DHI risking |
  
|
DHI risking |
|
This module can be used in REP where seismic amplitude anomalies and flat-spots may be Direct Hydrocarbon Indicators (DHIs). It is mainly intended for risking observed DHIs, but can also be used to risk the absence of a DHI where one should exist. The end result is a DHI-driven Chance of Success (DHI-COS) for the prospect, which can be compared with the geological Chance of Success (GCOS) to arrive at a final result (see below).
It is easy to make a list of the factors that ought to be considered when assessing a DHI (see e.g. Bacon, Simm & Redshaw, 3D seismic Interpretation, Cambridge University Press 2003) but there is little or no published information on how best to combine these factors to arrive at a prospect COS. The current module considers the DHI information under four headings:
1. Data: quality and quantity.
2. Amplitude Anomaly: is it convincing?
3. Flat-spot: is it convincing?
4. Analogues: what do drilled examples of similar DHIs tell us?
Empirically based on examples from several basins worldwide, REP assigns a calibrated weighting to these four headings of 15%, 35%, 20% and 30% respectively, when combining them into a DHI-COS.
There are of course many other types of possible DHI. These include both seismic (e.g. a gas chimney) and non-seismic (e.g. EM anomaly) observations. These are not currently included in REP's DHI risking but if enough calibration data become available, it may be possible to include these in a future version.
Local well control: To assess the relevance, consider the similarity of stratigraphy, facies, interval thickness, and geographical proximity of wells to the prospect. Log data: the ideal is to have nearby P and S sonic and density coverage over the same reservoir and for at least several hundred feet above and below it. Where applicable, logs in hydrocarbon-bearing reservoirs should ideally have been corrected for the effect of invasion by mud filtrate.
Seismic data: To assess the quality, consider the survey type (2D vs. 3D, single vs. multiple vintages), signal/noise, multiple suppression, amplitude preservation during processing, and distortion of amplitudes by overburden effects (especially shallow gas). Phase/polarity: if there is no direct well-tie, then the appearance of seabed or the top of a thick soft layer (e.g. Kimmeridge Clay/Draupne Fm. in the North Sea) or thick hard layer (e.g. basement) may give a good qualitative indication. AVO and impedance volumes: a complete set would ideally include angle gathers, angle stacks, intercept and gradient volumes, chi angle projections, acoustic impedance inversion and elastic inversion volumes.
In assessing how well it is understood what a DHI amplitude anomaly should look like, remember to include AVO effects (e.g. changes from near to far traces). In thinking about the relative size of fluid and lithology effects, consider not only the impedance difference but also the way that seismic response will be affected by bed thickness changes; tuning can increase or decrease the amplitude response, depending on the bed thickness relative to the dominant seismic wavelength. When considering whether the observed anomaly is consistent with being a DHI think about the relative amplitude change and polarity. Fit to structure is considered in the next section, where one should remember to allow for uncertainty in depth conversion.
In thinking about what a flat-spot should look like, one should include AVO behaviour as well as amplitude relative to lithological reflectors. Also consider possible departure from flatness in TWT due to lateral velocity variation in the overburden, or a genuine tilt in depth due to hydrodynamic and/or diagenetic effects. In assessing whether the flat-spot could be a multiple, consider its amplitude, relation to possible multiple generators, and discordance to bedding if applicable. Fit to structure combines two elements: whether the flat-spot terminates cleanly at the top seal (and bottom seal if relevant) and whether it is continuous within the reservoir across the entire mapped structure.
The first question is whether there are similar DHIs locally that have been drilled and for which you know the well outcome. Similarity of the analogue to the prospect combines geographical proximity, stratigraphy, trap type and prospect depth. The question on dry analogues should be taken to mean: among the analogues that show a similar DHI, what proportion when drilled did not have producible hydrocarbons in the interval of interest? Thus, say gas at low saturation would be regarded as a failure case.
Consider which of the two risks (DHI-COS and GCOS) you have more faith in and weight the overall COS accordingly using the slider bar. In REP this is done as a subjective exercise. It is recommended to be done by consensus and subjected to peer discussion after the two elements of risk assessment have been completed. No formal mechanism for combining the risks has been incorporated in REP.
Bayesian risk modification can be used as a more objective method of combining risks. It is an alternative approach that may be a helpful aid to thought for enthusiasts. To do Bayesian risk modification, consider all the elements that have been incorporated into the DHI risking, and estimate the values for:
P1: The probability of having a seismic appearance like that observed, given hydrocarbons are present (a true positive).
P2: The probability of having a seismic appearance like that observed, given hydrocarbons are not present (a false positive).
G: Geological Chance of Success.
[Note: Bayesian risking is a modification of G in the light of new information (P1 and P2). As such, this is a valid approach only if G has been arrived at without incorporating any of the DHI information in P1 and P2. This is not likely to be the case for many stratigraphic traps whose extent is already defined by a seismic amplitude or other DHI.]