WELL
LOG
NORMALIZATION AND COMPARATIVE VOLUMETRIC ANALYSES OF GAS HYDRATE AND FREE-GAS RESOURCES, CENTRAL NORTH SLOPE, ALASKA
Scott Geauner1, Justin Manuel1, Robert R.
Casavant1, Charles E. Glass1, and Ken Mallon2
1 Department of Mining and Geological Engineering, The University of Arizona, Tucson, AZ 85721
2 Petroleum Consultant, Houston, TX 77005
This presentation introduces some of the challenges and uncertainties that may be encountered during petrophysical and geological interpretation, and volumetric estimations of shallow gas hydrate and free-gas resources. Our area of interest (AOI) encompasses all of the Milne Point Unit (MPU), a large portion of the Kuparuk River Unit (KRU), and the western quarter of the Prudhoe Bay Unit (PBU) on the North Slope of Alaska. The analysis
includes
well
log
data from 90 wells across the AOI and a 3-D seismic volume over MPU. The major focus of our research is a comprehensive geological and geophysical reservoir characterization of the heterogeneous Sagavanirktok formation. Assessments of in-place volume and distribution of unconventional gas hydrate and associated free-gas were made within this formation. Elements of our research include regional
well
-
log
lithostratigraphic and chronostratigraphic correlation,
log
normalization, determination of reservoir sand units, and comparative assessment of in-place volumes of free-gas and gas hydrate within and below the ice and gas hydrate stability fields. Regional
well
log
correlations show that the Tertiary-age Sagavanirktok formation consists of several thick sequences of stacked fluvial, deltaic, and nearshore marine sands and interbedded terrestrial and marine shales. Geologic cross-sections and petrophysical analyses reveal significant lateral and vertical variation in reservoir continuity and quality. This is related to rapid structural-stratigraphic changes, and the presence of intraformational unconformities. Variations in the distribution and quality of the available
well
log
and seismic data also influence the accuracy of our reservoir description.
Normalization
Early correlation work and petrophysical reviews of available well
log
data in the AOI revealed enough variation in quality and
log
response to warrant a normalization of key
log
curves such as the gamma ray
log
(GR). This need was supported by observed inconsistencies in GR readings between several pairs of closely-spaced wells, wells that were logged by the same contractor within a reasonable timeframe. The GR was an appropriated first candidate for normalization because of its widespread availability (open-hole and cased-hole wells) and its role in characterizing facies types, assessing lateral continuity of reservoir sands, geological
analysis
and mapping, and determining net pay. GR normalization is relevant to all geologic and geophysical characterization activities. When additional core-
log
porosity-permeability relationships are developed, and fluid types better constrained, a rigorous normalization of key porosity logs will also be undertaken.
Log
normalization schemes typically work best when petrophysical and statistical analyses for curve shifting are "trained on" or derived from a relatively "homogeneous" geologic unit(s). Ideally, the unit should be correlative throughout the study area, as lithologically consistent as possible, and proximal in depth to the reservoir interval of interest (e.g. to minimize affects of
log
drift). Examples of candidate intervals include moderately thick evaporite sequences, dense subtidal or basinal marine limestone intervals, zones of relatively "clean" marine shale (low quartz component, platy), and volcanic ash units. Normalization based on a heterogeneous, sand-rich unit for instance would not be suitable.
Regional correlation of stratigraphic sequences above the Hue Shale reveal the presence of a widespread, relatively thick and lithologically consistent marine shale sequence
within the gas hydrate-prone Sagavanirktok
formation1. Quick-look assessment and isopach mapping of GR values in this shale-siltstone interval suggest that it serves as an adequate candidate for
log
normalization across the study area. Of the 90 wells available to our study, only 64 had sufficient
log
coverage over the marine shale
sequence
. GR values were averaged across the interval in each
well
and gridded using a nearest neighbor least squared algorithm. Net-gross was also determined for the interval. Net-gross contour maps provided a quick-look evaluation of regional variations in the geological "character" or sandiness of individual
sequence
intervals. The maps also helped to identify subregions of GR likeness and areas of abrupt change in lithology. Special vigilance is necessary in areas having closely-spaced
well
pairs. An abrupt and significant variation in GR values might be related to either
log
quality issues or geologic factors such as faulting and provide guidance as to whether or not the
well
should be included in the
analysis
.
In essence, pre-normalization GR mapping reveals trends and/or significant geological controls that might allay or complicate log
normalization. Following the mapping a histogram of
well
GR averages was constructed. The resulting normal distribution reflected (as the maps had) what was essentially a geologically consistent zone. Given little variation in the service company that logged the wells, and the date of the logging, the observed statistical randomness or noise was attributed to variation in tool response and borehole environments, and/or to local geologic factors such as faults. These results of the
analysis
lent support to moving forward with
log
normalization.
With the "area GR mean" established from the samples, curves in individual wells were bulk shifted to approximate that mean. While significant shifting was required in a few wells, the shifts in most of the wells were negligible. It is interesting to note though that many wells in the MPU area in which gas hydrates had been inferred by this and other studies, required a minimum amount of normalization. The significance of this observation is the object of additional study. A relationship to facies and shale ratios is suspected. Most of the statistical outliers related to wells located to the east and southeast of the MPU area. In only a few wells, could we assign the need for a moderate shift to a subdued GR response in a cased-hole log
. Further evaluation of the statistical variation included and excluded the outliers. The
analysis
revealed that the outliers had a negligible affect on the overall sample mean. Thus, outlier wells were also normalized at the time, but were not used later in mapping net sand for volumetric
analysis
.
