The early Miocene was a period of major continental margin progradation in the Gulf of Mexico Basin that accompanied prominent tectonic and climatic changes in North America. However, sediment pathways from continental upland sources to deep basinal sinks remain poorly constrained. This study presents 2192 new detrital zircon U-Pb analyses from 19 Lower Miocene samples spanning the entire northern Gulf of Mexico margin to elucidate early Miocene sediment provenance and paleodrainage systems. The U-Pb age patterns indicate that the Great Plains, southern Rocky Mountains, and mid-Cenozoic volcanic field were the major source terranes for the western-central Gulf of Mexico coast, whereas the Appalachian foreland basin and Appalachian Mountains mainly contributed sediment to the eastern Gulf of Mexico coast. Local source terranes included the Llano uplift and Edwards Plateau in central Texas and the Ouachita Mountains and foreland basin in Oklahoma and Arkansas. A comparison to previous detrital zircon studies around the Gulf of Mexico indicates that sediment recycling was important during the early Miocene.
Sediment associated with major paleorivers, including the Rio Bravo, Rio Grande, Houston-Brazos, Red, Mississippi, Tombigbee, and Apalachicola Rivers, can be differentiated using the detrital zircon U-Pb analyses. These data help to better define the early Miocene source-to-sink system in the northern Gulf of Mexico, by relating the basin fill to hinterland tectonic and geological evolution. In comparison to the Paleocene–Eocene Wilcox drainage system, the early Miocene drainage system of the northern Gulf of Mexico was smaller and received less input from western Mexico arc terranes and Archean basement in Wyoming. This drainage area reduction, related to regional thermal uplift and Basin and Range–Rio Grande rifting, likely explains the reduced sediment volume of the Lower Miocene strata in the Gulf of Mexico relative to the Paleocene–Eocene Wilcox Group.
Early Miocene (ca. 15–23 Ma) sediment input to the Gulf of Mexico, derived from erosion of North American interior highlands, formed deep-water reservoirs. Many studies have investigated sediment accumulation and depositional facies in this terminal basin sink (Galloway et al., 2000; Zeng and Hentz, 2004; Galloway, 2008; Loucks et al., 2011). However, few provenance analyses have been performed on the Lower Miocene strata of the Gulf of Mexico, and there is much to be learned about the linkage between the deep-water sedimentary record and the evolution of the sediment source terranes.
Previous provenance studies of the Lower Miocene strata in the Gulf of Mexico were largely carried out using petrographic methods applied to sandstones (McBride et al., 1988; Dutton et al., 2012). Resulting ternary diagrams of sandstone components (QFL, QmFLt) allow for discrimination among arc, low-grade metamorphic, high-grade metamorphic, and sedimentary sources (e.g., Dickinson and Suczek, 1979). However, subsurface diagenetic alteration of detrital feldspars and heavy minerals affects such provenance interpretations (Milliken, 1988, 2007). Furthermore, enhanced weathering and attrition under relatively humid conditions as well as limited resolution power can prevent standard sandstone petrography from illuminating complex source terrane histories (Mackey et al., 2012).
By contrast, the mineral zircon is highly resistant to both chemical and physical weathering and is ubiquitous in sandstones. Detrital zircon grains are therefore ideal for providing provenance information by relating zircon crystallization age to potential source terranes (e.g., Carrapa, 2010; Gehrels, 2014). The detrital zircon U-Pb analytical method has been proven to be a useful tool for provenance and paleodrainage analysis (e.g., Dickinson and Gehrels, 2003, 2008; Lawton et al., 2009; Dickinson et al., 2009, 2010; Saylor et al., 2012; Benyon et al., 2014; Blum and Pecha, 2014).
Previous detrital zircon work on the Cretaceous Cenomanian and Paleocene–Eocene Wilcox Group strata in the northern Gulf of Mexico indicated that a continental-scale drainage reorganization occurred in the Mid-Cretaceous in response to the tectonic formation of the western U.S. Cordillera. This reorganization changed the pre-established westward sediment routing (from the Appalachians to the western United States) to the southward sediment delivery pattern that persists today (Blum and Pecha, 2014). Mackey et al. (2012) documented a sediment source contribution from the western Mexico arc terranes by examining detrital zircon age spectra on the Paleocene–Eocene Wilcox Group strata in south Texas. Craddock and Kylander-Clark (2013) suggested that the majority of sediment supply to the Cenozoic Mississippi River Delta in Louisiana originated from the Sevier-Laramide region of the western United States.
However, the routing of Lower Miocene siliciclastic sediment from hinterland source to basinal Gulf of Mexico sink has received little attention, and there have been no systematic detrital zircon analyses on these important strata at a regional basin scale, despite major tectonic reorganization occurring in the late Oligocene–early Miocene. The late Eocene to Oligocene was characterized by widespread volcanism in the southwestern United States and deposition of voluminous amounts of volcanic ash on the Gulf of Mexico coastal plain (Galloway, 1977, 1981; Galloway et al., 2011). While the mid-Cenozoic volcanism waned in the early Miocene, mantle upwelling and dynamic surface uplift continued to induce deep erosion in the late Oligocene–early Miocene on the southern Great Plains, the southern Colorado Plateau, part of northeastern Mexico, and in central-western Texas (Cather, 2011; Cather et al., 2012). Simultaneously, the Rio Grande Rift created N-S–trending fault blocks and extensional basins in the late Oligocene that both trapped sediment and played an important role in reshaping the topography, drainage basin configuration, and sediment routing in the southwest United States (e.g., Chapin and Cather, 1994; Galloway et al., 2011). As a result, the pre-established Eocene paleodrainage system in the southwestern United States was profoundly reorganized during the late Oligocene–early Miocene (Cather et al., 2008, 2012). The middle–late Miocene was dominated by Basin and Range extension in the western United States and northern Mexico (e.g., Stewart, 1998). Hence, the early Miocene was a critical transitional period that involved changes in regional tectonics, paleogeography, and paleodrainage across the North American Cordillera. Arid climatic conditions were prevalent in the western United States (Cather et al., 2008; Chapin, 2008; Fan et al., 2014) and extended to the northwestern Gulf of Mexico (Galloway et al., 1982). Locally, the Llano uplift and Edwards Plateau were uplifted and began to contribute sediment to the Gulf of Mexico. Given these profound changes, there is a critical need for systematic detrital zircon analyses from the Lower Miocene sedimentary strata of the Gulf of Mexico basin in order to understand the influence of these evolving tectonic, paleogeographic, and climatic conditions in the continental interior and the sediment routing from hinterlands to the basinal sinks.
The Gulf of Mexico is one of most important hydrocarbon-producing basins in the world. Sediment eroded from an enormous portion of the continental United States prograded across the continental shelf and was transported into this deep-water sink during the Miocene. However, much of the Lower Miocene deep-water sediment lies beneath the salt canopy and is difficult to correlate seismically to source areas. Therefore, detrital zircon work on the Lower Miocene rocks of the Gulf of Mexico coastal plain can provide a critical link between paleogeographic evolution in the source area and sediment stored in the basinal sink. This work also provides new insights into the evolution of the regional drainage system of North America from the Paleocene to early Miocene.
This study presents 2192 new detrital zircon U-Pb analyses from 19 sample locations along the northern margin of the Gulf of Mexico with the objective of understanding the complex sediment sources and delivery pathways in the critical Lower Miocene strata (Fig. 1). We also integrated these new results with published detrital zircon age spectra from: (1) the Cordillera foreland basin (Dickinson and Gehrels, 2009), (2) Paleocene–Eocene Wilcox Group strata from the Gulf Coast (Mackey et al., 2012; Blum and Pecha, 2014), (3) Upper Cretaceous–Paleogene Difunta strata in Laramide foreland basins in Mexico (Lawton et al., 2009), and (4) Paleozoic strata in the Appalachian foreland basin (Park et al., 2010). Detrital zircon U-Pb age spectra from these various strata and areas are important for understanding pre-Miocene paleogeography and paleodrainage systems and the shifts that occurred during the Miocene. Comparison of detrital zircon age spectra between these published data and our Lower Miocene data allows examination of the influence of tectonism and drainage organization on the evolution of sediment routing to the northern Gulf of Mexico.
The Gulf of Mexico is a small oceanic basin created by rifting within Pangea (e.g., Pindell, 1985; Stern and Dickinson, 2010; Van Avendonk et al., 2015). The basin was initially filled with thick salt prior to, and during, Late Jurassic seafloor spreading as seawater filled the restricted basin, and then accumulated both carbonate and siliciclastic sediment through the rest of the Mesozoic (Galloway, 2008; Hudec et al., 2013). High sea level during the mid-Late Cretaceous across the Western Interior Seaway limited siliciclastic input to the basin (Galloway, 2008). In contrast, the Cenozoic history of the Gulf of Mexico was marked by a dramatic increase in siliciclastic influx, including episodes of high sediment supply in the Paleocene–Eocene (Wilcox Group), Oligocene (Frio/Vicksburg Group), early Miocene, middle Miocene, and Pleistocene (Galloway, 2008; Galloway et al., 2011; Fig. 2).
