excavation is located at approximately elevation -20 m or corresponding to 10 m below the in-situ seabed.

 The blue CPTU traces in Figure 13-3 are performed from base of excavation during June 2011, approximately two weeks after conclusion of the Phase 3 excavation.  The red CPTU traces are performed in April 2013 from the top of the Phase 4 infilled sand layer, placed in the bottom of the excavation. Only the part of the traces penetrating Palaeogene Clay has been included.  The shaded area "Range of all in-situ CPTUs in Palaeogene Clay" represents the measured approximate range of the mean values of CPTUs performed within the Fehmarnbelt area.

Figure 13-3 CPTU net cone resistance against depth below base of trial excavation

The blue traces show a reduction in strength (softening) within the upper 0.5 to 1.0 m as a result of the excavation. The red traces reveal that this effect has penetrated a further 0.75 m deeper within the time period between June 2011 and April 2013.

The red traces of qnet below approximately 2 m depth imply slightly higher qnet values compared to the blue traces. This is most likely due to challenges with maintaining saturated filters in the cone when penetrating the infilled sand but it may also reflect local variations in the ground conditions.

When the full effect of strength softening is reached for a one-dimensional case (no effects from e.g. slopes included), the corresponding qnet values expected are indicated by the shaded area marked with "Range of CPTUs at end of primary consolidation" being representative of a fully swelled condition.

The data compilation and data processing for the CPTUs have been described in /60/.

13.2.4 Stress and Stress History In the Quaternary period it is estimated that the glaciers of the Saale glaciations were approximately 1000 m thick above the Fehmarbelt area. This ice sheet will have pre- loaded the Palaeogene clay but the effective preload is difficult to determine as it is affected by the depth of frozen ground beneath the glacier, the basal water pressure and flow patterns.

Pre-consolidation pressures determined from laboratory testing are significantly lower than might be reasonably expected from consolidation under loading from the glacier. This phenomenon has been observed previously /58/ from testing of Lillebælt clay and therefore appears to be a characteristic of high plasticity Palaeogene clay. Because of these differences in values of pre-consolidation pressure it is important that in any interpretation that necessitates the use of 'pc clearly addresses the background of the value used.

The pre-consolidation pressure can be estimated from oedometer tests, but with the rate of secondary compression increasing with increasing stress, the virgin compression curve in a semi-logarithmic plot will generally not be a straight line. An uncritical use of standard methods for estimating ´pc may therefore over-estimate 'pc. The method of Janbu is the preferred method as this method does not rely heavily on the stress strain relationship at high effective stresses, /21/.

The conclusion on folded and intact Røsnæs clay is that the mean value of the pre-consolidation pressure derived from oedometer testing can be estimated by:

1.287 ′pc = ′v0[0.183qnet / ′v0]

Laboratory test data from Holmehus, Ølst and floes of the different formations show that ´pc for these deposits can be estimated using the same equation. The equation above can be approximated by ´pc = 0.25qnet.

Some soils exhibit a clear and well-defined “soil skeleton collapse” when ´pc is exceeded. A less distinct “soil-skeleton-collapse” mechanism is seen during Oedometer testing on Palaeogene clay combined with variations in the estimated ′pc from tests having been performed on the same core suggests that the cores are disturbed. Studying the CPTU measurements along different core runs, however, reveals an unexpectedly high variation in the net cone resistance, within depth intervals as little as 0.05 m. This finding supports the variation in laboratory test results within one and the same core. In addition, sample disturbance of the individual specimens have been evaluated using the NGI method. From this evaluation it appears that the majority of specimens tested are “very good to excellent”.

Based on the CPTU-evaluation and the investigation of sample disturbance of laboratory specimens, it is indicated that the material within the cores are well preserved and of quality comparable to the in-situ soil. It is, however, believed that the folded parts of the clay of Palaeogene origin do not behave as a “typical over-consolidated soil”. The folded parts carry their own degree of geological disturbance due to their folded nature and as such the geological history adds an additional consideration within the discussion on sample disturbance.

It may be expected that the soil encountered in the A-boreholes does not necessarily have the same properties as the soil in the corresponding depth of the neighbouring B-borehole (typical spacing 5 m apart). This is supported by the observation that ash layers within the Palaeogene Clay are inclined at up to 60 on some of the borehole profiles. Despite these concerns correlation between laboratory test results and cone resistance has been performed using the same depth in the A and B boreholes; it is expected that this approach will inevitably result in some scatter in the derived correlations.

Figure 13-4 shows a CRS test running to a maximum vertical stress of 7,500 kPa (in the range of the geological surcharge pressure under ice loading. Following the end of primary consolidation (EOP) the specimen is unloaded to approximately 200 kPa (EOP swelling has been reached).

Figure 13-4 CRS tests from borehole 09.A.002, 22.39 m depth (folded Røsnæs clay), where the pre-consolidation pressure is extracted along the reloading curve after the specimen has been consolidated for a vertical stress of 7,500 kPa

The figure illustrates that if the Casagrande approach is used to extract ´pc along the reloading curve AB, the estimated value will be approximately 725 kPa or 10 % of the maximum axial effective stress experienced by the specimen in the laboratory. At the stress of 725 kPa there is a marked change in stiffness (reduction in stiffness) during the reloading. In /57/ this stress has been called the “apparent pre-consolidation pressure” but within this document is deemed the pre-consolidation pressure (´pc) unless otherwise stated. It is remarkable that the clay, as per Figure 13-4, loses much of its memory of the maximum pressure applied within the same test. A similar ´pc was found along the reloading curve using the methods of Janbu and Becker.

Based on incremental loading oedometer tests where the lateral stress is measured, an estimate of the coefficient of earth pressure at rest, K0 is given during unloading, assuming that the specimen has been loaded to a normally consolidate stress state (K0  0.55). The estimate is an approximate mean value, coupling K0 with OCR, and where OCR is defined as the maximum axial effective stress applied within the test divided by the vertical effective stress at the relevant unloading step:

0.515 K0 = 0.548[OCR]

The principles addressed above are based on oedometer testing. Extracting the normalised cone resistance from the CPTUs at each location of the performed oedometer tests C implies that a normalised undrained shear strength cu /´v0 may be attached to each oedometer test, provided that a realistic cone factor is known. A SHANSEP relation has therefore been established from the oedometer tests directly and this relation fits well with the SHANSEP relation, established by running undrained triaxial tests in a way, where the individual tests are sheared from a known over-consolidation ratio. OCR is defined using ´pc as found in oedometer testing and also from the relationship ´pc  0.25 qnet.

Finally, the undrained triaxial tests consolidated for an effective in-situ stress state leads to a normalised shear strength that fits well with the normalised cone resistance, so the data are consistent.

13.2.5 Consolidation Properties

The coefficient of consolidation, ck is defined by:

ck = Eoedk / w where Eoed is the constrained oedometer modulus, k is the coefficient of permeability and w is the unit weight of water.

The measured heave within the trial excavation was back-calculated in /56/. The analyses showed that a reasonable fit could be obtained using the constrained tangential stiffness from oedometer testing with a coefficient of consolidation being approximately 100 times higher than the value measured in the laboratory. A recommended field value for k is 1.710-10 m/s = 5.410-3 m/year. A similar observation is also made in Appendix F, where back-calculations are performed for the old Lillebælt Bridge, using measured values of time versus settlement; the field value of the permeability is approximately 100 times higher than the corresponding laboratory value.

