The geochemical analysis and correlation of lateral and vertical units are based on the data of heavy-mineral concentrate-slide microscopy (raw-data, table 8.1.2d) and XRF (raw-data, 8.1.2a). This approach focuses a local scale, only including the area of Lac du Lauzanier and excluding the southern derived geochemical data, since no bed to bed correlation is possible between these far-distant areas. These data were engaged for identifying two or more minerals or elements that tend to show a resemblance either laterally or vertically, or both. This resemblance should reflect changes combined with the distribution of sediments and not the changes during weathering or diagenesis. Thus only stable minerals, rather than unstable ones, and elements were used, because they are capable of resisting modification during weathering- and diagenesis processes. Besides, studying the raw-data of stable minerals and elements, to identify resemblance, the data was imported to multivariate-based analyses for correlation coefficient calculation. Appropriate candidates were picked to establish reasonable ratios, which were plotted against other ratios to display geochemical fingerprinting. The ratios were established to include more than the ordinary two variables in a bivariate diagram, for providing higher resolution of changes of sediment distribution and for advancing their identification in the diagrams.

Furthermore, an effort was done in understanding the recognised stable mineral- and element ratio changes of sediment distribution in the deep marine siliciclastic environment of the Grès d'Annot Formation in the Lauzanier area.

Heavy minerals make up an accessory, yet diverse, component (<1%) of siliciclastic rocks. Transmitted light-analysis of these minerals theoretically aids in determining mapping sediment dispersal patterns and correlating sediment bodies and sequences on a regional and local scale. Heavy minerals strongly influence the concentrations of trace and rare earth elements, often generating distinctive geochemical fingerprints and geochemical marker horizons, both of which can be very useful for correlation (Pearce & Jarvis, 1995). Diagrams of various heavy-mineral ratios were plotted for identifying the geochemical fingerprints and geochemical marker horizons of the lateral Lauzanier samples (Diagram 11.2a, b and c). Furthermore an approach was done in finding vertically changes of heavy-mineral distribution in the Lauzanier succession (Diagram 11.2d, e and f). All six plots of the heavy-mineral ratios, whether lateral or vertical could not demonstrate any fingerprints at all since the plotting scatters strongly. There were no proximal - distal heavy-mineral ratio relationships or signatures of these in individual beds or units. Thus, no lateral bed-to-bed or unit-to-unit correlation is possible when applying count analysis of heavy-mineral ratios. Neither does the vertical analysis approach plot any signatures in the diagrams, which are based on vertical order of samples in the Lauzanier succession. While plotting the ratios, it was carefully paid attention to the differing grain-sizes of the samples. This observation revealed that between the sampled fine - coarse-grained arkoses, the differing grain-sizes are not responsible, or control the heavy-mineral distribution in the Lauzanier succession.

Additionally unsuccessfully was the heavy-mineral ratio plotting in the lateral diagrams, concerning the hypothesis of zircons and tourmalines are capable of being transported further by gravity flows than their physical resembling minerals (details, chapter 10.2.: Provenance). The lateral local scale diagrams (Diagram 11.2a, b and c) could not confirm this. Apparently, they do not fingerprint on a local scale, but only on a regional scale, or this feature might have other explanations (details, chapter 10.2.: Provenance).

The plotting indicates a homogeneous and uniform distribution and composition of heavy-minerals in the sampled fine - coarse grained arkoses. No matter where in the Lauzanier succession, this is always confirmed by the sample analysis, which includes Grès d'Annot Formation interval of more than 350m (Appendix 2: Log Correlation). This study confirms the heavy-mineral observations of Stanley (1961), which describes the rock composition in a single Bouma-sequence of the Grès d'Annot Formation as differing as much as in the whole succession. Maybe even the true differing heavy-mineral occurrence in successions are lower than the possible error being made while preparing the heavy-mineral concentrate-slide plus the error of their count. Therefore, it may be concluded that the outcrops of the Grès d'Annot Formation in the Lauzanier area, and also in the southern outcrops (Stanley, 1961), most likely only received sediments from a single source area that probably was sited at the southern Corsica-Sardinia-Esteral-Maures micro-plates assemblage (details, chapter 10.2.1.: Provenance).

