Open Geosci. 2017; 9:340–360

Research Article Open Access

Christopher Baiyegunhi*, Kuiwu Liu, and Oswald Gwavava Geochemistry of and shales from the , , in the Eastern Cape Province of : Implications for provenance, weathering and tectonic setting https://doi.org/10.1515/geo-2017-0028 Keywords: Geochemistry, provenance, weathering, tec- Received November 16, 2016; accepted April 11, 2017 tonic setting, Ecca Group Abstract: Geochemical compositions of twenty-four sand- stone and shale samples from the Ecca Group were anal- ysed to decipher their provenance, paleoweathering con- 1 Introduction ditions and tectonic setting. The shales have high Fe2O3, K2O, TiO2, Ce, Cu, Ga, La, Nb, Nd, Rb, Sc, Sr, Th and Y Clastic sedimentary rocks have vital information about the content more than the sandstones, whereas, sandstones composition, tectonic setting and evolution of continental are higher in SiO2, Hf and Zr than the shales. The posi- crust, mainly when the traditional petrographic methods tive correlations of Al2O3 with other elements as well as are unclear. Nonetheless, their chemical and mineralogi- the abundance of Ba, Ce, Th, Rb, Zn and Zr suggest that cal composition can be influenced by factors like source these elements are primarily controlled by the dominant rock characteristics, weathering, sorting processes during clay minerals. Tectonic discrimination diagrams revealed transportation, sedimentation and diagenetic processes to that the sandstones and shales are mostly of quartzose an extent [1–3]. Trace elements such as La, Y, Sc, Cr, Th, Zr, sedimentary provenance, suggesting that they were de- Hf, Nb and rare earth elements (REE) are thought to be use- rived from a cratonic interior or recycled orogen. The bi- ful indicators of provenance, geological processes and tec- nary plots of TiO2versus Ni, TiO2 against Zr and La/Th ver- tonic setting due to their relatively low mobility and insolu- sus Hf as well as the ternary diagrams of V-Ni-Th*10 in- bility during sedimentary processes [4–8]. Hence, the geo- dicate that the shales and sandstones were derived from chemistry of clastic sediments (i.e. and shale) felsic igneous rocks. A-CN-K (Al2O3-CaO-K2O) ternary dia- reflects a combination of provenance, chemical weather- gram and indices of weathering (CIA, CIW and PIS) suggest ing, hydraulic sorting, and abrasion [2, 4, 8–11]. that the granitic source rocks underwent moderate to high In geochemical provenance studies, fine grained sed- degree of chemical weathering. The CIA values range be- imentary rocks like shales are considered to be the most tween 24.41% and 83.76%, indicating low to high weather- useful rock because of their homogeneity before deposi- ing conditions. The CIW values for the studied sandstones tion, post-depositional impermeability and higher abun- and shales range from 25.90 to 96.25%, suggesting mod- dance of trace elements [4, 10, 12, 13]. Some relatively im- erate to high intensive chemical weathering. ICV values mobile elements like Sc, Th, Zr, Hf and rare earth ele- for the sandstones and shales vary from 0.71 to 3.6 (aver- ments (REE) show very low concentrations in natural wa- aging 1.20) and 0.41 to 1.05 (averaging 0.82), respectively. ters and are transported almost quantitatively throughout The K2O/Na2O ratios for the studied samples vary from the sedimentary process from parent rocks to clastic sed- 0.71 to 8.29, which reveal moderate to high maturity. The plot of CIA against ICV shows that most of the shales are geochemically mature and were derived from both weak *Corresponding Author: Christopher Baiyegunhi: Department of and intensively weathered source rocks. The tectonic set- Geology, Faculty of Science and Agriculture, University of Fort Hare, ting discrimination diagrams support passive-active con- Private Bag X1314, Alice, 5700, Eastern Cape Province, South Africa, tinental margin setting of the provenance. E-mail: [email protected], [email protected] Kuiwu Liu, Oswald Gwavava: Department of Geology, Faculty of Science and Agriculture, University of Fort Hare, Private Bag X1314, Alice, 5700, Eastern Cape Province, South Africa