Appendix D shows the correlation established between the secant value of the constrained oedometer modulus (loading), Eoed,sec and the CPTU net cone resistance valid for the floes and folded parts of the Palaeogene clays. This correlation can be used to estimate the primary consolidations settlements using a load spread of 1 horizontally and 2 verti- cally. Approximate values of Eoed,sec ranges from 10 to 60 MPa.

The rate of secondary compression, s amounts to approximately 0.1-0.5 % per log10 cycle of time (%/lct) when measured in the oedometer cell. The relation between s and the ratio 'v/'v0 is approximately linear (arithmetical axis), where 'v is the vertical effective stress at which s is extracted.

Primary and secondary consolidation settlements have been estimated for existing structures where the total settlement is known; see Appendix F. It is concluded that the correlations above lead to realistic estimates of the total measured settlement for structures placed on a horizontal surface of Palaeogene clay. If the structure is placed on a slope with an inclination of 5 to 10 (or more) relative to lateral, the estimated secondary consolidation settlement may be significantly underestimated. This observation is most likely due to the development of shear strains within the soil volume. The effect of shear strains on the consolidation properties of the Palaeogene clay was investigated in /62/, with a summary given in Appendix D. The summary in Appendix D includes recommendations that account for the underestimated secondary consolidation settlements.

Fill placed on Palaeogene clay will usually involve a slope at the extremities of the fill area. Laboratory testing in /62/ and the eastern breakwater in Puttgarden Harbour indicates that this slope should not have an inclination with horizontal,  steeper than tan ≤ 1/3. A more detailed discussion is given in Appendix D.

The range for the mean value of the compression ratio, Q for stress levels in excess of 2,500 kPa is 10-20 %.

During unloading of the soil, the constrained tangential oedometer modulus has been addressed. When the specimens are sufficiently unloaded, the stress-strain relationship maps into the swelling ratio, Qun resembling the compression ratio, but with an inclination of 4-8 %/lct. The tangential value of the constrained oedometer modulus is thus Eoed,tan = ´vloge(10)/Qun where ´v is the vertical effective stress. Supplementary oedometer testing was initiated to clarify whether the percentage of Smectite group clay minerals would influence the value of Qun, e.g. that Qun for Holmehus clay should be higher than for Røsnæs clay. Based on these tests it was concluded that one value of Qun can be used for the clays in the formations Røsnæs, Ølst and Holmehus.

The rate of secondary swell is dependent on the unloading stress relative to the estimated vertical effective in-situ stress; as the stress on the specimen decreases, the rate of secondary swell increases. Apparently the rate of secondary swell may exceed the rate of secondary compression by a factor of 2-4.

13.2.6 Static Shear Strength 13.2.6.1 Introduction The undrained shear strength and the effective shear strength properties have been addressed.

13.2.6.2 Undrained shear strength The undrained shear strength can be determined using correlations or measurements directly. Two correlations have been established: C  A SHANSEP correlation, where OCR, ´v and cu are linked. C  A CPTU correlation, cu = qnet / Nkt with Nkt = 25.5.

The SHANSEP correlation is defined by the normalised undrained shear strength versus the laboratory controlled over-consolidation ratio OCR:

C 0.78 cu /σ′v = 0.21[OCR]

For a loading scenario, OCR is estimated using the pre-consolidation pressure before the actual load is applied. For an unloading scenario the applied pre-consolidation pressure must be the reduced value after the effect of unloading has taken place. In both cases the correlation with ´pc = 0.25qnet can be used, provided that the value of qnet is representative for the loading case considered.

For an unloading case, a representative value of qnet may be extracted from representative CPTUs using a realistic stress level. As an example, the undrained shear strength at the bottom of an excavation where a 25 kPa vertical surcharge is placed may be found using a representative qnet value from a depth reflecting a vertical effective in-situ stress of 25 kPa, which typically will be c. 3 m depth in a similar deposit. This approach will lead to a similar undrained shear strength as found using Nkt = 25.5 on the representative qnet value directly.

In this way the two correlations (SHANSEP and CPTU) identify one and the same undrained shear strength. The approach has been documented in Appendix G when interpreting the four sliding tests where a skirted foundation is first consolidated for a vertical load and then pulled laterally until failure.

It must be emphasised that the suggested procedure with translating the CPTU measurements upwards and downwards depending on the representative stress level is valid as long as the CPTU measurements are representative (e.g. do not cross a boundary between folded and intact clay). CPTUs performed from the in-situ seabed represent a response for a soil having been allowed to swell freely for many years. For an excavation the effective stress corresponding to end of primary consolidation may be divided into three zones:

 Shallow zone where a representative stress state resembles what can be extracted from CPTUs, when the stress level is accounted for.  Deep zone, where no effects of the excavation is seen.  Transition zone where neither the shallow zone nor the deep zone are representative.

Depending on what is critical for the activities in question, the transition zone may be dealt with as either the shallow zone or the deep zone.

Anisotropy ratios have been investigated in the laboratory and it appears that the undrained shear strengths in compression, in extension and in direct simple shear are of the same average magnitude. This observation has also been confirmed by four vertical plate load test performed on 1.0 m diameter circular plates. The measured average resistance (233 kPa) was compared with an axi-symmetric Plaxis analysis using an undrained shear strength profile established using the CPTUs performed within the trial excavation and Nkt = 25.5. The computed resistance was 224 kPa using anisotropy ratios of 1.00.

Furthermore, the same shear strength profile was used to model the passive plate load test (two plates pushed away from each other with each plate being approximately 2 m high and 6 m long). The estimated passive force was 2,185 kN whereas the maximum value of the measured force was 1,655 kN.

Possible causes of the difference between the calculated and the measured failure loads are the additional softening of the clay between the time of the CPTU test and the plate load test, small uncertainties in level, variations in the CPTU profile and the possible range of Nkt value of which 25.5 is the best estimate. At the limit the best estimate values would suggest a scale effect reduction on undrained strength of 24 % when moving laboratory size specimens (0.004 m2) to a loaded area of 12 m2. However, such a reduction does not appear fully justified given the nature of the clay (laboratory strength is dominated by pre-existing shear planes) and the average results obtained from the vertical plate load tests with area of 0.8 m2.

Correlating CPTU measurements with the available undrained shear strength tests (CAUc, CAUe and DSS) implies that the undrained in-situ shear strength can be estimated using Nkt = 25.5 (geometrical mean value of 114 tests). The distribution is illustrated in Figure 13-5. These results cover Lillebælt, Røsnæs, Ølst and Holmehus formations, whereas the one measurement in Æbelø is excluded.

Comparing the estimated average Nkt for the different formations will show variations, but the magnitude of the variation is more or less directly linked to the number of tests performed within the relevant formations and it is therefore concluded that Nkt = 25.5 is applicable for the clays of Lillebælt, Røsnæs, Ølst and Holmehus formations.

Figure 13-5 The number of tests against the estimated cone factor. Floes, folds and intact deposits have been included. The figure includes the undrained in-situ shear strength as measured using CAUc, CAUe and DSS

The only exception is that the arithmetical mean value of Nkt, DSS (23.9 for 31 numbers) in the folded Røsnæs formation is approximately 10 % lower than the arithmetical mean value of Nkt, CAUc (26.1 for 45 numbers). This difference is believed to represent a strain rate in direct simple shear exceeding the strain rate in undrained triaxial testing by a factor of 20. It is therefore concluded that:  The Palaeogene clay soil exhibits a positive rate effect being 7 % shear strength increase per log10 cycle strain rate.  The undrained direct simple shear strength, as measured in this test program, shall be multiplied by 0.916 to comply with the undrained triaxial testing performed.

