In the late 1990s, Statoil invented the use of the controlled-source electromagnetic (CSEM) method for remote identification of hydrocarbons in the marine setting. The method, referred to as Seabed Logging (SBL), was tested offshore Africa in 2000, and subsequent pilot tests proved CSEM as a tool with the potential to predict hydrocarbons before drilling.

The marine CSEM method usually employs a horizontal electric dipole source with user-specified frequency content. The source is towed along a towline, normally as close to the seafloor as possible. In a conventional 2-D survey, receivers are positioned on the seafloor along the towline. In a 3-D survey, receivers are positioned in a grid on the seafloor.

In a conventional 2-D survey, receivers are positioned on the seafloor along the towline. In a 3-D survey, receivers

In a conventional 2-D survey, receivers are positioned on the seafloor along the towline. In a 3-D survey, receivers
are positioned in a grid on the seafloor. (Image courtesy of EMGS)

The data analysis now is to obtain one or more resistivity models of the subsurface that give responses that fit the acquired data, thus mapping CSEM responses back to resistivity distributions in depth. These resistivity models can be correlated to seismic data and geological knowledge about the area.

In the early days of CSEM exploration, data analysis was simpler and often consisted of interpreting so-called normalized plots where one considered variation of CSEM responses along a towline, thus having little or no depth control.

The early technical successes of the first test surveys, in particular the Troll calibration survey from 2003, created an enormous commercial interest. The industry quickly adopted the new technology and pushed applications into more challenging areas. Among the success cases, there were several vague and inconclusive results, and the total commercial interest declined. In retrospect, important questions are: What have we gotten out of the CSEM data so far? What is the prediction strength and accuracy? Is CSEM data worth the money?

Performance tracking
Performance tracking of CSEM is of high interest for its qualification as a robust exploration tool and for calibrating risk assessment. In 2009, Statoil evaluated CSEM performance by an internal review study of the CSEM database, where prediction strengths of all cases were measured by the degree of risk impact. The database consists of more than 60 CSEM cases in the period between 2001 and 2008. The database contains 32 cases with well control. In 21 of these cases, CSEM data were available before drilling. Surveys after 2008 were not included in the study.

A calibrated assessment of the risk impact of relevant targets in the database quantified the information strength in the CSEM data. Targets included leads and prospects as well as known discoveries and fields. The assessment was calibrated in the sense that all targets were evaluated after the same criteria, including cases with and without well control, and were ranked and compared to one another. The assessment was based on available information; the time frame of the study did not allow for reprocessing and new analysis of all the cases. Further work might change the assessment, particularly for cases with limited analysis effort. The assessment was based on the current view on the technology for all targets.

Cases were ranked from strongest negative to strongest positive using the same evaluation criteria. To summarize the results, cases were grouped into two classes: the cases with weak or inconclusive results and the cases with clear conclusions, either positive or negative.

The first observation was the high number of weak cases in the first years, and that the ratio of the weak cases relative to the conclusive cases decreases almost uniformly with time. The next observation was that cases with data from 2007 and 2008 were evaluated as conclusive, some positive and some negative. However, few of these recent cases were tested by drilling, and not all cases from that point forward will always give clear conclusions. Additionally, the evaluation was based on the available work, and further analysis could, in principle, change the results. Results reveal a positive trend with time.

Technology development
CSEM exploration has a short commercial history of seven to eight years, but during these years, a development and improvement of elements has been seen from acquisition and processing to data analysis and interpretation techniques. The accuracy of the measurements has improved, and the number of receivers has increased from about 20 in 2003 up to 140 or more per survey layout in 2009. Early surveys mainly were 2-D lines, while 3-D surveys constituted more than 90% of the global CSEM marked in 2009. Processing techniques have moved CSEM from a deepwater method (more than 3,300 ft or 1,000 m water depth) to a technique that today is applicable for 330 ft (100 m) water depth or shallower. The interpretation workflow now includes geological model building, forward modeling, and inversion as standard elements.

To better understand and explain the improvement in performance, three factors in the database review were evaluated by Statoil: the quality of CSEM data, the maturity and thoroughness of the performed analysis, and the feasibility determined by the geological setting. The relation between these components and the prediction strength was investigated.

Statoil found that the overall progress in CSEM data predictability is explained by improvements in all three components. One bad component can give inconclusive results. Alternatively, if, for example, the feasibility component is good, the results can be conclusive even if other components are bad. An example is the CSEM data on the Troll West Gas Province, where the data quality in 2003 and simple normalized magnitude curves were sufficient to obtain a convincing result. For more challenging cases, the development of the CSEM data quality and interpretation methodology has been important in improving the accuracy and robustness of predictions.

The value of CSEM data
The economical value of CSEM data can be predicted for specific exploration settings using standard decision analysis. The economical value is directly coupled to how the data affect decisions.

For a drill-or-drop setting, the CSEM data can have a large value if the data can change a decision from drop to drill based on positive CSEM results, and vice versa. Value generally increases with the prediction strength of the CSEM information, but risk modifications with no decision impact have, by definition, no economical value. Based on performance tracking and review of the prediction strength, the economical value of CSEM data can be more than 10 times above the typical costs for a CSEM survey and analysis. This value includes only the value related to the actual drill-or-drop decision, neglecting the potential value of the data in other settings, for example, mapping the outline of a prospect, decisions on drill location, ranking of prospects, and potential later use as calibration for other cases. Clearly, the real value of the CSEM data can be considerably higher.