The Paleogene Wilcox Group was largely fed from the central-southern Rocky Mountains, Cordillera arc terranes, and western Mexico basement terranes (Hutto et al., 2009; Mackey et al., 2012; Blum and Pecha, 2014). Oligocene Frio/Vicksburg Group strata are enriched in volcanic materials, mainly derived from southwestern U.S. and northern Mexico volcanic centers (Winker, 1982; Galloway et al., 2011). Therefore, western U.S. terranes were the dominant sediment sources for the Paleogene strata.
The early Miocene was a transitional period involving tectonic reorganization, climate change, and sediment redistribution in North America. Paleogeography in the western United States changed as a result of major tectonic events during the middle Cenozoic (Galloway et al., 2011). Regional crustal heating, uplift, and volcanism during the Oligocene were followed by Basin and Range extension in the middle Miocene (Stewart, 1998; Mack, 2004; Cather et al., 2012), leading to changes in sediment source provinces in western North America. Simultaneously, arid weather conditions extended to the northwest Gulf of Mexico margin, whereas a warm and humid climate began to characterize the eastern Gulf of Mexico coastal plain and Appalachian uplands (Galloway et al., 2011).
Lower Miocene strata in the northern Gulf of Mexico are marked by continental margin progradation that followed the Anahuac transgression (Fig. 2). These strata are capped by the major transgressive Amphistegina Shale. The Lower Miocene strata can be divided into two subunits, Lower Miocene 1 (LM1) and Lower Miocene 2 (LM2), separated by a transgressive shale dated at ca. 18 Ma that is continuously developed in the northwestern Gulf of Mexico (Galloway et al., 1986).
At the beginning of the early Miocene, sediment supply rate decreased relative to the underlying Oligocene Frio/Vicksburg strata (Fig. 2). However, by the late early Miocene, the rate of sediment supply began to increase in response to an increasing contribution from the Appalachians (Galloway, 2008; Galloway et al., 2011). During the middle Miocene, this eastern contribution peaked as rejuvenation of the Appalachians and consequent rapid erosion caused thick sediment accumulations to be deposited in both the eastern Gulf of Mexico and the western Atlantic (Poag and Sevon, 1989; Boettcher and Milliken, 1994; Galloway et al., 2011; Liu, 2014).
There are few provenance studies of the Lower Miocene sandstones in the Gulf of Mexico (McBride et al., 1988; Dutton et al., 2012). Published provenance work suggests a transition from volcanic-rock-fragment–rich Oligocene strata to more quartz-dominated Lower Miocene sandstone in the central Gulf of Mexico (Dutton et al., 2012). Petrographic analyses of sandstones from offshore Louisiana indicate a provenance similar to sediment carried by the modern Mississippi River, with source terranes from the Rocky Mountain, Appalachian, and Ouachita orogenic belts (McBride et al., 1988; Dutton et al., 2012).
STRATIGRAPHIC CONTEXT OF SAMPLES
Samples were collected from 16 outcrop exposures across the northern margin of the Gulf of Mexico (Fig. 1; Table DR1 in Data Repository1). In addition, three subsurface core samples were obtained from the core repositories of the Bureau of Economic Geology, University of Texas at Austin, and the Florida Geological Survey (Fig. 1; Table DR2 [see footnote 1]). Most of our well samples have limited biostratigraphic or geochronological control, but geophysical log correlations permit stratigraphic age assignments to sampled core intervals.
The stratigraphic ages of outcrop samples are constrained by published studies. Samples GOM2–7 (where GOM stands for Gulf of Mexico; Fig. 2) are from the Lower Miocene Oakville Formation of the Texas coastal plain, which was described by Galloway et al. (1982). GOM8 is from the basal part of the undifferentiated Miocene Fleming Formation that overlies the Oligocene Catahoula Formation in Jasper County, east Texas (Singleton, 2008). GOM9–13 are from the Miocene Catahoula, Lena, and Carnahan Bayou Formations in Louisiana. The Catahoula Formation in Louisiana is described as partly late Oligocene and early Miocene in age (Fig. 2), but our sample locations are mainly indicated as Lower Miocene by the Geological Map of Louisiana (Louisiana Geological Survey staff, 2008). GOM14–16 are from outcrops in Mississippi and are identified as belonging to the Lower Miocene Catahoula Formation based on published lithostratigraphies (Kolb and Durham, 1967; Dockery and May, 1981). Core samples GOM17 and 18 from the eastern Gulf of Mexico coast have been assigned to the undifferentiated Miocene Alum Bluff Group by the Florida Geological Survey. Samples were collected from the basal part of the Alum Bluff Group where it overlies the Oligocene Suwannee Limestone. The ages of these two samples are uncertain, and they could be middle Miocene. However, their zircon age spectra are similar to that of sample GOM19, obtained from the Chipola Formation along the modern Apalachicola River. GOM19 is the most precisely dated sample, with biostratigraphy and isotopic data indicating an early Miocene age, ca. 18.3 Ma (Bryant et al., 1992).
In total, 3 to 4 kg of material were collected from each outcrop, whereas 500–800 g were selected from subsurface cores. Core samples were obtained by cutting a 2.5-cm-wide longitudinal slice of core over a distance of several meters. Samples are from fluvial strata, mostly medium- to coarse-grained sandstones, with lesser fine-grained sandstones deposited in overbank environments. The samples collected from the northwestern Florida coastal plain are highly fossiliferous sandstones interbedded with limestone.
Standard mineral separation techniques, including heavy mineral and magnetic separation, were employed to extract zircon from surface and subsurface rock samples (see Data Repository for detailed procedures [footnote 1]). Separated zircon grains were sprinkled-mounted onto double-sided tape on 1 in. (2.54 cm) acrylic discs and analyzed at random using depth-profiling laser-ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) U-Pb geochronology. For each sample, at least 120 grains were analyzed to obtain a statistically robust provenance data set (Vermeesch, 2004). The analyses were completed using a PhotonMachine Analyte G.2 Excimer laser (30 μm laser spot size) with a large-volume Helex sample cell and a Thermo Element2 ICP-MS. GJ1 was used as the primary reference standard (Jackson et al., 2004), and Pak1 was used as a secondary zircon standard (thermal ionization mass spectrometry [TIMS] 206Pb/238U age of 43.0 Ma). Depth profiling of nonpolished, tape-mounted zircon grains enables the resolution of multiple zircon growth zones evident from core and rim ages (Stockli and Stockli, 2013). For analyzed detrital zircon grains, the 206Pb/238U age is used for grains younger than 1000 Ma, and 207Pb/206Pb age is used for grains older than 1000 Ma. For zircon U-Pb ages older than 1000 Ma, the discordance was calculated based on 206Pb/238U and 207Pb/206Pb ages. For zircon ages younger than 1000 Ma, the discordance was calculated based on 206Pb/238U and 207Pb/235U ages. All grains older than 1000 Ma with >20% discordance and grains younger than 1000 Ma with >10% discordance were discarded, as well as grains with >10% analytical error. The detailed analytical methods and raw data are incorporated in the ancillary Data Repository files (see footnote 1). All detrital zircon U-Pb geochronology analyses were carried out at the UTChron Geo- and Thermochronometry Laboratory at the University of Texas at Austin.
DETRITAL ZIRCON RESULTS
This study presents 2192 new detrital zircon U-Pb results from 19 samples. The data display large age variability, ranging from late Oligocene (ca. 24 Ma) to Archean (ca. 3600 Ma; Fig. 3). The spectra are divided into seven age components, based on a combination of ages of major basement provinces and magmatic events in North America and age peaks in our data (Fig. 3): Cordillera Magmatic Province (A; 24–280 Ma), Appalachian-Ouachita Province (B; 280–500 Ma), Pan-Africa Province (C; 500–700 Ma), Grenville Province (D; 950–1300 Ma), Midcontinent Province (E; 1300–1500 Ma), Yavapai-Mazatzal Province (F; 1600–1800 Ma), and Shield Province (G; >1800 Ma). Although the nomenclature for each group is interpretive, we do not necessarily make the assumption that zircon grains of the appropriate ages are from these provinces, but we simply use the province nomenclature to designate zircon age groups.
