2016 | 69/2 | 269–279 | 5 Figs. | 3 Tabs. | www.geologia-croatica Journal of the Croatian Geological Survey and the Croatian Geological Society Geochemistry of Lower Palaeogene bauxites – a unique signature for the tectonostratigraphic evolution of part of the Croatian Karst Zoran Peh and Erli Kovačević Galović Croatian Geological Survey, Sachsova 2, P.O. Box 268, HR-10000 Zagreb, Croatia; (corresponding author: zpeh@hgi-cgs.hr) doi: 10.4154/gc.2016.24 Abstract The origin and emplacement of Lower Paleogene Bauxites discussed here relies on recent per- ceptions concerning the development of the collision-induced diachronous discontinuity surfa- ces in certain parts of the Adriatic-Dinaric Carbonate Platform (ADCP), and in the Istrian Penin- sula in particular. This process, related to forebulge uplift during the initial stages of the orogeny, is amply recorded in the geochemical signature of related Istrian Lower Paleogene bauxites (ILPB) as was shown in previous investigations. As their self-evident extension, due attention is paid here to the effects of a flexural bulge in another part of the platform with its own suite of Lower Paleogene bauxite deposits (LPB), notably in its Dinaridic unit. Explication of the typical patterns of bauxite formation is grounded in the development of discriminant function models based on the Compositional Data (CoDa) analysis of geochemical data. This method provides the solid ground for separation of the various a priori defined bauxite groups deposited in the different subaerially exposed platform palaeoenvironments. In the final analysis, the discriminant model characteristic for the westernmost compartment of the Adriatic SW Unit (Istrian Karst) is cross-compared with the analogous model of the entire ADCP taken from earlier investigation but reconsidered within the CoDa framework. It is done with the purpose of assessing the geo- chemical correspondence between the two event-markers (ILPB and LPB) and clarifying the nature of the pertinent tectonostratigraphic constraints. The computer models show that the fore- bulge unconformity is not a widespread phenomenon, being distinctly recognized only in the Is- trian part of the ADCP. The most recent investigations of bauxites in Croatia had ad- dressed the root causes of geochemical variations observed in the bauxite deposits of Lower Palaeogene age (LPB) from the Croatian karst (KOVAČEVIĆ GALOVIĆ et al., 2012) (Fig. 1), focusing es- pecially on the Istrian Peninsula due to its specific geodynamic evolution through Cretaceous to Palaeogene times (PEH & KOVAČEVIĆ GALOVIĆ, 2014). These studies attempted to ex- plain regular variations in bauxite geochemistry both within re- gional (ADCP) and subregional (Istrian Peninsula) contexts high- lighting changes of depositional and diagenetic characteristics during subaerial exposure of the platform. Although Istrian Lower Palaeogene bauxites (ILPB) accommodate the overall geochemi- cal trend (marked primarily by the southeastward increase in chro- mium content), their stratigraphic position at the regional K–Pg unconformity occurring in the Adriatic part of the platform (Adri- atic domain, sensu KORBAR, 2009) must be treated with due cau- tion because of the implicit possibility of a unique geochemical signature resulting from tectonostratigraphic constraints. Essentially, the implied peculiarity of the ILPB revolves around the major hiatus between the Cretaceous and Palaeogene carbonate successions, which unlike the more south-eastern por- tions of the ADCP, lasted seemingly longer in Istria (e.g. VELIĆ et al., 1995; MATIČEC et al., 1996; TIŠLJAR et al., 1998; VLAHOVIĆ et al., 2005; OTONIČAR, 2007, 2008; KORBAR, 2009; CVETKO TEŠOVIĆ et al., 2011). As a result, the ILPB are underlain by carbonate rocks characterized by a wide age inter- val (from Albian to Santonian-Coniacian) suggesting the pre- dominance of tectonic over eustatic controls due to differential 1. INTRODUCTION In spite of extensive bauxite research and mining activities in Croatia which provided abundant mineralogical and chemical ar- chives, bauxite genesis, particularly in relation to the palaeokarst and its tectonostratigraphic constraints, has been underrated. Most certainly, this unfavourable situation can be ascribed to earlier pre- vailing perspectives on bauxites predominantly as an aluminum ore, with all research efforts concentrated on the assessment of conditions of their detection, exploitation and refinement. Occa- sionally, they were treated in the broader context of palaeoclimate- related geochemical and mineralogical processes (e.g. ŠINKOVEC, 1973, ŠINKOVEC & SAKAČ, 1982, 1991; SAKAČ & ŠINKOVEC, 1991) and of local tectonics (BLAŠKOVIĆ et al., 1989). However, their tectonically induced origin was considered more seriously only after the models of orogenic evolution of the Adriatic (Adri- atic/Dinaric) Carbonate Platform (ADCP) had been constructed with special reference to the External Dinarides of the NE Adriatic region, e.g. by VLAHOVIĆ et al., 2005; KORBAR, 2009; BRLEK et al., 2014. This approach, drawing also from „recognition of fore- bulge related unconformities associated with early stages of fore- land basin development” (CRAMPTON & ALLEN, 1995), stressed the significance of a number of stratigraphic gaps occur- ring in the Mesozoic-Tertiary successions of the Croatian Dina- rides. Considered from this perspective, the complex investigation of bauxites as „tectonic and climatic event markers at regional un- conformities” (D’ARGENIO & MINDSZENTY, 1995) stimulated a search for a new lithostratigraphic scheme setting the background for regional correlation studies over the entire platform. Article history: Manuscript received December 01, 2015 Revised manuscript accepted May 11, 2016 Available online June 28, 2016 Keywords: Bauxite geochemistry, Compositional Data, Discriminant function model, Forebulge unconformity, Adriatic-Dinaric Carbonate Platform, Croatia G eo lo gi a C ro at ic a Geologia Croatica 69/2270 erosion caused by uplift (MINDSZENTY et al., 1995), and also a prolonged period of terrestrial conditions necessary for the ac- cumulation and diagenesis of the protobauxitic material. Presuma- bly, the geochemical signal of the ILPB has responded to the coeval tectonostratigraphic constraints in a somewhat different way with respect to other LPB groups marked by different strati- graphic positions within the ADCP (Dinaric domain, sensu KORBAR, 2009; KOVAČEVIĆ GALOVIĆ et al., 2012). This premise infers a tacit antagonism between the ILPB and the other LPB groups, concerning their geotectonic setting – Adriatic vs. Dinaric domains of the ADCP – whereby the latter are associated with an unconformity formed during the supposedly shorter Late Cretaceous (Santonian-Early Campanian) to Palaeogene (Early- Middle Eocene) hiatus and, as a result, are underlain by carbo- nate bedrock of approximately the same age. The idea underlying this study was to assess the analogy be- tween the temporal sequences of the four ILPB groups (Adriatic) (PEH & KOVAČEVIĆ GALOVIĆ, 2014) and the spatial emplace- ments of other, presumably single-sequence „non-Istrian” LPB groups (Dinaric) (KOVAČEVIĆ GALOVIĆ et al., 2012), based on their corresponding geochemical signatures. It is conceivable to elucidate the behaviour of the PB geochemical signal in diffe- rent tectonostratigraphic settings, particularly with respect to de- velopment of a forebulge. This line of reasoning, in which time series (ILPB overlying a discontinuity surface at various stages of development) are substituted for spatially sampled data from a sufficiently large spatial domain (LPB from Istria to Dalmatia), is commonly termed „ergodic” (implying space-for-time substi- tution). In pursuit of an ergodic solution, the integrated analysis of spatial and temporal domains generally requires comparison of appropriate models. For this purpose, a multivariate statistical assessment of a primary geochemical pattern distinguishing be- tween the four groups of ILPB, (a priori separated according to bauxites overlying carbonate formations of different Cretaceous ages), was carried out using multiple discriminant analysis (MDA) as the most suitable mathematical technique. MDA is em- ployed as a method of data reduction and organization providing an efficient means of geochemical partitioning between several bauxite groups defined by an independent criterion, in this case the stratigraphic formations, or lithostratigraphic units. The de- veloped discriminant function model (DFM) is designed to ac- count for the geochemical changes in time reflected in the ILPB due to the existent (presumably long-lived) subaerial exposure of the Istrian part of the Adriatic platform. The other model describ- ing the variety of forms (bauxites) „frozen” in a distinct time- frame related to the regional (apparently shorter) hiatus over a large proportion of a platform (Dinaric), is taken from earlier re- search (KOVAČEVIĆ GALOVIĆ et al., 2012) and redefined in the spirit of the CoDa analysis for comparison purposes. The aforementioned spatial-temporal correspondence (if present) is expected to be substantiated by analysis and reflected in the matching geochemical signal from both models. Since the whole rock geochemical data analyzed here are of a compositional nature (represented in wt% and mg/kg), provid- ing information about relative magnitudes (as proportions of a whole), their statistical treatment is necessarily focused on the log-ratio approach introduced by AITCHISON (1982; 1986) and disseminated by contemporary proponents of compositional data analysis (e.g., TOLOSANA-DELGADO et al., 2005; AITCHISON & EGOZCUE; 2005; BUCCIANTI et al., 2006; REIMANN et al., 2012; BUCCIANTI, 2013). 