The event of the Grès d'Annot Formation in the Lauzanier outcrops and other outcrops in the northern region were supplied by conglomerates that most likely derived from a different source area, could not be verified by the heavy-mineral study. Apparently, the finer sediments of the conglomerate bearing gravity flow did not settle, or their deposition was of minor bulk importance, in the Lauzanier- or other northern parts of the Grès d'Annot basin.

Based on the poor fingerprinting in the high-resolution plots, no vertical or lateral correlation was possible to perform with heavy-mineral ratios (Diagram 11.2a, b, c, d, e anf f). Therefore this study may conclude that a local correlation (>3km) with heavy-minerals are not recommendable in sequences, where the mineral content among layers do not differ significant.

However, when focusing on a regional scale (probably <10km), the plots of the heavy-mineral ratios indicate that current is able to transport some heavy-minerals more distant than other (details, chapter 10.2.: Provenance and diagram 11.2a, b and c). Thus this indication suggest that heterogenuous hydraulic properties among stable heavy-minerals are interesting for further studies, since regional quantity comparement in a single bed might reveal direction of paleo-current, the current strength and proximal-distal status.

The approach of identifying major- and trace element fingerprints in lateral and vertical units was similar to those of the heavy-minerals. However fundamentally for the main approach, plotting element-ratios in bivariate diagrams, a recognisance-study was required for the purpose of establishing element-ratios with appropriate element candidates. This was achieved by calculating the correlation coefficient of all appearing XRF-element-data of all samples, whereby primarily, well correlating elements were of interest. The calculation is based on a multivariate calculation, whereas all lateral- and vertical samples are included as one individual group. Additionally, samples of all individual- beds and of every profile were calculated too, to reveal eventual geochemical changes between lateral- and vertical sampled arkoses. Besides identifying geochemical fingerprints of elements with bivariate diagrams, an application was advanced with ternary diagrams. Mainly to enable, when plotting ratios, the plotting of further two elements in one diagram. Previous authors have used spidergrams and dendograms. These alternatives did not enable any help in this study, mostly because they demand higher differing element-ratios to display meaningful plots. Furthermore the dendograms rely on a discrimination method for the establishment of groups with similar content of elements. Such groups were not possible to develop in this study.

To exclude geochemical signatures that were produced during weathering- and diagenesis processes, which modified minerals and resultant element modification, only the most immobile elements were applied. Fralick & Kronberg (1997) assigns Nb, Al, Ti, and Zr-, while Pearce, Jarvis, Moody & Cope (1993) additionally interpret Rb Cs, and Cr and the HREEs such as Lu and Yb-, furthermore Rollinson (1993) demonstrates that also, when low activities of F¯ then Y and P are immobile during weathering and diagenesis. These authors imply that most of these stable element concentrations are dominantly confined in non-authigenic detrital heavy-minerals.

Although, even when sampling was successfully restricted to only include homogeneous fine - coarse-grained arkoses, caution was assigned to grain-sizes of the samples while plotting the calculated XRF-data. This is advisable, due rapidly changing amounts of element quantities in siliciclastic sediments, which is particularly controlled by grain-size- and sediment homogeneity changes (Stanley et al., 1961). This fact is also valid even when sediments are derived from a single source area and beds are situated close to each other in a formation. Neither in this study, it was possible to sample homogeneous sizes of grains, which was especially a primer intention when sampling arkoses of an individual bed-to-bed correlation. Still these varying grain-sizes of the lateral samples of a single bed are only of minor differences, but severer between beds (Table 8.1.1b). The grain-sizes of the vertical samples, however, are varying much more than the lateral samples. This was purposely done, since they should enable interpretations of the importance of grain-sizes controlling the element content. Furthermore, not to miss the possibility of finding geochemical signatures vertically, a group of similar grain-sizes were applied (Table 8.1.1b). The homogeneity of the sediments appear as throughout homogeneous, except only one sample containing abnormal high- Ca and low SiO2 values (Table 8.1.2a: Sample number 39). Thus, differing homogeneity among the samples were ignored while approaching the identification of geochemical signatures.

