As an example, QPAC has been used to model the evolution of clay sediments at Searles Lake located near Death Valley, California. Searles Lake is an evaporite basin containing smectite mudstones that have been altered under mildly alkaline conditions (9 < pH < 10) over a 3 million year time period. The site is a useful natural analogue for the study of potentially deleterious alteration of clays by cementitious porewaters in geological repositories for radioactive wastes. Whilst the pH is lower than that associated with the use of fresh Ordinary Portland Cement (OPC) it is representative of conditions after fresh OPC has undergone partial reaction with geological materials, or of 'Low pH cements' that are being considered with the intention of reducing the potential amount of clay alteration.
The QPAC model was developed in collaboration with JAEA. The model simulates flows and reactions in the gradually forming sediment column, with flows varying with time as deeper-lying sediments undergo compaction and become more resistive to flow. To construct the model, the QPAC reactive transport module was coupled to a custom model for the evolving hydrogeology that includes a simplified effective stress model to simulate the compaction of the sediments. Due to the highly saline nature of the historic porewaters at Searles Lake, the Pitzer virial model for aqueous activities was incorporated into the model. Pitzer data from the Yucca Mountain database was used in the simulations.
Published mineralogical investigations have shown that up to 70% of detrital clay may be replaced by authigenic minerals such as Fe-illite, analcime, and K-feldspar over timescales <100 ka. Kinetic dissolution of detrital smectite under alkaline conditions was described using one of two models based on departure from thermodynamic equilibrium or by an empirical rate dependent upon aqueous Si concentrations. Secondary mineral growth was modelled with rates estimated using mineralogical data from the Searles Lake drillcore. The zonal pattern of smectite dissolution observed at Searles Lake was reproduced reasonably well by the Cama TST model of montmorillonite dissolution; the other models strongly overestimated clay dissolution.
The results of the simulations compare well with core samples taken from the site. An animated version of the results of one of the simulations can be seen below. Also see the associated paper Savage, Benbow, et al. (2010).
