GPR characterization of a naturally fractured siliciclastic reservoir on Svalbard, Arctic Norway K. Senger*, J. Tveranger, W. Wheeler Uni CIPR, Uni Research Allégaten 41, 5020, Bergen, Norway jan.tveranger@uni.no, walter.wheeler@uni.no *presently at EMGS AS (ksenger@emgs.com) A. Braathen University Centre in Svalbard/University of Oslo Sem Sælands vei 1, 0371, Oslo, Norway alvar.braathen@geo.uio.no B. Heincke GEOMAR Wischhofstrasse 1-3, 24148, Kiel, Germany bheincke@ifm-geomar.de Abstract—A naturally fractured siliciclastic aquifer on Svalbard is being evaluated as a potential CO 2 sequestration site for a local power plant. Well tests indicate substantial underpressure in the 700-1000 m deep reservoir, without a well-constrained cause. However, since the reservoir is gently dipping, and thus exposed at the surface 15 km up-dip from the planned injection site, this configuration requires pressure barriers to compartmentalize the reservoir. Outcrop mapping indicates the presence of m-scale offset faults (i.e. below the resolution of the existing seismic data) and igneous intrusions, both of which may act as impermeable structural discontinuities. Igneous intrusions are also imaged using regional geophysical data sets (2D seismic, magnetic). However, their orientation and dimension are poorly constrained by the present data set. In order to map and quantify these features, a 10 km 2 irregular grid of GPR profiles (100 km total length) on surface outcrops of the target formation was carried out in April 2013 to determine the feasibility of using GPR data to characterize the reservoir. Our study indicates that large amounts of geologically meaningful GPR data can be acquired cost- and time- effectively directly on the top of the targeted reservoir. Index Terms—Arctic, CO 2 sequestration, sub-seismic faults, igneous intrusions. I. INTRODUCTION As part of the UNIS CO 2 lab project, carbon dioxide sourced from Longyearbyen’s coal-fuelled power plant (annual emissions of ~ 60 000 tons) may be stored in a saline aquifer 700 m beneath this Arctic community. The targeted reservoir belongs to the Late Triassic-Middle Jurassic Kapp Toscana Group, and comprises a heterolithic shale-siltstone-sandstone succession locally affected by Early Cretaceous intrusions [1-3]. The reservoir crops out at surface 15 km to the north-east of the planned injection site (Fig. 1a). These outcrop exposures, often eroded plateaus, are very suitable for acquiring GPR data in order to improve the characterization of the target reservoir. On Svalbard, GPR studies to image “old” rocks, and particularly reservoir analogues, were restricted to a 2 km 2 , 200 line-km survey of a karstified Carboniferous succession at Wordiekammen, where a penetration of up to 40 m was achieved [4]. In this contribution, we aim to determine the feasibility of using GPR for mapping discontinuous reflectors (i.e. sub-seismic faults, igneous intrusions and sedimentary facies variation) in the exposed analogue of the target reservoir at Botneheia, and thus provide quantitative data to be used in reservoir modeling of the unconventional aquifer. II. GPR DATA ACQUISITION AND PROCESSING Data were acquired from 16 th to 22 nd of April 2013. The spring season on Svalbard was chosen specifically to allow data acquisition on snow-covered ground, using snowscooters (Fig. 2a). This allowed rapid data recording, as well as minimal environmental impact on the fragile tundra. Profile positions were recorded with differential GPS available for the majority of the profiles. More than 100 km of profiles were recorded during four days (Fig. 1b), with acquisition speed up to 15 km/h (Fig. 2b). In this contribution, we focus on processing and visualizing the regional profiles, in order to illustrate whether geological features (e.g. sedimentary strata, faults, igneous intrusions) may be mapped at Botneheia using GPR. Based on the results presented herein, additional surveys could be acquired within a dense 3D grid and with higher frequencies (e.g. 100 MHz) Profiles were oriented along numerous directions, to enhance the chance of intersecting structural discontinuities (i.e. faults), whose orientation was poorly constrained from previous work though presumed to reflect the orientation of small-scale fractures mapped at outcrops [1]. Furthermore, acquiring long GPR profiles was emphasized in order to gather significant discontinuity spacing measurements along different directions. Typical profile spacing across this irregular grid varied from 10 meters within a 100*100 m large study area to 250 m spacing between the regional profiles. Finally, numerous profiles were acquired directly over a 5 m thick doleritic dyke clearly exposed in the study area.