Primarily, the effort of identifying laterally and vertically geochemical signatures concentrated on plotting two single stable elements, which are correlating well. The calculated correlation coefficients of all samples, assigns correlation between xx/xx, xx/xx, xx/xx, and xx/xx as the most appropriate candidates for plotting in a bivariate diagram. The result of the xx/xx- and xx/xx-plotting obviously displays an assemblage or a linear trend (Diagrams 11.3a and b). The diagrams indicate that grain-size is not significantly responsible for the assemblage and linear trends, since grain-size averages of individual beds are not plotting in a grain-size order. Neither does the order of grain-size reveal meaningful fingerprints in an individual bed.

Bed 2, 6 and 7 are plotting throughout relatively widespread in the xx/xx-diagram. However, these beds do show a strong linear reliance, except bed 6, in the xx/xx-diagram. The unfortunate plotting of the samples of bed 6, whereby the two most southern derived samples are not indicating any reliance to its bed, is not surprising. During the sampling of this bed, it turned out as very complicated since it is a multi-cycle bed, sited in the upper super-cycle and predominantly among other multi-cycle beds. Thus, this bed was difficult to identify and sample over distances of several kilometres. The same problem was also expected of bed 7, which also is located in the upper super-cycle of the Lauzanier succession. Though, the sampling of this multi-cycle bed was concentrating at the base of the bed, which was a marker in the succession and simple to identify and trace laterally in the field. Similar sample conditions as of bed 7 are valid for the beds in the lower super-cycle (bed 1,2, 4 and 5). These beds are single- as well as multi-cycle beds, but comprise thinner beds as of those in the upper super-cycle; enabling excellent conditions for lateral sampling.

The approach in finding proximal and distal geochemical fingerprints was based on where the lateral samples do relatively plot in diagrams, combined with their true lateral position in a bed. Hereby, a diagram was generated, whereas a line connected the true lateral order of the sample positions in a bed; starting at the most proximal- and ending at the most distal located sample (Diagram 11.3c). Ideally, this line should be as straight as possible. This method was applied throughout all diagrams, which plotted lateral samples. There was no occasion where the lines were even roughly straight and connecting the samples in an order Thus, proximal - distal geochemical signatures were not found.

The effort in identifying vertical fingerprints was based on a similar manner like the proximal - distal approach. Also applying lines that connect the samples, whereby only samples of similar grain-sizes were applied. Ideally, these lines should display a systematic order of where the samples are sited, reflecting the true base - top position in the succession. Equal to the proximal - distal result, no lines were nearly straight in these diagrams; not indicating any vertically related geochemical signature (Diagram 11.3d and e).

The calculation of the correlation coefficient of all elements of all samples may not emphasise all well correlating immobile elements when considering; some elements may only correlate well laterally and not vertically, or opposite. This possibility can decrease the correlation coefficient values, resulting in non-discovered opportunities of plotting well correlating elements. Therefore, calculation of the element correlation coefficients was divided into two groups, one lateral and one vertical group. These tables show a conspicuously lack of varying correlation data between the elements of xx/xx- and xx/xx, which are the most correlating stable elements of the overall coefficient calculation. However, there are significant differences of other coefficient values. The coefficient table of the vertical related samples assigns the elements of xx/xx as correlating exceptionally well. This high coefficient value was not very high in the overall coefficient table, since the lateral samples have negatively influenced this value. Exactly the opposite is a fact when studying xx/xx, which is laterally correlating fair and vertically poor. When plotting the elements of xx/xx a slight correlation tendency of the beds is recognised. Though, this tendency is not as strong as of those in the xx/xx- and xx/xx-diagrams. Apparently the correlation coefficient of xx/xx is to low. In contrast, the vertical related correlation coefficient of xx/xx is very high. However, even with a very high correlation coefficient, the plotting of these elements is only showing a slight vertical trend. The slight trend is recognisable as the samples derived from the lower part of the succession plot with low values of both xx and xx while the samples of the upper super-cycle plot with high values (Diagram 11.3f). The approach, the connecting lines should reveal the true vertical position in the succession, is only partly realised in the diagram. The profiles with few samples show such a trend, however, the samples of the profiles 3 and 4 are lining disorderly. This feature suggests generating a diagram where the upper- and lower super-cycle should be plotted as individual groups. When plotting only samples with grain-sizes ranging from 200 - 300µm, a more convincing tendency is produced (Diagram 11.3g).

