WSAG
Lundi Log 001
Example below is a Deepwater GOM LWD triple-combo log suite with petrophysical parameters including interpreted fluid types. Oil on top of both reservoirs? Kind of doubt it. Why?
Scroll below for interpretation.
Figure 1. Deepwater Gulf of Mexico Triple-Combo Log Suite with various petrophysical curves. Curve scales are standard: GR, 0–150 api; Res, two cycles from 0.2 to 20 ohm-m; Dens, 1.65–2.65 g/cc; and NPhi, 60-0% in SS units. The 4 resistivity curves are inverted curves derived from the full suite of 32 raw phase and attenuation propagation resistivity curves to obtain fixed DOI (10 in [green], 20 in [red], 35 in [blue], and 60 in [black]) and vertical resolution matched (2 ft) curves. Well is deviated and strata are dipping, generating a 30–35° relative dip .
Lundi Log 001 Interpretation
The "oil" is actually due to polarization horns, an artifact of LWD propagation resistivity.
The petrophysical log indicates oil at the top of both reservoirs. While economically this potential “skim” of oil on top of water is uneconomical, we should consider updip potential. The moderate dips here (~20–30°) indicates that not too far from the present well position (~200–300 ft laterally in updip direction), full top-to-bottom saturation would occur. Given these circumstances, we’d likely be looking at lateral extent of the “full pay” domain to consider potential sidetrack or offset well. But, it’s not oil and a sidetrack or offset well might actually show “oil” at the top for the same reasons as described below.
The overall log characteristics of these two reservoirs indicates water-bearing intervals: Low GR, low Res, NPhi and Dens convergence due to less H and less Dens versus shale response. NPhi and Dens curves exhibit a “kissing” nature, suggesting liquid due to the clean nature of the sands. The two curves are not exactly on top of one another because the SS scaling of Nphi and Dens assumes quartz sand and water with 1.0 g/cc. The brine waters are heavier than 1.0 g/cc, which shifts the Dens curve rightward from where the 1.0 g/cc water line would be and the presense of some shale causes some separation of the NPhi and Dens curves (increasing Vsh increases Dens [curve moves right] and increases NPhi [curve shifts left due to more H in the clay minerals]). If these would be oil bearing strata, GR would basically stay the same, but resistivity overall would be higher and Nphi and Dens would be more closely overlapped and maybe even cross over slightly especially with light oils typical of the depths in the deepwater Gulf of Mexico. This is due to the density of oil being <1.0 g/cc so the Dens curve would be shifted leftward versus the water response and NPhi would be shifted rightward due to lesser H and thus lower NPhi in oil zones versus if water filled. If these would be gas-bearing strata, Nphi and Dens would show signficant crossover (gas has significantly less H than liquid so NPhi reads much lower and gas is much lighter than 1.0 g/cc so Dens would exhibit signficant leftward shift versus a water or oil response). Although the sands do contain some shale, and note that there are techniques that can determine how that shale is distributed in the reservoirs, the general patterns of water, oil, and gas responses on the logs holds, whereas high shale volumes for example can remove the gas crossover effect.
We need to investigate where the “oil” actually occurs in the two reservoirs. Due to curve resolution limitations a sharp bed boundary exhibits a transitional nature due to shoulder bed effects. Higher resolution lessens the shoulder bed effect and the transition appears sharper. Regardless, a bed boundary of a sharp contact is not at the first log response from one rock to another but is typically taken at the inflection point of the transition. In Figure 1, for both reservoirs, a high resistivity spike occurs above the inflection point for various curves, indicating that it is occuring within the shale, not the sand. Because GR, NPhi, and Dens curves are likewise transitioning from shale to sand, the petrophysical algorithm treated the high resistivity with an indication of sand (decreasing GR) and porosity (increasing Dens and decreasing NPhi) are flagged the resistivity spike as hydrocarbons. It is not. The resistivity spike is in the shale! Of course, oil shales exist but an interpretation of two rather thin oil shales above sandstone reservoirs in a deepwater depositional environment does not make sense geologically. Something else must be going on to explain the resistivity spike.
If the upper portion of both reservoirs did in fact contain hydrocarbons, an increase in resistivity relative to the overlying shale would appear BELOW the inflection point for these sharp transitions with associated decrease in Dens and NPhi. We don’t see this, as the resistivity spike is ABOVE the inflection point. If a thin zone in either the shale or the upper part of the sand would exhibit a decrease in porosity, then resistivity would spike up, but both NPhi and Dens would reveal the lower porosity (both curves shifting righwards). Again, however, we don’t see this.
Let’s have a look at the original triple-combo log (Figure 2) that shows not just the inverted resistivity curves but also the original phase and attenuation curves.
Figure 2. Deepwater Gulf of Mexico Triple-Combo Log Suite with phase and attenuation resistivity curves (left resistivity track) and the inverted resistivity curves (right resistivity track. Curve scales are standard: GR, 0–150 api; Res, three cycles from 0.2 to 200 ohm-m; Dens, 1.65–2.65 g/cc; and NPhi, 60-0% in SS units. The 4 phase and attenuation curves are high-frequency phase (red), high-frequency attenuation (green dashed), low-frequency attenuation (purple dotted), and low-frequency phase (blue dashed). The 4 resistivity inversion curves are fixed DOI (10 in [black], 20 in [red], 35 in [purple dashed], and 60 in [blue dashed]) and vertical resolution matched (2 ft) curves. Well is deviated and strata are dipping, generating a 30–35° relative dip .
