Wellbore Positions Obtained by MWD-IFR May Be Less Accurate Than Predicted


Various methods, often referred to as in-field referencing (IFR) techniques, are used to enhance the performance of traditional magnetic measurement while drilling (MWD) well-bore surveys. Simulation and field data studies of effects commonly not accounted for, like the fact that the geomagnetic field varies with depth, indicates an unrealistically high level of confidence in the IFR techniques, especially when run in conjunction with multi station corrections (MSC). The SPE wellbore position technical section (WPTS) is currently recommending the use of the standard MWD error model weighting functions, which do not reflect a horizontal east-west singularity associated with combined IFR-MSC surveys. The WPTS is further recommending the use of geographically dependent IFR azimuth uncertainty figures, which average globally to 0.22º (1s). The WPTS average is supposed to account for all errors that corrupt an IFR survey, not only those affected by the IFR correction itself which, according to WPTS, have an average of only 0.18º (1s). The WPTS figures do not account for the additional uncertainty generated by geomagnetic depth variation. Simulations show a global uncertainty of 0.38º (1s) from this error source alone. Geomagnetic depth corrections are possible, but not commonly applied, and are far from error free; residual errors after correction exceeding 30% appear to be common. The study described here indicates an underestimation of the real wellbore position (IFR based) uncertainty of 50% or more, a fact that represents a potential safety hazard to both human life and equipment of which the industry needs to be made aware.


Magnetic directional surveys use measurements of the Earth’s field derived from sensors in the survey tool to establish the orientation of the tool with respect to the directional reference defined by the Earth’s magnetic field vector. The accuracy of magnetic surveys is compromised as a result of variations and local distortions in the reference magnetic field, viz.

  • secular variations resulting from long term changes within the Earth’s magnetic core;
  • diurnal variations caused by solar wind and Earth rotation;
  • crustal variations owing to deep, magnetic basement rock giving rise to local variation, or other anomalies, in the ambient field.


The MWD quality is often improved through various techniques designed to reduce one or more of these distortions. The effects of secular variations can for example be corrected by the use of the British Geological Survey model of the Earth’s field (the BGGM model), while the effects of diurnal variations can be reduced by local monitoring of the time-dependent changes in the magnetic field, and the crustal variation through a pre-mapping of surface anomalies. There exist a number of different methods, and naming conventions for this type of service. For example, in-field referencing is sometimes used in connection with MWD corrections based on local monitoring only, and sometimes in connection with corrections based on all three types. This paper is based on the latter alternative.

The paper explores problems that are frequently encountered when attempting to survey certain well bore trajectories when using magnetic measurement while drilling (MWD) in conjunction with in-field referencing (IFR) techniques, and seeks to explain the causes of inconsistencies which have been found to arise. An example of this behaviour is given in Figure 1. Such problems seem to be particularly evident when surveying wells that are coincident with, or approach, the horizontal east-west direction. In such cases, unexplained differences between surveys conducted using IFR MWD and gyro survey equipment have been observed, differences which often exceed the magnitude of survey errors predicted by the combined IFR/gyro error model. IFR surveys without axial magnetic corrections, or the more powerful multi station corrections (MSC), will often be corrupted by magnetic interference that exceeds the total IFR error model budget.