The diagram demonstrates that the samples derived from the upper super-cycle predominantly plot in zones with higher values of xx and xx than the samples of the lower super-cycle. The same vertical tendency is displayed by xx/xx-diagram (Diagram 11.3h). Furthermore, plotting between these elements provides an even stronger tendency of isolating the samples of the lower super-cycle, since their xx content is throughout low. In both diagrams, the plotting of some samples, derived from the upper- and lower super-cycle, are mixing in a zone. When studying these samples more closely, they are dominantly sited in a rather moderate base - top position in the succession.

Additionally to the calculation of the correlation coefficients of all lateral and vertical samples, every bed and profile were divided into a group and individually calculated. This was an another effort to verify if no well correlating stable elements were ignored. When studying the correlation coefficients of these two new tables and the other remnant bed and profile tables, many, previous hidden, well correlating coefficients among immobile elements are discovered. Particularly xx and xx were able to correlate well with xx, xx also correlates with xx and further with xx and xx, xx correlates with xx. All these well correlating elements were plotted against each other. Some diagrams did demonstrate fingerprints that provided possibilities of geochemical correlation of individual beds. However, these geochemical fingerprints were not more obvious than those already plotted, whether vertically nor laterally.

Since many immobile elements are able to correlate well with several other immobile elements, an approach was implemented in plotting more than two variables. Thus, ternary diagrams were utilised, because they provide plotting of three variables at once. No matter which three elements were plotted in such a diagram, it was not possible to produce a better resolution of lateral or vertical geochemical fingerprints. Therefore, no ternary diagrams are shown in this work.

Another approach of plotting more than two elements was done by establishing a ratio of two element pairs. This method enables plotting of three or four elements in a bivariate diagram. Primarily, the ratios were generated by elements, which yielded the best correlation coefficients. This characteristic is especially emphasised through xx, xx and xx. Since xx and xx are already plotting a relatively strong tendency of lateral geochemical signatures (Diagram 11.3a), it seems natural to develop a ratio of xx/xx and xx/xx for plotting (Diagram 11.3i). The diagram of the plotted ratios, added with a trend line or regression line, obviously shows a tendency of a firm alignment of samples that derived from a distinct bed. These alignments are highlighted through the regression lines, which indicates high correlation coefficients among element ratios of a distinct bed-to bed correlation. Only four points plot strongly off-line. One pair of these samples derives from bed 6 and the other pair from bed 7, wherein both pairs are located as the two most southern ones. As described (details, text below diagram 11.3b), these off-line plotting samples may not be securely positioned in a bed-to-bed correlation.

Another diagram, which also shows the same trends and features as the plotting of xx/xx and xx/xx, is demonstrated when plotting xx/xx and xx/xx (Diagram 11.3j). This diagram may even exceed the firmness of sample aligning the regression lines. Again the two most southern derived samples of bed 6 are scattering and plotting far off the regression line. However, the previous off-line plotting element ratio pair of bed 7 is arranged moderately near the regression line in this diagram. The only less correlating bed of the xx/xx-xx/xx-diagram is bed 2, whereby the element ratio are somewhat distant the regression line, similar as in the xx/xx-xx/xx-diagram.