2. GEOLOGICAL SETTING The Croatian Dinaric Karst (NE Adriatic region) is an area of thick carbonate successions deposited from the Middle Permian to the Eocene on carbonate platforms of different ages, type and geotec- tonic settings. Some unresolved issues concerning its orogenic evolution in the later stages of development, particularly the rela- tionship between the end-member palaeogeographic and geody- namic models is still subject to differing interpretations (VLAHOVIĆ et al., 2005, and references therein; KORBAR, 2009, and references therein). These models postulate the existence of a single (e.g. VLAHOVIĆ et al., 2005) as opposed to two carbonate platforms (e.g. KORBAR, 2009), usually with Adriatic-Dinaric designation (ADCP), developed as a result of fragmentation of the Adriatic microplate (Adria) during the Upper Triassic and the Lower Jurassic, and subsequent events at its plate boundaries (e.g. VLAHOVIĆ et al., 2005; SCHMID et al., 2008; KORBAR, 2009; SISCIANI & CALAMITA, 2009; USTASZEWSKI et al., 2010 and references therein; MÁRTON, et al., 2010; ŠUMANOVAC, 2010; MÁRTON, et al., 2011; NOCQUET, 2012; HINSBERGEN et al., 2014). One of the themes highly relevant to the genesis of bauxites is associated with different tectonostratigraphic con- straints characterizing the two contrasting platform domains, namely Dinaric and Adriatic, both with their pertinent tectono- stratigraphic units, proposed in spite of the still uncertain NW ex- tension of the Budva basin and time of opening of the NE Adriatic trough (KORBAR, 2009) (Fig. 2). The first refers to the recent Ex- ternal Dinarides, a highly deformed fold-and-thrust belt of Alpine origins formed by the collision of Adria with the Austroalpine and Tisia domains. The other is the Adriatic foreland, a more stable part of Adria characterized by normal faulting and gentle compres- sional late-orogenic deformations, which are tectonic features typi- cal for Istria (MATIČEC, 1994; MATIČEC et al., 1996). Although the dynamic interplay of synsedimentary tectonic activity and eustatic changes had continually modified the ADCP Figure 1. Topographic sketch map showing areas included in the two bauxite- research campaigns relevant for this study: 1 – Geochemical investigation of Lower Palaeogene bauxites (LPB) from the Croatian part of the Adriatic-Dinaric carbonate platform (ADCP) (KOVAČEVIĆ GALOVIĆ et al., 2012); 2 – Geochemical investigation of Lower Palaeogene bauxites from the Istrian Peninsula (ILPB) (PEH & KOVAČEVIĆ GALOVIĆ, 2014). G eologia C roatica Zoran Peh and Erli Kovačević Galović: Geochemistry of Lower Palaeogene bauxites – a unique signature for the tectonostratigraphic evolution ... 271 until the final uplift of the Karst Dinarides reaching its maximum extent during the Oligocene-Miocene period, from the perspec- tive of bauxite formation only the period between the end of the Lower Cretaceous and the beginning of the Palaeogene is of in- terest for the scope of this investigation. This period is high- lighted by a major discrepancy in the nature of a regional uncon- formity developed on the carbonate platform, marked by recognition of the strong forebulge effect in the Adriatic domain, Istria in particular, a feature only hinted at previously in the southeastern part of Dinaric High Karst. As one of the five tectonostratigraphic units in the strati- graphic scheme of the entire ADCP, the Istrian Peninsula is reco- gnized as the Adriatic SW unit (ASWu) or Istrian Karst (KOR- BAR, 2009) (Fig. 3), sometimes called stable Istria (e.g. MÁRTON et al., 2003; MÁRTON et al., 2010). Most of it is distinguished by a prolonged emersion phase at the Cretaceous–Palaeogene (K–Pg) boundary which caused a major hiatus between the two carbon- ate platform depositional systems – the Cretaceous passive margin shallow-marine carbonate sequence of the Adriatic Car- bonate Platform and the Upper Cretaceous and/or Palaeogene shallow-marine sequences of the synorogenic carbonate platform, resulting from the specific evolution of the Adriatic foreland ba- sin complex (ASWu). A major regional subaerial unconformity resulting from flexural deformation (uplifting forebulge) in front of the approaching Dinaric orogen is locally denoted by prono-prono- unced palaeokarstic features (DURN et al., 2003; OTONIČAR, 2007; KORBAR, 2009; BRLEK et al., 2013). These clues firmly support the idea that a large part of present day Istria probably formed land until the onset of the Palaeogene transgression (MATIČEC et al., 1996). However, no direct relationship with bauxite formation can be established in the area since the oldest bauxite-hosting rocks are represented by limestones of Albian age while the scattered outliers of the Eocene carbonate rocks overlie still older Lower Cretaceous formations (of Valangian to Barremian age), indicating erosional vacuity devoid of bauxites (PEH & KOVAČEVIĆ GALOVIĆ, 2014). The complete strati- graphy of the ADCP carbonates from the Istrian Karst was de- scribed at length with reference to the typical Lower and Upper Cretaceous successions in VLAHOVIĆ (1999) as well as in GUŠIĆ & JELASKA (1990) and CVETKO TEŠOVIĆ et al. (2001), respectively. Albian to Campanian lithostratigraphic units relevant to this investigation include the following in ascending chronostratigraphic succession: 1) Pula formation (PU – Albian), 2) Rušnjak formation (RU – Cenomanian), 3) Sveti Duh forma- tion (SD – Upper Cenomanian-Lower Turonian) and 4) Gornji Humac formation (GH – Upper Turonian-Upper Santonian/Cam- panian) (MATIČEC et al., 2013). The latter three formations have been derived after the similar succession reported from the Island of Brač (GUŠIĆ & JELASKA, 1990; STEUBER et al., 2005), with modifications. In contrast to the Istrian Peninsula, the south-eastward por- tion of the Dinaric High Karst (Dinaridic SW unit, or DSWu, sensu KORBAR, 2009) is characterized by a Santonian (Early Campanian) to Late Palaeocene hiatus encompassing the Conia- cian to Santonian interval (earlier determined as Turonian to „Lower Senonian”), narrowing considerably the approximate age interval of the associated bauxites. The absence of an obvious tectonically induced bulging effect (diachronous uplift), ruled out differential erosion that would have occurred during the subae- rial exposure of the platform interior, introducing an entirely dif- ferent kind of unconformity in this area. Accordingly, the K–Pg boundary over the entire DSWu is marked by the underlying Cre- taceous bedrocks having similar ages, such that the Upper Cre- taceous strata underlying the bauxites generally shows bio- and lithofacies characteristics of the Gornji Humac (GH) formation Figure 2. Simplified tectonic map of the Adriatic-Dinaric carbonate platform (ADCP), modified from KORBAR (2009): 1) ASWu (Istrian Karst); 2) ANEu (Dalma- tian Karst); 3) DSWu (High Karst); 4) DNEu (Inner Karst); 5) Thrust front of the Ex- ternal Dinarides; 6) Thrusts of Adriatic and Dinaridic units; 7) Other major faults; 8) Investigated groups of Lower Palaeogene bauxites (LPB) from different quar- ters of the ADCP (after KOVAČEVIĆ GALOVIĆ et al., 2012): IST – Istria, NAI – North- Adriatic islands, NDA – North Dalmatia, CDA – Central Dalmatia. Figure 3. Simplified geological map of the Istrian Peninsula (compiled from the Geological map of the Republic of Croatia, 1:300,000, Croatian Geological Sur- vey, 2009) showing locations of the sampled Istrian Lower Palaeogene bauxites (ILPB): 1) Upper Jurassic formations; 2) Lower Cretaceous formations (including PU); 3) Upper Cretaceous formations (RU–GH); 4) Palaeocene–Middle Eocene formations; 5) Middle Eocene–Upper Eocene formations (flysch); 6) Thrust faults; 7) Investigated Istrian Lower Palaeogene bauxites (ILPB) grouped in accordance to underlying Cretaceous formations (after PEH & KOVAČEVIĆ GALOVIĆ, 2014). G eo lo gi a C ro at ic a Geologia Croatica 69/2272 (KORBAR, 2009; KORBAR et al., 2012). This is strong evidence in support of a major discrepancy between the bauxite deposi- tional environments concerning the two tectonostratigraphic units (ASWu vs. DSWu) in the Croatian karst. Although both are related to plate-interior settings (MINDSZENTY et al., 1995; D’ARGENO & MINDSZENTY, 1995), they differ according to the longevity of the hiatus and consequences of the arching effect which is typical only for Istria. 