contentsSeismic Facies Mapping: Part 2
Rob Kirk, Rob Kirk Consulting
This is the second of three articles where I would like to promote the making of facies maps from seismic to help with our geological understanding of our blocks.
The previous article (PESA News Resources October/November 2011) looked at what seismic facies are and gave one method for doing this. In this article we will look at other ways of doing it on the machine as well as giving a few tips and suggestions.
First, a few check points when you are getting started.
Figure 16 (numbering following on from previous article - references at end of third article, left column is time-depth pairs and right column is formation tops and times-names blanked as data is operational-Kingdom format) reminds you to personally check that your time depth curves are all "present and correct" in the project. Getting these in to shape can be a non-trivial matter!
Most jobs I find that I have to input my own 'formation tops' during the work as the originals usually end up needing modifying or at least infilling. (Do not use million year ages, or mfs or SB naming - as these usually change as the job commences. Use something 'sensible' that jumps in steps of say 100, so that when you add another horizon later you have numbers available).
Figure 17 is the 'final' tied gamma log on the seismic (with formation tops on but obscured here). You need this log, not a synthetic which is part of the geophysicist's job. All wells on your seismic need their gamma logs visible and you need to get a good feel for the seismic character of each unit.
Figures 18 and 19 show some stratigraphy with different phase rotations applied.
Occasionally the data 'prefers' to be picked at +90 degrees phase instead of normal zero phase (this may be considered as "poor man's" acoustic impedance). By 'prefer' I mean that geometries may be more easily seen by realigning rock boundaries with zero crossings rather than peak-troughs.
You will see some differences on these figures but it is not substantial so that job was done at zero phase.
Figures 20 and 21 (Neocomian deltas and fans from the Barrow Sub-basin) show a variety of facies - uninterpreted and interpreted (and calibrated by Novarra-1). The two legends are observational and interpretive. Get in to the habit of keeping your interpretation away from your observation for as long as possible.
Your observations are 'truth' but your interpretation may not be!
Figure 22 shows the legend that I use for facies mapping. Note that this has to be different from your lithofacies legend as no single seismic facies is unique in interpretation. High amplitude and continuity sheets can be deltas,
deep-water shales and silts, bedded basalts, platform carbonates, braided streams etc: - look for a calibrated association of facies - as mentioned before.
Figure 23 shows one way of trying to use the A-B/C method (described in Part 1 of the article) on the machine. Set up a facies horizon - e.g. sequence 5 SFM and then pick in a stair step fashion, as shown, the horizon (manual picking mode) so that you end up with a series of single colour swathes on the map. I jiggle the times that I pick so that I get a set range of colours and then just join up the colour swathes with a polygon on the map. I am not actually picking anything - only getting colours on the map, e.g. I have designated orange on the map to be onlapping and I pick a time that gives me orange - very simple.
How to get directionality though? Some facies mappers use faults and have the throw direction tick as the direction term for onlaps or downlaps. I put a bright colour, such as hot pink, on the end of my on or down-lapping pick so that on the map all the short pieces of pink are direction 'arrows'.
You will easily invent things that work for you when you go operational with this.
Figure 24 shows another trick for making the map. You usually need a separate indicator for the 'shelf edge' (or 'offlap break') as it always falls in the middle of a facies term as its downlapping extends past the shelf edge out in to the basin.
The shelf edge polygon that you end up with is useful for risking in that any play in this sequence landward of the line has a lower risk of good (= deltaic) reservoirs, while seaward reservoir risk will be greater as we are in 'fan territory'.
On recent jobs I have started picking the shelf edge as a 'continuous' horizon over the entire section (using several horizons as they double back over themselves).
Figure 25 is a facies map (from Angolan turbidites) made on 3D at a spacing of some 20 traces. Scale is no issue.
