* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Report of Investigation 195
Survey
Document related concepts
Transcript
GEOLOGICAL SURVEY OF FINLAND Report of Investigation 195 2012 3D modeling of polydeformed and metamorphosed rocks: the old Outokumpu Cu-Co-Zn mine area as a case study Edited by Eevaliisa Laine GEOLOGIAN TUTKIMUSKESKUS GEOLOGICAL SURVEY OF FINLAND Tutkimusraportti 195 Report of Investigation 195 Edited by Eevaliisa Laine 3D MODELING OF POLYDEFORMED AND METAMORPHOSED ROCKS: THE OLD OUTOKUMPU CU-CO-ZN MINE AREA AS A CASE STUDY Front cover: Vuonos ore body and Ni contents. 3D visualization is done by Paradigm GOCAD software package. Unless otherwise indicated, the figures have been prepared by the author of the publication. Layout: Elvi Turtiainen Oy Espoo 2012 Laine, E. (ed.) 2012. 3D modeling of polydeformed and metamorphosed rocks: the old Outokumpu Cu-Co-Zn mine area as a case study. Geological Survey of Finland. Report of Investigation 195, 77 pages, 66 figures and 1 table. The present work summarizes the results of the project “3D/4D modeling – Outokumpu area as a case study” conducted during 2007–2009 and 3D modeling of the Outokumpu assemblage in the project “Development of geological 3D modeling using structural geology and geostatistics” in 2010 at the Geological Survey of Finland (GTK). The aims of these projects were to apply and evaluate 3D/4D modeling software and formulate principles for visualizing and analyzing complicated geological structures typical of Finnish bedrock. The aim was to make recommendations for the use of 3D-modeling tools and processes for 3D/4D modeling at GTK. The 3D/4D modeling project resulted in suggestions for geological 3D modeling methods and processes presented by Eevaliisa Laine and Kerstin Saalmann and geological 3D models at different scales by Esko Koistinen, Kerstin Saalmann, Timo Tervo, Tapio Ruotoistenmäki and Eevaliisa Laine from GTK (Geological Survey of Finland), Gabriel Coirrioux from BRGM (France), Rosa Diaz from the University of Huelva and Noora Salminen from Helsinki University of Technology during 2007–2010, and the regional 3D model constructed by Tapio Ruotoistenmäki and Esko Koistinen using regional geophysical data. In addition, Noora Salminen completed her Masters thesis, entitled Microstructures in 3D modelling, Outokumpu area as an example site, for the Department of Civil and Environmental Engineering at Helsinki University of Technology. Keywords (GeoRef Thesaurus, AGI): bedrock, metamorphic rocks, structural geology, metal ores, ore bodies, three-dimensional models, Proterozoic, Outokumpu, Finland Eevaliisa Laine Geological Survey of Finland P.O. Box 96 FI-02151 Espoo FINLAND E-mail: [email protected] ISBN 978-952-217-182-5 (PDF) ISSN 0781-4240 Laine, E. (toim.) 2012. 3D modeling of polydeformed and metamorphosed rocks: the old Outokumpu Cu-Co-Zn mine area as a case study. Geologian tutkimuskeskus, Tutkimusraportti 195, 77 sivua, 66 kuvaa ja 1 taulukko. Raportti esittelee 3D/4D mallinnusmenetelmät – Outokummun alue esimerkkinä (2007–2009) ja sitä seuranneen geologisen mallinnuksen kehittäminen (2010) -hankkeiden tulokset. Hankkeiden tavoitteena oli soveltaa ja arvioida 3D/4D mallinnusohjelmistoja ja esittää Suomen kallioperälle tyypillisten monimutkaisten rakenteiden mallinnuksen periaatteita. Lisäksi tavoitteena oli antaa suosituksia 3D-mallinnuksen tekemisestä GTK:ssa. Hankkeiden tuloksena syntyi suositus Suomen kallioperän muodostumien 3D- mallinnusprosessista, eri mittakaavaisia Outokummun alueen malmien ja kivilajimuodostumien 3D-malleja sekä Noora Salmisen diplomityö mikrorakenteista. Asiasanat (Geosanasto, GTK): kallioperä, metamorfiset kivet, rakennegeologia, metallimalmit, malmiot, kolmiulotteiset mallit, proterotsooinen, Outokumpu, Suomi Eevaliisa Laine Geologian tutkimuskeskus PL 96 02151 Espoo Sähköposti: [email protected] CONTENTS Editor’s Preface by Eevaliisa Laine..........................................................................................5 Problems and challenges of 3D modeling of the Precambrian of Finland by Eevaliisa Laine and Kerstin Saalmann.................................................................................7 1 Introduction................................................................................................................. 8 2 3D geological models.................................................................................................. 10 3 Data used for 3D modeling......................................................................................... 12 4 Building a 3D geological model ................................................................................. 14 5 3D geological software packages................................................................................ 26 6 3D geological modeling process.................................................................................. 26 7 Practical considerations.............................................................................................. 28 8 Conclusions................................................................................................................ 29 Supplementary information on the web............................................................................ 29 References......................................................................................................................... 29 3D modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area by Eevaliisa Laine, Esko Koistinen, Kerstin Saalmann, Gabriel Couirrioux, Rosa Diaz, Noora Salminen and Timo Tervo............................................................................................ 31 1 Introduction............................................................................................................... 32 2 Geology of the Outokumpu region............................................................................. 33 3 Rock walls and microstructures in 3D........................................................................ 35 4 Ore models................................................................................................................. 43 5 3D modeling of the Outokumpu assemblage.............................................................. 52 6 3D modeling of the Sola serpentinites........................................................................ 60 7 3D model of the Vuonos area including detailed seismic FIRE-sections ................... 62 Supplementary information on the web ............................................................................ 66 References......................................................................................................................... 66 3D-pdf’s............................................................................................................................ 66 Three-dimensional modeling of the Outokumpu nappe area, SE Finland by Tapio Ruotoistenmäki and Esko Koistinen......................................................................... 67 1 Introduction............................................................................................................... 68 2 General geology of the study area.............................................................................. 68 3 Major geophysical formations in the study area.......................................................... 69 4 Major structures of the Outokumpu nappe area........................................................ 70 5 Preparation of the three-dimensional model............................................................... 72 6 Conclusions................................................................................................................ 74 References......................................................................................................................... 74 Acknowledgements........................................................................................................... 75 4 3D modeling of polydeformed and metamorphosed rocks: the old Outokumpu Cu-Co-Zn mine area as a case study Edited by Eevaliisa Laine Geological Survey of Finland, Report of Investigation 195, 2012 EDITOR’S PREFACE The present work summarizes the results of the project “3D/4D modeling – Outokumpu area as a case study” conducted during 2007–2009 and 3D modeling of the Outokumpu assemblage in the project “Development of geological 3D modeling using structural geology and geostatistics” in 2010 at the Geological Survey of Finland (GTK). The aims of these projects were to apply and evaluate 3D/4D modeling software and formulate principles for visualizing and analyzing complicated geological structures typical of Finnish bedrock. The aim was to make recommendations for the use of 3D-modeling tools and processes for 3D/4D modeling at GTK. Geological 3D modeling combines 3D visualization with the interpretation and interpolation of subsurface lithological and tectonic structures and their relation to each other. It differs from 2-dimensional mapping in the smaller amount of data that are often unevenly distributed and often restricted to clustered drill holes with large gaps between drilled areas. In cases of very sparse data sets, a subsurface model is more or less a visualization of present ideas of the subsurface. 4D modeling means the visualization of changes in rock formations as a function of time in 3D. The important task is to restore the original rock structures before deformation. The present report concentrates on geological 3D modeling, even though this is not possible without understanding the geological processes as a function of time. The tools and methods were tested by building geological 3D models of targets within the Outokumpu area by combining geological and geophysical data, applying geomathematical methods (e.g. statistics) and by using different geological software packages (Fig. 1). The Outokumpu area was chosen because it includes several old ore deposits that have been studied and mined since 1913. An important data source was the GEOMEX project from1998 to 2003 (http://vsa.gtk.fi/geomex/html/ source/start.html), which investigated the geological character, evolution and metallogenic potential of the Outokumpu region in Northern Karelia. 3D/4D project researchers came from GTK offices in Espoo, Kuopio and Rovaniemi. Juhani Ojala and Nicole Patison from Rovaniemi, Jouni Lerssi, Timo Tervo and Peter Sorjonen-Ward from Kuopio and Esko Koistinen, Kerstin Saalmann, Tapio Ruotoistenmäki and Eevaliisa Laine from Espoo worked in the project workshops and 3D modeling. Gabriel Coirrioux from BRGM (France) guided us in the use of GeoModeller software. Noora Salminen completed her Master’s Thesis on microstructures at the Horsmanaho and Sola open pits at Helsinki University of Technology. Gyorgy Havas from the Hungarian Geological Survey and Anneli Lindh from GTK Espoo office constructed the elevation model of the Outokumpu area. Finally, 5 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Eevaliisa Laine Rosa Diaz from Huelva University (Spain) modeled the Sola serpentinite bodies using Surpac mining software. Nils Gustavsson from the GTK Espoo office acted as a geomathematical expert by providing many useful ideas for the project. The geological modeling software included Gocad® by Paradigm, GeoModeller by Intrepid Geosciences and ISATIS from Geovariances, Surpac and Gems by Gemcom and Sirovision developed by CSIRO. These represent the main types of geological 3D modeling software among several similar software packages for geological 3D modeling. In addition general purpose softwares, such as Matlab (MathWorks), were used in 3D modeling. The main emphasis was on 3D modeling processes – not on the software. The 3D/4D modeling project resulted in suggestions for geological 3D modeling methods and processes presented by Eevaliisa Laine and Kerstin Saalmann and geological 3D models at different scales by Esko Koistinen, Kerstin Saalmann, Timo Tervo, Tapio Ruotoistenmäki and Eevaliisa Laine from GTK (Geological Survey of Finland), Gabriel Coirrioux from BRGM (France), Rosa Diaz from the University of Huelva and Noora Salminen from Helsinki University of Technology during 2007–2010, and the regional 3D model constructed by Tapio Ruotoistenmäki and Esko Koistinen using regional geophysical data. In addition, Noora Salminen completed her Master’s thesis, entitled Microstructures in 3D modelling, Outokumpu area as an example site, for the Department 3D of Geological Civil and Modeling Environmental Engineering at Helsinki University of Technology. Outokumpu Cu-Ni-Co mining area as a case study DATA GEOMEX project Other geological and geophysical data from the area Geological theories and hypothesis for the origin of the Outokumpu assemblage METHODS 3D software Geological inference Geomathematical methods 3D modeling workshops RESULTS A suggestion for a work flow for 3D modeling of Precambrian bedrock in Finland Geological 3D models at different scales based on geological and geophysical data from the Outokumpu area Discussion of problems and challenges in 3D geological modeling Fig. 1. 3D/4D modeling project, Outokumpu area as a case study. 6 3D modeling of polydeformed and metamorphosed rocks: the old Outokumpu Cu-Co-Zn mine area as a case study Edited by Eevaliisa Laine Geological Survey of Finland, Report of Investigation 195, 2012 Problems and chaLlenges of 3D modeling of the Precambrian of Finland by Eevaliisa Laine* and Kerstin Saalmann Laine, E. & Saalmann, K. 2012. Problems and challenges of 3D modeling of the Precambrian of Finland. Geological Survey of Finland. Report of Investigation 195, 17 figures. Geological 3D modeling of complex geological structures has many challenges beginning with data organization. Because the structures and lithologies have complex relationships with each other it is very difficult to predict the continuation of any structure to great depths, especially because there are practically no direct geological observations below 500 m. Large regional 3D models should be based on geological expertise and indirect geophysical data. Interpretations are sometimes difficult, even at a smaller scale, because of the locally high variability of lithologies. The use of different approaches and software is needed in a comprehensive data capture and 3D visualization of complex geology, such as present at the Outokumpu area. It is also evident that purely technical visualization without involving geological ideas and interpretations is impossible. Keywords (GeoRef Thesaurus, AGI): bedrock, structural geology, topography, three-dimensional models, data processing, Precambrian, Outokumpu, Finland *Eevaliisa Laine Geological Survey of Finland, P.O. Box 96, FI-02151 Espoo, Finland *E-mail: [email protected] 7 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Eevaliisa Laine and Kerstin Saalmann 1 Introduction A geological map displays geological features in a plane. Rock units are shown by color or symbols to indicate where they are exposed at the surface. Actually, a map is a 2D illustration of a 3D geometry. Structural features such as faults, folds and foliations are shown with strike and dip. These tectonic symbols in addition to elevation lines give the three-dimensional orientations of the features. In general, geological maps also show elevation lines, allowing the reconstruction of the 3D geometry. For this reason geological maps have been considered as 2½ dimensional. Hence, a geological map includes 3D information, and an experienced geologist is able to see the third dimension through a planar map. Vertical cross sections are often drawn in order to visualize geology in the third dimension. Because of recent developments in 3D modeling software, it is nowadays possible to visualize geological subsurface structures and construct solid models for further computation and modeling. Geological 3D modeling combines 3D visualization with interpretation and interpolation of subsurface lithological and tectonic structures and their relation to each other. It differs from 2D mapping in terms of limited data that is often unevenly distributed and often restricted to clustered drill holes with large gaps between drilled areas. In cases of very sparse data sets, the subsurface model is more or less a visualization of present ideas concerning the subsurface. These models may be improved by geophysical forward modeling and inversion. Despite