Clay Volume

Purpose

Clay volume, Clay volume, answers how much of the rock is clay. It matters because clay changes almost every other log measurement. Clay holds bound water, so it raises the neutron porosity and lowers the resistivity of the formation. A log analyst who ignores it will overestimate porosity and water saturation in shaly sands, and may write off pay as non-reservoir.

Clay volume is also the main lever for reservoir quality. It is the quantity most often used to separate reservoir from non-reservoir rock, so it goes into the cutoffs as well as the porosity and saturation models.

Position in the workflow

Upstream. Clay volume needs cleaned, depth-matched and normalized curves from Stage 1. A washout or a bad normalization shows up directly as a false clay signal.

Downstream. Clay volume feeds:

effective porosity, because the clay-bound water is removed from total porosity,
water saturation, in every shaly-sand model (Simandoux, Indonesian, dual water, Waxman-Smits),
the cutoffs analysis, where a clay volume limit often defines net reservoir,
mineral inversion, which takes a clay volume as an input, and
laminated sand/shale analysis and the geomechanical rock properties.

Error propagation. An error in clay volume is not diluted downstream. A clay volume that is too high lowers effective porosity and, through the shaly-sand correction, raises the hydrocarbon saturation that the model assigns. A clay volume that is too low does the opposite. Because net pay is a threshold on these results, a small clay error can move a bed from pay to non-pay.

Key concepts

Clay, shale and VWCL. Clay volume (Clay volume) is the volume of clay minerals. Shale volume (Shale volume) is broader: it is the volume of shale, which includes the silt and the water that the clay carries. The two are often used interchangeably, but they are not the same thing, and the downstream models are specific about which one they need. Always check which quantity a method returns and which one the next step expects.

Indicators. Each method uses a different physical indicator of clay: natural radioactivity (gamma ray), the electrochemical potential at the shale boundary (SP), the hydrogen content of bound water (neutron), or the departure of the neutron and density response from the clean-rock trend (crossplots).

Clean and clay endpoints. Every method is a map from an indicator value to a clay volume, anchored by two picks: the value in a clean, clay-free interval and the value in a 100% clay interval. For the gamma ray these are Clean gamma ray and Clay gamma ray. The result is only as good as these picks.

Linear index and non-linear transforms. The linear gamma ray index, Gamma ray index, is the straight-line interpolation between the endpoints. It is usually a maximum estimate of clay volume, because most gamma-ray responses are not linear in clay content. The non-linear transforms (Larionov, Clavier, Stieber) reduce the linear index by an amount that grows in cleaner intervals. They are empirical fits, not derived relations.

The table below shows the reduction at three values of the index, computed from the standard published forms of each relation:

import math
print(f"{'IGR':>5} {'Linear':>8} {'Stieber':>8} {'Clavier':>8} {'Larionov (Tertiary)':>20} {'Larionov (older)':>17}")
for igr in (0.25, 0.50, 0.75):
    stieber = igr / (3 - 2 * igr)
    clavier = 1.7 - math.sqrt(3.38 - (igr + 0.7) ** 2)
    lar_tertiary = 0.083 * (2 ** (3.7 * igr) - 1)
    lar_older = 0.33 * (2 ** (2 * igr) - 1)
    print(f"{igr:5.2f} {igr:8.3f} {stieber:8.3f} {clavier:8.3f} {lar_tertiary:20.3f} {lar_older:17.3f}")

Output

  IGR   Linear  Stieber  Clavier  Larionov (Tertiary)  Larionov (older)
 0.25    0.250    0.100    0.126                0.075             0.137
 0.50    0.500    0.250    0.307                0.216             0.330
 0.75    0.750    0.500    0.570                0.485             0.603

Method selection guide

Method	Inputs	Use when	Strengths	Weaknesses
GR, linear	GR	Only a gamma ray is available and a conservative (high) answer is acceptable	Simple, always available, easy to explain	Overestimates clay volume in most rocks
GR, Larionov (older rocks)	GR	Only a gamma ray; the rock age is known	Calibrated by rock age, reduces linear overestimate	Age split is a rule of thumb; fixed shape
GR, Clavier	GR	Only a gamma ray; a moderate reduction is wanted	Smooth, bounded	Fixed shape, empirical
GR, Stieber	GR	Only a gamma ray; a simple, one-constant reduction is wanted	Simple closed form, reaches 1 at the clay point	Fixed shape, empirical
SP	SP	Fresh mud over saline formation water, thick beds, no gamma ray quality	Not affected by radioactive minerals	Depends on mud and water salinity contrast, fails in oil-based mud and thin beds, suppressed by hydrocarbons
Neutron	Neutron	A clean reference neutron porosity is known	Sensitive to bound water	Gas lowers the neutron; needs a matrix and clay pick
Neutron-Density	Neutron, density	Both logs are good and gas is not dominant	Independent of the gamma ray; handles radioactive minerals	Poor in washouts, gas and heavy minerals; needs two endpoint pairs
Sonic-Density	Sonic, density	No neutron log	Independent of the gamma ray	Sonic is affected by compaction and gas
Neutron-Sonic	Neutron, sonic	No density log	Independent of the gamma ray	Same limits as the single curves

