lithologic modeling

lithologic modeling is an alternative geologic modeling concept that uses geostatistics to assign each cell’s lithologic material as defined in a pregeology (.pgf) file, to cells in a 3D volumetric grid.

There are two Estimation Types:

Nearest Neighbor is a quick method that merely finds the nearest lithology sample interval among all of your data and assigns that material. It is very fast, but generally should not be used for your final work.
Kriging provides the rigorous probabilistic approach to geologic indicator kriging. The probability for each material is computed for each cell center of your grid. The material with the highest probability is assigned to the cell. All of the individual material probabilities are provided as additional cell data components. This will allow you to identify regions where the material assignment is somewhat ambiguous. Needless to say, this approach is much slower (especially with many materials), but often yields superior results and interesting insights.

There are also two Lithology Methods when Kriging is selected.

The default method is block. This method is the quickest since probabilities are assigned directly to cells, and lithology is therefore determined based on the highest probability among all materials. However the resulting model is "lego-like" and therefore requires high grid resolutions in x, y & z in order to give good looking results.
The other method is Smooth. With Smooth, probabilities are assigned to nodes. In much the same way as analytical data, nodal data for probabilities provides an inherently higher effective grid resolution because after kriging probabilities to the nodes, there is an additional step where we "Smooth" the grid by interpolating between the nodes, cutting the blocky grid and forming a new smooth grid. MUCH lower grid resolutions can be used, often achieving superior results.

Module Input Ports

Input Geologic Field [Field] Accepts a data field from gridding and horizons to krige data into geologic layers.
Filename [String / minor] Allows the sharing of file names between similar modules.
Refine Distance [Number] Accepts the distance used to discretize the lithologic intervals into points used in kriging.

Module Output Ports

Geologic legend Information [Geology legend] Supplies the geologic material information for the legend module.
Output Field [Field] Contains the volumetric cell based indicator geology lithology (cell data representing geologic materials).
Filename [String / minor] Outputs a string containing the file name and path. This can be connected to other modules to share files.
Refine Distance [Number] Outputs the distance used to discretize the lithologic intervals into points used in kriging or displayed in post_samples as spheres.

Properties and Parameters

The Properties window is arranged in the following groups of parameters:

Grid Settings: control the grid type, position and resolution
Krig Settings: control the estimation methods
- NOTE: The Quick Method assigns the lithologic material cell data based on the nearest lithologic material (in anisotropic space) to your PGF borings. This is done based on the cell center (coordinates) and an enhanced refinement scheme for the PGF borings. In general the Quick Method should not be used for final results

Advanced Variography Options:

It is far beyond the scope of our Help to attempt an advanced Geostatistics course. The terminology and variogram plotting style that we use is industry standard and we do so because we will not provide detailed technical support nor complete documentation on these features, which would effectively require a geostatistics textbook, in our help.

However, we have offered an online course on how to take advantage of the complex, directional anisotropic variography capabilities in 3d estimation (which applies equally well to lithologic modeling and adaptive_indicator_krig), and that course is available as a recorded video class. This class is focused on the mechanics of how to employ and refine the variogram anisotropy with respect to your data and the physics of your project such as contaminated sediments in a river bottom. The variogram is displayed as an ellipsoid which can be distorted to represent the Primary and Secondary anisotropies and rotated to represent the Heading, Dip and Roll. Overall scale and translation are also provided as additional visual aids to compare the variogram to the data, though these do not affect the actual variogram.

We are not hiding this capability from you as the Anisotropic Variography Study folder of Earth Volumetric Studio Projects contains a number of sample applications which demonstrate exactly what is described above. However, we assure you that understanding how to apply this to your own projects will be quite daunting and really does require a number of prerequisites:

A thorough explanation of these complex applications
A reasonable background in Python and how to use Python in Studio
An understanding of all of the variogram parameters and their impact on the estimation process on both theoretical datasets as well as real-world datasets.

This 3 hour course addresses this issues in detail.

Discussion of Lithologic (Geologic Indicator Kriging) vs. Stratigraphic (Hierarchical) Geologic Modeling

Stratigraphic geologic modeling utilizes one of two different ASCII file formats (.geo and .gmf) which contain "interpreted" geologic information. These two file formats both describe points on each geologic surface (ground surface and bottom of each geologic layer), based on the assumption of a geologic hierarchy.

The easiest way to describe geologic hierarchy is with an example. Consider the example below of a clay lens in sand with gravel below. Some borings will see only sand above the gravel, while others will reveal an upper sand, clay, and lower sand.

$image\heir1.jpg$

The geologic hierarchy for this site will be upper sand, clay, lower sand, and gravel. This requires that the borings with only sand (above the gravel) be described as upper sand, clay, and lower sand, with the clay described as being zero thickness. For this simple example, determining the hierarchy is straightforward. For some sites (as will be discussed later) it is very difficult or even impossible.

$image\heir2.jpg$

For those sites that can be described using the above method, it remains the best approach for building a 3D geologic model. Each layer has smooth boundaries and the layers (by nature of hierarchy) can be exploded apart to reveal the individual layer surface features. In the above example, the numbers represent the layer numbers for this site (even though layers 0 and 2 are both sand). Two examples of much more complex sites that are best described by this original approach are shown below.

Geologic Example: Sedimentary Layers and Lenses

$image\evslayr1_wmf.jpg$

Geology Example & Figure: Outcrop of Dipping Strata

EVS is not limited to sedimentary layers or lenses. The figure below shows a cross-section through an outcrop of dipping geologic strata. EVS can easily model the layers truncating on the top ground surface.

$image\evslayr2_wmf.jpg$

However, many sites have geologic structures (plutons, karst geology, sand channels, etc.) that do not lend themselves to description within the context of hierarchical layers. For these sites, Geologic Indicator Kriging (GIK) offers the ability to build extremely complex models with a minimum of effort (and virtually no interpretation) on the part of the geologist. GIK can also be a useful check of geologic hierarchies developed for sites that do lend themselves to a model based upon hierarchical layers.

GIK uses raw, uninterpreted 3D borings logs as the input file. The .pgf (pre-geology file) format is used to represent these logs. PGF files contain descriptions of each boring with x,y, & z coordinates for ground surface and the bottom of each observed geologic unit. Consecutive integer values (e.g. 0 through n-1, for n total observed units in the site) are used to describe each material observed in the entire site.

NOTE: It is important to start your material ID numbering at zero (0) instead of 1.

Usually, materials are numbered based upon a logical classification (such as porosity or particle size), however the numbering can be arbitrary as long as the numbers are consecutive (don’t leave numbers out of the sequence). For the example given above, we could number the materials as shown in the figure below (even though it is not a numbering sequence based on porosity or particle size).

$image\heir3.jpg$

For a .pgf file, borings that do not see the clay (material 2 in the figure) would not need to consider the sand as being divided into upper and lower. Rather, every boring is merely a simple ASCII representation of the raw borings logs. The only interpretation involves classification of the observed soil types in each boring and assigning an associated numbering scheme.