Synthetic diagenesis – the key to carbonate systems in basin modelling

Carbonate rocks form significant hydrocarbon reservoirs, storing around 50-60% of the world’s petroleum reserves (e.g., Burchette, 2012). Basin modelling packages, which are used to predict hydrocarbon charging, need to represent basin evolution, diagenesis, as well as the more typical basin processes. These packages incorporate defined relationships between porosity and effective stress (or depth) as well as defined lithology descriptions that control many other properties (e.g., total thermal conductivity). The given relationships may work well for most siliciclastic rocks but not for carbonates. In fact carbonate rocks’ porosity evolution with burial/time is very sensitive to the chemical changes that are able to generate calcite cements and dissolution, dolomitisation (replacement and dissolution) as well as mechanical and chemical compaction. So porosity changes are not easily predicted by just the thickness and density of the overburden: the diagenetic evolution must also be represented. A range of porosity-depth curves for carbonates have been proposed, using experimental data (e.g., Shinn and Robbin, 1983) and measured subsurface data (e.g., Schmoker and Halley, 1982) or using representative equations (e.g., Goldhammer, 1997). Although these datasets are very valuable, they are not applicable everywhere because they are a response to the specific geological history, and inevitably to local changes in physical and chemical conditions. Giles et al (1998) proposed a workflow using 100 to 1000 samples from facies with similar initial properties taken over a wide depth range that have similar porosity/depth or porosity/effective stress behaviours and that are currently at their maximum effectiveness, in an area with similar geothermal gradient and without overpressure. Although their workflow is very useful, sometimes the diagenetic history of the reservoir (i.e. cementation and dissolution phases) is too different from that of the overburden rocks and so this workflow is not always appropriate. Here we use a synthetic diagenesis approach applied to a Lower Eocene carbonate reservoir with a complex diagenetic history. This requires thin sections or core photographs as initial input and treats diagenesis as a series of textural changes (e.g., cementation and dissolution). Then we apply the method of diagenetic backstripping proposed by van der Land et al., (2013), using Pore Architecture Models (PAMs) (Wu et al., 2006) to generate representative porosity values from the solid/pore geometry and distribution. These time-sequenced values are then attached to specific depths and times using basin modelling, initially with the pre-synthetic diagenesis basin model. The synthetic diagenesis approach is used to update the basin state calculations, with an iterative cycle, leading to a solution that honours the physics and chemistry of the system. Since we can make many textural observations, they provide a useful constraint, and may allow us to choose a small set of possible basin-history cases that are compatible with the observed rock data and the underlying process understanding.