When Forest Oil International acquired
a large block in the Orange Basin offshore South Africa from Anschutz
in 1998 it was nothing more than a big acreage play with a couple
of intriguing hints.
Those hints — coupled with state-of-the art 3-D
seismic technology — ultimately led to a major gas discovery and
the potential for several more.
"Thirteen wells had been drilled on the block by
Soekor (the South African state company) in the mid-1980s when the
firm launched an aggressive campaign to try and find internal resources,"
said Tim Berge, chief geophysicist with Forest's international group.
"All the wells were drilled on the basis of 2-D seismic into what
appeared to be structural closures."
Although a couple of those wells had shows, no commercial
accumulations were encountered.
But one well in particular, the A-K1, intrigued Berge
when he began mapping the block.
"That well had a combined test of 57 million cubic
feet of gas a day out of three sands," he said. "That's a pretty
good test. But, when I mapped the 2-D data I couldn't get the structure
to close — there was no way to account for the trapping of gas
based on the structural model."
Forest had a commitment to shoot 1,200 kilometers
of 2-D data in the block, but Berge argued internally that the A-K1
well was likely a stratigraphic trap and 3-D seismic would better
image the block's stratigraphic nature.
"You could see on the log character that the sands
were likely thick enough to be seismically resolvable," Berge said,
"and they had properties that we felt would lend themselves to direct
detection and AVO work if we had the right kind of data."
So Forest negotiated with the government to change
its commitment to 3-D data and in 2000 shot a 312-square-kilometer
survey in the A-K1 well area.
"When we got the 3-D data we could see all these
channels and bright spots," Berge said. "The area turned out to
be much more complicated than the structural model with a blanket
sand that had been developed from 2-D seismic."
3-D and Inversion Techniques
Forest, committed to drilling a well by the end of
2001, got busy working the 3-D data — and Berge said application
of 3-D seismic and inversion techniques were the technical keys
that made this play work.
"One of the first things we did was forward modeling
so we could understand, based on the one well we had, what the predicted
AVO response and stacked response would be from the productive sands,"
Berge said. "Once we had that in hand we started trying different
inversion techniques."
An intercept gradient was tried first.
"It was fairly easy to calculate," Berge said, "and
we found that we had a negative near trace response and a negative
AVO gradient, or a 'class three' AVO."
Forest then worked the seismic data with an elastic
inversion technique that uses pre-stacked time migrated data to
extract compression wave vector and an estimated shear wave vector,
which are then cross-plotted to determine lithology, porosity and
fluids.
"This method seemed to correlate best with the penetrated
anomaly, and tied best to my one point of control in the older A-K1
well," Berge said. "Based on this data we did a volumetric map,
which we based our first four well locations on."
The method turned out to be "very accurate" in predicting
reservoir, he added, and "fairly accurate" in predicting gas content.
"Inversion work was used to predict exactly where
and how thick the sands would be in the drilling program," he said.
"Depth estimates came right from that volume as well."
Forest also measured the in-place reserves from the
volume of anomalies.
"We used the volume not only for predicting and planning
the well campaign but also for estimating reserves from the field
as drilling progressed," he said. "We wanted to test the largest
reservoir compartments with the highest predicted porosity and lowest
water saturations first, and that was borne out by the drilling
program.
"Every inversion attempt, especially if it's independent
of other approaches, offers some risk reduction," he concluded.
"Every time you try something different you are looking at a somewhat
different part of the dataset, or looking at it in a different way
— and that has potential for giving you additional insight into
the real rock properties."
Test Time
The four-well drilling program was set to evaluate
the field and prove up a core area with enough reserves to be economically
developed. Wells tested individual compartments containing 28 to
520 billion cubic feet of gas for a total of 1.15 trillion cubic
feet.
- The first well, the A-K2, tested 30 million
cubic feet of gas and more than 600 barrels of condensate a day
from a 20-meter pay sand, according to Berge. Reservoir characteristics
were better than expected; the sands were clean and well sorted
with average porosity of 21 percent and almost no water saturation.
No water was produced and no significant
reservoir pressure draw-down was seen during the 12-hour test.
- The second well had a 15-meter gas bearing
sand of similar quality as the older A-K1 well.
Notably, the lowest gas sand in the well
is deeper than the lowest proven gas and highest proven water
in A-K1, clearly showing that this is a separate reservoir and
stratigraphic trap, he said.
- The third well targeted the largest and
brightest anomaly in the data set. It found two thick and porous
sands as predicted, but they contained low gas saturation water.
Additional elastic inversion showed that
these sands had less rigidity than others in the area. This
factor, combined with high porosity, accounts for its high values
in the elastic cross plot volume, according to Berge.
- The fourth well tested a feature that appeared
to be a preserved cut-off meander loop. The well tested 71 million
cubic feet of gas and 1,340 barrels of condensate daily from combined
tests of the upper two zones.
This well achieved the highest gas test
rate ever achieved in South Africa, he said.
"This drilling program uncovered a giant regional
stratigraphic trap and discovered the Ibhubesi commercial gas field,"
Berge said.
Ibhubesi produces from the Albian-Aptian, and internal
estimates for the proven, probable and possible reserves is one
trillion cubic feet of gas.
"The 3-D seismic was key," Berge said.
"We were completely successful in predicting the
presence of high reservoir quality sands on 10 occasions in five
wells," he said. "We predicted commercial gas content eight times
for a success rate of 80 percent; porosity predictions were always
within two PUs of the target interval; and thicknesses ranged from
about 30 percent less than predicted to about 30 percent more than
predicted."
Even the wet well was important to the overall drilling
campaign; after drilling it, Forest scientists went back and did
some revision and re-calibration of the inversion volumes.
"In fact, it was probably a good thing that we had
a wet well," Berge said, "because I feel we have a very good calibration
point now."
The Ibhubesi Field is a fluvial incised valley complex
with excellent reservoir properties:
- Porosity averages about 20 percent.
- Permeability ranges from 300 to 400 millidarcies.
- The field boasts high deliverability wells
with reserves of 50 to 300 billion cubic feet per channel.
- Application of 3-D seismic technology provides
excellent reservoir imaging and delineation, which allows for
efficient low risk field development and low finding and development
costs.
"I have been able to quantify the importance of 3-D
seismic and inversion methods," Berge said. "The mid-'80s drilling
campaign based on 2-D seismic included 14 wells — and being fairly
generous, three of that total were discoveries, for an overall success
rate of 21 percent," he said. "We were successful on three out of
four wells based on the 3-D seismic, or 75 percent. That translates
to a risk reduction attributable to the 3-D and inversion work of
about 54 percent."
"Based on our well costs and an estimated price of
gas of $1.50 per thousand cubic feet, the difference between the
two programs would be $15.2 million in dry hole savings and $216
million of added value from additional discovered reserves," he
said.
"All that for 3-D seismic acquisition and processing
costs that totaled $5 million."