Description:

_Journal of Structural Geology 33 (2011) 51e58_ _Contents lists available at ScienceDirect_ _Journal of Structural Geology_ _journal homepage: www.elsevier.com locate jsg_ _Do experts use idealised structural models? Insights from a deepwater foldethrust belt_ _Taija Torvela a,*, Clare E. Bond a,b_ _a Geology & Petroleum Geology, School of Geosciences, University of Aberdeen, Meston Building, King’s College, Aberdeen AB24 3UE, UK b Midland Valley Exploration, 144 West George Street, Glasgow G2 2HG, UK_ _article info_ _Article history: Received 24 June 2010 Received in revised form 27 September 2010 Accepted 11 October 2010 Available online 4 November 2010_ _Keywords: Foldethrust belt Structural model Kinematics Interpretation Seismic reflection Deepwater_ _abstract_ _Theoretical models are often used to aid interpretation of geological data. For foldethrust belts, structural and kinematic models have existed for over a century. While greatly contributing to our understanding of thrust systems, the usage of models can result in oversimplification and false kinematic interpretations. This paper investigates how and if experts use structural models in the interpretation of a seismic image from a deepwater foldethrust belt. The results show that in the majority of cases experts produced interpretations that were compliant with key features in existing structural models. Those interpretations that were less compliant to existing models, better accounted for features present in natural and experimental analogues. This has implications for the general applicability of structural models in interpretation._ _Г“ 2010 Elsevier Ltd. All rights reserved._ _1. Introduction_ _Understanding the evolution of foldethrust structures involves significant interpretation of geological data. Theoretical models have existed for over a century to aid in this (e.g., Willis, 1893; Suppe, 1983; Jamison, 1987; Suppe and Medwedeff, 1990; Erlsev, 1991), but the models define a simplified, mathematically constructible solution for a process that is not, in reality, simply explained. Theoretical models in geology create idealized analogues that can be further used in the interpretation of similar geological systems. This is, in general terms, a very useful approach but may also cause oversimplification of interpretations, especially as the geological system deviates from that for which the model was originally created. The models often fail to explain features observed in many natural foldethrust structures, such as strain localization, fault propagation and fault linkage._ _We present the results of an expert elicitation exercise, in which we have used experts to gather their collective geological interpretation knowledge (in the sense of "the Wisdom of Crowds"; Surowiecki, 2004). Explicit expert elicitation techniques (e.g., Meyer and Booker, 1991; Cooke, 1991) have been used in science, notably within the nuclear waste disposal sector to evaluate interpretational uncertainty and risk (Aspinall, 2010). In our example, rather than asking experts to risk assess their own or others' interpretations, we use the collective interpretations of experts to investigate how theoretical foldethrust models influence the interpretation of seismic data. We use the results to discuss the general usability of established theoretical models in the interpretation of foldethrust structures. This case study uses high quality seismic reflection data from the toe-thrust sector of a gravity-driven deepwater fold and thrust belt, but the conclusions are more generally applicable to the application of models in data interpretation._ _2. Data and experiment_ _2.1. The expert group_ _The exercise was performed at the American Association of Petroleum Geologists Hedberg Research Conference "DeepWater Fold and Thrust Belts" in October 2009. Hedberg Research Conferences are scientific meetings designed to gather scientists from both industry and academia with the aim of discussing state-of-the-art concepts, methodologies, case histories, and future directions relating to the conference subject (http: www.aapg.org education hedberg). Participation is selective and individuals apply, or are invited, to attend ensuring a diversity of key experts in_ _52_ _T. Torvela, C.E. Bond Journal of Structural Geology 33 (2011) 51e58_ _Fig. 1. Pie charts illustrating the level of self-assessed experience of the expert group in a) structural geology, b) interpretation of seismic data, and c) how often they interpret seismic images._ _Fig. 2. The uninterpreted seismic section used in the exercise. The black box indicates the central fold used in the analysis (Fig. 3). The inset shows the location of the seismic 3D volume from which the profile was extracted (Niger Delta)._ _T. Torvela, C.E. Bond Journal of Structural Geology 33 (2011) 51e58_ _53_ _seismic data used in the exercise is