The Chalk Group was first defined by Rhys (1974) in the UK southern North Sea, and later extended into the central and northern North Sea by Deegan & Scull (1977). The chalk Group forms important reservoirs for hydrocarbons in the Danish, Norwegian, Dutch and UK sector of the North Sea (Johnson, 1993). The Upper Cretaceous to Danian Chalk Group of the Danish Basin and the Danish North Sea has significantly impacted Danish society for centuries, as it constitutes the most substantial part of the Danish subsurface. Initially as fertilizer on the fields and more recently as reservoir for hydrocarbons and as an important drinking water aquifer. As Denmark moves towards a net zero climate target, which is to be reached by 2045, the Chalk Group may again be a key component. In particular, (depleted) oil- and gas fields of the North Sea are considered possible reservoirs in the context of reducing atmospheric concentrations of CO2 trough carbon capture and storage (CCS). Denmark has ambitions of becoming a European hub for CCS trough carbon storage in the Danish subsurface. In order to facilitate this process, it is imperative that we assess the chalk reservoirs from a fresh perspective. This entails integrating the wealth of existing data gathered from years of extensive research and production, and coupling it with new datasets made possible through technological advancements.

This thesis focusses on natural fracture system in the Upper Cretaceous Danish Chalk Group. Chalk is largely a dual porosity reservoir rock, which implies high porosity and low permeability. Fluid flow rates, directions and bafflers through the chalk therefore largely depends on the configuration of these fracture systems, which typically includes fractures in the size range of mm- to km. Consequently, the more accurately we can describe, predict and simulate fracture patterns, the more efficiently, cost-effective and environmentally sensible we can plan and execute storage-schemes.

Subsurface fracture characterization is fundamentally difficult, as mm-cm scale fractures are significantly below the resolution of 3D seismic data. Fractures in this size range have been known to enhance the effective permeability in the chalk reservoirs by 20 times. Well-scale borehole imagery (BHI) and core data provide a constraint on these fractures in subsurface scenarios. Both coring and BHI processes are however time consuming and expensive.

Moreover, in a typical scenario, the wells sample less than 1 % of the target reservoir. Onshore analogues therefore have the potential of significantly improving subsurface fracture characterization, as they are the only possible direct observations of natural fracture system of larger areal extents.

The Rørdal Quarry, situated in Aalborg, North Jutland is the primary study area in this thesis. It exposes chalks of Maastrichtian age that are considered analogues to the North Sea chalk reservoirs. Structural analysis and fracture characterization are enabled by an integrated dataset, based on remote sensing (digital outcrop models generated using drone imagery) and high resolution shallow geophysical data (ground penetrating radar and shallow seismic). Here, fracture characterization is approached from a geoscience perspective, with the ambition of collecting, integrating, interpreting and presenting data in a manner that is useful for the purposes of fluid flow simulations in the chalk. Additionally, familiarity with the fracture distributions of the subsurface chalk is imperative in predicting the migration of CO2 plumes after injection, which is a substantial part of monitoring storage reservoirs. The study of natural fractures on a broader scale has application to a range of other practical applications, including hazardous waste disposal and earthquake hazard assessment.

The Rørdal Quarry is under current excavations by Aalborg Portland. Excavation activities on two levels (the upper- and lower levels, respectively) allow for the repeated surveying of two NS oriented chalk walls with approximate lengths of 1 km and heights of 7-11 m. By surveying the quarry levels at different times of production, which corresponds to different positions in the Y (west-east) direction, a time series of digital outcrop models (DOMs) was generated. The combination of the DOM time series with the shallow geophysical data provides the rare opportunity of tracing structural elements in the chalk along strike. One of the DOMs is used for the extensive mapping of fractures, resulting in a dataset consisting of more than 27 400 fractures, covering three orders of magnitude. The volume of this dataset is unique in a global context: it cannot be obtained in the subsurface and few (published) efforts of this size have been made in onshore successions. The benefits of the dataset include: 1) possibility of study lateral variations in the exposed chalk and evaluate the implications for subsurface usage of this information, 2) evaluate fracture length and spacing distributions from an areal sampling perspective on a substantial global population 3) Consider the implications of the fracture system geometry on subsurface fluid flow and evaluate generic applicability of the data, 4) calibration of fracture modelling algorithms and 5) provide input for e.g. image recognition algorithms aimed at automatic fracture picking. Points 1-4 are executed in the work presented here.

Through the integrated structural analysis, three main structural elements are identified and evaluated along strike: a northern extensional fault zone, a southern strike-slip fault zone and subtle E-W oriented folding. The fault zones are steeply dipping towards the SE and are segmented along a NE-SW strike direction. By mapping of bedding a conceptual geological model and a structural framework is established. The structural framework is used for fracture simulation through a newly developed tool for dynamic fracture modelling (DFM), which provides an alternative to the traditional subsurface modelling efforts that rely heavily on extrapolation from sparse well data. Substantially more accurate fracture predictions are expected from the DFM approach, as it simulates fracture growth based on local stresses. Sensitivity analysis and calibration of the DFM tool are performed by evaluating implicit and explicit output against the detailed fracture dataset, mapped on the DOM.

The fracture characterization effort in the Rørdal Quarry indicates that fracturing in the exposed chalk occurs in a layer-bound system in which fractures are following a log-normal distribution of fracture trace lengths and close-to regular spacing, contrary to some published examples fracture growth patterns, interpreting smaller mixed datasets as fitting log-log distributions. Comparison with fracture populations in other Danish chalk exposures from published data suggests that this is a regional trend.

The fracture attributes of the chalk section exposed in the Rørdal Quarry is successfully reproduced through a newly developed tool for dynamic fracture modelling. In contrary to stochastic modelling approaches, which are the industry standard for subsurface fracture modelling, the DFM simulates fracture growth in response to an applied stress. This circumvents the need for extrapolation from sparse well data, and is expected to significantly improve the predictability of subsurface fracture networks and subsequent fluid flow simulations. Sensitivity analyses are performed on the DFM, trough evaluating three strain scenarios, related to the main structural elements defined in the Rørdal chalk exposure.