3. MATERIALS AND METHODS 3.1. Sampling, sample preparation and geochemical analysis Sampling was carried out on several occasions so that 50 LPB bauxite samples were collected from various sites all over the ADCP (including 21 early samples from Istria) (KOVAČEVIĆ GALOVIĆ et al., 2012) and 45 new Istrian samples added in a later campaign to complete the sampling scheme for the separate investigation of the ILPB (PEH & KOVAČEVIĆ GALOVIĆ, 2014) (Fig. 1). Samples from all locations were collected in a ran- dom fashion following the idea that bauxites from a single re- gional unconformity (K–Pg transition) should represent a bulk geochemical composition regardless of their particular horizons within the ore body (e.g. at the contact with the related palae- okarst or immediate cover), particularly in the case of small de- posits (100-300 m2) or occurrences typical for LPB in Croatia (Fig. 4). It is tacitly accepted in this work that multivariate analy- sis carried out on the bulk geochemistry of a number of cases is capable of identifying the most important geochemical processes by reducing all details along the vertical profile to a few mathe- matical functions. Bauxite samples weighing about 3 kg in total were crushed by hand, split into fractions by quartering, and finally ground in a tungsten carbide pestle mortar mill (Retsch Lab Equipment) and sieved to 0.063 mm in preparation for the analytical work. Chemical analysis was performed at the ACME Labs (ACME Analytical Laboratories Ltd.) in Vancouver, Canada, applying the Lithogeochemical Whole Rock Major and Trace Element ana- lytical method. Total abundances of the major oxides and a few minor elements were reported on a 0.2 g sample analyzed by ICP- emission spectrometry following a Lithium metaborate/tetrabo- rate fusion and dilute nitric digestion. Total trace elements were analyzed by ICP mass spectrometry – refractory elements went through the same decomposition as the major elements (addi- tional 0.2 g sample) while the rest were digested in hot Aqua Re- gia and analyzed by ICP MS (0.5 g sample). Analytical work was subjected to the strictest quality control in the ACME Labs, with blanks, duplicates and standard reference materials (STD SO-18) included in the CQ report, together with a certificate of analysis, as a measure of background noise, accuracy and precision. For the purpose of this work, the procedure for handling the data be- low the analytical detection limit (left-censored data) was ac- cepted from TARVAINEN et al. (2005) as the substitution method of assigning one-half the detection limit value to cen- sored data. This method is predominantly in use in geochemical and environmental studies (as per ANTWEILER & TAYLOR, 2008) developed from earlier works on analytical chemistry (e.g., Analytical Methods Committee, 1987). 3.2. Statistical considerations 3.2.1. Compositional data and log-ratio analysis Statistical analysis employed a selection of 32 elements including 10 major elements (represented as total abundances of major ox- ides) and 22 trace elements (Cr represented as Cr2O3) performing as predictor variables in MDA. Data from earlier research work embracing LPB from all over the ADCP had to be recalculated into log ratios in order to be compatible with the recent discrimi- nant function model (DFM) of Istrian bauxites. The strategy of analysis was adjusted to the objectives of this research so that ILPB data were installed as a foundation for the reference model against which LPB data were weighed (including identical pro- cedures described in chapter 3.1). This arrangement meant that the tables in most cases combined the results of analysis of both data sets albeit separately and LPB is discussed later in the text (denoted as a, and b). The analyzed dataset consists of 66 bauxite samples col- lected from ILPB deposits and occurrences, a priori separated into the four groups in accordance with rock formations (infor- mal lithostratigraphic units from ASWu) as their immediate bed- rock. The four groups include the Pula formation (PU), Rušnjak formation (RU), Sveti Duh formation (SD) and Gornji Humac formation (GH) in upward succession. In contrast, the associated LPB database contains a total of 50 samples previously collected from different quarters of the ADCP (including Istria) and cate- gorized in the four groups with a geographic designation: Istria (IST), North Adriatic Islands (NAI), North Dalmatia (NDA), and Central Dalmatia (CDA) (KOVAČEVIĆ GALOVIĆ et al., 2012); approximately in a NW-SW direction. The summary statistics for the whole dataset prior to data transformation and subsequent multivariate statistical procedure are displayed in Table 1 (minimum, maximum, median, and geo- metric mean). However, this information is relevant only if one is interested in relative rather than absolute values such as, for example, in the case of comparison with other similar investiga- tions, because the whole rock composition of bauxite samples represents the classical example of compositional data (CoDa) in geochemistry. Although the basic principles of CoDa are thoro- ughly explained in the previous work on bauxites (PEH & KOVAČEVIĆ GALOVIĆ, 2014) it is deemed necessary to reite- rate this issue again to some extent due to its relative novelty in the Croatian scientific literature, especially among geologists. The nature of CoDa implies a mathematical property that all vari- ables (compositions) in the analyzed sample must sum to a unit value (100% or 1.0). As a result, all geochemical, mineralogical, and other datasets in the geoscience world are heavily plagued by the constant-sum constraint (CSC). This problem interferes with procedures of traditional statistics since individual variables Figure 4. Outcrop of a small-scale Lower Palaeogene bauxite deposit (LPB) on Pag Island (Metajna-Zubovići) G eologia C roatica Zoran Peh and Erli Kovačević Galović: Geochemistry of Lower Palaeogene bauxites – a unique signature for the tectonostratigraphic evolution ... 273 are represented only as parts of some whole (or fractions of a con- stant sum) and, prevented from fluctuating independently (closed data), they are involved in spurious correlations. Formally, CoDa cannot be represented in their raw form as points in the open, Euclidean space, where the scale is absolute, not relative. They refer to a restricted sample space known as simplex (simplicial complex) consisting of D parts or compositions (e.g. geochemical variables). Thus a D-part composition (SD) is really a subset of D-dimensional real space (RD) (PAWLOWSKY-GLAHN & EGOZCUE, 2006) which can assume the Euclidean vector space structure only after the proper transformation of its components. From an array of transformations introduced in the literature, the centered log-ratio transformation (clr) of raw (compositional) data, originally proposed by AITCHISON (1986) is used here. The application of the centered log-ratio is deemed essential for processing CoDa in the multivariate statistical methods such as MDA since it preserves original distances between correspond- ing compositions allowing them to be handled in a straightfor- ward way (EGOZCUE & PAWLOWSKY-GLAHN, 2006; TO- LOSANA-DELGADO, 2012). The singularity problem inherent to a clr-transformed covariance matrix could be circumvented if MDA operates on its reduced form, which is not relying on a full rank of covariance (DAUNIS-i-ESTADELA et al., 2011). Since clr-transformed data represent unbounded real vectors in real space, in this case Mahalanobis distances (MD) stay invariant regardless of which component may be removed from the analy- sis (BARCELÓ-VIDAL & PAWLOWSKY-GLAHN, 1999). Non- essential clr-transformed variables may be amalgamated („other”) and removed from further analysis. Clr-coefficients are computed from the following expression: clr(x) = ( )  xg x1log , log ( )xg x2 , log ( )xg x3 , …., log ( )  xg xD : where x1, x2, x3,…xD represent parts (compositions), and g(x) rep- resents the geometric mean of the parts. As the compositional nature of geochemical data is ex- pressed either in wt %, or in mg/kg, affecting the scale, all measu- rements are converted into mg/kg before transformation by mul- tiplying wt% by 104. 3.2.2. Building a predictive discriminant model (DFM) Multiple discrimination analysis (MDA) is one of the most ex- ploited traditional multivariate statistical techniques particularly effective in building the predictive models of a multi-group dis- crimination established on the array of independent, or predictor variables. This method is amply described in earlier papers with similar targets and subjects of investigations (e.g. PEH & HALAMIĆ, 2010; PEH & KOVAČEVIĆ GALOVIĆ; 2014; GALOVIĆ & PEH, 2016) and will not be thoroughly discussed Table 1. Descriptive statistics of raw (compositional) geochemical data for a) ILPB, and b) LPB datasets. a) ILPB b) LPB Element Min Med Max g Min Med Max g SiO2 (%) 0.91 4.58 21.72 4.31 1.02 3.415 9.1 3.41 Al2O3 (%) 33.02 48.75 57.20 47.54 34.5 50.415 58.58 49.82 Fe2O3 (%) 4.13 24.58 42.21 25.02 14.76 25.955 45.96 25.62 MgO (%) 0.04 0.13 0.49 0.13 0.06 0.11 0.49 0.12 CaO (%) 0.03 0.08 7.80 0.10 0.04 0.115 0.64 0.10 Na2O (%) 0.005 0.01 0.33 0.01 0.005 0.01 0.33 0.01 K2O (%) 0.005 0.06 0.24 0.05 0.005 0.03 0.2 0.03 TiO2 (%) 1.30 2.90 3.67 2.76 1.6 2.955 3.88 2.84 P2O5 (%) 0.005 0.06 1.84 0.05 0.00 0.053 0.113 0.05 MnO (%) 0.005 0.05 0.12 0.05 0.02 0.05 0.46 0.06 Cr2O3 (%) 0.047 0.07 0.252 0.08 0.053 0.085 0.208 0.10 Ni (mg/kg) 52 200.00 524 186.31 88.0 224.5 1092.0 222.43 Sc (mg/kg) 30 47.50 123 47.43 35.0 54.0 91.0 56.07 Ba (mg/kg) 6 25.00 66 23.88 6.0 24.0 73.0 23.63 Co (mg/kg) 12 33.85 81 31.87 14.2 38.2 91.9 38.29 Ga (mg/kg) 25.3 49.85 66.4 48.54 33.6 53.55 65.1 52.49 Nb (mg/kg) 23.8 54.95 70.9 53.22 30.7 59.4 75.4 56.47 Sr (mg/kg) 27.6 96.95 992.7 96.56 27.6 74.7 220.4 74.59 Th (mg/kg) 24.8 43.50 71.1 42.69 28.5 45.25 60.7 45.25 U (mg/kg) 4.6 10.60 29.4 11.05 4.6 8.7 26.9 8.58 V (mg/kg) 202 425.50 1341 444.38 286.0 539.0 1876.0 565.45 Zr (mg/kg) 246.6 517.20 667.9 497.95 312.3 532.55 721.2 523.30 Y (mg/kg) 31.4 64.65 111.0 61.44 31.4 69.7 255.9 77.00 La (mg/kg) 40.7 87.75 146.0 87.11 40.70 93.8 179.3 94.00 Ce (mg/kg) 64.6 178.35 1096.9 172.74 64.6 193.8 645.3 203.49 Mo (mg/kg) 0.9 6.65 44.8 6.43 0.9 7.55 33.0 6.95 Cu (mg/kg) 14.5 95.35 247.6 87.75 21.9 82.3 250.5 87.40 Pb (mg/kg) 21.4 74.15 119.7 72.40 44.2 97.6 196.6 97.80 Zn (mg/kg) 10 74.00 633 78.65 12.0 90.0 422. 