Seismic time-lapse effects of solution salt mining – a feasibility study

This article addresses the question whether time-lapse seismic reflection techniques can be used to follow and quantify the effects of solution salt mining. Specifically, the production of magnesium salts as mined in the north of the Netherlands is considered. The use of seismic time-lapse techniques to follow such a production has not previously been investigated. For hydrocarbon production and CO2 storage, timelapse seismics are used to look at reservoir changes mainly caused by pressure and saturation changes in large reservoirs, while for solution mining salt is produced from caverns with a limited lateral extent, with much smaller production volumes and a fluid (brine) replacing a solid (magnesium salt). In our approach we start from the present situation of the mine and then study three different production scenarios, representing salt production both in vertical and lateral directions of the mine. The present situation and future scenarios have been transformed into subsurface models that were input to an elastic finite-difference scheme to create synthetic seismic data. These data have been analysed and processed up to migrated seismic images, such that time-lapse analyses of intermediate and final results could be done. From the analyses, it is found that both vertical and lateral production is visible well above the detection threshold in difference data, both at pre-imaging and post-imaging stages. In quantitative terms, an additional production of the mine of 6 m causes time-shifts in the order of 2 ms (pre-imaging) and 4 ms (post-imaging) and amplitude changes of above 20% in the imaged sections. A laterally oriented production causes even larger amplitude changes at the edge of the cavern due to replacement of solid magnesium salt with brine introducing a large seismic contrast. Overall, our preimaging and post-imaging time-lapse analysis indicates that the effects of solution salt mining can be observed and quantified on seismic data. The effects seem large enough to be observable in real seismic data containing noise.