these uncertainties, these models are important in many applications such as modeling of ore bodies, ore evaluation, hydrocarbon exploration and engineering applications. The models can be used for the calculation and simulation of geological processes such as fluid flow and thermal convection and conduction. It is also important to visualize 3D geology to aid research and present interesting geological sites to the public. Geological 3D models have to take into account geological structures and their contact relations. In addition, an important task is to be able to handle, measure and present the uncertainties associated with geological 3D models. The uncertainties depend on the amount, distribution and quality of data, bedrock heterogeneity and the accuracy of local geological interpretations. The uncertainty, especially related to geological data from Finland, has been discussed in depth by Nils Gustavsson (2010). The 3D visualization of uncertainties has recently been discussed, for exam- 8 ple, by Viard et al. (2011). Even though this is not the topic of the present report, the uncertainty will be touched on when different aspects of 3D modeling are discussed. During recent years there has been strong development of geological 3D modeling software. In addition, many books and articles have been written about the principles of geological 3D modeling and new concepts have been defined or suggested (e.g. Caumon et al. 2009; Kessler et al. 2009; Fernández et al. 2004; Gjøystdal et al. 1985; Saksa 1995; Groshong, 2006; Mallet 1997, 2002; Calcagno et al. 2008). In his book Geomodeling, Mallet (2002) proposed in his book Geomodelling the notion of geomodeling that emphasizes the specific nature of 3D modeling: “Geomodeling consists of the set of all the mathematical methods allowing to model in an unified way the topology, the geometry and the physical properties of geological objects while taking into account any type of data related to these objects.” The first 3D geological models were drawings and block models. In Finland,the most famous block models have been done by Adolf Metzger (e.g. Metzger, 1928). In addition, fence diagrams and sometimes small sculptures of ore bodies (e.g. a bronze model of the Keretti deposit shown in Fig. 11 by Laine et al. 2012 in this volume) have been used to visual subsurface structures before computerized 3D modeling. The first Finnish software for 3D geological modeling (Minenet) was developed by Outokumpu Oy in the 1980s (e.g. Brusila, 1983). In the 1990s it was replaced by mining software packages such as Surpac and Gems. The second Finnish software package is RockCAD that was developed by consulting and engineering company Pöyry in the 1990s for the nuclear waste studies (Saksa, 1995). RockCAD is a modelling system and 3D software tool for geological and environmental subsurface studies. It runs in the AutoCAD environment, one of most widely used CAD programs in the world. RockCAD utilizes a project-based approach. The sofware runs in standard Windows PC environment. RockCAD contains useful functions that allow the construction of 3D earth models and presentation of ground investigation data. Despite its good qualities, the software is not generally used in Finland. The AutoCAD software has been also used directly in 3D modeling of structural and metamorphic patterns by Timo Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Problems and challenges of 3D modeling of the Precambrian of Finland Kilpeläinen (1998). In Finland, the main applications of the 3D modeling have been related to ore deposit modeling (e.g. Parkkinen 2003), groundwater modeling (e.g. Luoma et al. 2010), nuclear waste site studies (Mattila et al. 2007) and, in general, geoenvironmental studies. Worldwide, 3D modeling is widely used in hydrocarbon exploration in sedimentary basins with a well known stratigraphy that is only weakly de- A) formed (mainly by faulting). Many 3D geological software packages have been developed for this purpose. Modeling of highly metamorphosed rocks is challenging because of the complex deformation and lack of established stratigraphic relationships. In the following we provide examples of Finnish bedrock modeling and discuss the challenges and possible solutions. The examples are from the Outokumpu region (Fig. 1). This region is located in the North Karelia Schist Belt, which was thrust on the late Archaean gneissic–granitoid basement of the Karelian craton during the early stages of the Svecofennian Orogeny between 1.92 and 1.87 Ga (Koistinen 1981). B) Fig. 1. A) Location of the Outokumpu area; B) Regional geology after Korsman et al. (1997). 9 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Eevaliisa Laine and Kerstin Saalmann 2 3D geological models 3D geological models differ in scale, application, use of geometry and applied interpolation methods. Regional 3D geological models cover areas of tens of kilometers. In very large 3D models, the vertical extent becomes small relative to horizontal dimensions, and in the visualization the vertical dimension has to be exaggerated. Depending on the thickness of the investigated sections, ranging from less than one to 60 km, the z-axis representing depth may become too small as compared to the horizontal dimensions when the horizontal scales are much larger. A value of 100 km may be kept as an upper limit for the horizontal extent, even for crustal-scale modeling, because the depth information otherwise becomes lost. There is also very little direct geological observation below 500 meters. In the case of a regional model, the geological structures are based on geological sections, potential fields and seismic surveys. Here, geological ideas and theories concerning the regional geology play an important role, for example in the case of the regional 3D model of the Outokumpu area, with a size of 120 km x120 km x 40 km (Fig. 2A), prepared by Ruotoistenmäki & Koistinen (2012) in this volume. In ore deposit and geoenvironmental studies, such as nuclear waste site studies, abundant drillhole data or other data are available and the 3D model may be data driven. Nevertheless, there is a need for interpretation. The 3D model of the Outokumpu assemblage, including the surrounding black schists (Fig. 2B), is based on vertical sections interpreted by Koistinen (Gaál et al. 1975) and drill-hole data. The smallest scale structures used in the project are 3D visualizations of rock textures and microstructures. The 3D visualization in these cases depends on technical tools that are applied in photographing. 3D modeling of microstructures of rock samples from the Horsmanaho open pit in 10 the Outokumpu area was performed by studying oriented thin sections cut differently relative to the tectonic directions under a polarizing microscope (Salminen 2009). The main difficulty was in the relative positioning of oriented thin sections. Geological 3D modeling simplifies the structures according to the scale. Lithological contacts are modeled by smooth surfaces, even if the boundary between different rock types is often gradational and usually very uneven because of minor folding and primary irregularities of geological formations. Faults are also modeled as surfaces of constant thickness, disregarding the fact that their width can change from meters to centimeters. A good example is the near vertical Kaasila fault cutting the Outokumpu ore (Koistinen 1981), the width of which varies from 10 cm to a few meters (pers. comm. T. Koistinen). Rock properties such as geochemical composition and petrophysical properties are usually presented as colored symbols in 3D and interpolated into a 3D grid. The most common approach is to use a rectangular grid and some mathematical interpolation such as the inverse distance method or geostatistical kriging interpolation, which is based on the stochastic model for spatial correlation (e.g. Chilés & Delfiner 1999). The latter method provides a possibility to study anisotropy structures and a kriging variance, which is a measure of variability and thus, a type of measure of the uncertainty (e.g. Rocha & Yamamoto 2000). The rectangular grid is not adequate for undulating or folded natural formations. A better approach is to use a stratigraphic grid defined by undulating layers. However, in Precambrian rocks the local orientations are so variable that it is not always possible to build a stratigraphic grid. The use of local orientations selected from the geological and geophysical data (Laine 1998) could be an alternative in the interpolation of rock properties. Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Problems and challenges of 3D modeling of the Precambrian of Finland A) B) Fig. 2. A) Regional model of the Outokumpu nappe based on geophysical data and interpretation of Tapio Ruotoistenmäki (Ruotoistenmäki & Koistinen 2012, in this volume). The width of the nappe formation is about 40 km B) Geological model of Outokumpu assemblage, including surrounding black schists, based on drill-hole data and several geological sections (Laine et al. 2012, in this volume). Surrounded mica gneisses are not included. The width of the Outokumpu formation is about 1.5 km. 11 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Eevaliisa Laine and Kerstin Saalmann 3 Data used for 3D modeling Data for 3D geological modeling are derived from various sources, including geological maps and cross sections, geological outcrop observations, drill-hole logs, geophysical maps and surveys. The obvious problems related to data from several sources are their different resolutions and scales. Drill-hole data provide information on changes in properties within a few centimeters, while large kilometer-scale structures are mainly interpreted from deep seismic sections. Geological maps and sections are interpretations of outcrop observations and geophysical maps. Thus, they do not present raw data, and different interpretations lead to different kinds of 3D geological models. On the other hand, the 3D modeling process may lead to new interpretations of the stratigraphic relationships, which may in turn change the geological map. This fact illustrates the difficulties when separating 2D and 3D geological databases. A better approach would be to construct two databases, one containing raw data (1D, 2D, 3D) and the other storing geological models that result from the interpretation and usage of raw data and which will be updated according to the availability of new data and geological theories. Outcrop data may be regarded as geological raw data, even though every geological observation includes interpretation. Geological observations also reflect the time during which they have been recorded. Drill-hole logs contain detailed information on rock types as well as geochemical and geophysical properties. Naturally, the form of data collected depends on the final purpose of the drilling. Lithologies are named in much more detail than found on the corresponding geological maps. Old drill-hole logs may also include rock types based on quite different interpretation from present classifications. The process of unifying rock names aiming to define rock units that can be modeled at the desired scale may not be an easy task. For example, the GEOMEX project (http:// vsa.gtk.fi/geomex/html/source/start.html) performed significant work in unifying the lithologies in the Outokumpu area according to the new interpretation of the alteration of the serpentinite body (Peltonen et al. 2008). The definition of rock units is accompanied by the selection and drawing of the main discontinuity structures such as faults and shear zones interpreted from geophysical maps and sections, elevation data and outcrops. Understanding of the 12 kinematics of the fracture system and the relative ages of fractures is needed in 3D modeling in order to properly draw lithologies between discontinuity surfaces. Geophysical data are used in many stages of 3D geological modeling, for example: 1.Geological maps are based on geophysical data in areas covered by Quaternary deposits. Lithologies observed from the outcrops are connected by using geophysical maps. For example, the geological map of the Outokumpu area (Fig. 3) has been drawn by using geophysical data especially magnetic low-altitude maps (GTK). 2.Geological discontinuities such as faults and shear zones are often interpreted from the geophysical data, and the nature of the discontinuity structure is determined by geological expertise. 3.Geological sections may be prepared by geophysical 2D inversion and used in 3D modeling together with geological sections. 4.The 3D geological models obtained may be improved by using 3D inversion (e.g., Scales et al. 2001) 5.3D inversion may also be used in analyzing the subsurface distribution of the petrophysical properties in order to build a 3D geological model together with surface observations. 6.In addition, sophisticated geophysical map processing has been used in order to understand bedrock structures in covered areas in Finland. One example is the study of black shales in Finnish bedrock by Airo et al. (2009). The above is not a complete list of available data, but it already gives an impression of the variety of data sources, and the data list shows the need to involve different experts in the 3D geological modeling process. The visualization of all the data in 3D may already help in understanding the subsurface structures. However, it can very often be confusing to consider too many different data sets at the same time. One of the major future challenges is to find working practices for handling and interpreting various 3D data sets. It is important to make a difference between measured data such as geophysical and geochemical data and interpretations such as geological maps and sections. Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Problems and challenges of 3D modeling of the Precambrian of Finland Fig. 3. The magnetic low-altitude map of Outokumpu area processed by Timo Tervo (GTK) (red mark the highest values and blue the lowest ones) and geologic map of the Outokumpu area compiled by T. Koistinen (Gáal et al. 1975). 13 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Eevaliisa Laine and Kerstin Saalmann 4 Building a 3D geological model Subsurface geology is often visualized or presented using cross sections. In the Outokumpu area cross sections were drawn across the Outokumpu assemblage rocks (Koistinen 1981, Fig. 3) . Data available from the drill-holes, as in the case of Outokumpu sections, help to contrain the subsurface geology in a cross section. Nevertheless, there is room for interpretation, especially when interpreting heterogenous and multiply deformed midcrustal rocks. In addition, the bedrock of the Outokumpu area is covered by Quaternary deposits, so there are very few direct observations of the bedrock surface and the geological map is strongly based on geophysical data. This is not an uncommon situation in Finland because, in most places, the bedrock is covered with the Quaternary deposits. Because folding, shearing and faulting strongly affect the midcrustal bedrock understanding of tectono-magmatic and metamorphic evolution is essential in drawing cross sections. In Figure 3 the Outokumpu sections have been presented together with the geological map. This is one option to visualize geology in 3D, and, is known as a fence diagram. Before computer applications, 3D visualizations were carried out using a block diagram that illustrates the relationship between the surface and subsurface structure by using an imaginary block cut out of the earth’s crust. The top of the block shows the areal geological map and the vertical sides represent geological sections. A block diagram presents a three-dimensional picture that helps to understand 3D structures. Figure 5 presents a block diagram of the Outokumpu assemblage from the Sola area. The applied projection is isometric, in which all of the distances are equal to the measurements along the corresponding directions in the maps and cross sections. In addition to different kinds of geometrical drawings and constructions (see eg. Groshong 2006), geologists like to make sketches in 3D. As an example, a schematic diagram of the structural position of the Vuonos orebody by Koistinen (1981) is presented in Figure 6. Diagrams of this kind are still important in illustrating geological ideas and theories. One of the challenges in computerized 3D geological modeling is to transfer these ideas into the digital world. Fig. 4. A Fence diagram of the Outokumpu assemblage after the geological map and cross sections by Koistinen (1981). Reproduced by permission of the Royal Society of Edinburgh from Transactions of the Royal Society of Edinburgh: Earth Sciences volume 72 (1981), 115–158. 