13.2.6.3 Effective shear strength properties Effective shear strength properties have been extracted using CADc tests and CAUc tests. When doing a detailed check of the test enclosures for CAD-testing within e.g. folded Røsnæs Clay, some tests keep dilating while others reveal a peak shear strength and some tests even show constant volume along the final part of the shearing phase (critical state). There appears to be no systematic trend in the results obtained. The mean value of the effective shear strength properties have thus been estimated by considering all CAD and CAUc tests allowing for an evaluation of effective shear strength properties as one po- pulation. The arithmetical mean values of the triaxial peak friction angle and the effective cohesion are 19.6° and 14 kPa.

13.2.7 Geophysical Properties In borehole 09.A.701 there are three different Palaeogene clay deposits present. Of those the Røsnæs clay has a high and increasing gamma radiation reading with depth. The transition to the underlying Ølst clay is marked by a natural gamma radiation peak and thereafter a saucer like depression trend. The transition from the Ølst clay to the

Holmehus clay is more subtle and is mainly seen as an increase in natural gamma radiation values.

From the log trends it seems possible to distinguish between intervals of intact clays and folded successions. An example is seen in 09.A.002, where the interval below 60 m depth from the B-boring is evaluated to be intact. The folded intervals seem to stand out with more fluctuating natural gamma radiation, induction conductivity and gamma density log responses; the explanation for this is unknown.

The Palaeogene clay unit has been divided into the Lillebælt, Røsnæs, Ølst, Holmehus and Æbelø Formations. The interpreted interval velocities and calculated physical soil properties from the VSP measurements have been correlated with the three Palaeogene clay deposits. The results are listed in Table 13-3. The change in measured velocities from the VSP does not always match the stratigraphy. This could be due to the limited precision of the VSP method and the difficulties with interpreting velocity intervals in grading lithologies.

Table 13-3 Derived Sonic and VSP interval velocities correlated with geology and age

Sonic P-VSP S-VSP Geology Int. Vel. [m/s] Int. Vel. [m/s] Int. Vel. [m/s] Range Average Range Average Range Average Røsnæs Fm 1490- 1575 1675- 2025 1850 250-450 300 1675 Ølst Fm Palaeogene 1570 1570 1700 1700 350 350 clay 09.A.701 Holmehus 1560- 09.A.701, 1580 1600 1600 250-350 300 1600 09.A.703

The small strain shear modulus (G0), Poisson’s ratio and Young’s modulus have been calculated using the results from the VSP measurements and the density log. The parameters have been calculated for the three Palaeogene clay formations; Røsnæs, Ølst and Holmehus Formations. The Røsnæs clay has been investigated in 4 borings, one on Lolland and three on Fehmarn. The high velocities around 2000 m/s are only measured, where the boring penetrates less than 6 m of the Palaeogene clay and where the overlying deposit is the high velocity till. According to these observations it should be noted that the measured velocity could be a little too high. It should also be noted that the velocities measured with the Sonic-log are a bit lower than the measurements from the VSP.

Table 13-4 Calculated parameters in the Palaeogene clay by VSP

G0 Poisson’s ratio Young’s modulus Geology [MPa] [-] [MPa] Range Average Range Average Range Average Røsnæs Fm 120-483 317 0.466-0.492 0.472 359-1423 932

Palaeogene Ølst 09.A.701 157-224 184 0.479 0.479 465-664 545 clay Holmehus 09.A.701, 90-155 120 0.483-0.492 0.486 265-455 360 09.A.703

It must be observed that the geophysical properties for the floes of clays of Palaeogene origin in the Glacial deposits have been included in the descriptions of the geophysical properties for Glacial deposits.

13.2.8 Small Strain Stiffness and Damping The small strain shear modulus has been measured in laboratory using resonant column testing and by using bender elements installed in triaxial and resonant column test specimens. Bender elements have not provided reliable estimates of shear wave velocity in the 1:1 height to diameter test specimens used for CAU triaxial testing. Bender element measurements made on 2:1 height to diameter specimens used for cyclic triaxial and resonant column testing appear more reliable and provide estimates of shear modulus that are consistent with values derived from resonant column testing. Bender elements in 2:1 height to diameter specimens also provide shear wave velocity values consistent with those obtained from geophysical downhole logging using Vertical Seismic Profiling within boreholes through the same geological deposit as that from which the specimens for testing were obtained.

The test results show that to a depth of about 30 m below ground level the small strain shear modulus (Gmax) within the Palaeogene clay increases approximately linearly with depth to a value of about 100 MPa at 30 m depth. The downhole logging indicates that below 30 m G0 appears to be essentially constant with depth at values ranging up to 400 MPa. Gmax increases if the consolidation stresses increase, but the increase is relatively modest. Increasing the vertical effective stress by 50 %, for example, results in an approximate 10 % increase in Gmax.

At depths less than c. 30 m below ground level the test data indicate a ratio of Gmax to net cone resistance, qnet, in the range 20 to 40. Downhole logging indicates G0/qnet ratio at greater depth in the range 50 to 200. Under cyclic loading conditions the same non- dimensional stiffness degradation curve relating G/Gmax to cyclic shear strain can be used for all Palaeogene clay materials.

13.2.9 Cyclic Undrained Shear Strength The folded Røsnæs and Ølst clays apparently have a cyclic shear strength higher than the static shear strength using the direct simple shear strength derived from qnet with Nkt = 25.5 as a reference, provided that the equivalent number of load cycles do not exceed 10.

The Holmehus clay deviates in this respect, apparently being significantly weaker. Cyclic loading will imply a shear strength degradation. The cyclic shear strength is 40 % lower than the static shear strength, provided that the equivalent number of load cycles does not exceed 10.

It must finally be emphasised that Deltares has performed the cyclic DSS tests for the Røsnæs and Ølst clays using active height control, whereas the tests on the Holmehus clay performed by Fugro/Houston used passive height control. 13.3 Special Investigations and Evaluations Special investigations and evaluations have been performed for clays of Palaeogene origin at the old Lillebælt Bridge, at the Fehmarnsund Bridge and at Puttgarden Breakwater. The ground conditions at these sites are comparable to the ground conditions in part of Fehmarnbelt Fixed Link alignment. Hence the borings performed on these sites serve as correlation boreholes for sites with similar ground conditions as in parts of Fehmarnbelt fixed link alignment and deliver information on documented behaviour during full scale and long term loadings on the actual deposits.

Background, evaluations and details of the geotechnical properties for clays of Palaeoge- ne origin at Lillebælt, Fehmarnsund and Puttgarden can be found in Appendix GDR 00.1- 001-F.

The geotechnical properties for the clays at these sites have been investigated through:  Classification testing by Fugro of samples from the type A-borings in Lillebælt (/25/).  In-situ testing (CPTU) by Fugro in the type B-borings in Lillebælt /25/).  Classification testing and field vane testing by GEO in the type A-boring below Pier 3 of the old Lillebælt Bridge (Appendix GDR 00.1-001-F).  Classification testing by GEO of samples from the type A-boring at Fehmarnsund (/34/).  In-situ testing (CPTU) by GEO in the type B-boring at Fehmarnsund (/34/).  Classification testing by GEO of selected samples from the type A-borings in Lillebælt and at Fehmarnsund for Advanced Laboratory Testing (/31/).  Advanced geotechnical testing by GEO of selected samples extracted from the type A-borings in Lillebælt and at Fehmarnsund, cf. /31/.