Open Access. © 2017 Christopher Baiyegunhi et al., published by De Gruyter Open. This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivs 3.0 License. Geochemistry of sandstones and shales from the Ecca Group Ë 341 iments [4, 14]. The relative distribution or enrichment of and sandstones in the study area in order to provide infor- these immobile elements in felsic and basic rocks have mation on the source rock characteristics, provenance, pa- been used to infer the relative contribution of felsic and leoweathering and tectonic setting using their major, trace basic sources in shales from different tectonic environ- and rare earth elements geochemistry. ments [15]. For example, La and Th are enriched in felsic rocks, whereas Sc, Cr, and Co are more concentrated in ba- sic rocks relative to felsic rocks. These elements are rela- 2 General geology and stratigraphy tively immobile during weathering [6, 16–18]. Hence, the ratios of La or Th to Co, Sc, or Cr are sensitive indicators of The word “Karoo” was derived from the Main Karoo Basin source rock compositions. Similarly, felsic igneous rocks of South Africa to describe sedimentary fill of all basins contain negative Eu anomalies (Eu/Eu* from chondrite- of akin age across Gondwana. The Main Karoo Basin of normalized plots of the REE), whereas basic igneous rocks South Africa is bordered in the southern part by a fold have little or no Eu anomalies, and the size of the negative belt (), while the northern part is held anomalies in the provenance seems to be preserved in fine- by Achaean [33]. It is a unique type of grained sediment [18]. Furthermore, some major elements basin of all the Karoo basins in southern Africa because such as alkali and alkali earth elements, which are water it contains the thickest and stratigraphically most com- mobile elements and very sensitive to climatic change, can plete mega-sequence of several depositories of the Permo- be used as a proxy of paleoclimate evolution [19, 20]. Geo- Carboniferous to Jurassic age sediments in southwestern chemical data on sandstones of unmetamorphosed sedi- Gondwana continent [33]. In addition, it serves as a datum mentary sequences deposited in epicratonic or intracra- for classifying Karoo basins in central and southern Africa. tonic basins also give important clues on paleoweathering The bulk of the Karoo strata occur in the Main Karoo Basin, conditions, variations in provenance composition and tec- with maximum preserved thickness adjacent to the Cape tonic settings [21–23]. Fold Belt in excess of 6 km [34]. The basin covers up to Several researchers such as [24] and [25] have pro- 700, 000 km2 and represents about 110 Ma of sedimenta- posed K O/Na O versus SiO tectonic setting discrimina- 2 2 2 tion spanning from 290 Ma to 180 Ma, and has its rocks cov- tion diagrams for sedimentary rocks in order to identify ering almost half of the area of South Africa [35]. The sed- tectonic setting of unknown basins. These diagrams are imentary succession reflects changing environments from still commonly used to deduce the tectonic setting of an- glacial to marine, deltaic, fluvial and finally aeolian [36]. cient basins. However, more detailed results can be ob- The Karoo Supergroup is believed to have originated tained using the calc-alkaline oxide ternary diagram (CaO- from the Gondwana Supercontinent [33]. The several litho- Na O-K O) of [26] and modified by [27]. [19, 28] docu- 2 2 spheric plates that separated to form the current Southern mented that the index of compositional variability (ICV), Hemisphere continents and India was once together as the K O/Al O ratio, chemical index of alteration (CIA) and 2 2 3 Gondwana supercontinent. The southern African remnant Al O -(CaO+Na O)-K O (A-CN-K) ternary plots are useful 2 3 2 2 of these continents contains the Karoo basins that include geochemical parameters for the study of provenance and the Main Karoo Basin and Great Kalahari Basin (Kalahari maturity of the rocks. In addition, [25, 29, 30] used al- Karoo, Aranos, and Mid-Zambezi Basins), as well as other kali metal oxides to reveal information about the prove- smaller basins in South Africa, Namibia, Zimbabwe and nance of clastic sediments. Recent geochemical investiga- Mozambique. The Karoo Supergroup has evolved from two tions on sandstones and shales have focused on decipher- distinct tectonic regimes sourced from the southern and ing the provenance and tectonic evolution of sedimentary the northern margin of Gondwana [37]. The southern tec- basins [22, 31, 32]. However, little attention has been paid tonic regimes are believed to be related to processes of sub- to the study of provenance in this region, despite the fact duction and orogenesis along the Panthalassan (paleo- that the targeted carbonaceous shales for shale gas explo- Pacific) margin of Gondwana, this resulted in the forma- ration in the region is hosted in the Ecca Group. Further- tion of a retro-arc foreland system known as the “Main more, it is also important in understanding the tectonic Karoo Basin” in association with the primary subsidence evolution of the southeastern Karoo basin. To date, the mechanism represented by flexural and dynamic load- geochemistry of fine-grained rocks of the Ecca Group inthe ing. The northern event was associated with extensional Eastern Cape Province of South Africa has not been stud- stresses that propagated southwards into the supercon- ied in detail to determine their source rock characteristics, tinent from the divergent Tethyan margin of Gondwana. provenance and tectonic setting. This study was therefore The Karoo Supergroup is a 12 km thick sequence of sedi- aimed at evaluating the geochemistry of the Ecca shales 342 Ë C. Baiyegunhi et al. mentary rocks that was deposited in a large intracratonic the following equation: retro-arc foreland basin in south-western Gondwana [38]. WeightA − WeightB Generally, the Karoo Supergroup is subdivided into five %LOI= × 100 WeightA − Weightcrucible main groups, namely; the Dwyka (Late Carboniferous), Several discriminatory plots of the major and trace el- Ecca (Late Carboniferous-Early Permian), Beaufort (Late ements were used to determine the provenance and tec- Permian-Middle Triassic), Stormberg (Late Triassic-Early tonic setting. The ternary plot of Al O -(CaO+Na O)-K O Jurassic) and Groups (Middle Jurassic) [39]; 2 3 2 2 (represented as A-CN-K) as well as the formulas for chem- Figure 1. The Drakensberg lavas are believed to have ter- ical index of alteration (CIA), chemical index of weather- minated sedimentation in the basin in the Middle Juras- ing (CIW) and plagioclase index of alteration (PIA) were sic [36]. used to quantify the degree of weathering. In the formu- The areal distributions of lithostratigraphic units in las, CaO* is the amount of CaO incorporated in the silicate the Main Karoo Basin is depicted in Figure 1 and Table 1. fraction of the rock. In this study, correction for CaO from The term “Ecca” for argillaceous sedimentary strata ex- carbonate contribution was not performed due to the ab- posed in the Ecca Pass, near Grahamstown in the East- sence of CO value. Thus, to compute for CaO* from the ern Cape Province, South Africa [37]. Stratigraphically, the 2 silicate fraction, the assumption proposed by [57] was em- Ecca Group in the study area can be subdivided into five ployed. Based on this, CaO values were accepted only if formations, namely, the Prince Albert, Whitehill, Colling- CaONa O, it was presumed that ham, Ripon, and Fort Brown Formations [40–43]; Fig- 2 2 the concentration of CaO is the same with that of Na O. ure ??]. It is estimated that the group attained a thickness 2 This procedure provides measure for the ratio of the sec- of about 3000 m in the southern part of the basin, while ondary aluminous mineral to feldspar, and forms a basis it is considerably thinner elsewhere in the northern part for the measure of intensity of weathering. Index composi- of the basin [44]. More details on the geology, stratigra- tional variation (ICV) proposed by [58] was used to deter- phy and tectonic setting of the Karoo Basin can be found mine maturity of the sediments. in [45–54].