14 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Problems and challenges of 3D modeling of the Precambrian of Finland Fig. 5. A block diagram of the Sola serpentinite massif and surroundings (Gaál et al. 1975). The Sola Cu-Co mineralization is shown by red spots. (Kontinen et al. 2006.) Fig. 6. A schematic diagram of the structural position of the Vuonos orebody (after Koistinen 1981). F1 and F2 are folding generations and -2—marks the scistocity associated with the F2-folding. Reproduced by permission of the Royal Society of Edinburgh from Transactions of the Royal Society of Edinburgh: Earth Sciences volume 72 (1981), 115–158. 15 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Eevaliisa Laine and Kerstin Saalmann Description of the geometries of the geological formations 3D modeling requires the description of the geologic formations and structures and their topological relationships. Some of the important features to be included into the 3D models are listed below by using examples from the Outokumpu Ore District, a metallogenic province about 100 km long x 60 km wide that is defined by the occurrence of bodies of massive sulfide and minor disseminations. • The topography gives the upper surface for the 3D model The topography of the Outokumpu area is rather flat, and an elevation model has been constructed by the Land Survey of Finland. In addition, elevation lines and water depths were used in order to build a more detailed elevation model for the Outokumpu area. This represents the upper surface of the Quaternary deposits, which cover over 80% of the land surface in Finland. In near future, the whole of Finland will be measured using laser scanning by the Land Survey of Finland in order to provide very detail elevation models. • Bedrock topography gives the upper surface for the bedrock, that is difficult to determine in areas with thick post glacial layers, as is case in large areas of Finland. In the Outokumpu area, drill-hole data can be used in estimating the thickness of the Quaternary deposits, and it is therefore possible to reconstruct the bedrock surface. In areas where there is no drilling, the bedrock surface may be defined by using geophysical methods such as gravity, seismic refraction and ground penetrating radar. This surface is erosional and cuts all the older bedrock structures. • Faulting and jointing is very common in the Precambrian bedrock in Finland. Locally it has a pronounced effect on the bedrock structures, and fault movements of several meters (e.g. the Kaasila fault) affect the ore continuity and, in turn ore exploration. Rock fracturing at different scales should be carefully modeled, preferably as the first step in 3D modeling. This may be very difficult in areas with only a few outcrops. Well-documented studies of rock fracturing exist from southern Finland (e.g. Pajunen et. al. 2008; Viola et al. 2011). The most difficult task is to trace these often undulating and variably thick structures to greater depths. • Finally, the lithological units may be lineated. If the lithological contacts are modeled by surfaces, the rock type classification need to be done so that it is possible separate different rock units by smooth surfaces representing lithological contacts. The Outokumpu assemblage that hosts the ore is separated from the surrounding mica schists and gneisses by black schist. On a large scale it is possible to model the combination of the Outokumpu assemblage and black schists as a single unit in the surrounding schists. Simplification is necessary because a 3D model using more detailed rock classification is almost impossible to construct as both host rocks and ores show the effects of polyphase deformation and polymetamorphism (Koistinen 1981). The above-described structural features and lithological contacts are modeled by surfaces in 3D modeling software. 3D modeling in 3D geomodeling software usually begins with the digitizing of geological features from geological and geophysical maps and cross sections (Fig. 7). 3D models are then built using these digitized points and lines, together with drill-hole data. Digitized points and lines are connected into surfaces representing lithological contacts, faults and shear zones using tectonic observations and taking into account the relative ages of formations. Geological understanding of the structural geology is essential because lithologies may be connected in many different ways based only on the drill-hole data (Fig. 8). In this stage the main interpretation is carried out. Every time a point is added outside the known objects (points, drill holes), an interpretation is made. There are several approaches to connecting the lines and points into surfaces in 3D. The main approaches applied in 3D geological modeling are described below in order to present the important differences that affect the whole 3D geological modeling process. Triangulation The most common method to build geological boundaries is triangulation. Triangulation is the division of a surface or plane polygon into a set of triangles, usually with the restriction that each triangle side is entirely shared by two adjacent triangles. There are numerous algorithms for com16 puting triangulations in 2D and 3D. Delaunay triangulation (Jones & Nelson 1992) is often used because it ensures that the circumcircle associated with each triangle contains no other point in its interior. Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Problems and challenges of 3D modeling of the Precambrian of Finland A) B) Fig. 7. A) Drill holes and B) the corresponding geological section interpretation (Kontinen et al. 2006) from Sola. 17 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Eevaliisa Laine and Kerstin Saalmann Fig. 8. Many different possibilities for section interpretation. Triangulation honors the known data values. Accordingly, the resulting model strongly depends on the amount of available data. Because the data points are often irregularly sampled, direct triangulation seldom produces a good quality mesh (Fig. 9). Geological interpretation of differently oriented cross sections provides more data and results in a more realistic 3D model on the assumption that the underlying geological model is meaningful. Instead of using digitized points, it is easier to model complicated structures by using digitized lines, as illustrated in Figure 8. Elongated ore bodies and faulted ore bodies such as the Outokumpu ore have been modeled by connecting closed curves enclosing the ore in sections using triangulation (Fig. 10). Fig. 9. The bedrock surface was constructed from the elevation points extracted from drill-hole data by triangulation. 18 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Problems and challenges of 3D modeling of the Precambrian of Finland Fig. 10. The thin Outokumpu ore bodies modeled by traditional section-by-section interpretation using geological sections. This seems to be the best way to delineate these thin and irregularly shaped orebodies. Modelling was performed by Esko Koistinen using Gemcom GEMS software. Interpolation of 3D surfaces using mathematical tools Different interpolation techniques such as inverse distance and kriging interpolation are often applied. These interpolation methods estimate the elevation z as a function of geographic coordinates (x and y). The interpolation usually consists of 1. the building of a regular grid according to the study area; 2. in the case of kriging interpolation, variogram modeling is needed in order to build a spatial statistical model for spatial correlation; and finally 3. calculation of the estimated values in a regular grid. This type of interpolation works with surfaces without overturned folding or faulting. There can be only one z-value for the same location. For example, the Outokumpu bedrock surface can be interpolated by using inverse distance (Fig. 11). The SE–NW trends possibly reflect the fracturing in the same orientation. Sometimes gridding shows better features that are not obvious on the triangulated presentation. 19 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Eevaliisa Laine and Kerstin Saalmann Fig. 11. The Outokumpu bedrock surface interpolated using inverse distance method and drill-hole data by ISATIS software. Indirect surface construction Caumon et al. (2009) have suggested one way to avoid the limitations of direct triangulation methods using Paradigm GOCAD® software and the following steps: 1. The surface boundary is first determined. 2. This boundary is then densified and filtered so that all the edges of the polygon have roughly the same length. 3. A surface is triangulated using this boundary. 4. This surface is fitted to the data points using discrete smooth interpolation. DSI optimizes all three spatial coordinates of mesh vertices under a large set of constraints (Mallet 1992, 20 1997, 2002). Briefly, DSI solves the optimal location of the surface nodes to minimize a weighted sum of the surface roughness and the constraint misfit. One shear zone within the Outokumpu assemblage was modeled by using indirect surface construction in Paradigm GOCAD® software (Fig. 12). The shear zone was interpreted by Koistinen (1981). These interpretations were digitized as lines from cross sections and these lines were used to build a surface representing the shear zone that cuts the Outokumpu Assemblage. Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Problems and challenges of 3D modeling of the Precambrian of Finland Fig. 12. Indirect surface construction by using GOCAD®software. A) a median plane for the digitized lines; B) indirect surface construction of the shear zone; C) the obtained shear zone (in white) visualized together with Vuonos ore body (in red), a boundary for the Outokumpu Assemblage (green) and a geological cross section reproduced by permission of the Royal Society of Edinburgh from Transactions of the Royal Society of Edinburgh: Earth Sciences volume 72 (1981), 115-158. 21 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Eevaliisa Laine and Kerstin Saalmann Implicit representations of surfaces Explicit representation means the previously described triangulation and other interpolation methods for estimating z=f(x,y) in building surfaces. The implicit representation is done by isosurfaces of a 3D scalar field (f(x,y,z)=constant). Carr et al. 2001) proposed a method by computing the parameters of Radial Basis Function for interpolating discrete data and reconstruct compex shapes. Ledez (2002) used Euclidean Distance Transforms computed on a Cartesian grid for reconstructing complex geometries. Discrete Smooth Interpolation (Mallet, 1992 and 2002) has also been used in building the scalar field on a discrete volumetric mesh (e.g. Frank et al. 2007). As an example of the use of implicit representation, the application of co-kriging in computing the scalar field (e.g. Calcagno et al. 2008) is presented. Calcagno et al. (2008) call the approach a potential-field interpolation method. The Sola sernetinite contact with the surrounded mica schist was modeled by using GeoModeller software (Fig. 13). In the case of sparse geological data, 3D models may be constructed by using lithological con- tact points and orientation data using a potential-field interpolation method by Calgacno et al. (2008) based on the earlier work of Lajaunie et al. (1997). The method is based on the following assumptions: 1. a geological interface limits two geological formations; 2. some orientation data sampled within geological formations are relevant to model the interfaces separating the formations; 3. the interfaces to be modeled may be regarded as belonging to an infinite set of surfaces aligned with the orientation field. Calgagno et al. (2008) described this interpolation method as being close to classical geological thinking in the sense that it attempts to reproduce the natural drawing of a geologist simultaneously guided by the knowledge of structural data. Structural data are used in the interpolation or building of the surfaces representing lithological contacts. This interpolation is performed by cokriging (null) increment data (i.e. the change in potential between any two points belonging to the Fig. 13. Sola serpentinite and enveloping black schist by using Geomodeller software. 22 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Problems and challenges of 3D modeling of the Precambrian of Finland same geologic interface is null) and their derivatives (orientation data representing the gradient or derivative of the field) according to Lajaunie et al. (1997). In the case where nearby layers share a geological history, combining the layers into a single series has the major advantage that data from one horizon can influence the shape of other nearby horizons, and vice versa. Measurements of strike and dip recorded anywhere within the series will all be taken into account at their point of measurement. In cases where the geological history is more complex, and geological horizons are not subparallel, separate potential interpolators must be used – one for each series of strata. In this case it is necessary to define the stratigraphic column, which records the chronological order of the strata, and also the relationships (either ‘onlap’ or ‘erode’). Where two geological surfaces from different potential interpolators intersect, an ‘erode’ surface cuts across any stratigraphically older horizons, whereas an ‘onlap’ surface would ‘stop’ against the older surface. The use of lithological contacts and stratigraphic layering is challenging in the case of midcrustal rocks, because lithological boundaries are not clear and often gradational. The use of any stratigraphy is also difficult because of shearing and faulting, which may be intense and difficult to model in 3D. Partial or total melting also makes it difficult to follow any particular horizon very far. In the Outokumpu area, the clear erode surface is the boundary between Quaternary deposits and the bedrock. All the other surfaces are ‘onlap’. It is not easy to decide when we can apply series and when to separate them. As an example, Sola serpentinite was modeled using this approach in GeoModeller software (Fig. 12). The black schist enveloping the serpentinite body could not be modeled separately but was included into serpentinite. This interpretation differs from the the one presented using the block model (Fig. 5). Other methods for surface construction There are several other methods for surface construction. For example, de Kemp & Sprague 2003 have presented a method to use sparse structural data in surface construction. Presently this tool is available as a SPARSE module in Paradigm Gocad®. The choice of particular surface construction method depends on geometries of geological formations and on the data density. In addition to lithological contacts, it is important to be able to model discontinuities and folding. 3D modeling of discontinuity surfaces faults and shear zones Faults and shear zones can be presented or visualized as surfaces, even though they are not necessarily continuous and their properties such as thickness and structure vary. This makes it difficult to connect them from drill hole to drill hole or between seismic sections (e.g. Markovaara-Koivisto et al. 2009). The relative ages of faults and shear zones are necessary but sometimes quite difficult to determine because of the scarcity of outcrops. Many of the 3D modeling software packages are capable of handling faults and shear zones by: 1. presenting a fault or a shear zone in 3D and cutting the lithological contacts with it; 2. using the information about which geological formations are cut by the fault or shear zone; 3. using the relative ages of faults and shear zones and showing these relationships in 3D by stopping the structures on the older ones; 4. using the the fault dimensions in 3D modeling, and 5. dividing the grid or voxet by a discontinuity surface. This is important in interpolation, so that the estimated values are not affected by the values from the area across the fault or shear. 