An important objective with the investigation in Lillebælt was to evaluate whether or not the bridge loading has achieved fully drained conditions in the clays of Palaeogene origin below Pier 3 or whether it is still in the primary consolidation phase. The existing bridge piers have since 1932 undergone considerable settlements from 0.3 m for Pier 1 to approximately 0.7 m for Pier 3, /41/.

Similarly, an important objective for the investigation at Fehmarnsund was to evaluate whether or not the loading from the northern embankment has achieved fully drained conditions in the clays of Palaeogene origin below or whether it is still in the primary consolidation phase. The observed total settlement from 1961 to 1971 of the northern embankment is around 0.8 m. The settlements within the last 40 years have not been moni- tored (cf . Appendix GDR 00.1-001-F and /41/). The model for estimating settlements, to be applied for the Fehmarnbelt Fixed Link, will be influenced by the mapped soil response at the two sites.

The overall objective with the investigation at Puttgarden was to evaluate the settlement model established by correlating oedoemeter testing with CPTUs.

In conclusion the test results indicate that:  The effective stresses below Pier 3 of the old Lillebælt Bridge and below the northern embankment at the Fehmarnsund Bridge are increased due to the load from the structures.  The measured settlement of the pier head at Puttgarden Harbour is 340 mm whereas the estimated settlement is approximately 320 mm.

Another special investigation of the clays of Palaeogene origin included laboratory mixing, cement stabilisation and unconfined compression testing of this material (/32/).

Additionally a special investigation with pump test with cuttings of Palaeogene clay have been performed (/64/). 13.4 Large Scale Properties The Geotechnical Large Scale Testing Area (GLSTA) is located in the shallow waters off the Fehmarn coast. In addition, a smaller area for ground anchor testing was established east of Rødbyhavn. The works commenced in year 2010 and have been completed as of December 2014. The scope of the GLSTA works is described in Chapter 9; the results are presented in more detail in Appendix GDR 00.1-001-G. Detailed results of the testing are presented in /49/ through /53/.

This section of the report summarises the results and findings of the GLSTA, namely:  Heave and porewater pressure measurements at and below the seabed in the trial excavation compared to predictions of heave based on soil models.  Bearing resistance of plate load tests (vertical, horizontal shear and passive failure modes) and comparison with other measurements of strength of the clay.  Pull out resistance of bored piles and driven piles; and  Pull out and creep resistance of ground anchors (located onshore in Lolland). 13.4.1 Instrumented Trial Excavation As presented in /49/ and in Appendix GDR 00.1-001-G the results of the trial excavation are in terms of extensometer heave data and porewater pressure response.

The results of the observations and measurements in the trial excavation have been used alongside the result from the advanced laboratory testing to develop soil models which can approximately account for the behaviour of the Palaeogene Clay at an engineering scale. The selected strategy has been to use the laboratory test data to develop stiffness and strength models of the clay and then to calibrate these against the results of the measurements in the trial excavation. This process is not without difficulty due to some apparent inconsistencies in the data; the conclusions reached apply only to the folded Røsnæs Clay which is present at the trial excavation location.

In Appendix GDR 00.1-001-G and /56/ four different soil models are presented; these are:  A model based solely on voids ratio variations with depth. This is a simple model but illustrates the order of magnitude of heave that could be expected below excavations or settlement for loading situations;  A model linking CPTU qt data and voids ratio to provide an indication of heave linked to a reduction in soil strength, again this is a simple model (/56/).  A 1D finite difference model (called NUMHEAVE) based on the advanced laboratory test data. This model is a heave model.  A 2D (or 3D) finite element model using the BRICK soil model based on the advanced laboratory test data, this model includes strength and stiffness calculations.

In the following paragraphs the base data from the trial excavation is presented along with a summary of the soil model calculations for heave.

Figure 13-6 Results from the trial excavation instrumentation (time in months after start of monitoring data collection)

A selection of the results from the three central installations (extenso-piezometers Nos 4, 5 and 6) is shown in Figure 13-6. The extensometers measure “heave” or “extension” between the base of the excavation and the instrumentation tip. The two deep extensometers (E6 at 9 m depth and E5 at 25 m depth) show similar movements that are greater than the shallow extensometer at 3 m depth (E4). This suggests that during the monitoring period no heave has occurred below 9 m depth; this is seen to be reasonable when considering the apparent porewater dissipation at piezometer P4 (3 m depth) following the equilibration stage and that no such dissipation can be seen on either the piezometers at 9 and 25 m depth (there is much noise in the data). The figure also show changes in the excess measured porewater pressure and incremental settlement when the Phase 4 and Phase 5 backfilling operations were carried out. It is noted that the resulting settlement from these backfilling events suggests that the soil below the bases of the trial excavation has experienced an cumulative settlement due to the works – this is not considered correct and must be considered in context of missing data between the initiation of the Phase 1 excavation and the turning on of the instrumentation prior to the Phase 2 excavation and possible disturbance of the instruments during the filling works. In Figure 13-6 the missing data is illustratively shown by dotted lines.

Figure 13-7 shows the results of the BRICK soil model predictions for the data shown in Figure 13-6. On this figure the calculated extension from the BRICK soil model have been set to zero on the day the first extensometer measurements were taken (30/09/2010), with movements prior to this date shown as negative data. It can be seen that the fit between the measured and predicted displacements is not perfect. The prediction suggests that the two shallow extensometers extend a similar amount but that the deeper instrument extends more; this is as a result of the shear induced movement associated with the extension of the excavation from the Phase 1 to Phase 2 geometries. Predicted porewater pressures (see appendix GDR 00.1-001-G) are based on the assumption that there is minimal porewater pressure dissipation during the 10 month period post excavation within the depths of the piezometers; this is in keeping with the deeper piezometer data in Figure 13-6 but slightly different for the shallow piezometer (P4 and 3 m depth) where there is an initial period of negative excess porewater pressure dissipation which then slows down prior to the Phase 3 excavation taking place. Accepting that the soil models developed from the advanced laboratory testing are reasonable, the permeability of the soil has been assessed by the matching of the rate of predicted heave with measured heave as per Figure 13-7. This procedure resulted in the soil mass permeability being close to two orders of magnitude greater than the laboratory permeability (field permeability k = 1.710-10 m/s compared to 510-12 m/s from the laboratory).

The magnitude of the movements that are predicted prior to the start of monitoring data collection is between 20 and 40 mm depending on depth. These movements include primary heave (likely to be the same for all instruments as the swelling depth is small over this short time period) and shear induced heave which will be larger at depth than in the shallow soil. These movements can be compared to approximately 30mm of heave assessed using the CPTU data (/56/) for the shallow soil (i.e. clay covered by instrument E4 which will not include significant shear induced heave). Taking these observations together it is reasonable to surmise that rapid softening and heave occurs immediately

after a formation in Palaeogene Clay is exposed under water and that this heave and local softening will result in settlement when the formation is reloaded.

Figure 13-7 Predicted versus measured extensometer heave (example from Brick model analyses)

Taking the data obtained from the back analysis of the trial excavation a prediction of total heave (long-term) has been made. The results are shown in Figure 13-8 for the 1D NUMHEAVE and 2D BRICK models presented in /56/. The different models show a marked similarity in computed results with the BRICK model providing slightly higher predicted heave than the NUMHEAVE model. Secondary heave movements shall be added to the primary heave to give the total heave.