3 Methodology 4 Results

Twenty-four rock samples of the shale and sandstone rep- 4.1 Major elements resenting various formations of the Ecca Group were anal- The major element compositions are quite variable but still ysed for the major and trace element (including rare earth comparable with the average compositions documented element) concentrations. X-ray Fluorescence (XRF) analy- by [59–61] (Table 2). All the samples show high concentra- sis was performed at the Council for Geoscience in Preto- tion of SiO , ranging from 55.94 to 87.99%. The Al O , CaO ria, South Africa. XRF analysis of the major and trace ele- 2 2 3 and Fe O contents are moderately high, ranging from 5.91 ments geochemistry of the samples was performed using 2 3 to 16.10%, 0.06 to 16.67% and 0.47 to 6.88%, respectively. MagiX Fast, XRF spectrometer. Bulk sample preparation The concentrations of TiO , MnO, MgO, Na O, K O and consists of drying where necessary, crushing to 10 mm, 2 2 2 P O are generally low, ranging from 0.13 to 0.80%, 0.004 splitting and milling in a tungsten carbide milling pot to 2 5 to 0.197%, 0.25 to 1.58%, 0.17 to 3.27%, 0.95 to 4.42% and less than 75 µm. Major element analyses were analysed 0.053 to 0.219%, respectively. The sandstones are higher on fused beads, while trace elements were executed on in SiO content more than the shales. On the other hand, pressed powder pellets. Major elements were analysed fol- 2 shales are higher in Fe O ,K O and TiO contents more lowing the procedure documented by [55], while the con- 2 3 2 2 than the sandstones, which reflect their association with centration of trace and rare earth elements were deter- clay-sized phases [62]. mined by ICP-MS (Inductively Coupled Plasma Mass Spec- The abundance of Al O was used as a normalization trometry) following the procedure described by [56]. Accu- 2 3 factor to make comparisons among the different litholo- racy and precision was estimated and monitored from the gies because of their immobile nature during weather- control samples and duplicates. The analytical precision is ing, diagenesis and metamorphism [63]. Major oxides of better than 5% for major and trace elements. Samples were the studied shales and sandstones were plotted against dried at 100 °C (weight A) and heated at 1000 °C (weight B), Al O as depicted in Figures 2 and 3. In addition, average to determine the percentage loss on ignition (LOI) using 2 3 UCC (Upper Continental Crust) and PAAS (Post-Archaean Geochemistry of sandstones and shales from the Ecca Group Ë 343

Figure 1: Areal distributions of lithostratigraphic units in the Main Karoo Basin (after [33, 40]). 344 Ë C. Baiyegunhi et al.

Table 1: Lithostratigraphy of the Karoo Supergroup in the study area [35]. Geochemistry of sandstones and shales from the Ecca Group Ë 345

Table 2: Comparing average chemical composition of the sandstones and shales from the Ecca Group with published average shales.

Oxides This study Average shale [59] Average shale [61] UCC PAAS NASC [60] Shale Sandstone

SiO2 (%) 68.72 76.43 58.10 58.50 66.60 62.40 64.82 TiO2 (%) 0.58 0.46 0.60 0.77 0.64 0.99 0.80 Al2O3 (%) 14.26 9.02 15.40 15.00 15.40 18.78 17.05 Fe2O3 (%) 4.30 3.61 6.90 4.72 5.04 7.18 5.70 MnO (%) 0.06 0.07 Trace – 0.10 0.11 – MgO (%) 1.28 0.59 2.40 2.50 2.48 2.19 2.83 CaO (%) 1.57 2.65 3.10 3.10 3.59 1.29 3.51

Na2O (%) 1.63 1.40 1.30 1.30 3.27 1.19 1.13 K2O (%) 3.08 1.57 3.20 3.10 2.80 3.68 3.97 P2O5 (%) 0.17 0.09 0.20 0.16 0.12 0.16 0.15

Australian Shale) values were extracted from [64] and CaO, Na2O, K2O and P2O5 are 1.16, 0.52, 0.62, 0.55, 0.59, [4], respectively and included in the plots for compari- 1.64, 1.27, 0.63 and 0.80, respectively, which are generally son purposes. In the shale samples, major elements like comparable with PAAS. Conversely, the average concen- TiO2, Fe2O3, MgO and K2O shows positive correlation with tration of MgO relative to PAAS is low, averaging 0.43. Al2O3, whereas MnO, CaO, Na2O and P2O5 shows no par- ticular trend (Figures 2). Similarly, in the sandstone sam- ples, TiO2, Fe2O3, MgO, K2O and P2O5 shows positive 4.2 Trace Elements correlation with Al2O3, while TiO2, MnO, CaO and Na2O shows no particular trend (Figure 3). The strong positive The trace element compositions are quite variable but still correlation of these major oxides with Al2O3 indicates that comparable with the average compositions documented they are associated with micaceous/clay minerals. by [60] and [66]. In the sandstones and shales, the con- The shale and sandstone samples were normalized tents of large ion lithophile elements (LILE) like Rb, Ba, to UCC [64] and PAAS [4] as depicted in Figures 4. Rel- Sr and Th vary from 39 to 197 ppm, 208 to 909 ppm, 23 ative to UCC, the average concentrations of SiO2, TiO2, to 340 ppm and 2.8 to 23 ppm, respectively. The content of Al2O3, Fe2O3, MnO, CaO, K2O and P2O5 in the sandstones high field strength elements (HFSE) like Zr, Y and Nb range are 1.09, 0.79, 0.73, 0.71, 0.65, 0.59, 0.83 and 1.07, respec- from 108 to 652 ppm, 9.1 to 35 ppm and 5 to 17 ppm, respec- tively, which are generally comparable with the UCC. On tively. Similarly, transition trace elements (TTE) like Sc, V, the other hand, the average concentration of Na2O and Cr, Ni and Zn range from 3 to 16 ppm, 6.4 to 96 ppm, 4.8 MgO relative to UCC are 0.36 and 0.37, respectively, which to 85 ppm, 2.1 to 31 ppm and 1.5 to 121 ppm, respectively. are generally low as compared to UCC. The depletion of Generally, shales have high Ce, Cu, Ga, La, Nb, Nd, Rb, Na2O(< 1%) in the Ecca rocks can be attributed to a rela- Sc, Sr, Th and Y content than the sandstones. On the other tively smaller amount of Na-rich plagioclase in them. K2O hand, sandstones are higher in Hf and Zr than the shales. and Na2O contents and their ratios (K2O/Na2O > 1) re- The high Sr content in shales indicate that Sr may be as- vealed that K-feldspar dominates over plagioclase (albite) sociated with calcite minerals. Based on the LILE average feldspar. K2O enrichment relates to the presence of illite values, except for Sr and Ba, almost all the shale samples as common clay mineral in the shales and sandstones. In exhibit similar LILE abundances relative to UCC and PAAS addition, the enrichment of CaO (averaging 0.59 relative to (Figures 5a and 5b). UCC) can be attributed to the presence of diagenetic calcite In contrast, the sandstones exhibit similar Th and U cement. Relative to UCC, the shales are low in MgO, CaO, contents relative to UCC and PAAS but are depleted in Rb, Na2O, MnO and high in SiO2, TiO2 and Al2O3. Al and Ti are Ba and Sr (Figure 5c and 5d). Th has very strong posi- easily absorbed on clays and concentrate in the finer, more tive correlations with Nb in the shales and sandstones. weathered materials [65]. In support of this, XRD analy- This possibly implies that it may have been controlled sis of the Ecca shales revealed that they are dominated by by clays and/or other phases associated with clay miner- kaolinite (Al2Si2O5(OH)4). In comparison with PAAS, the als. Rb and Ba are positively correlated in the sandstone average concentrations of SiO2, TiO2, Al2O3, Fe2O3, MnO, and shales, perhaps indicating a similar geochemical be- 346 Ë C. Baiyegunhi et al.