3D modeling of folding If fold structures have regular shapes they may be modeled using automated tools in some of the 3D geological software packages. For example, the axial surface may be defined based on orientation data and the axial surface may then be modeled. The shape of the fold can be modeled using the information whether it is anticline or syncline, and additional parameters including the interlimb angle. In the case of the Outokumpu formation, the 23 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Eevaliisa Laine and Kerstin Saalmann interpreted folding is so intense and unregular that it is not possible to make a simple mathematical model describing. However, locally using structural geological expertice the simplified fold patterns can be interpreted from the complicated looking bedrock surfaces (Fig. 14). Fig. 14. Horsmanaho open pit, the folded layers visible and the sample location x. Photo: Esko Koistinen, GTK and drawing: Kerstin Saalmann. 24 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Problems and challenges of 3D modeling of the Precambrian of Finland Interpolation of rock properties Geohemical compositions and petrophysical properties are often interpolated into a regular grid. The volume for interpolation grid should be divided using the obtained 3D model surfaces into volumes within which it may be assumed that geochemical and petrophysical properties are continuous, i.e. the applied mathematical or statistical model is valid for interpolation inside these subvolumes. Large changes in rock properties occur across the lithological boundaries and tectonic discontinuities. The rock properties change more across the layering than parallel to it. As the original or metamorphic layering is not planar, rectilinear coordinates are seldom an appropriate choice for interpolation. The rock properties should be investigated parallel to lithological or structural features. This can be carried out by using a so-called stratigraphic grid or by unfolding the data, i.e. transforming data into plane in order to be able to use rectilinear coordinates. The Figure 15 presents the stratigraphic grid along the slightly undulating Vuonos ore. Fig. 15. Stratigraphic grid illustrating Nidistribution along the ore body colored red. 25 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Eevaliisa Laine and Kerstin Saalmann 5 3D geolOgical software packages The present 3D geological software packages are designed for different purposes, and accordingly they use slightly different approaches. In the present study, different 3D software packages were applied. There are several other software packages, and new ones will be developed along with new computer technologies. 3D mining software packages aim to construct ore solids. These software packages are able to analyze and composite drill-hole data. An essential part is ore evaluation and often mine planning. In addition to mining applications, these software packages are also used in general geological 3D modeling, such as Surpac in geological modeling at the Olkiluoto nuclear waste site (e.g. Mattila et al. 2007). Many ‘general geological’ 3D software packages were originally planned for hydrocarbon exploration in sedimentary basins with a well-known stratigraphy, which is then visualized using surfaces cut by discontinuities such as faults and unconformities. Presently, these are being applied in modeling deformed rocks. Geological software packages construct 3D models using different approaches described in the previous section. This is a critical property that should be taken into account. Sometimes it might be quite difficult to apply an approach based on the known stratigraphy for strongly deformed rocks lacking any clear stratigraphic sequence. These software packages also have individual ways to save, import and export data. There might be advantages in using software that directly uses a general purpose database. In an ideal case, 3D modeling software would work with a general purpose database and be able to export 3D models in ASCII-format. Most of 3D modeling software packages also include numerical tools for statistical analysis, and some contain modules for the simulation of groundwater flow for the obtained model. Based on our experience with the Outokumpu case, a combination of different software packages is needed to construct a satisfactory model. We have prepared the following list of wishes for future 3D modeling tools in geology: • Easier inclusion of drawings into the modeling process. The best alternative could be a tool for drawing directly into a model. • Easier tools for changing and modifying parts of models. • It would also be good to be able to select modules that work independently in order to save the computer capacity. 6 3D geological modeling process The starting point of a 3D modeling process is the interpretation or knowledge of the subsurface geological structures. Before any further steps, the major discontinuity structures, faults and shear zones, should be known and their relative ages defined. These structures are key elements because they divide the bedrock into separate units. The collection and georeferencing of geological data is perhaps one of the most time-consuming parts of 3D modeling. Unifying the geological rock units from maps with those appearing in drill holes may be difficult. The geological map of the Outokumpu area (Gáal et al. 1975) contains six rock types, whereas about 55 different rock names were found for drill-hole lithologies. The rock units used in our example are based on the division of the GEOMEX project (Peltonen et al. 2008). Further simplification of the complex geology was achieved by defining the main structures such as faults and folds. 26 The Outokumpu area has been affected by a multiple-phase tectono-magmatic history with polyphase folding and shearing of different orientations causing complex interference structures followed by polyphase fault tectonics. The Outokumpu association was mod- eled by using the geological model of Koistinen in Gáal et al. (1975) (Fig. 2B). The modeling would have been very difficult using solely drill-hole data alone. Knowledge of the general trend and dip of the structures and the geological map helps to connect the lithologies in sections. Areas built up of strongly metamorphosed rocks, in particular in Precambrian terrains, often lack an established stratigraphy. Nevertheless, rules can be determined for how the rock types are distributed relative to each other, for example cross-cutting relations and rock-type associations related to certain tectonic, alteration or metamorphic events. Geometrical difficulties are caused Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Problems and challenges of 3D modeling of the Precambrian of Finland Fig. 16. Metamorphic rocks often show a layered structure due to intense deformation. The picture is from Hanko, illustrating strongly deformed mafic and felsic layersin. by overturned tight folds and lens-shaped bodies. Relatively regular metamorphic layered structures are found in Finnish bedrock, although these rarely represent the primary layering (Fig. 16). 3D geological models are seldom fully realistic, and are not intended to be. This is because of the lack of data. The correct idea for the form of geological bodies helps in constructing a reliable model. The way the software connects the points to form a model also affects the results. Polygonal shapes are caused by triangulation, while smooth interpolated surfaces may leave out important details. It should be kept in mind that all model points outside the hard data points are a result of interpretation. In the case of Finnish bedrock and available data, the following work flow could be recommended (Fig. 17): 1. Compilation of data and georeferencing. 2. Modeling of the surfaces of the youngest discontinuities that cut the geological formation under study Digitizing of sections in the case of abundant data and fairly linear bodies. In the case of complicated bodies, a sophisticated approach using tectonic orientations and contact points should be used. The latter approach is also recommended if there is very little knowledge of the subsurface and the model is solely based on geological intuition or theory. 3. The use of geophysical data through forward modeling in checking whether the obtained model reproduces the measured potential fields or signals. This is especially important in the case of seismic surveys. 4. Improvement of the geological model using geophysical inversion, i.e. optimization of the model. 5. Saving the model together with hard data and detailed documentation of the modeling process. A 3D model result should not be misunderstood to represent hard data. As we are continuing a model outside the actual data points, the model is uncertain. It is a result of interpolation, extrapolation and interpretation, and also reflects the ideas and theories at the time when it is constructed. Therefore, it is more important to save the associated hard data and to document the 3D modeling process together with underlying theories and a discussion of uncertainties rather than saving the geometric 3D objects alone. In Figure 17 a workflow for a general 3D geological modeling process is presented. It shows also possible applications of 3D geological models. 27 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Eevaliisa Laine and Kerstin Saalmann 7 Practical considerations In 3D modeling it is important that geologists, 3D modelers, and geophysicists work together. The organization of the team work is not just a process of collecting people together, but demands the learning of new methods of team work, communication, and dealing with ideas. 3D modeling demands a considerable amount of time. Fast computers may create an illusion of easy and rapid building of 3D geological models. However, in order to build as reliable a 3D model as possible, a considerable amount of data collection, georeferencing and digitizing work have to be done, accompanied by many steps involving cross-checking and continuous modification of interpreted structures. Fig. 17. A workflow diagram for a geological 3D modelling process. 28 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Problems and challenges of 3D modeling of the Precambrian of Finland 8 Conclusions The biggest problem in the 3D modeling process is the lack of data. The deepest data sources are from depths above 1 km. Geophysical data can be used to obtain some information on the spatial distribution of petrophysical properties at depth. However, geophysical fields and sections can be interpreted in many ways. Several alternative explanations may exist for how the bedrock of the certain area has been formed. Without extensive drilling, the interpretation of the subsurface geology is more or less ‘informed guessing’, or simply the visualization of someone’s geological ideas. It would be also wise to use different geological ideas as starting points in 3D modeling when searching for new ores or looking for possible nuclear waste sites. The few outcrop data may not be sufficient to build a unique model of the subsurface geology. In the case of Outokumpu area a lot of drill cores are available reaching a depth of about 500 metres. Below that point structures are continued using seismic sections. But the structures apparent on the seismic sections that do not continue to the surface may not be correlated with any of the surficial structures without deep drilling. Still the modeling and 3D modeling is the best way to get some idea about how the subsurface structures look like or what kind of structures are possible using all the available geological and geophysical information. Various software packages have certain strengths, but also specific limitations. Complex geology, the use of different data types and data sets as well as different modeling approaches makes it necessary to apply a number of software packages. We conclude that the use of different approaches is necessary for comprehensive data capture and 3D visualization of complex geology, such as present in the Outokumpu area. It is also evident that purely technical visualization without involving geological ideas and interpretations is impossible. It is important to understand that all models contain a large degree of uncertainty, and very different models may be constructed from the same geological surface structures, geophysical fields and reflections. Supplementary information on the web GEOMEX JV-project 1999–2003. Available at: http://vsa. gtk.fi/geomex/html/source/start.html Scales, John A., Smith, Martin L. & Sven Treitel 2001. Intro- ductory Geophysical Inverse Theory. Available at: http:// samizdat.mines.edu References Airo, M.-L., Loukola-Ruskeeniemi, K. & Hyvönen, E. 2009. Geophysical databases supporting large- and detailed scale mineral exploration: black shale formations in Finland. In: Smart science for exploration and mining: proceedings of the 10th Biennial SGA Meeting, Townsville, Australia, 17–20 August 2009. Townsville: James Cook University, 781–783. Brusila, J. 1983. Kemin kromiittiesiintymän geostatistinen analyysi. (Geostatistical analysis of the Kemi chromite deposit). Unpublished Master Thesis. Helsinki University of Technology. Calcagno, P., Chilès, J. P., Courrioux, G. & Guillen, A. 2008. Geological modelling from field data and geological knowledge, part I — modelling method coupling 3D potential field interpolation and geological rules. Physics of the Earth and Planetary Interiors 171, 147–157. Carr, J. C., Beatson, R. K., Cherrie, J. B., Mitchell, T. J., Fright, W. R., McCallum, B. C. & Evans, T. R. 2001. Re- construction and representation of 3D objects with radial basis functions. In Proceedings of SIGGRAPH 2001, ACM Press/ACM SIGGRAPH, E. Fiume, Ed., 67–76. Caumon, G., Collon-Droiaillet, P., Le Carlier de Veslud, C., Viseur, S. & Sausse, J. 2009. Surface-based 3D modeling of geological structures. Math. Geosci 41, 927–945. Chilés, J. P. & Delfiner, P. 1999. Geostatistics: Modeling Spatial Uncertainty. Wiley-Interscience. 720 p. Da Rocha, M. M & Yamamoto J. K. 2000. Comparison between kriging variance and interpolation variance as uncertainty measurements in the Capanema iron mine, State of Minas Gerais – Brazil. Natural Resources Research, Volume 3, 223–235. De Kemp, E. A. & Sprague, K. B. 2003. Interpretive tools for 3D structural geological modeling parti: Bézier-based curves, ribbons and grip frames. GeoInformatica, 7(1), 55–71. Fernández, O., Mûnoz, J.A., Arbués, P., Falivene, O. & Marzo 29 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Eevaliisa Laine and Kerstin Saalmann M. 2004. Three-dimensional reconstruction of geological surfaces: an example of growth strata and turbidite systems from the Ainsa basin (Pyrenees, Spain). AAPG Bull 88(8), 1049–1068. Frank, T., Tertois, A. L. & Mallet J-L. 2007. 3D-reconstruction of complex geological interfaces from irregularly distributed and noisy point data. Comp Geosci 33 (7), 932–943. Gáal, G., Koistinen, T. & Mattila, E. 1975. Tectonics and stratigraphy of the vicinity of Outokumpu, North Karelia, Finland: including a structural analysis of the Outokumpu ore deposit. Geological Survey of Finland, Bulletin 271, p. 67. Gjøystdal, H., Reinhardsen, J. E. & Asteböl, K. 1985. Computer representation of complex three-dimensional geological structures using a new solid modeling technique. Geophysical Prospecting 33 (8), 1195–1211. Groshong, R. H. 2006. 3-D structural geology, 2 nd edn. Berlin: Springer. Gustavsson, N. 2010. Geologisen paikkatiedon epävarmuus. Uncertainty of geological spatial data. Geological Survey of Finland, archive report TA/2010/7. Jones, N. L. & Nelson, J. 1992. Geoscientific modeling with TINs. Geobyte 7, 44–49. Kessler, H., Mathers, S. & Sobisch, H.-G. 2009. The capture and dissemination of integrated 3D geospatial knowledge at the British Geological Survey using GSI3D software and methodology. Computers & Geosciences 35, 1311– 1321. Kilpeläinen, T. 1998. Evolution and 3D modeling of structural and metamorphic patterns of the Palaeoproterozoic crust in the Tampere–Vammala area, southern Finland. Geological Survey of Finland, Bulletin 397. 