Inversion of EM data:
gas and tar sands exploration


Shallow gas exploration

This work is summarized from Shallow Gas Exploration with Electromagnetic Systems: Frequency Domain EM and DC Resistivity surveys, Scott Napier, 2003, unpublished B.Sc. Thesis, UBC.

A methodology for inverting airborne frequency domain data to recover models of electrical conductivity has been developed as an undergraduate research project.

A first step is to invert high frequency data to correct altitude information for errors due to vegitation and other causes. Then the complete set of frequencies can be inverted for a layered Earth structure (a 1D model) under every measurement location. All 1D inversion results from data collected along one flight line are then gathered and concatenated to produce a 2D image. This is all done using the UBC-GIF code, "EM1DFM".

The geologic situation being investigated is shown in the figure above. Result of inversion along one line is shown in the figure below. Data are shown in the top of this panel, then misfit, model norm and altitude are plotted (blue-green, red, and black graphs respectively), and finally the conductivity structure under the line is plotted as a colour contour plot with "warm" colours representing more conductive geologic units. The effect of correcting altitude this way is visible by moving your mouse over the figure below.

Results of performing this type of process on a large data set gathered along many lines can be concatenated into a 3D model. The model of conductivity structure shown below has had the most conductive regions removed from the image. The "Z" shaped region is a pipeline between two wells, and the resistive region in the upper right could be a gas reservoir.

Other aspects of this work:

  • DC resistivity data have also been gathered in this environment and 2D inversion has been performed.
  • Future research directions are described next.

Inversion of time domain EM for characterization of tar-sands resources

This work is summarized from Geotem Surveying And Inversion Applied To Oil Sands Exploration Kearl Lake, Athabasca Region, Alberta, Jamin Cristall, 2003, unpublished B.A.Sc. thesis.

Objectives of this project: (i) To determine whether airborne time domain electromagnetic surveying is a viable method for imaging shallow oil sands deposits; (ii) to develop an inversion methodology for interpreting the data in terms of complex geology; (iii) to demonstrate the practical use of this methodology in a real-life oil sands exploration setting.

The challenge: For the huge areas containing tar sands resources, map the:

  • depth to top of McMurray Fm.;
  • thickness of McMurray Fm.;
  • depth to base of McMurray Fm.;
  • quality of oil sands.

Forward modeling: It is useful to explore and compare the system response to various aspects of the geology. Properties of the formation rocks are obtained from well logs in the areas (example to the right) and from literature on tar sands geology. Well log data in blue in the figure to the right are averaged into a few layers associated with each real formation, and this simplified model is used to predict how sensitive the TEM surveys will be to characteristics of these geologic situations. Conclusions include:

  • The data are most sensitive to shallow, conductive units and are less sensitive to deep, resistive units;
  • GEOTEM perceives the average conductivity of the formation when resistivities change rapidly within the formation - i.e. fine structure is not resolvable.

Inversion testing: Several aspects of inversion methodology were investigated to assess which procedure is most likely to recover useful characteristics of the formations. Conclusions include:

  • Use of the L1-Norm rather than more conventional L2-Norm (smooth models) may offer improved resolution of layer boundaries and bulk resistivities.
  • Tradeoff parameter can be chosen automatically (L-curve or GCV methods) or it can be fixed. Experiments for this work suggested that fixed tradeoff parameter produced models with structures that were more appropriate for this context.
  • It is possible for other layers to be averaged in to the recovered McMurray Formation.
  • Recovered bulk resistivities for the McMurray Formation may well be lower than its true resistivity.

Inversion of field data: Performing 1D inversions at every location along the flight line, and then concatentating the 1D models to produce a section works well in this environment, where structures are predominantly flat-lying. An example is shown below. Moving your mouse over the image displays a more conventional interpretation image called a conductivity depth transform generated from the same data set.


  • GEOTEM surveying combined with formal inversion provides a powerful tool for oil sands exploration.
  • Careful interpretation is required.
  • Inversion provides the means to interpret depth to top, thickness, and relative quality of oil sands at shallow to moderate (~200 metres) depths.


  • Use current methodology for surveying vast areas for potential SAGD (steam assisted gravity drainage) reservoirs.
  • Use MEGATEM system (~6.5? dipole moment) for imaging the McMurray Fm. at depths exceeding ~200 metres.
  • Develop methodology for quantitatively incorporating borehole data and seismic constraints to improve resolution further.