Cordillera Magmatic Province (component A) is composed of late Paleozoic to Cenozoic zircon grains and makes up 31.8% of the total population in all samples. It is the largest component for samples from the western Gulf of Mexico (GOM1–13; Fig. 3; Table 1). However, the percentage of Cordillera Magmatic Province grains gradually decreases eastward, and it constitutes only a minor component in samples GOM14–19. The Cordillera Magmatic Province component can be further subdivided into six subgroups, A1–A6 (Fig. 4). Middle Permian to Late Triassic zircon grains (subgroup A6; 270–210 Ma) account for 10.3% of total Cordillera Magmatic Province grains and have age peaks around 230 Ma in samples GOM1, 2, 6, and 7, and peaks around 214 Ma in samples GOM4 and 5 (Fig. 4). Jurassic grains (subgroup A5; 180–150 Ma) are relatively abundant in samples GOM1–13 with several peaks around 160 Ma, 165 Ma, and 173 Ma. Zircon ages from 140 to 110 Ma are rare (3.7%), resulting in a significant age gap between subgroups A5 and A4 (Fig. 4). Early Late Cretaceous zircon grains (subgroup A4; 110–85 Ma) are an important component in samples GOM1, 4, 5, and 7, with peaks around 95 Ma. Late Late Cretaceous zircon grains (subgroup A3; 85–65 Ma) show a common age peak around 75 Ma, while Paleogene zircon grains (subgroup A2; 65–40 Ma) show age peaks around 57 Ma and 46 Ma. Late Eocene to Oligocene zircon grains (subgroup A1; 40–24 Ma) are represented by ages peaks around 35 Ma and 27 Ma (Fig. 4).
Appalachian-Ouachita Province (B; 280–500 Ma) constitutes 8.3% of the total population in all samples. It is an important component for samples GOM15–19 from the eastern Gulf of Mexico (Fig. 3; Table 1). The Pan-Africa Province (C; 500–700 Ma) is a minor component in all samples (Table 1). The Grenville Province (D; 950–1300 Ma) is prevalent in all samples and features multiple age peaks; it accounts for 28% of the total population in all samples. Grenville-age zircon grains are most abundant in samples GOM14–19 from the eastern Gulf of Mexico. It is also an important component of several samples from the western Gulf of Mexico; e.g., GOM4 contains 29% Grenville-age zircon grains (Table 1). However, the proportion of Grenville-age zircon grains in western Gulf of Mexico samples is much lower than in samples from the eastern Gulf of Mexico. Midcontinent (E; 1300–1500 Ma) and Yavapai-Mazatzal (F; 1600–1800 Ma) zircon grains are present in most samples. They are most abundant in GOM8–13 from the central Gulf of Mexico and less abundant in samples from both the eastern and western Gulf of Mexico. Shield Province (G; >1800 Ma) zircon grains are a minor component, and few zircon grains display this age (<6% of each sample).
The multidimensional scaling (MDS) ap-proach was used to detect dissimilarity between the U-Pb age spectra of the samples. The MDS method is based on the widely used Kolmogorov-Smirnov (K-S) test, but it provides better visualization of the similarity or dissimilarity between samples (Vermeesch, 2013). As indicated by the MDS plot (Fig. 5), 19 samples fall within four main groups. These groups correlate with the geographic positions of samples in known depocenters (Fig. 1) and are assigned the following names: Greater Rio Grande Embayment, Houston Embayment, Mississippi Embayment, and Eastern Gulf of Mexico Embayment. Samples GOM1 and GOM4 plot outside the Greater Rio Grande Embayment group (Fig. 5). This is because of differences in their components of Appalachian-Ouachita (GOM1), Pan-Africa (GOM1), and Grenville (GOM4; Fig. 3; Table 1) Provinces. However, given the geographic locations of these samples and high similarity of Cordillera Magmatic Province, Midcontinent, and Yavapai-Mazatzal components to other Great Rio Grande Embayment samples (Fig. 3; Table 1), we regard all of samples GOM1–7 as a single group. The dissimilarities of GOM1 and 4 are discussed in the following.
The Greater Rio Grande Embayment samples are characterized by high proportions of Cordillera Magmatic Province zircon grains and moderate proportions of Grenville, Midcontinent, and Yavapai-Mazatzal components (Fig. 3; Table 1). The Houston Embayment group includes six samples (GOM8–13) and shows a similar age pattern to the Greater Rio Grande Embayment group, but it shows a trend of decreasing Cordillera Magmatic Province component and increasing content of Midcontinent and Yavapai-Mazatzal components from west to east (Fig. 3; Table 1).
Significant changes occur between the Houston Embayment and Mississippi Embayment groups. The Mississippi Embayment consists of two samples (GOM14–15), showing a rapid increase of Grenville component and decrease of Cordillera Magmatic Province component relative to the Houston Embayment (Fig. 3; Table 1). In comparison to both the Greater Rio Grande and Houston Embayment, the Cordillera Magmatic Province becomes only a moderate or even minor component. Midcontinent and Yavapai-Mazatzal inputs still play important roles, but they lose importance eastward. The Eastern Gulf of Mexico Embayment consists of samples GOM16–19 and is dominated by Appalachian-Ouachita and Grenville components (Fig. 3; Table 1).
Cordillera Magmatic Province (A; 24–280 Ma)
Late Paleozoic to middle Cenozoic zircon grains (24–280 Ma) coincide with documented magmatic activity in the Western Cordillera arc and its predecessors (Chen and Moore, 1982; Ducea, 2001; DeCelles, 2004; DeCelles et al., 2009; Fig. 6). The oldest igneous activity (245–280 Ma, corresponding to subgroup A6) is recorded in isolated Permian igneous centers across northern Sonora in Mexico and plutons in the East Mexico arc (Fig. 7; Torres et al., 1999; Dickinson et al., 2010). Two younger peaks of magmatic activity at 150–160 Ma (subgroup A5) and 85–100 Ma (subgroup A4) are documented in the Sierra Nevada Batholith (Fig. 7; Chen and Moore, 1982; Ducea, 2001; DeCelles et al., 2009).
The Coast Mountains Batholith in the northwestern United States reflects magmatism in the Late Jurassic (150–160 Ma), Early to Late Cretaceous (85–110 Ma), and Late Cretaceous to Paleogene (50–80 Ma), encompassing subgroups A2–A5 (DeCelles et al., 2009; Paterson et al., 2011). Early Cretaceous (110–150 Ma) igneous activity mainly occurred in southernmost California and northern Baja Peninsula (Fig. 7; Dickinson and Lawton, 2001), but it is not common in the southwestern United States. Late Cretaceous (80–100 Ma) magmatism can be also found in the Baja Peninsula (Mackey et al., 2012, and references therein).
Late Cretaceous to middle Eocene (40–80 Ma, subgroups A2 and A3) zircon grains are related to the Laramide orogeny, coinciding with rapidly increasing convergence between the North American and Farallon plates (DeCelles, 2004). Magmatic intrusions (dated at 40–75 Ma) are documented in the Laramide porphyry copper province in southern Arizona, southwestern New Mexico, northern Mexico, and the Colorado mineral belt (Fig. 7; Tweto and Sims, 1963; Cunningham et al., 1994; Chapin et al., 2004; Barra et al., 2005; Chapin, 2012). The youngest subgroup (A1; 24–40 Ma) corresponds to the mid-Cenozoic ignimbrite flare-up that extends from Trans-Pecos Texas to central Colorado, including the San Juan, Mogollon-Datil, Trans-Pecos, and Sierra Madre Occidental volcanic fields (Ferrari et al., 1999; Chapin et al., 2004; Ferrari et al., 2007; Fig. 7).
Appalachian-Ouachita Province (B; 280–500 Ma)
Paleozoic zircon grains (280–500 Ma) correspond to the Appalachian-Ouachita orogeny along the eastern margin of North America, which formed during assembly of the supercontinent Pangea (Fig. 6). The age peaks in the Appalachian-Ouachita Province correspond to three major tectonic accretion events, including the Taconic orogeny (440–465 Ma), Acadian orogeny (350–380Ma), and Alleghanian orogeny (265–327 Ma; Eriksson et al., 2003; Thomas et al., 2004; Hatcher, 1987, 2010; Park et al., 2010; Thomas, 2011). Although the Appalachian-Ouachita terranes are widely distributed, extensive exposures only occur in the Central and Southern Appalachian Mountains, whereas more restricted exposures can be found in the Ouachita Mountains in Oklahoma and Marathon uplift in west Texas.
Pan-Africa Province (C; 500–700 Ma)
Late Proterozoic to early Paleozoic zircon grains (500–700 Ma) may be derived from Iapetan rifting, magmatism along the eastern Laurentia margin, and/or the Pan-African orogeny, which built supercontinent Gondwana (Hoffman, 1989; Park et al., 2010). Zircon grains originating from these tectonic processes are usually from the Suwannee, Carolina, and Avalon terranes in the eastern United States (Park et al., 2010). Another possible source of component C is the Amarillo-Wichita uplift, which features a sharp Cambrian age peak at ca. 516 Ma (Riggs et al., 1996; Dickinson et al., 2010).