124 p. Koistinen, T. J. 1981. Structural evolution of an early Proterozoic strata-bound Cu-Co-Zn deposit, Outokumpu, Finland. Transactions of the Royal Society of Edinburgh: Earth Sciences, 72, 115–158. Kontinen, A., Peltonen, P. & Huhma, H. 2006. Description and genetic modelling of the Outokumpu-type rock assemblage and associated sulphide deposits. Geological Survey of Finland, archive report M 10.4/2006/1. Korsman, K., Koistinen, T., Kohonen, J., Wennerström, M., Ekdahl, E., Honkamo, M., Idman, H. & Pekkala, Y. (eds.) 1997. Bedrock Map of Finland 1:1 000 000. Geological Survey of Finland. Laine, E. 1998. Geostatistical, geological and geophysical modelling of subsurface structures of Precambrian Bedrock in Finland. Helsinki University of Technology, Laboratory of Engineering geology and Geophysics, Research Report. 101 p. Lajaunie, C., Courrioux, G. & Manuel, L. 1997. Foliation Fields and 3D Cartography in Geology: Principles of a Method Based on Potential Interpolation. Mathematical Geology, Vol. 29, No. 4. Ledez, D. 2002. Euclidean distance mapping: Geological applications. in Terra Nostra, (Proceedings of IAMG, Ber- 30 lin), 25–30. Luoma, S., Backman, B., Valjus, T., Klein, J. 2010. Threedimensional geological model and its application for groundwater flow model and management of the Hanko Aquifer, south Finland. Department of Geosciences and Geography, C1. Mallet, J.-L. 1992. Discrete smooth interpolation in geometric modeling. Comp Aided Design 24, 178–191 Mallet, J.-L. 1997. Discrete modeling for natural objects. Math Geol 29(2), 199–219. Mallet, J.-L. 2002. Geomodeling. Oxford University Press, New York. 593 p. Markovaara-Koivisto, M, Laine, E., Wennerstrom, M., Hokkanen, T. & Selonen, O. 2009. Deriving rock quality parameters from a 3D discontinuity model; case study from a dimension stone quarry Finland (FIN). Finnish Tunnelling Association, Get underground seminar 4-5 Nov. 2009, Espoo, Finland. Mattila, J., Aaltonen, I., Kemppainen, K., Wikström, L., Paananen, M., Paulamäki, S., Front; Kai, Gehör, S., Kärki, A. & Ahokas, T. 2007. Geological model of the Olkiluoto Site. Version 1.0. Posiva Oy, Working Report 92. Metzger, A. 1928. Über die Tektonik des Grundgebirges um Svartå in SW-Finnland. Fennia 50 (17) . Pajunen, M., Airo, M.-L., Elminen, T., Niemelä, R., Salmelainen, J., Vaarma, M., Wasenius, P., Wennerström, M. 2008. Construction suitability of bedrock in the Helsinki area based on the tectonic structure of the Svecofennian crust of southern Finland. In: Tectonic evolution of the Svecofennian crust in southern Finland - a basis for characterizing bedrock technical properties. Geological Survey of Finland. Special Paper 47. 309 p. Parkkinen, J. 2003. Spatial statistical and structural modeling of the Keivitsa Ni-Cu-PGE deposit, northern Finland. Geophysical Research Abstracts. EGS-AGU-EUG joint assembly, Nice, France, 06-11 April 2003. Peltonen, P., Kontinen, A., Huhma, H. & Kuronen, U. 2008. Outokumpu revisited: New mineral deposit model for the mantle peridotite-associated Cu–Co–Zn–Ni–Ag–Au sulphide deposits. Ore Geol. Rev. 33, 559–617. Saksa, P. 1995. ROCK-CAD – computer aided geological modelling system. Nuclear Waste Commission of Finnish Power Companies. Report YJT-95 (18). Salminen, N. 2009. Microstructures in 3D modelling, Outokumpu area as an example site. Master’s thesis, Department of Civil and Environmental Engineering, Geoenvironmental technology. 59 p., 9 app. Viard, T., Caumon, G. & Lévy, B. 2011. Adjacent versus coincident representations of geospatial uncertainty: Which promote better decisions? Computers & Geosciences 37, 511–520. Viola, G., Mattila, J., Zwingmann, H., Todd, A. & Raven, M., 2011. Structural and K/Ar illite geochronological constraints on the brittle deformation history of the Olkiluoto region, southwest Finland. Posiva Working Report 37. 3D modeling of polydeformed and metamorphosed rocks: the old Outokumpu Cu-Co-Zn mine area as a case study Edited by Eevaliisa Laine Geological Survey of Finland, Report of Investigation 195, 2012 3D MODELING OF POLYDEFORMED AND METAMORPHOSED ROCKS AT DIFFERENT SCALES USING GEOLOGICAL AND GEOPHYSICAL DATA FROM OUTOKUMPU AREA by Eevaliisa Laine*, Esko Koistinen, Kerstin Saalmann, Gabriel Coirrioux, Rosa Diaz, Noora Salminen and Timo Tervo Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012. 3d modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area. Geological Survey of Finland. Report of Investigation 195, 41 figures. 3D models of the Outokumpu assemblage, the Keretti and Vuonos ore bodies, and open pit rock walls at Horsmanaho have been constructed by using a variety of geological and geophysical data and by applying different approaches. Software packages used in the modeling process have different means for interpolation and inference of geological structures such as lithological interfaces and faults. Different approaches are needed in order to model the complex contact relationships within Precambrian rocks. It is not always possible to find geological cross sections for realistic 3D modeling of complicated fold patterns. On the other hand, structural data in drill core loggings are usually missing. The possibility of importing different data sets and models into the same software platform facilitates the 3D interpretation of structures. However, the use of different approaches and data from several sources requires detailed planning and documentation of the 3D modeling process. Keywords (GeoRef Thesaurus, AGI): metal ores, ore bodies, metamorphic rocks, three-dimensional models, structural geology, deep seismic sounding, Paleoproterozoic, Keretti, Vuonos, Outokumpu, Finland *Eevaliisa Laine Geological Survey of Finland, P.O. Box 96, FI-02151 Espoo, Finland *E-mail: [email protected] 31 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012 1 Introduction The Outokumpu area was selected as a case study for the 3D/4D project (at GTK 2007–2009) because of a long history of research, and hence the availability of data from several sources. The mining history of this area extends from 1913 to 1988, involving the exploitation of three major deposits: Outokumpu, Vuonos and Luikonlahti. Ou- tokumpu and Vuonos deposits are considered in more detail in this study. The use of different geological modeling techniques is presented through examples according to the modeling scale (Fig. 1). The smallest scale in this study was a few centimeters, i.e., 3D modeling of microstructures. Despite the problems with geocoding, pleochroism A) B) Fig. 1. 3D modelling of the Outokumpu area at different scales. A) Regional geology Korsman et al. (1997) and the area for the 3D geological model of Ruotoistenmäki & Koistinen (2011 in this volume). B) Geological map of the vicinity of Outokumpu by Tapio Koistinen in Gáal et al. (1975). 32 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 3d modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area and interference colors, some planar structures were modeled using thin section photographs in the geological modeling software Paradigm Gocad® (http://www.earthdecision.com/index.html). The results were presented by Salminen (2009) in her Master’s thesis entitled ‘Microstructures in 3D modeling, Outokumpu area as an example site.’ Rock walls in the open pit from Horsmanaho and Sola represent the scale of tens of meters. 3D models were constructed by transforming stereo photographs with Sirovision software. Keretti and Vuonos ores were modeled using section drawings from the mines and the traditional section-by-section interpretation method. This seems to be the best way to delineate these thin and irregularly shaped ore bodies, which were found to be rather continuous, although locally truncated by vertical faults. The triangulation into solid models was performed using Gems software in the case of Outokumpu ore and Paradigm Gocad® in the case of Vuonos ore. The ores are situated in the so-called Outo- kumpu assemblage (Peltonen et al. 2008). This assemblage and the associated black schists were modeled by using geological sections and a geological map (Gáal et al. 1975) with GeoModeller software (http://www.geomodeller.com). The resulting 3D models were combined and visualized using Paradigm Gocad® (http://paraviewgeo.mirarco.org/index.php/Main_Page) and finally exported into 3DPDF formats. The use of data from different sources was demonstrated by visualizing the seismic sections of the Outokumpu area together with together with geological and geophysical maps using Paradigm Gocad®. A regional 3D model of the area was inferred from geophysical data by Ruotoistenmäki and Koistinen (2010) and is presented in a chapter of the present report. The aim of this paper is to demonstrate the possibilities and limitations of 3D modeling with these case studies. 2 Geology of the Outokumpu region The Outokumpu mining district hosts a Palaeoproterozoic sulfide deposit characterized by an unusual lithological association. It is located in the North Karelia Schist Belt (Fig. 2), which was thrust on the late Archaean gneissic–granitoid basement of the Karelian craton during the early stages of the Svecofennian Orogeny between 1.92 and 1.87 Ga (Koistinen 1981). Two major tectono-stratigraphic units can be distinguished, (1) a lower, parautochthonous ‘Lower Kaleva’ unit and (2) an upper, allochthonous ‘upper Kaleva’ unit or ‘Outokumpu allochthon’. The latter consists of tightly-folded deep marine turbiditic mica schists and metagraywackes containing intercalations of black schist, and the Outokumpu assemblage, which comprises ca. 1950 Ma old, serpentinized peridotites surrounded by carbonate–calc-silicate (‘skarn’)–quartz rocks. The ore body is enclosed in the Outokumpu assemblage, which is thought to be part of a disrupted and incomplete ophiolite complex (Vuollo & Piirainen 1989) that can be traced to the Kainuu schist belt further north where the well-preserved Jormua ophiolite is exposed (Kontinen 1987, Peltonen & Kontinen 2004). 33 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012 A) B) Fig. 2. A) Location of the Outokumpu area; B) Regional geology after Korsman et al. (1997). 34 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 3d modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area 3 Rock walls and microstructures in 3D Rock walls were mapped and photographed in the Horsmanaho quarries. Open pits are ideal for understanding structures in 3D. For example, it is easy to make observations of structures such as faulting and schistosity from extensive rock walls (Fig. 3). 3.1 Microstructures in 3D Along with mapping of the quarry walls oriented samples were taken. Thin sections are most commonly made from unoriented rock samples. In this case, microstructures such as schistosity and layering may be observed, as well as their mutual relationships. However, the orientation of these is not related to any macroscopic features in the rock. Oriented thin sections, in contrast, allow these features to be related to macroscopic structures such as schistosity or folds. The rock slices chosen for the thin sections were oriented parallel to the main planes of the strain ellipsoid; therefore, two section planes, XZ and YZ were investigated for most rock samples; for 1 - 2 samples, the third plane, XY, was also included. Microstructures alone generally do not provide sufficient information without being linked with field observations. A major aim was to visualize the 3-dimensional microstructures coupled with the 3D geometry on the outcrop scale in order to illustrate how the microstructures reflect the outcrop patterns using 3D- modeling and 3D-visualization software. The oriented sections from Horsmanaho are presented as an example (Salminen 2009). The rock type is black schist with a fairly high graphite content and strongly altered fracture surfaces. Figure 4B illustrates the position of the sample relative to the macroscopic folding at Horsmanaho. The photographs were taken using crossed nicols in a polarizing microscope, and were georeferenced and imported into a Paradigm Gocad® project. The schistose structure was due to the thin graphite layers (Fig. 4). It can be seen that the shear zones and schistosity continue from the vertical to the horizontal section, so that the planes could be fitted and measured. The obtained orientations are local because of the scale of the study. In folded rocks the orientation of layering is variable, so these orientations are only meaningful when used in association with outcrop observations. The study of oriented sections may provide an explanation for structures observed macroscopically. The whole process demands careful orientation during sampling and thin section preparation. Fig. 3. An example photograph from the Horsmanaho open pit. Photo: Esko Koistinen, GTK. 35 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012 Interference colors and phleochroism cannot be observed in 3D view, and the interpretations of structures could not therefor be carried out solely using thin section photographs in 3D. However, opaque minerals and microjoints look similar in different orientations, and may be studied from the thin section photographs in 3D. In the future, it could be interesting to study larger rock samples in 3D, for example by using photographs of polished rock surfaces. A) B) Fig. 4. A) Thin sections (the size of 2 cm x 3 cm) in 3D and the schistocity plane created from traces on two differently oriented sections (Photo: Noora Salminen); B) Horsmanaho open pit, the folded layers visible, and the thin sections were made from the sample near the hinge of the fold. Photo: Esko Koistinen, GTK and drawing: Kerstin Saalmann. 36 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 3d modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area 3.2 Rock walls using the Sirovision® 3D imaging system The Sirovision® 3D imaging system creates 3D images from digital pho tographs. These photographs were taken in Horsmanaho and Sola pits, and the exact camera position and orientation as well as the position of marker points on the photographed pit walls was surveyed by GPS and using a tachymeter. With SiroJoint®, spatial measurement of planes could be mapped from the 3D images and checked and combined with field measurements of structural data. Photogrammetry is essentially a process of triangulation using measurements of angle obtained from images. The position of an object point ‘P’ in Figure 5 can be calculated from measurements of the positions of the image(s) of the point in the left and right images, that is the positions of points ‘PL’ and ‘PR’. With these measurements and using both the position and orientation of the camera for each image, and the focal length of the camera the position in space of point ‘P’ can be determined. If the true positions of the camera(s) are known the actual spatial position of the object point ‘P’ is determined. If only the relative positions of the camera(s) are known the spatial position of the object point ‘P’ relative to the cameras is determined. By repeating this process for a set of objects, or points, a 3D map can be generated. The photogrammetry system developed by CSIRO Exploration and Mining performs these calculations automatically for a dense grid of points in both images thus generating a detailed 3D map of the scene being analyzed. Siro 3D: 8 steps to create a 3D image: 1. Plan The 3D Imaging Task 2. Acquire The Images 3. Correct The Images (Remove distortion) 4. Set up 3D Imaging Support Data (<=Acquire Control Point & Image (Camera) Position (Survey) Data and setup camera position and orientation) 5. Set up The Imaging Task 6. Run The Matching 7. Create The 3D Image 8. Save The 3D Image During a short field period in 2007, walls from Horsmanaho and Sola open pits were photographed. Faulting and jointing of the obtained 3D models may be analyzed using SiroJoint® software. SiroJoint® analyses 3D images of the surface of a rock mass (topographic maps of the surface of a rock mass integrated with visual images of the rock) to determine, for instance, the dip, orientation and spacing of discontinuity sets. It provides facilities for the visualization and analysis of rock slopes using accurate spatial data registered in a real world coordinate system. Figure 6 illustrates the use of a digital camera and control points on the rock wall. The Horsmanaho open pit consists of black schists, serpentinite and talc schist. The example Figure 7 presents a photograph and 3D image of the rock wall consisting of black schist. Talc schist has been mined away from the Sola serpentinite. The rem- Fig. 5. Geometry for the determination of the position of a point in object space 37 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012 nants were photographed over the water- filled open pit. The obtained stereo photos were combined into 3D image (Fig. 8). Discontinuities such as lithological contacts were not easy to discern because of the uneven and often dirty rock walls. In addition, lithologi- cal contacts as well as fault surfaces are often curved and may not be defined by one tectonic measurement. Consequently, the main use of 3D images of the rock walls is the analysis of jointing and planar faults. Fig. 6. Digital Camera and control points. Photo: Esko Koistinen, GTK. 38 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 3d modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area Fig. 7. A SiroVision example from Horsmanaho. Photo: Esko Koistinen, GTK. 39 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012 Fig. 8. Sola open pit as SiroVision 3D visualization. Photo: Esko Koistinen, GTK. 40 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 3d modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area As an example, the jointing of one rock wall was analyzed using SiroJoint® software. In order to measure all the joint orientations, the jointing of rock walls was studied by examining the rock wall from different directions (Fig. 9). The fracture orientations were calculated by SiroJoint® and presented in a stereographic projection (Fig. 10A) and in a Rose diagram (Fig. 10B). SiroJoint® also calculates the mean directions of the clusters, and it is possible to show corresponding joint planes in 3D (Fig. 10C). On the example rock wall at Horsmanaho, the jointing is nearly vertical and parallel and mainly perpendicular to the strike of the Outokumpu assemblage rocks. The same orientation of brittle structures is evident on the geophysical maps, in the elevation model, and has been confirmed by geological observations (See Chapter 5). Geological observation is needed to define of relative ages and movements of faults. Furthermore, fracture properties such as fracture fillings and openness should be observed directly on the wall. Fig. 9. Rock fractures studied on a 3D image from Horsmanaho pit from two different directions. Photo: Esko Koistinen, GTK. 41 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012 A) B) C) Fig. 10. A) Jointing on the stereographic projection; B) Presented with the rose diagram and C) Fracture plains visualized in 3D using SiroJoint®. 