1. Introduction
Opaque minerals (OM) such as ilmenite (FeTiO3), titanomagnetite defined as a solid solution between ulvöspinel (Fe2TiO4) and magnetite (Fe3O4) can be used as provenance tracers in sedimentary terranes due to their relative stability in volcanic and coastal sands (Darby and Tsang, 1987; Grigsby, 1990; Kasper-Zubillaga et al ., 2013). Additionally, OM are useful in provenance and coastal process studies because of their variable concentrations in major and trace elements, especially in TiO2, and their relative stability during sand transport, which provides information on source rocks and coastal processes (Grigsby, 1990; Butler, 1992; Fletcher et al ., 1992; Bradley et al ., 2002; Kasper-Zubillaga et al ., 2008a, b; Armstrong-Altrin et al ., 2012; Nallusamy et al ., 2013; Makvandi et al ., 2015).
The purpose of this paper is to determine the provenance of OM in beach and dune sands of the western coast of the Gulf of Mexico surrounded by mafic volcanic, volcanoclastic rocks and subordinate exposures of felsic intrusive bodies. OM in the beach and dune sands from the western coastal area of the Gulf of Mexico were studied due to their similar geomorphological characteristics to coastal areas elsewhere, i.e ., wide sandy strandplains, some with vegetated dune systems (Komar and Wang, 1984; Bradley et al ., 2002; Angusamy, 2006; Dill, 2007). In this paper we discuss the extent to which OM in beach and dune sands from the Gulf of Mexico are geochemically controlled by the geology of the surrounding area. In order to prevent bias focusing on one single detrital-mineral population (in our case OM) representing a minimal part of the total sediment input seawards (Garzanti, 2016), our approach considered the bulk composition of the coastal sands, including: a) the grain size characteristics of the coastal sands, b) the petrographic analysis of light minerals (i.e. , quartz, feldspars, and rock fragments) (Kasper-Zubillaga et al. , 2013), c) the geochemistry of the OM, for provenance purposes, and d) the roundness of the grains, in order to observe the degree to which the subordinate OM are close to the source rocks or have experienced longshore transport away from the source rocks.
2. Study areas
The study area is located in the State of Tamaulipas, on the western coast of the Gulf of Mexico, Mexico (Figure 1). The three beaches sampled, from south to north, were Miramar, Barra del Tordo, and Tepehuajes (Figure 1). Sand dunes are of the vegetated-transverse type (Livingstone et al ., 1999). The three beach sites are located 74 to 51 km apart with an average width of 130 ± 29.9 m, 50 ± 19.6 m, and 43.3 ± 6.5 m (ρ = 0.05; n = 3) for each beach, respectively (INEGI, 2012).
The sedimentary rocks exposed in the area belong to the Gulf of Mexico Coastal Plain between the Sierra Madre Oriental (Tpal (lu)) and the Gulf of Mexico coast, composed of Upper Cretaceous and Tertiary shales and sandstones along the three beach sites studied. The Volcanic rocks, i.e ., the Aldama and Maratínez extrusive igneous rocks (T (Igeb) and Q (Igeb)) of mid to late Tertiary and/or Quaternary ages, respectively, are exposed near the Barra del Tordo and Tepehuajes beach sites (Treviño-Cázares et al ., 2005; Aranda-Gómez et al ., 2007). Their chemical composition ranges from trachytes to alkali basalts, i.e. , ~ 47 to 60 % SiO2; ~ 5 to 15 % Na2O + K2O (Le Maitre et al ., 2002; Aranda-Gómez et al ., 2007) (Figure 1). The Sierra de Tamaulipas intrusive igneous rocks intruded in Cretaceous limestones west of the Aldama and Maratínez rocks. They are composed of alkaline intrusions, i.e ., gabbro to granite (T (Igia)) (Ortega-Gutiérrez et al ., 1992; Aranda-Gómez et al ., 2007).
Tides in the study area are semidiurnal, averaging 50 and 30 cm in amplitude for the three beach sites (Yáñez-Arancibia and Day, 2005). Longshore currents are 13 cm/s and 7 cm/s during the dry and wet seasons, respectively (Fernández-Eguiarte et al ., 1992). Waves measure on average 0.6 to 1.2 m in height, with periods of 6 to 8 s, and they move in a northerly direction.
Winds come from the north, east, northeast, and southeast with average velocities of 4, 3, 5, and 4 m/s per year, respectively (Pérez-Villegas, 1990). Average annual precipitation ranges from 300 to 2000 mm (INEGI, 2012). Average temperatures range from 28 to 30 °C in summer and from 14 to 20 °C in the winter (Yáñez-Arancibia and Day, 2005).
3. Equipment and methods
3.1. Field sampling
Beach and dune sands were collected from Miramar (n = 35), Barra del Tordo (n = 55), and Tepehuajes (n = 35) during the wet season i.e ., August 2012. Samples were taken from the uppermost centimeter of the inshore, foreshore, and backshore of the beach. Samples were also taken from the windward side and the crest of the dune sands. This was performed in order to prevent bias in grain size and mineral composition of the sands since each beach and dune subenvironment are controlled by a combination of coastal processes, i.e ., longshore currents, wind deflation, and wave energy (Komar and Inman, 1970; Abuodha, 2003; Kasper-Zubillaga et al ., 2013, 2015).
3.2. Sand petrography
Compositional modes from the Miramar and Barra del Tordo beach sites were taken from a previous study conducted by Kasper-Zubillaga et al. (2013). Sand samples from Tepehuajes beach and dune sands were point-counted in this study following the method described by Franzinelli and Potter (1982) . This was done because of a) the size (Mz ɸ) and b) the sorting (σ ɸ) character of the sands. Further results regarding these two parameters are shown in subsection 4.1. Normalized rock fragment point-counting results were plotted in a ternary diagram with Rv and Rs (volcanic and sedimentary lithic fragments plotted separately) were associated with a supracrustal influence, while the Rm + Rp (metamorphic and plutonic lithic fragments grouped together) pole was associated with a deep-seated crust (see Kasper-Zubillaga et al ., 2013). Modal analysis of the whole bulk sand composition, i.e ., point counting of 200 to 300 grains per slide, included: total quartz (Qt) = mono and polycrystalline quartz; total feldspar; (Ft) = K-feldspar (Fk) + plagioclase (P); and total rock fragments (Rt) = volcanic + sedimentary + metamorphic + plutonic lithics; heavy minerals (HM), mainly magnetite with few zircon grains; and biogenic components (B) like shells.
3.3. OM separation
OM fractions were subsampled using the sieving sets between 2.00 ɸ to 3.75 ɸ. This was done prior to the grain-size parameter determinations for the whole bulk sand sample due to their usefulness in selecting the sand fractions for heavy mineral separation. Heavy mineral fractions were obtained using 0.2, 0.4, 0.8, and 1.2 amperage at tilting variable 10° to 20° with ~ 40 g of weighted sand sample at the start of the magnetic separation (Bradley et al ., 2002).
The opaque minerals/non-opaque minerals (OM/NO) ratio was calculated for beach and dune sand samples in order to determine the concentration of ilmenite, titanomagnetite, and magnetite grains for the three beach sites. The OM/NO ratio was produced as the quotient of the fractions expressed in grams obtained with the magnetic separator. Samples displaying the highest values for the OM/NO ratio were selected to be randomly analyzed with a JEOL SuperProbe Microprobe Analyzer (MASP). Details of the analytical results are given in Tables 1 and 2.
Average sand sizes in ɸ = phi units used for magnetic separation. M = Miramar, B = Barra del Tordo, T = Tepehuajes. Beach sands F = foreshore, B = backshore. Dune sands W = windward, C = crest. Number of trace minerals > 1 are between brackets.
M= Miramar ; B= Barra del Tordo; T= Tepehuajes; F= foreshore, B= backshore, W= dunes windward, C= dunes crest
OM compositions were plotted in the TiO2-FeO-Fe2O3 ternary diagram (Butler, 1992) (Table 1; Figures 2, 3, 4). Normalization of the data was based on Fe, Ti, and trace element content reported as oxides based on the MASP results (Table 2). Confidence regions of the population mean (CRPM) were drawn at 99 % confidence levels to determine if the means of the three OM groups, i.e. , ilmenite, titanomagnetite, and magnetite, were significantly different (Weltje, 2002).