Grenville Province (D; 950–1300 Ma)
Middle Mesoproterozoic to early Neoproterozoic zircon grains (D; 950–1300 Ma) are derived from Grenville basement in southern and eastern Laurentia (Fig. 6), formed during a Mesoproterozoic orogeny, contemporary with the assemblage of supercontinent Rodinia (Hoffman, 1989; Dickinson and Gehrels, 2009). The Appalachian terranes in the eastern United States have exposures of Grenville basement. Zircon grains from this area usually feature two major age peaks at 1020–1090 Ma and 1140–1190 Ma, which correspond to the Ottawan orogeny and Shawinigan orogeny, respectively (Rivers et al., 2002; Rivers, 2008).
In addition, the Llano uplift in Texas is a major Grenville basement exposure in southern Laurentia with three major phases of magmatic activity: 1232–1288 Ma, 1120–1150 Ma, and 1070–1120 Ma (Mosher, 1998). There are also limited exposures of Grenville basement in the Franklin Mountains in west Texas (Amato and Mack, 2012).
Midcontinent Province (E; 1300–1500 Ma)
Zircon grains of 1300–1500 Ma represent Mesoproterozoic A-type granite-rhyolite in midcontinental North America (Fig. 6; Bickford et al., 1986; Van Schmus et al., 1996). However, Phanerozoic cover buried most of the basement except for an exposure in the St. Francois Mountains (Bickford et al., 1986). Therefore, more likely sources for zircon grains of this age are the numerous granitic intrusions in the southwestern United States (Fig. 6; Bickford et., 1986; Hoffman, 1989; Karlstrom et al., 1997, 2004).
Yavapai-Mazatzal Province (F; 1600–1800 Ma)
Yavapai-Mazatzal Province corresponds to Yavapai (1700–1800 Ma) and Mazatzal (1600–1700 Ma) basement in southwestern Laurentia (Hoffman, 1989; Karlstrom et al., 2004; Whitmeyer and Karlstrom, 2007). Yavapai-Mazatzal basement was elevated during the Laramide orogeny and formed the core of the southern Rocky Mountains, which are therefore the most likely direct source for zircon grains with this age range. Alternative sources of Yavapai-Mazatzal grains would be recycling from older sedimentary strata in the western United States, e.g., Carboniferous–Permian (Gehrels and Pecha, 2014) and Permian–Jurassic (Dickinson and Gehrels, 2003) rocks, which had originally received zircon grains from the local Ancestral Rocky Mountains province during the late Paleozoic.
Shield Province (G; >1800 Ma)
Pre–1800 Ma zircon grains were formed by Trans-Hudson, Wyoming, and Superior orogenic events that formed the Precambrian Canadian Shield and the North American craton (Whitmeyer and Karlstrom, 2007; Fig. 6). Modern exposures of these Proterozoic and Archean basements are in the northern United States and Canadian Shield region (Laskowski et al., 2013). Uplifts in Wyoming, such as the Wind River Range (2500–2700 Ma; Frost et al., 2000), could be possible sources for zircon grains in this age range, because the adjacent Wind River Basin records rapid uplift and erosion of these highlands in the Paleocene–early Eocene (Fan et al., 2011).
Detrital zircon U-Pb age spectra of 19 Lower Miocene samples from the Gulf of Mexico display high west-to-east variability. The geochronological data show that each of the four distinct groups (Greater Rio Grande Embayment, Houston Embayment, Mississippi Embayment, and Eastern Gulf of Mexico Embayment) represents a unique provenance and thus provides key insights into drainage system development and sediment routing.
Greater Rio Grande Embayment Provenance
Detrital zircon U-Pb data for the Greater Rio Grande Embayment consist of seven samples from the south-central Texas coastal plain and are characterized by a high proportion of Cordillera Magmatic Province component and moderate proportions of Grenville, Midcontinent, and Yavapai-Mazatzal components (Fig. 3; Table 1). Zircon grains of 24–40 Ma (subgroup A1) correspond to the period of mid-Cenozoic magmatism in the southwestern United States. Volcanic fields, including the San Juan volcanic field in southwestern Colorado, Mogollon-Datil volcanic field in western New Mexico, Trans-Pecos volcanic field in western Texas and northern central Mexico, and Sierra Madre Occidental in western Mexico, are the likely ultimate sources (Chapin et al., 2004; Ferrari et al., 2007; Cather et al., 2008). However, the Rio Grande Rift extension developed since the late Oligocene would have precluded large volumes of volcanic materials from volcanic fields to the west entering the Gulf of Mexico. Only the Trans-Pecos volcanic field is a likely direct volcanic zircon source (Fig. 8). However, some older zircon grains within this component might be derived from Eocene–Oligocene strata on the Great Plains and Texas coastal plain, to which large volumes of volcanic ash were transported by northeastward-blowing winds (e.g., Larson and Evanoff, 1998). The presence of mid-Cenozoic volcanic zircon grains is also evident by the abundance of volcanic rock fragments in sediment deposited in the Rio Grande Embayment (Galloway et al., 1982, 2000, 2011; Dutton et al., 2012).
Pre–40 Ma Cordillera Magmatic Province zircon grains correspond to major Permian–Triassic arc, Sierra Nevada Batholith, and Laramide-related magmatic centers to the west of Laramide uplifts. However, the uplifted mid-Cenozoic volcanic fields along the eastern Colorado Plateau margin formed a topographic barrier to eastward sediment transportation (Cather et al., 2008, 2012). Consequently, the drainage systems, which previously carried sediment eastward to the Gulf of Mexico in the Paleocene and Eocene, were reorganized by the mid-Cenozoic volcanism, resulting in a relatively closed drainage system with a fluvial outlet only to the Pacific Northwest (Cather, 2011; Galloway et al., 2011; Cather et al., 2012; Fig. 8). In addition, the initiation of the Rio Grande Rift in the late Oligocene–early Miocene created series of extensional basins, which stored sediment eroded from the southern Colorado Plateau (Chapin and Cather, 1994). Although the volume of Lower Miocene sediment stored in the Rio Grande Rift basin is limited (Chapin and Cather, 1994), it is likely that large volumes of sediment eroded from various Cordillera arc terranes were no longer able to reach the Gulf of Mexico as result of the rift-related drainage basin dismemberment.
All of these factors suggest that the late Paleozoic to Paleogene zircon grains (40–280 Ma; subgroups A2–A6) in Lower Miocene Gulf of Mexico strata were probably not directly sourced from western Cordillera arc terranes. This interpretation is supported by dramatic differences in detrital zircon U-Pb age spectra between the Lower Miocene Gulf of Mexico coastal plain strata and Mesozoic eolian strata on the Colorado Plateau (Dickinson and Gehrels, 2009; Fig. 9A). Major differences exist in several age components, especially for the 300–700 Ma component, which is abundant in Colorado Plateau Mesozoic strata but nearly absent on the Gulf of Mexico coast. Therefore, it is unlikely that a large volume of sediment reached the Gulf of Mexico from the Colorado Plateau region.
The zircon grains with 40–280 Ma ages are instead more likely recycled from strata previously deposited on the southern Great Plains, western and central Texas, and northeastern Mexico. These strata are Cretaceous–Early Cenozoic erosional products that were derived from the Permian–Triassic arc, Sierra Nevada Batholith, and Laramide-related magmatic terranes and transported by eastward-flowing paleorivers (Dickinson et al., 1988; Cather et al., 2012). Deep erosion occurred on the southern Great Plains and southern Rocky Mountains during the late Oligocene–early Miocene in response to increased mantle buoyancy induced by the mid-Cenozoic volcanism in southwestern North America (Roy et al., 2004; Eaton, 2008; Cather, 2011; Cather et al., 2012). This deep erosion removed ∼1.5 km of Upper Cretaceous–Paleogene strata on the southern Great Plains of Texas, New Mexico, and northeastern Mexico (Kelley and Chapin, 1995; Cather et al., 2012, and references therein). Sediment eroded from such older strata probably comprise an important component of the Lower Miocene strata along the Texas coastal plain, because of the high similarity between U-Pb age spectra of the Lower Miocene strata and Paleocene–Eocene Wilcox Group strata (Fig. 9B).
