PAPUA-NEW-GUINEA MT - LOOKING WHERE SEISMIC IS BLIND

Hydrocarbon exploration in the Papuan fold belt is made extremely diff icult by mountainous terrain, equatorial jungle and thick karstified M iocene limestones at the surface. The high-velocity karstified limesto nes at or near the surface often render the seismic technique useless for imaging the subsurface. In such areas magnetotellurics (MT) provid es a valuable capability for mapping subsurface structure. The main st ructural interface which can be mapped with MT, due to the large elect rical contrast, is the contact between the resistive Darai limestone a nd the underlying conductive sediments of the Ieru Formation. In some areas the base of the Darai can be mapped with reasonable accuracy by fitting 1D models to the observed MT data. However, in many cases wher e 2D and 3D effects are severe, 1D interpretations can yield dramatica lly incorrect results. Numerical and field data examples are presented which demonstrate the severity of the 1D errors and the improvements in accuracy which can be achieved using a 2D inverse solution. Two MT lines over adjacent anticlines, both with well control and seismic dat a, are used to demonstrate the application of 1D and 2D inversions for structural models. In both cases the seismic data provide no aid in t he interpretations. The example over the Hides anticline illustrates a situation where 1D inversion of either TE or TM mode provides essenti ally the same depth to base of Darai as 2D inversion of both TE and TM . Both models provide base Darai depth estimates which are within 10% of that measured in the well. The example over the Angore anticline il lustrates the inadequacy of 1D inversion in structurally complex geolo gy complicated by electrical statics. The TE mode fits a 1D Darai thic kness of 800 metres while the TM mode fits a 1D Darai thickness of 350 0 metres, bracketing the thickness of 2450 metres observed in the well . The final 2D inversion model provides a depth estimate of 2250 metre s. Four MT lines along the Angore anticline have been interpreted usin g 2D inversion. A high degree of correlation exists between lineaments observed on an airborne radar image and zones of low resistivity with in the high-resistivity material interpreted as Darai limestone. These low-resistivity zones are interpreted as fault zones. Three-dimension al modelling has been used to simulate 3D statics in an otherwise 2D e arth. These data were used to test the Groom-Bailey (GB) decomposition for possible benefits in reducing static effects and estimating geoel ectric strike in the Papua New Guinea (PNG) field data. It has been fo und that the GB decomposition can provide improved regional 2D strike estimates in 3D contaminated data. However, in situations such as PNG, where the regional 2D strike is well established and hence can be fix ed, the GB decomposition provides apparent resistivities identical To those simply rotated toi strike.