Figure 13-8 Prediction of long-term heave for the trial excavation

The results illustrated in Figure 13-8 show that a heave generated at the base of the 10 m deep excavation after 120 years is in the order of 325 mm primary heave (average of data “Primary heave, fully excavated”) plus an additional 115 mm secondary (blue “secondary heave” line) heave, giving a total heave of approximately 440 mm. In the case where there is a partial backfilling of the excavation of 25 kPa (this is compared to the 85 kPa excavation unloading) then the total heave will be reduced to approximately 250 mm (175 mm primary from the average of the “Primary heave, partly backfilled” data plus 75 mm secondary heave from the red “Secondary heave” data). These movements compare to the “void ratio model” predictions, which is for full primary heave and several thousand years of secondary heave of 550 mm for no back fill and 400 mm for the 25 kPa back fill case. All data suggest that heave movements are significant in the long-term.

13.4.2 Plate Load Tests The plate load tests were carried out to provide data for comparison with the laboratory test results which were carried out on small samples (typically 0.07 m diameter); the plate load tests were:  4 number vertical plate load tests, 1 m diameter, typically carried out 0.3 m below general excavation level;  4 number horizontal shear plate load tests, 0.5 m by 1.0 m in plan area and with 0.1 m deep skirts, typically carried out at 0.3 m below general excavation level. The ground below the plates was allowed to come into equilibrium with vertical effective stresses of 32 and 57 kPa prior to shearing in undrained conditions;  1 number parallel passive plate load test, 2 m by 6 m installed down to 2.4 m below general excavation level.

The following comparisons are possible between the plate load test results and the laboratory and in-situ borehole-test data:  Investigation of anisotropy of strength.  Investigation of scale effects on strength.  Investigation of undrained strength as a function of stress level.

In the main part the investigation has been directed to measurement of in-situ shear strength of the folded Røsnæs clay as encountered at the base of the 10 m deep trial excavation with its base at 20 m water depth. It became apparent from the CPTU tests carried out in the base of the excavation that rapid softening occurred in the upper 1 m or so of clay as shown in Figure 13-9 where at shallow depth the cu line derived from conditions prior to the excavation taking place diverged from the strength measured after the excavation took place. It is anticipated that with time the depth of the softened zone will increase as porewater pressures dissipate at greater depth (see also Figure 13-3 above). 0,0 BS UU triaxials 0,2 11.C.461 11.C.462 0,4 11.C.463 11.C.463-1 0,6 11.C.464

0,8 cu profile

1,0 [m] 1,2 1,4 1,6

1,8 Depth below base excavation excavation base Depthbelow 2,0 0 20 40 60 80 100 120 140 cu [kPa] Figure 13-9 cu derived from CPTU (cu = qnet / Nkt, Nkt = 25.5) data and UU tests

An example of the plate load test results (vertical plate load tests) is shown in Figure 13- 10, it is clear that there is significant scatter in the results and this must be remembered when considering the conclusions reached.

0 V1 average

V2 average V3 average 50 V4 average

100

150 Vertical displacement [mm]displacement Vertical 200 0 50 100 150 200 250 Vertical applied load [kN]

Figure 13-10 Vertical plate load test results

The implication of this rapid clay softening at the base of the excavation on the plate load tests (vertical and passive only as the horizontal shear test was carried out after a hold period to achieve an equilibrium condition) is that the tests were carried out in clay with strength rapidly varying with depth in much of the failure zone.

Accepting the situation of rapidly varying clay strength in the zone of the plate load tests the following conclusions were reached: o For the folded Røsnæs clay a value of Nkt = 25.5 stated in Chapter 13.2.6 has been used to compute the profile of undrained shear strength against which the vertical and passive plate load tests have been correlated. The calculated capacities are sufficiently close to the values recorded in the field to indicate that this value is suitable for calculation of cu from CPTU data for engineering scale considerations. o The assessment of plate load tests is based on the correlation largely based on CAUc c test data (cu ) and shows no systematic divergence suggesting that anisotropy effects are small. o The resistance of the horizontal plate load tests was found to be similar to that calculated using soil strengths using the SHANSEP approach presented in Chapter 13.2.6 so long as the pre-consolidation pressure is based on σ´pc = 0.25·qnet. Importantly, the value of qnet needs to be obtained from soil at the same effective stress as a plate load tests is carried out at; i.e. for a plate test sheared with a vertical effective confining stress of 30kPa the value of qnet needs to be obtained from the part of the CPT profile where the in-situ vertical effective stress is also 30kPa (at approximately 3.5m depth below seabed level). The link of σ´pc with qnet is peculiar to this material as it implies significantly lower pre-consolidations pressures at shallow depth than at deeper depth. 13.4.3 Pile and Ground Anchor Tests The pull-out testing of piles and ground anchors has the following objectives:  Identify the relationship between soil strength and pull-out resistance and how these vary with pile and anchor types;  Investigate the increase in capacity with time;

 Investigate, for ground anchors, creep performance under static and cyclic loading.

13.4.3.1 Pile Test Results The preliminary results of the pile tests are summarised in Figure 13-11 and Table 13-5 for bored piles, Figure 13-12 and Table 13-6 for driven piles and Table 13-7 for ground anchors.

Figure 13-11 Pull-out resistance for bored piles versus pile head displacement

Table 13-5 Summary of bored pile results Pile test Average assessed pile capacity for all Average cumulative Adhesion test piles (including pile weight and pile head displacement factor, α end effects, approx. 280 kN) [kN] for all test pile [mm] 1st test * 3450 28 1.1 2nd test 3000 45 0.95 3rd test 2650 62 0.83 4th test 2400 70 0.74 5th test 2200 85 0.67 * Bored pile BP3 which has not been included in this data gives the significantly higher capacity recorded in the result.

For the bored piles (bored in the dry) there is no systematic increase in capacity with time. A design adhesion factor, α, of 1.0 is preliminarily recommended for design for situations where the pile is not allowed to fail (post peak performance shows a rapid softening of the pile resistance). When using this adhesion factor the undrained shear strength of the folded Røsnæs clay should be assessed with cu = qnet / Nkt where Nkt = 25.5.

3500

First loading

3000

2500 Second loading Fifth loading 2000

DP1 Third loading Fourth loading 1500 DP2 DP3 1000 DP4

500 DP5

Assessed ultimate capacity [kN] capacity ultimateAssessed Trendline 0 0 20 40 60 80 100 120 Cummulative displacement [mm]

Figure 13-12 Pull-out resistance for driven piles versus pile head displacement

Table 13-6 Summary of driven pile results Pile test Average assessed pile capacity for all Average cumulative pile Adhesion test piles (including pile weight and head displacement for all factor, α end effects, approx. 250 kN) [kN] test pile [mm] 1st test 3125 12 1.0 2nd test 2250 36 0.7 3rd test 2040 61 0.7 4th test 2000 86 0.6 5th test 2000 110 0.6

For the driven piles there is a small systematic increase in capacity with time with capacity increasing from 3,000 kN for the load test at 1 month to 3,200 kN at 6 months; thereafter the capacity is seen to be constant with time. This increase in capacity is associated with the reduction in installed excess pore pressures seen on the pile side piezometers. A design adhesion factor, α, of 1.0 is preliminarily recommended for design for situations where the pile is not allowed to fail (post peak performance shows a rapid softening of the pile resistance). When using this adhesion factor the undrained shear strength of the folded Røsnæs clay should be assessed with cu = qnet / Nkt where Nkt = 25.5.