93.01 As (mg/kg) 0.25 51.15 269.60 45.18 1.7 36.15 154.9 28.57 Cd (mg/kg) 0.05 0.85 4.40 0.79 0.1 1.2 11.0 1.19 Hg (mg/kg) 0.03 0.29 2.23 0.29 0.03 0.225 1.02 0.22 Note: Min = minimum; Med = median; Max = maximum; g = geometric mean G eo lo gi a C ro at ic a Geologia Croatica 69/2274 here. This also relates to the process of model building – from the structural to the functional, by assigning the geological meaning to discriminant functions (labeling) – which is the stepping stone of each MDA. It suffices to say that here, the scope of this study is focused on building the two predictive discriminant models with maximum classification efficiency, based on the two sets of bauxite groups defined earlier as: a) ILPB (containing PU, RU, SD and GH groups) and b) LPB (containing IST, NAI, NDA and CDA groups); and the 32 log-ratio transformed CoDa chemical variables (whole rock chemical analysis). The latter set (b) in- cludes the log ratio adjusted LPB model already defined in the earlier project (KOVAČEVIĆ GALOVIĆ et al., 2012) and added to this investigation for comparison purposes and discussion on tectonostratigraphic constraints with possible ergodic implica- tions. To this end, a discriminant analysis from the statistical software package of STATISTICA, Release 7.1 (Statsoft, Inc., 2006) was used in order to achieve the best separation between the groups. 4. RESULTS AND DISCUSSION Results of the MDA are briefly recapitulated in Table 2. contain- ing: a) the multivariate test for the overall significance of discrimi- nation and; b) the tests of residual roots (discriminant functions). The Wilks’ l statistical test, employed routinely in the analysis, shows the vanishingly low probability (p<0.000) confirming that all bauxite groups have the same multivariate mean, which is compulsory in order to safely proceed with computing discrimi- nant functions (DFs). Since the number of groups in the studied cases is only four (K=4) the between-group variation is com- pletely explained by the three discriminant functions (K–1). How- ever, examination of Table 2. reveals that only the first two dis- criminate functions justify closer inspection. DF1 and DF2 together account for more than 94% and almost 93% of the total between-group difference in ILPB and LPB models, respectively, allowing the third one to be easily disregarded. For the sake of convenience and prevention of possible confusion, the interpre- tation of the two models will proceed separately allowing, as it were, the LPB model to be „integrated” with the already built tectonographic scheme of the Istrian Karst (PEH & KOVAČEVIĆ GALOVIĆ, 2014). 4.1. Labeling the functions The function labeling for the ILPB dataset has been presented in PEH & KOVAČEVIĆ GALOVIĆ, 2014, so that the pertinent dis- criminant model of the Istrian karst will only be briefly revisited since the comparison of the two models (ILPB+LPB) is indispen- sable for their integration into the greater ADCP scheme. It suf- fices to say that in the computed mathematical model of the ILPB (Tab. 2a; Fig. 5a-b), the first discriminant function DF1 highlights the simple enrichment/depletion master relationship while DF2 provides additional insight both into the process of ferralization experiencing manganese enrichment in the final stages, and into various anaerobic processes affecting organic-rich argillaceous sediments in the poorly drained protobauxitic material. For the LPB model previously constructed by classical statistical tech- niques all MDA paraphernalia had to be recomputed in the CoDa spirit (Tab. 2b; Fig. 5c-d). 4.2. LPB model revisited – the CoDa approach Previous investigation of Lower Palaeogene bauxites revealed an exceptional role for Cr in their disposition over the Dinaric karst (KOVAČEVIĆ GALOVIĆ et al., 2012). However, from the stand- point of statistical modeling, the LPB model discussed in earlier work suffered from a variety of problems arising from the tradi- tional raw data approach. One of the greatest shortcomings was the separate analysis of major and trace elements resulting in two different DFMs, each encumbered by the problem of sub-com- positional incoherence (incoherence of the correlation between the two raw components as a measure of dependence) (e.g. PAW- LOWSKY-GLAHN & EGOZCUE, 2006; AITCHISON, 2008). Here, this issue is resolved by the compositional data approach which allows working on the full set of geochemical data in terms of ratios (such as clr coefficients) always preserving sub-compo- sitional coherence. Also, one of the major fallacies related to „classical” statistical methods is exposed by the CoDa analysis. Specifically, in some cases, including the previously established LPB model (KOVAČEVIĆ GALOVIĆ et al., 2012), it may turn out that the new methodology validates what is already known through the use of traditional methods with raw compositional data. However, the straightforward answer to this dilemma is that more often than not „either the researcher was lucky using his traditional methods or at least the new methodology must be cor- rect in that case” (AITCHISON, 2008). The CoDa DFM is strikingly similar to its classical counter- part particularly in the portion applying to the group arrangement based on the latter’s trace element dataset (KOVAČEVIĆ GALOVIĆ et al., 2012). However, correspondence also exists for the major-element suite so that the new model integrates the whole set of transformed compositional data in Euclidean space allowing for more coherent interpretation of geochemical data in the bauxite-forming environment. This is to say that specific per- formance of Cr is reinstated by its complementary positioning with regard to bauxitic/clay components (primarily K2O, fol- lowed by Al2O3, TiO2 and others). As a strongly bipolar function, DF1 clearly separates the North Adriatic IST and NAI groups against the Dalmatian NDA and CDA groups (Fig. 5c-d), show- Table 2. Multivariate test for overall significance of discrimination, and tests of residual roots for a) ILPB discriminant function model, and b) LPB discriminant function model. a) No. of variables 32 No. of samples 66 Wilks’ lambda 0.015 Approximate F ratio 2.972 Degrees of freedom [96; 93] p-level p < 0.0000 DF Eigen value Eigen (%) Canon. R Wilks’ l chi2 df p- level 1 10.339 77.23 0.955 0.015 196.6 96 0.000 2 2.289 17.10 0.834 0.173 82.5 62 0.042 3 0.759 5.67 0.657 0.568 26,6 30 0.647 b) No. of variables 32 No. of samples 50 Wilks’ lambda 0.00036 Approximate F ratio 6.2687 Degrees of freedom [96; 45] p-level p < 0.0000 DF Eigen value Eigen (%) Canon. R Wilks’ l chi2 df p- level 1 54.510 81.51 0.991 0.000 245.8 96 0.000 2 7.417 11.09 0.939 0.020 121.3 62 0.000 3 4.948 7.40 0.912 0.168 55.3 30 0.003 Note: LPB data clr-transformed from KOVAČEVIĆ GALOVIĆ et al., 2012. G eologia C roatica Zoran Peh and Erli Kovačević Galović: Geochemistry of Lower Palaeogene bauxites – a unique signature for the tectonostratigraphic evolution ... 275 ing the latter as a repository for the ingress of Cr from some ex- ternal source – most probably the Dinaridic ophiolite belt with ultramafic massifs and genetically related sedimentary forma- tions (PAMIĆ et al., 2002) in the Internal Dinarides. This input is so strong that, in all probability, the LPB deposited further to the NW simply reflects the sharp decline in Cr content putting to the forefront other bauxite components mentioned above. With the introduction of DF2 (only 11.1% of total variability) the over- all picture is only embellished by separating the NAI and NDA groups as more enriched in aluminum and related components (Al2O3, TiO2, Ga, Nb) as well as rare earths and associated ele- ments (Ce, La, P2O5, Sc, Y). This leaves the IST group as genera- lly the most kaolinitic (K2O) among the investigated bauxites (upper left quadrant of Fig. 5c-d). 