Skip to main content
Version: 3.4.8 (unreleased)

Measuring fracture intensity

Orientation tells you which way the fractures run. Intensity tells you how many there are, and it is the number that drives everything predictive — flow, strength, and the DFN.

VRGS measures it two ways: along a line you place, and across the whole outcrop surface as a map. They answer different questions and it is worth being clear which one you need.

The measures

DefinitionUnitsMeasured from
P10fractures per unit lengthm⁻¹a scan line
P21trace length per unit aream⁻¹fracture traces on the mesh
P32fracture area per unit volumem⁻¹orientation glyphs
P33fracture volume per unit volume (≈ porosity)P33 = aperture × P32

The one to keep hold of: P10 and P21 are what an outcrop can show you. P32 is what a model needs. A 2D exposure cannot directly measure area per unit volume, so P32 is always derived — either stereologically from P21, or by fitting a DFN. That derivation is where most of the uncertainty in a fracture model lives.

P10: scan lines

If you made scan lines in step 1, the results are already in their properties:

PropertyMeaning
P10Fracture count divided by scan-line length.
Set SpacingMean raw 3D distance between consecutive fractures along the line.
Normal Set SpacingThe same distance projected onto the set's mean pole — true perpendicular spacing.
Fisher KConcentration of the logged fracture orientations.
Mean Dip / Mean AzimuthThe mean orientation of the fractures the line crossed.
Raw P10 depends on where you put the line

A scan line perpendicular to a set crosses every fracture in it. A scan line parallel to that set crosses almost none. Both give a "correct" P10 for a completely different rock.

This is the Terzaghi bias, and VRGS corrects for it — but only when each logged fracture has a measured orientation, which means a From intersections scan line. A From Points scan line has no per-fracture orientations, so the correction, the Fisher statistics and the normal spacing are all undefined and reported as zero.

If a scan line shows a suspiciously round zero for those fields, that is what happened.

Place several scan lines in different orientations rather than one. Comparing their corrected P10 is a far better estimate than trusting any single line.

P21 / P32 / P33: intensity maps

Instead of one number for one line, this computes intensity at every vertex of the mesh and stores it as a colour-mappable attribute — so you can see where the outcrop is more fractured.

In the Data tree, right-click the triangulated mesh:

  • P21 Map — the full command. Prompts for every parameter, and produces P21, P32 and P33.
  • Radial Basis Function (RBF) Area — a quick path that asks only for the radius and uses sensible defaults.

Start with the quick path to see whether the radius is sensible, then re-run the full command once you know.

The one parameter that matters

Every value is computed inside a search sphere of radius R centred on each vertex. R is a bandwidth, not a detail setting:

  • Small R — high spatial detail, but noisy, with few fractures per sphere.
  • Large R — smooth and stable, but blurs the variation you were looking for.

A reasonable starting point is a few times the mean fracture spacing, which your scan lines have just given you. That is the practical reason to do the scan lines first.

Both halves of every ratio are clipped exactly to the sphere — a trace that only partly crosses it contributes only the length inside, and a triangle straddling it contributes only the area inside — so the result is a true local intensity that varies continuously as the sphere sweeps across the mesh.

The trap in P32

P21 is measured from your traces, and traces are real observations. P32 is measured from the orientation glyphs — the virtual ellipse VRGS draws for each measurement — and the glyph's size is the modelled fracture extent.

So P32 depends directly on the glyph display settings. Auto Glyph Size sizes each glyph to the extent of the vertices actually digitised for that measurement, which ties it back to how big the fracture was on the outcrop. With glyph sizes left at some arbitrary default, P32 is an arbitrary number with units.

Check the glyph settings before you read a P32 map, not after.

P33 needs an aperture

Leave the aperture at 0 and the P33 layer is all zero — which is correct rather than broken, since P33 is aperture × P32 and you have not supplied one. Fill it in only if you have an aperture measurement worth using.

Displaying the result

The maps come back as attribute layers on the mesh. Display them exactly like any other attribute — see Attribute generation and analysis: pick a perceptually uniform colour map, and use Clip to stretch the colour scale across the range that actually matters rather than letting one hot spot flatten everything else.

See also

Next: building a DFN.