No matter how mature the field, geoscientists
seem to always come up with either a new technology or another twist
on the tried-and-true to pull more hydrocarbons out of the reservoirs.
When commodity prices are looking good, producers have an added
incentive to use the technology to go after deeper targets and also
explore for smaller ones in producing trends that historically have
been exploited at shallow depths.
The heavily drilled Oligocene Vicksburg sands in
south Texas are a prime example of a mature area where moderate-potential
fault block drilling targets in productive zones and higher-potential
targets in deeper, untested intervals offer the lure for continued
drilling.
Over the trend's 80-year productive life, more than
three TCF of gas and 100 million barrels of oil have been produced
in Starr and Hidalgo counties. A large non-exclusive 3-D seismic
survey by WesternGeco (nee Western Geophysical) beginning in 1994
spurred a whole new round of interest in pursuing the hydrocarbons
that likely remain.
Edge Petroleum and Carrizo Oil and Gas seized the
opportunity to better explore in the area when they licensed 450-square-miles
of the 3-D survey in 1995.
The ensuing drilling program using 3-D imaging to
pinpoint untested pockets in productive zones between 5,500 and
7,500 feet deep yielded a disappointing 20 percent success ratio.
"Two of the wells were only a few miles apart
and had been determined to have analogous stratigraphy, structure,
timing of trap formation and source proximity," said John Hastings,
vice-president of exploration at Edge. "But even though both
reservoirs proved to have the predicted facies, only one well had
a commercial gas accumulation."
Undaunted, the exploration team returned to the drawing
board to determine how to better risk future drilling prospects.
The One-Two Punch
Subsurface geology along with structural and stratigraphic
interpretation of seismic data historically have been the tools
used to explore the Vicksburg. There is a dearth of bright spot
examples because the trend is not an amplitude-supported play.
But Hastings and his colleagues soon found that a
novel seismic processing technique combined with close attention to rock properties provided the one-two
punch needed to beat the odds in their drilling program. Indeed,
the resulting activity yielded six commercial discoveries (including
two stratigraphic traps) and two dry holes for a 75 percent success
rate.
The turnabout from the early drilling disappointment
began when the group acquired comprehensive log suites that included
dipole sonic data in order to better understand rock properties
of the target zones in the two initial wells.
To predict AVO behavior, synthetic common reflection
point, or common depth point (CDP) gathers were modeled using log
data. Small contrasts in acoustic impedance between both the gas-bearing
and wet sands and the encompassing shales were evident. Poisson's
ratio - a measure of a rock's rigidity - in the gas sand was observed
to be significantly lower, however, than in either the encasing
shales or the wet sands, Hastings noted.
On the modeled gathers,
the small acoustic impedance contrast at the top of the gas reservoir
resulted in a weak reflection at near offsets, while the strongly
negative Poisson's ratio contrast yielded a strong negative reflection
at far offsets - a Class 2 AVO anomaly, according to exploration
team member Mark Gregg, formerly with Edge.
The onset of the anomaly is at an offset equal to
reservoir depth, or an incident angle of about 26 degrees. Positive
reflections at near-offsets exhibited at the tops of the water sands
weaken with offset.
"Because of the modeling study, we were motivated
to undertake a pilot prestack reprocessing project to test the hypothesis
that Vicksburg gas fields exhibit Class 2 AVO anomalies," Hastings
said.
"The reprocessing technique that was used yielded
useable data at incident angles more than 40 degrees to enhance
observation of the targeted anomalies."
Seeing Something Different
In the two early wells, which were used as test sites,
reprocessed CDP gathers with non-hyperbolic moveout revealed a distinct
difference in AVO character between the wet sands and the gas-bearing
sands. The gas zone produces a clearly defined Class 2 AVO anomaly,
but the wet sands do not.
The pronounced far-offset reflectivity that characterizes
this type anomaly is realized only at incident angles greater than
26 degrees or so. Hastings noted the most highly developed part
of the anomaly would be muted on a stack processed with conventional
normal moveout.
Angle stacks through the two wells were used as a
tool to identify and assess the weak near-offset and strong far-offset
reflections of the targeted Class 2 anomalies. This evaluation provided
the impetus to reprocess a larger part of the 3-D survey.
Near-angle and far-anglestacks in the reprocessed volume were compared, providing invaluable
insight into the Vicksburg trend, according to Hastings:
- In the study area, roughly half of the 100 or so Vicksburg
gas wells with cumulative production of more than one BCF were
associated with Class 2 AVO anomalies.
- About 65 percent of the approximately 70 anomalies drilled
that appeared to be geologically valid targets had commercial
gas accumulations.
- Thicker, better developed reservoirs had the most distinctive
AVO anomalies.
- Most of the productive anomalies occurred at depths between
5,000 and 10,000 feet.
The exploration team members concluded that near-
and far-angle stacks appeared to be a valid exploration tool to
identify prospective Class 2 AVO anomalies in the Vicksburg, with
an anticipated success rate of 65 percent.
They proceeded to conduct reconnaissance exploration
within the reprocessed data by visualizing anomalies in the far-angle
stack data set.
Anomalies visually "popped out" of the
data, according to Gregg. He said they catalogued known gas reservoirs
as productive analogs and readily identified untested anomalies
as prospective targets.
The initial well - drilled on the basis of information
derived from the reprocessing exercise - had both stratigraphic
and structural trapping components, adding an element of risk over
a relatively simple structural trap.
Drilling commenced based on the strength of the AVO
anomaly, and the bit encountered a 100-foot gross interval with
72 feet of net pay at the anomaly. Initial production was three
MMCFD. The conventional NMO stack showed no evidence of an anomaly.
"If we hadn't had the AVO tool," Gregg
said, "the prospect most likely would have been missed because
of its stratigraphic nature."
'Dramatic Improvement'
A couple of the successful AVO-based wells involved
a small upthrown fault trap associated with two Class 2 AVO anomalies.
"During the pre-AVO program, the small trap
wasn't deemed prospective because it was unattractive economically,
with a 20 percent probability of success," said Charles Bukowski,
formerly with Edge. "Using the statistical AVO success rate
of 65 percent, Edge decided to drill, and the combined EUR for the
two producers is 4.5 BCF."
One of the unsuccessful AVO attempts was drilled
where a Class 2 AVO anomaly was interpreted as a stratigraphically-trapped
gas reservoir. With no well control to validate the anomaly, the
prospect was drilled based on the anomaly, and the bit encountered
105 feet of clean, low-gas saturation sand.
Hastings noted this dry hole highlights two pitfalls
of the AVO method:
- Anomalies can be caused by reservoirs having either commercial
or non-commercial gas deposits.
- Great-looking anomalies can be tempting to drill even though
they don't really satisfy criteria for conventional prospect evaluation.
Still, the positives vastly outweigh the negatives.
"The 75 percent success rate we realized using
the AVO method is similar to the 65 percent rate predicted by statistical
analysis," Gregg said, "and it represents a dramatic improvement
over the 20 percent we achieved with conventional subsurface and
structural mapping evaluation."