Figure 2: Major elements versus Al2O3 graph showing the distribution of shale samples from the Ecca Group. Average data of UCC and PAAS from [64] and [4], respectively are also plotted for comparison. Geochemistry of sandstones and shales from the Ecca Group Ë 347

Figure 3: Major elements versus Al2O3 graph showing the distribution of sandstone samples from the Ecca Group. Average data of UCC and PAAS from [64] and [4], respectively are also plotted for comparison. 348 Ë C. Baiyegunhi et al.

Figure 4: Spider plot of major elements showing (a) Ecca shales normalized against UCC, (b) Ecca sandstones normalized against UCC, (c) Ecca shales normalized against PAAS, (d) Ecca sandstones normalized against PAAS (after [64] and [4]).

Figure 5: Spider plot of trace elements showing (a) Ecca shales normalized against UCC, (b) Ecca shales normalized against PAAS, (c) Ecca sandstones normalized against UCC, (d) Ecca sandstones normalized against PAAS (after [4, 64]). haviour. These correlations indicate that their distribu- ally, Zr and Hf have high positive correlations and the ra- tions are mainly controlled by illites and other minor clays. tio of Zr to Hf in the analysed samples range from approxi- HFSE elements are enriched in felsic rocks rather than mately 21-43. This suggests that the elements are controlled mafic rocks [67]. The concentrations of Zr, Hf and Yinthe by zircons, since the values are similar or nearly the same shales and sandstones are comparable with the UCC con- with those documented by [68] for zircon crystals. The av- tents (Figure 5a), whereas Nb is depleted. Relative to PAAS erage contents of Zr in the shales are lower than those content, the concentration of Y and Nb in the shales and in the sandstones, which perhaps indicate that the min- sandstones are depleted, while Zr and Hf concentrations eral zircon tend to be preferentially concentrated in fine- are relatively similar to PAAS contents (Figure 5b). Gener- grained sands. TTE in the Ecca sandstones and shales are Geochemistry of sandstones and shales from the Ecca Group Ë 349

depleted in comparison with UCC and PAAS (Figure 5a), alkali contents (K2O + Na2O) and K2O/Na2O ratios should except for Cu and Zn, which is relatively comparable with be considered as reliable indicators of the intensity of the UCC and PAAS in the shales. The TTE in the studied source material weathering. In order to determine the de- sandstones do not behave uniformly. Among TTE, Sc cor- gree of source rock weathering, a few indices of weathering related positively with Ni which indicates that it is mainly have been proposed based on the molecular proportions concentrated in the phyllosilicates. of mobile and immobile element oxides (Na2O, CaO, K2O and Al2O3). Thus, the chemical composition of weathering products in a sedimentary basin is expected to reveal the 4.3 Provenance mobility of various elements during weathering [76]. The indices of weathering/alteration include chemical index of The composition of major element or oxides in sandstones alteration (CIA), chemical index of weathering (CIW) and and shales has also been used to determine sedimentary plagioclase index of alteration (PIA). Chemical index of provenance by the application of discriminant function alteration (CIA) proposed by [77] is the most widely used analysis [69]. This discriminant function analysis distin- chemical index to determine the degree of source area guishes between four major provenance fields, namely, weathering. [77] defined the CIA formula to evaluate the mafic igneous, intermediate igneous, felsic igneous and degree of chemical weathering as: quartzose sedimentary or recycled. In Figure 6a, the sand- [︁ * ]︁ CIA = Al2O3 /(Al2O3 + CaO + Na2O + K2O) × 100 stone samples plotted in the quartzose sedimentary prove- nance field. In contrast, the shale samples are scattered in Where CaO* is the content of CaO incorporated in silicate both quartzose sedimentary provenance and intermediate fraction. igneous provenance field, but they are mostly within the The value of CIA gives a measure of the ratio of origi- quartzose sedimentary provenance field (Figure 6b). The nal/primary minerals and secondary products such as clay binary plot of TiO2 versus Zr shows that all the shale and minerals. CIA values range from almost 50 in case of fresh sandstone samples are from felsic igneous rocks (Figure 7). rocks to 100 for completely weathered rocks. Thus, CIA val- ues increase with increasing weathering intensity, reach- Again, the TiO2-Ni diagram of [71] revealed that the ing 100 when all the Ca, Na and K have been leached from source area for most of the samples are predominantly of weathering residue. The CIA values in the Ecca sandstone acidic magmatic nature (Figures 8), despite the fact that and shale samples range from 24.41 to 83.76 (averaging a few samples plotted outside the field assigned for felsic 66.30), and 53.77 to 78.28 (averaging 69.74), respectively. source. The bivariate plot of La/Th against Hf (Figure 9) These CIA average values revealed relatively moderate to and ternary diagram of V-Ni-Th*10 (Figure 10) indicate that high degree of chemical weathering in the source area. In the studied sandstones and shales are derived from felsic addition to CIA, chemical index of weathering (CIW) also source rocks. provides information on the intensity of chemical weather- ing the sediments have undergone. In comparison to other weathering indices, the CIW is a superior method involv- 4.4 Paleoweathering conditions ing restricted number of components that are well-known with consistent geochemical behaviour during weather- Intensity of chemical weathering of source rocks are ing. The CIW formula as expressed by [78] is shown below: mainly controlled by the composition of the source rock, [︁ ]︁ CIW = Al O / (Al O + CaO* + Na O) × 100 duration of weathering, climatic conditions and rates of 2 3 2 3 2 tectonic uplift of source region [9]. Several researchers The CIW values of the studied sandstone and shale sam- like [4, 19, 73, 74] have documented that about 75% of la- ples range from 25.90 to 92.70 (averaging 75.20), and 76.23 bile materials in the upper crust are composed of feldspars to 96.25 (averaging 82.36) respectively. These CIW values and volcanic glass. Chemical weathering of these mate- point to moderate-high intensive chemical weathering. As rials resulted in the formation of clay minerals. During documented by [74], source area weathering and elemen- chemical weathering, Ca, Na and K are largely removed tal redistribution during diagenesis also can be assessed from source rocks and the amount of these elements sur- using the plagioclase index of alteration (PIA). PIA moni- viving in sediments derived from the rocks served as indi- tors and quantifies progressive weathering of feldspars to cator of the intensity of chemical weathering [2]. Accord- clay minerals [22, 74]. The maximum value of PIA is 100 ing to [75], if siliciclastic sedimentary rocks are free from for completely altered materials (i.e. kaolinite and gibb- alkali related post-depositional modifications, then their site) and weathered plagioclase has PIA value of 50. [74] 350 Ë C. Baiyegunhi et al.

