The purpose of this post is to provide a deeper insight into the potential OIP in the Jurassic and to explain why the Jurassic alone, never mind the Triassic, has a significant probability of containing well in excess of 10bn bbl, i.e. well above the DGA P10 and P1 estimates. The post is quite long but is worth your time.
There has been much discussion (and misunderstanding) about the differences in OIP estimates for Shaikan shown in DGA’s Report 1 (Preliminary Jurassic Evaluation) and those quoted by TK at the Nov 30th presentation to analysts. The DGA report presents a mean OIP of 2.79bn bbl for the Jurassic (based on well results to 2055m MD) whilst TK said, somewhat confusingly, that there was a possible 10-15bn bbl in place. Some posters on this board, and possibly the city analysts also, have assumed TK’s figures must be wildly optimistic since they exceed the P10 and even P1 estimates for the Jurassic. Others have concluded that his comments appear to have taken into account the Triassic and an analysis on this section is awaited in DGA Report 2.
I have built a Monte Carlo model of the two Jurassic Sections (Barsarin thro’ Mus and Butmah reservoirs) using the input data provided in Table 4 of DGA Report 1. The model was used to understand why the report produced a modest OIP (only in Iraq is 2.8bn modest!), given the huge net pay figures and very substantial size of the structure.
Reproduced below are the key OIP resource estimates in MM bbl. The first number is that from DGA report, the second number is from the model.
Mean: 2790, 2723
P90, 956, 744 (think proven)
P50: 2217, 2043 (think probable)
P10: 5266, 5183 (think possible)
Note the excellent agreement with the P50 estimate and the Mean, the latter being the most important of all the figures. The small divergence at the P90 level is to be expected since the two models probably handle the end point distributions slightly differently and in any event the P10 and P90 distributions are not sampled so frequently giving rise to more statistical variation between model runs.
Having established the model’s ability to replicate the DGA model, it was used to answer the key question that a GKP employee should have asked when presented with this report. “What does the OIP distribution look like if the structure is filled to the closing contour of 2200m SS, ref Fig4 DGA report?”
This was simulated by setting the reservoir area equal to that within the 2000m SS contour (approx 140 sq km) and keeping the other parameters unchanged. This discounts the roughly 20 sq km at the periphery where the base of the Butmah dips below the closing contour. In this area there is a reduction of the net pay above the oil water contact (wedge effect). Pay reduces from the DGA input figures of 167m at 2000m SS down to 0m at the spill point. The model is conservative in this respect.
Key results were as follows:
Mean: 13.8 bn
P90, 8.8 bn
P50: 13.3 bn
P10: 19.4 bn
How, you might ask, is it possible for this Mean to exceed the P1 from the DGA report and be credible?
It is all down to the key assumptions regarding area of the reservoir which has by far the widest range of any parameter in the model. For example DGA use a P90 of 6.5 sq km for the top four Jurassic reservoirs which they say is 25% of the mapped area of lowest known oil (LKO). In other words the well has found oil down to a certain depth but their model doesn’t give the area associated with this a 90% probability of being sampled when the model calculates OIP. In fact their P50 area of 20.6 sq km is less than the area associated with the LKO in the 4th reservoir unit (Mus).
Consider how their upside areas have been handled. The full structure area within closure is sampled only 1 in 100 times and only 10 times in 100 samples is the oil filled reservoir area considered to be larger than 1/3 of the total area under closure. Also the lognormal distribution of the area parameter results in the modes (most frequently sampled areas in the models) being less than 20 sq km in the Barsarin thro’ Mus and Butmah models.
Note durby was making a similar point about the area assumptions in his post of 2/12/09 to Hub when discussing distance from Shaikan 1 to site of Shaikan 2. My models quantify the impact of the area assumptions.