4.3. A geochemical signal – setting the limits of the bauxite age? As suggested before, the specific tectonostratigraphic constraints characterizing different ADCP units imply the possibility of a hidden coherence underlying the unique geochemical signatures of both the ILPB and LPB in general. This particular line of cor- respondence can be scanned from analogous discriminant models developed for both instances. In this regard the Mahalanobis dis- tances (MD) computed for the ILPB model can be compared with the rebuilt model from the earlier geochemical investigations of the LPB over the entire ADCP (KOVAČEVIĆ GALOVIĆ et al., 2012) as illustrated in Table 3. In the former case the multivariate distances increase successively from the PU upward, putting the SD at the upper end of the sequence indicating the greatest MD values, more than twice those of the youngest, GH group (Tab. 3a). RU is also closer, that is, more similar by its geochemical signa- ture, to the GH than to the SD. In the latter however, a somewhat different response is observed, in this instance with regard to the spatial rather than temporal relationships – four LPB groups distributed along the NW-SE directed K–Pg regional uncon- formity, from Istria (IST) to central Dalmatia (CDA) (spatially „ascending”, or easting), showed a regular tendency to increase the Cr-signal towards the SE with a characteristic acme at the central Dalmatian (CDA) group, the last in the sequence (IST ® NAI ® NDA ® CDA) (Tab. 3b). Hence, spatial and temporal arrows have characteristic peaks that correlate the chronological position of the bauxite-hosting SD formation in Istria (relatable to the supposed Cr–event) with the spatial position of the CDA group of the LPB along the same dis- continuity surface at the K-Pg boundary. The rapport between the two models provides a rationale for comparing the controls of subaerial exposure (tectonic uplift in the plate-interior settings during collision along the distant plate margins, according to Figure 5. Comparison be- tween variables and groups in the ILPB and LPB discrimi- nant function model (clr- transformed data): scatter- plots of variable loadings (a,c), and individual objects (samples) (b,d) in reduced discriminant space of the first two discriminant func- tions (DF1-DF2). Table 3. Mahalanobis distances (MD) for a) ILPB discriminant function model, and b) LPB discriminant function model. a) ILPB b) LPB GROUP PU RU SD GH GROUP IST NAI NDA CDA GH up w ar d su cc es sio n à 0.00 CDA N W à S E 0.00 SD 0.00 31.49 NDA 0.00 83.16 RU 0.00 70.32 28.52 NAI 0.00 161.64 274.32 PU 0.00 26.49 149.98 71.52 IST 0.00 40.68 202.30 309.22 à upward succession NW à SE Note: LPB data clr-transformed from KOVAČEVIĆ GALOVIĆ et al., 2012. G eo lo gi a C ro at ic a Geologia Croatica 69/2276 D’ARGENIO & MINDSZENTY, 1995), invoking an ergodic hypothesis as a working tool. Although some authors hold it inappropriate to swap spatial for temporal variability in ergodic approaches to climate-tectonics-erosion studies (BALCO & STONE, 2005) it can, in truth, be the only way to model stochas- tic processes in case if no sample helps meaningfully to predict values that are very far away in time of that sample. This line of reasoning assumes that spatially sampled data can be substituted for time series (PAINE, 1985) which is particularly useful in con- trasting the two discussed DFM. A singular, (probably the same) event (Cr–increase) can be expressed in two models: a) in the tem- poral model where the ILPB deposited on the SD formation (Up- per Cenomanian-Lower Turonian) indicates a specific temporal point during the major hiatus associated with subaerial exposure of the Istrian Karst (ASWu), and; b) in the spatial model where the LPB (IST excluded), are related to a major unconformity be- tween the Uppermost Cretaceous and the Early Palaeogene strata developed under distinctly different tectonostratigraphic con- straints prevailing over the High Karst area (DSWu, sensu KOR- BAR, 2009). In the latter case, the rocks underlying the erosional surface show characteristics of the GH formation originally de- scribed on the Island of Brač (GUŠIĆ & JELASKA, 1990). How- ever, tectonic deformation over the DSWu are more intricate with respect to the Istrian Karst (ASWu), except for the Učka Mt which was strongly affected by the latest orogenic wrench-fault tecton- ics (PRELOGOVIĆ et al., 1995; PRELOGOVIĆ et al., 2003), rep- resenting an area struck by both strike-slip and compressive tec- tonics (SCISCIANI & CALAMITA, 2009). In this regard it is worth mentioning that subaerially exposed ILPB deposits are pre- served even on the Učka Mt about 800 m above the Istrian main- land, suggesting that their immediate cover (Middle Eocene fo- raminiferal limestones; ĆOSOVIĆ et al., 2004) has been stripped off much later prohibiting the complete removal of the ILPB by advancing denudation of the underlying Cretaceous bedrock in spite of the vigorous uplift. This, together with the fact that Brač Island (a model for Upper Cretaceous formations division; GUŠIĆ & JELASKA, 1990) pertains to the Dalmatian Karst tectono- stratigraphic unit (Adriatic NE, or ANEu, sensu KORBAR, 2009) (Fig. 2), showing a somewhat different tectonic pattern, further complicates the clear-cut comparison of the bauxite-hosting litho- logic units along the entire K–Pg boundary within the ADCP do- main. Provided that the Cr–event which filtered through both the ILPB and the LPB was more or less synchronous over the entire karst Dinarides during the K–Pg interval, it would demonstrate different responses of the particular fragments from the Adriatic and Dinaric domains in Croatia to the tectonic deformations caused by the differential spatial influence of the advancing Di- naridic thrust front (KORBAR, 2009; BRLEK et al., 2013). 4.4. The ILPB model of the Istrian Karst (ASWu) As mentioned earlier, results from the previous investigations concerning the bauxites on the Istrian karst (ASWu) (PEH & KOVAČEVIĆ GALOVIĆ, 2014) ought to be briefly reiterated in order to present the link to the rest of the carbonate platform more clearly. The Istrian Karst seems to be of particular importance since the hiatus in that part of ADCP is associated with a promi- nent forebulge zone induced by the approaching Dinaridic thrust front, bringing into direct focus the significance of the erosional vacuity. The PU-GH depositional sequences are clearly truncated forming an unconformity which is directly overlain by Eocene foraminiferal limestones and occasionally by the Liburnian for- mation, (but only in the farthest SE part of the Istrian Peninsula). According to OTONIČAR (2007) the initial uplift of the fore- bulge occurred at the end of the Campanian, which resulted in differential erosion of the subaerially exposed Cretaceous car- bonate successions. Controlled by topography, uplift, relief, rate of erosion, vacuity and total stratigraphic gap are all mutually and positively correlated (MINDSZENTY et al., 1995) – the youngest deposits were first to emerge at the crest of a subaeria- lly exposed forebulge and they were also the first to experience erosion. In this process, after complete destruction of the as- sumed Campanian strata, erosional truncation advanced deeper, affecting the successively older formations. As a result of the mi- gration of the forebulge and the progradation of the associated foreland basin system through time, away from the approaching Dinaric thrust front (in conformity with the model promoted by CRAMPTON & ALLEN, 1995, and DeCELLES & GILLES, 1996), both the age of the bedrock immediately underlying the evolved unconformity and the spread of the stratigraphic gap ex- panded steadily in SW direction (OTONIČAR, 2007). In other words, the K-Pg unconformity in Istria is formed as a „dia- chronous erosion surface” (sensu YANG, 2011) with the young- est bedrock formation (GH) closer to the orogenic wedge and the older formations successively further away. Naturally, the deve- lopment of the palaeokarst and the accumulation of the proto- bauxitic material began with the exhumation of the GH and were brought to an end with the PU, in reverse order with respect to their deposition. With this in mind, the aftermath of the Cr–event as reflected in the DF1 of the ILPB model, can be dated as post- Campanian, subsequent to the erosional truncation which cut suf- ficiently deep to exhume the SD formation. The structure of the DF1 can be attributed to the uplift-related change from phreatic to vadose conditions during karstification of the emerging fore- bulge similar to that reported by MINDSZENTY et al. (2000) from the Cretaceous of the Transdanubian Range, Hungary. The originally high content of both Cr and Ni changes gradually in favour of Cr due to the loss of Ni through the process of leaching which intensified towards the crest of the forebulge during uplift. In the incipient stages of evolution of the forebulge the deposi- tional/diagenetic environment probably favoured the settling of phreatic conditions since the newly formed low-level karst planes did not support formation of a deep unsaturated zone with unim- peded drainage. A vadose environment would have been estab- lished later when the region of the uplifting forebulge became sufficiently high (generally less than 200 m, according to CRAMPTON & ALLEN, 1995) to permit free downward circu- lation of water and the appearance of streams draining the ex- posed highlands. That period is coeval with exhumation of the SD formation at the crest of the forebulge and subsequent redis- tribution of the DF1 suite of elements (Cr and Ni in particular) in the proto-bauxite material in accordance with Eh/pH conditions controlled by the position of the groundwater table (D’ARGENIO & MINDSZENTY, 1995). Simultaneously, the diminishing trace element signal in the bauxites towards the PU and RU formations with reemerging significance of Ni relative to Cr may indicate a decrease in amplitude of the relief and also of the rate of erosion on the distal slope of the forebulge towards the back-bulge basin (SW Istria) (OTONIČAR, 2007) which restored phreatic condi- tions. Also, the diagenetic environment was likely to change dur- ing the early phase of burial, turning the initially vadose condi- tions into phreatic ones (D’ARGENIO & MINDSZENTY, 1995). The fact that Cretaceous formations older than Late Albian (PU), extending to the Valanginian, are devoid of ILPB deposits, is in- dicative of intermittent episodes of subaerial exposure in the Cre- taceous. They affected relatively thin carbonate succes sions in SW Istria due to recurrent brief emersions during the Early Cre- G eologia C roatica Zoran Peh and Erli Kovačević Galović: Geochemistry of Lower Palaeogene bauxites – a unique signature for the tectonostratigraphic evolution ... 