Show button Hide button Support

Facies Classification

Integration of Facies Classification into the Interpretation Suite 

A unique way to gain a deeper understanding of data through facies analysis.

Seismic facies analysis provides an alternative to seismic attribute analysis, when amplitude behavior is not the main driver for analyzing geological settings or reservoir condition changes related to lithology variations, fluid content changes or thickness variations. The different classification approaches embedded in Emerson's SeisEarth™ software solution enable better delineation and detailed analysis of prospects. 

This approach can simply be based on the analysis of the seismic signal, in order to provide qualitative information about an interval of interest.  If used for reservoir characterization purposes, seismic facies classification can be applied in an exploration context, as it provides images that contribute to the interpretation of geological settings, as well as more information than amplitude maps.

When we start drilling a prospect, however, more reliable information, such as well logs, is made available to us, and it is mandatory to be able to calibrate the facies models to well data.  Emerson E&P Software offers two different methods for facies-to-well calibration:

  • By supervising facies volume creation with facies logs 
  • By supervising waveform classification with 2D modeling (e.g. wedge modeling)

Facies Classification Benefits:

  • Streamlined workflow in the interpretation platform.
  • Faster, more accurate calibration between facies volumes and facies logs.
  • Increases geophysical meaning of facies results (“connecting the dots”).
  • Easy parameter testing and sharing.
  • No learning curve thanks to workflow approach.
  • Integration of QC steps into workflows promotes data validation within the data generation process.



Winner of Hart Energy’s Meritorious Awards for Engineering Innovation (MEAs) in Exploration/Geoscience

Click here to learn more about Rock Type classification.

The Rock Type Classification workflow uses a brand new algorithm which generates rock type volumes calibrated to facies logs using a probabilistic approach, to assess uncertainty in rock type quality and distribution.  Users will have the option to access the Rock Type Classification workflow directly within the interpretation platform.  The probabilistic approach results in less guesswork when quantifying uncertainty in rock type distribution. Results are interactively generated in a 2D and 3D environment for in-depth analysis.

Rock Type Classification Benefits:

  • Enables better risk management and well planning optimization
  • Brings new potential about seismic data reliability for predicting reservoir facies away from wells, especially when referring to prestack data, which carry more information with any type of seismic attributes.
  • Provides faster images of the subsurface while still maintaining accuracy, thus helping to accelerate the decision-making process in the drilling location determination.


Waveform Classification uses the same proven neural network method as used in Stratimagic for classifying seismic wave shapes. The implementation of this methodology as a plug-in to SeisEarth provides access to quality control tools as well as advanced wave shape constraining from the 2D Modeling utility, to map the impact of changes in reservoir conditions on amplitude data.  For example, users may map thickness variations or fluid effects on seismic data by constraining seismic facies mapping with synthetic traces representing those effects. With such an embedded workflow, the geoscientist can perform the calibration of waveform classification to different scenarios, and play the “what if” game to evaluate prospects.  

In a 4D analysis context, waveform classification models help to compare the seismic response between two different seismic vintages. This provides an easy way to analyze the evolution of the seismic response through time, as a result of production and injection.


Attribute clustering is a next-generation unsupervised classification method, which extends Machine Learning beyond facies analysis.  It supports multiple outputs: volumes, map and wells.

This method is valid for various uses, including:

  • Facies analysis
  • Structural delineation from multiple seismic attributes
  • AVO analysis from prestack data and AVO attributes
  • Outlier detection