This similarity is most marked in central Texas, although Yavapai-Mazatzal Province and Shield Province components are more abundant in the Wilcox strata (Fig. 9B). In contrast, detrital zircon age spectra of the Wilcox strata in southwestern Texas (Mackey et al., 2012) are dissimilar to those of the Lower Miocene strata (Fig. 9C). For example, the Wilcox strata U-Pb data from Mackey et al. (2012) show a continuous pattern from 50 Ma to 280 Ma without significant age gaps, while 110–150 Ma grains are almost entirely missing from the Lower Miocene strata. In addition, the Grenville Province component is much less common in the Wilcox strata of southwestern Texas (Fig. 9C). Large differences also exist between the Upper Cretaceous–Paleogene Difunta strata in the foreland basin of the Sierra Madre Oriental (Lawton et al., 2009) and the Lower Miocene strata (Fig. 9D). The former contains abundant 110–150 Ma and 230 Ma zircon grains, whereas the Lower Miocene strata have rare zircon grains in those ranges. Therefore, the Lower Miocene strata may be partially recycled from the Wilcox Group strata in central Texas (Fig. 9B), but the Laramide foreland basins in Mexico and the Wilcox Group of southwestern Texas are unlikely to be significant sources (Figs. 9C and 9D).
The original source of 110–150 Ma zircon grains is most likely the Peninsular Ranges Batholith in southernmost part of California, northern Baja Peninsula, and north-central Sonora (Dickinson and Lawton, 2001; Lawton et al., 2009; Mackey et al., 2012). Magmatism in this age interval is not common in other terranes in the southwestern United States. Therefore, the absence of zircon grains of 110–150 Ma age in the Lower Miocene strata suggests a reduced catchment size compared to the Paleogene Wilcox Group, with cutoff of sediment supply from southernmost California, northern Baja Peninsula, and western Mexico (Fig. 8).
Grenville zircon grains (D; 9500–1300 Ma) in the Greater Rio Grande Embayment probably indicate a recycled source. Although Grenville basement exposures in the southwestern United States are limited, Grenville-affinity grains are commonly found in several sedimentary basins, e.g., the Late Jurassic to Early Cretaceous Bisbee and McCoy basins (Dickinson et al., 2009), Mesozoic eolian basins on the Colorado Plateau (Dickinson and Gehrels, 2003, 2008, 2009), and Paleocene–Eocene Wilcox Group in the Gulf of Mexico (Blum and Pecha, 2014). Most Rio Grande Embayment samples (GOM1–3 and 5–7) do not show a prominent increase of Grenville zircon grains in the Lower Miocene strata relative to the Paleogene Wilcox Group (Blum and Pecha, 2014; Fig. 9B), suggesting that their Grenville-component zircon grains are possibly recycled from these strata. Only sample GOM4 displays a high proportion of Grenville component (Fig. 3; Table 1), indicating local input of Grenville grains. The main such primary contributor of Grenville grains in Texas was the Llano uplift. About 1.3 km of the Llano uplift and Edwards Plateau were unroofed starting in the early Miocene (Corrigan et al., 1998; Ewing, 2005).
Midcontinent (E; 1300–1500 Ma) and Yavapai-Mazatzal (F; 1600–1800 Ma) are additional important components in the Greater Rio Grande Embayment samples. They can either come from Yavapai-Mazatzal basement exposed by the southern Laramide uplift, or be recycled from sedimentary strata such as the Mississippian–Permian (Gehrels and Pecha, 2014), Permian–Jurassic (Dickinson and Gehrels, 2003), and Paleocene–Eocene Wilcox Group strata (Blum and Pecha, 2014). Compared to the Paleogene Wilcox Group strata (Blum and Pecha, 2014; Fig. 9B), the Yavapai-Mazatzal and Shield components in the Lower Miocene samples are smaller. This contrast probably indicates that the Lower Miocene and Paleocene–Eocene Wilcox Group strata shared similar sediment source terranes, but with a decreased sediment supply from Laramide uplifts and Archean basement exposures in Wyoming. It is also possible that the Lower Miocene strata are largely recycled from the Paleocene–Eocene Wilcox strata, with an extra input of younger grains that dilutes the Yavapai-Mazatzal and Shield components.
Few zircon grains are present from the Appalachian-Ouachita Province (B; 300–500 Ma), Pan-Africa Province (C; 500–700 Ma), and Shield Province (G; >1800 Ma). No basement sources are available nearby for direct supply of these grains. Therefore, those present are probably partially recycled from older strata, e.g., the Wilcox Group (Blum and Pecha, 2014). Sample GOM1 has a small U-Pb age peak around 413 Ma, which might be recycled from the Marathon region in west Texas, where this peak is recorded in the Pennsylvanian Haymond Formation (Gleason et al., 2007). Therefore, it is possible that sample GOM1 records slightly different source terranes, perhaps including a tributary from the Marathon uplift in west Texas.
In summary, the Lower Miocene samples from the Greater Rio Grande Embayment record large volumes of zircon grains that were likely recycled from older strata on the southern Great Plains, western and central Texas, and northeastern Mexico. The mid-Cenozoic volcanic fields, particularly the Trans-Pecos volcanic field, probably provided the majority of first-cycle volcanic zircon grains. The Llano uplift in central Texas contributed sediment in a limited area. By comparison with two previous studies (Lawton et al., 2009; Mackey et al., 2012) on the Paleogene strata in the Gulf of Mexico region, the lack of 110–150 Ma zircon grains indicates that sediment sourced from western Mexico arc terranes was cut off by the early Miocene (Figs. 9C and 9D). Compared to the Paleocene–Eocene Wilcox strata deposited on the central Texas coastal plain (Fig. 9B; Blum and Pecha, 2014), the Lower Miocene strata have fewer grains from Yavapai-Mazatzal and Wyoming-Superior terranes, indicating either that sediment supply from the Rocky Mountains and Wyoming-Superior terranes had decreased or that Lower Miocene strata were largely recycled from older strata on the southern Great Plains (Fig. 9B). Future application of detrital zircon U-Pb and (U-Th)/He double dating to individual zircon crystals may help to resolve this uncertainty.
Houston Embayment Provenance
The Houston Embayment (samples GOM8–13), centered near the Texas-Louisiana boundary, is distinguished from the Greater Rio Grande Embayment by a decrease of Cordillera Magmatic Province zircon grains and a marked increase of Midcontinent and Yavapai-Mazatzal zircon grains (Fig. 3; Table 1). Cordillera Magmatic Province zircon grains are still an important component of the Houston Embayment group, indicating similar sources to those of the Greater Rio Grande Embayment. However, the influence of these sources is reduced in the Houston Embayment.
Midcontinent and Yavapai-Mazatzal zircon grains become pronounced components of the Houston Embayment group (Fig. 3; Table 1). Midcontinent and Yavapai-Mazatzal zircon grains are most likely derived from Laramide basement uplifts in the southwestern United States (Bickford et al., 1986; Hoffman, 1989; Karlstrom et al., 1997, 2004; Fig. 6). A significant increase of Yavapai-Mazatzal grains relative to the Greater Rio Grande Embayment indicates that source terranes for the Lower Miocene strata may have shifted northward to the major southern Laramide uplifts where the Paleozoic and Mesozoic cover strata had been eroded and Precambrian Yavapai-Mazatzal basement core exposed (Tweto, 1975).
Similarity exists between the Lower Miocene strata and the Paleogene Wilcox Group in the Houston Embayment (Fig. 10A). This indicates that the Lower Miocene strata are either recycled from the Wilcox strata or that they come from similar source terranes to those of the Wilcox. However, the proportion of Yavapai-Mazatzal zircon grains in the Lower Miocene strata (24.8%) is larger than the proportion of Yavapai-Mazatzal component in the Wilcox Group (19.5%). The Proterozoic basement cores of Laramide uplifts were exposed during the late Paleocene to Eocene, as indicated by the Precambrian igneous and metamorphic clasts preserved in adjacent lacustrine basins (e.g., Tweto, 1980; Carroll et al., 2006). Apatite fission-track thermochronology studies on the southern Rocky Mountains indicate that >2 km of material were eroded during the late Oligocene–early Miocene from the Sangre de Cristo Mountains in New Mexico (Kelley and Chapin, 1995; Pazzaglia and Kelley, 1998). These observations suggest that erosion of uplifted Laramide cores, particularly the Sangre de Cristo Mountains, added increasing amounts of Yavapai-Mazatzal zircon grains to the Lower Miocene strata relative to the Paleocene–Eocene Wilcox strata.
However, differences in Yavapai-Mazatzal components between the Wilcox and Lower Miocene strata are not great, and resulting interpretations should be treated with caution. For example, the variation could be caused by the limited data set (i.e., ∼100 grains in each sample). Yavapai-Mazatzal component zircon grains may be derived from a mixed source that included both Laramide basement and sedimentary strata in the western United States.
The Shield component is an important portion (13.5%) of the Paleocene–Eocene Wilcox strata, indicating a sediment contribution from basement uplifts in Wyoming. In contrast, the Lower Miocene strata lack a large Shield component (Fig. 10A; Table 1).