277 taceous (MATIČEC et al., 1996), which prevented the formation of bauxite deposits. It is obvious that the Cr-Ni polarity relationship described above represents the keynote striking the answering chord in the whole set of processes related to bauxitization in the forebulge zone during its development. In this regard, another set of geo- chemical processes defined by the DF2 may elucidate the progres- sive changes and maturation of the proto-bauxitic material during the movement of the forebulge and the progressive formation of the unconformity. It is functional within the same matrix of tec- tonostratigraphic constraints explained by the DF1, exhibiting, however, a slight shift in time. Due to the reducing environment established in the early stages of the forebulge uplift (swampy lowlands across the emerged GH formation), the process of late- ritization (ferralitization) of the proto-bauxitic material could not have progressed very far while chalcophile elements were co- precipitated with Fe-sulfides or, as suggested from Fig. 5a-b (Fe2O3 opposing chalcophiles), occurring in the form of discrete metal sulfide phases (KORETSKY et al., 2006), with both poten- tially being adsorbed on clay minerals. Following exhumation of the SD formation, vadose conditions were established, which re- inforced the process of karstification and chemical reworking of the proto-bauxite. However, only with the emergence of the RU formation did the protobauxitic material achieve its full maturity with the removal of all mobile elements (Ni was removed earlier during the SD phase) and accumulation of Al, Fe, and Mn oxy- hydroxides. Finally, due to flexural subsidence in the distal parts of the forebulge, similar conditions were restored in regard to the PU formation. It was characterized by the reinstatement of the reducing environment, accommodating the same suite of com- ponents as in the initial stage of the uplift (Fig. 5a-b). 4.5. Integrating the models into the broader tectonostratigraphic scheme Subaerial exposure of the Istrian part of the ASWu and associated bauxitization probably lasted until the Early Eocene when plat- form conditions were restored burying the previously formed un- conformity under foraminiferal limestones (ĆOSOVIĆ et al., 2004). A similar forebulge effect in approximately the same pe- riod was also hypothesized in the central part of the High Karst (DSWu) where the underlying carbonate bedrock mostly refers to the GH formation (Coniacian to Santonian) (KORBAR, 2009). However, in this area the discontinuity surface may not exhibit clear diachronism which in Istria is plainly indicated by the con- trasting position of Cr and Ni in the bauxite geochemical signa- ture. In this sense the LPB lithofacies in a spatially confined geo- tectonic setting of the External Dinarides (NAI–CDA groups) cannot be squarely associated with a specific age interval between emersion and final drowning of both the ASWu and DSWu. It is particularly evident in the case of LPB groups outside Istria where bauxite deposits are regularly „sandwiched” between Santonian (Early Campanian) footwall and Early Eocene hanging wall car- bonate rocks betraying the apparent synchronicity along the entire K-Pg boundary in this part of the investigated area. The proposed models demonstrate that the prominent position of the SD baux- ites in Istria is brought about by the bulging effect manifested in the inverse Cr/Ni relationship (Fig. 5a-b) in a fashion previously described. However, a similar Cr/Ni status is missing in the LPB model where Ni stands relatively close to Cr (positive association) for both Dalmatian groups (NDA and CDA) (Fig. 5c-d) indica- ting the absence of apparent bulging in the NE part of the DSWu. In all probability, the close Cr/Ni relationship in the easternmost portion of the DSWu has been developed as a net feedback from slow uplift prohibiting effective draining and elimination of phreatic conditions which resulted in the distinctly gibbsitic chara- cter of the CDA proto-bauxite material during the entire emergent stage (KOVAČEVIĆ GALOVIĆ et al., 2012). Both rich in Cr, the aforementioned groups are, nonetheless, not quite close together with regard to their general geochemical signal, as is clearly dem- onstrated by Mahalanobis distances (MD) calculated for both models. MD values locate the CDA group to the far end of the LPB spatial sequence (Tab. 3) while the SD group is placed some- where in the „middle” of the ILPB temporal succession. Model comparisons based on MD values demonstrate that both promi- nent groups (SD and CDA) are located farthest from their respec- tive „reference points”, that is, the PU and IST groups, suggesting thereby the period of most intense Cr-enrichment as the correlat- ing event. When established, the general statistical reciprocity between the two bauxite lithofacies reveals (in light of the ergodic assumption), the space-time symmetry in the process of the profound geochemical change and reworking of the initial geo- chemical signal (input of Cr/Ni-rich material) across the entire ADCP, that is, both its Adriatic and Dinaric domains (cf. Fig. 5). It implies that the SD is closest in time while the CDA is closest in space to the supposed Cr-event (input of source material). How- ever, it is essential that the supposed event cannot be decoupled from geochemical processes affecting the Cr/Ni partition in pro- tobauxites during subaerial exposure of the karst environment. This means that the major contrast between the bauxite lithofacies from two ADCP domains, exposed through divergent geochemi- cal fates of the two dominant element pair, is defined primarily by the presence/absence of the forebulge dynamics and ensuing environmental control over the carbonate platform. Last but not least, the geochemical signature of both bauxite domains (ILPB and LPB), as communicated through the all-important DF1 in both CoDa DFMs, most probably indicates the temporal inequa- lity between the duration of subaerial exposure of Istrian GH and its High Karst platform-interior counterpart. 5. CONCLUSION The origin and emplacement of Lower Palaeogene bauxites on the Adriatic-Dinaric carbonate platform was explored in light of the recent advancements in the modeling of the foreland basin sys- tems. In the Istrian part of the platform (ASWu) bauxitization was coeval with expansion of the collision-induced diachronous dis- continuity surface, a process assumed to have been developed un- der strong tectonic control (uplift of a flexural bulge) during the initial stages of orogeny. To this purpose, a discriminant function model of the ILPB compositional data (whole rock geochemical data) was constructed beforehand, providing the solid ground for the separation of bauxite samples (based on lithostratigraphic units or formations) derived from the developing subaerial expo- sure of the Istrian Cretaceous/Palaeogene palaeoenvironment. At- tention was then paid towards finding distinctive geochemical signatures in the bauxites (LPB) from the neighbouring tectonos- tratigraphic unit (DSWu) in order to recognize the possible deve- lopment of forebulge unconformities elsewhere across the plat- form. A similar model of the entire ADCP (based on geo graphical deployment of the bauxite groups) was adopted from earlier in- vestigations and restructured after the principles of compositional data analysis, in order to assess the geochemical correspondence between the two event-markers (ILPB and LPB) at the same re- gional unconformity. The need for clarification of the nature of tectonostratigraphic constraints, especially those involved in fore- bulge dynamics, has occurred as a logical net effect associat ed G eo lo gi a C ro at ic a Geologia Croatica 69/2278 G eo lo gi a C ro at ic a with the proposed modeling of the bauxite lithofacies in terms of the spatial/temporal relationships. The most important results of these investigations can be summarized as follows: a) Both models heavily rely on the importance of the first discriminant function which comprises 77.2 and 81.5% of the to- tal variability in the ILPB and LPB models, respectively, albeit with different implications in terms of geochemical