Alpine Fault fluid composition and transport in proximity to the fault, using analyses of out-crop samples and gases entrained in drill fluids

Abstract

The chemistry of fluids and gases within fault zones can provide information about their sources/origins, flow rates and paths, and further fluid-rock interactions along those paths. New Zealand’s tectonically active Alpine Fault Zone is an ideal system within which to study fluids because there is a wealth of supporting prior research about it. Furthermore, globally there are very few systems that are similarly both late in the seismic cycle, and known to rupture in large events. Phase 2 of the Deep Fault Drilling Project (DFDP-2) aimed to drill to, and through the Alpine Fault at ~1 km depth to obtain information about its physical state and structure before it next ruptures. Online gas analysis (OLGA) was employed to monitor gases (N2, O2, Ar, CO2, CH4, He, H2, hydrocarbons, and Rn) entrained in the circulating drill mud, which were derived from crushed rock at the drill bit, and from permeable layers within the drilled formation. 33 complimentary offline samples collected at the time of drilling over the depth range 236 to 892 m were later analysed for standard % of gas by volume, and He and carbon isotopes. He isotope data were also acquired from 14 core samples of outcropping Alpine Fault mylonites from four different locations. In both surface and offline DFDP-2B samples, 3He/4He ratios of bulk rock showed R/Ra values of 0.46 (±) and 0.57 (±), increasing with proximity to the fault. This indicates interaction with mantle derived fluids at depth, and supports a previous suggestion there is a present day fluid flux from the mantle along the fault. However, hangingwall and footwall He ratios differ sufficiently to support previous suggestions based on Sr isotopes that the Alpine Fault is an effective barrier to cross-fault fluid flow. OLGA data were commonly of low quality due to various drilling-related disturbances, or equipment problems. However, some useful data were collected. The most notable signal in OLGA was variation in Rn, inferred to reflect interception of particularly permeable horizons in the surrounding formation.