Figure 6: Major element Discriminant Function diagram for sedimentary provenance (a) shale, (b) sandstones (after [68]). The discriminant functions are: Discriminant Function 1 = (−1.773 TiO2) + (0.607 Al2O3) + (0.760 Fe2O3) + (−1.500 MgO) + (0.616 CaO) + (0.509 Na2O) + (−1.224 K2O) + (−9.090); Discriminant Function 2 = (0.445 TiO2) + (0.070 Al2O3) + (−0.250 Fe2O3) + (−1.142 MgO) + (0.438 CaO) + (1.475 Na2O) + (−1.426 K2O) + (-6.861).

favourably removed than K-feldspars [19]. Feldspathic ma- terials in the sandstones and shales were subjected to vari- able intensities of weathering during the different evolu- tion stages. The individual bivariate plots of K2O/Na2O, K2O + Na2O, Na2O, K2O and CaO against PIA can be used to unravel the mobility of elements during the final stage of chemical weathering of previously altered feldspars. In the bivariate plot of K2O/Na2O against PIA (Figure 11a), the values of K2O/Na2O in the sandstone commonly increases with increasing value of PIA. On the contrary, in the plot of K2O + Na2O versus PIA (Figure 11b), the total content of alkalis in most of the samples decrease with increasing value of PIA. Figures 11 and 12 (c-e) shows the behaviour of Na, Ca and K during progressing weathering of feldspars in Figure 7: TiO2–Zr plot of shales and sandstones samples from the the sandstones and shales. Ecca Group (background field after [70]). Ternary plot of A-CN-K proposed by [19] is another method that can be used to assess the composition of defined the PIA formula to evaluate the amount of chemi- original source rock as well as the mobility of elements cal weathering as: during the process of chemical weathering of source ma- terial and post-depositional chemical modifications. The [︁ * ]︁ PIA = (Al2O3 − K2O) /(Al2O3 + CaO + Na2O − K2O) ternary plot of Al2O3-(CaO+Na2O)-K2O (represented as A- × 100 CN-K) is useful for identifying compositional changes of shales and sandstones that are related to chemical weath- The PIA values of the studied sandstone and shale samples ering, diagenesis and source rock composition. Geochemi- range from 21.910 to 91.82 (averaging 72.23) and 54.92 to cal data of the sandstones and shales from the Ecca Group 95.14 (averaging 78.76), respectively. Again, The PIA values were plotted in an A-CN-K diagram (Figure 13). The arrows suggest moderate-high or intense destruction of feldspars 1 to 5 in Figure 13 represent the weathering trends of gab- during source weathering, transport, sedimentation, and bro, tonalite, granodiorite, adamellite and granite, respec- diagenesis. During the initial stages of weathering, Ca is tively [79]. In the A-CN-K diagram (Figure 13), the sand- quickly leached than Na and K. With increasing weath- stones and shales plotted above the line joining plagio- ering, the total alkali content (K2O + Na2O) decreases clase and potash feldspar. The plots define a narrow lin- with increase in K-Na ratio (K2O/Na2O). This is due to ear trend which runs slightly at an angle parallel to the A- destruction of feldspars among which plagioclase is more Geochemistry of sandstones and shales from the Ecca Group Ë 351

Figure 8: TiO2 versus Ni bivariate plot for (a) sandstones, (b) shales from the Ecca Group (after [71]). The majority of the samples plot near the acidic source field.

Figure 9: Plot of Hf versus La/Th for the Ecca (a) shales, (b) sandstones (background field after [72]).

Figure 10: V-Ni-Th*10 triangle diagram for the Ecca (a) sandstones, (b) shale samples (background field after [13]). Shaded area represents composition of the felsic, mafic, and ultramafic rocks. 352 Ë C. Baiyegunhi et al.

CN edge. This is possibly due to the fact that the removal rate of Na and Ca from plagioclase is generally greater than the removal rates of K from microcline [19]. The plot trend towards illite on the A-K edge and does not show any inclination towards the K apex, thus indicating that the sandstones and shales are free from potash metasomatism during diagenesis. The trend line when extended back- ward intersects the plagioclase-potash feldspar join near arrow 3, which is the field of granodiorite (potential ulti- mate source). Linear weathering trend point to steady state of weathering conditions where material removal matches with production of weathering material [2].

4.5 Climatic conditions and sediment maturity

The original character and maturity of sediments as well as the prevailed climatic conditions can be determined by calculating the index compositional variation (ICV) pro- Figure 11: Bivariate diagrams depicting mobility of elements during posed by [58]. The ICV tends to be highest in minerals that weathering of feldspars in the sandstone samples from the Ecca are high in weathering intensity and decreases in more sta- Group. (a) (K2O/Na2O) wt. % versus PIA. (b) (K2O + Na2O) wt. % ble minerals (less weathered minerals). The ICV decreases versus PIA. (c) Na2O wt. % versus PIA. (d) CaO wt. % versus PIA. (e) further in the montmorillonite group clay minerals and K2O wt. % versus PIA. lowest in the kaolinite group minerals [58]. In addition, more mature shale tends to have low ICV values (< 1.0).