processes. In the ILPB model DF1, it is interpreted as reflecting the intricate environmentally-controlled Ni/Cr enrichment-depletion relation- ship whereas in the DF2 it is a supplementary process (17.1%) explained as manifesting the interchange between vadose and phreatic bauxite-forming environments due to changing redox conditions. Both functions, however, appeal to the effects of in- traplatform tectonic activity with the purpose of clarifying the specified geochemical controls. In the LPB model DF1 the spe- cific performance of Cr is established by its complementary po- sitioning with regards to bauxitic/clay components (primarily K2O) while the DF2 (11.1%) reflects enrichment (or reduction) in Al2O3 and related components typical for bauxites. b) The discontinuity surface underlying the ILPB is charac- terized by diachronism caused by the development of a flexural forebulge on the Istrian part of the ASWu. Forebulge evolution is recorded in the ILPB geochemical signature by assigning to the SD bauxite group a pivotal position in the significant geochemi- cal (Cr/Ni) partitioning after the initial input of source material. Bauxite-hosting SD and RU formations mark the maximum pos- sible cumulative uplift and erosion, facilitating maturation of the proto-bauxitic material under oxidizing, vadose conditions. In contrast, the PU group, formed on the distal flank of the forebulge (away from the orogenic wedge), shows evidence of the geochemi- cal signal characteristic of reducing conditions. In the case of the LPB, the North Adriatic IST and NAI groups are clearly sepa- rated from the Dalmatian NDA and CDA groups in such a man- ner that the latter participate as a repository for the chromium ingress from some extraneous source – most probably the Dinari- dic ophiolite belt). c) Comparison of the ILPB and LPB models via an ergodic approach brings into the foreground the pivotal status of the SD (ASWu) and CDA (DSWu) bauxite groups dictated by existing ADCP tectonostratigraphic constraints. In light of a supposed Cr–event it indicates both divergent geochemical fates of the most important descriptor variables (Cr and Ni) with respect to develo- pment of a flexural bulge, and the temporal inequality of the Is- trian GH formation (ASWu) and its High Karst counterpart (DSWu). ACKNOWLEDGMENT This study was supported by The Ministry of Science, Education and Sports, Republic of Croatia (MZOS) via the former Scientific Project (zProject): The Geological Maps of the Republic of Croatia. Their support is greatly appreciated. The authors wish to express their gratitude to the reviewers for their thoughtful suggestions and comments. Thanks are also due to all who have contributed to the execution of this research. REFERENCES AITCHISON, J. (1982): The statistical analysis of compositional data (with discussion).– J. Roy. Stat. Soc. Series B, 44, 139–177. AITCHISON, J. (1986): The Statistical Analysis of Compositional Data.– Chapman and Hall, London-New York, 416 p. doi:10.1002/bimj.4710300705 AITCHISON, J. (2008): The single principle of compositional data analysis, continuing fallacies, confusions and misunderstandings and some suggested remedies.– In: DAUNIS-i-ESTADELLA, J. & MARTÍN-FERNÁNDEZ, J.A. (eds.): Proceedings of CoDaWork’08, the 3rd Compositional Data Analysis Workshop, May 27–30, University of Girona, Girona (Spain), 1–28. AITCHISON, J. & EGOZCUE, J.J. (2005): Compositional data analysis: where are we and where should we be heading?– Math. Geol., 37, 829–850. doi:10.1007/s11004- 005-7383-7 ANALYTICAL METHODS COMMITTEE (1987): Recommendations for the defini- tion, estimation and use of the detection limit.– Analyst, 112, 199–204. doi:10.1039/ AN9871200199 ANTWEILER, R.C. & TAYLOR, H.E. (2008): Evaluation of statistical treatments of left-censored environmental data using coincident uncensored data sets: I. Sum- mary statistics.– Environ. Sci. Technol., 42, 3732–3738. doi:10.1021/es071301c. BALCO, G. & STONE, O.H.J. (2005): Measuring Pleistocene erosion rates with cos- mogenic nuclides in buried alluvial sediment, Fisher Valley, southeastern Utah.– Earth Surf. Proc. Land., 30, 1051–1067. doi: 10.1002/esp.1262. BARCELÓ-VIDAL, C., & PAWLOWSKY-GLAHN, V. (1999): Letter to the Editor: Comment on „Singularity and Nonnormality in the Classification of Composi- tional Data“ by BOHLING, G.C., DAVIS, J.C., OLEA, R.A. & HARFF, J.– Math. Geol., 31, 581–585. BLAŠKOVIĆ, I., DRAGIČEVIĆ, I. & POKRAJČIĆ, I. (1989): Tectonic control of the origin of the paleorelief of bauxite deposits in the Western Herzegovina, Yugosla- via.– Travaux ICSOBA, 19/22, 231–238. BRLEK, M., KORBAR, T., CVETKO TEŠOVIĆ, B., GLUMAC, B. & FUČEK, L. (2013): Stratigraphic framework, discontinuity surfaces and regional significance of Campanian slope to ramp carbonates from central Dalmatia, Croatia.– Facies, 59, 779–801. doi: 10.1007/s10347-012-0342-0 BRLEK, M., KORBAR, T., KOŠIR, A., GLUMAC, B., GRIZELJ, A. & OTONIČAR, B. (2014): Discontinuity surfaces in Upper Cretaceous to Paleogene carbonates of central Dalmatia (Croatia): Glossifungites ichnofacies, biogenic calcretes, and stratigraphic implications.– Facies, 60, 467–487. doi:10.1007/s10347-013-0378-9. BUCCIANTI, A. (2013): Is compositional data analysis a way to see beyond the illu- sion?– Comp. Geosci., 50, 165–173. doi:10.1016/j.cageo.2012.06.012. BUCCIANTI, A., MATEU-FIGUERAS, G. & PAWLOWSKY-GLAHN, V. (eds.) (2006): Compositional Data Analysis in the Geosciences: From Theory to Prac- tice.– Geol. Soc. Spec. Publ., The Geological Society London, 264, 224 p. CRAMPTON, S.L. & ALLEN, P.A. (1995): Recognition of forebulge unconformities associated with early stage foreland basin development; example from the North Alpine foreland basin.– AAPG Bull., 79, 1495–1514. CVETKO TEŠOVIĆ. B., GUŠIĆ, I., JELASKA, V. & BUCKOVIĆ, D. (2001): Strati- graphy and microfacies of the Upper Cretaceous Pućišća Formation, Island of Brač, Croatia. Cretaceous Res., 22, 591–613. doi: 10.1006/cres.2001.0279. CVETKO TEŠOVIĆ, B., GLUMAC, B. & BUCKOVIĆ, D. (2011): Integrated bios- tratigraphy and carbon isotope stratigraphy of the Lower Cretaceous (Barremian to Albian) Adriatic-Dinaridic carbonate platform deposits in Istria, Croatia.– Cre- taceous Res., 32, 301–324. doi:10.1016/j.cretres.2010.12.011. ĆOSOVIĆ, V., DROBNE, K. & MORO, A. (2004): Paleoenvironmental model for Eocene foraminiferal limestones of the Adriatic carbonate platform (Istrian Penin- sula).– Facies, 50, 61–75. doi:10.1007/s10347-004-0006-9. D’ARGENIO, B. & MINDSZENTY, A. (1995): Bauxites and related palaeokarst: tec- tonic and climatic event markers at regional unconformities.– Eclogae Geol. Helv., 88, 453–499. DAUNIS-i-ESTADELLA, J., THIÓ-HENESTROSA, S. & MATEU-FIGUERAS, G. (2011): Two more things about compositional biplots: quality of projection and inclusion of supplementary elements.– In: EGOZCUE, J.J., TOLOSANA-DEL- GADO, R. & ORTEGO, M.I. (eds.): Proceedings of the 4th International Workshop on Compositional Data Analysis (CoDaWork’11), Girona, 1–14. DeCELLES, P.G. & GILES, K.A. (1996): Foreland basin systems.– Basin Res., 8, 105– 123. DURN, G., OTTNER, F., TIŠLJAR, J., MINDSZENTY, A. & BARUDŽIJA, U (2003): Regional Subaerial Unconformities in Shallow-Marine Carbonate Sequences of Istria: Sedimentology, Mineralogy, Geochemistry and Micromorphoplogy of As- sociated Bauxites, Paleosols and Pedo-Sedimentary Complexes. Field Trip P8, 209-254.– In: VLAHOVIĆ, I. & TIŠLJAR, J. (eds): Field Trip Guidebook of 22nd IAS Meeting of Sedimentology, Opatija, 256 p. EGOZCUE, J.J. & PAWLOWSKY-GLAHN, V. (2006): Simplicial geometry for com- positional data.– In: BUCCIANTI, A., MATEU-FIGUERAS, G. & PAW- LOWSKY-GLAHN, V. (eds): Compositional Data Analysis in the Geosciences: From Theory to Practice. London: Geol. Soc. Spec. Publ., 264, 145–158. GALOVIĆ, L. & PEH, Z. (2016): Mineralogical discrimination of the pleistocene loess/ paleosol sections in Srijem and Baranja, Croatia.– Aeolian Res., in press. doi: 10.1016/j.aeolia.2016.04.006 GUŠIĆ, I. & JELASKA, V. (1990): Stratigrafija gornjokrednih naslaga otoka Brača u okviru geodinamske evolucije Jadranske karbonatne platforme (Upper Cretaceous stratigraphy of the Island of Brač within the geodynamic evolution of the Adriatic carbonate platform).– Djela Jugoslavenske akademije znanosti i umjetnosti, 69, Institut za geološka istraživanja, Zagreb, 160 p. G eologia C roatica Zoran Peh and Erli Kovačević Galović: Geochemistry of Lower Palaeogene bauxites – a unique signature for the tectonostratigraphic evolution ... 279 HINSBERGEN, D.J.J., MENSINK, M., LANGEREIS, C.G., MAFFIONE, M., SPALLUTO, L., TROPEANO, M. & SABATO, L. (2014): Did Adria rotate rela- tive to Africa?– Solid Earth, 5, 611–629. doi:10.5194/se-5-611-2014 HRVATSKI GEOLOŠKI INSTITUT (CROATIAN GEOLOGICAL SURVEY) (2009): Geološka karta Republike Hrvatske 1:300000 (Geological Map of the Republic of Croatia 1:300000), Zagreb. KORBAR, T. (2009): Orogenic evolution of the external Dinarides in the NE Adriatic re- gion; a model constrained by tectonostratigraphy of Upper Cretaceous to Paleogene carbonates.– Earth-Sci. Rev., 96, 296–312. doi: 10.1016/j.earscirev.2009.07.004. KORBAR, T., GLUMAC, B., CVETKO TEŠOVIĆ, B. & CADIEUX, S.B. (2012): Response of a Carbonate Platform to the Cenomanian–Turonian Drowning and OAE 2: a Case Study from the Adriatic Platform (Dalmatia, Croatia).