Controls on natural gas migration in the western Nile Delta fan

Abstract

The aim of this study is to combine petrophysical and geochemical data in order to reconstruct the migration history and pathways of mixed microbial-thermogenic gases drilled on the Nile Delta fan, offshore Egypt. While general interest lies in understanding migration routes, rates and mechanisms special attention is dedicated to understanding (1) the origin of gas in both reservoir and non-reservoir units using chemical and isotopic fingerprints and (2) whether a free gas phase supports relatively rapid leakage via bulk flow in non-reservoir units, both above and below commercial accumulations. The Pilocene section in this study is a classic slope environment comprising channels, mud-rich turbidites, mass transport complexes and hemipelagites. Data from seismic and drilled wells suggest that the channel and levee reservoirs are rarely full to spill, implying either a lack of charge and leakage rates which precludes complete filling of the structures. The provided data set enables a quantitative assessment of gas distribution and its genetic fingerprint in the context of both stratigraphic position and lithology. Data is reported from 25 wells, each provided with a conventional wireline log suite and some with borehole images and high-quality core images. Gas concentration data, plus compositional and isotope data are available for isotubes and headspace gas for both reservoir and non-reservoir units. Small-to-medium scale linear and non-linear depth shifts between different techniques (core recovery, core logging, wireline logging) in conjunction with scale and resolution issues demanded logical/stochastic depth synchronisation and well as harmonisation of signal resolution (typically up-scaling). Accordingly, great care was taken to depth-match core, log and gas data. In general, there is evidence of leaking thermogenic and partly biodegraded gas from the reservoirs under investigation, while some microbial methane appears to be retained in the cap rock. Careful assessment of the maturity of the thermogenic gas charge suggests that in a given structure, maturities are similar throughout the sampled section of underseal, reservoir and top seal. Furthermore, compositional temperature stratification suggests a balance between influx of fresh gas and microbial metabolism rates, supporting the concept of a dynamic charge-leak scenario that is sustaining hydrocarbon fermenting microbial communities in the deep biosphere. It was found that microbial recycling of hydrocarbons at depth enables the identification of diffusive gas mixing pathways in the combined analysis of methane and ethane compositional and isotopic data. The proposed diffusion pattern supports the idea of a widely present coupling between both methanogenic and biodegrading microbial communities that exhibit strong carbon isotopic dis-balances at gas-water contacts (GWC) where nutrient supply is in favour of the biodegrading metabolism. Although the hypothesis of coupled diffusive/microbial gas overprints complies with (1) various literature reports that microbial attack on free gas phases is hindered by restricted physical access and (2) segregative isotope fractionation as a consequence of differences on methane and ethane diffusivity, it is conditional to the nature of gas mixing patterns along borehole trajectories in the context of lithology and pore fluid saturations. Undoubtedly, the ubiquitous presence of microbial gas has consequences for vertical net leakage. As classic empirical wireline models for hydrocarbon saturation (i.e. free gas phase volumetrics) are not suited for clay-dominated cap rock sections, an alternative approach presented in this study is based on total gas (TG) modelling from nuclear logs and its solubility in the formation of brine. The calibrated saturation model is scale-independent and implies that free gas occurs on the most of the clay-dominated non-reservoir sections. However, model resolution is not sufficient to capture the suspected filamentary network of free gas phase within the mudrock pore space that enables relatively rapid leakage via Darcy flow. In an unique attempt to validate manual and thereby subjective lithofacies allocations to core images a subset of rock sample properties such as grain size fractions and porosity were successfully modelled using quantitative core image properties. However, model validity appears to be restricted to clay-rich lithofacies due to both an absence of calibration data for sands and occurrence of abnormally dark sandstone units. Further, an artificial neural network (ANN) was trained to propagate the calibrated core fancies along the entire wireline logged borehole section to set the lithological context for a detailed fluid flow analysis. Reproducibility of input (core) facies by output (wireline) facies is similar to the reproducibility by fellow geoscientists but could not be significantly improved to 60-80% of reliability by reduction of facies types. The study shows that a combination of geochemical data with lithological and petro-physical information generates detailed insights into rates, mechanisms, and pathways of two phase flow through the deep biosphere of gas-charged basins. Vertical, geologically rapid flow through mud-rich sequences is a viable migration route for gas if the influence of cap rock bypass systems (permeable faults, sandstone intrusions, mud volcanoes etc.) is restricted. It was found that an adequate quantification of both thermogenic gas fraction and diffusive gas mixing fingerprints is crucial to identity stratigraphic intervals that are not dominated by advective leakage through the pore space and are consequently bypassed.