42 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 3d modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area 4 Ore models The Outokumpu (Keretti) and Vuonos Cu-CoZn-Ni sulfide ores in eastern Finland are hosted by the distinctive Outokumpu assemblage, which comprises serpentinized peridotite massifs that have been altered and metamorphosed to quartz rocks, Cr-silicate skarns and carbonate rocks. The Outokumpu assemblage, and the enclosing metamorphosed black shales and turbidites belong to an allochthonous nappe complex emplaced onto the Karelian Craton margin during the early stages of the Svecofennian Orogeny. The Outokumpu ore body, which was discovered in 1910, contained 28 Mt of ore, of which 23 Mt was mined before closure in 1989. The Vuonos ore body, discovered in 1965, some 10 km NE of Keretti, contained 6 Mt of ore, of which 4 Mt was extracted. Recent geochemical models (Peltonen et al. 2008) explain the close association of peridotites with carbonate-quartz rocks as result of silicification and carbonation alteration of the ultramafic rocks. Ni and Co have been derived from a different source, and a polygenetic development is suggested, including the formation of a Cu-rich proto-ore within the 1950 Ma seafloor followed ca. 40 Ma later by mobilization of Ni and redeposition in disseminated Ni sulfides in response to alteration during obduction of the upper Kaleva unit. Later deformation finally caused mixing of the Cu and Ni ore components, leading to the uncommon metal association of Outokumpu. 4.1 Outokumpu ore model In 1954 the 3D model of the Keretti (also called Outokumpu) ore body was constructed using bronze by Outokumpu Oy (Fig. 11). During the project, Esko Koistinen carried out the modeling of the Outokumpu ore using mine cross sections and other maps stored at GTK’s archive in Outokumpu. The sections around Keretti area were scanned in the GEOMEX project (http://vsa.gtk. fi/geomex/html/source/start.html). The other sections around Mökkivaara and the Old Mine were scanned in the ongoing 3D/4D project at GTK’s Eastern Finland office. The Outokumpu ore has been surveyed on a local co-ordinate system and vertical sections have been drawn over a distance of forty meters in the longitudinal SW-NE direction of the ore. In the Keretti area, geological sections have a distance of twenty meters. The sections are numbered 0 to 98 from NE to SW. The sections have been drawn viewing towards NE. The coordinate system has been changed twice during the production history of the mine. Hence, the vertical sections are located in three different directions (Fig. 12). The relation between the local mining coordinate system and the national coordinate systems could not be found out in archives. The location of the sections was derived by georeferencing the mining area map of Figure 12 based on drill hole locations on the map and their known national KKJ coordinates. A maximum error in horizontal locations is estimated to a couple of metres. The oldest sections in the Kaasila area are numbered from 0 to 12, the second system near the old shaft is numbered from 13 to 31 and the third system from the old mine through Mökkivaara to Keretti is numbered from 32 to 98. Geological drawings of the oldest sections, numbered from 0 to 31, were prepared at a scale of 1:800. An example of these drawings is given in Figure 13. Section 23 is located at the old shaft and section 24 is 40 m SW of the shaft. It can be seen that the drawings do not describe the final stage of ore caving. Sections 31 to 46 in the Mökkivaara area have been named “drill sections” (syväkairausprofiili, Fig. 14). The scale is 1:800. Sections 47 to 98 in the Keretti area have been Fig. 11. A bronze-made 3D model of the Outokumpu ore body at a scale of 1:4000 1954. Photo: Jari Väätäinen. 43 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012 A) B) Fig. 12. The mining area of Outokumpu. Vertical sections are situated in three different directions, A, B and C (Figs. A and B). Mining started from the Kaasila ore body and finished in the Lietukka ore body. (GTK archives Outokumpu). 44 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 3d modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area Fig. 13. An example of the oldest section drawings numbered from 0 to 31. Section 23 is located at the old shaft and section 24 is 40 m SW of the shaft. drawn at a scale of a 1:400 (Fig. 15). In the oldest sections 0 to 31 only the ore is shown, while in the other sections the country rocks are also presented. All sections have been hand drawn. Some occasional sections were lacking. The sections were georeferenced using Gemcom GEMS software. KKJ coordinates were not given in the section drawings, but sectional x coordinates (P100, P200 etc.) and the main working levels of the mine (+200, +250, +280, +320) were drawn. The sections were georeferenced according to these lines using certain construction lines to assist work in the 3D space. In GEMS software the ore body was digitized to 3D rings using the georeferenced sections as references. The 3D rings and the georeferenced section 23 are shown in Figure 16. The 3D rings were connected to produce a 3D solid model of the ore deposit (Fig. 17). Fig. 14. Section 44 as an example of the drill section drawings from 31 to 46 in the Mökkivaara area. 45 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012 Fig. 15. Section 71 as an example of the section drawings from 47 to 98 in the Keretti area Fig. 16. Cross sections of the Outokumpu ore body digitized as 3D rings. Section 23 in the figure is located at the old mine shaft. 46 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 3d modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area Fig. 17. The 3D rings are connected to a 3D solid model of the Outokumpu ore deposit. Fig. 18. 3D model of the Outokumpu ore seen from SE. The vertical section 23 in the figure is located at the old mine shaft. Kaasila on the right, Lietukka on the left. Fig. 19. 3D model of the Outokumpu ore seen downwards from NW proves how thin the ore body is. The vertical section 23 in the figure is located at the old mine shaft. Kaasila on the left, Lietukka on the right. 47 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012 Fig. 20. A detail of the 3D model of the Outokumpu ore. A fault divides the Kaasila ore body (on the left) and the Kumpu ore body from each other. Mining started as open pit mining at the outcrops of the Kaasila ore body. The vertical section 23 in the figure is located at the old mine shaft. Fig. 21. A detail of the 3D model of the Outokumpu ore. A fault divides the Kumpu ore body (on the left) and the Lietukka ore body from each other. Vertical section 71 in the figure is located at the Keretti head frame. The Poikanen ore body grows out of the Lietukka body and is seen below the main body. Appendices 1 and 2 contain PDF3Ds files for this model. The 3D model of the Outokumpu ore deposit produced from the old mine sections is presented in Figures 18 to 21. The model demonstrates well the known main features of the ore. The length of the deposit is 3686 m and the maximum width 406 m measured in 3D from the model. In the conventional 2D presentations the Outokumpu ore body gives the impression that the ore body is much thicker than it in fact is (Fig.18). In 3D, looking downwards from NW, the body appears very thin (Fig. 19). The major, well-known faults, which divide the body into the Kaasila, Kumpu and Lietukka bodies, could also be modeled (details in Figs 19 and 20). The Poikanen ore body is seen below the main ore body in the Keretti area (Fig. 21). The model also demonstrates separated small ore blocks and roughness of the ore surfaces, which are seldom present in illustrations of the Outokumpu ore. The volume of the 3D model is 8118 Mm3, which gives an ore tonnage of 28 Mt with a density of 3.45. 4.2 Vuonos ore The Vuonos ore was modeled using cross sections scanned during the 3D/4D project at GTK’s Eastern Finland office. The approach was similar to the one applied in Outokumpu ore modeling. In the Vuonos case, sections were imported into Paradigm Gocad® software. Gocad® is an extensive software package for 3D geological modelling. The ore was digitized along the vertical mining sections and then combined into a 3D model of the Vuonos ore using triangulation. The modeling process is illustrated in Figure 22. Vuonos ore is gently folded, but not as strongly 48 as the surrounding serpentinite and its alteration products (Fig. 23). It may be interpreted to be associated with a shear zone between two serpentinite bodies, because in many sections sheared rocks have been found near the ore layer based on the section drawings. This shear zone is also present in the cross sections across Outokumpu assemblage drawn by Koistinen (1981). In addition to gentle folding of the ore body along the strike, the flat ore body is a flexure together with the Outokumpu assemblage rocks on the section drawings (Fig. 24). The ore body may Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 3d modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area have folded together with the Outokumpu assemblage rocks, or simply filled the space conforming to the roundish serpentinite bodies. These kinds of features may not be modeled in detail by using triangulation to connect vertical sections, because the connection outlines between sections produce a modeled thicker ore body instead of a curved plane. Curving may be better modeled using potential fields (e.g. Lajaunie et al. 1997) if structur- al measurements of the orientations of subsurface structures are available. This type of thin ore layer is a challenge to any currently existing 3D modeling software. The Vuonos ore body gently dips to the northeast. The middle part of the Vuonos ore body is more strongly deformed (Fig. 25) caused by faulting interpreted from geophysical data (Laine & Saalmann 2008, Fig. 31). A) B) C) Fig. 22 A) Digitized ore outlines from the section drawings; B) digitized outlines connected to form an ore body; C) the whole Vuonos ore body. 49 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012 Fig. 23. An example of Vuonos mine sections. Green colors refer to serpentinite, red to ore, yellow to calc-silicate rocks, violet to black schist, and blue to mica schist. 50 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 3d modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area Fig. 24. A mining section showing a detail of the Vuonos ore body (colored red). 51 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012 Fig. 25. Vuonos ore body colored red, three faults interpreted from seismic section (Laine & Saalmann, 2008). Seismic section (to the depth of 2 700 m) is done by high resolution seismics (Heinonen et al. 2011) and further prosessed and interpreted by Laine & Saalmann (2008). 5 3D modeling of the Outokumpu assemblage The about 100-m-thick Outokumpu assemblage enclosing the ore bodies was visualized from georeferenced maps and sections (Gaál et al. 1975), based on structural geological interpretations. In the 3D model, the Outokumpu assemblage and enveloping black schists are grouped together. The software used was GeoModeller, which enabled the use of constraints from structural observations, even though it was not possible to define a reliable stratigraphy within these complexly de formed rocks. The present-day sequence of rock layers was therefore used as an informal “pseudostratigraphy”. Despite this problem, the resulting model appears structurally realistic (Fig. 27). GeoModeller software applies an interpolator method (McInerney et al. 2005) that is based on potential field theory. A set of smoothly curving, sub-parallel geologic surfaces, such as lithological contacts (Fig. 26), in 3D space are treated analogous to a set of iso-potential surfaces of a scalar (potential) field. It is assumed that lithological contact data lie on a potential field surface and orientation vectors, e.g. schistosities parallel to lithological layers, are orthogonal to a local tangential plane to the potential field. The field increment between any two points belonging to same geologic interface is null. Orientation data rep- 52 resent the derivative of the field. The scalar field is then interpolated by cokriging the increment data and their derivatives (Lajaunie et al. 1997). This method was applied on geological maps and sections from the Outokumpu area, even though very little tectonic orientation data exists of the subsurface structures. 3D modeling of the Outokumpu assemblage surrounding the Vuonos ore body began during the Outokumpu workshop 2007. It was supervised by Gabriel Couirrioux from BRGM who acted as a teacher and helped us in 3D modeling of the Outokumpu assemblage surrounding the Vuonos ore body. Taking into account the complexity of the zone, it was decided to model the boundary between the Outokumpu assemblage surrounded by black schists and mica gneiss. Black schists envelope the Outokumpu assemblage according to the interpretation on the geological map. Disharmonic folding makes modeling from cartographic data alone very improbable, and extrapolation of the geometry at depth or even laterally is highly unpredictable without any information about subsurface structures. In the Vuonos area, such information was obtained from three geological sections (Gáal et al. 1975). Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 3d modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area Fig. 26. Possible lithological contacts for iso-potential surfaces. The section was drawn by Koisinen in Gáal et al. (1975). Fig. 27. Geological map, 3 sections, digitizing and the model. (1) On the left, digitalization of data in maps and sections relying on the georeferenced images. (2) On the right, the corresponding black-schists model. 