Repositorio Institucional de Documentos

2013 147 Teresa Ubide Garralda The Cretaceous alkaline magmatism in northeast Iberia: igneus processes and geodynamic implications Departamento Director/es Ciencias de la Tierra Arranz Yagüe, Enrique Lago San José, Marceliano Director/es Tesis Doctoral Autor Repositorio de la Universidad de Zaragoza – Zaguan http://zaguan.unizar.es UNIVERSIDAD DE ZARAGOZA Departamento Director/es Teresa Ubide Garralda THE CRETACEOUS ALKALINE MAGMATISM IN NORTHEAST IBERIA: IGNEUS PROCESSES AND GEODYNAMIC IMPLICATIONS Director/es Ciencias de la Tierra Arranz Yagüe, Enrique Lago San José, Marceliano Tesis Doctoral Autor 2013 Repositorio de la Universidad de Zaragoza – Zaguan http://zaguan.unizar.es UNIVERSIDAD DE ZARAGOZA Departamento Director/es Director/es Tesis Doctoral Autor Repositorio de la Universidad de Zaragoza – Zaguan http://zaguan.unizar.es UNIVERSIDAD DE ZARAGOZA THE CRETACEOUS ALKALINE MAGMATISM IN NORTHEAST IBERIA. IGNEOUS PROCESSES AND GEODYNAMIC IMPLICATIONS TERESA UBIDE GARRALDA PhD THESIS 2013 UNIVERSIDAD DE ZARAGOZA FACULTAD DE CIENCIAS DEPARTAMENTO DE CIENCIAS DE LA TIERRA GRUPO DE INVESTIGACIÓN GEOTRANSFER Supervisors Dr. Enrique Arranz Yagüe Dr. Marceliano Lago San José A mis padres Contents Agradecimientos Resumen Summary Chapter 1. Introduction ........................................................................................... 1 1.1. Interest ....................................................................................................................................... 3 1.2. Objectives ................................................................................................................................... 7 1.3. Outline ........................................................................................................................................ 8 Chapter 2. Materials and Methods ........................................................................ 11 2.1. Samples .....................................................................................................................................13 2.2. Methodology .............................................................................................................................13 2.2.1. Field work .......................................................................................................................13 2.2.2. Laboratory work .............................................................................................................13 2.2.3. Data treatment ...............................................................................................................18 2.2.4. Presentation of results ...................................................................................................18 Chapter 3. Age of the Cretaceous alkaline magmatism in northeast Iberia: implications for the Alpine cycle in the Pyrenees .................................................. 21 3.1. Introduction ...............................................................................................................................23 3.2. The Cretaceous magmatism in northeast Iberia: geological context, characterisation and previous dating results ..............................................................................................................24 3.2.1. Pyrenees .........................................................................................................................24 3.2.2. Catalonian Coastal Ranges ..............................................................................................26 3.3. Samples and methods ...............................................................................................................27 3.3.1. Location and description of the samples ........................................................................27 3.3.2. 40Ar/39Ar methodology ....................................................................................................30 3.4. Results .......................................................................................................................................32 3.5. Discussion ..................................................................................................................................36 3.5.1. The Triassic-Jurassic rifting in the Central Pyrenees ......................................................36 3.5.2. The Cretaceous opening of the Bay of Biscay and rotation of Iberia .............................37 3.5.3. The onset of Alpine compression in the Late Cretaceous ..............................................43 3.6. Conclusions ................................................................................................................................45 3.7. Appendix ....................................................................................................................................47 Chapter 4. Petrogenesis of the Cretaceous alkaline magmatism in northeast Iberia: new insights into the evolution of the mantle source ................................ 51 4.1. Introduction .............................................................................................................................. 53 4.2. Geological context, igneous typologies and previous studies .................................................. 53 4.3. Samples and methods .............................................................................................................. 55 4.4. Petrology .................................................................................................................................. 62 4.5. Geochemistry ........................................................................................................................... 65 4.5.1. Classification and affinity ............................................................................................... 65 4.5.2. Major and trace element variations .............................................................................. 67 4.5.3. Sr-Nd-Pb isotope composition ....................................................................................... 71 4.6. Discussion ................................................................................................................................. 73 4.6.1. Source enrichment and mantle heterogeneity ............................................................. 73 4.6.2. Evolution of the mantle source since the beginning of the Alpine cycle ...................... 78 4.7. Conclusions ............................................................................................................................... 83 Chapter 5. The Cretaceous lamprophyre suite in the Catalonian Coastal Ranges: complex antecryst populations and their effect on whole rock compositions ...... 85 5.1. Introduction .............................................................................................................................. 87 5.2. Geological context .................................................................................................................... 88 5.3. Field appearance of the lamprophyres .................................................................................... 88 5.4. Samples and methods .............................................................................................................. 95 5.5. Petrography .............................................................................................................................. 97 5.5.1. The macrocryst assemblage .......................................................................................... 97 5.5.2. The groundmass .......................................................................................................... 100 5.6. Mineral chemistry ................................................................................................................... 101 5.6.1. Major element composition ........................................................................................ 101 5.6.2. Trace element composition ......................................................................................... 111 5.7. Whole rock chemistry ............................................................................................................. 113 5.7.1. Major element composition ........................................................................................ 113 5.7.2. Trace element composition ......................................................................................... 116 5.8. Discussion ............................................................................................................................... 117 5.8.1. Origin of mineral compositional variations ................................................................. 117 5.8.2. Origin of whole rock compositional variations ............................................................ 123 5.8.3. Reconstruction of the magmatic plumbing system ..................................................... 126 5.9. Conclusions ............................................................................................................................. 129 Chapter 6. The influence of crystal settling on the compositional zoning of a thin lamprophyre sill: a multi-method approach ................................................. 131 6.1. Introduction ............................................................................................................................ 131 6.2. Field observations .................................................................................................................. 134 6.3. Samples and methods ............................................................................................................ 136 6.4. Petrography ............................................................................................................................ 138 6.5. Mineral chemistry ................................................................................................................... 139 6.6. Whole rock chemistry ............................................................................................................. 139 Resumen Durante la apertura del océano Atlántico, responsable de la ruptura del supercontinente Pangea, la apertura del Golfo de Vizcaya provocó la rotación y separación de Iberia respecto de Europa. Este evento condicionó el desarrollo de un magmatismo alcalino, previo al inicio de la colisión alpina que desencadenó la elevación de la cadena pirenaica entre Iberia y Europa. Dicho magmatismo alcalino es de edad cretácica y hoy se reconoce en los Pirineos y en el margen septentrional de las Cadenas Costero Catalanas. Sin embargo, no existen, hasta la fecha, estudios completos ni comparativos de estas rocas. Esto contrasta con su enorme interés científico, pues suponen una preciada fuente de información acerca del controvertido contexto geodinámico, así como de los procesos ígneos relacionados con la generación de fundidos en el manto y su ascenso a través de la corteza. Así, la presente Tesis Doctoral comprende la petrología, mineralogía, geoquímica y geocronología del magmatismo cretácico en todo el noreste de Iberia, incluyendo afloramientos en los Pirineos y con especial énfasis, en el margen septentrional de las Cadenas Costero Catalanas, ya que este dominio ha sido menos estudiado. Se han obtenido nuevas edades 4 0 Ar/ 39 Ar que actualizan los datos previos y delimitan temporal y geográficamente la provincia ígnea. Además, la edad del magmatismo, asociado a un rifting continental, aporta nuevos argumentos a considerar para evaluar las teorías sobre la evolución geodinámica del límite Iberia-Europa, tanto respecto a la apertura del Golfo de Vizcaya como al inicio de la compresión alpina. El estudio de la composición de las rocas en elementos mayores, elementos traza y relaciones isotópicas de Sr-Nd-Pb revela una signatura heterogénea y enriquecida del manto en este dominio, que podría haber perdurado desde el final de la orogenia varisca hasta el cuaternario. La composición de los minerales en elementos mayores y traza, junto con las estimaciones barométricas realizadas, informan de manera muy detallada acerca del origen de las poblaciones de cristales, complejas, que se observan en las rocas, revelando procesos de recarga y mezcla de magmas en profundidad y la incorporación de cristales heredados en magmas ascendentes hacia la superficie terrestre. Estos datos permiten reconstruir la historia de los procesos que tuvieron lugar en el sistema de alimentación durante el ascenso de los magmas y su evolución en cámaras magmáticas. En este trabajo se realiza la primera cuantificación de la influencia de cristales heredados (tipo antecristal) en la composición global de las rocas en las que se encuentran. El modelo geoquímico desarrollado muestra que la acumulación de antecristales máficos provoca modificaciones significativas en las composiciones de roca total, y esto cuestiona la interpretación tradicional de las composiciones magmáticas “primitivas”, así como las ideas sobre procesos de acumulación gravitatoria de cristales y zonación composicional en cuerpos ígneos. La aplicación, en esta tesis, de métodos matemáticos como la elaboración de modelos cuantitativos y el análisis estadístico de componentes principales al tratamiento de datos geoquímicos, es novedosa en este campo científico y abre el camino a una mejor evaluación de hipótesis petrológicas en base a datos geoquímicos completos, proporcionando una metodología de trabajo potencialmente aplicable al estudio de sistemas ígneos. Finalmente, es importante destacar que los datos geoquímicos cobran todo su significado únicamente cuando se comprende, en detalle, la petrología de las rocas y las condiciones de equilibrio entre los minerales y el fundido que los engloba. Esto, además, lleva al cálculo, por primera vez, de un conjunto internamente consistente de coeficientes de reparto de elementos traza para clinopiroxeno y anfíbol en fundidos de tipo lamprófido alcalino (camptonita). Summary During the opening of the Atlantic Ocean, which was responsible for the break-up of the Pangea supercontinent, the opening of the Bay of Biscay triggered the rotation and separation of Iberia with respect to Europe. This event conditioned the development of alkaline magmatism, prior to the onset of Alpine collision eventually leading to the formation of the Pyrenees between the Iberian and European domains. The alkaline magmatism is Cretaceous in age and is found in the Pyrenees and in the northernmost Catalonian Coastal Ranges. However there have been to date neither complete nor comparative studies of these rocks. This is in spite of their important scientific significance as a valuable source of information on what is a controversial geodynamic context, as well as on igneous processes related to melt generation in the mantle and its ascent through the crust. Given this lack of information, the present PhD Thesis addresses the petrology, mineralogy, geochemistry and geochronology of the Cretaceous magmatism along northeast Iberia, including igneous occurrences in the Pyrenees and especially in the, less studied, northernmost Catalonian Coastal Ranges. New 40 Ar/ 39 Ar age determinations constrain temporally and geographically previous notions about the alkaline magmatism associated to continental rifting, and provide new data to evaluate the theories about the Iberia-Europe geodynamic evolution regarding both the opening of the Bay of Biscay and the onset of Alpine compression. The study of the major element, trace element and Sr-Nd-Pb isotope concentrations in the rocks reveals a heterogeneous and enriched mantle source signature that might have prevailed in this area from Late Variscan to Quaternary times. The major and trace element composition of the minerals, together with calculated barometric data, inform in great detail about the origin of complex crystal populations observed in the rocks, revealing deep magma recharge and mixing processes and entrainment of inherited crystals into magma batches ascending towards the Earth’s surface. This leads the way to reconstruct the history of processes taking place in the plumbing system during magma ascent and evolution in magma chambers. The novel quantification of the influence of inherited crystals (antecrysts) on the bulk composition of the rocks they are included in, should be especially noted. It shows that the accumulation of mafic antecrysts modifies significantly whole rock compositions, and this brings into question the traditional interpretation of “primitive” magmatic compositions, as well as notions about crystal settling processes and compositional zoning in igneous bodies. The application in this thesis of mathematical methods such as quantitative modelling and statistical principal component analysis to the treatment of geochemical data is innovative in this field of science. It opens the way to a better evaluation of petrological hypotheses on the basis of full compositional datasets, providing a methodology potentially applicable to the study of igneous systems. Finally, it is well recognised that geochemical data can only be fully understood when the petrology of the rocks and mineral-melt equilibrium conditions are well known. Such an in-depth analysis in this work has moreover allowed the calculation, for the first time, of a consistent partition coefficient dataset of trace elements for clinopyroxene and amphibole in alkaline lamprophyre (camptonite) melts. Theory guides, Experiment decides Izaak M. Kolthoff 1 1 Introduction Platja Fonda, Costa Brava Chapter 1 · Introduction 2 Introduction · Chapter 1 3 1.1. Interest The Cretaceous Era was characterised by widespread alkaline igneous activity on widely separated peri-Atlantic continental margins (Matton and Jébrak, 2009; Fig. 1.1). Magmatism concentrated in the period between 125 and 80 Ma has received the name Peri-Atlantic Alkaline Pulse (PAAP; Matton and Jébrak, 2009). In the Iberian Peninsula, it includes outcrops in southern Portugal and northern Spain (e.g., Rock, 1982; Solé et al., 2003). Cretaceous pulses in alkaline magmatic activity were closely associated with major Atlantic tectonic events (Matton and Jébrak, 2009). In northern Spain, the opening of the Bay of Biscay (Fig. 1.1) led to the anticlockwise rotation and separation of Iberia with respect to Europe (Vissers & Meijer, 2012a and references therein). This was followed by the onset of Alpine collision between Iberia and Europe from the Late Cretaceous, eventually leading to the formation of the Pyrenees (Vissers & Meijer, 2012b and references therein). The rifting scenario generated by the opening of the Bay of Biscay thinned the crust between the Iberian and European domains (e.g., Jammes et al., 2009). This enabled not only the ascent of alkaline magmas, but also the development of high temperature and low pressure metamorphism, hydrothermalism and mantle exhumation (Golberg et al., 1986; Lagabrielle and Bodinier, 2008; Lagabrielle et al., 2010; Montigny et al., 1986; Poujol et al., 2010; Schärer et al., 1999). Today, Cretaceous alkaline magmatism is recognised as extrusive and intrusive, mafic and felsic rocks along the Pyrenees and as isolated lamprophyre intrusions in the northern part of the Catalonian Coastal Ranges (e.g., Azambre et al., 1992; Solé et al., 2003). In detail, this magmatism crops out in four main domains (Fig. 1.2): Basque-Cantabrian Basin in the western Pyrenees, North Pyrenean Basins, Eastern Pyrenees and northern Catalonian Coastal Ranges. In addition, there is a geographically restricted alkaline magmatism in the Central Pyrenees (Galé et al., 2000; Mey, 1968; Fig. 1.2) that could be Cretaceous on the basis of the geochemical affinity of the rocks and preliminary 87 Rb/ 86 Sr data (Galé and Arranz, 2001). The Cretaceous alkaline magmatism in northeast Iberia has a threefold interest. Firstly, the magmatism has traditionally been related to the opening of the Bay of Biscay and allied rotation of Iberia (e.g., Montigny et al., 1986; Rock, 1982) and predates the beginning of Iberia-Europe collision (Solé et al., 2003; Vissers and Meijer, 2012a). Although these geodynamic events have attracted considerable interest in geoscientists, they remain poorly understood (e.g., Jammes et al., 2009; Rosenbaum et al., 2002; Vissers Chapter 1 · Introduction 4 Fig. 1.1. Reconstruction of the peri-Atlantic realm at 100 Ma with occurrences of Cretaceous alkaline igneous rocks denoted as circles (modified after Matton and Jébrak, 2009). The magmatism studied in this PhD thesis (red circle) is related to the opening of the Bay of Biscay and predates the onset of Alpine collision between Iberia and Europe. 11 2 Materials and Methods SUMMARY This chapter includes a summary of the samples used as well as a general description of the methodologies applied. Field work was followed by petrological and mineralogical studies and a diversity of geochemical and geochronological analyses, for which several laboratories were used in Spain and abroad. The large amount of new data, together with a thorough review of previous studies, led to interpretations on magmatic processes, mantle sources and geodynamic conditions. During this Thesis, new sample preparation and data treatment techniques have been developed and geochemical interpretations have been closely linked to a deep understanding of the rocks. Punta des Mut, Costa Brava Chapter 2 · Materials & Methods 12 Materials & Methods · Chapter 2 13 2.1. Samples The Cretaceous magmatism in northeast Iberia crops out in several sectors across the Pyrenees and in the northern Catalonian Coastal Ranges (see Fig. 1.2). Previous studies on this magmatism were carried out with different levels of detail and the magmatism in the Catalonian Coastal Ranges was particularly unstudied. Therefore, sampling covered all the sectors but focused especially on those in the Catalonian Coastal Ranges. The samples are representative of the diversity of magmatic products, with special emphasis on mafic compositions as they are closer to the mantle source and primary magmatic processes. A summary of samples is presented in Table 2.1. This table includes location, classification and methods applied to each sample. On the basis of macroand microscopic examinations, the most representative and unaltered samples were selected for geochemical and geochronological analyses. 2.2. Methodology 2.2.1. Field work After a careful review of previous studies and cartographic material, numerous field campaigns were undertaken across northeast Iberia. Field observations on the petrology, structure and emplacement characteristics of the magmatic outcrops leaded to sampling of the least altered material in each case. Field work was especially relevant in the Catalonian Coastal Ranges, given that this domain was poorly studied and several outcrops had never been reported before. 2.2.2. Laboratory work This section mentions all the techniques applied during this PhD project and Chapters 3 to 7 include a detailed description of the specific samples and methods used in each study. The samples were first studied macroscopically to select one or several areas of interest for microscopic examination. The preparation of thin sections of the samples (3 0 µm thickness) was ordered to the Servicio General de Apoyo a la Investigación-SAI of the Universidad de Zaragoza (Spain). The thin sections were studied under the petrographic microscope of the Servicio General de Apoyo a la Investigación-SAI of the Universidad de Zaragoza (Spain) to describe the petrography of the samples (texture, mineral assemblage, crystallisation sequence, crystal stratigraphy) and select minerals and rocks for geochemical and geochronological studies. Chapter 2 · Materials & Methods 14 Table 2.1. Summary of samples studied and methods applied. Domain Sector Sample Classification Thin sections EMP a LA-ICP-MSa ICP-AES ICP-MS b Rb/Sr Sm/Nd Pb/Pb b Ar/Ar AGE c Basque-Cantabrian Basin Bilbao SGR gabbro 1 1 1 1 Guernica FR-1 trachybasalt 3 1 1 FR-3 trachybasalt 1 FR-4 trachybasalt 1 FR-5 trachybasalt 3 FR-6 trachybasalt 2 1 FR-7 trachybasalt 1 FR-8 trachybasalt 2 1 FR-9 trachybasalt 1 1 1 Vergara ELG-01 picrite 1 ELG-02 picrite 1 ELG-03 picrite 1 ELG-04 picrite 3 1 1 ELG-05 picrite 1 ELG-06 picrite 1 ELG-07 picrite 1 ELG-08 basalt 1 1 1 1 ELG-09 basalt 1 1 MALZA-1 trachyte 1 1 1 MALZA-B basalt 3 1 1 URR gabbro 3 1 2 URR-2 gabbro 1 URR-3 gabbro 1 1 1 8057 basaltic trachyandesite 1 1 1 8058 basaltic trachyandesite 1 8059 basaltic trachyandesite 2 8060 basaltic trachyandesite 1 North Pyrenean Basins OloronSainteMarie AUB monzonite 1 1 1 1 COG-3 basaltic trachyandesite 1 1 COG-4 basaltic trachyandesite 1 COG-5 monzonite 1 CP: Central Pyrenees. Classification criteria according to Le Maitre (2002). (a) number of polished thin sections analysed; (b) number of whole rock powders analysed; (c) number of separated fractions analysed; (d) equivalent to samples 1, 2, 3, 4 and 5 in Chapters 6 and 7. Materials & Methods · Chapter 2 15 Table 2.1 (cont) Domain Sector Sample Classification Thin sections EMP a LA-ICP-MS a ICP-AES ICP-MS b Rb/Sr Sm/Nd Pb/Pb b Ar/Ar AGE c Eastern Pyrenees Corbières CORB-1 monchiquite 1 1 CORB-1B monchiquite 1 CORB-1C monchiquite 2 CORB-2 monchiquite 2 1 1 1 CORB-3 monchiquite 1 1 CORB-4 monchiquite 1 CORB-5 monchiquite 1 CORB-6 monchiquite 2 1 VIN-1 sannaite 1 1 VIN-2 sannaite 4 1 1 1 VIN-3 sannaite 1 VIN-4 sannaite 2 FIT-1 syenite 1 1 FIT-2 syenite 2 1 1 1 FIT-3 syenite 1 1 CP Denuy DEN1 basalt 1 1 DEN2 basalt 4 4 1 Catalonian Coastal Ranges East AIG-04 camptonite 2 2 1 AIG-05 camptonite 1 1 1 2 AIG-06 camptonite 1 1 AIG-08 camptonite 1 1 1 AIG-12 camptonite 1 CAL-1 d camptonite 1 1 1 1 1 CAL-2 d camptonite 1 1 1 1 1 CAL-3 d camptonite 1 1 1 1 CAL-4 d camptonite 1 1 1 CAL-5 d camptonite 1 1 1 Centre SAG-03 camptonite 1 1 SAG-03B camptonite 1 1 1 1 LLAG-01 camptonite 2 2 1 1 2 West SFB-1(r) camptonite 1 1 1 SFB-2(r) camptonite 1 1 1 SFB-3 camptonite 2 1 1 1 SFB-4 camptonite 2 1 1 SFB-5 camptonite 1 1 1 SFB-6 camptonite 2 1 1 1 1 1 Chapter 2 · Materials & Methods 16 Fig. 2.1. Photomicrographs of a complexly zoned clinopyroxene macrocryst set in a fine-grained groundmass, which includes amphibole microcrysts. Plane-polarised (A) and cross-polarised (B) transmitted light. The area is circled in red for spot analyses. The different zones in the macrocryst (white circles) reveal different melt compositions and a complex magma history (see Chapters 5 and 7 in this volume). The rock is a Cretaceous lamprophyre from the Catalonian Coastal Ranges (sample SFB-4). The mineral chemistry study included mineral classification and especially, a careful analysis of crystal populations. This was particularly relevant in porphyritic lamprophyres from the Catalonian Coastal Ranges, which include macrocrysts with complex zoning patterns as well as groundmass microcrysts (Fig. 2.1). Mineral compositions were analysed on polished thin sections ordered to the Servicio General de Apoyo a la Investigación-SAI of the University of Zaragoza (Spain). Major element concentrations were determined by electronic microprobe (EMP) at the Centro Nacional de Microscopía Electrónica of the Universidad Complutense de Madrid (Spain) and trace element concentrations were determined by inductively coupled plasma mass spectrometry with laser ablation (LA-ICP-MS) at the Centro de Instrumentación Científica-CIC of the Universidad de Granada (Spain). Samples for whole rock chemistry were carefully selected considering the nature of the rocks (primitive or evolved, porphyritic or microlitic, type of macrocryst assemblage; Fig. 2.2). Whole rock elemental and Sr-Nd-Pb isotope compositions were analysed on powders obtained by crushing, milling and splitting representative aliquots of sample at the Servicio General de Apoyo a la Investigación-SAI of the University of Zaragoza (Spain). Elemental analyses were ordered to the Service d'Analyse des Roches et des Minéraux (SARM) of the Centre des Recherches Pétrographiques et Géochimiques in Nancy (France). Major element concentrations were obtained by inductively coupled plasma atomic emission spectroscopy (ICP-AES), also referred to as Fig. 2.2. Scanned thin sections of microlitic (left) and porphyritic (right) lamprophyres from the Catalonian Coastal Ranges (the width of each image is 2.5 cm). Whole rock compositions are strongly influenced by the volume fraction and type of macrocrysts (see Chapters 5 and 6 in this volume). Materials & Methods · Chapter 2 17 inductively coupled plasma optical emission spectrometry (ICP-OES). Trace element concentrations were obtained by inductively coupled plasma mass spectrometry (ICPMS). Sr-Nd-Pb isotope analyses were ordered to the SGIker-Geochronology and Isotopic Geochemistry facility of the Universidad del País Vasco UPV-EHU (Spain), using thermal ionisation mass spectrometry (TIMS) and multi collector inductively coupled plasma mass spectrometry (MC-ICP-MS). Additionally, one of the samples was analysed at the Isotope Geochemistry & Mass Spectrometry Laboratory of the Miami University (Ohio, USA), using TIMS. 40 Ar/ 39 Ar geochronology was carried out on mineral and groundmass separates at the Argon Geochronology laboratory of the VU University Amsterdam (The Netherlands). For this end, two research stays were done under the supervision of Prof. Jan R. Wijbrans. During the first research stay (January 2010), the samples were prepared for analysis by crushing, sieving and separating the fractions of interest (groundmass, amphibole or biotite, depending on the sample) using heavy liquids, Frantz® magnetic separator and Faul vibration table, as appropriate. The separated fractions were purified by HF-leaching and hand-picking under the microscope. To visually differentiate amphibole grains from clinopyronoxene, the fractions were observed over a white surface (Fig. 2.3; see details in Ubide et al., 2010a and Chapter 3 in this volume). The final separates were sent for irradiation to the High Flux Reactor (HFR) at Petten (The Netherlands). During the second research stay (September – December 2010), Ar isotope ratios were determined on the irradiated separates to obtain crystallisation ages. The separates were first preheated and then analysed by incremental heating with laser ablation. The gas was purified in an ultrahigh vacuum extraction line and analysed with a magnetic sector noble gas mass spectrometer (see details in Chapter 3). Fig. 2.3. P ho to mi cr o gr a ph s of a 25 0 – 300 µm sample fraction including clinopyroxene and amphibole. Only amphibole was of interest for 40 Ar/ 39 Ar geochronology due to its higher concentrations in K 2 O. However, it is hard to distinguish from clinopyroxene, as they have similar physical properties and a black, shiny appearance (left image). This was solved by observing the separates over a white surface and best results were obtained adding acetone (right image): brown grains are amphibole and pale-green fragments are clinopyroxene. Chapter 2 · Materials & Methods 18 2.2.3. Data treatment Compositional data were mainly treated with ad-hoc built spreadsheets. Mineral compositions were recalculated to structural formulae for classification purposes. After a careful evaluation of mineral-melt equilibrium, clinopyroxene and amphibole compositions were used for thermobarometric estimates, as these minerals are unaltered and frequent in the rocks. Recalculations and normalisations were applied to whole rock compositions as appropriate, and geochemical parameters were calculated from mineral and whole rock compositions. Large geochemical datasets were constructed with new and previously published whole rock data for discussion on the mantle source. Geochemical modelling was applied to trace element concentrations in order to quantify natural processes and therefore test petrological and mineralogical hypotheses (see Chapter 6). Specifically, the contribution and effect of macrocrysts to the whole rock composition of porphyritic rocks is quantified for the first time. Principal component analysis was applied to trace element datasets in order to statistically test hypotheses and models (see Chapters 6 and 7). This is a method poorly exploited in igneous mineralogical-geochemical studies and provides stimulating and unbiased results. The PAST software (Hammer et al., 2001) was used. Mineral-melt partition coefficients were calculated for clinopyroxene and amphibole in alkaline lamprophyre melts (see Chapter 7). They were tested by c ur v e fitting to the lattice strain model by Blundy and Wood (1994) using the TableCurve 2D version 5.0.1 software and compared to previous datasets in basic rocks. Data reduction for 40 Ar/ 39 Ar geochronology (see Chapter 3) was undertaken with the ArArCalc2.5 software package (Koppers, 2002). Previously published age data were compiled and discussed together with new data in terms of geodynamic evolution. 2.2.4. Presentation of results Petrological, mineralogical, geochemical and geochronological data was interpreted and discussed in terms of magmatic processes, mantle sources and geodynamic conditions in the context of the Iberia-Europe boundary. Some of the results have been presented in numerous national and international conferences with oral and poster presentations, several research papers in peer-reviewed journals and this Thesis dissertation. The most relevant contributions related to the research carried out during this PhD Thesis are the summarised in the following page. Materials & Methods · Chapter 2 19 Ubide et al. (2010). A visual method for separating amphibole (kaersutite) and Mgclinopyroxene. Macla 13, 217–218. Spanish Mineralogical Society Meeting (poster). Ubide et al. (2010). The lamprophyric sub-vertical dyke swarm from Aiguablava (Catalonian Coastal Ranges): petrology and composition. Geogaceta 49, 83–86. Ubide et al. (2011). Magma differentiation and recharge processes: evidence from clinopyroxene compositions (Catalonian Coastal Ranges, northeast Spain). Geophysical Research Abstracts 13, EGU2011-6888. EGU General Assembly (oral comm.). Ubide et al. (2011). Subduction-related signature in late-Variscan lamprophyres from the Catalonian Coastal Ranges (northeast Spain). Abstract #1924. IUGG General Assembly (poster). Ubide et al. (2012). Unravelling the influence of antecryst settling on the composition of a lamprophyre sill: results from geochemical modelling and principal component analysis. Geophysical Research Abstracts 14, EGU2012-116. EGU General Assembly (oral comm.). Ubide et al. (2012). Enriched mantle source for the Cretaceous alkaline lamprophyres from the Catalonian Coastal Ranges (NE Spain). Mineral. Mag. 76 (6), 2475. Goldschmidt Conference (poster). Ubide et al. (2012). The influence of crystal settling on the compositional zoning of a thin lamprophyre sill: A multi-method approach. Lithos 132-133, 37–49. Ubide, T. (2013). New 40 Ar/ 39 Ar ages for the Cretaceous magmatism in northeast Iberia: implications for the rotation of Iberia and the onset of Pyrenean compression. Abstracts book, 61. IPGP Doctoral Students Congress Paris (invited oral comm.). Ubide et al. (2013). Magmatic processes revealed by heterogeneous crystal populations in a lamprophyre system. Mineral. Mag. 77 (5), 2371. Goldschmidt Conference (poster). Ubide et al. (under review a). Age of the Cretaceous alkaline magmatism in northeast Iberia: implications for the Alpine cycle in the Pyrenees. Tectonics. Ubide et al. (under review b). Clinopyroxene and amphibole crystal populations in a lamprophyre sill: a record of magma history and a window to mineral-melt partitioning. Lithos. Ubide et al. (in prep.). The relevance of crystal transfer on magma mixing processes: a case study in the Maladeta Plutonic Complex (Pyrenees, Spain). Contrib. Mineral. Petrol. Chapter 2 · Materials & Methods 20 Age & Geodynamics · Ch a p ter 3 27 silica undersaturated (Enrique, 2009; Ubide et al., 2012a; Chapter 4 in this volume). They share a common mantle source (Ubide et al. 2012b; Chapter 4 in this volume). The available isotopic ages (Solé et al., 2003) correspond to three different lamprophyre intrusions and vary from ca. 80 to 69 Ma. Two of the intrusions were dated by 40 K/ 40 Ar and only one by 40 Ar/ 39 Ar. 3.3. Samples and methods 3.3.1. Location and description of the samples Three outcrops from the Pyrenees and three outcrops from the Catalonian Coastal Ranges were selected for dating with 40 Ar/ 39 Ar (Fig. 3.1; 3.2). A total of nine samples were analysed (Table 3.1; UTM coordinates in the electronic supplement). The Pyrenean samples comprise one sample from the Basque-Cantabrian Basin (SGR-Bt), one sample from the North Pyrenean Basins (AUB-Amp) and two samples from the Central Pyrenees (DEN1-Gr and DEN2-Gr) because it is the least accurately dated domain so far (Galé and Arranz, 2001). No samples were included from the Eastern Pyrenees, as these are the only Cretaceous alkaline rocks in the Pyrenees already dated by the 40 Ar/ 39 Ar method (Golberg et al., 1986; Vitrac-Michard et al., 1977) and most of the studies yield very consistent ages (Golberg et al., 1986; Montigny et al., 1986). In the Catalonian Ranges previous data (Solé et al., 2003) are scarcer than in the Pyrenees. Accordingly, a total of five samples (CALAmp, CAL-Gr, LLAG-Amp, LLAG-Gr and MSPC-Amp) were analysed from three different, geographically scattered lamprophyre intrusions that have never been dated before. Sample SGR-Bt is a biotite separate from a gabbro located in the Basque-Cantabrian Basin near San Groniz (Bilbao). This gabbro (Fig. 3.2A) has a mediumto coarse-grained granular holocrystalline texture composed of clinopyroxene, plagioclase, opaque minerals, biotite, apatite and secondary phases including calcite, mixed phyllosilicates and analcime. Sample AUB-Amp is a separate of large amphibole crystals from a monzonite located in the North Pyrenean Basins near L’Aubisqué, south from Pau. This monzonite can be correlated to the teschenites described previously in this area (e.g., Azambre et al., 1992; Cabanis and Le Fur-Balouet, 1990). It shows a coarse-grained holocrystalline doleritic texture (Fig. 3.2B), defined by up to cm-sized crystals of feldspars, brown amphibole (kaersutite), which sometimes show green rims, clinopyroxene and opaque minerals with associated biotite; the intergranular spaces are occupied by fine-grained crystals of prehnite, chlorite, analcime and accessory epidote, green amphibole, apatite, acicular feldspars, opaque minerals and titanite. Samples DEN1-Gr and DEN2-Gr are groundmass separates from two basaltic dykes located in the Central Pyrenees near Denuy (Huesca). Chapter 3 · Age & Geodynamics 28 These basalts (Fig. 3.2C, 3.2D) are highly vesicular fine-grained hypocrystalline porphyritic rocks, where the vesicles are filled with calcite. They carry clinopyroxene, spinel and apatite xenocrysts and spinel peridotite xenoliths. The groundmass contains plagioclase, biotite, opaque minerals, clinopyroxene, minor amounts of altered glass and accessory apatite. Age & Geodynamics · Ch a p ter 3 29 Table 3.1. Summary of 40 Ar/ 39 Ar data; preferred ages are marked in bold. Full data tables are available in electronic format. Bt: biotite; Amp-M: amphibole macrocryst; Gr: groundmass; GAB: gabbro; MNZ: monzonite; BAS: basalt; LAMP: lamprophyre. Geological setting Sample Separated phase Rock type Plateau age ±2σ MSWD %39Ar, n steps Inverse isochron age ±2σ MSWD K/Ca ±2σ Pyrenees SGR-Bt Bt GAB 104.82 ±0.66 7.64 65.09 101.39 ±1.24 3.21 5.602 ±1.824 ±0.63% 8 ±1.23% AUB-Amp Amp-M MNZ 102.63 ±0.75 2.06 99.78 102.77 ±0.63 2.29 0.082 ±0.005 ±0.73% 8 ±0.61% DEN1-Gr Gr BAS 231.83 ±5.22 2.31 62.59 236.27 ±5.74 2.45 0.067 ±0.030 ±2.25% 5 ±2.43% DEN2-Gr Gr BAS 179.68 ±1.39 1.32 53.23 180.65 ±1.79 1.62 0.170 ±0.018 ±0.78% 4 ±0.99% Catalonian Coastal Ranges CAL-Gr Gr LAMP 80.73 ±1.28 2.43 74.12 79.07 ±1.77 2.40 0.454 ±0.048 ±1.58% 7 ±2.23% CAL-Amp Amp-M 80.37 ±1.39 37.44 88.70 78.98 ±0.50 2.34 0.130 ±0.007 ±1.73% 3 ±0.63% LLAG-Gr Gr LAMP 80.42 ±0.79 2.00 64.30 79.20 ±0.81 1.58 0.134 ±0.017 ±0.98% 11 ±1.02% LLAG-Amp Amp-M 78.06 ±0.66 7.26 93.46 77.39 ±0.86 7.57 0.128 ±0.007 ±0.84% 3 ±1.12% MSPC-Amp Amp-M LAMP 79.53 ±0.94 16.81 99.50 78.21 ±0.45 1.63 0.111 ±0.004 ±1.18% 6 ±0.58% The Catalonian Coastal Ranges lamprophyres are subvolcanic sills located in the province of Girona (Chapter 5). One of the selected sills is located near Calella de Palafrugell (see Chapters 6 and 7 and Ubide et al., 2012a for a full petrological and geochemical characterisation of this intrusion). The second one is located in Llagostera and is reported here for the first time; although there are a few meters of exposure, a less weathered underground section was studied and sampled during an excavation in the neighbouring plot, undertaken for building purposes. The third intrusion is located in the area between Santa Coloma de Farners and Sant Feliu de Buixalleu, near the Monasterio de Sant Pere Cercada (Pallí et al., 1993). The three lamprophyres are classified as camptonites and share similar petrographical features (Fig. 3.2E, 3.2F, 3.2G). They are also very similar to some of the lamprophyres dated by Solé et al. (2003). These rocks are holoor hypocrystalline, porphyritic due to the presence of mmto cm-sized crystals (macrocrysts and minor megacrysts) of brown amphibole (kaersutite), clinopyroxene, opaque minerals and pseudomorphosed olivine. The groundmass is fine-grained, especially at the chilled margins of the intrusions where it gets almost aphyric and rich in vesicles. The groundmass is mainly composed of plagioclase, brown amphibole (kaersutite), opaque minerals, acicular apatite and recrystallised glass (where present). In the lamprophyre Chapter 3 · Age & Geodynamics 30 from Llagostera there are evolved patches of groundmass that show K-rich, Ca-poor feldspars and lower volume fractions of mafic minerals. Large amphibole crystals were separated for dating from the three lamprophyres (samples CAL-Amp, LLAG-Amp and MSPC-Amp). These crystals display homogeneous cores in all the cases, slightly rounded by magmatic corrosion and overgrown by subidiomorphic rims (Fig. 3.2E, 3.2F, 3.2G). The groundmass of the lamprophyres from Calella de Palafrugell and Llagostera is especially fine-grained (Fig. 3.2E, 3.2F) and in these cases groundmass was separated as well for dating experiments (samples CAL-Gr and LLAG-Gr). 3.3.2. 40 Ar/ 39 Ar methodology The selected phases (groundmass, amphibole or biotite) were separated at the mineral separation laboratory of VU University Amsterdam. Groundmass and amphibole were previously analysed in order to assure high potassium contents: K 2 O varies from 0.54 to 2.27 wt. % in groundmass samples –whole rock analysesand from 1.21 to 2.07 wt. % in amphibole crystals –spot analyses- (authors’ unpublished data). Alteration zones in bulk samples were removed by sawing, and the remaining materials were crushed and sieved to select the 400-500 µm fractions, which were washed with demineralised water. The LOC-50 (Liquid Overflow Centrifuge model-50; IJlst, 1973) was used for heavy liquid separations. Heavy liquids were Diiodomethane diluted with 1.2Dichlorobenzene. The Frantz® separator was used for magnetic separations. A Faul vibration table was also used when separating biotite. Subsequently, all samples were purified by handpicking under the microscope. Because amphibole and clinopyroxene have very similar physical and optical properties, a method was developed to exclude clinopyroxene grains from amphibole separates (Ubide et al., 2010a). The method relies on the different colour of these minerals under transmitted light: mineral separates were observed under the microscope using a transparent tray over a white surface, so that the thinnest rims of the grains revealed the colour of the mineral under transmitted light; best results were obtained adding acetone to the studied fraction, as refraction of light inside the liquid helped transmitting the light (Ubide et al., 2010a). Finally, the samples were cleaned in an ultrasonic bath with demineralised water. Groundmass and amphibole samples were additionally leached with HF (5 %, 10 minutes) to remove alteration from groundmass and overgrowths from amphibole crystals, and cleaned again. A summary of the mineral separation methods applied to each sample is presented in Table 3.2. For irradiation, each sample was wrapped in an 8 mm diameter Al-foil package. About 30 mg of sample were used for biotite, ca. 50 mg for amphibole and ca. 60 mg for groundmass. Sample packages and 8-10 mg aliquots of laboratory standard sanidine DRA (25.42 Ma, calibrated following Kuiper et al., 2008) wrapped in Cu-foil were sealed in a 9 mm diameter ID quartz vial, with one standard package loaded at the top and bottom Age & Geodynamics · Ch a p ter 3 31 Table 3.2. Separated fractions for 40 Ar/ 39 Ar analyses and mineral separation methods applied. Sample Rock sample Separated phase Faultabel Heavy liquids (g/cm3) Franz magnet (mA) Sieve (µm) Handpicking (g) Leaching HF Final weight (mg) SGR-Bt SGR Biotite Yes (P) >2.9 200-400 400-500 0.083 No ~ 30 AUB-Amp AUB Amphibole No 3.05-3.3 410-480 400-500 0.119 Yes ~ 50 DEN1-Gr DEN1 Groundmass No 2.9-3.05 400-500 0.195 Yes ~ 60 DEN2-Gr DEN2 Groundmass No 2.93.05 magnetic 400-500 0.167 Yes ~ 60 CAL-Gr CAL-1 Groundmass No 2.76 -2.9 400-500 0.203 Yes ~ 60 CAL-Amp CAL-2 Amphibole No 3.2-3.3 410-500 400-500 0.127 Yes ~ 50 LLAG-Gr LLAG-01 Groundmass No 2.8-2.9 400-500 0.184 Yes ~ 60 LLAG-Amp LLAG-01 Amphibole No 3.2-3.3 400-500 400-500 0.115 Yes ~ 50 MSPC-Amp SFB-6 Amphibole No 3.2-3.3 440-525 400-500 0.162 Yes ~ 50 positions and between every 4 sample packages. The quartz vial was sealed in a standard Al-irradiation-capsule and irradiated for 20 hrs in a Cd-lined rotating facility (RODEO) at the NRG-Petten HFR facility in The Netherlands. 40 Ar/ 39 Ar analyses were carried out at the geochronology laboratory of VU University Amsterdam. The experiments were undertaken on multigrain samples, ca. 9 months after their irradiation, allowing for the decay of 37 Ar in Ca-rich samples (groundmass and amphibole). The samples were loaded in 6 mm wide, 3 mm deep holes of 60 mm diameter copper discs, with 21 holes per disc (half of the irradiated material was loaded per sample). The standards were loaded in 3 mm wide, 6 mm deep holes of a 185 holes copper disc (2-3 grains per hole and 7 replicates per standard). The samples and the standards were preheated to ca. 200 ºC to remove undesirable atmospheric argon. After this preheating step, they were placed in an ultrahigh vacuum extraction line. A CO 2 laser heating system was used for all the experiments; single fusion was applied to the standards and incremental heating to the samples. Positioning of the laser beam was achieved using an analogue Raylease scanhead fitted with a dual mirror system for X-Y adjustment and a ZnS 300 mm focusing lens. The beam delivery system achieved ca. 300 µm at the focal point and for the samples the beam was diffused in the y-direction with a 200 Hz frequency triangular current causing a +1 and -1 mm beam deflection. To ensure even heating of all the grains, the beam was applied in circles increasing in diameter for the standards and in a raster pattern for the samples. The gas was analysed isotopically with a Mass Analyzer Products LTD 215-50 noble gas mass spectrometer. Beam intensities of Ar isotopes were measured in a magnet field controlled peak-jumping mode over the mass range 40–36 on a secondary electron multiplier using switchable amplifier resistors with relative gains of 5x10 2 , 5x10 and 5 with respect to the Faraday collector (10 11 Ω resistor on the Faraday collector amplifier). For data collection, the mass spectrometer is operated with a modified version of standard MAP software. System blanks were measured every 2 (biotite) or 4 (groundmass and amphibole) sample measurements, and Chapter 3 · Age & Geodynamics 32 at the beginning and the end of each step. The total system blanks were in the range of 6x10 -2 – 9x10 -1 volts for mass 40, 5x10 -5 – 6x10 -3 volts for mass 39, 8x10 -6 – 7x10 -4 volts for mass 38, 1x10 -5 – 2x10 -4 volts for mass 37 and 2x10 -4 – 4x10 -3 volts for mass 36. Mass discrimination (ca. 1.002-1.010 per atomic mass unit and 1.0045 for the main series of experiments) was monitored by frequent analysis of 40 Ar/ 36 Ar air pipette aliquots. Data reduction was performed with the originally in-house developed ArArCalc2.5 software package (Koppers, 2002) (http://earthref.org/tools/ararcalc/). Each analysis was corrected for mass discrimination; system blanks were negligible compared to sample signals (often better than a factor of per mil). The irradiation parameter J for each sample was determined by interpolation using a second-order polynomial fitting between the individually measured standards. For the decay constants and the abundance of 40 K the values recommended by the IUGS Subcommission on Geochronology (Steiger and Jäger, 1977) were used. 3.4. Results A summary of 40 Ar/ 39 Ar incremental heating results is presented in Table 3.1. Full data tables can be found in the electronic supplement, where the results are reported following recommendations by Renne et al. (2009). Age spectra, K/Ca spectra and inverse isochron plots are presented in Figs. 3.3 to 3.7, with uncertainties quoted at 2σ. The experiments showed good, consistent results with age calculations that in most cases meet commonly accepted reliability criteria (McDougall and Harrison, 1999). Several age spectra showed elevated ages in the initial steps which may either point to loosely bound excess 40 Ar (some amphibole separates) or, alternatively, to recoil loss of 39 Ar from fine grained alteration phases (some groundmass separates) (Koppers et al., 2000; Wijbrans et al., 2007). The first steps were not considered in the age calculations if they contained abundant atmospheric argon, excess argon, were affected by recoil loss of 39 Ar or presented large uncertainties. The last steps were not considered in the age calculations if they revealed extraneous (excess or inherited) argon. Always three or more concordant, contiguous, high temperature steps were selected, representing more than 50 % of the 39 Ar released. If the 40 Ar/ 36 Ar intercept for the trapped argon derived from the isochron treatment of the data was not significantly different from the atmospheric ratio of 295.5, excess argon was ruled out. If excess argon was detected (biotite and some amphibole separates), the inverse isochron age was preferred over the plateau age (e.g., McDougall and Harrison, 1999). If not (groundmass separates and sample AUB-Amp), the preferred age was Age & Geodynamics · Ch a p ter 3 33 Fig. 3.3. 40 Ar/ 39 Ar incremental heating results for samples SGR-Bt and AUB-Amp, presented in age and K/Ca spectra and inverse isochron plots. In the age and K/Ca spectra, the thickness of the steps reflects the associated uncertainty; the horizontal line represents the plateau. In the inverse isochron diagrams, the dashed line represents the ideal atmospheric-radiogenic mixing line, whereas the solid line represents the inverse isochron calculated with the plateau defining steps (black squares); square size includes uncertainty. The preferred crystallization age is marked in bold and graphically with a thicker line. chosen according to the mean squares of weighted deviates (MSWD) values, which measure the extent to which the scatter of data can be explained by analytical uncertainty. Preferred crystallisation ages (plateau or inverse isochron ages) are marked in bold in Table 3.1 and Figs. 3.3 to 3.7. All preferred ages show MSWD values below the cut-off value of 2.5 (O´Connor et al., 2007) excepting sample SGR-Bt, which nevertheless presents an acceptable inverse isochron age. From the amounts of 39 Ar and 37 Ar released during the experiments some information may be obtained on the chemical composition of the mineral phases contributing to the spectrum (Wijbrans et al., 2007). This effect is shown in the K/Ca plots (Figs. 3.3 to 3.7). The variation in K/Ca is never larger than one order of magnitude, highlighting the purity of the separates. Meanwhile, the radiogenic component of the argon in the selected steps ranges from less than 10 % to almost 100 %. It is generally very high for biotite and amphibole separates and rather low for groundmass separates, although this is not reflected in larger analytical uncertainties in the groundmass experiments. Chapter 3 · Age & Geodynamics 34 Fig. 3.4. 40 Ar/ 39 Ar incremental heating results for samples DEN1-Gr and DEN2-Gr. Explanation is as in Fig. 3.3. Fig. 3.5. 40 Ar/ 39 Ar incremental heating results for samples CAL-Gr and CAL-Amp, which represent different separates from the same lamprophyre. Explanation is as in Fig. 3.3. For the amphibole experiment, the age of the last step is included in the age spectrum. Age & Geodynamics · Ch a p ter 3 35 Fig. 3.6. 40 Ar/ 39 Ar incremental heating results for samples LLAG-Gr and LLAG-Amp, which represent different separates from the same lamprophyre. Explanation is as in Fig. 3.3. For the amphibole experiment, the age of the last step is included in the age spectrum. Fig. 3.7. 40 Ar/ 39 Ar incremental heating results for sample MSPC-Amp. Explanation is as in Fig. 3.3. The age of the last step is included in the age spectrum. Chapter 3 · Age & Geodynamics 36 The detailed interpretation of the age results is in the Appendix (section 3.7). Sample SGR-Bt from the Basque-Cantabrian Basin and AUB-Amp from the North Pyrenean Basins yielded similar, Albian ages at 101.39 ± 1.24 and 102.63 ± 0.75 Ma, respectively (Fig. 3.3; Table 3.1). In the Central Pyrenees, samples DEN1-Gr and DEN2-Gr yielded different ages neither of them Cretaceous (Fig. 3.4; Table 3.1). The age of DEN1-Gr is 231.83 ± 5.22 Ma (Late Triassic) and that of DEN2-Gr is 179.68 ± 1.39 Ma (Early Jurassic). Finally, in the Catalonian Coastal Ranges all the experiments yielded similar, Campanian ages at ca. 79 Ma (Figs. 3.5 to 3.7; Table 3.1). Groundmass and amphibole separates yielded ages indistinguishable from each other within the uncertainty. The preferred crystallisation age for each intrusion was 78.98 ± 0.50 Ma for the lamprophyre from Calella de Palafrugell (CAL-Amp sample), 79.20 ± 0.81 Ma for the lamprophyre from Llagostera (LLAG-Gr sample) and 78.21 ± 0.45 Ma for the lamprophyre from the Monasterio de Sant Pere Cercada (MSPC-Amp sample). 3.5. Discussion 3.5.1. The Triassic-Jurassic rifting in the Central Pyrenees The age of samples DEN1-Gr and DEN2-Gr is Late Triassic (ca. 232 Ma; Carnian) and Early Jurassic (ca. 180 Ma; Toarcian), respectively (Table 3.1). These data are in contrast with preliminary geochronological results (Galé and Arranz, 2001) and dismiss a Cretaceous age for the alkaline dykes in the Central Pyrenees. It follows that the Cretaceous magmatism in the Pyrenees is restricted to the European domain, to the north of the North Pyrenean Fault Zone (Fig. 3.1). Two types of magmatic events are recognised in northeast Iberia during the Triassic and Jurassic. On the one hand, there is widespread tholeiitic magmatism (“ophites”) emplaced during the Triassic-Jurassic boundary in the Western and Central Pyrenees (e.g., Alibert, 1985; Azambre et al., 1987; Béziat et al., 1991; Montigny et al., 1982; Rossi et al., 2003). On the other hand, there is more restricted, geographically scattered, alkaline magmatism forming part of a Late Triassic alkaline igneous province (Lago et al., 1996; Sanz et al., 2013). Within the Pyrenean chain, the latter magmatism has only been recognised in the Eastern Pyrenees (Corbières area; e.g., Azambre and Rossy, 1981) where it has a Late Triassic, Norian age (Montigny et al., 1982). The Triassic-Jurassic geochronological results reveal for the first time, an alkaline magmatism in the Central Pyrenees for that period. Given the affinity of these rocks, one could think they have a common origin within the alkaline igneous province. The differences in age between the samples suggest a long-living (ca. 50 Ma), although geographically restricted, magmatism. If it were related to the alkaline province, the Age & Geodynamics · Ch a p ter 3 43 geochronological and petrological information is provided to contribute to better constrain the opening of the Bay of Biscay and related rotation of Iberia. 3.5.3. The onset of the Alpine compression in the Late Cretaceous Samples from the Catalonian Coastal Ranges yield the youngest ages, constraining the magmatic activity to the Campanian, at ca. 79 Ma (Fig. 3.8; Table 3.1). This datum is very well defined as macrocryst and groundmass separates produce consistent ages in all the studied intrusions (see section 3.4). The previous geochronological study by Solé et al. (2003) provided three ages, including an especially young one which is not supported by the new data and as noted by those authors, must be considered a minimum estimate. The Cretaceous magmatism in the Catalonian Coastal Ranges is the only located in the Iberian domain, to the south of the North Pyrenean Fault Zone, and is significantly younger than the ca. 95-90 Ma age of the igneous rocks from the Eastern Pyrenees (Fig. 3.8). Strikingly, its age at ca. 79 Ma coincides with the age of magnetic anomaly 33o (e.g., Gee and Kent, 2007; Walker and Geissman, 2009; Fig. 3.8). Solé et al. (2003) stated that these magmas were emplaced in zones of crustal relaxation and/or thinning in the relatively stable tectonic transition between the rotation of the Iberian Peninsula and the Alpine collision with respect to Europe. It is however unclear when exactly Alpine compression commenced. The change to a transpressional regime is well recorded in the Pyrenean geology by inversion of previous extensional faults and deposition of growth sequences in the southcentral Pyrenees (Beaumont et al., 2000; Choukroune, 1992; García-Senz, 2002; GarridoMegías and Ríos-Aragüés, 1972; McClay et al., 2004; Muñoz, 1992; Puigdefàbregas and Souquet, 1986). However, these authors report slightly different ages within the Late Cretaceous for the onset of compression: some give a Santonian age whereas others give a Campanian age, and others a Santonian-Maastrichtian or Campanian-Maastrichtian age range. According to plate-kinematic reconstructions based on the fit of sea-floor magnetic anomalies, Rosenbaum et al. (2002) concluded that the relative motion of Iberia and Europe changed to convergence at anomaly A34 (85 Ma, Santonian, according to the timescale by Walker and Geissman, 2009; Fig. 3.8). The recent kinematic studies by Vissers and Meijer (2012a, 2012b) emphasised that there is no geological evidence in the Pyrenees for any significant crustal shortening prior to the Campanian. Vissers and Meijer (2012b) proposed that the onset of collision is coeval with the latest stages of spreading in the Bay of Biscay, which involved first extension then shortening in the Pyrenean domain with a sinistral strike-slip component. Specifically, the youngest oceanic crust of the Bay of Biscay spread in the interval between anomalies A34 and 33o and provoked strike-slip extension in the Pyrenean realm; only from anomaly 33o did strike-slip shortening start in the Pyrenees (Fig. 3.10). Chapter 3 · Age & Geodynamics 44 Fig. 3.10. Kinematics of the Bay of Biscay axis during the final stages of opening and the initial stages of Alpine shortening in northeast Iberia. From Vissers and Meijer (2012b). The new geochronological results consistently point to a magmatic event at 79 Ma (Fig. 3.8). This is inconsistent with Alpine shortening starting before that time. Indeed, the timing of the magmatism coincides exactly with the abovementioned change of tectonic regime in the Pyrenean domain, related to the end of sea-floor spreading in the Bay of Biscay at anomaly 33o (Vissers and Meijer, 2012b). Moreover, detrital thermochronometry of foreland basin sediments gives insights on the early stage of orogenic development (e.g., Filleaudeau et al., 2011) and recent studies in the southcentral Pyrenees identify a first exhumation signal at ca. 78 Ma (Whitchurch et al., 2011) and ca. 80 Ma (Filleaudeau et al., 2011), well in line with the age of the magmatism as well. Hence, the Cretaceous lamprophyres in the Catalonian Coastal Ranges could be markers of the tectonic change. In line with this hypothesis, Scarrow et al. (2011) recently concluded that lamprophyres potentially act as tectonomagmatic markers of changes in geodynamic conditions. These authors studied alkaline lamprophyres of Permian age in the Central Iberian Zone, which also are geographically restricted and small in volume. In that case, the lamprophyres were emplaced during the final stages of the Variscan orogen related to an important shift in the regional tectonic regime from transpressional to transtensional. In the case studied here, the lamprophyres appear to be related to the onset of strike-slip shortening (or to the latest stages of strike-slip extension), as lamprophyre intrusions are commonly associated to strike-slip shear zones (Pirajno, 2010 Age & Geodynamics · Ch a p ter 3 45 and references therein). These data suggest that Alpine shortening in the Pyrenees started in the Campanian at ca. 79 Ma. Since Alpine compression started at the Iberia-Europe plate boundary and was propagated intraplate afterwards, the age of the magmatism in the Catalonian Coastal Ranges could alternatively be subsequent to the onset of compression at the plate margin. However, tectonic inversion in the Catalonian Coastal Ranges is considered Late Cretaceous as well (Juez-Larré and Andriessen, 2002, 2006) and the magmatism is located in the Paleozoic basement at the northernmost tip of the Ranges (see Fig. 3.1), so probably within or very close to the Pyrenean realm during Late Cretaceous times. Moreover, the Pyrenean orogen developed diachronously from east to west (Whitchurch et al., 2011 and references therein) and values for the total shortening across the belt tend to increase towards the east (e.g., Vergés et al., 2002), consistently with a small component of anticlockwise rotation of Iberia during convergence (Vissers and Meijer, 2012b). These data lend further support to considering the lamprophyres from the Catalonian Coastal Ranges coeval with the onset of Alpine shortening in northeasternmost Iberia. 3.6. Conclusions New 40 Ar/ 39 Ar ages on the magmatism of the Pyrenees and Catalonian Coastal Ranges form three age groups: a Late Triassic (ca. 232 Ma) – Early Jurassic (ca. 180 Ma) magmatism in the Central Pyrenees that was previously thought to be Cretaceous, a Cretaceous ca. 102 Ma magmatism in the Basque-Cantabrian Basin and the North Pyrenean Basins and a Late Cretaceous ca. 79 Ma magmatism in the northern part of the Catalonian Coastal Ranges. These results, combined with a detailed evaluation of previous data, provide new insights into the geodynamic evolution of northeast Iberia during the Alpine cycle: • The Late Triassic and Early Jurassic ages represent the first report of an alkaline magmatism in the Central Pyrenees for that time. The new ages widen temporally and geographically the alkaline magmatism associated to the continental rift setting developed in northeast Iberia and in general, southwest Europe at the beginning of the Alpine cycle. • The Cretaceous alkaline magmatism in the Pyrenees took place from the Albian to the Santonian (mostly between 105 and 85 Ma). In contrast with traditional thinking, magmatism is not coeval but post-dates the Aptian opening of the Bay of Biscay and allied rotation of Iberia with respect to Europe. Geochronological, petrological and compositional data for the magmatism are in disagreement with Chapter 3 · Age & Geodynamics 46 both the classical scissor-type opening model and the strike-slip opening model, and suggest that an alternative interpretation should be considered. Magmatism as well as contemporaneous metamorphism, hydrothermalism and mantle exhumation developed in a tectonically unclear stage between the rotation of Iberia and the onset of Alpine collision. Magmatic activity is especially recorded between ca. 100 and 90 Ma. It is restricted to ca. 95-90 Ma in the Eastern Pyrenees, whereas multiple stages develop towards the west with the longest time span recorded in the Basque-Cantabrian Basin. The new ages at ca. 102 Ma agree with those of amphibole in veins that cross-cut mantle exposures, supporting the interpretation that the veins represent melt conduits for the Cretaceous alkaline magmas. • The Late Cretaceous alkaline magmatic event in the Catalonian Coastal Ranges is precisely dated at ca. 79 Ma (Campanian). This age is inconsistent with shortening starting before that time as previously considered, and could actually represent the onset of the Alpine shortening in northeast Iberia. It is noteworthy that the alkaline magmatism in the Central Pyrenees has turned out not to be Cretaceous, so it will not be considered for interpretations on the Cretaceous magmatic province in northeast Iberia (see Chapter 4). Age & Geodynamics · Ch a p ter 3 47 3.7. Appendix In this section a detailed description of 40 Ar/ 39 Ar experiment results is presented. This includes the interpretation of the age spectra, K/Ca spectra and inverse isochrons as well as the criteria for the selection of crystallisation ages. Sample SGR-Bt yielded a disturbed age spectrum with abundant atmospheric argon in the first steps and excess argon in the highest temperature steps (Fig. 3.3). The isochron treatment of the data confirmed the presence of excess argon, as the non-radiogenic intercepts were lower than atmosphere. Hence, the inverse isochron age of 101.39 ± 1.24 Ma was considered the most accurate estimate of the crystallisation age. The MSWD value was relatively high, as the result of low individual analytical step uncertainties. The K/Ca spectrum indicates a relative compositional heterogeneity, either due to compositional variations among the biotite grains or, more probably, due to contamination of the grains with other phases (e.g., other Ca-rich phases; Fig. 3.2A). The age defining fractions are the most K-rich, therefore indicating the highest biotite proportions, which retained most of the radiogenic argon (> 85 %). Sample AUB-Amp produced linear age and K/Ca spectra where almost 100 % of 39 Ar was considered for the plateau age calculation (Fig. 3.3). No excess argon was found as the non-radiogenic intercepts were indistinguishable from atmosphere and the plateau and inverse isochron ages are concordant. The plateau age of 102.63 ± 0.75 Ma is preferred due to its lower MSWD value. Most of the gas was released in the last step, indicating the high retentivity for argon of amphibole. The K/Ca flat spectrum pointed out the purity of the sample and confirmed the compositional homogeneity of the amphibole grains, in agreement with petrographical observations (Fig. 3.2B). Samples DEN1-Gr and DEN2-Gr yielded different ages neither of them Cretaceous (Fig. 3.4). The lowest temperature steps were somewhat irregular and showed very low K/Ca values, probably related to contamination of some groundmass grains by calcite-filled vesicles (Fig. 3.2C), which are hard to leach if they are surrounded by groundmass. The last steps were apparently affected by contamination with a Ca-rich phase (probably xenocrysts or xenoliths present in the rock, see Fig. 3.2D) which started degassing at high temperatures, as reported elsewhere for groundmass samples (e.g. De Beni et al., 2005; Koppers et al., 2000; Veldkamp et al., 2007). The fractions included in the age calculation are the most K-rich, probably related to the biotite microcrysts of the groundmass (Fig. 3.2C, 2D). No excess argon was detected and the plateau and inverse isochron ages overlapped within the uncertainty. Plateau ages were preferred due to their lower MSWD values. Therefore, DEN1-Gr and DEN2-Gr were inferred to be 231.83 ± 5.22 Ma and 179.68 ± 1.39 Ma old, respectively. Chapter 3 · Age & Geodynamics 48 Samples CAL-Gr and LLAG-Gr yielded L-shaped age spectra with very high ages for the first step (Fig. 3.5, 3.6). Given that no excess argon was found in the inverse isochron treatment of the data, these L-shaped age spectra must be explained by recoil loss of 39 Ar from fine grained phases (Koppers et al. 2000). These are probably secondary minerals (mainly chlorite) present as late magmatic phases and, to a lesser extent, as alteration products (recrystallisation) of the glass of the rock. The total fusion ages were slightly higher than the plateau ages, also supporting recoil loss of 39 Ar (Koppers et al., 2000). Besides, sample LLAG-Gr showed a K-rich last step. This cannot be related to contamination by large crystals because they are mafic minerals (Fig. 3.2F) which present low K/Ca ratios. This step might therefore be related to evolved, K-rich patches in the groundmass (see section 3.3.1) which started degassing at high temperatures. Inverse isochron ages show lower MSWD values than plateau ages. Samples CAL-Amp, LLAG-Amp and MSPC-Amp (Fig. 3.5, 3.6, 3.7) produced apparently flat age spectra. In detail however, the age spectra have decreasing staircase morphology and high MSWD values and reveal excess argon in the low temperature steps. The inverse isochron treatment of the data confirmed excess argon in the samples. The excess argon was probably diffused into the crystals from the magma chamber, as reported elsewhere (e.g., Harrison and McDougall, 1980; Wijbrans and McDougall, 1988). This agrees with the compositional homogeneity of the amphibole grains, as recognised petrographically (Fig. 3.2E, 2F, 2G) and confirmed by the K/Ca spectra. If argon diffusion into crystals is recognised, a plateau age cannot be defined; however, the purely radiogenic component from the inverse isochron corresponds to a geologically meaningful age (e.g., McDougall and Harrison, 1999). Sample LLAG-Amp however, did not produce a well defined age in any case. The steps selected for each inverse isochron represented high percentages of the gas released and high percentages of radiogenic argon in each step. Besides, the inverse isochron ages were similar to the age of the last step, presumably the least affected by argon diffusion because it most likely represents the cores of the crystals; the last step contained more than 60 % of the gas released and it was enriched in radiogenic argon over 90 %. Groundmass and amphibole separates analysed in the lamprophyres from Calella de Palafrugell and Llagostera yielded ages indistinguishable from each other within the uncertainty. The preferred crystallisation age for each intrusion was that with the lowest associated MSWD value, namely: 78.98 ± 0.50 Ma for the lamprophyre from Calella de Palafrugell, which corresponds to the inverse isochron age from sample CAL-Amp, and 79.20 ± 0.81 Ma for the lamprophyre from Llagostera, which corresponds to the inverse isochron age from sample LLAG-Gr. The amphibole separate from the Monasterio de Sant Age & Geodynamics · Ch a p ter 3 49 Pere Cercada lamprophyre yielded a very accurate inverse isochron age with almost 100 % of 39 Ar considered: 78.21 ± 0.45 Ma. Chapter 3 · Age & Geodynamics 50 51 4 Petrogenesis of the Cretaceous alkaline magmatism in northeast Iberia: new insights into the evolution of the mantle source SUMMARY The Cretaceous magmatism in northeast Iberia (Pyrenees and Catalonian Coastal Ranges) constitutes an alkaline igneous province with magma differentiation from ultrabasic to intermediate-acid compositions. Increasing magmatic activity and melting rates towards the west can be related to the context of opening of the Bay of Biscay that controlled the development of this magmatism. Major element, trace element and Sr-Nd-Pb isotope geochemistry reveals a common, enriched/metasomatised, heterogeneous (HIMU-EM) and asthenospheric mantle source. Comparison with a large geochemical database of magmatisms occurring in Iberia throughout the Alpine cycle indicates that the Cretaceous signature might have been present in large areas of the mantle from Permian to Quaternary times and was tapped according to favourable tectonic conditions. Landscape near Elgoibar, Basque Cantabrian Basin Chapter 4 · Petrogenesis 52 Petrogenesis · Chapter 4 59 Table 4.1. (cont.) Domain CCR Domains: BCB - Basque-Cantabrian Basin; NPB - North Pyrenean Basins; EP - Eastern Pyrenees; CCR - Catalonian Coastal Ranges. Rock types: Pi - Picrite; B - Basaltic rock; G - Gabbroic rock; Tr - Trachyte; L - Lamprophyre; Sye – Syenite. Classification criteria according to Le Maitre (2002). PIC – picrite; BAS – basalt; TRB – trachybasalt; BTR – basaltic trachyandesite; TR – trachyte; GAB – gabbro; MNZ – monzonite; SYE – syenite; MCH – monchiquite; SAN – sannaite; CAMP – camptonite. Fe2O3T: total iron expressed as Fe 2O3. LOI: Loss On Ignition. Mg# represents magnesium number: 100 MgO / (MgO + FeO) and FeO = 0.9 Fe2O3T. < DL: below the detection limit. (a) UXT: replicate analysis of URR. (b) CAL-Gr: bulk trace element composition of the groundmass in the Calella de Palafrugell lamprophyre, modelled in Ubide et al. (2012a) –see Table 6.3-. Sector West West West West West West UTM coor. 31T 31T 31T 31T 31T 31T 465860 465860 466218 466218 470170 467387 4629223 4629223 4629029 4629029 4630359 4631372 Rock type L Classification CAMP CAMP CAMP CAMP CAMP CAMP Sample SFB-1(r) SFB-2(r) SFB-3 SFB-4 SFB-5 SFB-6 SiO2 43.29 41.73 43.38 42.39 42.60 44.08 TiO2 2.80 2.93 1.91 2.72 3.09 2.93 Al2O3 11.36 11.56 8.24 10.64 14.09 14.11 Fe2O3T 11.26 11.24 9.78 11.66 12.09 11.78 MnO 0.18 0.13 0.15 0.14 0.17 0.19 MgO 10.03 9.92 14.03 11.21 4.39 6.54 CaO 11.74 12.37 13.15 11.87 10.38 8.64 Na2O 2.51 1.99 1.78 2.29 2.87 3.91 K2O 1.40 1.42 0.76 1.37 2.12 2.18 P2O5 0.42 0.45 0.26 0.38 0.64 0.62 LOI 4.26 5.98 5.62 4.25 7.33 4.80 Total 99.24 99.72 99.07 98.92 99.77 99.79 Mg# 50 50 61 52 29 38 Rb 35.24 34.21 23.17 34.83 58.66 59.52 Cs 19.15 33.60 11.15 15.17 2.71 12.94 Be 1.16 1.11 < DL 1.06 1.68 1.52 Sr 492.50 357.80 293.10 444.70 692.70 507.60 Ba 435.70 604.60 627.90 395.60 644.30 586.70 V 314.20 330.70 239.30 327.00 301.50 286.50 Cr 629.20 606.80 1304.00 744.80 140.30 220.70 Co 51.31 56.18 56.34 55.89 35.38 34.03 Ni 167.00 167.80 294.30 213.50 55.03 85.08 Cu 92.59 100.60 160.80 99.56 24.53 51.56 Zn 99.95 139.10 77.34 101.80 102.70 155.60 Ga 18.45 18.86 13.47 17.94 23.16 22.03 Y 19.76 21.27 14.71 19.97 28.54 26.80 Nb 40.86 42.84 27.85 36.20 59.24 56.65 Ta 3.15 3.31 2.18 2.83 4.56 4.45 Zr 208.50 216.70 144.90 194.50 306.00 294.40 Hf 5.11 5.32 3.58 4.96 7.35 7.13 Mo 1.45 1.57 1.18 8.82 8.17 9.78 Sn 2.49 2.42 1.75 2.35 2.85 2.81 Pb 3.64 3.84 3.14 3.95 8.73 4.58 U 1.12 1.34 0.74 0.95 1.52 1.52 Th 3.82 3.82 2.58 3.32 5.49 5.43 La 35.49 37.66 24.55 33.18 55.05 53.44 Ce 76.45 82.08 52.78 72.24 116.40 113.60 Pr 9.33 10.12 6.49 8.92 14.04 13.59 Nd 36.85 40.13 26.05 36.00 55.48 52.77 Sm 7.08 7.67 5.08 7.04 10.16 9.67 Eu 2.16 2.36 1.57 2.18 3.06 2.87 Gd 5.74 6.26 4.20 5.73 8.09 7.63 Tb 0.79 0.86 0.59 0.80 1.12 1.05 Dy 4.18 4.48 3.11 4.16 5.93 5.60 Ho 0.72 0.76 0.53 0.73 1.04 0.98 Er 1.82 1.97 1.38 1.88 2.71 2.56 Tm 0.25 0.26 0.19 0.25 0.37 0.34 Yb 1.56 1.62 1.14 1.53 2.30 2.18 Lu 0.22 0.24 0.17 0.23 0.34 0.32 Chapter 4 · Petrogenesis 60 (SARM) in Nancy (France). The samples were analysed by ICP-AES for major elements and ICP-MS for trace elements. Details of the analytical procedures and detection limits are available at http://www.crpg.cnrs-nancy.fr/SARM/. The analyses cover 10 samples from the Basque-Cantabrian Basin, 2 samples from the North Pyrenean Basins, 9 samples from the Eastern Pyrenees and 16 samples from the Catalonian Coastal Ranges. A replicate of one of the samples from the Basque-Cantabrian Basin (URR) was sent for analysis with a different name (UXT) to test the reproducibility of the methods; the results were very similar (Table 4.1). In addition, the bulk composition of the groundmass of one of the lamprophyres in the Catalonian Coastal Ranges (CAL-Gr) is available from Ubide et al. (2012a). From among the sample powders analysed for elemental concentrations, 13 were selected for Sr, Nd and Pb isotope analyses (results in Table 4.2). These included 4 samples from the Basque-Cantabrian Basin, 3 samples from the Eastern Pyrenees and 6 samples from the Catalonian Coastal Ranges. Measurements were carried out at the SGIker-Geochronology and Isotopic Geochemistry facility of the Universidad del País Vasco UPV/EHU (Spain). Regarding Sr and Nd isotope analyses, chemical procedures for sample preparation are described in Pin and Bassin (1992), Pin et al. (1994) and Pin and Santos Zalduegui (1997). Sm and Nd concentrations were determined by isotope-dilution TIMS using a mixed 149 Sm150 Nd tracer. The 147 Sm/ 144 Nd values are precise to ± 0.2 % at the 95 % confidence level. 143 Nd/ 144 Nd ratios were measured by TIMS in a Finnigan Mat-262 instrument in static multicollection mode, and corrected for mass fractionation by normalization to 146 Nd/ 144 Nd = 0.7219 (Wasserburg et al., 1981; Thirlwall, 1991). The La Jolla isotopic standard and its uncertainty (2 SD) were measured under the same conditions and gave 143 Nd/ 144 Nd = 0.512101 ± 0.000003 (n = 2). In the calculations of εNdT, 143 Nd/ 144 Nd CHUR =0.512638 and 147 Sm/ 144 Nd CHUR =0.1967 (Jacobsen and Wasserburg, 1984) were used. 87 Sr/ 86 Sr ratios were measured by MC-ICP-MS using a high-resolution Thermo Fisher Scientific Neptune instrument in static multicollection mode, and corrected for mass fractionation by normalization to 88 Sr/ 86 Sr = 8.375209 (Steiger and Jäger, 1977). The reported 87 Sr/ 86 Sr ratios were adjusted to the NBS 987 standard, measured under the same conditions with a value and uncertainty (2 SD) of 87 Sr/ 86 Sr = 0.710270 ± 0.000011 (n = 3). The 87 Rb/ 86 Sr values were calculated considering the 87 Sr/ 86 Sr values and the elemental concentration in Rb and Sr (e.g., Faure, 1991). Pb isotope analyses were performed using a Neptune MC-ICP-MS instrument. Chemical procedures for Pb extraction followed those described by Manhès et al. (1987). Pb isotopes were corrected for instrumental mass bias by internal second element normalisation using known amounts of Tl and the exponential mass fractionation law, a 205 Tl/ 203 Tl ratio of 2.3889 was used. The average ratios for the NBS981 Pb standard and their uncertainties (2 SD) measured under the same conditions are 206 Pb/ 204 Pb = 16.9439 ± 0.0038, 207 Pb/ 204 Pb = 15.5011 ± 0.0030, 208 Pb/ 204 Pb =36.7288 ± 0.0071, 208 Pb/ 206 Pb = 2.16767 ± 0.00017 and 207 Pb/ 206 Pb = 0.91484 ± 0.00007 (n = 67). Petrogenesis · Chapter 4 61 Table 4.2. Sr-Nd-Pb isotope composition of selected samples. Uncertainties are quoted at 2σ. Domain BCB BCB BCB BCB EP EP EP Sample ELG-08 FR-9 8057 MALZA-1 CORB-2 VIN-2 FIT-2 87Rb/86Sr 0.09421 0.21763 0.14133 0.32276 0.09539 0.10886 0.64184 87Sr/86Sr 0.70631 0.70575 0.70692 0.70476 0.70520 0.70502 0.70484 ± 0.00001 ± 0.00001 ± 0.00001 ± 0.00001 ± 0.00001 ± 0.00001 ± 0.00001 87Sr/86Sr t 0.70618 0.70544 0.70672 0.70431 0.70508 0.70488 0.70401 εSr 25.50 15.06 33.22 -1.10 9.75 6.91 -5.39 147Sm/144Nd 0.1143 0.1103 0.1150 0.0857 0.1049 0.0929 0.0709 143Nd/144Nd 0.512766 0.512761 0.512776 0.512782 0.512750 0.512726 0.512698 ± 0.000003 ± 0.000003 ± 0.000003 ± 0.000003 ± 0.000003 ± 0.000003 ± 0.000003 143Nd/144Nd t 0.512691 0.512689 0.512701 0.512726 0.512688 0.512671 0.512656 εNd 3.55 3.50 3.74 4.23 3.24 2.91 2.62 TNdDM (Ga) 0.59 0.57 0.58 0.56 0.54 206Pb/204Pb 20.257 20.659 19.540 20.050 20.627 20.974 21.152 ± 0.004 ± 0.007 ± 0.004 ± 0.004 ± 0.004 ± 0.004 ± 0.004 207Pb/204Pb 15.664 15.668 15.700 15.632 15.680 15.698 15.704 ± 0.003 ± 0.005 ± 0.003 ± 0.003 ± 0.003 ± 0.003 ± 0.003 208Pb/204Pb 40.229 40.709 39.624 40.314 40.428 40.818 41.037 ± 0.008 ± 0.013 ± 0.007 ± 0.007 ± 0.007 ± 0.007 ± 0.007 208Pb/206Pb 1.986 1.971 2.028 2.011 1.960 1.946 1.940 207Pb/206Pb 0.773 0.758 0.803 0.780 0.760 0.748 0.742 Table 4.2. (cont.) Domain CCR CCR CCR CCR CCR CCR CCR Sample AIG-05 AIG-05 a CAL-2 SAG-03B LLAG-01 SFB-3 SFB-6 87Rb/86Sr 0.17088 0.15627 0.42897 0.10907 0.22863 0.33914 87Sr/86Sr 0.70387 0.70385 0.70376 0.70594 0.70402 0.70552 0.70582 ± 0.00001 ± 0.00002 ± 0.00001 ± 0.00001 ± 0.00001 ± 0.00001 ± 0.00001 87Sr/86Sr t 0.70368 0.70358 0.70546 0.70390 0.70526 0.70543 εSr -10.30 -11.73 14.91 -7.19 12.11 14.55 147Sm/144Nd 0.1112 0.1160 0.1097 0.1282 0.1149 0.1086 143Nd/144Nd 0.512775 0.512792 0.512787 0.512777 0.512827 0.512764 0.512751 ± 0.000003 ± 0.000007 ± 0.000003 ± 0.000003 ± 0.000003 ± 0.000003 ± 0.000003 143Nd/144Nd t 0.512717 0.512726 0.512720 0.512760 0.512704 0.512694 εNd 3.55 3.73 3.60 4.39 3.29 3.10 TNdDM (Ga) 0.56 0.57 0.55 0.58 0.60 0.58 206Pb/204Pb 20.918 21.002 19.364 19.062 19.300 19.239 19.457 ± 0.004 ± 0.015 ± 0.004 ± 0.006 ± 0.005 ± 0.005 ± 0.004 207Pb/204Pb 15.779 15.766 15.696 15.677 15.651 15.651 15.662 ± 0.003 ± 0.020 ± 0.003 ± 0.005 ± 0.004 ± 0.004 ± 0.003 208Pb/204Pb 40.518 40.563 39.794 39.280 39.290 39.288 39.816 ± 0.007 ± 0.060 ± 0.007 ± 0.013 ± 0.009 ± 0.010 ± 0.007 208Pb/206Pb 1.937 2.055 2.061 2.036 2.042 2.046 207Pb/206Pb 0.754 0.811 0.822 0.811 0.813 0.805 Domains: BCB - Basque-Cantabrian Basin; EP - Eastern Pyrenees; CCR - Catalonian Coastal Ranges. (t) Age of the rocks: ca. 100 Ma in the Basque-Cantabrian Basin and North Pyrenenan Basins, ca. 90 Ma in the Eastern Pyrenees and ca. 80 Ma in the Catalonian Coastal Ranges (see Chapter 3). TNdDM: T-depleted mantle model ages for mafic rocks. (a) A second aliquot of sample AIG-05 was analysed in a different laboratory. A second aliquot of sample AIG-05 from the Catalonian-Coastal Ranges was prepared and analysed by P. Larrea and D. Kuentz at the Isotope Geochemistry & Mass Spectrometry Laboratory of the Miami University (Ohio, USA). Sr, Nd and Pb isotope ratios were determined by TIMS following the methods described in Larrea et al. (submitted). The results (Table 4.2) were very similar to those obtained at the SGIker-Geochronology and Chapter 4 · Petrogenesis 62 Isotopic Geochemistry facility of the University of the Basque Country UPV/EHU (Spain), and most of them agree within the uncertainty. In order to keep data in the figures comparable, only data from the IBERCRON laboratory are plotted. For comparison, a large geochemical database was constructed including elemental and isotopic compositions of magmatisms of Late Variscan to Quaternary age recorded in Iberia. A significant amount of data corresponds to previous studies carried out by the research group, and there are also numerous data from other authors. 4.4. Petrology The samples selected for geochemical analyses cover diverse petrological typologies. The classification of all the samples is presented in Table 4.1 and detailed mineral assemblages and crystallisation sequences are available in the Electronic Supplement. Volcanic and hypovolcanic rocks occur in the Basque-Cantabrian Basin and the North Pyrenean Basins. There are basaltic-andesitic rocks, locally picritic, as well as trachytes. Special care was exercised to obtain fresh samples, as the studied volcanic outcrops show widespread alteration. Basaltic-andesitic samples (ELG-08, FR-1, FR-9, MALZA-B and 8057 from the BasqueCantabrian Basin, and COG-3 from the North Pyrenean Basins) have porphyritic or intergranular textures (Fig. 4.2A), mainly composed of feldspars (30 – 75 vol. %) and clinopyroxene (Ti-rich diopside), with volume fractions that vary from very low or 0 % up to 60 %. The relative proportions of these two minerals vary with the more or less mafic character of the lavas. The mineral assemblage is completed with Fe-Ti oxides (2 – 5 vol. %), olivine pseudomorphs (0 – 6 vol. %), amphibole (mainly kaersutite; 0 – 5 vol. %), and minor amounts of biotite, titanite, apatite, epidote, calcite and chlorite. Some of the samples include chlorite and other phyllosilicates occupying intercrystalline spaces that may be interpreted as destabilised glass. In these cases, the original texture would be intersertal and thus the rock hypocrystalline. Sample ELG-08 was extracted from a large sill that becomes locally picritic due to the accumulation of olivine crystals. The picrite, sampled as ELG-04, has a volume fraction of olivine of 70 – 80 %. This rock displays a cumulate (heteroadcumulate) texture where olivine crystals are the main cumulus phase (Fig. 4.2B). Most olivine crystals are transformed into an assemblage of serpentine group minerals and opaque grains; they frequently have a rim of bluish chlorite. Olivine relicts display homogeneous Petrogenesis · Chapter 4 63 Fig. 4.2. Photomicrographs of representative samples. Cross-polarised transmitted light. A: Basalt ELG-08 from the Basque-Cantabrian Basin. B: Picrite ELG-04 from the Basque-Cantabrian Basin. C: Trachyte MALZA1 from the Basque-Cantabrian Basin. D: Gabbro URR from the Basque-Cantabrian Basin. E: Syenite FIT-2 from the Eastern Pyenees. F: Monchiquite CORB-2 from the Eastern Pyrenees. G: Sannaite VIN-2 from the Eastern Pyrenees. H: Camptonite from the Catalonian Coastal Ranges. Mineral abbreviations are after Whitney and Evans (2010). Chapter 4 · Petrogenesis 64 compositions, with forsterite contents of Fo 81-83 . Opaque minerals (Cr-rich spinel and Fe-Ti oxides) are also present as cumulus phases (2 vol. %). Clinopyroxene (Ti-rich diopside; 10 – 15 vol. %) represents the principal intercumulus phase, developing poikilitic crystals (Fig. 4.2B). The remaining intercumulus spaces are filled with amphibole, biotite and Fe-Ti oxides that crystallised from the residual melt. The mineral assemblage is completed with minor amounts of feldspars, epidote, apatite and green chlorite. The trachyte sample (MALZA-1 from the Basque-Cantabrian Basin) is composed almost entirely of oriented feldspar crystals (97 vol. %) with a bimodal size distribution, producing a porphyritic and trachytic texture (Fig. 4.2C). There are also opaque minerals (2 vol. %), calcite (1 vol. %) and accessory apatite. Intrusive rocks on the other hand, are present in all the sectors. They become progressively more abundant towards the east and represent the only Cretaceous magmatics in the Eastern Pyrenees and the Catalonian Coastal Ranges. There are gabbromonzonite-teschenite rocks as well as syenites and lamprophyre intrusions. Mafic intrusive sam pl es include gabbros SGR, URR/UXT and URR-3 from the BasqueCantabrian Basin and monzonite AUB from the North Pyrenean Basins, which can be correlated to the teschenites described in previous studies (Azambre et al., 1992 and Cabanis and Le Fur-Balouet, 1990). These rocks have granular to intergranular textures (Fig. 4.2D; see also Fig. 3.2A, 3.2B) composed of feldspars (35 – 60 vol. %), clinopyroxene (Ti-rich diopside; 6 – 40 vol. %), amphibole (mainly kaersutite; up to 20 vol. %), opaque minerals (Fe-Ti oxides and rarely Cr-rich spinel; 3 – 6 vol. %) and minor abundances of olivine pseudomorphs, biotite, titanite, analcime, prehnite, apatite, epidote, calcite and chlorite. Accessory zircon is recognised in sample URR/UXT. Mafic minerals represent an important fraction (30 – 40 vol. %) of the mineral assemblage, so that samples with low proportions of clinopyroxene are especially rich in amphibole (e.g., sample AUB; see the Electronic Supplement). Syenite samples (FIT-1, FIT-2 and FIT-3 from the Eastern Pyrenees) have an intergranular texture dominated by feldspars (ca. 70 vol. %) (Fig. 4.2E). Analcime is relatively frequent (10 vol. %) and followed in abundance by amphibole (hastingsite; 8 vol. %), clinopyroxene (aegirine-augite and minor aegirine; 4 vol. %), nepheline (3 vol. %), biotite (2 vol. %), titanite (2 vol. %) and lavenite (1 vol. %). Opaque minerals and zircon are accessory phases. Lamprophyres crop out in the Eastern Pyrenees and the Catalonian Coastal Ranges. The Eastern Pyrenean lamprophyres are classified as monchiquites (samples CORB-1, CORB-2, CORB-3 and CORB-6) and sannaites (samples VIN-1 and VIN-2). The monchiquites show a porphyritic texture and a mineral assemblage dominated by clinopyroxene (Ti-rich Petrogenesis · Chapter 4 65 diopside; 50 – 55 vol. %) and phlogopite (25 – 30 vol. %) (Fig. 4.2F). Opaque minerals (FeTi oxides and minor Cr-rich spinel) represent ca. 6 vol. % of the rock, whereas olivine pseudomorphs vary between 2 and 5 vol. %. There are also minor amounts of feldspars, nepheline, analcime, apatite, calcite and glass transformed into chlorite and other secondary products. These rocks enclose ultramafic xenoliths linked to ultramafic bodies cropping out in the Pyrenees (Azambre and Fabriès, 1989). The sannaites on the other hand, are microporphyritic and mainly composed of amphibole (25 – 40 vol. % kaersutite and minor hastingsite) and clinopyroxene (Ti-rich diopside; 20 – 25 vol. %) (Fig. 4.2G). FeTi oxides (2 – 6 vol. %), feldspars (6 – 10 vol. %), glass transformed into cryptocrystalline assemblages (10 – 25 vol. %) and minor proportions of analcime, titanite, apatite and calcite complete the mineral assemblage. In the Catalonian Coastal Ranges, lamprophyre intrusions are the only representatives of the Cretaceous magmatism. These samples (AIG-04, AIG-05, AIG-08, CAL-1, CAL-2, CAL-3 , CAL-4, CAL-5, CAL-Gr, SAG-03B, LLAG-01, SFB-1(r), SFB-2(r), SFB-3, SFB-4, SFB-5 and SFB-6) are camptonites with a strongly porphyritic texture in most cases (see Chapter 5) (Fig. 4.2H; see also Fig. 3.2E, 3.2F, 3.2G, 5.8, 5.9). They consist of clinopyroxene (mainly Ti-rich diopside), which is a minor component in certain intrusions but reaches up to 50 vol. % in others, amphibole (mainly kaersutite; 25 – 55 vol. %), Fe-Ti oxides (6 – 14 vol. %), olivine pseudomorphs (up to 8 vol. %; relicts have forsterite contents of Fo 88-89 ), feldspars (6 – 60 vol. %) and minor amounts of biotite, apatite, analcime, calcite and chlorite. The amount of glass, which is mostly recrystallised to chlorite and other secondary products (frequently cryptocrystalline), reaches ca. 30 vol. % of the rock at the chilled margins of certain intrusions (see detailed mineral assemblages and crystallisation sequences in the Electronic Supplement). 4.5. Geochemistry 4.5.1. Classification and affinity The classification of the samples (Table 4.1) follows criteria by Le Maitre (2002). Whole rock major element and trace element compositions are presented in Table 4.1. Loss On Ignition (LOI) values are relatively high (1.40 – 8.58 wt. %). This is due to the alteration of certain samples (e.g., the picrite sample ELG-04), but more frequently due to the abundance of hydrous minerals like amphibole or biotite, or due to the presence of calcite and hydrous minerals filling vesicles, which are widespread in lamprophyres (Ubide et al., 2012a –see Chapter 6-). Irrespective of their volcanic, subvolcanic or intrusive origin, all the samples have been plotted in the Total Akalis vs. Silica diagram (TAS diagram; Le Bas et al., 1986) to evaluate Chapter 4 · Petrogenesis 66 their distribution (Fig. 4.3). Based on their SiO 2 concentrations, most samples are basic to ultrabasic. The syenite samples from the Eastern Pyrenees (FIT-1, FIT-2 and FIT-3) are intermediate whereas the trachyte sample from the Basque-Cantabrian Basin (MALZA-1) is intermediate-acid. The high alkalis concentration in all the rocks indicates their alkaline geochemical affinity. The distribution of the samples agrees with previous studies and defines two evolutionary trends (Fig. 4.3). Rocks from the Basque-Cantabrian Basin and the North Pyrenean Basins Fig. 4.3. Total Alkalis vs. Silica – TAS diagram (Le Bas et al., 1986; analyses recalculated to 100% on an anhydrous basis). Filled symbols represent data obtained in this study and empty symbols represent data from the literature included for comparison: data from the Basque-Cantabrian Basin and North Pyrenean basins are from Azambre et al. (1992) and Cabanis and Le Fur-Balouet (1990), data from the Eastern Pyrenees are from Azambre et al. (1992) and data from the Catalonian Coastal Ranges are from Enrique (2009). Mafic intrusives include gabbros, teschenites, syenoteschenites, dolerites and monzonites; mafic extrusives include basalts, andesites, trachybasalts and basaltic trachyandesites; picritic rocks include picrites and cortlandites; lamprophyres include monchiquites, sannaites and camptonites. Fig. 4.4. SiO 2 vs. Nb/Y diagram (Winchester and Floyd, 1977). Data from the literature lack many trace elements including Nb and Y, so they cannot be included for comparison. Petrogenesis · Chapter 4 67 define a SiO 2 -mildly-saturated or -undersaturated trend (with normative olivine and hypersthene or normative olivine and nepheline or leucite) with a saturated rock (trachyte MALZA-1) in the silica-rich compositions. On the contrary, rocks from the Eastern Pyrenees and the Catalonian Coastal Ranges define a steeper, SiO 2 - undersaturated trend. Given the high LOI values of many samples, the classification scheme proposed by Winchester and Floyd (1977), which introduces immobile elements, has been also applied. According to the SiO 2 vs. Nb/Y diagram (Fig. 4.4), mafic samples are classified in the basanite/nephelinite field or the alkali basalt field. Syenite samples plot outside the phonolite field whereas the trachyte sample plots in the trachyte field. In the V vs. Ti diagram (Fig. 4.5), most samples plot within the OIBtype field defined by Shervais (1982). This agrees with the alkaline geochemical affinity of the rocks, as also supported by normative nepheline and no normative quartz in most compositions, high concentrations in Nb and Ta in all of the samples (Table 4.1), and mineral compositions, particularly Ti-rich mafic minerals (see section 4.4). Inferences on the geodynamic context must not be derived. 4.5.2. Major and trace element variations The MgO concentrations of most samples vary between 4 and 14 wt. % (Table 4.1). The picrite sample from the Basque-Cantabrian Basin has an extremely high MgO concentration of ca. 25 wt. %, whereas the trachyte sample from the Basque-Cantabrian Basin and the syenite samples from the Eastern Pyrenees have MgO concentrations below 1 wt. %. Following these differences, most samples plot relatively grouped in bivariate diagrams vs. decreasing MgO, while the picrite sample plots to the left and the trachyte and syenite samples plot the right of the diagrams (Fig. 4.6, 4.7). Compositional variations define evolutionary trends where SiO 2 , Al 2 O 3 , alkalis (Na 2 O, K 2 O) and incompatible trace elements like the Large Ion Lithophile Elements (LILE) and the High Field Strength Elements (HFSE) increase with decreasing MgO, whereas CaO and compatible elements like the transition elements decrease with decreasing MgO, and Fig. 4.5. V vs. Ti diagram (Shervais, 1982). Symbols are as in Fig. 4.3 and 4.4. Chapter 4 · Petrogenesis 68 TiO 2 , Fe 2 O 3T , P 2 O 5 , Sr and Y show an uneven distribution (Fig. 4.6, 4.7). Overall, these variations agree with magma fractionation from more primitive to more evolved compositions. The picrite sample, however, shows the effect of olivine accumulation for many elements besides MgO, as it has anomalously low concentrations in TiO 2 and CaO as Fig. 4.6. Variations in major element concentrations with decreasing MgO. Symbols are as in Fig. 4.3 and 4.4. Petrogenesis · Chapter 4 75 Fig. 4.12. Bivariate diagrams of 87 Sr/ 86 Sr vs. major and trace element parameters. Symbols are as in Fig. 4.3 and 4.4. Rossy et al. (1992) showed that the Nd isotope composition of the Cretaceous magmatism is significantly less radiogenic than that of the ultramafic bodies cropping out in the Pyrenees (Downes et al., 1991). It is also different from the Nd isotope composition of ultramafic xenoliths included in Quaternary lavas in the Catalonian Coastal Ranges, analysed by Bianchini et al. (2007). Since these ultramafic rocks can be taken as samples of the local lithosphere, these observations agree with an asthenospheric origin for the studied magmas. The Pyrenean ultramafic bodies are cross-cut in certain areas by a late generation of amphibole-pyroxenite veins (Lagabrielle and Bodinier, 2008 and references therein) that are considered to represent melt conduits for the Cretaceous alkaline magmatism (Golberg et al., 1986; Bodinier et al., 1987). These veins are coeval with the Cretaceous magmatism (Albarède and Michard-Vitrac, 1978; Golberg et al., 1986; Henry et al., 1998; Ubide et al., under review a –see Chapter 3-) and have 143 Nd/ 144 Nd isotope ratios (Downes et al., 1991; Mukasa et al., 1991) of 0.512698 to 0.512874 (recalculated to 100 Ma). These values overlap with the recalculated 143 Nd/ 144 Nd ratios obtained in this study (0.512656 – 0.512760; Table 4.2), lending further support to the genetic link between the amphibole veins and the Cretaceous magmas. In order to obtain additional constraints on the mantle source, incompatible trace elements from samples closest to primary melts need to be regarded. Only samples with MgO > 7 wt. % have been considered. In addition, rocks with high volume fractions of Chapter 4 · Petrogenesis 76 large crystals have been dismissed as their composition is strongly influenced by the accumulation of the crystals (see Chapters 5 and 6). The primitive mantle-normalised multielemental patterns of the samples that meet these criteria are presented together in Fig. 4.13. Fig. 4.13. Multielemental patterns of selected samples normalised to the primitive mantle composition by McDonough and Sun (1995). Note that only certain samples from the BasqueCantabrian Basin, the Eastern Pyrenees and the Catalonian Coastal Ranges fit the criteria outlined in the preceding paragraph. The rocks show positive anomalies for Nb-Ta and generally smaller ones for Ba; sam ples from the Basque-Cantabrian Basin and Eastern Pyrenees also present a negative anomaly in Pb (Fig. 4.13). These data point to an asthenospheric enriched mantle source similar t o EM-1 and/or HIMU (Hofmann, 1997). Pb systematics confirms mixing of HIMU and EMtype components in a heterogeneous source (Fig. 4.10). The fact that samples from the Catalonian Coastal Ranges show no anomalies or slightly positive anomalies for Pb in trace element patterns (Fig. 4.13) may reflect a small contribution of an EM-2 component, which could be related to the incorporation of subducted crustal components into the mantle (Zindler and Hart, 1986), since crustal contamination was not important in the genesis of these magmas (Fig. 4.11, 4.12). The involvement of an EM-2 component may also contribute to the 207 Pband 208 Pb-enrichments of the samples (Orejana et al., 2008; Fig. 4.10). The enrichment in incompatible elements and volatiles, as revealed by the trace element patterns and the abundance of lamprophyre rocks and hydrous phases, suggest a metasomatic modification of the mantle source. The nature of the metasomatic phases can be investigated through the composition in diagnostic trace elements. The samples show positive correlations in Rb/La vs. K/La plots, as well as in Ba/Nb vs. Ba/Ce plots (not shown). As explained in Orejana et al. (2008), these features might account for the presence of phlogopite in the source during partial melting as this mineral may preferentially incorporate LILE when compared to REE and HFSE. On the other hand, the Petrogenesis · Chapter 4 77 relative abundances of LILE provide alternative constraints on the composition of the source. Melts in equilibrium with phlogopite are expected to present Rb/Sr > 0.1 and Ba/Rb < 20, whereas melts in equilibrium with amphibole typically have Rb/Sr < 0.06 and Ba/Rb > 20 (Furman and Graham, 1999; McCoy-West et al., 2010). The studied mafic samples have Rb/Sr ratios of 0.00 – 0.15 and Ba/Rb ratios of 6.52 – 65.46. Although these ratios are not conclusive, they may suggest that amphibole could predominate over phlogopite in the mantle source. The high concentrations in Nb and Ta support amphibole as the main relevant phase (e.g., Ionov et al. 1997), whereas the role of phlogopite is probably less important, as this is not a frequent phase in the rocks. Finally, rocks with moderate P 2 O 5 contents such as the studied samples (0.2 – 1 wt. %; Table 4.1) indicate that a P-rich phase, such as apatite, might have been a stable metasomatic mineral in the mantle source (e.g., Orejana et al., 2008). Even though the trace element patterns across northeast Iberia are similar to each other, there are small differences between samples from different domains (Fig. 4.13). Samples from the Eastern Pyrenees are the most enriched in LILE and HFSE and show a deep negative anomaly for K and Pb. Their patterns are therefore the closest to HIMU (e.g., Woodhead, 1996). Their Pb isotope systematics has clear HIMU signature as well (Fig. 4.10). Regarding the alkaline character of the rocks, (La/Lu) N values are highest for the Eastern Pyrenees and Catalonian Coastal Ranges, and in the TAS diagram, these rocks define more SiO 2 -undersaturated trends than rocks from the Basque-Cantabrian Basin and the North Pyrenean Basins (Fig. 4.3). It follows that the alkaline character decreases from east to west. Accordingly, mantle melting rates may have been higher or mantle melting pressures lower towards the west (Kushiro, 2001). Moreover, magmatic activity is more enduring and developed in the west, and progressively scarcer and more intrusive towards the east (Ubide et al., under review a –see Chapter 3-). All of these differences are in agreement with the geodynamic scenario of opening of the Bay of Biscay and anticlockwise rotation of Iberia that enabled the development of the studied magmatism, as this kind of setting would trigger extensional conditions increasing towards the west (Ubide et al., under review a –see Chapter 3-). The strong structural control on the development of the magmatism favours a non-plume scenario, as proposed by Matton and Jébrak (2009) for other Cretaceous magmatisms of the Peri-Atlantic Alkaline Pulse. The discussion of geochemical data on the Cretaceous magmatism in northeast Iberia leads to the conclusion of a common, asthenospheric, enriched/metasomatised and heterogeneous mantle source with the involvement of HIMU and EM-1 components and a possible contribution of EM-2 component. Melting took place in the garnet stability field and responded to higher tectonic extension rates towards the west. The resulting melts have similar trace element and isotopic signatures, and it is interesting to note that elemental compositions are homogeneous whereas isotope compositions are heterogeneous, although still comparable among the different domains. Chapter 4 · Petrogenesis 78 4.6.2. Evolution of the mantle source since the beginning of the Alpine cycle During the Alpine cycle, the northeast of the Iberian realm was the locus of several tectono-magmatic events associated to the relative movements of Iberia (e.g., Arranz et al., 2011 and references therein). The study of the magmatisms can provide insights into the evolution of the mantle sources in the area. First, as the Variscan compression progressively ceased favouring the emplacement of calc-alkaline plutons (Arranz and Lago, 2004; Castro et al., 2002), the tectonic regime evolved to a transcurrent and finally extensional context which resulted in thinning of the crust (Lago et al., 2004a and references therein). This enabled partial melting of the lithospheric mantle, triggering the emplacement of hypabyssal or even extrusive calkalkaline and mildly alkaline magmas during the Permian. Today, these rocks are observed in the Pyrenees, the Catalonian Coastal Ranges and the Iberian Chain (Bixel, 1988; Bonin, 1988; Cabanis and Le Fur-Balouet, 1989; Debon and Zimmermann, 1993; Galé, 2005; GilImaz et al., 2012; Lago et al., 2004a, 2004b, 2012; Ubide et al., 2010b). The Triassic rifting stage, related to the opening and westward extension of the Tethys Ocean (e.g., Salas and Casas, 1993), enabled widespread tholeiitic magmatism (“ophites”) in the western and central Pyrenean basins (Azambre et al., 1987). In addition, a Triassic alkaline magmatism was developed in the eastern Pyrenean margin (Azambre and Rossy, 1981). Ubide et al. (under review a –see Chapter 3-) have reported alkaline extrusive rocks of Triassic age in the Central Pyrenees as well. Alkaline mafic rocks are also recognised in the southern Catalonian Coastal Ranges and the Iberian Chain (Lago et al., 1996, 2000; Sanz et al., 2013). Most of the Jurassic period constituted a post-rift stage (Salas and Casas, 1993) in northeast Iberia. The magmatic activity in this period is restricted to isolated, alkaline extrusive rocks cropping out in the Central Pyrenees (Ubide et al., under review a –see Chapter 3-) and the Iberian Chain (Martínez González et al., 1997; Ortí and Vaquer, 1980). During the Cretaceous, the opening of the Bay of Biscay created a rifting setting favourable for the development of an alkaline magmatic province in northeast Iberia (Azambre et al., 1992; Cabanis and Le Fur-Balouet, 1990; Montigny et al., 1986; Rock, 1982; Solé et al., 2003; Ubide et al., under review a –see Chapter 3-). This magmatism crops out in the Pyrenees and northern Catalonian Coastal Ranges and is the focus of the present study. Finally, the most recent magmatic activity took place during the Quaternary. An alkaline volcanic province developed to the north-northwest of the Catalonian Coastal Ranges (south of the Pyrenees) as the Iberian realm was subjected to extensional conditions after Petrogenesis · Chapter 4 79 the Alpine compression (e.g., Araña et al., 1983; Cebriá et al., 2000). This province is frequently cited as the Garrotxa Volcanic Field (e.g., Aulinas et al., 2013; Gasperini et al., 2013; Gisbert et al., 2009, 2011) and is related to the European Cenozoic rift system (Gasperini et al., 2013). The diversity of magmatisms recorded in northeast Iberia, most of which are dominated by mafic compositions, suggests a heterogeneous mantle beneath this area. Nevertheless, when comparing the geochemistry of these magmatisms, it turns out that the asthenospheric, enriched and heterogeneous HIMU-EM signature defined for the Cretaceous magmas (see section 4.6.1) is, strikingly, present in rocks of very different age. The primitive mantle-normalised multielemental patterns of the magmatisms with similarities to the Cretaceous magmas are presented in Fig. 4.14 (for the sake of simplicity, magmatisms with distinct compositions are not shown). In order to avoid evolved as well as cumulative compositions, only samples with MgO concentrations between 7 and 9 wt. % are plotted. Many of the oldest samples have high LOI values, but only those with contents below 5 wt. % are plotted. For comparison, the mafic compositions plotted in Fig. 4.13 are included as a grey field in the background (Fig. 4.14). Significant similarities with the Cretaceous melts are found: in the Permian magmatism of the SE Iberian Chain (Lago et al., 2012), in the Triassic alkaline magmatisms of the eastern Pyrenean margin (Azambre and Rossy, 1981) and Iberian Chain (Lago et al., 2000), in the Jurassic magmatism of the Iberian Chain (unpublished data by the research group; see Martínez González et al., 1997 for a synthesis of this magmatism), and in the Quaternary magmatism of the Catalonian Coastal Ranges (Cebriá et al., 2000). The Permian magmatism of the SE Iberian Chain is most similar to HIMU (e.g., Woodhead, 1996), whereas the magmatisms closest in time to the Cretaceous (Jurassic and Quaternary) are also closest in composition to the Cretaceous magmatism. Other than the northeast part of Iberia, the Cretaceous compositional features are observed in the Permian magmatism of the Spanish Central System (Orejana et al., 2008) and also in the Triassic alkaline magmatism of Majorca (unpublished data by the research group; see Lago et al., 1996 for a synthesis of this magmatism). The similarities observed in normalised multielemental patterns are reproduced in the SrNd and Pb-Pb isotope systems, when measured (Fig. 4.15). Regarding the Sr-Nd isotope plot, Jurassic and Quaternary magmatisms agree with the least radiogenic Cretaceous data; the Permian and Triassic magmatisms discussed above have more scattered Sr-Nd isotope ratios, but some data agree with the Cretaceous compositions as well. Other than that, it is clear that the crustal signature decreased since the end of the Variscan orogeny, marked by the development calc-alkaline plutonism and volcanism with high 87 Sr/ 86 Sr Chapter 4 · Petrogenesis 80 Fig. 4.14. Primitive mantle (McDonough and Sun, 1995) normalised multielemental patterns of Permian to Quaternary magmatisms in Iberia, to be compared to the Cretaceous compositions (see data in Fig. 4.13). Permian compositions from the Spanish Central System and the SE Iberian Chain are from Orejana et al. (2008) and Lago et al. (2012), respectively; Triassic alkaline compositions from the eastern Pyrenean margin, Iberian Chain and Majorca are from Azambre and Rossy (1981), Lago et al. (2000) and unpublished data by the research group (see section 4.6.2), respectively; Jurassic compositions from the Iberian Chain are unpublished data by the research group (see section 4.6.2); Quaternary compositions from the Catalonian Coastal Ranges are from Cebriá et al. (2000). From among the compositions provided in these studies, only those fitting the criteria outlined in the preceding page are plotted. HIMU composition from Woodhead (1996). Petrogenesis · Chapter 4 81 Fig. 4.15. Isotope systematics of Iberian magmatisms developed since the beginning of the Alpine cycle. SrNd data are recalculated considering the age of the rocks. Carboniferous (Variscan) calc-alkaline plutonism in the Pyrenees from Roberts et al. (2000); Lower Permian (Vari scan) calc-alkaline magmatism in the Iberian Chain from Lago et al. (2004b); Middle-Upper Permian magmatism in the Pyrenees, Spanish Central System and SE Iberian Chain from Galé (2005), Orejana et al. (2008) and Lago et al. (2012), respectively; Triassic tholeiitic magmatism in the Pyrenees and Triassic alkaline magmatism in the eastern Pyrenean margin from Alibert (1985); Triassic alkaline magmatism in the Central Pyrenees and Iberian Chain as well as Jurassic magmatism in the Iberian Chain are unpublished data by the research group; Cretaceous magmatism in the Pyrenees and Catalonian Coastal Ranges from this study and Rossy et al. (1992) (see data in Fig. 4.9; 4.10); Quaternary magmatism in the northern Catalonian Coastal Ranges from Cebriá et al. (2000). Mantle component fields compiled from GEOROC database. Chapter 4 · Petrogenesis 82 ratios and low 143 Nd/ 144 Nd ratios (Fig. 4.15), probably due to increasing extensional conditions. Pb isotope analyses are scarcer in previous studies (Fig. 4.15). The Permian, Triassic, Jurassic and Quaternary magmatisms discussed above are similar in terms of Pb isotopic composition to the least radiogenic Cretaceous data, supporting the petrogenetic link suggested on the basis of trace elements and Sr-Nd isotopes. Cretaceous Pb isotope ratios, however, reach most radiogenic concentrations, not recorded in any other magmatism in Iberia. The Permian magmatism in the SE Iberian Chain has normalised multielemental patterns very similar to HIMU (Fig. 4.14; Lago et al., 2012), but Pb isotope analyses have not been carried out yet. T NdDM model ages have been calculated for the magmatisms with available 143 Nd/ 144 Nd and 147 Sm/ 144 Nd isotope ratios. The results are presented in Fig. 4.16 in relation to the age of the rocks. The Jurassic and Cretaceous magmatisms yield consistent model ages that overlap with each other (0.53-0.58 Ga and 0.54-0.60 Ga, respectively; see Cretaceous data in Table 4.2). Older magmatisms yield more heterogeneous model age values. Still, the youngest model ages related to the Permian and Triassic magmatisms considered above agree with the Cretaceous values. These results may reflect a Cadomian fractionation event in the mantle related to the genesis the petrogenetically linked magmas. In contrast, magmatisms unrelated to the Cretaceous signature have older model ages (Fig. 4.16). According to the variability of model age data, the Cadomian event may have affected older mantle domains, rejuvenating their isotopic compositions and probably modifying their overall composition. Fig. 4.16. Age vs. T-depleted mantle model ages (T NdDM ) plot of Iberian magmatisms developed since the beginning of the Alpine cycle. Literature data from Alibert (1985), Galé (2005), Lago et al. (2004b, 2012), Orejana et al. (2008) and Roberts et al. (2000). Triassic alkaline magmatism in the Central Pyrenees and Iberian Chain as well as Jurassic magmatism in the Iberian Chain are unpublished data by the research group; Cretaceous magmatism in the Pyrenees and Catalonian Coastal Ranges from this study (see data in Table 4.2). Symbols are as in Fig. 4.15. Petrogenesis · Chapter 4 83 The evaluation of trace element and isotope compositions of Iberian magmatisms since the beginning of the Alpine cycle indicates that the mantle source related to the studied Cretaceous melts probably represents a common, asthenospheric and enriched mantle domain, of considerable areal extension and long-living since Permian to Quaternary times. It is equally important to note that there are also many other signatures unrelated to this source, such as crustal signatures in the Carboniferous and Permian magmatisms (e.g., Galé, 2005; Lago et al., 2004a, 2004b; Ubide et al., 2011) or the MORB-like magmatism related to the widespread development of tholeiites during the Triassic (Azambre et al., 1987). It follows that the mantle below Iberia is heterogeneous on a large scale as well. The different mantle sources were tapped throughout the Alpine cycle according to the reigning tectonic conditions and therefore, plume-related scenarios as proposed in certain studies (e.g, Cebriá et al., 2000; Piromallo et al., 2008) are less likely. In line with this, the Cretaceous magmatism is related to extension of the lithosphere, rather than to mantle upwelling (Ubide et al., under review a –see Chapter 3-). 4.7. Conclusions The Cretaceous alkaline magmatism in northeast Iberia is recorded in numerous outcrops in the Pyrenees and the Catalonian Coastal Ranges. The petrology and geochemistry of the rocks reveals that: • There is a wide petrological variability including basic and acid, intrusive and extrusive rocks. Towards the west, the magmatism (and specifically volcanism) is more developed and related to higher melting rates of the mantle source. This can be correlated with extensional conditions increasing towards the west given the related geodynamic context of opening of the Bay of Biscay and anticlockwise rotation of Iberia. • The Cretaceous melts were extracted from a heterogeneous and metasomatised, asthenospheric mantle source with the involvement of HIMU and EM-type components. The model ages reflect a Cadomian fractionation event in the mantle. • This mantle signature is recorded in other magmatisms in Iberia w it h age s spanning from Permian to Quaternary. It therefore may represent an extensive and long-living source beneath the region, tapped throughout the Alpine cycle under favourable tectonic conditions. Chapter 4 · Petrogenesis 84 The lamprophyre suite · Chapter 5 91 Fig. 5.2. Field photographs of the Aiguablava lamprophyre. A-D) Sa Planassa – Punta des Mut outcrop, where the sill cross-cuts a Late Variscan lamprophyre. E) Cap Rubí outcrop. F) Platja Fonda outcrop, where the wedge-shaped ends of the sill are observed (white arrows). All of the sub-horizontal intrusions show a characteristic vertical zoning, as schematised in Fig. 5.7. Well developed chilled margins appear at both sides, in agreement with the difference in age between the Cretaceous magmas and the Paleozoic country-rock (e.g., Ubide et al., 2008). The chilled margins are mostly aphyric and sometimes laminated. The central facies on the other hand, displays a remarkable accumulation of large crystals towards the bottom, due to gravitational settling before solidification (Ubide et al., 2012a –see Chapter 6-). Vesicles are frequent in all the intrusions and reach millimetre sizes. Chapter 5 · The lamprophyre suite 92 Fig. 5.3. Field photographs of the Calella de Palafrugell lamprophyre. A-D) Ses Negres outcrop. White arrows in A mark the ends of the sill in the picture. E) The sill crops out above the nature trail along the coast (dotted line), although it is hardly accessible and more altered than in the Ses Negres outcrop. Vesicle alignments are commonly recognised parallel to the upper chilled margin. The abundance of vesicles together with the subhorizontal dip of most intrusions indicates a shallow level of emplacement for these magmas (Gimeno, 2002; Ubide et al., 2012a). The nicest examples of vertical zoning are found in the subhorizontal sills cropping out by the shoreline, where they are clearly exposed given the absence of soil and vegetation. The sketch in Fig. 5.7 is from the Aiguablava sill. This intrusion crops out in the Aiguablava The lamprophyre suite · Chapter 5 93 Fig. 5.4. Field photograph of the S’Agaró lamprophyre, located in Cala de la Font. Note that this is the only sub-vertical intrusion of the whole lamprophyre suite. It intersects a Late Variscan lamprophyre. Fig. 5.5. Field photograph of the Llagostera lamprophyre. Although there are a few meters of exposure in the neighbouring plot, we could study and sample this less weathered underground section dug for building purposes. Nevertheless, this sill is more altered than those cropping out in the coast. beach area, in three different locations that are less than 1 km away from each other. In the Sa Planassa – Punta des Mut outcrop the sill reaches its maximum thickness (ca. 2 m; Fig. 5.2A, 5.2B). Emplacement mechanisms are very clear in this outcrop (Gimeno, 2002), including fragments of country-rock dragged inside the intrusion during emplacement (Fig. 5.2C) and gravitational accumulation of large crystals (Fig. 5.2D). In Cap Rubí outcrop Chapter 5 · The lamprophyre suite 94 Fig. 5.6. Field photographs of the lamprophyres of the western sector, which is intensely vegetated. A) The Molí d’en Ponç lamprophyre. B-C-D) The Molí de Dalt lamprophyre. E-F) The Monasterio de Sant Pere Cercada lamprophyre, which crops out in a river bed. These sills are more altered than those cropping out in the coast. the sill has an intermediate thickness and a staircase morphology indicating that the magma emplaced along pre-existing fractures (Fig. 5.2E). Finally, in Platja Fonda outcrop the wedge-shaped end of the sill can be observed (Fig. 5.2F). The groundmass crystal size decreases progressively from the Sa Planassa – Punta des Mut outcrop to the Platja Fonda outcrop, correlating with decreasing magma cooling times from the thickest to the thinnest parts of the sill (Ubide et al., 2008). The lamprophyre suite · Chapter 5 95 Fig. 5.7. Sketch and field photographs of the Aiguablava lamprophyre in Cap Rubí outcrop. Note the vertical structure of the sill, including: 1) sharp contacts with the country-rock; 2) well developed, mostly aphyric chilled margins that sometimes show lamination; 3) central facies with gravitational accumulation of large, mafic crystals towards the bottom; and 4) vesicle alignments parallel to the upper chilled margin. This structure is common to all the Cretaceous lamprophyres in the Catalonian Coastal Ranges excepting the S’Agaró lamprophyre, which is the only sub-vertical and microlitic intrusion. The other clear example of vertical zoning is the Calella de Palafrugell sill (Fig. 5.3). This is moreover the thinnest intrusion of the whole Cretaceous lamprophyre suite (Table 5.1), allowing for a detailed study of the physical processes involved in the generation of the vertical zoning and its effects on the composition of the lamprophyre (Ubide et al., 2012a –see Chapter 6-). 5.4. Samples and methods Several samples were collected from all lamprophyre intrusions (Table 5.1). Samples from the Aiguablava lamprophyre cover the three outcrops of the sill (see Ubide et al., 2008), whereas samples from the Calella de Palafrugell lamprophyre cover the vertical section of the sill (see Ubide et al., 2012a; Fig. 6.2). After a careful macroscopic examination of the samples, several sections were selected from each intrusion for microscopic examination. Chapter 5 · The lamprophyre suite 96 Polished thin sections were prepared at the Servicio General de Apoyo a la InvestigaciónSAI, Universidad de Zaragoza (Spain). After a detailed petrographic study, thin sections were selected for in situ, spot analyses of major and trace elements on minerals. Major elements were analysed for all of the lamprophyres, whereas trace elements were obtained on thin sections from the Calella de Palafrugell lamprophyre and Monasterio de Sant Pere Cercada lamprophyre. Finally, whole rock analyses including major and trace element concentrations were carried out for all of the lamprophyres. Mineral major element compositions were determined by electron microprobe (EMP) at the Centro Nacional de Microscopía Electrónica of the Universidad Complutense de Madrid (Spain), using a JEOL JZA-8900M electron microprobe equipped with four wavelength dispersive spectrometers. Analyses were performed using an accelerating voltage of 15 kV and an electron beam current of 20 nA, with a beam diameter of 5 μm. Elemental counting times were 10 s on the peak and 5 s on each of two background positions. Corrections for inter-elemental effects were made using a ZAF (Z: atomic number; A: absorption; F: fluorescence) procedure. Mineral trace element compositions were determined by laser ablation ICP-MS (LA-ICPMS) at the Centro de Instrumentación Científica-CIC of the Universidad de Granada (Spain), using a 213 μm Mercantek Nd-YAG laser coupled to a quadrupolar Agilent 7500 ICP-MS with a shielded plasma torch. The ablation was carried out in a He atmosphere, using a laser beam with a diameter fixed at 80 µm, a repetition rate of 10 Hz and an output energy of 0.3 mJ per pulse. The spot was preablated for 45 s using laser output energy of 50 %. The spot was then ablated for 60 s with laser output energy of 70 %. NIST610 glass was systematically analysed as an external standard. Concentration values were corrected using the silicon concentrations obtained by electron microprobe as an internal standard. Further details on technical methods can be found in Bea et al. (2005). Data reduction was carried out by the CIC laboratory staff. Samples were reduced to powder for whole rock analyses. Given the strongly porphyritic nature of most lamprophyres, large amounts of sample material were used to assure an accurate representation of the analysed powders. Samples were crushed in a manganese steel jaw-crusher, milled in an agate vibrating cup mill and successively split down to few grams at the Servicio General de Apoyo a la Investigación-SAI, Universidad de Zaragoza (Spain). Major and trace element concentrations were determined at the Service d'Analyse des Roches et des Minéraux (SARM) in Nancy (France). The samples were analysed by ICP-AES for major elements and ICP-MS for trace elements. Details of the analytical procedures and detection limits are available at http://www.crpg.cnrsnancy.fr/SARM/. In addition, the bulk composition of the groundmass of the Calella de Palafrugell lamprophyre was modelled in Ubide et al. (2012a –see Chapter 6-). The lamprophyre suite · Chapter 5 97 5.5. Petrography The Cretaceous lamprophyres in the Catalonian Coastal Ranges are classified as camptonites according to criteria by Le Maitre (2002): they have abundant large crystals of mafic minerals (mostly pale-pink clinopyroxene and brown amphibole) in a groundmass where plagioclase predominates over alkali feldspar. The complete mineral assemblage consists of brown amphibole, pale-pink clinopyroxene, feldspars, opaque minerals, pseudomorphosed olivine and accessory biotite, analcime, apatite, calcite, chlorite and recrystallised glass. Most lamprophyres have porphyritic textures, with large mafic crystals embedded in a fine-grained groundmass (Fig. 5.8). According to their size, the large crystals are considered “macrocrysts” (1 – 8 mm, commonly ca. 3 mm) or “megacrysts” (ca. 1 cm). Macrocrysts are ubiquitous, whereas megacrysts occur exclusively in the Llagostera and Molí d’en Ponç sills (Fig. 5.8); therefore, large crystals will be referred to as macrocrysts hereafter. The groundmass on the other hand, contains “microcrysts” (< 5 0 – 500 µm). Microcrysts are further subdivided into “microphenocrysts” (up to 500 µm in size) and “microlites” (commonly 100 – 200 µm in size). The S’Agaró lamprophyre is the only intrusion with a microlitic texture, solely composed of microcrysts, with no macrocrysts (Fig. 5.8). 5.5.1. The macrocryst assemblage The macrocryst assemblage includes clinopyroxene, amphibole, opaque minerals and pseudomorphosed olivine. They sometimes cluster together into glomerocrysts. Xenocrysts of quartz and plagioclase occur occasionally. The macrocrysts accumulate towards the bottom of the central facies in all the sub-horizontal intrusions (see Fig. 5.7). Their volume fraction in the accumulation zone varies among the different intrusions and is highest for the Llagostera and Molí d’en Ponç lamprophyres, where large crystals reach more than 40 vol. % of the rock (Fig. 5.8). The proportion of each mineral phase in the macrocryst assemblage varies among the different lamprophyres (Table 5.1; Fig. 5.8). Clinopyroxene predominates in the Calella de Palafrugell lamprophyre and in the rounded blocks that occur near Molí d’en Ponç, whereas amphibole predominates in the Aiguablava, Llagostera and Monasterio de Sant Pere Cercada intrusions. Clinopyroxene and amphibole appear in similar proportions in the Molí de Dalt lamprophyre. Finally, clinopyroxene and olivine predominate in the Molí d’en Ponç intrusion. Opaque minerals are always a minor macrocryst phase, but the highest proportions occur in the Calella de Palafrugell lamprophyre (see Fig. 5.8). Clinopyroxene and amphibole macrocrysts have complex zoning patterns and an average size of 3 – 4 mm. Clinopyroxene macrocrysts occur in all the porphyritic lamprophyres. Chapter 5 · The lamprophyre suite 98 Fig. 5.8. Scanned thin sections of the studied lamprophyres (the width of each image is 2.5 cm). They represent the zone of accumulation of macrocrysts in each intrusion, excepting CAL-5 which belongs to the upper chilled margin of the Calella de Palafrugell lamprophyre and SAG-03B which belongs to the microlitic lamprophyre from S’Agaró. The pictures in the lower row were obtained by sandwiching the thin sections between two polarising sheets with perpendicular polarisation directions, to obtain “crossed polars” images. The thin sections from samples CAL-5 and SFB-5 show circles for mineral analysis by EMP. The lamprophyre suite · Chapter 5 99 Most of them are pale-pink-coloured under plane polarised light (Fig. 5.9A). They have complexly zoned subhedral to anhedral cores overgrown by euhedral to subhedral rims. In the central and western sectors, where crystals reach largest sizes (megacrysts of ca. 1 cm), there are colourless, anhedral inner cores with highly irregular contours (Fig. 5.9B). These zones of colourless clinopyroxene under plane-polarised light show second order interference colours under cross-polarised transmitted light. Particularly in the western sector, crystals with colourless inner cores show colourless thin mantle s around the cores. These colourless mantles are further overgrown by pale-pink clinopyroxene, generating a striking alternation of colours (Fig. 5.9B). Moreover, there is an enclave of ca. 2 cm diameter composed of colourless clinopyroxene crystals. Meanwhile in the eastern sector, there is a pale-green, anhedral core in the Calella de Palafrugell lamprophyre; it presents resorption features, as it is corroded and displays an engulfed irregular contour (Fig. 5.9C). There are pale-green cores in the sills from the western sector as well, but they are subhedral and do not show resorption features (Fig. 5.9D). Amphibole macrocrysts occur in all the porphyritic lamprophyres, with the exception of the rounded blocks that occur near the Molí d’en Ponç. Amphibole macrocrysts normally have homogeneous, anhedral and rounded cores and euhedral to subhedral rims. They are reddish brown-orange under plane polarised light, with different hues for the core and rim areas (Fig. 5.9E). In the eastern and western sectors, some macrocryst cores have an anhedral and rounded inner core, brown in colour (Fig. 5.9F). These inner cores are darker in the sills from the western sector than in the sills from the eastern sector. Strikingly, brown inner cores have not been recognised in the Llagostera lamprophyre (central sector), despite the fact that in this intrusion, amphibole makes up very large crystals (megacrysts of ca. 1 cm). Opaque minerals and olivine pseudomorphs are usually smaller than clinopyroxene and amphibole macrocrysts. Macrocrysts of opaque minerals occur in all the porphyritic lamprophyres. They sometimes appear as single macrocrysts of ca. 1 mm size, especially in the Calella de Palafrugell lamprophyre (Fig. 5.8). More frequently, they are associated with macrocrysts of clinopyroxene and/or amphibole. Furthermore, alignments of sulphide inclusions occur in the inner zones of some clinopyroxene and amphibole macrocrysts. This type of inclusions has been interpreted as droplets of immiscible sulphide liquids adhering to the surfaces of rapidly growing silicate crystals (Demant et al., 2010; Peterson and Francis, 1977; Woodland and Jugo, 2007). Olivine macrocrysts occur in all the porphyritic lamprophyres, with the exception of the Llagostera intrusion. Although olivine macrocrysts normally have a small size of ca. 1 mm, this mineral makes up megacrysts of ca. 1 cm in the Molí d’en Ponç lamprophyre. Olivine is completely transformed to secondary assemblages and unaltered relicts have only been found in a macrocryst of sample SFB-4 from the Molí d’en Ponç lamprophyre. Chapter 5 · The lamprophyre suite 100 Fig. 5.9. Photomicrographs of clinopyroxene (A-D) and amphibole (E-F) macrocrysts. Plane-polarised (left) and cross-polarised (right) transmitted light. The same magnification was used in all the cases. Compositional types are summarised in Table 5.2. 5.5.2. The groundmass The groundmass is very similar in all of the porphyritic lamprophyres and equivalent to the microlitic lamprophyre from S’Agaró (Fig. 5.10). It is composed of microcrysts of feldspars, amphibole, opaque minerals, acc es sory apatite and glass. The latter is mostly recrystallised to chlorite and other secondary products that are indistinguishable under the petrographic microscope. Clinopyroxene is present in minor proportions in all the lamprophyres, although in the central sector it is almost completely replaced with The lamprophyre suite · Chapter 5 107 Fig. 5.14. Amphibole major element composition in bivariate diagrams vs. decreasing Mg#. simpler than in the case of clinopyroxene, has a correlation with compositional changes as well (Table 5.2; Fig. 5.14). 3 compositional types have been defined in macrocryst cores: J, K and P. In contrast with clinopyroxene, most amphibole macrocryst cores, including those of the largest crystals at Llagostera (Fig. 5.8), are unzoned and composed entirely of P compositions. Macrocryst rims have a composition that has been named as Chapter 5 · The lamprophyre suite 108 R, following the same criterion as in clinopyroxene. In the groundmass, microphenocrysts have a simple zoning into P-type cores and R-type rims. They therefore have a zoning pattern equivalent to that of most macrocrysts. Finally, the microlites are unzoned and have an R composition. The temporal relationships between the different compositional types in amphibole are presented in Table 5.4, according to their relative stratigraphy in the analysed crystals. Note that R compositions are the last to crystallise. Compositional trends in amphibole are not as clearly defined as in clinopyroxene (compare Fig. 5.12 and Fig. 5.14) and R compositions show a significant scatter because they have a slightly different slope in the lamprophyres from the central sector relative to the rest. Yet, P-R compositions define a normal zoning trend of decreasing Mg# and MgO (Fig. 5.14). Ptype macrocryst cores and microphenocryst cores are similar in composition, especially considering the Mg# values. However, P compositions in macrocryst cores reach highest MgO, K 2 O and SiO 2 concentrations and lowest TiO 2 contents. The rounded, brown-coloured inner cores observed in some of the macrocrysts in the eastern and western sectors (Fig. 5.9F) have a distinct composition. In the western sector, they have a dark brown colour and a composition named as J. In the eastern sector, they have a lighter brown colour and a composition named as K. Both the Jand K-type inner cores are overgrown by orange-coloured P compositions that constitute the rest of the macrocryst core. K-type inner cores have lower Mg# and MgO contents than P compositions (Fig. 5.14). Jtype inner cores reach even lower Mg# and MgO contents and moreover, higher SiO 2 and lower TiO 2 and Al 2 O 3 concentrations. J-P and K-P macrocryst cores therefore define reverse zoning patterns. In contrast with the other amphibole compositions, J-type inner cores are mainly classified as pargasite and in a lower proportion as kaersutite (Fig. 5.13). The difference with the other compositions is especially clear for TiO 2 , with concentrations down to ca. 3.5 wt. % in J compositions (Fig. 5.14). Feldspars More than 100 analyses were obtained from felspars, which occur as microlites in all o f the lamprophyre intrusions. The analyses are presented in Table 5.III of the electronic supplement. Table 5.4. Amphibole crystal stratigraphy (as in Table 5.3). Note that J and K compositions are always inner cores, that P and R compositions overgrow most of the others and that R compositions are the last to crystallise. Mantling phases J K P R Mantling pahses J   K   P  R The lamprophyre suite · Chapter 5 109 The ternary classification diagram anorthite – albite – orthoclase is shown in Fig. 5.15. Feldspar compositions vary continuously from labradorite to sanidine (An 0169 , Ab 30-72 , Or 01-63 ). Plagioclase-type compositions are progressively enriched in the orthoclase component as their anorthite content decreases, a feature typical of alkaline rocks. The range of compositions is similar among the different lamprophyres. Nevertheless, most analyses from the eastern sector range from labradorite to oligoclase whereas all analyses obtained from the S’Agaró lamprophyre are classified as sanidine (see Table 5.III of the electronic supplement). Opaque minerals More than 70 analyses were obtained from macrocrysts and microcrysts of opaque minerals. They are presented in Tables 5.IV and 5.V of the electronic supplement. The macrocrysts occur in all of the porphyritic lamprophyres and the microcrysts occur in all of the lamprophyres. Opaque minerals are mainly Fe-Ti oxides. Most of them are spinel group minerals, but there are also minor amounts of ilmenite. Cr 2 O 3 concentrations are below 2 wt. %. Spinel group compositions are plotted in bivariate diagrams vs. decreasing MgO in Fig. 5.16. Magnetite – titanomagnetite compositions are frequent in the microcrysts and have variable TiO 2 concentrations between 2 and 18 wt. %. The rest of compositions, including all macrocrysts and some microcrysts, have significant amounts of MgO and Al 2 O 3 and a correlative decrease in FeO T (not shown), indicating the involvement of the spinel endmember. Olivine Olivine pseudomorphs occur as macrocrysts in the eastern and western sectors. Analyse s were carried out on the unaltered relicts found in a macrocryst core from the Molí d’en Ponç lamprophyre (sample SFB-4) and are presented in Table 5.VI of the electronic supplement. Fig. 5.15. Felspar classification diagram anorthite – albite – orthose. Chapter 5 · The lamprophyre suite 110 Fig. 5.16. Spinel group minerals major element composition in bivariate diagrams vs. decreasing MgO. The compositions are highly homogeneous, with forsterite contents of Fo 88-89 , Mg# val ues of 80 – 82, NiO concentrations varying in the range 0.16 – 0.26 wt. %, and CaO concentrations varying from 0.22 to 0.26 wt. %, away from compositions in peridotite (Fig. 5.17). Fig. 5.17. Olivine major element composition in bivariate diagrams vs. decreasing forsterite. Fields for abyssal peridotites are from Sakyi et al. (2012; and references therein). Accessory minerals Analyses on accessory phases were carried out for biotite, analcime and apatite. They are presented in Tables 5.VII, 5.VIII and 5.IX of the electronic supplement. The lamprophyre suite · Chapter 5 111 Biotite has concentrations in MgO from 6 to 16 wt. % and concentrations in FeO T of 9 to 24 wt. %. TiO 2 and K 2 O concentrations vary in the ranges 6 – 9 and 8 – 9, respectively. The ratio Fe 2+ / (Fe 2+ + Mg) in atoms per formula unit varies between 0.24 and 0.69, with most compositions classified as biotite and few compositions classified as phlogopite according to the Mg:Fe = 2:1 division defined by Deer et al. (1966). Analcime has SiO 2 and Al 2 O 3 concentrations that range 52 – 55 and 23 – 24 respectively, with constant values for Na 2 O of ca. 12 wt. %. Apatite is mainly classified as fluorapatite, with F contents of 0.12 – 2.20 wt. % and Cl contents of 0.22 – 0.45 wt. %. CaO concentrations vary from 27 to 56 wt. % and P 2 O 5 concentrations vary from 22 to 42 wt. %. 5.6.2. Trace element composition LA-ICP-MS trace element analyses were undertaken on clinopyroxene and amphibole, given the complex compositional changes recorded in the crystals of these minerals. Macrocrysts and microphenocrysts were analysed; microlites were too small to provide reliable spot analyses. A lamprophyre from the eastern sector (Calella de Palafrugell) and a lamprophyre from the western sector (Monasterio de Sant Pere Cercada) were selected for analysis. In the case of clinopyroxene, A-, B-, Gand X-type macrocrysts cores were analysed, together with P-type compositions in macrocryst cores and microphenocryst cores. In the case of amphibole, analyses were obtained from Jand K-type inner cores in macrocrysts and P-type compositions in macrocryst cores and microphenocryst cores. Moreover, some amphibole macrocrysts in the Monasterio de Sant Pere Cercada lamprophyre have thick rims so that rim analyses could be carried out as well. The obtained dataset includes a total of 23 compositions and is presented in Table 5.X of the electronic supplement. Primitive mantle (McDonough and Sun, 1995) normalised trace element patterns are presented in Fig. 5.18. Clinopyroxene Rare Earth Element (REE) concentrations are up to ca. 60 times enriched over the primitive mantle (Fig. 5.18A). REE normalised patterns are convex-upwards, with the exception of the X-type green core, which presents convex-downwards patterns. Among the convex-upwards normalised patterns, B-type compositions are the least REEenriched, Pand certain A-type cores have an intermediate composition and Gand certain A-type cores are the most REE-enriched. Chapter 5 · The lamprophyre suite 112 Fig. 5.18. Trace element composition of clinopyroxene (A, B) and amphibole (C, D), normalised to primitive mantle (McDonough and Sun, 1995). Normalised values below 0.01 were screened out. Considering more trace elements (Fig. 5.18B), the X-type core again shows a normalised pattern different from the rest. All of the analyses however, display a negative anomaly for Sr and low concentrations in Rb and Ba, significantly below the primitive mantle in the case of X-, G-, Band some A-type cores. Finally, the P-composition obtained from a microphenocryst shows a distinct positive anomaly for U. Amphibole is up to ca. 100 times enriched over the primitive mantle for the REE (Fig. 5.18C). REE normalised patterns are convex-upward and fairly parallel to each other. They are similar to convex-upwards patterns in clinopyroxene (compare with Fig. 5.18A). J-type compositions are clearly the most REE-enriched, closely followed by rim compositions. Kand P-type compositions on the other hand, have lower concentrations in REE. P-type microphenocryst cores reach relatively enriched compositions. Multielemental normalised patterns for amphibole (Fig. 5.18D) are fairly parallel with slight enrichment differences between the different compositional types, similarly to the REE-normalised patterns. Moreover, Jand R-type compositions show a negative anomaly in Sr, whereas Kand P-type compositions show a positive anomaly for this element. All of the normalised patterns show a strong negative anomaly for the U-Th pair and high normalised values for Ba. The lamprophyre suite · Chapter 5 113 5.7. Whole rock chemistry All of the lamprophyres were analysed for whole rock compositions. The lamprophyres located inland (Llagostera and the western sector) are partly altered, especially at the chilled margins, so the analyses were carried out on the zone of accumulation of mafic macrocrysts (see Fig. 5.8). Thus, most samples have a porphyritic texture. Microlitic whole rock compositions include the lamprophyre from S’Agaró and the groundmass of the Calella de Palafrugell lamprophyre modelled in Ubide et al. (2012a –see Chapter 6-). The whole rock composition dataset is presented in Table 4.1 –see Chapter 4and includes 17 analyses. 5.7.1. Major element composition The lamprophyres are basic and ultrabasic rocks, silica undersaturated and rich in titanium and alkalis (Table 4.1). They are olivineand nepheline-normative. The alteration of the samples is low and restricted to the replacement of olivine macrocrysts and glass, which are present in minor proportions, with secondary mineral assemblages (Ubide et al., 2012a). Nevertheless, Loss on ignition (LOI) values are relatively high due to the abundance of hydrous minerals and vesicles. Whole rock magnesium number (Mg# = 100 MgO / (MgO + FeO) and FeO = 0.9 x total iron as Fe 2 O 3 ) shows a wide variation range, between 27 and 61. However, only some sam ples from the central and western secto rs have values over 40 and only the Molí d’en Ponç lamprophyre (western sector) has values over 50. MgO concentrations vary between 4 and 14 wt. %, with the highest values for the Molí d’en Ponç lamprophyre (Table 4.1). Fig. 5.19 shows whole rock major element compositions plotted in bivariate diagrams vs. decreasing MgO. The samples do not define a consistent variation trend. Overall, CaO concentrations increase and Al 2 O 3 , Na 2 O, K 2 O and P 2 O 5 concentrations decrease with increasing MgO. However, SiO 2 , TiO 2 and Fe 2 O 3 define different patterns depending on the sector considered. Samples from the eastern and central sectors become more TiO 2 - rich and SiO 2 -poor with increasing MgO, whereas samples from the western sector show the opposite trend for TiO 2 and an irregular variation for SiO 2 . Meanwhile, samples from the eastern sector and, especially, the Calella de Palafrugell lamprophyre show a significant increase in Fe 2 O 3 with increasing MgO, whereas samples from the central sector show little variation and samples from the western sector decrease in Fe 2 O 3 with increasing MgO. The distribution of the samples correlates with the petrology and mineralogy of the rocks. Samples with high volume fractions of macrocrysts have high concentrations in MgO, with Chapter 5 · The lamprophyre suite 114 Fig. 5.19. Whole rock major element compositions in bivariate diagrams vs. decreasing MgO. White symbols represent macrocryst-free compositions: Gr denotes the groundmass composition of the lamprophyre from Calella de Palafrugell (Ubide et al., 2012a –see Chapter 6-) and Mi denotes the composition of microlitic lamprophyre from S’Agaró. The vertical dashed line may represent an average composition for the groundmass of the whole lamprophyre suite. The lamprophyre suite · Chapter 5 115 Fig. 5.20. Whole rock trace element compositions in bivariate diagrams vs. decreasing MgO. Explanation for the symbols is as in Fig. 5.19. The microlitic lamprophyre from S’Agaró (sample SAG-03B) has a Ni concentration below the detection limit, so it has been plotted at 0 ppm for that element. The vertical dashed line may represent an average composition for the groundmass of the whole lamprophyre suite. Chapter 5 · The lamprophyre suite 116 the highest values for the samples with an olivine-rich macrocryst assemblage (compare Fig. 5.8 and Fig. 5.19). In contrast, microlitic compositions plot together with samples with low proportions of macrocrysts at low MgO values. Microlitic compositions share an MgO concentration of ca. 6 wt. % and an Mg# value of ca. 35 (Fig. 5.19; Table 4.1 –Chapter 4-). 5.7.2. Trace element composition Representative trace elements are plotted in bivariate diagrams against decreasing MgO (Fig. 5.20). Most incompatible trace elements, including Large Ion Lithophile Eleme nt s (LILE) and High Field Strength Elements (HFSE) show an overall decrease for increasing MgO contents. Transition elements show distinct variations depending on the sector considered. Lamprophyres from the eastern and central sectors show a sharp increase in V and a smooth increase in Ni and Cr when MgO increases. In contrast, lamprophyres from the western sector show a sharp increase in Ni and Cr and irregular variations for V when MgO increases. REE and multielemental normalised patterns are presented in Fig. 5.21. REE concentrations are enriched ca. 10 to more than 100 times over the chondrite composition by McDonough and Sun (1995). Incompatible elements are enriched up to ca. 100 times over the primitive mantle composition by McDonough and Sun (1995). All of the lamprophyres show normalised patterns similar in shape, suggesting a common magma source. REE normalised patterns have a high slope, with (La/Lu) N values between 11 and 19. Multielemental normalised patterns show positive anomalies for Nb and Ta and minor ones for Ba; some of the samples present a negative anomaly in Pb. These data are typical of alkaline rocks with an asthenospheric, enriched mantle source similar to EM-1 and/or HIMU (Hofmann, 1997; see Chapter 4). Strikingly, the microlitic or macrocryst-poor samples are very similar to each other in the three sectors, whereas the macrocryst-rich samples are less enriched in incompatible elements (Fig. 5.21). In the central sector, the macrocryst-rich sample (Llagostera lamprophyre) has a positive anomaly for Ti. This anomaly is also present, although smaller, for the porphyritic samples from the eastern sector. The lamprophyre suite · Chapter 5 123 5.8.2. Origin of whole rock compositional variations The Cretaceous lamprophyre suite in the Catalonian Coastal Ranges has a common magma source (Ubide et al., 2012b; see also Chapter 4) but shows whole rock compositional variations from 4 to 14 wt. % MgO (Table 4.1 –see Chapter 4-). Samples with MgO concentrations up to ca. 7 wt. % have overall similar compositions (Fig. 5.19, 5.20), whereas samples with higher MgO contents show distinct variation trends with the increase in MgO. In the eastern and central sectors the increase in MgO is coupled with an increase in TiO 2 and V, a minor increase in Ni and Cr and a decrease in SiO 2 . In the western sector the increase in MgO is stronger and coupled with a dramatic increase in Ni and Cr, a decrease in TiO 2 and an irregular variation for SiO 2 and V. Fe 2 O 3 shows a different evolution with the increase of MgO; it increases in the eastern sector, especially for the Calella de Palafrugell lamprophyre, remains fairly constant in the central sector and decreases in the western sector. Regarding the other elements, the increase in MgO is coupled with an increase in CaO and a decrease in Al 2 O 3 , Na 2 O, K 2 O, P 2 O 5 and incompatible trace elements (Fig. 5.19, 5.20, 5.21). Whole rock variation trends can hardly respond to magma fractionation given the petrologic and mineralogic resemblance between the samples, especially regarding their groundmass. Furthermore, clinopyroxene and amphibole compositions cogenetic with the groundmass (P and R) are common to all the lamprophyres and have a similar composition in all of them. Even the inherited antecrysts are compositionally similar, when present, among different lamprophyres. Rather than magma fractionation, the significant whole rock compositional variations appear to correlate with the proportion and mineralogy of macrocrysts in the rock (see Fig. 5.8). The influence of macrocrysts on the bulk composition of porphyritic rocks has been recently highlighted in lamprophyres (Scarrow et al., 2011; Ubide et al., 2012a –see Chapter 6-) and basaltic rocks from ocean island settings like Hawaii (Sakyi et al., 2012) or Azores (Larrea et al., 2013). The microlitic lamprophyre from S’Agaró and the groundmass composition from Calella de Palafrugell have very similar MgO concentrations of ca. 6 wt. %. These macrocryst-free samples, together with macrocryst-poor samples from Aiguablava and the western sector, make up the relatively homogeneous group of compositions of MgO < 7 wt. % (Fig. 5.19, 5.20), with highest concentrations in incompatible elements (Fig. 5.21). In contrast, macrocryst-rich samples have MgO > 7 wt. % and lowest concentrations in incompatible elements (Fig. 5.19, 5.20, 5.21). The volume fraction of macrocrysts is highest in the accumulation zone of the Molí d’en Ponç and Llagostera lamprophyres (Fig. 5.8) and this correlates with high MgO concentrations. In the western sector the macrocryst assemblage is olivine-rich and this accounts for the highest MgO, Ni and Cr enrichments. Amphibole is more common in the macrocryst assemblage of the eastern sector and most frequent in the Llagostera lamprophyre and that explains the TiO 2 and V enrichments and Chapter 5 · The lamprophyre suite 124 the relative SiO 2 depletions for these samples. Finally, opaque minerals are a minor macrocryst phase but are especially frequent in the Calella de Palafrugell lamprophyre, correlating with the Fe 2 O 3 -enrichment shown by the macrocryst-rich samples from this sill. Clinopyroxene and amphibole show normalised incompatible element patterns less enriched than whole rock samples (compare Fig. 5.18 with Fig. 5.21), so that increasing volume fractions of macrocrysts in the sample may result in decreasing whole rock trace element concentrations. The relationship between whole rock variations and macrocryst accumulation was quantitatively modelled for the Calella de Palafrugell lamprophyre in Ubide et al. (2012a) and Chapter 6 in this volume. In the present study, whole rock compositions have been plotted together with macrocryst core compositions in representative major element diagrams (Fig. 5.23). Fig. 5.23. Major element composition of whole rocks and representative macrocryst core types in bivariate diagrams vs. decreasing MgO. To make iron concentrations comparable, FeO T contents in whole rock are calculated as 0.9 x Fe 2 O 3T . Whole rock compositions are presented keeping the symbols from previous figures and mineral compositions are presented as compositional fields (or arrows when outside the diagram). The lamprophyre suite · Chapter 5 125 From the plots in Fig. 5.23, it is clear that clinopyroxene macrocrysts, which appear in high proportions in all the porphyritic lamprophyres (Fig. 5.8; Table 5.1), control most of the whole rock compositional variations. Amphibole is the main macrocryst phase in the Llagostera lamprophyre and it is also relevant in the Calella de Palafrugell lamprophyre and triggers TiO 2 -enrichments and SiO 2 -depletions in macrocryst-rich samples from these intrusions. Amphibole is also the main macrocryst phase in the Aiguablava lamprophyre and the most MgO-rich sample from Aiguablava varies along with the Llagostera and Calella de Palafrugell lamprophyres. This sample is nevertheless relatively poor in macrocrysts (Fig. 5.8) and in consequence, the macrocryst accumulation effect is minor. Olivine is especially frequent in the macrocryst-rich samples from the western sector and controls the MgO-enrichment in these samples. Opaque minerals have extreme compositions as well, but they are always a minor macrocryst phase excepting the Calella de Palafrugell lamprophyre, explaining the highest concentrations in iron of these samples. In conclusion, the significant variations in whole rock composition among the samples from the Catalonian Coastal Ranges lamprophyres are controlled by the proportion and mineralogy of accumulated macrocrysts. The microlitic groundmass is equivalent for the whole lamprophyre suite and has MgO concentrations of ca. 6 wt. % (Fig. 5.19, 5.20) and Mg# values of ca. 35 (Fig. 5.22B, 5.22D). It represents the magma that hosted the recycled crystal populations and carried them up to the shallow emplacement levels. The variable entrainment of antecrysts in the different intrusions is the driving force that pushes whole rock geochemical variations away from groundmass compositions. The magma represented by the groundmass of the lamprophyre suite must have undergone some fractionation prior to emplacement, either in a magma chamber or en route to shallower depths, as its composition cannot represent a primary magma in equilibrium with its mantle source, that would typically show Mg# > 60 – 70, MgO > 11 wt. %, Cr > 500–1000 ppm and Ni > 200–500 ppm (Frey et al., 1978). Again, the few samples with these characteristics are strongly influenced by the accumulation of macrocrysts, so they must not be considered for the interpretation of primary petrogenetic processes. The groundmass composition therefore suggests fractionation processes taking place at depth. Mineral compositions actually call for fractionation, together with mafic recharge, to explain the origin of deep, recycled antecrysts (see sections 5.8.1 and 5.8.3). At shallower depths, slight fractionation probably affected the groundmass magma triggering minor compositional differences between the samples with MgO < 7 wt. %, as well as the development of evolved patches of groundmass in the Llagostera lamprophyre. Chapter 5 · The lamprophyre suite 126 5.8.3. Reconstruction of the magmatic plumbing system The preceding discussion has shown that the Cretaceous lamprophyre suite in the Catalonian Coastal Ranges basically represents a partly fractionated melt carrying a mush of large, recycled crystals from deep crystallisation areas. 40 Ar/ 39 Ar geochronology applied to the Calella de Palafrugell, Llagostera and Monasterio de Sant Pere Cercada lamprophyres constrained the time of emplacement to ca. 79 Ma (Ubide et al., under review a –see Chapter 3-). This age is Campanian and agrees with the end of the Mesozoic extension in the Late Cretaceous (see Chapter 3), when the continental crust in this area was thinned to ca. 30 km (Salas and Casas, 1993). Petrologic and geochemical data obtained in the present study make it possible to hypothesise the nature of the magma feeding system related to the lamprophyre suite, as illustrated in Fig. 5.24. Fig. 5.24. Conceptual sketch of the magmatic plumbing system and magmatic processes involved in the generation of the Cretaceous lamprophyre suite in the Catalonian Coastal Ranges. Octagons represent clinopyroxene crystals in equilibrium with the melt and hexagons represent amphibole crystals in equilibrium with the melt. Rounded shapes, in contrast, represent crystals in disequilibrium with the melt. The green resorpted core represents the clinopyroxene xenocryst. Magma history in the eastern sector is further explored in Chapter 7 (see Fig. 7.12). Because all the lamprophyres share a common groundmass that hosts different recycled crystals per sector, a large magma reservoir located below the level of crystallisation of recycled crystals might be inferred. Barometric calculations reveal that A-type clinopyroxene antecrysts and the X-type clinopyroxene xenocryst from the Calella de The lamprophyre suite · Chapter 5 127 Palafrugell lamprophyre crystallised in the lower part of the crust, close to the crustmantle boundary (Ubide et al., under review b –see Chapter 7-). It follows that the reservoir of groundmass magma was located at a greater depth, maybe at the upper mantle-lower crust interface that constitutes a rheological discontinuity favourable for magma stagnation (e.g., Annen, 2011). The reservoir may have supplied magma to the whole lamprophyre suite and according to whole rock multielemental normalised patterns, it was fed by astenosphere-derived melts (see also Chapter 4). The lamprophyres from the central sector are the simplest regarding mineral compositional types. In the Llagostera lamprophyre, all clinopyroxene macrocrysts have B-type cores, all amphibole macrocrysts have P-type cores and there are no olivine macrocrysts. The B1-B2-B3-R crystal stratigraphy in clinopyroxene macrocrysts can be associated with progressive evolution of the magma from more primitive to more evolved compositions. The irregular contours of B1-type inner cores, together with the presence of a diffused B2 zone between B1 and B3, suggest that B1-inner cores could have been carried with the magma filling the reservoir from lower levels. The central sector likely represents magma directly coming from the large groundmass reservoir, where macrocrysts had room and time enough to reach very large sizes. This magma may have been especially water-rich according to the abundance of amphibole macrocrysts and the lack of olivine. The S’Agaró lamprophyre is the only sub-vertical and macrocryst-free intrusion and may represent a subvertical apophysis of the Llagostera or other related sill, unreached by the deep, large and relatively heavy crystals considering the high density of clinopyroxene and amphibole. Alternatively, it could represent a shallow section of a feeder conduit with macrocrysts sinking to lower levels before solidification of the groundmass magma. The lamprophyres from the western and eastern sectors have olivine macrocrysts as well as clinopyroxene and amphibole macrocrysts with more complex zoning patterns that include P-type zones. Moreover, amphibole macrocrysts are smaller. These differences relative to the lamprophyres from the central sector might indicate a possible recharge of the groundmass reservoir with a more primitive and water-poor magma, as detailed in the following paragraphs. Magma recharge would also explain the slight differences in mineral composition between the microlites of amphibole and plagioclase in the central sector and those in the western and eastern sectors (see section 5.6.1). In the western sector, there are cm-sized clinopyroxene macrocrysts with B-type cores, similarly to the central sector. However, in the western sector B-type cores are surrounded by C compositions, which represent a step backwards in the evolutionary trend, back to more primitive compositions, and could be related to recharge of the groundmass reservoir with a more primitive magma. Given the high MgO and Cr 2 O 3 concentrations in B1, B2 and C compositions, the recharging magma might be relatively Chapter 5 · The lamprophyre suite 128 similar to the original magma entering the magma reservoir and mixed with the magma in equilibrium with B3 clinopyroxene to give rise to an intermediate C composition, similar to B2. C zones overgrow B3 and are overgrown by P, resuming the normal evolution towards more evolved compositions. Moreover, olivine macrocrysts were formed first in the crystallisation sequence, as they are in equilibrium with a liquid with Mg# similar to that producing B1 clinopyroxene and their compositions do not fit a mantle origin. In addition, the western sector lamprophyres have G-type clinopyroxene macrocryst cores and J-type amphibole macrocryst cores, which crystallised from magmas with similar Mg# contents, lower than the other antecrysts. This suggests the existence of another, more evolved magma chamber located above the groundmass reservoir. Gand J-type crystals were sampled by the groundmass magma in equilibrium with P-type clinopyroxene and amphibole ascending towards the emplacement level from the deep reservoir. In the eastern sector there are no B-type clinopyroxene cores, indicating that the magma did not carry clinopyroxene crystals from the deep reservoir. The groundmass magma is however in equilibrium with P-type clinopyroxene and amphibole, indicating that it left the reservoir after the mafic recharge event. There are abundant macrocryst cores with A-clinopyroxene and K-amphibole compositions, both exclusive to the eastern sector lamprophyres. This suggests the presence of another evolved magma chamber located above the deep groundmass reservoir. In this case, the A1-A2-A3-P evolution in clinopyroxene macrocryst cores reveals successive inputs of magma from the deep reservoir recharging the evolved magma chamber and mixing with the resident melts. The magma in this chamber was initially more evolved and water-poor and contained a xenocryst (X-clinopyroxene) dragged from lower levels. Progressive recharge with a more primitive and water-rich magma followed by mixing with the magma resident in the crystallising chamber controlled the linear evolution of A-type clinopyroxene towards P compositions. A3-clinopyroxene crystallised coevally with K-amphibole once the water content of the mixture magma was high enough, and this was followed by the crystallisation of P-clinopyroxene and P-amphibole related to the last recharge event, which likely triggered the ascent of the groundmass magma (see also Ubide et al, under review b –Chapter 7-). In all the lamprophyres, microphenocrysts and microlites crystallised in response to progressive fractionation of the groundmass magma. Microphenocryst cores may have crystallised in polybaric conditions during the ascent of the magmas, whereas rims and microlites probably crystallised at the emplacement level. The emplacement level was probably close to the surface in all the cases, as revealed by the widespread presence of vesicles, the subhorizontal dip of most intrusions implying that the magma had to support the weight of the overlying rocks and the barometric calculations in Calella de Palafrugell suggesting magma emplacement at ca. 1 km depth. (Ubide et al., under review b –see Chapter 7-). The lamprophyre suite · Chapter 5 129 5.9. Conclusions The in-depth petrologic, mineralogic and geochemical study of the Cretaceous lamprophyre suite in the Catalonian Coastal Ranges has led to the following conclusions: • Most of the studied lamprophyres are apparently simple porphyritic rocks and yet, they are composed of highly complex crystal populations that include phenocrysts, antecrysts and xenocrysts revealing deep magmatic processes. Antecrysts crystallised from earlier accumulated magmas and are predominant in the macrocryst assemblage of the lamprophyres. The obtained results imply that large crystals in porphyritic rocks must not be considered phenocrysts cognate with the host groundmass before a detailed petrographic and mineralogic study is carried out. Such assumption may otherwise lead to incorrect petrogenetic interpretations. Moreover, a deep knowledge of crystal stratigraphy and composition unveils magmatic processes that would remain misunderstood if whole rock compositions were considered alone. • The whole rock composition of the porphyritic lamprophyres is strongly influenced by the proportion and type of macrocrysts in the sample. All of the samples have very similar groundmass compositions and yet different whole rock compositions due to the variable accumulation of macrocrysts. Importantly for petrogenetic interpretations, it follows that MgO-rich samples are not necessarily indicative of primitive compositions close to the mantle source. Rather, MgO-rich samples may arise from relatively MgO-poor melts with high volume fractions of MgO-rich antecrysts. Again, the obtained results show that a deep understanding of the samples must precede petrogenetic interpretations. • The Catalonian Coastal Ranges Cretaceous lamprophyres are shallow exposures of a common, intricate magmatic system. Astenosphere-derived melts probably stagnated at the upper mantle – lower crust interface constituting a large magma reservoir where the largest crystals formed and from which lamprophyre magmas were extracted. Crystals from this reservoir record recharge by a new batch of primitive melt. Some of these or related magmas stagnated in the lower crust and underwent fractionation giving rise to more evolved melts. The evolved magma chambers could receive repeated influxes of primitive magma from the deeper reservoir, and the recharging magma repeatedly mixed with the magma resident in the evolved chambers. Magma upwelling from the reservoir could entrain already formed large crystals and, on its way to the shallow emplacement levels, could incorporate crystals from the evolved chambers. The result of this system is a lamprophyre family composed of a great variety of inherited large crystals in a common groundmass. Chapter 5 · The lamprophyre suite 130 131 6 The influence of crystal settling on the compositional zoning of a thin lamprophyre sill: a multi-method approach. SUMMARY The Cretaceous lamprophyre from Calella de Palafrugell (Catalonian Coastal Ranges) is a subhorizontal sill with well developed chilled margins, vesicle alignments and abundant mafic macrocrysts (antecrysts) accumulated towards the bottom. These characteristics, together with a small thickness (< 0.5 m), make this intrusion an exceptional laboratory for the study of compositional zoning of igneous intrusions. Classical petrology and geochemistry are combined with mathematical modelling and multivariate statistics to quantify the influence of antecrysts on whole rock compositions. This leads to unravel the origin of the compositional zoning of the sill, as being due to gravitational settling of the antecrysts after emplacement of a single pulse of magma. View from ses Negres close to Calella de Palafrugell, Costa Brava Chapter 6 · Antecryst accumulation 132