Decision guidance

If you only have a gamma ray, start with a non-linear gamma ray transform and treat the result as an upper bound.
If the interval contains radioactive minerals (feldspar-rich arkoses, micaceous sands, glauconite, uranium-rich organic shales), avoid gamma ray methods as the primary indicator and use neutron-density.
If the mud is oil-based, or the beds are thin, do not use SP.
If the interval is gas-bearing, treat neutron-based methods with caution. Gas lowers the neutron response and pulls the estimate toward a cleaner rock.
If more than one indicator is available, compute them all and compare, as described below.

Shared parameter picking

All gamma ray methods use the same two picks, Clean gamma ray and Clay gamma ray. They belong here so that they are chosen once.

Pick per zone, not per well. Clean-sand and shale responses change with formation. Use a separate pair for each major zone.
Read them from the histogram. For Clean gamma ray, use a low percentile of the GR histogram in clean reservoir intervals (the 5th to 10th percentile is a common choice). For Clay gamma ray, use a high percentile in thick shales (the 90th to 95th percentile). Avoid the absolute minimum and maximum, which are often noise or bad data.
Check against geology. Clean gamma ray should correspond to intervals you believe are clean sand from cores, cuttings or other logs. Clay gamma ray should be taken from a shale that is representative, not from a hot or a bad-hole shale.
Pick on normalized curves. If the curves have been normalized across wells, use the same picks across the wells.

Other methods use their own endpoints. SP uses an SP clean and an SP clay, and neutron and crossplot methods use a clean and a clay point on the neutron and density scales. The same principles apply.

Recommended default approach

Absent other information, a careful generalist would:

Clean and normalize the gamma ray, then pick Clean gamma ray and Clay gamma ray per zone from the histogram.
Compute the linear index and one non-linear gamma ray clay volume.
If neutron and density are available and of good quality, compute the neutron-density clay volume.
Compare the two, investigate any large differences, and take the lower of the two as the working clay volume unless there is a reason to prefer one.

Combining methods

When more than one indicator is available, combine them rather than choosing arbitrarily.

Minimum of several. Each method has its own way of reading too high: the gamma ray with radioactive minerals, the neutron with bound water in non-clay minerals. Taking the minimum is a standard way to discard the biases that push the estimate up. It is biased low if one indicator reads too low, for example neutron in gas.
Average. Appropriate when the methods are individually well calibrated and the errors are random rather than systematic.
Reconciliation. Where the methods disagree, find out why before combining. The disagreement is information: a gamma ray that is high relative to the neutron-density clay volume is a signature of radioactive minerals.

QC of results

A good result:

lies between 0 and 1, with values near 0 in clean reservoir and near 1 in thick shale,
follows the visible shale and sand pattern on the logs,
agrees reasonably with an independent indicator, with differences that can be explained, and
ties to core or XRD clay content where available.

Signs of a bad result:

clean sands reading 10 to 20% clay or more,
clay volume close to 1 over a thick clean interval, which suggests a wrong clay pick,
clay volume that tracks the caliper, which suggests borehole effects, and
a gamma-ray-based result that is much higher than a neutron-density result, which suggests radioactive minerals.

Common pitfalls

Using the same endpoints for all zones and all wells.
Treating the linear gamma ray index as the clay volume without a correction.
Picking endpoints from the data minimum and maximum instead of percentiles.
Ignoring radioactive minerals.
Using neutron-based methods through gas.
Using SP with oil-based mud or in thin beds.
Mixing up clay volume and shale volume when passing the result to the next step.

Going Deeper

The gamma ray methods dominate practice because the gamma ray is cheap, nearly universal and available in old wells. The non-linear transforms were developed to correct a known bias in the linear index, and they are calibrated to particular formations and ages, which is why they should be validated locally rather than adopted by default. Spectral gamma ray, which separates potassium, uranium and thorium, can remove the contribution of uranium or potassium and give a cleaner clay indicator where it is available. Core-derived clay content from XRD is the best available check and the one that should anchor any calibration.