The observation and findings of the pile load tests are as follows:  Five reinforced concrete bored piles were installed in the test area to 25m below seabed level. The piles were bored with temporary casing between above sea level to 10 m below seabed level. The pile was bored in the dry.  Five open-ended steel tube piles were driven to 25 m below seabed level. The piles were driven without allowing a plug to form inside the steel tube.  The back calculated peak values of the shaft adhesion factor (α) for first time loading are (shaft friction, qs = α∙cu and cu based on Nkt = 25.5): o Bored piles: α = 1.1 o Driven piles α = 1.0  Both bored and driven piles showed large reductions in α when the piles were loaded beyond peak resistance as a result of strain softening. At pile head displacements of 100 mm the values of α reduced to approximately 60 % of the peak values.  Neither the bored or driven piles showed significant increases in capacity with time after installation.

13.4.3.2 Ground Anchor Test Results Ground anchor tests were carried out on the types of anchor as stated in Table 13-7 and 13-8. All anchors had a 7 m fixed length with a nominal diameter of 0.152 m installed in the Røsnæs clay.

Table 13-7 Ground Anchor tests 2011 season No. Anchor type Test Age at Test type date test (days) G7 Non-post air flush 07/09/11 57 Investigation test grouted G8 Non-post air flush 13/09/11 63 Creep test

grouted

G11 Post grouted air flush 06/09/11 56 Investigation test G13* Post grouted air flush 13/09/11 64 Creep test G15 Dummy air flush 12/09/11 66 Free length friction test

anchor 2011 works 2011 G17 Post grouted water 24/10/11 62 Investigation test flush G18 Non-post water 25/10/11 63 Investigation test grouted flush * This anchor showed a free length shorter than tolerable, the results are not considered valid for the folded Røsnæs clay.

Table 13-8 Ground anchor tests 2012 season No. Anchor type Test Age at Test type data test (days) G8 Non-post air flush 14/11/12 491 Investigation test grouted G9A Non-post air flush 11/09/12 383 Investigation test

grouted

G10A Non post air flush 18/09/12 391 Creep test grouted G12 Post grouted air flush 13/09/12 429 Investigation test

G14 Post grouted air flush 18/09/12 435 Creep test 2012 works 2012 G16 Dummy air flush 13/09/12 433 Free length friction test anchor G19 Post grouted air flush 18/09/12 70 Creep test G20 Post grouted air flush 13/11/12 126 Investigation test

An example of the load tests results from anchors subjected to investigation tests is shown in Figure 13-13, creep tests in Figure 13-14 and cyclic tests in Figure 13-15.

900 G7 NPG AF 800 G8 NPG AF G11 PG AF 700 G12 PG AF G17 PG WF 600 G18 NPG WF 500 G20 PG AF

400

300

Anchor head displacement [mm] displacement head Anchor 200

100

0 0 50 100 150 200 Load [kN]

Figure 13-13 Investigation test results, ground anchors

Figure 13-14 Creep test load loss results corrected for reaction system settlement

Figure 13-15 Cyclic loading displacement results

The observation and findings of the ground anchors tests are as follows:  Only ground anchors with 7 m nominal fixed length and a distal end load transfer were tested.  Approximately 50 % of the post grouting injection tubes were unable to be used as the water pressure was insufficient to crack open the tubes. As such the results of the post grouting tests show a reasonable lower bound to the benefit of post grouting.  Anchor installation using water flushing rather than air flush allowed for faster drilling speed within the Palaeogene clay but lower anchor capacity.  With the exception of the non-post grouted water flush anchors the back-calculated value of adhesion factor, α, were in excess of 1.0 where α is the ratio of anchor skin friction to undrained shear strength where undrained shear strength is based on CPTU qnet / Nkt where Nkt = 25.5. For preliminary design (location specific testing will be required to validate this), the following values of α may be considered for distal end load transfer anchors with fixed lengths of 7 m or less: o Non-post grouted, air flush: α = 1.15 o Post grouted, air flush: α = 1.50 o Non-post grouted, water flush: Method not considered suitable  Creep tests were carried out by locking off the anchor load for a period of approximately 20 days. The lock off load was approximately 50 % of the geotechnical failure load. The three successful tests suggest that the creep load loss (i.e. the % loss of lock off load at constant anchor head displacement) is in the order

of 0.5 to 1.0 %/lct where lct is log10 of time in days. There is inadequate data to distinguish between post and non-post grouted anchor constructed with air flush.  Cyclic load tests were carried out to assess the cumulative displacement resulting from a load cycle from 100 % to 60 % back to 100 % of the working load (approximately 30 to 50 % of the ultimate capacity). The three successful tests suggest that the creep heave for cyclic stress is in the order of 0.5 to 2.0 mm/lcs where lcs is log10 of the number of cycle of stress. There is inadequate data to distinguish between post and non-post grouted anchor constructed with air flush.

14 Cretaceous Chalk 14.1 Geological Description Twelve of the 2009/2010 borings reached the chalk with borehole boring 09.A.008 taken almost 75 m into the deposit. As expected from the 1996-investigation, the chalk is typical Danish white chalk as known from Møns Cliff and a number of other locations. The degree of induration is almost consistently H2. Only in boring 09.A.008 a significant layer of H3 chalk has been described from the lowermost part of the borehole.

The chalk has typically average flint content less than 5 %. The flint appears to be present as nodules and not as plates, but the nodules are typically concentrated in layers. How- ever, one of the borings has passed a flint body on 0.25 m.

The age of the chalk on the boring profiles is written as Maastrichtien. A sample has been taken for dating from the uppermost chalk and from the bottom of boring 09.A.008. The upper sample was from middle Maastrichtien, the lowermost one was dated to the upper Campanian. Two further samples from the upper and lower part of the chalk in boring 09.A.019 have been analysed; they were respectively of upper and middle Maastrichtien age. As it is not possible by visual inspection with certainty to observe a difference between chalk of the two ages, and as it is not considered to be important from an engineering perspective to distinguish between chalk from the two periods, it has been decided to maintain the age description as Maastrichtien for all the chalk in the borings.

Observations of “black matter” in the chalk has been described and discussed in Chapter 10.2. 14.2 Geotechnical Properties 14.2.1 General The geotechnical properties of the chalk deposit have been investigated through:  Classification testing by Fugro of samples from the type A-borings (/2/ and /23/).  In-situ testing (CPT) by Fugro in the type B-borings (/2/ and /23/).  Classification testing by GEO of selected samples from the type A-borings for Advanced Laboratory Testing (/30/).  Advanced geotechnical testing by GEO of selected samples from the type A-borings (/30/).

Details of the geotechnical properties for Cretaceous chalk can be found in Appendix GDR 00.1-001-E.

14.2.2 Classification Properties All chalk cores have been logged and described in accordance with the Danish Geotech- nical Society Bulletin 1, /54/ with the degree of induration as detailed in Table 14-1. The classification in accordance with the International Society of Rock Mechanics (ISRM) has been included for comparison only. The distribution of the degree of induration for the recovered cores appears also from the table.

Table 14-1 Rock classification system and %o of cores with varying levels of induration- Chalk ISRM Rock Classification DGS Bulletin 1 Classification Distribution of Rock Degree of degree of indu- Description Description Grade Induration ration [%] R0 Extremely weak H1 Unlithified 6.0 R1 Very weak H2 Slightly indurated 87.8 R2 Weak H3 Indurated 5.8 R3 Medium strong Strongly H4 0.4 R4 Strong indurated R5 Very strong Very strongly H5 0.0 R6 Extremely strong indurated

The chalk is predominantly slightly indurated H2 (88 % of the recovered cores) with a small percentage of H1 and H3 material. Results of all classification testing on chalk specimens are summarised in Table 14-2.