– J. Sediment. Res., 82, 163–176. doi: 10.2110/jsr.2012/17 KORETSKY, C.M., HAAS, J.R., MILLER, D, & NDENGA, N.T. (2006): Seasonal variations in pore water and sediment geochemistry of littoral lake sediments (Asy- lum Lake, MI, USA).– Geochemical T., 7/11, 1–26. doi:10.1186/1467-4866-7-11 KOVAČEVIĆ GALOVIĆ, E., ILIJANIĆ, N., PEH, Z., MIKO, S. & HASAN, O. (2012): Geochemical discrimination of Early Palaeogene bauxites in Croatia.– Geol. Croat, 65, 53–65. doi: 104154/gc.2012.04 MÁRTON, E., DROBNE, K., ĆOSOVIĆ, V. & MORO, A. (2003): Palaeomagnetic evi- dence for Tertiary counterclockwise rotation of Adria.– Tectonophysics, 377, 143–156. doi:10.1016/j.tecto.2003.08.022 MÁRTON, E., ZAMPIERI, D., GRANDESSO, P., ĆOSOVIĆ, V. & MORO, A. (2010): New Cretaceous paleomagnetic results from the foreland of the Southern Alps and the refined apparent polar wander path for stable Adria.– Tectonophysics, 480, 57–72. doi:10.1016/j.tecto.2009.09.003 MÁRTON, E., ZAMPIERI, D., KÁZMÉR, M., DUNKL, I. & FRISCH, W. (2011): New Paleocene–Eocene paleomagnetic results from the foreland of the Southern Alps confirm decoupling of stable Adria from the African plate.– Tectonophysics, 504, 89–99. doi:10.1016/j.tecto.2011.03.006. MATIČEC, D. (1994): Neotectonic deformations in Western Istria, Croatia.– Geol. Croat., 47/2, 199–204. MATIČEC, D., VLAHOVIĆ, I., VELIĆ, I. & TIŠLJAR, J. (1996): Eocene limestones overlying Lower Cretaceous deposits of western Istria (Croatia): did some parts of present Istria form land during the Cretaceous?– Geol. Croat., 49, 117–127. MATIČEC, D., FUČEK, L., PALENIK, D. & OŠTRIĆ, N. (2013): Geological setting of the Istrian County.– In: MIKO, S., KRUK, B., DEDIĆ, Ž., KRUK, Lj., PEH, Z., KOVAČEVIĆ GALOVIĆ, E. & GABRIĆ, A. (eds.): Mining-Geological Study of Potentiality and Mineral Resources Management in the Area of Istrian County.– Croatian Geological Survey, Zagreb, Croatia, 19–48. (In Croatian). MINDSZENTY, A., D’ARGENIO, B. & AIELLO, G. (1995): Lithospheric bulges re- corded by regional unconformities. The case of Mesozoic – Tertiaty Apulia.– Tec- tonophysics, 252, 137–161. MINDSZENTY, A., CSOMA, A., TÖRÖK, Á., HIPS, K. & HERTELENDI, E. (2000): Flexura jellegű előtéri deformációhoz köthető karsztbauxitszintek a Dunántúli- középhegységben (Rudistid limestones, bauxites, paleokarst and geodynamics. The case of the Cretaceous of the Transdanubian Range).– Bull. Geol. Soc. Hung., 131/1–2, 107–152 (in Hungarian with ext. English abstract). NOCQUET, J.-M. (2012): Present-day kinematics of the Mediterranean: A comprehen- sive overview of GPS results.– Tectonophysics, 579, 220–242. doi: 10.1016/j. tecto.2012.03.037. OTONIČAR, B. (2007): Upper Cretaceous to Paleogene forbulge unconformity associ- ated with foreland basin evolution (Kras, Matarsko podolje and Istria; SW Slov- enia and NW Croatia).– Acta Carsologica, 36/1, 101–120. OTONIČAR, B. (2008): Evolution of paleokarst related to a forebulge unconformity; an example from the late Cretaceous and Paleogene of Kras, Matarsko Podolje and Istria, southwest Slovenia and northwest Croatia.– In: SASOWSKY, I.D., FEA- ZEL, C.T., MYLROIE, J.E., PALMER, A.N. & PALMER, M.V (eds): Karst from Recent to Reservoirs, Karst Waters Institute Special Publication 14, Leesburg, Vir- ginia, 221 p. PAINE, A.D.M. (1985): ‘Ergodic’ reasoning in geomorphology: time for a review of the term? – Prog. Phys. Geog., 9, 1–15. doi: 10.1177/030913338500900101. PAMIĆ, J, BALEN, D. & HERAK, M. (2002): Origin and geodynamic evolution of Late Paleogene magmatic associations along the Periadriatic–Sava–Vardar magmatic belt.– Geodin. Acta, 15, 209–231. PAWLOWSKY-GLAHN, V. & EGOZCUE, J.J. (2006): Compositional data and their analysis: an introduction.– In: BUCCIANTI A, MATEU-FIGUERAS, G. & PAW- LOWSKY-GLAHN, V. (eds): Compositional Data Analysis in the Geosciences: From Theory to Practice. Geol. Soc. Spec. Publ., 264, 1–10. DOI: 10.1144/GSL. SP.2006.264.01.01 PEH, Z. & HALAMIĆ, J. (2010): Discriminant function model as a tool for classifica- tion of stratigraphically undefined radiolarian cherts in ophiolite zones.– Journal. Geochem. Explor., 107, 30–38. DOI: 10.1016/j.gexplo.2010.06.003. PEH, Z. & KOVAČEVIĆ GALOVIĆ, E. (2014): Geochemistry of Istrian Lower Palaeogene bauxites – Is it relevant to the extent of subaerial exposure during Cretaceous times?– Ore Geol Rev., 63, 296–306. PRELOGOVIĆ, E., KUK, V., JAMIČIĆ, D., ALJINOVIĆ, B. & MARIĆ, K. (1995): Seizmotectonic activity of the Kvarner area (in Croatian).– In: VLAHOVIĆ, I., VELIĆ, I. & ŠPARICA, M. (eds.): Proceedings of the First Croatian Geological Congress, Vol. 2., Croatian Geological Society and Institute of Geology, Zagreb, 487–490. PRELOGOVIĆ, E., PRIBIČEVIĆ, B., IVKOVIĆ, Ž., DRAGIČEVIĆ, I., BULJAN, R. & TOMLJENOVIĆ, B. (2003): Recent structural fabric of the Dinarides and tec- tonically active zones important for petroleum-geological exploration in Croatia.– Nafta, 55, 155–161. REIMANN, C., FILZMOSER, P., FABIAN, K., HRON, K., BIRKE, M., DEMETRIA- DES, A., DINELLI, E. & LADENBERGER, A. & The GEMAS Project Team (2012): The concept of compositional data analysis in practice – Total major ele- ment concentrations in agricultural and grazing land soils of Europe.– Sci Total. Environ., 426, 196–210. doi: 10.1016/j.scitotenv.2012.02.032. SAKAČ, K. & ŠINKOVEC, B. (1991): The Bauxites of the Dinarides. Travaux ICSOBA, 20–21, 1–11. SCHMID, S. M., BERNOULLI, D., FÜGENSCHUH, B., MATENCO, L., SCHEFER, S., SCHUSTER, R., TISCHLER, M. & USTASZEWSKI, K. (2008): The Alpine- Carpathian-Dinaridic orogenic system: correlation and evolution of tectonic units.– Swiss J Geosci, 101, 139–183. SCISCIANI, V. & CALAMITA, F. (2009): Active intraplate deformation within Adria: Examples from the Adriatic region.– Tectonophysics, 476, 57–72. doi: 10.1016/j. tecto.2008.10.030. StatSoft, Inc. (2006): STATISTICA (Data Analysis Software System), Releaase 7.1; www.statsoft.com2006 STEUBER, T., KORBAR, T., JELASKA, V. & GUŠIĆ, I. (2005): Strontium-isotope stratigraphy of Upper Cretaceous platform carbonates of the island of Brač (Adri- atic Sea, Croatia): implications for global correlation of platform evolution and biostratigraphy.– Cretaceous Res., 26/5, 741–756. ŠINKOVEC, B. (1973): The origin of Early Palaeogene bauxites of Istria, Yugoslavia.– Travaux ICSOBA, 3, 151–164. ŠINKOVEC, B. & SAKAČ, K. (1982): The Paleogene bauxites of Dalmatia.– Travaux ICSOBA, 12, 293–331. ŠINKOVEC, B. & SAKAČ, K. (1991): Bauxite deposits of Yugoslavia – the state of the Art.– Acta Geol. Hung., 34/4, 307–315. ŠUMANOVAC, F. (2010): Lithosphere structure at the contact of the Adriatic microplate and the Pannonian segment based on the gravity modelling.– Tectonophysics, 485, 94–106. doi:10.1016/j.tecto.2009.12.005 TARVAINEN, T., REEDER, S. & ALBANESE, S. (2005): Database management and map production.– In: SALMINEN, R., BATISTA, M.J., BIDOVEC, M., DEME- TRIADES, A., De VIVO, B., De VOS, W., DURIS, M., GILUCIS, A., GRE- GORAUSKIENE, V., HALAMIĆ, J., HEITZMANN, P., JORDAN, G., KLAVER, G., KLEIN, P., LIS, J., LOCUTURA, J., MARSINA, K., MAZREKU, A., O’CONNOR, P.J., OLSSON, S.Å., OTTESEN, R.-T., PETERSELL, V., PLANT, J,A,, REEDER, S,, SALPETEUR, I,, SANDSTRÖM, H., SIEWERS, U., STEEN- FELT, A. & TARVAINEN, T. (eds): Geochemical Atlas of Europe, Part 1, Back- ground Information, Methodology and Maps. Espoo: Geological Survey of Fin- land, 1–4. TIŠLJAR, J., VLAHOVIĆ, I., VELIĆ, I., MATIČEC, D. & ROBSON, J. (1998): Car- bonate facies evolution from the Late Albian to Middle Cenomanian in southern Istria (Croatia): influence of synsedimentary tectonics and extensive organic car- bonate production.– Facies, 38, 37–151. TOLOSANA-DELGADO, R., OTERO, N. & PAWLOWSKY-GLAHN, V. (2005): Some Basic Concepts of Compositional Geometry.– Math. Geol., 37, 673–680. doi: 10.1007/s11004-005-7374-8. TOLOSANA-DELGADO, R. (2012): Uses and misuses of compositional data in sedi- mentology.– Sediment. Geol., 280, 60–79. doi: 10.1016/j.sedgeo.2012.05.005. USTASZEWSKI, K., KOUNOV, A., SCHMID, S.M., SCHALTEGGER, U., KRENN, E., FRANK, W. & FÜGENSCHUH, B. (2010): Evolution of the Adria-Europe plate boundary in the northern Dinarides: From continent-continent collision to back-arc extension, Tectonics, 29, TC6017. doi:10.1029/2010TC002668 VELIĆ, I., TIŠLJAR, J., MATIČEC, D. & VLAHOVIĆ, I. (1995): A review of the Ge- ology of Istria.– In: VLAHOVIĆ, I. & VELIĆ, I. (eds): Excursion Guidebook, First Croatian Geological Congress, Croatian Geological Society and Institute of Geology, 21–30. VLAHOVIĆ, I. (1999): Carbonate facies of shallow water depositional systems from Kimmeridgian to Late Albian in western Istria:– Unpubl. PhD thesis (in Croatian with English summary).– University of Zagreb, 327 p. VLAHOVIĆ, I., TIŠLJAR, J., VELIĆ, I. & MATIČEC, D. (2005): Evolution of the Adri- atic carbonate platform: palaeogeography, main events and depositional dyna- mics.– Palaeogeogr. Palaeocl., 220, 333–360. doi: 10.1016/j.palaeo.2005.01.011. YANG, Y. (2011): Tectonically-driven underfilled–overfilled cycles, the middle Creta- ceous in the northern Cordilleran foreland basin.– Sediment. Geol., 233, 15–27. doi: 10.1016/j.sedgeo.2010.10.002.