1] This study investigates vertical flow and pesticide transport along fractures in water saturated unoxidized clayey till. From two experimental fields, each 40 m 2 , 96% and 98%, respectively, of total vertical flow was conducted along fractures in the till, while the remaining 2-4% of flow occurred in the clay matrix at very slow flow rate. An applied dye tracer was observed only along 10-26% of the total fracture length measured on the horizontal surface of the experimental fields. In vertical sections the dyed fracture portions constituted root channels, which penetrated the till vertically along the fractures into the local aquifer at 5 m depth. No dye tracer was observed in the fractures without root channels or in the unfractured clay matrix, suggesting that root growth along the fracture surfaces was the principal agent of fracture aperture enhancement. Using hydraulic fracture aperture values determined from large undisturbed column (LUC) collected from one of the experimental fields, it was estimated that 94% of flow in the fractures was conducted along the fracture root channels, while only 6% of flow was conducted along the fracture sections without root channels. For natural vertical hydraulic gradients (0.8-2.3 at the site), flow rates of 0.8-2 km/d were determined for a fracture root channel, while fracture sections without root channels revealed flow rates of 9-22 m/ d. Corresponding flow rates in the unfractured matrix were 7-19 mm/yr. For infiltrated bromide (nonreactive tracer) and mobile pesticides mecoprop (MCPP) and metsulfuron, very rapid migration (0.28-0.5 m/d) and high relative breakthrough concentrations (30-60%) into the aquifer were observed to occur along the fracture root channels using a constant hydraulic gradient of 1. Only traces were measured from infiltration of the strongly sorbed pesticide prochloraz. The concentrations of the bromide and pesticides in the monitoring wells were modeled with a discrete fracture matrix diffusion (DFDM) model coupled with a single porosity model (SP) for the till and aquifer, respectively. Using effective fracture spacings and mean fracture apertures for the fracture channel sections as modeling input parameters for the till, the concentrations observed in the wells of the aquifer could be reasonably approximated.

When using monitoring wells for investigation of contaminant sources in clayey till, there is a high risk that fractures may cause mobile contaminants to bypass the monitoring wells. This paper indicates that the probability of interception between monitoring wells and hydraulic conductive fractures is often significantly less than 50%. Based on a field experiment and application of a calibrated discrete fracture matrix diffusion numerical model (FRAC3Dvs), the paper also evaluates pesticide-monitoring results for different positions of monitoring well screen relative to fractures. For well screens situated 0.25 and 2 m from a conductive fracture, the first concentrations of the pesticide metabolite (2,6 dichlorobenzamide, "BAM") would be measured two years and 18 years, respectively, after the contaminant had been transported into an underlying aquifer. In this way, underlying aquifers may be subjected to contamination by downward moving contamination without being observed in monitoring wells in the till.

No abstract

The Halfdan Northeast gas field is a widespread, thin accumulation in high porosity, low permeability chalk of Danian age. Successful development relies on implementation of long single- or multilateral horizontal wells in thin reservoir zones and securing the maximum number of reservoir feet for each well bore. Consequently, detailed knowledge of reservoir geometry and properties is essential. This paper describes how the detailed reservoir architecture was resolved by geophysical reservoir characterization based on rock physics understanding using optimized seismic elastic inversion products. The chalk reservoir is difficult to distinguish from the overlying shale on conventional seismic as the effects of porosity and gas/fluid mixture cause the reservoir chalk to exhibit similar acoustic impedance as the shale above. The placement of wells is further challenged by the fact that erosional depressions, ranging in size from ca. 400 to 6,000 metres across, are present in the chalk. Optimized elastic inversion allowed mapping the top of the continuous chalk and the base of the overburden shales, whereby the infill of these structures could be characterized, consisting of marl and disturbed chalk. A well targeting the disturbed chalk infill confirmed this interpretation and found it to be of reservoir quality. The further development strategy is to tap both the laterally continuous chalk and the disturbed chalk bodies to maximize recovery. Introduction Halfdan Northeast is a chalk gas field offshore Denmark, comprising a thin but widespread accumulation with very low relief. The reservoir is coccolith chalk of Danian age (Ekofisk Formation), characterized by extremely small grain size (average around 1 micrometer), high porosities (30–45% in reservoir zones) with a cyclical development and very low permeabilities (typically 0.1 - 1 mD). Figure 1 shows the field location and a typical vertical well log. The gas bearing chalk is generally less than 70 ft in thickness, and overlies water bearing chalk of variable porosity. Target zones for reservoir drilling are typically less than 20 ft thick. The reservoir chalk is capped by marl and shale of Tertiary age. Ensuring recovery from Halfdan Northeast relies on placing long horizontal wells in the thin high porosity reservoir zones. Due to the thinness of the reservoir and the presence of mobile water directly below it, fracturing is not viable. Because of the very low permeability, the production from wells is strongly correlated to reservoir contact, i.e. the amount of reservoir feet (Reference 1). For well implementation, the key to success is therefore a detailed understanding of reservoir geometry and characteristics combined with the ability to optimally place these horizontal wells using geosteering. The interface between Danian chalk and overlying Tertiary marl and shale may often be mapped on seismic as an increase in acoustic impedance. However, if the porosity of the chalk is sufficiently high and the chalk is gas-bearing, the interface may be mapped as a decrease in acoustic impedance2–3. In the case of Halfdan Northeast, the reservoir is gas-bearing, but of a porosity that makes it impossible to map top reservoir from near stack amplitude seismic, because the acoustic impedance of the reservoir is similar to the impedance of the overburden. This paper describes how this challenge has been met by a multidiciplinary, iterative approach comprising rock physics, geophysical reservoir characterization, geological interpretation and modeling and well implementation.