Table 14-2 Basic geotechnical classification properties for Chalk No. of Arithmetical Standard Property Depth range results mean value deviation Water content, w 0–30 m 327 33.3% 3.1% Below 30 m 61 28.8% 3.4% Carbonate content, - 62 95.0% 1.7% CaCO3 Specific gravity of - 126 2.69 0.04 solids, ds 3 3 Dry density, ρd - 310 1.41 Mg/m 0.08 Mg/m Void ratio, e - 310 0.89 0.11 Porosity, n - 310 0.47 0.03 Saturated unit weight, - 456 18.7 kN/m3 0.8 kN/m3 γ sat Effective unit weight, - 456 8.7 kN/m3 0.8 kN/m3 γ′

14.2.3 CPTU Nine of the type B-borings with CPTU have penetrated into the chalk. The majority of the chalk material is generally categorised by a net cone resistance (qnet) between 10 and 15 MPa and a friction ratio (Rf) between 1.5 and 3.0%. Based on the borehole logging these parameters are assumed to represent the H2 material.

There are local layers with higher qnet values in excess of 30 MPa, and occasionally there are instances where refusal was met during testing. These are likely to be the stronger H3 and H4 layers and the flint bodies identified on the borehole logs.

The data compilation and data processing for the CPTUs have been described in /60/.

It appears that the corrected and correlated qnet in the chalk generally varies between 8 and 18 MPa with a slightly increasing trend for qnet values down to approximately 50 m below seabed. It shall be noted that these qnet values have been sorted and filtered as described in /60/.

14.2.4 Stress and Stress History Yield stress data derived from CRS oedometer testing is summarised in Table 14-3. Yield stress values σyield are high relative to the in-situ vertical effective stress σ'0 and when correlated against a lower bound CPTU qnet profile the ratios of qnet/σyield are as given in Table 14-3. The minimum yield stress determined was 1900 kPa.

Table 14-3 Yield stress data for chalk Property No. of data Average Standard points value deviation

Yield stress (σyield) 22 3905 kPa 1399 kPa

σyield/σ´0 22 7.7 2.8 qnet/σyield 18 3.2 0.9

14.2.5 Consolidation Properties Oedometer modulus values derived from CRS tests on specimens of intact chalk are summarised in Table 14-4. There is a relatively large scatter in the measured values as indicated by the high standard deviations relative to the average values. This is most likely due to the variable nature of the chalk as evidenced by the CPTU data accentuated by the relatively small dimensions of the oedometer test specimens.

Correlations between oedometric modulus and net cone resistance (qnet) are also summarised in Table 14-4. These correlations are based on lower bound values to the qnet profiles, because this is considered to be more representative of the chalk matrix.

Table 14-4 Constrained oedometric modulus values for chalk Property No. of data Arithmetical Standard points mean value deviation

Eoed,tan 21 370 MPa 172 MPa E1,reload,sec 22 2346 MPa 1545 MPa E1,reload,sec/Eoed,tan 21 6.6 3.1 Eoed,tan/qnet 17 34 16 E1,reload,sec/qnet 18 210 141

14.2.6 Static Shear Strength

The compressive strength (σc) and tensile strength (σt) of intact specimens of chalk have been measured by uniaxial compressive tests and Brazil tests respectively. Test results are summarised in Table 14-5 below.

100

Table 14-5 Summary of UCS and Brazil tensile strength test results

UCS test Brazil test Strength ratio σc/σt Number of results 19(*) 19 18(*)

Arithmetical mean value σc = 1.73 MPa σt = 0.22 MPa 8.4 Standard deviation 0.22 MPa 0.05 MPa 2.2 (*) Excludes low strength from 09.A.008 at 16.1 m below top of chalk qnet/c and qnet/t ratios are summarised in Table 14-6. The qnet values used to derive these relationships are based on the lower bound envelope to the relevant CPTU data.

Table 14-6 qnet/c and qnet/t ratios

qnet/c qnet/t Number of results 17 17 Arithmetical mean value 6.8 54.6 Standard deviation 2.0 15.6

Undrained shear strengths determined from the anisotropically consolidated undrained triaxial compression and extension tests are summarised in Table 14-7.

Table 14-7 Undrained shear strengths from CAU compression and extension tests C E Compression Extension Strength ratio cu /cu Number of results 26 16 11 C E Arithmetical mean value cu = 1087 kPa cu = 781 kPa 1.4 Standard deviation 223 kPa 101 kPa 0.2 qnet/cu values (equivalent to cone factor Nkt) derived from the triaxial compression and extension test results are summarised in Table 14-8. The average qnet/cu for triaxial com- C pression is twice the average qnet/c ratio consistent with the definition of c = 2 cu .

Table 14-8 qnet/cu for triaxial compression and extension strength

qnet/cu ratios Triaxial compression Triaxial extension Number of results 24 14 Arithmetical mean value 13.7 18.2 Standard deviation 3.4 4.6

Effective stress strength parameters derived from drained triaxial compression tests and undrained triaxial compression and extension tests with pore water pressure measurements are summarised in Table 14-9. In this instance upper and lower bound envelope parameters are provided.

Table 14-9 Effective stress strength parameters for triaxial compression and extension Triaxial compression Triaxial extension c′ [kPa] φ′ [] c′ [kPa] φ′ [] Upper bound 360 42 330 42 Lower bound 0 34 0 31

101

For lower effective stress levels the effective strength parameters in triaxial compression for chalk are assessed to be: φ' = 40° and c' = 50 kPa.

Modulus of elasticity values derived from UCS and CAU triaxial compression tests are summarised in Table 14-10.

Table 14-10 Modulus of elasticity values from UCS and CAU tests

Elvdt from UCS tests E50,sec from CAU tests Number of results 20 9 Arithmetical mean value 970 MPa 1007 MPa Standard deviation 525 MPa 263 MPa

Modulus of elasticity values normalised by the lower bound net cone resistance at the position of the test specimen are summarised in Table 14-11.

Table 14-11 Normalised modulus of elasticity values from UCS and CAU tests

Elvdt/qnet from UCS tests E50,sec/qnet from CAU tests Number of results 18 9 Arithmetical mean value 85 103 Standard deviation 50 34

14.2.7 Small Strain Stiffness and Damping The results of the acoustic velocity measurements in 09.A.007 were:  P-wave velocity: 2230-2464 m/s.  S-wave velocity: 1748-1807 m/s.  Eacoustic: 10847-11553 MPa.  Gacoustic: 6063-6268 MPa.

The Young’s moduli from acoustic measurements are approximately 10 times higher than the Young’s moduli from the triaxial and UCS tests. This is in agreement with the general experience from testing on slightly indurated (H2) chalk of Maastrichtien age.

The P-wave velocity is also similar to the sonic P-wave velocity determined in the geophysical borehole logging in that boring (/5/).

102

15 References

Standards are not included in the list of references.

