Amplitude Vs Offset (AVO) attributes were evaluated for discriminating fluid type and lithology using well-log curves from the northern and southern areas of offshore West Cameron, Gulf of Mexico. This dataset included approximately 600 well-log suites which were sampled into approximately 30,000 200-ft intervals that each contained 30 seismic attributes and environment properties. To achieve this lithology and pore-fluid interpretation, multiple rock-property cross plots were generated to establish depth trends above and below the onset of abnormal pressure (geopressure depth). Commonly applied reflectivity and layer attributes were successful in robustly discriminating lithology and pore-fluid type in the study area both above and below geopressure.
This project is focused on northern and southern blocks of offshore West Cameron in the Gulf of Mexico and this is a domain that has undergone extensive and complicated geological processes through its 165 million-year history. It has been the receiving basin for most part of the North American continent.
“Rapid burial of older, commonly muddy sediment caused buildup of fluid pressure within the thick basin fill. This geopressurization decreased mechanical strength of the sediment, facilitating structural deformation. It also generated strong pressure gradients that directed fluids up and out of the deep basin towards the shallow sand bodies of the basin margin” (William, 2009). The chronostratigraphic (time) lines become deeper from the shoreline to the shelf edge.
“In most exploration and reservoir seismic surveys, the main objectives are, first, to correctly image the structure in time and depth and, second, to correctly characterize the potential reservoirs based on the reflection amplitudes. If the amplitudes are accurately rendered, a host of additional features can be derived and used in interpretation. Collectively, these features are referred to as seismic attributes.” (PetroWiki, 2015)
The stacked seismic data volume volumes (both full range and partial angle stacks) are commonly used for interpretation of geologic structure and seismic attributes. Full-stack amplitude is a common seismic amplitude but it’s interpretation in thin-layered beds is not necessarily linear. Usually, amplitude has a strong correlation with porosity, liquid saturation (oil/water and/or gas) and water saturation as these reservoir properties have a strong effect on both velocity and density, because there are unconsolidated sediments in the study area, the effective pressure and the mechanism of any abnormal pressure are important attributes to consider.
In early investigations of seismic amplitude attributes, Castagna and Smith (1994) found, the reflection coefficient difference (Rp – Rs) to be a more universal indicator than the AVO product (Intercept A * Slope B) in clastic stratigraphic intervals, where Rp is then normal-incident P-wave reflectivity while Rs is the normal-incident S-wave reflectivity. Also, Amplitude Versus Offset (AVO) concepts can be used to generate weighted stacking schemes which can be used to display information about rock properties in standard seismic data (Smith and Gidlow, 1987). Numerous crossplots based on recent publications were generated to determine pore fluid, lithology and porosity indicators.
The project focused on establishing diagnostic trends for reservoir properties in the West Cameron region. The areas of study include TILE 2 AND TILE 19 (Geophysical Development Corporation nomenclature) seen in Figure A, within the West Cameron area of the Gulf of Mexico.
This project was divided into checking dataset assumptions and quality, using the dataset to calculate and derive further attributes and creating crossplots to evaluate AVO sensitivity trends. Initially, data were tested to better understand the assumptions that created the attributes used throughout the study, like observing average attribute trend over depth. Key objectives for this project were to determine AVO attribute trends with depth particular to this part of Gulf of Mexico to help in grouping data into Class I, II, III,or IV, which will help the interpreter to understand how to look for hydrocarbons in regional seismic data. By adding understanding to AVO analysis, companies can avert risk and additional costs when processing and interpreting seismic.
Project data were prepared by Dr. Fred Hilterman of Geokinetics. Using sonic, density, gamma and resistivity logs, an algorithm created by Hilterman derived shear wave velocities, seismic attributes (normal incidence, far offset gradients, acoustic impedance), and elastic rock properties (bulk modulus of the dry rock, shear modulus of the dry rock). These attributes were calculated for shale, water-wet sandstone, gas-saturated sandstone, fizz-saturated sandstone, and oil-saturated sandstone using the Batzle-Wang fluid substitution transforms. Overall, 33 attributes were calculated for every 200 ft interval, accompanied by associated mud weight and temperature readings. The dataset contains 16864 200ft intervals from the northern portion of offshore west Cameron and 12362 200-ft intervals from the southern portion which were derived from about 300 wells in each section, under normal pressure and overpressure.
The analysis was carried out on data above and below geopressure. Effective pressure (approximately overburden minus pore pressure) is a large influence on rock velocities (Hilterman, 2001), and therefore both pressure conditions were considered.
To add to the dataset provided, all calculated attributes were sorted into above and below geopressure. Additional calculations to derive acoustic impedance (AI), which is P-wave velocity times density (?VP) were carried out. Average attribute values across 1000-ft depth intervals were calculated and they were only considered statistically significant if more than 5 data points existed for each depth interval. Plots of the average properties were then plotted against depth to quality check the data and create empirical relationships.
EXCEL was the main tool used in my catalogue study, as it could manipulate the large dataset, present graphical figures, and capture statistical information. Developing diagnostic attribute trends was a primary objective. Numerous crossplots that will be further discussed in this paper analyzing attribute changes including the following;
1- Average Acoustic Impedance against Depth – Above and below geopressure.