53 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012 Figure 27 illustrates the digitalization of data from georeferenced maps and cross section images as well as the obtained Outokumpu assemblage model. It was decided to model the envelope of the Outokumpu assemblage rocks and black schists by a single boundary. In this case, singularities such as fold hinges and sharp angles are very difficult to link from one section to another. This is due to the shape complexity, and the existence of narrow elongated zones that cannot be resolved at the chosen scale. These zones can only be retrieved by adding intermediate sections. Another alternative is to consider the different lithologies as an organization of formations with Outokumpu assemblage rocks surrounded by black schists and mica schist. It is then necessary to correctly assign the drawing lines to the top of formations. The pile presented here is an example of what is possible to do in that case. This is, of course, not a unique solution. The only drawback is the creation of additional sub-formations. However, the result is quite acceptable. Figure 28 shows the informal stratigraphy and data organi- Fig. 28. An alternative way to organize the pile and data. 54 zation, while Figure 29 presents different views of the model. A further step during the 3D modeling process was to add possible fracture zones into the model. Faulting and jointing have already been described by Väyrynen (1939). During the present project, brittle structures were interpreted from an areal elevation model interpolated from elevation lines (Fig. 30) and geophysical maps (Fig. 31). Fracture orientations are mainly oriented perpendicularly to the strike of Outokumpu assemblage rocks. They are assumed to be subvertical such as Kaasila fault (B) in Keretti area. The Outokumpu ore is cut by this vertical fault (B) with a displacement of 100 m in the vertical direction (e.g. Fig. 18). A probable fracture zone (C) subdivide the Outokumpu assemblage into the Vuonos and Keretti areas. In addition, structure A is a discontinuity structure that borders the Outokumpu assemblage rocks to the southwest. The Vuonos area is cut by a probable fracture zone D and bounded to the North by the discontinuity structure E (Fig. 31). Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 3d modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area Step by step, more sections and drill hole data were included in the model of the Outokumpu assemblage. The applied stratigraphy consisted of two rock types: 1.The Outokumpu assemblage, in which the enveloping black schists (i.e. metamorphic black shales) were included in a model because they are part of the alteration zone, and it is difficult to exclude them geometrically. 2.Mica schist The rock types along the drill holes were classified according to this classification. The geological map (Gáal et al. 1975) and geological sections were digitized by using this classification (Fig. 32). 3D modeling was performed using GeoModeller softw are, and the 3D model presents the Outokumpu assemblage associated with the black schists as an elongated body with an undulating surface (Fig. 33) caused by folding, faulting and the original structure of the serpentinite bodies. Furthermore, the Outokumpu assemblage was modeled using the present layering of the rock formations from the southeast to north west: 1. Mica schist (above) 2. Black schist 3. Mica schist (middle) 4. Outokumpu assemblage 5. Mica schist (below) The resulted 3D model is illustrated in Figure 34. The probable fracture zones A, B, C, D and E are also included in the model. Fig. 29. Outokumpu assemblage rocks and mica schist. 55 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012 Fig. 30. An elevation model of the Outokumpu area overlaid by the geological map of the Outokumpu area (Gaál et al. 1975). Yellow lineaments were interpreted from geological and geophysical maps,and blue lineaments were interpreted from the elevation model. 56 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 3d modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area Fig. 31. Low-altitude magnetic map (processed by Timo Tervo); lines A, B, C, D and E mark approximate locations of possible fault or fracture zones. 57 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012 Fig. 32. Lithologies, sections and faultsdigitized from the geological map (Gáal et al. 1975). Fig. 33. Outokumpu assemblage (including black schists). 58 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 3d modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area Fig. 34. A model of the Outokumpu association and cross cutting fracture zones. The width of the formation is about 1.5 km. 59 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012 6 3D modeling of the Sola serpentinites At Sola and Horsmanaho (Fig. 1) the Outokumpu assemblage rocks occur as roundish bodies slightly elongated in the SW-NE direction. The Sola drill-hole data were analyzed in the Surpac mining software, which uses triangulation to build 3D surfaces. The rock types were classified according to the GEOMEX project (http://geomaps2.gtk.fi/ Geomex-tiedostot/kaytto/geomex_geology.html). The rock types included in the 3D model were black schist, serpentinite, mica schist, and quartz veins (Table 1). The mafic dykes cutting the serpentinite body were not modeled separately, but were included with the serpentinite. The first stage in Surpac is to construct sections Table 1. Rock types and corresponding rock codes and colors.Lisää alkuperäiset koodiselitykset Code BS OUM OBA WA QTZVE Rock type Black schist Outokumpu Assemblage OKU Metabasites Mica schist Quartz vein Fig. 35. Digitizing serpentinite bodies on the sections. 60 Colors Purple Green Green Blue Yellow Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 3d modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area from drill-hole data. The outlines of serpentinite bodies were digitized on each section (Fig. 35). This is not a trivial task, and generally requires previous knowledge of the form and orientation of the ore bodies. Three serpentinite bodies were modeled from sections and finally combined into three serpentinite models (Fig. 36). Geological sections and maps were used in improving the obtained model (Fig. 37): Fig. 36. Surpac solid models of Sola serpentinites. A) B) Fig. 37. Geological block model of the Sola serpentinites by Gáal et al. (1975); Kontinen (2008). 61 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012 7 3D model of the Vuonos area including detailed seismic FIRE-sections Seismic reflection profiles from the Outokumpu region measured in the FIRE project (Heinonen et al. 2011) reach a depth of 2700 m. The migrated sections were imported into Paradigm Gocad® software. The amplitude values represent the seismic signal that obtains positive and negative values. In order to discern the reflectors the absolute value was calculated. The reflectors were seen better (Fig. 36) when the indicator coding was applied: I (|amplitude|) = 1, if |amplitude|>2.5, 0 otherwise. The seismic sections were studied visually in 3D using Paradigm GOCAD® software and compared with geological maps. The strength of the GOCAD® software is its ability to combine data from different sources. The few examples of possible interpretations of the OKUFIRE-sections are listed below: • The first observation from the seismic sections is that the Outokumpu bedrock can be divided in two parts according to reflectivity (Fig. 38). In the western side of this discontinuity surface the reflectors are angular, in contrast to the roundish forms on the other side. The boundary is marked with a green plane dipping to the southeast (Fig. 39A). Reflectors in the southeast give a visual impression of movement that has turned and thickened the layers. • It can also be seen that the reflections in the 62 seismic section OKUFIRE3 do not continue in the sections OKUFIRE1 and OKUFIRE2. The reason may be inclined, folded and faulted structures. The location of the actual reflecting structures or layers may be at some distance from the sections. The reflections on OKUFIRE3 are subhorizontal, indicating that the section crosscuts the inclined reflectors (layering or shearing) (Fig. 38). In addition, faults interpreted from seismic section OKUFIRE3 (Laine and Saalmann 2008; Fig. 25) cut the Vuonos ore at the point where the ore body has been faulted or sheared. These faults have also been interpreted from the geological and geophysical maps (Fig. 31). • The subsurface formations appear to be continuous and parallel to the Outokumpu assemblage (Fig. 40). • It can be intepreted several possible orientations for discontinuity structures (Fig. 41). Because seismic sections may be interpreted in many ways. Deep drilling is needed to verify the existence of shear zones or locations of the rocks belonging to the Outokumpu assemblage. Drilling was first carried out near the OKUFIRE1, and the petrology was described in the Deep Drilling Project at GTK (Västi 2011). In general, 3D visualizations facilitate the interpretation of seismic profiles, especially, when other geological data sets can be visualized and compared with seismic reflectors. Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 3d modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area A) B) C) Fig. 38. A) FIRE-sections OKUFIRE1, OKUFIRE2 and OKUFIRE3 on the geological map (Gáal et al. 1975); B) The original amplitude value presented in grey scale, and C) the indicator-transformed absolute value of the amplitude. The value of 1 is marked with a white color. 63 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012 A) B) Fig. 39. A) Seismic sections OKUFIRE 1 and 2 to the depth of 2 700 m and the surface, which divides the bedrock into differently reflective parts. The continuity of the structures is also evident. B) Seismic sections OKUFIRE 2 and 3, geological map (Gáal & Koistinen, 1973), Outokumpu ore body, Vuonos ore body and the dividing surface. Fig. 40. OKUFIRE 1 and 2 to the depth of 2 700 m together with the Keretti and Vuonos ore bodies. 64 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 3d modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area Fig. 41. OKUFIRE 2 section to the depth of 2 700 m and the Vuonos ore body (in red) surrounded by Outokumpu Assemblage (bounded surface in green) that is cut by a shear zone (in white); possible large discontinuity structures (in white, green and yellow ) were interpreted from the seismic sections (Laine and Saalmann 2008). The corresponding 3DPDF is found in Appendix 3. 8 Conclusions The Outokumpu area has been studied since 1910, and the data source consists of old maps and section drawings. Because the mining areas are covered by thick layers of Quaternary deposits, very few outcrops are available. Hence, the section drawings and drill-hole data formed the main geological data source in this study. The Keretti and Vuonos mines are closed and inaccessible. The complicated structures affected the 3D geological modeling in many ways: 1. It is difficult to correlate planar fabrics inferred from microstructures with those observed macroscopically, because their orientations vary over short distances. 2. The complex folding and shearing patterns cause repetition of data in such a manner that it is difficult to discern or model lithologies separately. Simplification of rock type classification is usually needed for geometric 3D modeling. 3. The oldest structures are complicated and their continuation to greater depths is sometimes very difficult to predict. Modeling of these structures should be preceded by 3D modeling of large faults or shear zones. Even though the latter features are fairly planar or only gently curved surfaces their subsurface orientation and dimensions may be difficult to model. In 3D modeling of these subplanar features, geophysical data are essential and sometimes the only data source. 4. Geophysical data, such as seismic sections, may be interpreted in many different ways. However, they are sometimes the only data sources and provide some idea of subsurface structures and their continuities, especially when visualized and examined in 3D modeling environments. Because the Outokumpu area has a complex geology, a purely technical visualization using data sets without involving geological ideas and interpretations and considering different view points is impossible. It is also evident that the Outokumpu 3D model strongly depends on the interpretation of the regional deformation and alteration history, so that a different interpretation of the geological history could lead to quite a different 3D model. However, structural features are more easily interpreted by using modern 3D modeling software. The data and models from different sources may be studied inside the same platform. Tectonic orientations may be used directly in 3D modeling. Nevertheless, data from different sources at various scales and resolutions do not make the modeling process easier, and demand a team of researchers from different fields in geology and geophysics rather than an individual specialist working alone. The understanding and documentation of each step in the modeling process is essential. 65 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Laine,E., Koistinen, E., Saalmann, K., Coirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012 Supplementary information on the web GeoModeller, available at: http://www.geomodeller.com GEOMEX-project, available at: http://vsa.gtk.fi/geomex/ html/source/start.html Paradigm GOCAD®, available at: http://www.earthdecision. com/index.html ParaviewGeo, available at: http://paraviewgeo.mirarco.org/ index.php/Main_Page) References Gaál, G., Koistinen, T. & Mattila, E. 1975. Tectonics and stratigraphy of the vicinity of Outokumpu, North Karelia, Finland: including a structural analysis of the Outokumpu ore deposit. Geological Survey of Finland, Bulletin 271. 67 p. Heinonen, S., Kukkonen, I. T., Heikkinen, P. J. & Schmitt, D. R. 2011. High resolution reflection seismics integrated with deep drill hole data in Outokumpu, Finland. In: Outokumpu Deep Drilling Project 2003–2010 Kukkonen, I. T. (ed.) Geological Survey of Finland, Special Paper 51, 105–118. Koistinen, T. J. 1981. Structural evolution of an early Proterozoic stratabound Cu-Co-Zn deposit, Outokumpu, Finland. Trans. Royal Soc. Edinb., Earth sciences 72 (2), 115–158. Kontinen, A. 1987. An early Proterozoic ophiolite – the Jormua mafic-ultramafic complex, northeastern Finland. Precambrian Research 35, 313–341. Kontinen, A., Peltonen, P. & Huhma, H. 2006. Description and genetic modeling of the Outokumpu-type rock assemblage and associated sulphide deposits. GEOMEX / Final Technical Report, Archive report M 10.4 /2006/1. Korsman, K., Koistinen, T., Kohonen, J., Wennerström, M., Ekdahl, E., Honkamo, M., Idman, H. & Pekkala, Y. (Eds.) 1997. Bedrock Map of Finland 1:1 000 000. Geological Survey of Finland. Kukkonen, I. T. & Lahtinen, R. (eds.) 2006. Finnish Reflec tion Experiment FIRE 2001–2005. Geological Survey of Finland, Special Paper 43. 247 p. Lajaunie, C., Courrioux, G. & Manual, L. 1997. Foliation fields and 3D cartography in geology: principles of a method based on potential interpolation: Mathematical Geology v. 29 (4), 571–584. Laine, E. & Saalmann, K. 2008. 3D modeling – Outokumpu area as a case study: discussion of aims and approach. In: Third International Geomodelling Conference 22–24 September 2008 Villa la Pietra Firenze, Italy, Extended abstracts (edited by Giacomo Corti). Bollettino di Geofisica teorica ed applicata, Vol. 49 – N. 2 supplement, 512–514. Laine, E. (ed.), 2012. 3D modeling of polydeformed and metamorphosed rocks: the old Outokumpu Cu-Co-Zn mine area as a case study. Geological Survey of Finland, Report of Investigation 195. Laine, E., Koistinen, E., Saalmann, K., Couirrioux, G., Diaz, R., Salminen, N. & Tervo, T. 2012. 3D modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from the Outokumpu area. In: Laine, E. (ed.) 3D modeling of polydeformed and metamorphosed rocks: the old Outokumpu Cu-Co-Zn mine area as a case study. Geological Survey of Finland, Report of Investigation 195. McInerney, G., Courioux, C. & Lees, 2005. Building 3D geological models directly from the data? A new approach applied to Broken Hill, Australia. In: Soller, D. R. (ed.), Digital Mapping Techniques ‘05; Workshop Proceedings. US Geological Survey, Open-File Report 2005-1428, Baton Rouge, USA: 119–130. Peltonen, P. & Kontinen, A. 2004. The Jormua Ophiolite: a mafic-ultramafic complex from an ancient ocean-continent transition zone. In: Precambrian ophiolites and related rocks. Developments in Precambrian geology 13. Amsterdam: Elsevier, 35–71. Peltonen, P., Kontinen, A., Huhma, H. & Kuronen, U. 2008. Outokumpu revisited: New mineral deposit model for the mantle peridotite-associated Cu–Co–Zn–Ni–Ag–Au sulphide deposits. Ore Geol. Rev. 33, 559–617. Salminen, N. 2009. Microstructures in 3D modelling, Outokumpu area as an example site. Master’s thesis, Department of Civil and Environmental Engineering, Geoenvironmental technology. 