Creating multiple realizations of salt structures in the North Sea using the pseudo distance field approach

Sascha Goerne and Heidrun Stueck and Fabian Jähne-Klingberg and Frithjof Bense. ( 2020 )
in: 2020 RING Meeting, ASGA

Abstract

Knowledge of the exact shape", position in space and dimensions of geological bodies like salt diapirs or igneous intrusive bodies is essential for various use of the sub-surface. Uncertainties in seismic interpretation together with the lack of hard data such as borehole data or field observations as well as multi-z geometries are challenges in the explicit modeling of these structures. In the TUNB project the geological survey organizations (GSOs) of the north German federal states and the BGR jointly develop a 3D structural geological model of the North German Basin (NGB) with SKUA-GOCAD for subsequent Open Access use (Müller et al., 2016). This structural model will incorporate prominent stratigraphic horizons and unconformities as well as faults and salt structures in a regional scale. It will serve as basis for the analyses and evaluation of subsurface use potentials and potential conflicts of use. The construction of the 3D model is based on the evaluation and processing of all relevant data and information from map series," deep wells and reflection seismic data. BGR is focusing on the German North Sea sector with a variety of Zechstein salt structures. The explicit modeling of salt bodies from sparse data in SKUA-GOCAD can be a challenge due to their complex Multi-Z geometries and uncertainties regarding the exact position of the boundaries of the salt body picked from seismic data. The focus of the presented work is both on the time-efficient semi-automatic generation of the """"best fitting"""" geometry and on the creation of multiple realizations of the salt boundary to take uncertainties into account. In order to realize these goals", a macro was written that allows the modeler to obtain results quickly with as little interaction as possible. First, an explicit reference structural model is built by extracting points from seismic picks, filtering this pointcloud using horizontal slices and transforming it to a set of parallel curves that can be used to compute an initial surface. In an iterative process of interpolation and optimization of the control points shooting directions, this surface is optimally adapted to the initial data. Afterwards this reference surface is used to generate an uncertainty envelope by calculating a spatial buffer using a distance property on a grid object. Based on Nicolas Clausolles's approach (Clausolles et al. 2019) we calculate a normalized distance field and multiple normalized random fields (Sequential Gaussian Simulation) from which several equally probable stochastic realizations of the salt bodies are derived. Spatial trends and observed data points can be taken into account by variing the variogram and distribution model as well as interpolation constraints. The concept was tested with several salt domes in the North Sea. The results are promising and show that complex geological bodies can be modelled from sparse data including less certain interpretation data in a short time. In the future we will test this approach within other geological settings for example to build magmatic intrusions.

Download / Links

    BibTeX Reference

    @INPROCEEDINGS{GOERNE_RM2020,
        author = { Goerne, Sascha and Stueck, Heidrun and Jähne-Klingberg, Fabian and Bense, Frithjof },
         title = { Creating multiple realizations of salt structures in the North Sea using the pseudo distance field approach },
     booktitle = { 2020 RING Meeting },
          year = { 2020 },
     publisher = { ASGA },
      abstract = { Knowledge of the exact shape", position in space and dimensions of geological bodies like salt diapirs or igneous intrusive bodies is essential for various use of the sub-surface. Uncertainties in seismic interpretation together with the lack of hard data such as borehole data or field observations as well as multi-z geometries are challenges in the explicit modeling of these structures. In the TUNB project the geological survey organizations (GSOs) of the north German federal states and the BGR jointly develop a 3D structural geological model of the North German Basin (NGB) with SKUA-GOCAD for subsequent Open Access use (Müller et al., 2016). This structural model will incorporate prominent stratigraphic horizons and unconformities as well as faults and salt structures in a regional scale. It will serve as basis for the analyses and evaluation of subsurface use potentials and potential conflicts of use. The construction of the 3D model is based on the evaluation and processing of all relevant data and information from map series," deep wells and reflection seismic data. BGR is focusing on the German North Sea sector with a variety of Zechstein salt structures. The explicit modeling of salt bodies from sparse data in SKUA-GOCAD can be a challenge due to their complex Multi-Z geometries and uncertainties regarding the exact position of the boundaries of the salt body picked from seismic data. The focus of the presented work is both on the time-efficient semi-automatic generation of the """"best fitting"""" geometry and on the creation of multiple realizations of the salt boundary to take uncertainties into account. In order to realize these goals", a macro was written that allows the modeler to obtain results quickly with as little interaction as possible. First, an explicit reference structural model is built by extracting points from seismic picks, filtering this pointcloud using horizontal slices and transforming it to a set of parallel curves that can be used to compute an initial surface. In an iterative process of interpolation and optimization of the control points shooting directions, this surface is optimally adapted to the initial data. Afterwards this reference surface is used to generate an uncertainty envelope by calculating a spatial buffer using a distance property on a grid object. Based on Nicolas Clausolles's approach (Clausolles et al. 2019) we calculate a normalized distance field and multiple normalized random fields (Sequential Gaussian Simulation) from which several equally probable stochastic realizations of the salt bodies are derived. Spatial trends and observed data points can be taken into account by variing the variogram and distribution model as well as interpolation constraints. The concept was tested with several salt domes in the North Sea. The results are promising and show that complex geological bodies can be modelled from sparse data including less certain interpretation data in a short time. In the future we will test this approach within other geological settings for example to build magmatic intrusions. }
    }