Generally, the Houston Embayment records a decrease in the impact of sources from the southern Great Plains, western and central Texas, northeasternmost Mexico, and the Trans-Pecos volcanic field. The Houston Embayment source terranes extended to the major southern Laramide uplifts, as reflected by the large component of Yavapai-Mazatzal zircon grains contained in samples GOM8–13. Compared to the Wilcox Group, the Houston Embayment contains fewer zircon grains from basement uplifts in Wyoming.
Mississippi Embayment Provenance
The Mississippi Embayment group shows a dramatic difference in U-Pb age patterns with respect to the Houston Embayment, with a sharp increase in the Grenville component and decrease in the Cordillera Magmatic Province component (Fig. 3; Table 1), indicating a large contribution from the Appalachian terranes in the east and reduced contribution from the western U.S. terranes. Although Grenville zircon grains are common in Mesozoic–Cenozoic basins in western North America (e.g., Patchett et al., 1999; Dickinson and Gehrels, 2008, 2009; Leier and Gehrels, 2011; Blum and Pecha, 2014), dominance to the degree found in the Mississippi Embayment (>50%) is mostly found in the Appalachian regions (e.g., Eriksson et al., 2003, 2004; Park et al., 2010).
This change is coupled with an increase in the Appalachian-Ouachita component (300–500 Ma), which is likely derived from exposures of the Paleozoic fold-and-thrust belt in the Appalachian terranes. Exposure of the Appalachian-Ouachita terranes in the Ouachita Mountains could be an alternative source for these grains. Detrital zircon studies on the Cenomanian Woodbine-Tuscaloosa Formation in the Gulf of Mexico region (Blum and Pecha, 2014) and the Cenomanian Dakota Formation in western Iowa and eastern Nebraska (Finzel, 2014) suggest that the Ouachita Mountains and foreland basin are potential major sources for both Appalachian-Ouachita and Grenville zircon grains. Previous petrographic studies on the Lower Miocene strata also indicate the Ouachita Mountains as an important local source terrane (e.g., Galloway et al., 2000, 2011; Dutton et al., 2012).
Midcontinent and Yavapai-Mazatzal components in the Mississippi Embayment are also reduced relative to the Houston Embayment (Fig. 3; Table 1). However, Cordillera Magmatic Province, Midcontinent, and Yavapai-Mazatzal components still make up ∼30% of the total detrital zircon population in the Mississippi Embayment, indicating a continued strong western U.S. influence. Paleostream and sub-Ogallala paleovalley trend (as discussed in next section) studies indicate fluvial transport of such sediment from Laramide uplift regions eastward to the Great Plains (Bart, 1975; Scott, 1975; Fig. 11), where it was finally collected by the paleo–Mississippi River (Galloway et al., 2011).
A comparison of the Mississippi Embayment of the Lower Miocene strata to the Wilcox Group strata located to the north (updip; Blum and Pecha, 2014; Figs. 8 and 10B) shows them to be very similar, indicating that the Lower Miocene strata were probably either sourced from recycled Paleocene–Eocene Wilcox strata or originated from the same drainage area. The Shield component is more abundant in the Wilcox Group strata, possibly indicating a stronger connection to Wyoming basement uplifts in the Paleocene–Eocene than during the early Miocene. This can be partially explained by climatic changes. Paleostream analysis of Miocene Arikaree strata of the High Plains of western Nebraska and eastern Wyoming suggests that paleorivers carried sediment from the Laramie Range (Archean basement) eastward to the Great Plains (Bart, 1975). However, it is probable that the arid climate that prevailed in the western United States during the late Oligocene and early Miocene diminished the capacity for fluvial sediment transport (Galloway et al., 2011).
In summary, the Lower Miocene strata in the Mississippi Embayment record a sharp increase in sediment delivered from the Appalachian terranes relative to the Houston Embayment (Fig. 8), but Great Plains and Laramide uplifts were still important contributors of sediment. Lower Miocene strata might also include recycled Wilcox strata, but probably with less contribution from the Archean basements uplifts of the northern Rocky Mountains than is the case for the Wilcox Group strata.
Eastern Gulf of Mexico Embayment Provenance
The Appalachian-Ouachita and Grenville components (Fig. 3; Table 1) dominate the Eastern Gulf of Mexico Embayment samples, with only minor Pan-Africa, Midcontinent, Yavapai-Mazatzal, and Shield components, and Cordillera Magmatic Province zircon grains are completely absent. The predominance of the Appalachian-Ouachita and Grenville components indicates a proximal source from the Grenville basement and Paleozoic fold-and-thrust belt in the Appalachian terranes (Fig. 8).
Our Grenville age spectra basically match the four pulses of orogeny in the Appalachian terranes: Elzeverian (1190–1250 Ma), Shawinigan (1140–1190 Ma), Ottawan (1020–1080 Ma), and Rigolet (980–1010 Ma; Rivers et al., 2002; Rivers, 2008). Eriksson et al. (2004) suggested that Grenville zircon grains in Neoproterozoic through Pennsylvanian sedimentary rocks of the Central Appalachian Basin have a bimodal distribution with distinct age peaks between 1020 and 1080 Ma (Ottawan) and 1140–1190 Ma (Shawinigan). This age pattern can also be seen in Paleozoic strata in the Appalachian foreland basin (Park et al., 2010) and the Paleogene Wilcox Group strata on the eastern Gulf of Mexico coastal plain (Blum and Pecha, 2014). These major Ottawan and Shawinigan peaks are also present in our data, and the overall patterns are similar (Figs. 3, 10C, and 10D).
Paleozoic grains are the second important component in samples from the Eastern Gulf of Mexico Embayment and constitute 17%–31% of each sample (Table 1). Sample GOM16 has a single prominent age peak corresponding to the Taconic orogeny (ca. 430–500 Ma), while GOM17–19 record two additional peaks related to the Acadian (ca. 350–400 Ma) and Alleghanian (ca. 265–325 Ma) orogenies. Taconic plutons are mainly exposed in the Chopawamsic and Milton terranes of the Central Appalachians, while Acadian and Alleghanian rocks are widely developed in the Inner Piedmont of the Southern Appalachian terranes (Eriksson et al., 2003). The increase of Acadian and Alleghanian components in GOM17–19 indicates more materials sourced from the proximal Southern Appalachian terranes.
Marked similarity can be observed between the Eastern Gulf of Mexico Embayment Lower Miocene and Paleozoic strata of the Appalachian foreland basin (Park et al., 2010; Fig. 10C) and the Wilcox samples from the eastern Gulf of Mexico coastal plain (Blum and Pecha, 2014; Fig. 10D). These similarities likely indicate multiple recycling events of older strata. However, the two high peaks exhibited by samples GOM17–19 and corresponding to the Acadian and Alleghanian orogenies indicate a proximal source from the Southern Appalachian terranes (Fig. 8). Because the Pan-Africa component (C; 500–700 Ma) is rare, the peri-Gondwanan terranes (e.g., Carolina terrane) are excluded as a possible source of the Lower Miocene strata in the Eastern Gulf of Mexico Embayment.
Drainage Systems of the Early Miocene in the Gulf of Mexico Basin
The early Miocene was an important transitional interval of tectonic reorganization in North America for which the drainage systems are not well known (Cather et al., 2012). Major drainage reorganization occurred on the Colorado Plateau as the mid-Cenozoic volcanism disrupted the preexisting eastward-flowing Eocene drainage system and reversed the flow direction to northwestward toward the Pacific Northwest (Spencer et al., 2008; Cather et al., 2008, 2012). The Rio Grande Rift probably also played a role in preventing western rivers from entering the northern Gulf of Mexico (Fig. 12). However, the continental drainage divide was probably west of the rift, along the axis of volcanic centers in southwestern North America during the earliest early Miocene (LM1; Fig. 12; Galloway et al., 2011). As a result, volcaniclastic detritus from the mid-Cenozoic volcanic fields to the west of the Rio Grande Rift (e.g., San Juan volcanic field and Mogollon-Datil volcanic field) could still reach the Rio Grande Embayment in the Gulf of Mexico during this period, as indicated by the presence of abundant volcanic-rock-fragment–rich sediment stored in the Rio Grande Embayment (Galloway, 1981; Galloway et al., 1982). Rapid progradation of the Rio Grande Embayment shelf margin in LM1 indicates a continuity of sediment input to the Gulf of Mexico from the Oligocene to early Miocene. In contrast, retrogradation of the shelf margin in the late early Miocene (LM2) suggests a decline of sediment influx associated with increased sequestering of sediment by the Rio Grande Rift and disruption of the regional drainage basin by extensional tectonics (Fig. 1; Galloway, 2005). Other studies have suggested that sedimentation in the Rio Grande Rift began in the early Miocene, but that rapid basin subsidence and sediment accumulation in the rift clearly occurred in the middle–late Miocene (Chapin and Cather, 1994; Cather et al., 1994; Connell, 2004; Chapin, 2008). Therefore, it is probable that there was a connection between volcanic fields in New Mexico and southern Colorado and the Gulf of Mexico in the earliest early Miocene. As volcanic activity waned and Rio Grande Rift extension accelerated, transport of sediment from these sources to the Greater Rio Grande Embayment began to decline (Fig. 12).