59 p. , 9 app. Vuollo, T. & Piirainen, J. 1989. Primary minerals in the serpentinites of the Outokumpu Complex, North Karelia, Finland: contributions to the origin. In: 28th International Geological Congress, Washington, D.C. USA, 9–19 July 1989. Abstracts, p. 313. Västi, 2011. In: Kukkonen, I.T. (ed.). High resolution reflection seismics integrated with deep drill hole data in Outokumpu, Finland. Geological Survey of Finland, Special Paper 51, 105–118. Väyrynen, H. 1939. On the geology and tectonics of the Outokumpu ore field and region. Bulletin de la Commission Géologique de Finlande. 124. 3D-pdf’s Appendix 1: http://arkisto.gtk.fi/3d/outokumpu_geo_3d.pdf Appendix 2: http://arkisto.gtk.fi/3d/outokumpu_topo_3d.pdf 66 Appendix 3: http://arkisto.gtk.fi/3d/vuonos_ore_body_ and_structures_interpreted_from_geological_and_seismic_sections.pdf 3D modeling of polydeformed and metamorphosed rocks: the old Outokumpu Cu-Co-Zn mine area as a case study Edited by Eevaliisa Laine Geological Survey of Finland, Report of Investigation 195, 2012 THREE-DIMENSIONAL MODELING OF THE OUTOKUMPU NAPPE AREA, SE FINLAND by Tapio Ruotoistenmäki and Esko Koistinen, GTK Ruotoistenmäki, T. & Koistinen, E. 2012. Three-Dimensional Modeling Of The Outokumpu Nappe Area, SE Finland. Geological Survey of Finland. Report of Investigation 195, 8 figures. The three dimensional geometry of the Outokumpu nappe area can be evaluated using a combination of existing geophysical data and maps. The trend of the outcropping trace of the buried nappe structure at depth can be constrained from the long, continuous Haaralanniemi anomaly zone, discernible in both electromagnetic and magnetic maps. The deep cross section is determined from seismic and gravimetric interpretations. In the Juojärvi area the magnetic anomalies define the interfering fold structure of nappe basement dipping NE. In the Miihkali area the Haaralanniemi anomaly borders the local Miihkali basin, representing the uppermost parts of the nappe. The base of the nappe structure is interpreted to be cut and thrusted over itself below the Kylylahti-Sola anomaly zone. The model used here provides a valuable qualitative tool for extrapolating the position of Outokumpu association rocks at depth. However, the base of the nappe has been cut by Maarianvaara granitic rocks. Moreover, the repeated stacking of thrust folds and faults is apparent in the whole Outokumpu area in vertical sections. The model data were digitized and prepared using Gemcom GEMS software and imported into Geosoft Oasis montaj system for further modeling. Keywords (GeoRef Thesaurus, AGI): tectonics, nappes, three-dimensional models, structural geology, geophysical methods, Precambrian, Outokumpu, Finland Tapio Ruotoistenmäki Geological Survey of Finland, P.O. Box 96, FI-02151 Espoo, Finland E-mail: [email protected] Esko Koistinen Geological Survey of Finland, P.O. Box 96, FI-02151 Espoo, Finland E-mail: [email protected] 67 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Tapio Ruotoistenmäki and Esko Koistinen 1 Introduction During 1998–2003, the geology, geophysics, evolution and metallogenic potential of the Outokumpu region in Northern Karelia, Finland, were reassessed through the Geomex project, which represented a joint venture between the Geological Survey of Finland and Outokumpu Mining Oy. The purpose of the project was to collect, reprocess and archive existing geological and geophysical data and material and locate new target areas for exploration. The geophysics subproject of the Geomex project was reported in Ruotoistenmäki and Tervo (2006). The report also presented regional and local-scale tectonic models based on geophysical and petrophysical data from the study area. In this article, a ‘three-dimensional’ model is presented to demonstrate the non-outcropping, deeper characteristics of the regional tectonic model for the Outokumpu nappe area given in Ruotoistenmäki and Tervo (2006). 2 General geology of the study area The general geological features of the Outokumpu area are illustrated in Figure 1 by Sorjonen-Ward and Luukkonen (2005). They classified the area into roughly four major units: Proterozoic granitoids, allochthonous Paleoproterozoic supracrus- tal rocks, parautochthonous Paleo-proterozoic sedimentary-volcanic rocks and Archaean basement windows. For a more detailed geological description of the area, see also Kontinen & Peltonen (2003). Fig. 1. Simplified local-scale lithological map of the Outokumpu area studied in the Geomex project (outlined by yellow lines). Semitransparent gray shades relate to the total magnetic intensity recorded by regional airborne surveys (high anomalies are darker; reproduced from databases of the Geological Survey of Finland). Modified from the map by Sorjonen-Ward and Luukkonen (2005). 68 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Three-Dimensional Modeling Of The Outokumpu Nappe Area, SE Finland 3 Major geophysical formations in the study area Some major regional (= continuous) electromagnetic anomaly zones in the study area are illustrated in Figure 2. In the map, the NE-trending Outokumpu anomaly zone (OZ) is the main ore potential zone in the area, rising to the SW against the NW-trending Juojärvi anomaly (JZ), and together forming a T-shaped anomaly group. Moreover, the Miihkali ‘basin’ (MB) and Kylylahti-Sola thrust zone (KS-T) form ore potential structures on the northern part of the nappe. A very important feature is the generally continuous Haaralanniemi anomaly zone (HA), which represents the outcropping edges of the basin of the whole Outokumpu nappe structure. In the Fig. 2. Electromagnetic real (in-phase) component map of the study area. The map has been compiled from data recorded by the Geological Survey of Finland. OZ = Outokumpu anomaly zone, JZ = Juojärvi anomaly, MB = Miihkali ‘basin’, HA = Haaralanniemi anomaly, KS-T = Kylylahti-Sola thrust zone, SW-D = southwest ‘deep’, WNB = western nappe basin (breakup of the Haaralanniemi anomaly). Map adopted and modified from Ruotoistenmäki and Tervo (2006). 69 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Tapio Ruotoistenmäki and Esko Koistinen southwest ‘deep’ (SW-D), the base of the nappe is not visible, but disappears below outcropping sedimentary rock cover. On the west, at WNB, however, the Haaralanniemi anomaly disappears, exposing the uplifted base of the nappe cut by younger, Proterozoic Maarianvaara granite. 4 Major structures of the Outokumpu nappe area When interpreting the geometry of the regional structures in the Geomex area, an important key feature is the Juojärvi magnetic anomaly (JZ in Fig. 3) in the SW part of the Outokumpu zone. In the anomaly zone, a striking feature is its sinuous NW–SE trend. The geometry of the Juojärvi anomaly zone can be best explained by the schematic fold model in the inset in the figure. In the model, the geometry of the upper edge of the folds emphasized by the blue line is analogous to that of the Juojärvi magnetic anomaly. The geometry suggests the presence of (at least) two perpendicular regional fold groups in the Juojärvi area, one with a SW–NE trend and another with a SE–NW-trending fold axis, as shown in Figure 4. From the geometry of the magnetic anomaly patterns, it can also be concluded that SW–NE folds preceded the SE–NW folds. In Figure 4, the geometry of the main Outo- kumpu zone is largely defined by the SW–NEtrending folds, represented by blue lines for the synformal fold axis and red lines for antiforms. The SE–NW folds include the regional synform NE from Juojärvi (JS) and the more deeply eroded Juojärvi antiform (JA). The SE-trending magnetic anomaly zone southwest of antiform JA in the figure represents the SW limb of the fold, which could therefore contain Outokumpu assemblage rocks. It must also be emphasized that the pattern of the folds in the Juojärvi anomaly zone actually reflects the geometry of the vertical cross-section of the SW–NE-trending folds (i.e. the geometry of the base of the nappe), as is evident from the schematic model in Figure 3. The intersecting SW–NE and SE–NW-trending antiforms southwest of the Juojärvi area also explain the locations of the Archaean basement windows SW of Juojärvi (Fig. 1). The antiforms Fig. 3. Schematic model to explain the Juojärvi anomaly (JZ). The synform marked by red in the lowermost inset refers to the Outokumpu anomaly zone (OZ). 70 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Three-Dimensional Modeling Of The Outokumpu Nappe Area, SE Finland interfere at locations emphasized with red circles in in Figure 4, thus generating antiformal domes and exposing the Archaean basement, as depicted in in Figure 1. Moreover, the geological dip obser- vations agree with the inferred domal interference patterns as shown in Ruotoistenmäki and Tervo (2006). Fig. 4. General tectonic features interpreted from geophysical maps of the study area. The base map is the magnetic low-altitude map (for abbreviations see text and Figure 2). The map is adopted and modified from Ruotoistenmäki and Tervo (2006). 71 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Tapio Ruotoistenmäki and Esko Koistinen 5 Preparation of the three-dimensional model The preparation of the three-dimensional model was based on the two-dimensional interpretation shown in Figure 4. and seismic refraction and reflection profiles crossing the nappe area (Penttilä 1967 and Kukkonen et al. 2006). In the model, the nappe area was divided in cross-sections shown in in Figure 5 a and b. The model consists of three basic sub-blocks: 1) the northernmost block north of the Kylylahti-Sola thrust zone (‘Oku-North’), 2) the main nappe zone (‘OkuSouth’) and 3) the Miihkali basin (MB). In Figure 5b, violet curves define the modeled basin of Oku-North, blue curves refer to Oku-South and red defines the course of Juojärvi antiformal interference domes. The blocks were modelled in 3D by the LaPlace Fig. 5. (a): Location of selected cross-sections used for constructing the three-dimensional model. Yellow curves define the locations of the outcropping edges of the modeled structures. (b): The inferred variations of the geometry of the Outokumpu nappe basin based on the model presented in Figure 4. (c): The three-dimensional model derived by the LaPlace gridding method of Gemcom GEMS. The Outokumpu basin surface bordered by the Haaralanniemi anomaly is colored in magenta and the Miihkali basin is in green. Antiforms are shown by blue curves, synforms by red and thrust zones by black curves. 72 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Three-Dimensional Modeling Of The Outokumpu Nappe Area, SE Finland gridding method of Gemcom GEMS. The model data were further imported into the Geosoft Oasis montaj system, in which the data were compiled in the three-dimensional images presented in Figures 6 to 8. In the figures, the magnetic map is projected on the model consisting of three submodels built in succession from the lowermost Oku-North ‘layer’ to the uppermost Miihkali ba- sin. In the model, the yellow lines define the approximate outcropping of the sub-models. It must be noted that the Oku-North and Okusouth layers do not necessarily imply primary vertical succession. More probably, the nappe has been ‘broken’ during the thrust processes along the KS-T line and the southern part has been thrust over the separated northern block. Fig. 6. Three-dimensional model of the northern- and lowermost section (Oku-North) of the Outokumpu nappe (below Miihkali basin in Figure 4) Fig. 7. Three-dimensional model of the main SW block of the Outokumpu nappe (Oku-South) thrust over the NW section in Figure 6. 73 Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Tapio Ruotoistenmäki and Esko Koistinen Fig. 8. Three-dimensional model of the Miihkali basin overlying the NW and SW sections in Figure 6 and Figure 7. The approximate location of the Haaralanniemi anomaly (HA, Fig. 2), Kylylahti-Sola thrust zone (KS-T) and southwest ‘deep’ (SW-D) have also been marked in the figure. 6 Conclusions The three-dimensional geometry of the Outokumpu nappe area can be derived by carefully combining information from existing surface geophysics, geology and deep seismic studies. In the Outokumpu area, a very important geophysical key anomaly is the long continuous Haaralanniemi anomaly representing outcropping edges of the base of the nappe, and thus defining the borders of the nappe area. The Juojärvi anomaly defines the SW geometry of the nappe and its interfering fold structure. Moreover, the Kylylahti-Sola anomaly represents the NE thrust zone where the Outokumpu nappe has been cut into two major blocks, where the southern block is thrusted over the NE block. The Miihkali basin represents the uppermost part of a deep interference synform on top of the Outokumpu nappe structure. These anomalies of the Outokumpu nappe area are good examples of two-dimensional features that can provide valuable three-dimensional information when carefully studied. References Kontinen, A. & Peltonen P. 2003. Description and genetic modelling of the Outokumpu type rock assemblage and sulphide ores – a GEOMEX subproject. GEOMEX / Preliminary Technical Report. Geological Survey of Finland. Kukkonen, I.T., Heikkinen, P., Ekdahl, E.,Hjelt, S.- E., Korja, A., Lahtinen, R., Yliniemi, J. & FIRE Working Group. 2006. FIRE reflection seismic transects: New images of the crust in the Fennoscandian Shield. Abstract in: 27th Nordic Geological Winter Meeting, Oulu, 9.–12.01.2006. Bulletin of the Geological Society of Finland, Special Issue 1, 2006. 74 Penttilä, E. 1967. Outokumpu jakson räjäytysseismologisesta tutkimustyöstä. On the seismological study of the Outokumpu zone. Report by Outokumpu Oy (in Finnish). Ruotoistenmäki, T. & Tervo, T. 2006. Geophysical Characteristics of the Outokumpu Area, SE Finland. Geological Survey of Finland, Report of Investigation 162, 37 p. Sorjonen-Ward, P. & Luukkonen E. J. 2005. Archean rocks. In: Lehtinen, M., Nurmi, P. A., Rämö, O. T. (eds.), 2005. Precambrian Geology of Finland – Key to the Evolution of the Fennoscandian Shield. Developments in Precambrian Geology 14. Amsterdam: Elsevier, 19–99. Geologian tutkimuskeskus, Tutkimusraportti 195 – Geological Survey of Finland, Report of Investigation 195, 2012 Three-Dimensional Modeling Of The Outokumpu Nappe Area, SE Finland Acknowledgements Many thanks to referees for reading this report and offering valuable comments and to Roy Siddall for English checking. I owe also thanks to my daughter Carmela Lönnqvist for helping with illustrations. Copyright owner The Royal Society of Edinburgh/ RSE Scotland Foundation and Tapio Koistinen Geological Survey of Finland The publishers and authors listed below are gratefully acknowledged for giving their permission to use original and redrawn figures based on illustrations in journals they hold the copyright. Journal Transactions of the Royal Society of Edinburgh: Earth Sciences Figure number Part I (Problems and challenges of 3D modeling of the Precambrian of Finland): Figures 4, 6 and 12, the cover page Part I (Problems and challenges of Geological Survey of Finland. Bulletin 3D modeling of the Precambrian of Finland): Figures 3, 5, 7 Part II (3D modeling of polydeformed and metamorphosed rocks at different scales using geological and geophysical data from Outokumpu area): Figures 1, 26, 27, 30, 31, 32, 37, 38, 39 75 www.gtk.fi [email protected] The present work summarizes the results of the project “3D/4D modeling – Outokumpu area as a case study” conducted during 2007–2009 and 3D modeling of the Outokumpu assemblage in the project “Development of geological 3D modeling using structural geology and geostatistics” in 2010 at the Geological Survey of Finland (GTK). The aims of these projects were to apply and evaluate 3D/4D modeling software and formulate principles for visualizing and analyzing complicated geological structures typical of Finnish bedrock. The aim was to make recommendations for the use of 3Dmodeling tools and processes for 3D/4D modeling at GTK. Several 3D geological models were constructed at different scales using data from the Outokumpu mining area. ISBN 978-952-217-182-5 (PDF) ISSN 0781-4240