/1/ GDR 17.0-002, Seabed CPT Campaign 2009, May 2010, prepared by Fugro. /2/ GDR 17.0-001, Boring Campaign 2009, January 2011, prepared by Fugro. /3/ GDR 03.0-001, Geophysical Surveys Onshore, April 2013, prepared by RA. /4/ GDR 03.0-002, Geophysical Surveys Offshore, December 2010, prepared by RA. /5/ GDR 04.0-002, Geophysical Borehole Logging, 2009 and 2010 Boring Cam- paign, December 2010, prepared by RA. /6/ GDR 01.3-002, Summary of Geological Conditions, July 2013, prepared by RA. /7/ GDR 01.3-003, Seismicity, May 2013, prepared by RA. /8/ GDR 01.3-001, Fehmarn Dome, May 2013, prepared by RA. /9/ GDR 01.5-001, Lolland and Fehmarn, Land Connection Areas, Ground water conditions, December 2011, prepared by RA. /10/ Eurocode 7- Geotechnical design- Part 1: General rules (DS/EN 1997-1:2007). /11/ Eurocode 7- Geotechnical design- Part 2: Ground investigation and testing (DS/EN 1997-2: 2007). /12/ Femern's Geo Nomenclature, prepared by RA. /13/ GDR 00.0-001, Geological/geotechnical Investigations, 1995/96, September 1996 (Rambøll/Haas 1996-rapport). /14/ Luke, K. 1996: Cone factors from field vane and triaxial tests in Danish soils, NGM-96 Reykjavik, pp. 203-208. /15/ Mortensen, J.K., Hansen, G., Sørensen, B., 1991: Correlation of CPT and Field Vane Tests for Clay Tills. Danish Geotechnical Society, Bulletin No. 7. /16/ Lunne, Robertson and Powell: Cone Penetration Testing in Geotechnical Practice, 1997. /17/ IRTP (1999): International reference test procedure for the cone penetration test (CPT) and the cone penetration test with pore pressure (CPTU) issued by the International Society of Soil Mechanics and Geotechnical Engineering (ISSMGE) in Proceedings of the twelfth European Conference on Soil Mechanics and Geotechnical Engineering, Amsterdam, edited by Barends et al., Vol. 3 pp 2195- 2222. Balkema. /18/ Akai, K. (1960). Die Strukturellen Eigenschaften von Schluff, Mitteilungen Heft 22, Die Technische Hochschule, Aachen. /19/ Becker, D.E., Crooks, J.H.A., Been, K. and Jefferies, M.G. (1987), Work as a cri- terion for determining in-situ and yield stresses in clay, Canadian Geotechnical Journal, Volume 24, p. 549. /20/ Casagrande, A. (1936), The determination of the pre-consolidation load and its practical significance, Proceedings of the First International Conference on Soil Mechanics and Foundation Engineering, Volume 3, Discussion D-34, p. 60, Boston, June 22 to 26, 1936. /21/ Janbu, N. (1969). The resistance concept applied to deformation of soils. Proceedings of the 7th International Conference on Soil Mechanics and Foundation Engineering, Volume 1, p. 191, Mexico City, August 1969. /22/ Femern A/S. Geo Information System. /23/ GDR 17.0-003, Boring Campaign 2010, January 2011, prepared by Fugro.

103

/24/ GDR 17.0-004, Onshore Fehmarn CPT Campaign 2010, January 2011, prepared by Fugro. /25/ GDR 17.0-005, Boring Campaign Lillebælt, January 2011, prepared by Fugro. /26/ GDR 18.0-002, Advanced Laboratory Testing, Laboratory Procedures, March 2011, prepared by GEO. /27/ GDR 18.0-003, Advanced Laboratory Testing, Postglacial and Lateglacial Deposits, March 2011, prepared by GEO. /28/ GDR 18.0-004, Advanced Laboratory Testing, Glacial Deposits, January 2013, prepared by GEO. /29/ GDR 18.0-005, Advanced Laboratory Testing, Clays of Palaeogene Origin, April 2013, prepared by GEO. /30/ GDR 18.0-006, Advanced Laboratory Testing, Cretaceous Chalk, March 2011, prepared by GEO. /31/ GDR 18.0-007, Advanced Laboratory Testing, Fehmarnsund and Lillebælt, March 2011, prepared by GEO. /32/ GDR 18.0-008, Advanced Laboratory Testing, UCS-testing on cement stabilised soil, March 2011, prepared by GEO. /33/ GDR 19.0-001, Large Scale Testing, Cone penetration testing and geotechnical sampling, May 2013, prepared by GEO. /34/ GDR 19.0-002, Large Scale Testing, Boring Campaign, Fehmarnsund Brücke, March 2011, prepared by GEO. /35/ GDR 19.0-101, Large Scale Testing, Trial excavation, instrumentation and surveys, December 2013, prepared by Aarsleff, GEO. /36/ GDR 19.0-201, Large Scale Testing, Driven pile installation and tension load testing, November 2012, prepared by Aarsleff. /37/ GDR 19.0-301, Large Scale Testing, Bored pile installation and tension load testing, November 2012, prepared by Aarsleff. /38/ GDR 03.0-003, Large Area Bathymetry, June 2010, prepared by RA. /39/ GDR 14.0-001, Geophysical Surveys, Nearshore, May 2010, prepared by GEUS / DHI. /40/ GDR 05.0-002, Positioning System, Outline and guide, January 2014, prepared by RA. /41/ GDR 01.4-001, Fact Finding, Existing Structures, June 2013, prepared by RA. /42/ GDR 01.6-001, Munitions, May 2013, prepared by RA. /43/ Femern Legend and Definitions, prepared by RA. /44/ GDR 19.1-001, Large Scale Testing, Boring campaign, Production site Rødby- havn, January 2012, prepared by Aarsleff/GEO. /45/ GDR 10.0-001, Production Site, Rødbyhavn, May 2012, prepared by RA. /46/ GDR 19.1-002, Large Scale Testing, Alignment Borings Lolland and Fehmarn, March 2013, prepared by Aarsleff/GEO. /47/ GDR 19.0-501, Large Scale Testing, Ground anchorage installation and testing, January 2013, prepared by Aarsleff. /48/ GDR 19.0-601, Large Scale Testing, Plate load testing, December 2011, prepared by Aarsleff. /49/ GDR 09.0-101, Large Scale Testing, Trial Excavation and monitoring, January 2014, prepared by RA. /50/ GDR 09.0-201, Large Scale Testing, Driven pile tension load testing, July 2013, prepared by RA.

104

/51/ GDR 09.0-301, Large Scale Testing, Bored pile tension load testing, July 2013, prepared by RA. /52/ GDR 09.0-501, Large Scale Testing, Ground anchor testing, January 2014, prepared by RA. /53/ GDR 09.0-601, Large Scale Testing, Plate load testing, July 2013, prepared by RA. /54/ DGF Bulletin no. 1, A guide to engineering geological soil description, prepared by Danish Geotechnical Society, May 1995. /55/ GDR 05.0-003, Local Plane Fehmarn Transformation, May 2011, prepared by RA. /56/ GDR 09.0-701, Large Scale Testing, Soil Models Palaeogene Clay, January 2014, prepared by RA. /57/ Krogsbøll A, Hededal O, Foged N (2012). Deformation properties of highly plastic fissured Palaeogene clay – lack of stress memory? NGM 2012 Procee- dings, Copenhagen, 2012, pp 133-140. /58/ Hansen, J.B. and Mise, T. (1964). An Empirical Evaluation of Consolidation Tests with Little Belt Clay, Danish Geotechnical Institute, Bulletin No. 17. /59/ GDR 01.4-004, Typical Values of Geotechnical Properties, May 2013, prepared by RA. /60/ GDR 00.1-003, Geotechnical Data Compilation and Data Processing, August 2013, by RA. /61/ IHO Standards for Hydrographic Surveys. International Hydrographic Organi- zation. Special Publication N° 44, 5th Edition. February 2008. /62/ GDR 08.0-002, Advanced Laboratory Testing, Clays of Palaeogene Origin, DSS Testing 2013, August 2013, prepared by RA. /63/ GDR 05.0-004, Fehmarnbelt Positioning System, Getting Started, December 2013, prepared by RA. /64/ GDR 18.0-009, Advanced Laboratory Testing, Pump test with clay cuttings, December 2013, prepared by GEO.

105