Figure 2 shows curve separation in the phase and attenuation resistivities. In a wireline environment, due to the long time since drilling, invasion may occur in porous and permeable reservoirs and is a dominant cause of curve separation on wireline logs (other factors can and do cause curve separation). In the LWD environment, however, the typical short time since drilling often limits the amount of invasion and other factors may dominate curve separation, such as different vertical resolutions of individual curves, anisotropy, eccentricity, dielectric effects, etc. Note that the shales in particular exhibit more curve separation, likely due to anisotropy. The process of resistivity inversion attempts to correct for these effects, for example by vertical resolution matching (all processed curves having the same resolution) and other factors. The inverted curves track one another quite well indicated that these effects have largely been removed. However, the inversion algorithms “struggle” at sharp boundaries, similar to other filtering mechanisms, oftentimes yielding anomalously high or low resistivities. Figure 2 shows slightly elevated resistivities versus shale response of especially the high-frequency phase curve above the upper shale-sand inflection point and below the lower sand-shale inflection point. The inverted logs have increased the resistivities to an upper sharp spike and a lower bump. It is this increased resistivity that triggered the oil interpretation at the tops. The lower bumps were insufficient to trigger the oil flag.
So what is causing the resistivity increase in the shale in the phase/attenuation curves? We have to consider the assumptions of the resistivity model and log response. The response of propagation resistivity logs is based on a cylindrical borehole model, with the tool centered in a hole perpendicular to strata, treating the borehole and the formation as discrete cylinders. This assumption used in the generation of the phase and attenuation curves can fail due to many circumstances. For example, a tool straddling a bed boundary now effectively has two stacked formation cylinders, yielding an intermediate response, and since each curve has a different resolution their formation cylinders may incorporation different proportions of the two formation, yielding curve separation as described above. Similarly anisotropy, eccentricity, etc. “foul up” the assumed model, thus causing differential effects.
These effects can be worsened as relative dip increases from 0 to 90° (i.e., from well and strata being mutually perpendicular to mutually parallel, for example, in reservoir navigation). Now consider near a bed boundary that not only is the formation cylinder actually two cylinders (in an ideal situation) but also with relative dip now the boundary between those formation is dipping relative to the borehole. This introduces a new effect, so-called polarization. Effectively a buildup of capacitance charge occurs at bed boundaries of sharply contrasting resistivity when the tool exhbits non–0° relative dip, progressively increasing as relative dip increases. The modeled response does not account for this and resulting phase and attentuation curves exhibit an increase in resistivity on the resistive side of the bed boundary, yielding a so-called polarization horn. A major concern in hydrocarbon-bearing reservoirs would be failure to recognize that the increased resistivty is an artifact of processing and instead think Rt is increased, yielding lower water saturation. In particular, high-frequency and phase measurements are more strongly influenced by polarization effects versus low-frequency and attenuation measurements. Polarization effects increases as relative dip increases—polarization effects have been used in reservoir navigation because the high anomalous values can indicate impending reservoir exit into a conductive zone (shale or water leg)! Additionally, greater resistivity differences across the boundary increase polarization effects as does the log-parallel sharpness in the resistivity change from one layer to another.
Polarization effects can be quite significant in resistivity hydrocarbon-bearing strata adjacent to conductive shales, due to the large resistivity contrast, but the effect can occur on the resistive side of any sharp contrast. Figure 3 shows a Gulf of Mexico example with high relative dip (~70°). Note the upper part of both reservoirs exhibits strong polarization horns due to the large difference in resistivity compared to the overlying shale and sharp transition from shale to sand. The lower part of the upper reservoir exhibits a gradual coarsening-upward pattern, thus the resistivity change is gradual and no horn develops. The lower sand exhibits a basal sand bounded up in relatively sharp contact with shaly sand. There is less resistivity contrast between sand and shaly sand and thus although horns develop on its top and bottom, the effects are smaller than at the tops of the main reservoirs.
Figure 3. Gulf of Mexico Triple-Combo Log Suite showing polarization horns at sharp boundaries between hydrocarbon-bearing reservoirs and overlying shale.
In the Lundi Log 001 example (Figures 1 and 2), a sharp resistivity contrast exists at the top of each water-bearing sand relative to the overlying shale with a moderate relative dip. The resulting polarization horns develop in the shale due to its higher resistivity compared to the wet sand. They are clearly not as well developed as in the Figure 3 example, due to lower relative dip and less resistivity contrast. The resistivity inversion further amplifies the polarization artifact to the point of triggering the oil flag. The basal contacts exhibit lesser resistivity contrast due to shale and porosity variations and thus form more of a polarization bump rather than horn.
Literally in this case the devil is in the details! In hydrocarbon zones with both sharp top and sharp base, the polarization horns can even look like devil horns. The worse devil though is falsely thinking hydrocarbons as in the Lundi Log example! Sidetrack or offset updip to catch “full pay” may well yield the same horn effect!