2- Average P-wave velocity versus Depth – Above and below geopressure.
3- Average Density versus Depth – Above and below geopressure.
4- Average Normal Incidence versus against Depth – Above and below geopressure.
5- P-wave velocity versus S-wave Velocity.
6- Average Normal Incidence versus Depth – Above and below geopressure.
7- Average Normal Incidence versus average Slope – Above and below geopressure.
8- P-wave and S-wave ratio against Acoustic Impedance.
9- LamdhaRho versus MuRho
Some plots were used primarily to quality check the data, and others offer guidance for interpreters working on AVO studies.
Results of Study
Acoustic Impedance values at 200-ft depth intervals were averaged into 1000-ft depth intervals and were then plotted as a function of depth for the Northern (TILE2) and Southern (TILE19) areas. In each area the data were separated by the constraint of being above or below geopressure.
TILE2 data above geopressure indicates a gradually increasing average AI values for all lithologies (Figure 1). Gas, Fizz, and Oil sand AI trends lose separation with increasing depth. Shale and brine-saturated sand have similar trends at shallow depths and having greater separation with depth. Below geopressure data indicate no separation between Oil and Fizz sand trends across all depths (Figure 2). There is a good separation between Gas sands for both Fizz and brine-saturated sand. The shale trend becomes harder to separate from hydrocarbon bearing units with increasing depth (TILE2). Brine-saturated sand has the highest average AI trends.
TILE19 above geopressure shows clear separation between Gas, Fizz, Oil, brine¬-saturated sands, and shale across all depths (Figure 1). It becomes more complex to separate Oil and Fizz sand trends with depth in TILE19 above geopressure crossplot (Figure 1). TILE19 below geopressure values also indicates clear separation between Gas sand, Brine-saturated sand and Shale. While Fizz and Oil sand trends with depth becomes increasingly similar for most depths (Figure 2). For TILE2 below geopressure, there is a sharp decrease in AI trends at about 14,000′ while for TILE19 below geopressure, sharp increase in values can be observed at depths of 13,000′. The difference in Acoustic Impedance between rock layers affects reflection coefficient. This change could be related to differing pressure mechanism. The abnormal pressure mechanism in TILE2 is due to smectite to illite transformation while in TILE 19, the mechanism is disequilibrium compaction.
Plots were created evaluating P-wave velocity, density, and normal incidence over depth. In TILE19 average shale P-wave velocity, for above and below geopressure (Figures 3 ; 4), is approximately the same as the brine-saturated sandstone. In TILE2, the sands being slightly older, have experienced more diagenetic change and thus are faster both above and below geopressure (Figures 3 ; 4). For the density-depth crossplots in TILE2 and TILE19, (Figures 5 ; 6), the hydrocarbon zones have distinctly different densities, even at shallow depths and this also includes above and below geopressure. It’s observed that, the shale density is significantly greater than sand densities, regardless of fluid saturation.
The normal-incident crossplots for TILE2 and TILE19, above geopressure (Figure 7) , indicate the normal incidence values for oil are similar to fizz for depths below 6000ft and begin to approach brine-saturated sand NI at the shallower depths. Oil sand normal incidence approaches brine-saturated sand reflectivities at shallow depths because up shallow there is little to no gas in the oil and both should exhibit similar velocities. TILE2 and TILE19 normal-incident trends below geopressure, (Figure 8), show Oil and Fizz sand havings nearly identical trends across all depths.
A petrophysical interpretation was carried out with plots that differentiate lithology and pore fluid, a method proposed by Krief (1990). By plotting the square of P-wave velocity against the square of the S-wave velocity, linear trends emerge for each lithology and each pore fluid. By calculating the slope and y-intercept for known lithologies and fluid substitutions, the interpreter has values to differentiate lithologies and pore-fluids in more questionable zones. By plotting measured squared P-wave and S-wave data within a zone of interest, the intercept will line up to that of shale or one of the sand scenarios. For TILE2 (Figure 9), the graph shows good separation for all scenarios except oil and fizz saturated sands. To differentiate these would require further investigation. TILE19 (Figure 10) also shows good distinction between shale and brine-saturated sand. Again, oil and fizz saturated sands have similar trends. In both TILE2 and TILE19, the krief method will work better in shallower depths to distinguish between pore fluids, as there is little separation between scenarios at higher velocities. This method assumes that linear trends are due to porosity only and may be skewed if there are gradual changes in lithology or increases in shale content (Krief, 1990).
The second part of this project looked at using petrophysical data to guide seismic interpretation and AVO analysis. The first example of this was plotting normal incidence data against depth to identify what class fluids were expected throughout the area. It was assumed that Class II boundaries fell between 0.03 and -0.03, with Class I data greater than 0.03 and Class III/IV data less than -0.03 (Hilterman, 2017). Looking at TILE 2 trends (Figure11), for depths above geopressure, the area transitions from Class III to Class II around 9,000′ and from Class II to Class I around 13,000′. The core of Class II data lies at around 10,000′, however. If hydrocarbons exist in this section, the best way to prove such would be evaluating far-offset seismic data through AVO interpretation. A small presence of Class I data exists below 12,000′ for TILE2 below geopressure (Figure 12); the trend indicates mostly class III and II sands with the transition to Class II at about 13,000′. TILE19, above and below geopressure (Figures 11 ; 12), have average NI trends that are Class III.