ICV = (Fe2O3 + K2O + Na2O + CaO + MgO + MnO)/Al2O3

As documented by [58], sandstones or shales with ICV > 1 are compositionally immature with the first cycle of sed- iments deposited in tectonically active settings. On the other hand, those with ICV < 1 are compositionally ma- ture and were deposited in the tectonically quiescent or cratonic environment where sediment recycling was ac- tive. For the studied sandstones and shales, the ICV val- ues range from 0.71 to 3.6 (averaging 1.20) and 0.41 to 1.05 (averaging 0.82).Based on the average ICV values, it can be inferred that the sandstones are compositionally imma- ture whereas the shales are compositionally mature and deposited in the tectonically quiescent or cratonic environ- ment. The K2O/Na2O ratios for the sandstones vary from 0.71 to 7.16 (averaging 1.61) and 1.04 to 8.29 (averaging 3.26) for the shale samples. These ratios revealed moderate to high maturity of the shales, which agrees with the ICV val- ues [9]. The K2O/Na2O ratios are comparable to those of sediments from passive margins, which increase with ma- Figure 12: Bivariate diagrams depicting mobility of elements during weathering of feldspars in the shale samples from the Ecca Group. turity of rocks [24]. The binary plot of CIA against ICV for (a) (K2O/Na2O) wt. % versus PIA. (b) (K2O + Na2O) wt. % versus PIA. the studied samples (Figure 14) shows that most of the (c) Na2O wt. % versus PIA. (d) CaO wt. % versus PIA. (e) K2O wt. % shales are geochemically mature and were derived from versus PIA. both weak and intensively weathered source rocks. Geochemistry of sandstones and shales from the Ecca Group Ë 353

Figure 13: A-CN-K ternary diagram of molecular proportions of Al2O3-(CaO+Na2O)-K2O for (a) sandstone, (b) shale samples from the Ecca Group (background field after [19]). The dotted arrow shows the actual weathering trend for the samples. The CIA scale shown attheleft side is for comparison.

tic sedimentary rocks, the proposed plot of SiO2 against (Al2O3 + K2O + Na2O) after [81] was used to classify the maturity of Ecca sandstones and shales as a function of climate. Figure 15 shows that the sandstones and shales mostly plot in the field of arid climate with few samples plotting in the humid climate field with varied maturity.

Figure 14: Binary plot of CIA against ICV for the Ecca shale and sand- stone samples.

Alternatively, SiO2/Al2O3 ratios of siliciclastic rocks are sensitive to sediment recycling and weathering pro- cess and can serve as an indicator of sediment maturity. With increasing sediment maturity, quartz survives pref- erentially to feldspars, mafic minerals and lithics [25, 80]. The average SiO2/Al2O3 ratios in unaltered igneous rocks range from ˜ 3.0 (basic rocks) to ˜ 5.0 (acidic rocks). Val- Figure 15: Chemical maturity of the Ecca sandstones and shales (background field after [81]). ues of SiO2/Al2O3 ratio > 5.0 in sandstones and shales point to progressive maturity [80]. The SiO2/Al2O3 ratios of the sandstones vary from 4.16 to 14.77 (averaging 9.45), while those of the shales range from 3.94 to 14.89 (aver- aging 5.50). The K2O/Na2O ratios of the sandstones range 4.6 Tectonic setting of the source area from 0.76 to 7.16 (averaging 1.61), while those of the shales vary from 1.04 to 8.29 (averaging 3.26). The low values of Several researchers like [5, 24, 25] documented that the K2O/Na2O as well as the high values of SiO2/Al2O3 in- chemical compositions of siliciclastic sedimentary rocks dicate low to moderate sediment maturity. To constrain are considerably controlled by plate tectonic settings of the climatic condition during sedimentation of siliciclas- their provenances and depositional basins. Thus, sili- 354 Ë C. Baiyegunhi et al. ciclastic rocks from different tectonic settings possess riched in the passive margin setting. La-Th-Sc tectonic dis- terrain-specific geochemical signatures. Tectonic setting crimination diagram demonstrates the passive margin set- discrimination diagrams give reliable results for siliciclas- ting of the depositional basin for the Ecca sandstones and tic rocks that have not been strongly affected by post- shales (Figure 20). depositional weathering and metamorphism [8]. Bivari- ate plots of major and trace element geochemistry have been used by several researchers to determine the tectonic 5 Discussion and conclusion setting of sandstones and shales [5, 24, 25, 27, 30, 82]. Among the various tectonic setting discrimination dia- Selected trace elements such as La, Ce, Nd, Y, Th, Zr, Hf, grams, the major element-based discrimination diagrams Nb, Sc, Co and Ti are useful in discriminating provenance of [5] and [24] are widely used. [24] divided series of tec- and tectonic setting of sedimentary basins because of their tonic plots to differentiate between four main tectonic set- relative immobility during sedimentary processes [5, 83]. tings, namely, oceanic island arc (OIA), continental island These elements are present in very low concentrations in arc (CIA), active continental margin (ACM) and passive sea and river water, chiefly transported as particulate mat- continental margin (PM). Chemical analyses data of the ter and reflect the signature of the parent material [5, 24, sandstones and shales have been plotted on 4 tectonic set- 83]. The trace element data of sandstones and shales from ting discrimination diagrams of [5, 24, 25, 27]. Bivariate the Ecca Group are generally comparable with the trace plots of TiO versus (Fe O + MgO) and Al O /SiO against 2 2 3 2 3 2 element data of UCC and PAAS. In the studied samples, Fe O + MgO shows that most of the studied samples plot- 2 3 the concentration of high field strength elements like Hf ted in the passive margin, active continental margin and and Zr are higher in the sandstones than the shale sam- continental island arc fields (Figures 16a and 16b). In Fig- ples. The low content of these elements and elemental ra- ure 16b, only one sample plotted in the oceanic island arc tios like La/Sc and Th/Sc points to the presence of fraction- field. The Bivariate plot ofK O/Na O versus (Fe O + MgO) 2 2 2 3 ated source rocks with lower compatible element contents and Al O /(CaO + Na O) versus (Fe O + MgO) revealed 2 3 2 2 3 and recycled sediments in the source area. The variation in that the sandstones and shales are related to passive mar- chemical composition depicts change in the supply of ma- gin and active continental margin (Figures 16c and 16d). terial and a variation in physic-chemical environment of deposition. The pattern of geochemical behaviour of indi- Most of the shale samples represent the active con- vidual element shows that most of the trace elements that tinental margin and few samples fall in the passive con- found their way into the ancient sediments seems to have tinental margin field (Figure 17a). In addition, only one invaded the lattices of the silicates and clay minerals and sample from the Prince Albert Formation plotted in the is- structurally combined with them. land arc field. Active continental margins are subduction The geochemical data of major and trace elements related basins, continental basins and pull-apart basins show that the studied sandstone and shales have the same associated with strike-slip fault zones. On the other hand, source. Based on the discriminant function plots, it can passive continental margins are basins on continental be inferred that the sandstones and shales are mostly of crust and basins associated with ocean floor spreading, quartzose sedimentary provenance, suggesting that they failed rifts and Atlantic-type continental margins. Most of were derived from a cratonic interior or recycled orogen. the sandstone samples plotted in the passive continen- The binary plots of TiO versus Ni suggest that most of the tal margins and two samples each falls in the island arc 2 Ecca samples are sourced from an acidic magmatic nature and active continental margin fields (Figure 17b). Sim- (Figure 8). Furthermore, the binary plots of TiO versus Zr ilarly, more detailed results can be obtained using the 2 (Figure 7), La/Th against Hf (Figure 9) and the ternary di- calc-alkaline ternary diagram (CaO-Na O-K O). The calc- 2 2 agrams of V-Ni-Th*10 (Figure 10) shows that the shale and alkaline ternary diagram depicted in Figure 18a shows that sandstone samples were derived from felsic igneous rocks. most sandstone samples are related to passive continen- The results of the trace elements correlate well with that of tal margin. Shale samples are also represented in both ac- the major elements. tive continental margin and passive continental margin The CIA values (24.41-83.76%) in the Ecca rocks indi- (Figure 18b). Th-Sc-Zr/10 tectonic discrimination diagram cate low to high weathering degree of the source rocks. The of [5] revealed that the source area for most of the Ecca relatively high CIA values (53.77 to 78.28%) in the shales samples is predominantly of passive continental margin probably reflect the presence of clay minerals and low per- (Figure 19). Several researchers like [5, 24, 25] have doc- centage of detrital feldspars. Alternatively, the low CIA val- umented that immobile elements like La, Zr and Hf are en- Geochemistry of sandstones and shales from the Ecca Group Ë 355

Figure 16: Bivariate plots of (a) TiO2 (wt.%) versus (Fe2O3 + MgO) (wt.%), (b) Al2O3/SiO2 versus Fe2O3 + MgO (wt.%), (c) K2O/Na2O versus (Fe2O3+ MgO) (wt.%), (d) Al2O3/(CaO + Na2O) versus (Fe2O3+ MgO) (wt.%) of the Ecca rocks on the tectonic setting discrimination diagram of [24]. PM: Passive Margin, ACM: Active Continental Margin, CIA: Continental Island Arc, OIA: Oceanic Island Arc.

Figure 17: K2O/Na2O versus SiO2 tectonic-setting discrimination diagram for Ecca (a) shales, (b) sandstones (background field after [25]).

Figure 18: Na2O-CaO-K2O ternary plot for (a) sandstone, (b) shale samples from the Ecca Group (background field after [27]). OIA = oceanic island arc, CIA = continental island arc, ACM = active continental margin, PM = passive continental margin. 356 Ë C. Baiyegunhi et al.

Figure 19: Th-Sc-Zr/10 tectonic discrimination diagram for the Ecca (a) sandstones, (b) shales (background field after [5]). A = oceanic island arc B = continental island arc, C = active continental margin, D = passive continental margin.

Figure 20: La-Th-Sc tectonic discrimination diagram for the Ecca (a) sandstones, (b) shales (background field after [5]). A = passive and active continental margin, B = continental island arc, C = oceanic island arc.

ues in some of the shales (< 40%) point to a low weathering tent (K2O + Na2O) decreases with increase in K-Na ratio condition in the source area and perhaps reflect cool con- (K2O/Na2O). This is due to destruction of feldspars among ditions. The variations in CIA reflect changes in the prop- which plagioclase is more favourably removed than K- erties of feldspar versus aluminous clay minerals. The ob- feldspars [2, 19]. served changes in the CIA values indicate that the sedi- The bivariate plots of Na2O versus PIA, K2O against ments are moderately to highly weathered, which possi- PIA and CaO versus PIA shows weak correlation which bly suggests that the sediments were derived from source could be attributed to the presence K-bearing minerals rocks that have been subjected to both chemical and physi- (i.e. muscovite and biotite) as well as the retention of cal weathering. The CIW values for the studied sandstones most of the mobilized K by aluminous material resulting and shales range from 25.90 to 96.25%, suggesting mod- in the formation of illite. The above described and doc- erate to high intensive chemical weathering. In the anal- umented imprints of progressing chemical weathering of ysed samples, the CIW values are higher than those of detrital feldspars perhaps give an impression (may be mis- CIA due to exclusion of K2O from the index. Based on the leading) that major event of chemical weathering of de- CIW values, the Ecca sandstones and shales are inferred trital feldspars has taken place essentially in the termi- to have undergone moderate to high chemical weathering. nal basin prior to the lithification of the detritus and dur- The PIA values of the studied samples range from 21.910 ing diagenesis. The observed systematic depletion of Na2O to 95.14%. Again, The PIA values suggest moderate to high content with increasing degree of chemical weathering of or intense destruction of feldspars during source weather- the feldspars (Figures 11c and 12c) tend to support above ing, transport, sedimentation and diagenesis. During the assumption. However, it is known that detrital feldspar initial stages of weathering, Ca is quickly leached than Na grains can survive more than one sedimentary cycle. The and K. With increasing weathering, the total alkali con- A-CN-K diagram of the sandstone and shale samples show Geochemistry of sandstones and shales from the Ecca Group Ë 357 that paleoclimate of the source area was relatively warm, structive comments, which really help in improving the which caused chemical weathering of source rocks reduc- manuscript. ing some initial feldspar in source rocks. Fairly long dis- tance of transport, (i.e., perhaps hundreds of kilometres) Declaration of interest The authors declared that there is also an important process responsible for further reduc- is no any actual or potential conflict of interest including tion of feldspars. Diagenetic alteration of feldspar was the any financial, personal or other relationships with other most important factor reducing feldspar in the sandstone. people or organizations regarding the publication of this The ICV values for the Ecca sandstones and shales original manuscript. range from 0.41 to 3.6, while K2O/Na2O ratios vary from 0.71 to 8.29. The K2O/Na2O ratios are comparable to those of sediments from passive margins, which increase with References maturity of rocks (Bhatia, 1983). The binary plot of CIA against ICV (Figure 14) shows that most of the shales are [1] Nesbitt, H.W., Young, G.M., Petrogenesis of sediments in the ab- geochemically mature and were derived from both weak sence of chemical weathering: effects of abrasion and sorting and intensively weathered source rocks. The bivariate plot on bulk composition and mineralogy. Sedimentology, 1996, 43, of K2O/Na2O versus (Fe2O3+ MgO), Al2O3/(CaO + Na2O) 341-358. [2] Nesbitt, H.W., Fedo, C.M., Young, G.M., Quartz and feldspar sta- against (Fe2O3+ MgO) and log (K2O/Na2O) versus SiO2 re- bility, steady and non-steady state weathering and petrogen- vealed that the sandstones and shales are generally re- esis of siliciclastic sands and muds. Journal of Geology, 1997, lated to passive margin and active continental margin. The 105, 173-191. Na2O-CaO-K2O ternary plot after [27] suggests passive con- [3] Armstrong-Altrin, J.S., Provenance of sands from Cazones, Aca- tinental margin and active continental margin provenance pulco and Bahía Kino beaches, Mexico. Revista Mexicana de for the Ecca shales and sandstones. Triangular Th-Sc-Zr/10 Ciencias Geológicas, 2009, 26(3), 764-782. tectonic discrimination diagram of [5] revealed that the [4] Taylor, S.R., McLennan, S.M., The Continental Crust: Its Com- position and Evolution. Blackwell Scientific Publications, 1985, source area for most of the Ecca samples are predomi- 312 pp. nantly of passive continental margin with a minor contri- [5] Bhatia, M.R., Crook, K.A.W., Trace element characteristics of bution from continental island arc and active continen- and tectonic setting discrimination of sedimentary tal margin sources (Figure 19). The La-Th-Sc ternary plot basins. Contributions to Mineralogy and Petrology, 1986, 92, also suggests passive and active margin settings for the 181-193. [6] Cullers, R.L., Barrett, T., Carlson, R., Robinson, B., Rare earth Ecca samples. The similarity of the La/Th versus Hf dia- element and mineralogic changes in Holocene soil and stream gram (Figure 9) and the La-Th-Sc tectonic discrimination sediment: a case study in the Wet Mountains, Colorado, USA. diagram (Figure 20) is a clear indication of a passive-active Chem. Geol., 1987, 63, 275-297. margin depositional basin. The data from these immobile [7] Cullers, R.L., Basu, A., Suttner, L., Geochemical signature of provenance in sand-size material in soils and stream sediments trace elements correlate well with Figure 17 (K2O/Na2O near the Tobacco Root batholith, Montana, USA. Chem. Geol., versus SiO2) as well as Figure 18 (Na2O-CaO-K2O). 1988, 70, 335-348. In summary, the study of paleoweathering conditions [8] McLennan, S.M., Hemming, S., McDaniel, D.K., Hanson, G.N., based on chemical index of alteration (CIA), plagioclase Geochemical approaches to sedimentation, provenance and index of alteration (PIA) and A-CN-K (Al2O3-CaO+Na2O- tectonics. In: M.J. Johnson and A. Basu, (Eds.), Processes Con- K2O) indicate that perhaps chemical weathering in the trolling the Composition of Clastic Sediments. Geological Soci- source area and recycling processes have been more signif- ety of American Special Paper, 1993, 32, 21-40. [9] Wronkiewicz, D.J., Condie, K.C., Geochemistry of Archean icant in the shales and sandstones. The CIA values indicate shales from the Supergroup, South Africa: low to high weathering conditions of the samples and the source-area weathering and provenance. Geochim. Cos- paleoclimate of the source area was probably warm. The mochim. Acta, 1987, 51, 2401-2416. tectonic setting discrimination diagrams support passive- [10] Condie, K.C. Chemical composition and evolution of the upper active continental margin setting of the provenance. continental crust: contrasting results from surface samples and shales. Chemical Geology, 1993, 104, 1-37. [11] Yan, Y., Xia, B., Lin, G., Cui, X., Hu, X., Yan, P., Zhang, F., Acknowledgement: Thanks to National Research Foun- Geochemistry of the sedimentary rocks from the Nanxiong dation (NRF) of South Africa (UID: 101980) and Govan Basin, South China and implications for provenance, paleo- Mbeki Research and Development Centre (GMRDC) of the environment and paleoclimate at the K/T boundary. Sedimen- University of Fort Hare for financial support. The authors tary Geology, 2007, 197, 127-140. thank the editors and anonymous reviewers for their con- [12] McLennan, S.M., Simonetti, A., Goldstein, S.L., Nd and Pb iso- topic evidence for provenance and post-depositional alteration 358 Ë C. Baiyegunhi et al.

of the Paleoproterozoic Huronian Supergroup, Canada. Precam- [27] Toulkeridis, T., Clauer, N., Kröner, A., Reimer, T., Todt, W., brian Research, 2000, 102, 263-278. Characterization, provenance, and tectonic setting of Fig Tree [13] Bracciali, L., Marroni, M., Pandolfi, L., Rocchi, S., Geochem- from the Archaen Barberton Greenstone Belt, South istry and petrography of Western Tethys Cretaceous sedimen- Africa. Sedimentary Geology, 1999, 124, 113-129. tary covers (Corsica and Northern Apennines): From source area [28] Weaver, C.E., Clays, Muds and Shales. Amsterdam, Elsevier, to configuration of margins, in Arribas, J., Critlli, S., Johanson, 1989, 818pp. M.J., (eds.), Sedimentary provenance and petrogenesis: Per- [29] Crook, K.A.W., Lithogenesis and geotectonics: the significance spectives from petrography and geochemistry. Geological So- of compositional variation in flysch arenites (greywackes). So- ciety of America, Special Paper, 2007, 420, 73-93. ciety of Economic, Paleontology and Mineralogy Special Publi- [14] Condie, K.C., Another look at rare earth elements in shales. cation, 1974, 19, 304-310. Geochim. Cosmochim. Acta, 1991, 55, 2527-2531. [30] McCann, T., Petrological and geochemical determination of [15] Wronkiewicz, D.J., Condie, K.C., Geochemistry and mineralogy provenance in the southern Welsh Basin. In: A. C. Morton, S. of sediments from the Ventersdorp and Transvaal Supergroups, P., Todd and P. D. W. Haughton (Editors), Developments in Sed- South Africa: cratonic evolution during the early Proterozoic. imentary Provenance. Geological Society Special Publication, Geochim. Cosmochim. Acta, 1990, 54, 343-354. 1991, 57, 215-230. [16] Condie, K.C., Wronkiewicz, D.J., The Cr/Th ratio in