3.4. OM morphology (roundness)
In order to estimate the roundness of the OM, between 50 to 200 OM grains for beach and dune sands were counted. This was done to a) show the degree of roundness of the OM for beach and dune sands at each site studied, and b) determine the degree to which the OM were under the influence of long and/or short transport since roundness is a good indicator of transport and proximity to possible source rocks (Folk, 1980; Sagga, 1993; Kasper-Zubillaga et al ., 2005). Fifty-four thin sections were used based on the visual chart presented by Powers (1953). OM were classified and counted as VA = very angular, A = angular, SA = subangular, SR = subrounded, R = rounded and VR = very rounded. The roundness ratios VA/A (very angular/angular), SA/SR (subangular/subrounded), and R/VR (rounded/very rounded) were determined to be used as poles in a ternary diagram (Figures 3a, b) (see Kasper-Zubillaga, 2009). OM were undifferentiated in terms of their geochemical composition, which implies that the OM were considered as one group without isolating ilmenites, titanomagnetites, and magnetites separately during the roundness analysis.
4. Results
4.1. Grain size
The coastal sands were moderately to well-sorted fine sands with Mz = 2.09 ± 0.16, σ = 0.53 ± 0.06; Mz = 2.17 ± 0.08, σ = 0.48 ± 0.08; and Mz = 2.08 ± 2.08, σ = 0.81 ± 0.14 for Miramar, Barra del Tordo, and Tepehuajes, respectively. The beach and dune sands showed relatively similar grain size and sorting values for the three beaches with the exception of the sorting values for Tepehuajes, which were slightly higher compared to the Miramar and the Barra del Tordo sands.
4.2. Modal analysis
Previous petrographic analyses showed that Miramar (Q90 F2 R8 ; n = 46), Barra del Tordo (Q92 F3 R5 ; n = 41) (Kasper-Zubillaga et al ., 2013), and Tepehuajes (Q96 F1 R3; n = 16) beach and dune sands are quartzose composed of quartz, then rock fragments and feldspar. The sedimentary rock fragments were composed of sandstones, shales, and chert. Chert was dominated by a fine cryptocrystalline texture and showed no rhythmic banding resulting from hydrothermal volcanic activity (He and Zhou, 2014). The sedimentary fragments dominate the samples collected at the Miramar, Barra del Tordo, and Tepehuajes beaches with average ranges of 53 - 60 %, 52 - 65 %, 57 - 64 %, respectively (Kasper-Zubillaga et al ., 2013). Basaltic lithic fragments are characterized by the absence of conspicuous phenocrystals. The presence of phaneritic textures, polycrystalline, quartzo-feldspathic grains are observed in intrusive fragments, which are subordinate in all analyzed samples, displaying abundances of medium-sized polycrystalline grains with little recrystallization, an absence of sutured and/or crenulated boundaries and straight crystal-to-crystal boundaries as intrusive derived fractions (Voll, 1960; Blatt, 1967). Quartz-rich river loads are close to the Miramar and Barra del Tordo beach sites, i.e. , Q94 F2 R4 and Q96 F1 R3, respectively, partially controlling the high quartz content in the coastal sands (Kasper-Zubillaga et al ., 2013). The Tepehuajes beach site is influenced by the Soto La Marina river input but influenced by longshore currents with a northwesterly component (Fernández-Eguiarte et al ., 1992; Kasper-Zubillaga et al ., 2013).
4.3. OM at the Miramar beach site
The OM/NO ratios for beach and dune sands are 0.03 ± 0.00, 0.03 ± 0.00; 0.03 ± 0.00, 0.06 ± 0.02; 0.11 ± 0.05, 0.19 ± 0.06 at Miramar, Barra del Tordo, and Tepehuajes, respectively. The TiO2-FeO-Fe2O3 ternary diagram shows the ilmenite fraction slightly above the 50 % ilmenite mark, i.e. , 1/2 FeTiO3. Comparisons of the ilmenite fraction were made between the gabbros from east Greenland, and ilmenite included as subordinate minerals in basalts from eastern India (Vincent and Phillips, 1954; Nageswara-Rao et al ., 2012). Additionally, the ilmenite CRPM plots slightly above the 50 % ilmenite mark. Titanomagnetite single grain data and their CRPM overlap the ulvöspinel-titanomagnetite mark, but also overlap the titanomagnetite values for basalts formed at high and low temperatures from east India (Nageswara-Rao et al ., 2012), and extend towards the wüstite-FeO pole (Grigsby, 1990). Depletion of magnetite grains is observed at the Miramar beach site. No overlapping of the CRPM areas for the ilmenite and titanomagnetite data was observed (Figure 2).
4.4. OM at the Barra del Tordo beach site
The TiO2-FeO-Fe2O3 ternary plot shows that the ilmenite data set plot slightly above the 50 % ilmenite mark, as observed for the ilmenite fraction at the Miramar beach site. Similar to the Miramar beach site trends, titanomagnetite single grain data and their CRPM overlap the ulvöspinel-titanomagnetite mark. The titanomagnetite data set tend towards the wüstite-FeO pole (Grigsby, 1990) and overlap with the reported values for basalts formed at high and low temperatures in east India (Nageswara-Rao et al ., 2012) The magnetite samples plot close to the wüstite pole and detrital magnetite grains from intrusive acidic igneous rocks (Grigsby, 1990). The CRPM areas for the ilmenite and titanomagnetite data do not overlap (Figure 3).
4.5. OM at the Tepehuajes beach site
The highest concentration of OM is observed in samples from the Tepehuajes site. The TiO2-FeO-Fe2O3 ternary diagram shows that some ilmenite samples tend towards the rutile pole with high titanium content and slightly above the 50 % ilmenite mark. Some ilmenite samples overlap with the ilmenite in basalts formed at high and low temperatures in eastern India, and slightly above the ilmenite-gabbro association from east Greenland (Vincent and Phillips, 1954; Nageswara-Rao et al ., 2012). The titanomagnetite minerals plot closer to the wüstite pole while four magnetite grains overlap the detrital magnetite samples comprising granites from the USA (Grigsby, 1990). The rest of the magnetite is spread out towards the magnetite mark and the hematite pole. The CRPM areas for the ilmenite, titanomagnetite, and magnetite grains do not overlap (Figure 4).
4.6. Roundness in OM
The VA/A-SA/SR-R/VR triangular diagram showed that beach sand samples from Miramar and Barra del Tordo tend towards the right-hand side of the triangle, i.e ., average values > 60 % and 49 %, respectively, near the R/VR pole. The Tepehuajes beach sand samples plot in the middle of the ternary diagram, i.e. , average value > 45 % towards the R/VR pole (Figure 5a). The dune sand samples showed more grouping towards the R/VR pole for Miramar and Barra del Tordo than beach sand samples from the same locations, but the Tepehuajes dune sands showed undefined tendencies towards a specific pole with the exception of two dune sand samples that tended towards the SA/SR pole, i.e. , average value ~ 40 % towards the SA/SR pole (Figure 5b).
5. Discussion
5.1. Grain size
The average grain-size parameters for the three beach sites do not differ significantly. Grain size values similar to those of this study have been reported for the western coast of the Gulf of Mexico close to the area of influence of the Trans-Mexican Volcanic Belt (TMVB) where a wide coastal plain is exposed (Kasper-Zubillaga et al ., 1999). In contrast, coarse grain-sized beach sands from the southwestern Mexican Pacific coast are influenced by volcanic outcrops in a narrow coastal plain (Carranza-Edwards et al ., 2009). In our study, the grain sizes were determined by a wide coastal plain that controlled their distribution in beach and dune sands.
5.2. Modal analysis
The quartzose character of the beach and dune sands from the western coast of the Gulf of Mexico is a response to the control exerted by the sedimentary units, i.e. , Upper Cretaceous and Tertiary shales and sandstones, exposed along the coast from the Miramar to the Tepehuajes beach sites. This interpretation is supported by the presence of sedimentary rock fragments like sandstones, chert, monocrystalline quartz with straight extinction, K-feldspars, amphiboles, and pyroxenes as non-OM fractions from the whole bulk sand composition (Kasper-Zubillaga et al ., 2013). Furthermore, it is likely that the gabbro and especially granite outcrops from the Sierra de Tamaulipas (Ortega-Gutiérrez et al ., 1992; Aranda-Gómez et al ., 2007) are the remnants of denudation processes that influenced the supply of sediments that contribute to the sedimentary units along the coast, i.e ., the Upper Cretaceous and Tertiary shales and sandstones. Elsewhere, modern quartz-rich beach and dune sands are found derived from granite, i.e. , K-feldspar, albite Na-plagioclase enrichment, amphiboles, and mica, e.g ., on the eastern coast of South America, the northwestern coast of Mexico and the northern coast of New Zealand, where sedimentary and acidic intrusive rock units control the major composition (Potter, 1986; Kasper-Zubillaga et al ., 2005, 2007 a, b; Kasper-Zubillaga and Zolezzi-Ruiz, 2007).
5.3. OM content and roundness
The OM content of the three beach sites increases from Miramar to Tepehuajes, from south to north. The increase in OM at the Tepehuajes beach sites may be due to the longshore transport of OM. This is supported by the fact that beach and dune sands are composed mainly of subrounded and rounded to very rounded ilmenite, titanomagnetite, and magnetite grains. This interpretation is based on the data for OM thin sections that were plotted in the VA/A-SA/SR-R/VR triangular diagram (Figures 5a, b). The roundness of the OM at each beach site is similar, tending towards the SA/SR and R/R/VR poles. This suggests a long time of transport along the coast due to longshore currents, breaking waves, and backflow concentration of OM landwards, as well as mixing with sea-floor sediments during river water-borne sand settlement seawards (Bradley, 2002; Mohapatra et al ., 2015). Additionally, modern river-sand discharges exert little control on the current deposition of OM along the beaches studied (Kasper-Zubillaga et al ., 2013).
5.4. OM at the Miramar beach site
Ilmenite is probably primarily derived from pre-existing eroded mafic sources, i.e. , alkali basalts close to the Miramar beach site (Cantagrel and Robin, 1979; Nageswara-Rao et al ., 2012). This interpretation is closer to the basalt and alkali basalt outcrops near to the Miramar beach site. Evidence of these statements are based on a) the presence of alkali basalts as part of the outcrops of the remnants of the Aldama volcanic field, b) the increase of K-feldspar content as "light minerals" in the sands, c) the presence of quartz and pyroxenes in the Miramar sands, and d) the presence of light rare earth element (LREE) enrichment, i.e. , chondrite normalized and the positive Eu anomalies (Chakraborty et al ., 1980; Nageswara-Rao et al ., 2012; Kasper-Zubillaga et al ., 2013). The ilmenite has not been subjected to chemical alteration, since its average TiO2 and FeO values are below and above the average values obtained for altered ilmenite (intermediate phases) of 67.36 ± 7.50 and 27.31 ± 8.00, respectively, in sedimentary rocks (Morad and Aldahan, 1986). Additionally, relict, unaltered ilmenite detrital grains, i.e. , mostly subrounded, rounded, and very rounded homogeneous grains, from offshore marine sands may be transported onshore via longshore currents, breaking waves and wind deflation (Bradley et al ., 2002) (Figures 5a, b).
Titanomagnetite is derived from sedimentary units such as the Upper Cretaceous and Tertiary shales and sandstones exposed as a result of the denudation of mafic/basalt rocks from the subalkaline to alkali basalts from the Aldama volcanic rocks (Vasconcelos et al ., 2002) since the non-OM minerals included K-feldspars, amphiboles and pyroxenes and the chemical composition overlaps on the ulvöspinel mark. Titanomagnetite is a solid solution between the magnetite and the ulvöspinel end members (Sauerzapf et al ., 2008). The main source of titanomagnetite is probably composed of alkali basalts and/or tholeiitic basalts, i.e. , subalkaline basalts (Vasconcelos et al ., 2002) (Table 3). Furthermore, the Ti-Fe composition values (%) are close to values from tholeiitic basalts from eastern India and titanomagnetite homogeneous grains (Deer et al ., 1966; Prévot and Mergoil, 1973; Nageswara-Rao et al ., 2012) (Table 3). The presence of subordinate concentrations of sphene and zircon is derived as an exogenous process of granites forming part of the Sierra de Tamaulipas (Ortega-Gutiérrez et al ., 1992; Aranda-Gómez et al ., 2007) as part of the composition of the sedimentary units described previously, i.e ., Upper Cretaceous and Tertiary shales and sandstones.
Average data in % from the semiquantitative MASP procedure. *Basalts = Tholeiitic basalts. Data taken from Nageswara-Rao et al . (2012).
5.5. OM at the Barra del Tordo beach site
Ilmenite grains at the Barra del Tordo site are also derived from sedimentary sources such as the Upper Cretaceous and Tertiary shales and sandstones as part of the outcrops of the remnants of the Aldama volcanic field. These grains are homogeneous, rounded to subrounded, with an average range of 40 to 48 % of FeO and 49 to 56 % of TiO2, suggesting fresh ilmenite grains that come close to the composition of unaltered ilmenite minerals from sedimentary rocks (Morad and Aldahan, 1986). Beach processes such as breaking waves, longshore currents, and wind deflation control the transport and deposition of ilmenite grains onshore.
Titanomagnetite is also derived from the sedimentary units resulting from the erosion of subalkaline rocks (Vasconcelos et al ., 2002). The primary sources of titanomagnetite are probably subalkaline basalts low in K-feldspar with increased Ca-plagioclase (Aranda-Gómez et al ., 2007; Nageswara-Rao et al ., 2012; Kasper-Zubillaga et al ., 2013) (Tables 1 and 3). A subordinate low concentration of magnetite is observed in the Barra del Tordo beach and dune sands due to the absence of felsic volcanic mafic but dominantly acidic rocks as primary sources of the beach and dune sands, which can contribute to homogeneous magnetite grains (Grigsby, 1990).
5.6. OM at the Tepehuajes beach site
The OM concentration at the Tepehuajes beach site is the highest for beach and dune sands compared to the Miramar and Barra del Tordo beach locations. Major dispersion in the ilmenite plots is observed in the TiO2-FeO-Fe2O3ternary diagram, suggesting a large variation in the ilmenite grains probably due to the variable concentration of Fe-Ti oxides.
Ilmenite, titanomagnetite, and magnetite as subordinate components from the whole bulk composition of the beach and sand studied are derived from the Upper Cretaceous and Tertiary shales and sandstones units originated by the denudation of the Sierra de Tamaulipas, Aldama, and Sierra de Maratínez dominated by granites, trachytes, alkali basalts, and subalkaline basalts, i.e., toleiitic basalts (Ortega-Gutiérrez et al ., 1992; Vasconcelos et al ., 2002; Aranda-Gómez et al ., 2007) (Figures 6a, b).
A forward stepwise linear discriminant function analysis (LDFA) for the three beach sites showed that TiO2 is the variable that contributes to the separation of some mineral samples from the three sites (significance ρ = 0.00; Wilks' lambda = 0.97; tolerance = 0.90) (Figure 7). The variable concentration of TiO2 controls the dispersion in the composition of TiO2 and FeO in the ilmenite, titanomagnetite, and magnetite grains for the three beaches studied based on the LDFA. The LDFA performed for the three sites shows how some of the Tepehuajes OM samples plot slightly away from the rest of the overlapped samples compared to the rest of the beach samples.
6. Conclusion
Grain sizes in the sands from the western coast of the Gulf of Mexico are characterized by moderately to well-sorted fine sands controlled by a wide coastal plain. The beach and dune sands from the western coast of the Gulf of Mexico come from the Upper Cretaceous and Tertiary shales and sandstones exposed along the coast from the Miramar to the Tepehuajes beach sites. It is likely that eroded cycles from the Sierra de Tamaulipas composed of intrusive rocks produced the shale and sandstone units near the coast. Grain fractions such as sandstones, chert, monocrystalline quartz with straight extinction, K-feldspars, amphiboles and pyroxenes support the previous statement. The OM are composed of subordinate ilmenite, titanomagnetite, and magnetite derived from pre-existing eroded mafic sources, i.e ., alkali and subalkaline basalts from the Aldama and Sierra Maratínez volcanic field sites respectively and also from granites from the Sierra de Tamaulipas. Subrounded and rounded to very rounded OM grains like ilmenite, titanomagnetite, and magnetite, suggest longshore transport, breaking wave influence, and the wind deflation effect, from the source rock to the beach site.