Further AVO identification of pore fluid was performed by the plotting normal incident reflection coefficient versus the slope B reflectivity, which is a function of the P-wave, S-wave, and density. In the TILE2 NI vs B crossplot, (Figure 13), the polygon clusters indicate the fluid type with clusters farther from the origin representing more compressible pore fluids. Gas and Fizz sands are farthest from the origin in a southwest direction. In figure 14, TILE 19 well data also shows similar cluster trends with fizz and gas sand having the lowest NI and slope values. Values for brine-saturated sand is closest to the origin. Plots like this can help guide groupings of classes in unknown data, and assist the interpreter in understanding which seismic data will be the best indicator of hydrocarbon zones.
Crossplots of Vp/Vs ratio against Acoustic Impedance were made as a fluid type indicator. “Introducing shear wave traveltime is helpful in determining mechanical rock properties. It is found that compressional wave velocity is sensitive to the saturating fluid type, while shear-wave velocity is less sensitive to pore-fluid type. Therefore, the ratio of compressional wave velocity to shear wave velocity, Vp/Vs, is a good tool in identifying fluid type. The fact that compressional wave velocity decreases and shear wave velocity increases only slightly with the increase of light hydrocarbon saturation, makes the ratio of Vp/Vs more sensitive to change of fluid type than the use of Vp or Vs separately. Field examples have shown that shear traveltime decreases while compressional traveltime increases when the water saturated points become gas or light oil. The decrease of shear traveltime (increase of shear wave velocity) is due to the decrease of density. The increase of compressional traveltime (decrease of compressional wave velocity) is due to the decrease of bulk modulus of reservoir rocks which compensates the decrease of rock density”. (Hamada, 2004).
This is accurately depicted in the resulting crossplot for both TILE2 (Figure 15) and TILE19 (Figure 16) data. Shale has the highest Vp/Vs value with Gas sand having the lowest. Shale and Wet sand have a parabolic trend over a wide range of values while the hydrocarbon containing sands have increasingly linear range with increasing gas content.
LMR crossplots provide better separation of different pore-fluid clusters. LambdaRho is a sensitive indicator of water vs gas saturation and MuRho is used to distinguish pure rock fabric or lithology. Together, both derived parameters provide good lithology and fluid indicators.
For TILE2 (figure 17), LMR crossplots over 6000ft intervals were derived to better indicate lithology and fluid responses with increasing depth. Across shallow to intermediate depth intervals in TILE2, (Figure 17), Shale and Gas sand have the lowest lamdha-rho values respectively. There is a good separation among all lithologies except Gas sand and Fizz sand which have overlapping responses. This is to be expected since the fluid content is similar. But between 12001-18000ft, there is better separation between Gas and Fizz sand while Fizz sand becomes similar to Oil sand. While it appears that increasing depths leads to easier identification of Gas sand from Fizz sand, it also points towards similar Fizz and Oil sand trends at depth. TILE19 (Figure 18) LMR crossplots were generated over 5000ft intervals. At depths below 5000ft, there is good separation for all lithologies except Gas and Fizz sand. Between 5001-10000ft, Gas, Fizz and Oil sands are slightly overlapping each other respectively. At depths beyond 10000ft, Fizz and Oil sand responses become identical with all other lithologies having good separation.
This project created a petrophysical catalogue for two regions in the West Cameron area of the Gulf of Mexico using attributes derived from well log data. The workflow applied to the project is one that can be used in any Gulf of Mexico petrophysical evaluation. The analysis led to the following conclusions;
1. TILE2 and TILE19 data represent areas susceptible to overpressure mechanisms. TILE2 being overpressure results from a smectite to illite transformation while TILE19 is due to disequilibrium compaction. This is indicated by the fact that in TILE19, the shale velocity and density essentially remain the same as the onset values at geopressure. While in TILE2, the shale density keeps increasing as it goes from normal to abnormal pressure with depth and the shale velocity decreases going from normal to abnormal pressure.
2. In TILE2 and TILE19, the effect of geopressure on the velocity and density trends is undeniable as you move from depths above geopressured to below geopressure. Knowledge of prevailing pressure at depth is important to avoid inaccurate interpretation of the pore fluid.
3. TILE2 Gas sands are mostly Class III-II sands above and below geopressure while TILE19 Gas sands are predominantly Class III sands.
4. Lamdha-Mu-Rho crossplots provided good lithology and fluid discrimination for TILE2 and TILE19 data except between Gas and Fizz sands. For TILE2, at depths above 12,000ft, Fizz and Oil sands have identical responses and for TILE19 data, at depths above 10,000ft, Fizz and Oil sand responses overlap.