DGA’s approach is understandable. They don’t want to stick their neck out and say there is 10-15 bn bbl in place after just one well and then find for some reason we don’t yet understand that it doesn’t work out. The 2.8bn is a fair estimate of the Mean OIP based on a one well result. But what the report doesn’t do is show the full potential of the Shaikan structure. This is because the DGA report does not make reference to important pressure information obtained from the well tests and the offset well that was shown in the November presentation to the analysts. It is possible that this information was not provided to DGA prior to release of their preliminary evaluation. The DGA Report is dated Oct 21st, 2009.
The information is contained in Slide 15 of the presentation to the analysts. It shows the oil column pressure gradient obtained from DSTs 1-3 and the local aquifer pressure gradient in the Jebel Kand 1 well which lies about 20km away in the Hunt oil block. In the situation where the oil water contact has not been encountered in a well, the industry has commonly used pressure information from the aquifer in adjacent wells to infer the depth at which the OWC is downdip. The intersection of the well’s oil gradient with the local/regional aquifer pressure gradient can predict the depth of the Free Water Level which in most reservoirs of reasonable permeability is close to the oil water contact. (Let’s not get into the differences here.) GKP are well aware of this and used the intersection to predict that the Free Water Level/OWC is 2230m SS.
Slide 15 is indicating that the structure contains oil down to, or close to 2230mSS. Does the data look reasonable or anomalous? Well, based on the following it does look perfectly plausible.
1. The aquifer gradient of 0.44 psi/ft corresponds to that of a fresh water/low salinity aquifer.
2. The oil gradient of 0.359 psi/ft is about that which would be expected of a medium to low gravity crude. From experience it is not obviously too high, say 0.4 psi/ft, neither is it obviously too low, say 0.3 psi/ft.
3. Correlation of fit (R value) of the straight line for both gradients is exceedingly good, i.e. R ~ 1.0.
4. Although there are only 3 points on the oil gradient line they span about 350m which is very large in comparison with most of the rest of the world’s fields. Often extrapolations are done on lines created from more pressure points but with a smaller vertical separation range.
5.The intercept of 2230 mSS may also be suggesting the data is valid. It is very close to the closing contour of 2200 mSS, implying that the structure is filled to closure. The difference of 30 m may be the result of having to extrapolate the oil gradient line 1200 m downdip from the lowermost test pressure and the depth errors inherent in seismic interpretation. Alternatively the oil column may extend below the closing contour indicating the Shaikan structure is one culmination of a larger oil accumulation as has been discussed. Note the Kirkuk Field has 3 separate culminations within one super giant accumulation.
What could cause the OWC to be shallower?
1. If the JK well and Shaikan well are not in pressure communication it is possible that the Shaikan aquifer pressure line is displaced slightly to the right, i.e. has higher pressure at the same depth locally in the Shaikan area. The structure would not be full. Unfortunately GKP do not present any background material to support the use of the JK-1 pressures.
2. If the three Shaikan test intervals are not in pressure communication, or the pressures have not been interpreted correctly, and each interval is one point on three separate oil gradient lines each of steeper slope (e.g. say about 0.33 psi/ft) then the OWC would be shallower. However I believe this is unlikely. Given the low GOR reported if any change is warranted I would expect the gradient to be marginally less steep and hence the contact to be deeper.
My view is that Slide 15 is telling us that the structure is filled to at least the 2200m contour and that there therefore a significant probability that the Jurassic contains 10 – 15 bn OIP. This is of course not guaranteed. For example even if the OWC is at 2200 mSS or deeper the net to gross could deteriorate significantly down dip from the existing well leading to lower net pay thicknesses down dip.
In conclusion two things to consider:
1. When the next DGA report is released look out for whether they have revised their Jurassic interpretation in the light of the pressure information. If not then their revised estimate is still likely to contain considerable upside.
2. Make no mistake the Shaikan 2 well has the potential to prove up very substantial volumes of oil. It will have a bigger effect on OIP than the extended test in the short term. If GKP were to sell out before Shaikan 2 I would be disgusted. It has so much upside if located correctly to investigate the oil column sufficiently down dip.
Let me know if you have any questions.
All the best for 2010 GKP investors.
Hub thanks for your 2010 hot list.
Cheers,
Gramacho
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment