5. Work methods

To advance bed correlation and for the detection of distinct markers for the sampling of homogeneous fine - coarse-grained sandstones of the Lauzanier sandstone body, a geological mapping was carried out to provide an overview of the distribution of the Grès d'Annot sandstone body in the area (Appendix 1: Geological Map). Furthermore, four panorama pictures (Appendix 3, Panoramic Pictures: 3.1, 3.2, 3.3 and 3.4) were generated and five logging lines were recorded (Appendix 2: Log Correlation).

For provenance reasons, other localities besides the Lauzanier area were included. These were the area of Annot and Peira Cava. The analysed samples from those areas served as a comparison between the sandstone contents.

The maps used for the geological mapping of the Grès d'Annot Formation in the Lac du Lauzanier area are:

• the topographic map of Michelin, 1: 200,000, 245 France, Provence - Côte d'Azur
• the topographic map of IGN (Institut Gèographique National), 1: 25,000; 3639 ouest, Larche
• the geological map of BRGM (Bureau de Recherches Gèologiques et Minières), 1: 50,000, 896 Larche

The topographic IGN map was used as a base map for the processing of the geological map of the Grès d'Annot Formation in the Lac du Lauzanier area. Originally the base map displayed a 1: 25,000 scale which was enlarged by copying to a 1: 10,000 scale. However, the base map had to be redrawn due to shading of the flanks in the original topographic map, which caused colour discrepancies. The redrawn base map was then scanned with black & white - 200 dpi and imported in to the graphic programme of Corel Draw 7.0 and processed to a geological map (Appendix 1: Geological map). The generated geological map is not displaying the surrounding rock, but only the exposures of the Grès d'Annot Formation and its eroded products of unconsolidated rock, because only the Grès d'Annot Formation was of interest for this work. The unconsolidated rocks were divided into either slope scree or glacial scree (Appendix 1: Geological Map). The surrounding rock of the Grès d'Annot in the south is the prior, underlying, Eocene Schistes à Globigèrines (belonging stratigraphically to the Marnes Bleues, figure 4.2c: Stratigraphy chart). These are overlying the Eocene (Priabonien) Calcaires Nummulitique, while in the north, the Grès d'Annot Formation is bordered and unconformably underlying the autochthon Sub-Brianconaisse Nappe (Figure 3.2: Locality map). This thrusted nappe comprises olisthostromes, belonging to the Formation of the Schistes à Blocs (Figure 4.3c: Startigraphy chart). To ease the geological mapping the BRGM geological map of Larche served as guidance. Additionally the sampled and correlated beds were also mapped and displayed in the geological map, whereas an individual bed is drawn with the same colour (Appendix 1: Geological Map, or appendix 2: Log Correlation). An altimeter benefited the positioning in the Lauzanier area due to the lack of fix points in the topographic map. The strike and dip data of the bedding, the linear data of the transverse faults and the strike and dip of the three directional joint system were obtained with a geologist compass. Yet an extra structural map of the Lac du Lauzanier area was not generated, because the geological map is still visually clear even when the structural map is integrated.

The pictures gave an overview of the beds, advancing the bed correlation between logs and for identifying extensional- and well exposed markers in the Grès d'Annot outcrops. They also served as a pattern for the sedimentological and tectonic interpretation (Appendix 3, Panoramic Pictures: 3.1, 3.2, 3.3 and 3.4).

The four panorama pictures were taken with a Pentax ESPIO 115M Zoom (with an average of about 50mm lens) from three flanks of Lac du Lauzanier area. One was taken from the west flank- (Appendix 3.1, West Flank of Lac du Lauzanier), two of the east flank (Appendix 3.2, Northern East Flank of Lac du Lauzanier and Appendix 3.3, East Flank of Lac du Lauzanier) of Lac du Lauzanier and one of the east flank (Appendix 3.4, East Flank of L'Enchastraye Valley - French/Italian border) of the French/Italian border. It was necessary to develop two panorama pictures of the east flank of Lac du Lauzanier, because the flank is a bulging body, which led to an approach in covering the flank with only one panorama picture impossible in respect to the picture fitting. The developed pictures were scanned with RGB colours - 720 dpi, then imported to the CorelPhoto-Paint 7 programme, all pictures were cut at their ends to avoid scale providing a better picture fit. Further the pictures were manipulated in dark- and brightness in the same programme mentioned above, and additionally enlarged. All panorama pictures were finally imported to CorelDraw 7.0 and assembled on a large-format sheet (Appendix 3, Panoramic Pictures: 3.1, 3.2, 3.3 and 3.4).

Five field logs were recorded and generated of the Grès d'Annot sandstone body in the Lac du Lauzanier area for the following reasons (Appendix 2: Log Correlation): advancing an effort in correlating beds, to resolve high resolution sandstone/heterolithics ratio for an approach in predicting relatively deeper areas on local basis, to gain a record of the complete Grès d'Annot succession, to obtain a record of the sedimentological and fossil features of the succession to provide an overview of the sediment paleo-current trend and finally, for interpreting the depositional environment of the Grès d'Annot Formation in the Lauzanier area.

The primary intention of the positioning of field log lines should first of all engage a distant as possible correlation between beds, in respect to sampling purposes which are demanding extensions of individual beds on a kilometre scale (Appendix 1: Geological Map). However, the Grès d'Annot sandstone body and its beds are dipping with an average of 35°, allowing only a maximum of 2.9km extensive, northwest - southeast directed, individual bed to be exposed in the outcrops (Appendix 1: Geological Map, valid for bed 1 and 2). Three logs were recorded and generated from incomplete successions of the east-flank of Lac du Lauzanier. Only one log was processed from the incomplete succession of the west-flank, because correlation of the cycles near the top of the succession was difficult and therefore not sampled. The field log from the French-Italian border is in contrary nearly covering the complete Lauzanier succession. The field logs were recorded from base to top. Starting and ending point of a log line is illustrated as log flags in the Geological Map (Appendix 1). The displayed log lines in the geological map are showing the logging positions. They were chosen as perpendicular to the bedding plane as possible, but sometimes outcrop situation was poor and did not expose solid rock. Thus, lateral displacement was unavoidable for further logging continuation.

Several implementations were used for obtaining field data. Hydrochloric acid was utilised in testing calcareous content, which seldom was caused by low calcium carbonate contents. The sorting and texture of the bedrock on a microscopic scale was defined with help of a lens. A folder rule served for the measuring of the bed thickness, while the positioning was controlled through an altimeter combined with the topographic map. Furthermore the altimeter became a function as of a control instrument of the true field logged altitude metres. The field data was recorded in a field book, then evaluated and hand drawn into a 1: 200 log pattern. These patterns were scanned with 200 dpi and imported to the software programme of CorelPhoto-Paint 7.0, in which they were cut and aligned for tie purposes. Then exported to the graphic software programme of Corel Draw 7. The scans were digitalised and generated (Appendix 2: Log Correlation), whereas the former scale was decreased to 1: 400 scale.

An approach in correlating beds was primarily made by pre-selecting distinct markers from the low resolution panorama pictures which served as a favourable orientation for the high resolution log correlation (Appendix 2: Log Correlation). The correlation was furthermore advanced through the adding of bed contents and sedimentological features like: texture, sediment structure, colouring of bedrock and fossil activity (Appendix 2: Log Correlation). Sorting index up to 0.2cm of the beds were determined after the Wentworth scale (Wentworth, 1922). However, if a bed showed grain sizes more than 0.2cm, they received the terminology of a pebble. These pebbles vary from a range of 0.2 to 8cm in the Lauzanier area. The grain sizes were generated in the logs with an individual column (Appendix 2: Log Correlation), which were divided into clay, silt, very fine-, fine-, medium-, coarse-, very coarse sand and pebbles. A sinful method for illustrating the log motive, which mostly is showing an upward fining, except when sub-amalgamation - amalgamation, or inverse grading, occur. The grain sizes are divided into three colour codes; the green heterolithics, range from clay to very fine sand, the yellow sand, range from fine sand to very coarse sand and the orange conglomerates (Appendix 2: Log Correlation). Additionally the column is illustrating erosional- and depositional channels, which cut into substrata. Further, the column shows sediment structures like, ripple current lamination, load casts and undulating base or top of beds. The following column, the lithology column, display the sorting of the rock (see legend in appendix 2, Log Correlation, for further details) and the heterolithics, sands and conglomerates by colouring code, equal to grain size column. The next column, the description column, is describing; texture, sediment structure, colouring of bedrock and fossil activity (Appendix 2: Log Correlation). A comment column was added for comments regarding; no exposure, amalgamation, channel position in bed and its width and fill, paleo-current direction and other comments (Appendix 2: Log Correlation). The final column, sample column, is listing the sampling. Red colour is marking analysed samples, black colour indicates taken but not analysed samples (Appendix 2: Log Correlation). The numbering is dividable in two main sample groups, one vertical connected- (with "V") and one (without "V") lateral correlated group. "P" and the following number associates the sample to the log number ending with a dot, whereas the split numbers indicate from between which logs the sample is positioned. The next letter is related to in which direction the sample is taken from the log. Those without letters are taken from within the log. The final number is illustrating from which bed the sample belongs to, though this is only valid for the lateral (without "V") samples. The end number of the vertical (with "V") samples, which only are taken within logs, means from were the sample is relatively located from the base beginning number. The number in brackets, after the sample numbering, is a short way, which is displayed, in the geological map for the positioning of the samples (Appendix 1: Geological Map). To comprehend a better overview of the Log Correlation referring to the Geological Map and the opposite, they were gathered on a single large-format sheet. However the Geological Map is still kept as an own appendix due to the largeness and unwieldiness of the Log Correlation sheet.

An approach in predicting former relatively deeper basin areas on local Lac du Lauzanier area basis from the high resolution Log Correlation sheet, was possible with calculation of the sandstone/heterolithics ratio. The ratio was determined through applying an equal interval on three different logs. The compared logs are; log 4, the west-flank-, log 2 and log 3, the east-flank of Lac du Lauzanier and the log 5 from the French-Italian border (Appendix 2: Log Correlation).

The geochemical analysing approach was dependent on lateral sampling in a single bed on a kilometre-scale, because of a fundamental consideration; a turbidite will rather show differential mineral composition over far lateral- contrary to short distances. Therefore, sampling was first of all a concern of finding the most extensive beds in the field (Appendix 1: Geological Map, valid for bed 1 and 2). They should additionally represent an easily recognisable marker, comprise a homogeneous fine- to coarse-grained sandstone and show bright colours. These characteristics qualify sandstone to be more extensive than sandstones with other characteristics. Important during sampling of an individual bed was only to take samples of a distinct grain size, providing increasing representativeness, for particularly avoiding compositional changes that are controlled by grain-sizes. The samples from Annot and Peira Cava were also included in the sampling. Because they were important for provenance efforts, since analysis of the mineral composition of different sandstones from different localities could either support the theory of, same sediment composition concludes a joint source area, or alternatively, different composition consequently different sources areas. 106 samples were taken (Figure 5.1.4), though only 48 samples were analysed. The analysed samples are closest to the criteria mentioned above.

Beside a hammer, a chisel become helpful and enabled the breaking of solid rock for the sampling in the Lac du Lauzanier area, due to the rock hardness which shows significant fabric maturity. On the contrary, the samples from Annot and Peira Cava are less consolidated and break easily. Principally only fresh rock was taken about hand piece big. Weathering parts of rock were crushed away, because they could mislead the results of geochemical analyses. Samples were orientated to the northern- and upside direction and numbered. To hinder oxidisation the samples were immediately put in an air tight sample bag and numbered again.

A correlation between the sampled beds and other markers between Log 1 - 4 were successfully accomplished (Appendix 2: Log Correlation). Unfortunately the sampled beds were not possible to correlate on firm basis between log 5 and Log 1 - 4. This led to the only possible consequence of excluding the samples from log 5 of the geochemical analysing (Figure 5.1.4), because the analysing demands secure sampling since lateral samples must strictly be collected from distinct beds. Despite the insecure correlation between log 5 and the other logs, a correlation was interpreted and displayed (Appendix 2: Log Correlation). Vertical samples also had to be excluded since they need lateral samples as a reference of geochemical analysing results. This excluding the log 5 samples resulted in a 2.5km decrease of the east-west extending sampling, thus the true distance of sampling now is reduced to 3.5km in east - west direction (Appendix 1: Geological Map). Despite this shortage, the disadvantage was not severe, since the paleo-current mainly trended northwest- to north-northwestwards, it seemed obviously that the north - south expanding sampling was more important. A turbidite will rather comprise different mineral composition parallel to the current rather than in perpendicular directions. The outcrops of the Grès d'Annot Formation in the Lac du Lauzanier area enabled sampling, nearly parallel to the main paleo-current, over a distance of 2.9km (Appendix 1: Geological Map). Approximately the lateral samples were taken with 600 - 700m intervals (Appendix 1: Geological Map), whereas the vertical sampling was taken more sporadically. However, an effort was applied to keeping the vertical sampling connected to the sampling criteria, similar to sandstone characteristics (mentioned in chapter 5.1.3. Work methods for field logging).

The ultimate aim of this thesis was to find an additional correlation tool, particularly for supporting the correlation of beds in barren sedimentary sequences To obtain such a geochemical correlation tool, it was necessary to identify two or more geochemical indicators, which reveal lateral and vertical dependence in deep marine systems. A further effort was made through an attempt in understanding the factors that are controlling the distribution of eventual dependent indicators. The approach in qualifying and quantifying the rock components, which includes major- and heavy-minerals and major- and trace elements, were determined through optical- and geochemical instrument analysing. The implementation for the optical- and geochemical instrument analysing are: thin sections, heavy-mineral concentrate-slides (HM), x-ray diffraction (XRD), x-ray fluorescence (XRF), Electron Probe Microanalysis (EPMA) and induction-coupled-plasma mass-spectometry (ICP-MS).

48 of the 106 samples are used for analysis (Figure 5.1.4), whereas all 48 samples adopt the same analysing procedures. Though, for the EPMA a minor amount of samples were required.

The primary intention for applying thin sections was to attain petrographic characteristics of the Grès d'Annot sandstones, such as mineral determination, texture and structure. Additionally they were used to control the geochemical instrument analysing data and to contribute to the not identified minerals by the XRD-application.

The 48 hand piece-big samples were sawn to approximately 3.5cm width- and 4cm long- and 1 cm thick blocks. They were finished at the Hälbich Petrographisches Labor in Rollshausen, near Salzgitter. Only three thin sections, one from Lac du Lauzanier area, Annot area and Peira Cava area, show blue porosity colouring. The low amount is chosen because the samples from Lac du Lauzanier area are homogenous throughout with very poor porosity values (normally below 5%), and since the porosity colouring disturbs the carbonate colouring. The half side of the residually 45 thin sections received a differentiated carbonate colouring: calcite (red); ankerite (bluish-green); Fe-calcite (bluish-violet) and dolomite (colourless). Eight thin sections were high quality polished, and not covered, to enable microprobe utilisation. The thin sections were not used for determining the quantitative mineral content, since the risk of high failure quotes by point counting minerals. However, the thin sections served as a control implementation for the evaluation of the XRD values.

As guidance for determining minerals the books of Tröger (Optische Bestimmung der gesteinsbildenden Minerale, Teil 1, 1952), MacKenzie & Guilford (Atlas of rock-forming minerals in thin section, 1980) and Pichler & Schmitt-Riegraf (Gesteinsbildende Minerale im Dünnschliff, 1987) were used.

The appliance of heavy-mineral analysing was part of the ultimate aim of this study, finding two or more dependent geochemical indicators in deep marine siliciclastic environment. Furthermore their analysing results were significant for the provenance effort. Quantitative content of the heavy-minerals from the three sampled localities was compared, implicating either diverse sources or same source. It also, occasionally, benefited the replacement of the x-ray fluorescence data, providing securer data for the provenance study.

The heavy-mineral preparation and analysing were accomplished by supervision of A. C. Morton at the British Geological Survey in Keyworth/Nottingham. Additional preparations were performed at the Department of Petroleum Geology/Technical University of Clausthal. To benefit heavy-mineral determinations the atlas of Mange & Maurer (Schwerminerale in Farbe, 1991), Boenigk (1983. Schwermineralienanalyse) and the book of Tröger (Optische Bestimmung der gesteinsbildenden Minerale, Teil 2, 1967) were applied.

The 48, mostly, well-cemented hand piece big samples (Figure 5.1.4) were crushed to gravel sized rock fragments by a stone crusher machine. No chemical treatment was involved, since it is introducing a potential problem: acids can and do modify heavy-mineral suites, apatite being particularly susceptible (Morton, 1984). Then the crushed remnants were sieved to a 63-125µm fraction, at a grain size range where heavy-minerals are very well populated (Morton, pers. comm.), and cleaned by ultra sonic to avoid mineral aggregate residuals. Approximately 15g of the fragments were segregated to light- and heavy-mineral fraction through density segregation in bromoform. The heavy-mineral fractions were embedded in slides with araldite or canada balsam.

Processes operating during transport, deposition and diagenesis may severely influence the composition of heavy-mineral assemblages in sandstones. As a consequence, conventional heavy-mineral data may not be a reliable guide to the nature of sediment source material. Certain features of heavy-mineral suites, however, are inherited directly from the source area without significant modification, such as the varietal characteristics of individual mineral species (Morton, 1994). An alternative approach to varietal studies concentrates on relative abundances of minerals that are characterised by, comparatively immune to alteration during the sedimentary cycle, similar density and hydraulic behaviour, comparable habit and stable during diagenesis (Morton, 1994). The heavy-minerals complying these characteristics in this study are: zircon, rutile, monazite, Cr-spinel, garnet, apatite, tourmaline, whereas, also recognised but excluded were: anatase, epidote, hematite, titanite, mica, chlorite, brookite, staurolite, opaques. The immobile heavy-minerals were further divided into two groups, depending on similar density and habit, which results similar hydraulic behaviour. Representatives for the first group are: zircon, garnet, monazite, Cr-spinel and for the second group are: tourmaline and apatite (Morton 1994). Both groups were counted to an amount of 200 or more each. Hereby, a group of apatite/tourmaline (Ati-ratio was calculated as: 100 x number of apatite grains/total number of apatite plus tourmaline grains) and a group of garnet/zircon (GZi), rutile/zircon (RuZi), monazite/zircon (MZi), Cr-spinel/zircon (CZi) illustrate mineral indexes. This approach avoids some of the practical problems associated with varietal studies, such as the need to make subjective decisions about mineral properties or to use advanced analytical techniques that may not be accessible to the analyst (Morton 1994). It also makes use of more components of the heavy-mineral suite and thus provides a more balanced view of provenance characteristics (Morton, 1994).

The qualitative content of the major mineral (Figure 5.2.3) and the semi-quantitative content of quarz, calcite and dolomite was resolved from the 48 samples (Figure 5.1.4: Sampling overview) by application of x-ray diffraction.

About 10g of the gravel sized rock fragments were inserted into a bin with a steel ball and centrifuged to powder size in a centrifuge machine. 0.8g of the powder was mixed with 0.2g flourite, because the adding served as a standard for the quantitative evaluation, and pressed together in a press-tablet. The qualitative and semi-quantitative major mineral content could be interpreted form the XRD-plots (Figure 5.2.3). The positioning on the lateral x-axis (reflection angle) of the peaks reflected type of mineral and the hyperbola surface of the peaks yielded the indirect semi-quantitative content, whereby the added 20% fluorite hyperbola served as a reference. The true content was calculated by a Philips XRD-programme (Philips PW1877 Automated Powder Diffraction, version 3.5b). Not all of the minerals could be identified by the XRD-application since the higher amounts of major minerals superimpose those with the lower amounts. Thus, some mineral amounts were estimated through the thin section microscopy. Furthermore, the XRD-application did not present any possibilities of detecting any accessory minerals, because of the same reason as mentioned above.