Numerous other diagrams did show similar features as the xx/xx-xx/xx- and xx/xx-xx/xx-diagrams, but non of them were able to demonstrate the features better than the previous plotted diagrams.

The element ratio diagrams did not show any further, more convincing than already plotted, vertical geochemical signatures and thus, they are neither shown in this study.

Remarkably, neither the approach of plotting four variables, i.e. plotting a xx/xx-xx/xx-diagram, could hinder thexx/xx-xx/xx- and xx/xx-xx/xx-diagrams as occupying the exceptional position of emphasising possibilities of geochemical bed-to-bed correlation in this work.

When considering the shown diagrams, for the aim of identifying lateral or vertical geochemical signatures, this approach was fairly successful. Vertically, the diagrams can roughly categorise the position of the samples in three zones (Diagram 11.3g and h). A further, more detailed determination of the relatively true base - top position of a sample among other samples in the Lauzanier succession, could not be found in any diagram of the immobile element plots (example, diagram 11.3d, e and f). Apparently, it is not possible to record a detailed change of the depositing sediment of the Grès d'Annot basin. Based on the roughly categorised three zones in the diagrams, the basin in the Lauzanier area was initially filled with sediments of low xx, xx and xx contents. Additionally, the diagrams demonstrate an increase of these contents, while deposition was progressing in time. Conclusively, since the diagrams were not able to order the close-spaced vertical samples, the change of immobile element contents were distributing irregularly (example, diagram 11.3d, e and f). This feature implies that the sediments, which most likely derived from the Corsica, Sardinia, Esteral and Maures assemblage of micro-plates (details, chapter 10.2. and 10.2.1.), were supplied by barely changing rock-types during the erosion-period of the source region. The deposition of the sediments, whereas an span of more than 350m were analysed (Appendix 2: Log Correlation), might even have been influenced by an additional source area and causing the slight increase of i.e. the primarily depositing sediments with low xx, xx and xx values (details, chapter 10.2. and 10.2.1.). Another explanation for these rising values might be explained through the ability of the heavy-minerals, which are comprising the high immobile element concentrations, rather tend to settle in a more proximal environment as of the upper super-cycle.

The petrography throughout the analysed part of the Lauzanier succession, and most likely the other successions across the Grès d'Annot basin too, comprise sediments of a nearly constant composition (Chapter 8.1. including the sub-chapters 8.1.1. and 8.1.2., chapter 10.2. and 10.2.1. and chapter 11.2.). The lack of the nearly devoid changes of sediment composition is hindering the approach of identifying vertical fingerprinting in the diagrams. Thus, the area of Lac du Lauzanier is difficult for such an approach. Theoretically, geochemical signatures should be much easier to identify in sequences where the depositing sediments have derived from differing rock-types of more than a single source area, or of varying rock-types of a single source that eroded more selectively.

When studying the geochemical signatures of the bed-to-bed correlation in the plotted diagrams, the immobile elements demonstrate their high sensitiveness. The elements and element ratios predominantly plot obvious fingerprints (Diagram 11.3a, b, i and j). Some samples of individual beds plot well assembled, but a more convincing geochemical fingerprint is the ability of the stable elements to plot aligned, mostly very near a regression line. By all means, this feature cannot be a coincidence, neither can these convincing results be caused by a changing grain-size or sediment homogeneity of the sampled beds. The homogeneity is evidently displayed in the XRF-data table (Table 8.1.2a). If the change of grain-sizes, which is only ranging between an average of 130 - 550µm (Table 8.1.1b), influence or control the plotting is declined by all diagrams (Diagram 11.3a, b, i and j). Hereby, the beds, comprising various grain-sizes, are not plotting any signature at all. The same evidence is featured by the study of the samples of an individual bed, whereby the occurring slight grain-size changes neither are controlling nor plotting signatures in the diagrams.