Mixed HC Columns – do’s and don’ts
In many basins there are hydrocarbon fields containing both oil and gas. Often with a wide range of relative amounts of gas and oil: from only oil to only gas, and every conceivable gas-oil mix in between these two endmembers.
How should we do Risk & Volume Assessment of prospects in those settings?
There are three approaches that I have seen being used:
- Treating gas as a risk. In areas where we are after oil only, gas is sometimes taken as a risk. The prospect POS is then the probability to find oil.
- Merging the R&V results of an oil case and a gas case. Because of the inherent uncertainties of predicting the percentage of oil and gas, two cases (scenarios) are run: an oil case and a gas case. These are carried through the R&V assessment separately and then they are merged for final prospect POS and volumes.
- Subdividing the HC column in a gas and oil column. The percentage of the total column that is expected to be gas (or the height of the gas cap in metres) can be specified (with appropriate uncertainty ranges). Or, for a gas case an oil rim can be input in the same way.
My preference is clearly for approach 3). Before going into how this works, I’ll explain why approaches 1) and 2) are to be avoided.
In approach 1 (treat gas as a risk), it is assumed that the structure will be filled with oil only, even though a mixed column is expected. In a mixed column, part of the trap will be filled with gas, reducing the trap volume available for oil. This is ignored in this approach, and the oil volumes will therefore be overestimated.
This approach confuses uncertainty for risk. Volume is an uncertainty element, not a risk. Colloquially we may speak of a ‘gas risk’, but in Risk & Volume assessments we should calculate volumes of what we expect to find, not of what we hope to find or of what we can commercialise.
Finally, the development of an oil field without free gas is fundamentally different from developing an accumulation with a mixed gas-oil column. If the prospect’s commercial viability is assessed based on the R&V assessment of an oil case without a gas cap, it will be overly optimistic and will have little to do with reality.
Approach 2 (merging oil and gas cases) is often selected when it is too difficult to decide on percentages of oil and gas in the HC column. R&V assessments are then done for two scenarios (cases): oil only and gas only. These scenarios are then merged for a total prospect R&V assessment.
Merging of scenarios requires indicating scenario weights. Let’s say that we have a 60% oil case and 40% gas case. Probabilistic R&V applications will then pick randomly 60% of the cases from the oil only assessments and 40% from the gas only assessments. Based on those, the Expectation Curves for oil and gas are plotted and Mean Success Volumes for oil and gas are calculated.
In fact, none of the probabilistic iterations will then have a mixed oil-gas column! Just like approach 1, this approach has little to do with what we expect to find. It is in fact a bit like ‘throwing in the towel’. We find it (too) difficult to predict the percentages of the column that are oil and gas, and therefore we leave it to the computer to sort this out for us. This is not the way to go!
Approach 3 is clearly the one to follow. On the next page I try to explain how this works.
The correct approach: subdividing the HC column is gas and oil (as percentage or in metres)
The more complete R&V applications all offer the possibility to input for an oil case a gas cap, and for a gas case an oil rim. This works as follows:
- Input the probability of finding a gas cap or oil rim: a percentage between 0% and 100%.
- Indicate either the:
- The percentage of the column formed by the gas cap or oil rim as an uncertainty distribution: minimum – mode – maximum (a beta distribution is generally recommended)
- The length of the gas cap or oil rim in metres as an uncertainty distribution: minimum – mode – maximum (a beta distribution is generally recommended)
Based on the column length uncertainty a specific column length will be picked for each probabilistic iteration. The percentage of the HC column formed by the gas cap (e.g. 0 – 35 – 80%) is then applied to each iteration. Note that the uncertainty for gas cap can be truncated at 0%. This is recommendation for cases when there is a probability of no gas cap.
When the input of a gas can (or oil rim) is done in metres, the length of the gas cap (or oil rim) is independent of the column length. Note that when using this option, for some of the iterations a trap may result with only gas: when a short column length is picked and a larger gas cap length. This is a difference with the option where the gas cap is always a percentage of the total HC column.
The probabilistic tools at our disposal allow us to input mixed gas-oil columns with sufficient degrees of freedom to cater for all situations.
It may be difficult to predict what the oil-gas mix will be. If that’s the case, large uncertainty ranges will have to be assumed, and the best guess is the mode of the uncertainty distribution.
Leaving the decision to the computer is never a good option.
Read the next volume of the blog series:
- Rethinking Trap Distribution Through Seal Effectiveness
- Reporting Risk & Volume Assessments: Key Insights and Best Practices
- The Bull’s Head Risk Matrix: A Data-Driven Approach to Risk Assessment
- Prospect Police versus the Socratic Method
- Data Acquisition for Exploration: Risk Reduction or POS Polarisation?
- Exploring Stratigraphic and Combination Traps: Definitions, Risks, and Applications
View upcoming related courses:
Trap & Fault-Seal Analysis, Modeling for Oil & Gas and CO2