excellent, thus minimising the impact of seismic artefacts and noise on the interpretations. The dataset comes from the contractional zone of the gravity-driven Niger Delta deepwater foldethrust belt where a shaly, poorly reflective package lies between the oceanic basement below and a reflective, sand-dominated sediment package with a thickness in the order of several km's on top (Fig. 2; Avbovbo, 1978). The shales represent the pro-delta marine phase of the depositional history (the Akata formation; e.g., Avbovbo, 1978), followed and overlain by the units deposited in the distal part of the delta (the Agbada formation; e.g., Avbovbo, 1978; Deptuck et al., 2003). The still active foldethrust belt formed within the Agbada formation, driven by a gravitational collapse of the delta along major detachments within the overpressurised shale unit (Morley and Guerin, 1996; Briggs et al., 2006)._ _2.3. The experiment_ _Fig. 3. The theoretical models that were compared with the collected interpretations. a) The breakethrust fold model (Willis, 1893); b) The trishear fault propagation fold model (Erslev, 1991); c) The fault-bend fold model (Suppe, 1983); d) The "simple" fault propagation fold model (Suppe and Medwedeff, 1990); e) The detachment fold model (Jamison, 1987). The breakethrust fold redrawn from Willis (1893), the other models redrawn from Shaw et al. (2004)._ _the subject area required to achieve the Hedberg Research Conference aims. In this case, experts in foldethrust belts, possessing a range of knowledge in the associated geological systems and processes (representing a range of disciplines including structural geology, geophysics and seismic interpretation) attended the conference._ _Our expert group included scientists with extensive experience in structural interpretation of foldethrust belts, with authors of heavily cited papers on the topic (h-factors of up to 21; source Scopus, May 2010). The participants were asked to fill out a questionnaire attached to the seismic image, in order to map their background and expertise. 67% of the 24 people who returned an interpretation had more than 5 years of experience after their degree, the average experience of all participants being 12.6 years. 92% and 87% indicated that they were specialists or had a good working knowledge in structural geology and/or seismic interpretation, respectively (Fig. 1a,b). 83% of the group interprets seismic images at least on a monthly basis (Fig. 1c)._ _2.2. The dataset_ _The experiment used a high-resolution 3D seismic dataset, from which a single vertical 2D profile was chosen. The quality of the_ _Participants were asked to complete a paper interpretation of the 2D line using colouring pens and to provide some information about their professional background. No further instructions or information about the location or the stratigraphy were given. The aim of the exercise, unknown to the participants, was to collect a range of interpretations of the seismic image, largely following the principles for capturing the widest possible range of interpretations presented in Bond et al. (2008), and to compare them to the existing theoretical models to see whether the interpreters were influenced, consciously or subconsciously, by the models. The ultimate goal was to see whether the experts produced model-driven interpretations or image-observation-driven interpretations, to compare the interpretations with natural and experimental examples, and to discuss the implications of both interpretational styles in the light of the comparison. A total of 24 interpretations were collected. All interpretations, along with the original seismic image, can be viewed at and downloaded from the Virtual Seismic Atlas VSA (www.seismicatlas.org)._ _2.4. The theoretical model set_ _The suite of structural (kinematic) models used here to define the theoretical model set, against which we have compared the experts' interpretations, are: the breakethrust fold model by Willis (1893), the fault-bend fold model by Suppe (1983), the detachment fold model by Jamison (1987), the "simple" fault propagation fold model by Suppe and Medwedeff (1990), and the trishear fault propagation model of Erlsev (1991). The resultant foldethrust geometries for each of the models are shown in Fig. 3. The five models have been characterised by key features which the authors would expect to see in an interpretation if the interpretation is compliant with a specific theoretical model (Table 1). It is worth noting that most of the theoretical models, bar the detachment fold_ _Table 1 The key features of the five structural end-member models used in the analysis of the interpreted seismic images._ _Model Breakethrust fold Trishear Fault-bend fold Fault propagation fold_ _Fault-detachment linkage Soft-to-hard-linked Hard-linked Hard-linked Hard-linked_ _Detachment fold No faults above_ Ключевые слова: e, r, o