Indeed, I have gone in to a new basin and done facies mapping on three scales. The first was 1: 100,000 over the whole basin at a coarse 2D line spacing and that identified corridors of interest. Then the corridors were mapped at 1:50,000 scale on a tighter 2D grid and finally when we were at prospect level and had the 3D seismic surveys in the same was repeated at 1:25,000 scale on the 3D.
This is the stage that 'automated' facies mapping should be started - only after the 'manual' mapping has been done when you have a reasonable idea of what is going on geologically - rather than the usual 'random' automated work done to see 'what shows up'.
Figure 26 (from Prather et al, 1998, 2000) shows the Shell facies scheme for use in the deepwater Gulf of Mexico. I will leave you to chase the reference up (see next article for references) but basically it is an amplitudecontinuity scheme, as there are few good geometries in the deep water GOM. These authors do not apply this scheme outside of deep water.
Figure 27 (from the Mobil scheme from a few years ago) uses three letters as shown. The first letter is amplitude, the second is continuity and the third is geometry. These terms were 'hard wired' into a formal colour palette - Figure 28 - and picks were made that resulted in these colours on the map a bit like the scheme I described earlier.
Figure 29 shows another very simple method for mapping (Kingdom format). The sequence boundary candidates have been named from older to younger, Reg100, Reg200, etc. A sequence then is, for example, Reg_70_65 and we set up a facies mapping horizon to map, for example, channels in that sequence. Then just pick the base of the channel and let the colour swathes build up on the map (Figure 30). When completed draw a polygon around them and name it sensibly. The suite of resulting polygons (Figure 29) will then become the palaeogeography map - with suitable well calibration.
You will note that I have used faults in this case, instead of horizons, because Kingdom has a bug that they do not seem to be fixing whereby on 3D surveys you cannot readily see your horizon picks on the map as they are too thin. Faults, however, do show up on the base map so I do my facies work using them.
You will encounter issues like this as our software companies have not been pushed to develop better manual seismic facies mapping tools. We have great structural tools but poor tools for geology. Simple onlaps etc. would be nice to have.
Figure 31 is a simple facies map made in the days of discovering Jabiru and Challis in the Timor Sea (when we had to book time on the single Landmark machine and worked into the wee small hours).
Note the onlap, and downlap arrows and the well logs. A key in making these maps is to include on the map the relevant log window of just the sequence you are mapping.
This is usually coloured with a formal lithology scheme - such as Oxy's colours of orange for fluvial, gold for deltas, brown for turbidites. These windows help you visualise what is happening to your geology. Note that this seismic facies mapping goes hand in hand with a significant amount of well sequence stratigraphic work - which rides on the back of what the biostratigraphers have done.
Figure 32 shows some facies mapping of fan sequences in the deep water Gulf of Mexico.
Figures 33 and 34 are facies maps from Devonian carbonates in the Canning Basin of northwest Australia. The unique aspect of exploring here (along with having quite a bit of well data for a change) is that just 50 km away are the coeval rocks in a magnificent, undeformed outcrop.
The facies mapping done for this work (in conjunction with Mobil and Rick Sarg) required a trip to the field where we calibrated the subsurface work and were able to de-risk new plays that were developed.
Figure 33 shows all the systems tracts from an entire carbonate sequence. There were fluvial incised valleys, basinal 'fans' and prograding carbonate platforms and aggrading reefs.
Figure 34 was an attempt at drawing the palaeogeography at lowstand time - using half of the elements from the facies map.
Figure 35 shows a pair of facies maps made prior to, and leading up to, the discovery of the Harriet group of turbidite fields by Occidental (Kirk, 1985).
The first map shows the highstand of the last Barrow sequence but with a large canyon cut in to the delta plain.
The latter was the feeder for the subsequent turbidites- mapped on the right figure. These fans are the lowstand units of the Zeepaard sequence, as defined by Peter Arditto.
In the next article, Part 3, we will look at some examples of 'automated' seismic facies mapping.
N.B. Any resolution issues are due to the quality of the author's images.