Determination of Reservoir Sand
Following the GR normalization, a sand/shale cutoff for distinguishing potential reservoir sandstone for mapping was determined. Sand and shale baselines were defined on a per well
basis and a number of different GR cutoffs were examined in a representative sample of wells to determine an optimal GR cutoff for the AOI. Statistical and
well
analysis
showed that a 55 API GR cutoff was best at distinguishing between reservoir and non-reservoir sands in both the better quality normalized cased-hole and open-hole logs. The cutoff was successfully tested against wells that had been withheld in the initial
analysis
. With the reservoir sand units now identified, a series of gross isopach maps., net isopach maps, and sand/shale ratio maps were produced for all lithostratigraphic and later, chronostratigraphic units within the Sagavanirktok formation. Regional structural and stratigraphic trends were noted. These maps as
well
as net-gross relationships across the AOI were quantified and used in seismic attribute analyses. The geologic maps also helped define the limits of reservoir rock used in our volumetric calculations. Details of geologic map interpretations are explored further in the presentation.
Comparative Volumetric Analysis
Following determination of net sand intervals, an average and range of porosity values within each chronostratigraphic sequence
were determined.
Well
log
responses for resistivity, acoustic transit time, bulk density and neutron porosity were evaluated down to a depth of 4000 ft., which is
well
below the Sagavanirktok formation. Intervals of gas hydrate and associated free-gas were first manually determined. A comparison of UA's determinations with those from earlier studies
1 was completed. The
analysis
confirmed that below the base of the ice-bearing permafrost (IBPF),
log
response for gas hydrate generally included an increase in resistivity and an associated decrease in acoustic transit time relative to water-bearing reservoir intervals. In some wells we noted that bulk density and compensated neutron
log
curves appeared to begin converging, but did not crossover as is common in deeper gas zones within the AOI. This "behavior" might be reflecting a certain amount of continuing destabilization associated with the penetration of a gas hydrate-bearing sand. Below the base of the gas hydrate stability field, in-situ pressure and temperature conditions are not sufficient for gas hydrate formation
1 in the presence of free gas and water. Indications of free-gas units in the MPU were commonly marked by a similar increase in resistivity, crossover of bulk density/neutron porosity or noticeable merging (influenced by thin bed phenomenon?), and an increase in acoustic transit time. For all interpreted gas hydrate- and gas-bearing zones, a maximum and conservative thickness of total net pay was summed in each
sequence
for mapping and volumetric analyses.
In an effort to automate fluid prediction, a preliminary log
-based algorithm for estimating different fluid types (e.g. ice, gas hydrate, gas) and their probability of occurrence was developed. Fluid type determinations worked best below the IBPF
2. They were derived from resistivity, acoustic and density
log
measurements. With the algorithm we were able to automatically calculate fluid type thickness at varying probabilities across reservoir sand intervals. A preliminary comparison of cutoffs with previous "quick-look"
log
analyses indicated that a 25% chance of occurrence (defined by the algorithm) equated best to gas hydrate and gas-bearing zones. Using this probability cutoff, additional net pay maps and reservoir rock volumes were derived for each chronostratigraphic unit. Using average porosities for reservoir sands in each unit, preliminary volumetric calculations were completed. Gas hydrate and gas expansion factors were taken from previous
studies1 to complete a comparative volumetric assessment across both the AOI and within the Milne Point Unit. Volumes estimates from earlier
studies1 were then compared to volumes derived from our seismic mapping, the algorithm discussed above, and manual
well
-
log
interpretations. The results of this comparative study demonstrated the range and uncertainty in calculating volumetrics across the AOI and will be addressed in the presentation.
Future work related to continuing volumetric assessment includes a detailed analysis
of
well
-
log
and integration with seismic-based reservoir facies and fluids. In conjunction with structural analyses (e.g. faulting), the identification and mapping of net pay in discrete sand bodies should improve our modeling of resource quality, quantity, distribution, and continuity. This work will be important to refining volume estimates, reservoir models, and forecasting recovery factors and production.
Acknowledgements and Disclaimer:
The University of Arizona contribution is part of a larger collaborative program that includes researchers from the University of Alaska Fairbanks and the U.S. Geological Survey. BP Exploration (Alaska), Inc. provides overall project coordination and provided data for the characterization and modeling efforts. Interpretation and processing software was made available through support from the University Grants Program of Landmark Graphics Corporation and from GeoPlus Corporation. This research was funded by the Department of Energy (Award # DE-FC-01NT41332). The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
References:
1. Collett, T. S., K. J. Bird, K. A. Kvenvolden, and L. B. Magoon, 1988, Geologic interrelations relative to gas hydrates within the North Slope of Alaska: United States Geological Survey Open-File Report, v. 88-389, n. 150) (The Eocene-age shale unit is equivalent to USGS correlation unit 16-17 and the UA unit 36-36a.)
2. Glass, C. E. 2003, Estimating pore fluid concentrations using acoustic and electrical log
attributes, Interim Report, UA Gas Hydrate Project.