The paleo–Missouri and Ohio Rivers were probably not fully integrated into a southward-flowing system until the Pliocene–Pleistocene (Galloway et al., 2011; Blum and Roberts, 2012). Instead, the paleo–Ohio River, which drained the Appalachian foreland basin, flowed northeastward in the early Miocene as a tributary to the preglacial St. Lawrence River (Hoagstrom et al., 2014, and reference therein). Similarly, the Miocene upper Missouri River flowed northward and joined the preglaciation “Bell River” in Canada (Howard, 1958; Sears, 2013). The size of the early Miocene drainage system feeding the Gulf of Mexico was therefore much smaller than the modern system, and this is reflected in the contrast between the sediment volumes reaching the Gulf of Mexico in the early Miocene and in the Pleistocene (Fig. 2).
Analysis of the early Miocene drainage systems is helpful for understanding how bulk sediment was transported from upland source terranes to basinal sinks. Previous paleodrainage analyses have been mainly constrained by petrographic data and logical deduction where the sedimentary record was limited (Galloway et al., 1982, 2011). Detrital zircon U-Pb analysis greatly enhances our understanding of paleodrainage systems supplying the northern Gulf of Mexico by allowing us to distinguish individual fluvial systems.
Early Miocene Drainage on the Great Plains
Our detrital zircon analysis strongly suggests a major sediment source from the Great Plains and southern Rocky Mountains. Large paleorivers carried the bulk of sediment to the Gulf of Mexico. However, the courses of these rivers across the Great Plains are difficult to define because their sedimentary records were mostly eroded during the late Cenozoic. Major erosion occurred before the deposition of the Middle to Upper Miocene Ogallala Group, cutting deeply into Permian, Triassic, Jurassic, and Lower Cretaceous rocks on the southern Great Plains, and less deeply into Upper Cretaceous–Cenozoic rocks on the northern Great Plains (Fig. 11). In contrast to poor preservation of Lower Miocene strata on the southern Great Plains, the aquifer made up of the Middle-Upper Miocene Ogallala Group is well preserved and studied. The contours of the base of this aquifer from Weeks and Gutentag (1981) guide reconstruction of the ancestral drainage pattern. The courses of the major early Miocene rivers can be interpreted based on the trend and configuration of these contours (Fig. 11).
Some Lower Miocene strata are preserved on the northern Great Plains where erosion was shallow. Paleostream analysis based on heavy mineral assemblages and cross-bedding in the Lower Miocene strata of the Arikaree Formation in western Nebraska and southeastern Wyoming suggests that northeastward-flowing streams carried materials eastward from the Laramie Mountains and Front Range (Bart, 1975). Condon (2005) also documented paleocurrents on the northern Great Plains oriented east-northeastward from a major source in the Front Range. This flow pattern might be related to regional structural tilting to the northeast induced by mantle buoyancy (Roy et al., 2004; Condon, 2005). The trend of basal Ogallala aquifer contours also displays a similar east-northeast trend (Fig. 11).
There are few records of the early Miocene paleodrainage trends on the southern Great Plains. Basal Ogallala aquifer contours suggest that the early Miocene rivers generally flowed east-southeastward (Fig. 11; Weeks and Gutentag, 1981; Fallin, 1988; Chapin, 2008). The sediment derived from the Great Plains was deposited in the Rio Grande, Houston, and Mississippi Embayment depocenters of the Gulf of Mexico (Galloway et al., 2011).
Paleo–Rio Bravo, Rio Grande, and Houston-Brazos Drainage Systems
While detrital zircon U-Pb age spectra of GOM1–7 display similar overall age patterns, indicating a shared common source terrane, these samples represent three different drainage systems influenced by local structures. These drainage systems can be differentiated by both sample petrographic features and detailed analysis of the detrital zircon age spectra. Samples GOM1–4 lie in the Rio Grande Embayment (Fig. 1). However, the U-Pb age pattern of GOM1 differs from those of samples GOM2–4, which are presumed to reflect paleo–Rio Grande delivery. Sample GOM1 has an age peak around 412 Ma that is reduced or absent in samples GOM2–4 (Fig. 3; Table 1). Sample GOM1 may have collected tributary contributions from the Marathon uplift by recycling strata from Pennsylvanian Haymond Formation, where zircon grains of Paleozoic age are well documented (Gleason et al., 2007). We interpret sample GOM1 to indicate a distinct fluvial system, paleo–Rio Bravo, rather than the paleo–Rio Grande (Fig. 12).
Although detrital zircon U-Pb signals of samples GOM2–7 are similar (Fig. 3), they are from two different fluvial systems, separated by local structures (e.g., the San Marcos Arch). The paleo–Rio Grande fluvial system (GOM2–4) south of the San Marcos Arch is thought to have been a coarse-grained bed-load river system, whereas intrabasinal fluvial systems (GOM5–7; named paleo–Houston-Brazos here) to the north of the arch were considered more mixed-load fluvial (Galloway et al., 1982). In addition, the paleo–Rio Grande was enriched in volcanic rock fragments that were delivered by an extrabasinal river from the southern Laramide uplifts and southern Great Plains, whereas the paleo–Houston-Brazos was more carbonate-rock-fragment–rich because of proximity to the Edwards Plateau in central Texas (Galloway et al., 1982, 2011; Dutton et al., 2012; Fig. 12).
Major fluvial rerouting occurred in the paleo–Houston-Brazos system to the northeast during this period. The paleo–Houston-Brazos was a large extrabasinal fluvial system in Eocene–Oligocene time, but the river discharge diminished during the early Miocene, and it became a minor sediment contributor to the Gulf of Mexico with no large deltaic depocenter formed in front of the fluvial axis (Spradlin, 1980; Galloway et al., 1986, 2011). This change may have been related to uplift of the Edwards Plateau in the early Miocene. This uplift deflected a paleoriver from the southern Great Plains, which had previously flowed southeast to join the Houston-Brazos, eastward to merge with the paleo–Red River instead (Galloway et al., 2011; Fig. 12).
Paleo–Red River and Mississippi Drainage System
The paleo–Red River had not yet been integrated with the Mississippi River in the early Miocene and was thus a separate major fluvial system delivering sediment to the deep-water Gulf of Mexico (Galloway et al., 2000, 2011). Two major deltaic depocenters were mapped by Galloway et al. (2000, 2011): one in eastern Texas (paleo–Red River) and the other in Louisiana (paleo–Mississippi River). The paleo–Red river and paleo–Mississippi River axes have similar mineral compositions: only mapping at the Gulf of Mexico margin differentiates the two. However, detrital zircon U-Pb age spectra from the two systems (GOM8–13 and GOM14) differ significantly and help to differentiate these fluvial systems (Fig. 3; Table 1).
Samples in the Houston Embayment have a reduced component of Cordillera Magmatic Province zircon grains and a high percentage of Yavapai-Mazatzal grains, indicating major southern Rocky Mountains and southern Great Plains sources (Fig. 12). Mineral assemblages (e.g., chert pebbles, garnet, and chromite) in Lower Miocene paleochannel deposits indicate that tributaries draining the southern Ouachita Mountains flowed southward to join the Red River in the early Miocene (McBride et al., 1988). In addition, Permian–Cretaceous vertebrate fossils were reworked from southwestern Oklahoma to paleochannel deposits in eastern Texas and western Louisiana by the paleo–Red River (Manning, 1990). The Red River may also have collected tributaries that earlier connected to the Houston-Brazos paleoriver in central Texas, as noted above.
In contrast, sample GOM14, near the Louisiana-Mississippi state boundary, strongly suggests transport by the paleo–Mississippi River (Fig. 12), because its U-Pb age data consist of large numbers of zircon grains from both the Appalachian terranes to the east (Grenville and Appalachian-Ouachita grains) and western U.S. sources (Western Cordillera arc, Yavapai-Mazatzal, and Midcontinent grains).
However, the early Miocene paleo–Mississippi River drainage area was much smaller than that of the Paleocene–Eocene, as well as that of the modern Mississippi River, because the paleo–upper Missouri and Ohio Rivers did not drain to the south until diverted by continental glaciation in the Pliocene–Pleistocene (Fig. 12; Galloway et al., 2011; Blum and Roberts, 2012). This is also evidenced by the lack of Archean-aged zircon grains from northern Wyoming or the Canadian Shield in Lower Miocene strata (Figs. 3 and 12; Table 1). In contrast, Archean-aged zircon grains are an important component of both the Paleocene–Eocene Wilcox Group (Fig. 10B; Blum and Pecha, 2014) and modern samples (Iizuka et al., 2005). Sample GOM15, from central-eastern Mississippi, is a fine-grained coastal and shelf sandstone and does not represent direct fluvial axis deposition. It displays a similar detrital zircon U-Pb age pattern as sample GOM14, suggesting eastward longshore current reworking of the paleo–Mississippi River deposits to the west.