The measuring of the powder preparations was accomplished by a x-ray diffractometer of Philips (PW - 1710-Basis) at the Department of Petroleum Geology/Technical University of Clausthal. The measuring was subjected to the following equipment and configuration of the instrument: CU Ka - pipe, automated aperture, 30 {kV} voltage, 20 {mA} amperage, a1=1.54060 - a2=1.54439 {Å} wave length differential, 2.010° start angle and 69.990 end angle, 1 °/min goniometer velocity, 2x10 amplitude, decay of 4. Besides the software of Philips XRD-programme, for the determination of qualitative and semi-quantitative mineral content, the ASTM (Joint Commitee of Powder Diffraction Standards, 1974) was accessorily applied in qualitative determinations.

The qualitative- and quantitative content of elements was analysed with a x-ray fluorescence content meter. A significant application for the ultimate aim of this study, to find two or more dependent geochemical indicators in deep marine siliciclastic environment. Independently, it contributed to identify sedimentary provenance and the sediments assigned to former tectonic settings. Furthermore, the XRF-data occasionally could be replaced with heavy-mineral data, providing securer data. The XRF-data also benefited the classification of the Grès d'Annot sandstone. The glass specimen preparations of the melt method started with the mixing of 3.6g tetraborate and 0.6g of sample powder (1/6 ratio), which was carefully homogenised and inserted into crucibles. These crucibles experienced four pre-heating, heating and cooling interval procedures, whereby the heating was enabled by two opposite placed flames. Finally, the extract from the crucibles were poured into ingot moulds and cooled. The glass specimen topside were numbered and carefully released from the ingot moulds, without damaging the measurable bottom side. The glass specimen analysing was measured at the Department of Mineralogy, Geochemistry and Salt Deposits/Technical University of Clausthal with a Philips (type PW 1480 X-Ray Spectrometer) x-ray diffractometer. This diffractometer was calibrated with USGS-Standard SDC-1 at the IMMR (Institute of Mineralogy and Mineral Resources), which gives reliable data from a range of XRF measurable elements (Figure 5.2.4).

During the heavy-mineral analysing specific dissolved- and stable garnets were encountered, which only occurred in upper levels of the Grès d'Annot succession in the Lac du Lauzanier area. However, only the stable garnets appeared at lower levels in the succession, which lead to an approach in defining the specific dissolved garnet for reconstructing the burial depth of the sandstone. The effort in defining these dissolved- and stable garnets was provided through the application of an EPMA-device. The arranged thin sections for EPMA could not be implemented due poor appearance of garnets, causing difficult circumstances for their localisation. Therefore, six new preparations of heavy-mineral concentrate slides were generated for the EPMA. It requires high quality polishing- and a carbon evaporation of the slide faces without covering, engaging conduction ability for the EPMA-ion-beam. The analysing of the garnets was applied with a digital CAMECA (type SX100 Electron Microprobe) electron probe micro-analyser at the Department of Mineralogy, Geochemistry and Salt Deposits/Technical University of Clausthal. Even after high quality polishing a crude surface of the garnets was not possible to hinder, thus, a beam of 20µm was necessary to reflect enough beam for the collector absorption. The beam was adjusted to 25kV and 20nA, whereby the analysis was attributed with fixed elements of: Ca, Fe, Mn, Mg, Cr, Ti, Al, Si.

The implementation of SEM and an additional device of electron dispersive x-ray analysis (EDX) were mainly utilised for determining the occurring clay-minerals in the samples. An approach was also made with this device in identifying the various appearing garnet-types. However, this device could not be successfully applied in identifying garnet-types since this device was not capable of locating the garnets.

An approximately 1cm-big piece of rock was covered with a grease, which is able to provide settling of gold during evaporation. The gold is the target for an electron bombardment, which besides the SEM, provides an EDX-element identification device. This analysing was accomplished at the Department of Petroleum Geology/Technical University of Clausthal.