Paleo–Tombigbee River and Apalachicola Drainage System
Lower Miocene strata from the eastern Gulf of Mexico margin are mainly thin marine carbonates and barrier island and beach sand deposits (Galloway et al., 2000; Galloway 2008). The river that preceded the middle Miocene Tennessee River was relatively small, and sediment was reworked into coastal and shelf sand bodies. The onshore fluvial deposits identified here are local intrabasinal fluvial systems, which provided limited sediment influx to the Gulf of Mexico.
We interpret sample GOM16 as representative of the proximal, intrabasinal paleo–Tombigbee River. Sample GOM16 contains high proportions of Grenville- and Appalachian-Ouachita–aged zircon grains with no grains younger than 280 Ma. Zircon grains derived from western U.S. sources were collected by the paleo–Mississippi River to the west and could not pass eastward of this barrier. Most of the paleo–Tombigbee River sediment was recycled from the Paleocene–Eocene Wilcox Group located to the north (updip) or Paleozoic strata from the Appalachian foreland basin (Figs. 8, 10C, and 10D). Overall, the sediment influx rate in this region was low, and only in the middle Miocene did the emerging paleo–Tennessee River begin to carry large volumes of sediment to the Gulf of Mexico continental shelf and basin (Galloway et al., 2000, 2011). Boettcher and Milliken (1994) suggested a change of exhumation rate in the Southern Appalachians from only 0.4 ± 0.2 km eroded within 60 m.y. before 20 Ma, to 1.5 ± 0.5 km eroded from 20 Ma to present. The increase in sediment erosion rate was postulated to be principally related to epeirogenic uplift of the Appalachians induced by dynamic mantle activity (Gallen et al., 2013; Miller et al., 2013; Liu, 2014).
We also interpret a paleo–Apalachicola River based on distinct detrital zircon U-Pb age spectra of samples GOM17–19 and their locations near the modern Apalachicola Embayment. The zircon components of the Eastern Gulf of Mexico Embayment provenance are dominated by the Grenville and Appalachian orogeny terranes, suggesting that the paleo–Apalachicola River drained the eastern Blue Ridge and Inner Piedmont of the Southern Appalachian terranes that were similar to those seen today. At the GOM19 outcrop, the fossiliferous quartz sand (ca. 18.3 Ma; Bryant et al., 1992) directly overlies Oligocene Suwanee Limestone, suggesting a pronounced early Miocene transition from a Paleogene carbonate environment to a Neogene siliciclastic environment.
SUMMARY AND CONCLUSIONS
Nineteen Lower Miocene samples were collected from the northern Gulf of Mexico margin, and 2192 reliable detrital zircon U-Pb ages were obtained. The detrital zircon data display large variability in age distributions, ranging from late Oligocene (ca. 24 Ma) to Archean (ca. 3600 Ma). Major zircon components include Cordillera Magmatic Province (24–280 Ma), Appalachian-Ouachita Province (300–500 Ma), Grenville Province (950–1300 Ma), Midcontinent Province (1300–1500 Ma), and Yavapai-Mazatzal Province (1600–1800 Ma). Only a few grains appear to come from the Pan-Africa (500–700 Ma) and Shield (>1800 Ma) Provinces.
Detrital zircon age data from the 19 samples generally cluster into four regional components: Greater Rio Grande Embayment, Houston Embayment, Mississippi Embayment, and Eastern Gulf of Mexico Embayment. Each of these groups represents a distinct provenance. The Greater Rio Grande Embayment strata were mainly derived from the southern Great Plains and Trans-Pecos volcanic field. Local source terranes include the Llano uplift and Edwards Plateau in central Texas, both of which were uplifted and exhumed during the early Miocene. The Houston Embayment strata were likely derived from the southern Rocky Mountains and the southern Great Plains, where deep erosion occurred during the late Oligocene–early Miocene. The Mississippi Embayment on the central Gulf of Mexico coast contains sediment sourced from both western (mainly Great Plains and southern-central Rocky Mountains) and eastern U.S. terranes (Appalachian foreland basin and Paleozoic Appalachian orogenic terranes). Locally, the Ouachita Mountains and Ouachita foreland basin could have been sediment contributors. The Eastern Gulf of Mexico Embayment suggests a locally proximal source from the eastern Blue Ridge and Inner Piedmont of the Southern Appalachian terranes. Therefore, the detrital zircon age spectra document a continental-scale sediment provenance shift across the northern Gulf of Mexico from highlands in the western United States to the Appalachian Mountains and foreland basin in the eastern United States.
The sediment associated with the paleo–Rio Bravo, Rio Grande, Brazos, Red, Mississippi, Tombigbee, and Apalachicola Rivers can be differentiated based on detrital zircon analysis, improving definition of sediment transport pathways from upland sources to basin sinks (Fig. 12). Future detrital zircon U-Pb dating of deep-water samples could provide a complete sediment routing pathway from hinterland source terranes to basin terminal sinks. Complementary onshore and deep-water detrital zircon analyses that reveal sediment pathways from fluvial sources to basinal sinks would also aid in prediction of petroleum reservoir quality in the deep basin.
Comparison with published detrital zircon provenance studies yields improved understanding of drainage system evolution through time. The absence of zircon grains with ages of 110–140 Ma in the Lower Miocene strata indicates a cutoff of drainage from western Mexico terranes at this time. A smaller component of grains older than 1800 Ma in the Lower Miocene, compared to the Paleogene Wilcox strata, probably indicates a reduced contribution from the central-northern Rocky Mountains, if the contrast is not biased by the limited detrital zircon age data (∼100 grains in each sample). Finally, a reduced Lower Miocene drainage area compared to that of the Paleogene Wilcox Group can be inferred (Fig. 12). This may partially explain the decreased volume of sediment in the Lower Miocene Gulf of Mexico strata relative to the Wilcox Group (Galloway et al., 2011; Fig. 2). This decrease in sediment volume may also have been related to arid conditions in the western United States, which reduced water discharge and thus the capacity of rivers to deliver sediment to the Gulf of Mexico.
A comparison with published detrital zircon provenance studies also suggests that recycling is common in Lower Miocene strata. The prominent Cordilleran Magmatic Province component is probably a product of erosion of earlier Mesozoic–Cenozoic strata rather than sourced directly from arc terranes to the west of the Rio Grande Rift. Other important components, like Appalachian-Ouachita, Grenville, Midcontinent, and Yavapai-Mazatzal grains, might have been partially recycled from the Paleocene–Eocene Wilcox Group strata. These recycling events may have been highly responsive to tectonic and paleodrainage evolution, e.g., thermal uplift and development of the Rio Grande Rift in the late Oligocene–early Miocene in the southwestern United States. Future application of combined detrital zircon U-Pb and (U-Th)/He analyses on single zircon crystals would further refine sediment provenance and help to resolve remaining uncertainties surrounding sediment sources and routing.
We would like to thank the members of the Gulf Basin Depositional Synthesis Project (GBDS) industrial associate program and GBDS Project Manager Patricia Ganey-Curry. This research was also funded by an internal Jackson School of Geosciences Energy Theme seed grant to Stockli and Snedden. We are grateful for an Ed Picou Fellowship from the Gulf Coast Section of the Society for Sedimentary Geology, which helped fund the field work. We also thank the Institute for Geophysics, University of Texas at Austin, for providing an Ewing-Worzel Graduate Fellowship. Field sample collection was facilitated by Gary Kinsland and Robert Hatcher. The Florida Geological Survey and Bureau of Economic Geology allowed us to sample subsurface cores. We also thank Lisa Stockli and Spencer Seman for assistance with U-Pb dating, and Daniel Arnost, Timothy Shin, and Shanping Liu for their help with mineral separation. Constructive comments by reviewers Mike Blum and Andrew Leier and Editors Timothy Lawton and Aaron Cavosie are appreciated and helped to greatly improve this manuscript.
↵1GSA Data Repository item 2016246, including detailed description of the methods used for detrital zircon U-Pb age analyses, Lower Miocene sample location data, and zircon U-Pb data, is available at http://www.geosociety.org/pubs/ft2016.htm or by request to .
Science Editor: Aaron J. Cavosi
Associate Editor: Timothy Lawton
- Received 29 November 2016.
- Revision received 27 June 2016.
- Accepted 31 July 2016.
- © 2016 Geological Society of America