THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT Tracking Barrier Island Response to Early Holocene Sea-level Rise: High Resolution Study of Estuarine Sediments in the Trinity River Paleovalley

Understanding how barrier islands respond to factors such as variations in sediment supply, relative sea-level rise, and accommodation is valuable for preparing coastal communities for future impacts of climate change. Increasingly, the underlying antecedent topography has been observed to have a significant control on the evolution of the barrier island system by providing increased elevation, decreased accommodation, and sediment supply for the barrier to rework and anchor upon. However, less attention has been focused on how back barrier sediments respond to this decreased accommodation, and how this may affect barrier island evolution. Additionally, the control in which the geometry of the underlying valley itself has on the initiation of barrier islands is poorly understood. Here we examine the stratigraphic THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT framework of the Trinity River incised valley, offshore Galveston, Texas in order to investigate the role of antecedent topography in the evolution of an ancient barrier island system. We present high-resolution imaging of the Trinity incised valley fill using over 1200 km of 3D seismic, <700 km of 2D envelope and full waveform chirp data, along with 2 piston cores, 4 gravity cores, 1 platform boring, with associated grain size, foraminiferal, and radiocarbon data. We find that the geometry and elevation of the underlying antecedent topography plays a central role in the evolution of the barrier island system, promoting both initiation and stabilization. This study provides a methodology to investigate the evolution of a relict barrier island system where little to none of the barrier is preserved. With this methodology, we revise the established Holocene paleoshoreline model for the Trinity incised valley.


Introduction
Estuaries are uniquely dynamic depositional systems controlled by the interplay of river, wave, and tidal forces. They develop within incised valleys, which are formed through fluvial incision during periods of lower relative sea-level Nordfjord et al., 2006).

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
Preserved estuarine sediments within incised valley fills enable comprehensive studies investigating estuarine response to relative sea-level rise (Thomas and Anderson, 1994;Storms et al., 2008;Simms et al., 2008;Rodriguez et al., 2004Rodriguez et al., , 2005Anderson et al., 2004Anderson et al., , 2014Anderson et al., , 2016Reijenstein et al., 2011;Ronchi et al., 2018Ronchi et al., , 2019Shawler et al., 2020). In particular, the Holocene estuarine record within the Gulf of Mexico incised valley fills have been extensively studied due to excellent preservation and well-constrained rates of sea-level rise during this time period (Milliken et al., 2008;Anderson et al., 2016). These studies have led to the commonlyapplied stratigraphic model of estuarine and barrier island response to episodic changes in relative sea-level rise (Thomas and Anderson, 1994;Anderson et al., 2008;Anderson et al., 2010), changes in climate, which regulates sediment supply (Anderson et al., 2008), and flooding of antecedent fluvial topography (Rodriguez et al., , 2005Anderson et al., 2008) in the form of flooding surfaces. In response to increased relative sea-level rise, decreased sediment supply, or increased surface area of fluvial terraces, estuarine systems will experience rapid landward back-stepping, wherein fluvial and bayhead delta sequences are overlain by bay, tidal, barrier, and marine deposits. This results in an incised valley fill characterized by discontinuous, landward-stepping, deepening-upward successions (Thomas and Anderson, 1994;Anderson et al., 2016).
Antecedent topography is another factor that is increasingly understood to play a pivotal role in controlling the evolution of estuarine and barrier island systems in a multitude of ways.
Previous authors found that flooding of antecedent fluvial terraces within the incised valley results in increased surface area for rapid bay expansion (Rodriguez at al., 2005;

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
2008; Simms et al., 2008), while also serving as local sand sources to supply the transgressing barrier . Through a comprehensive field and morphodynamic modelling study, Shawler et al. (2020) found that antecedent topographic highs play a central role in how barrier islands respond to sea-level rise. They found that steep antecedent slopes and decreased back-barrier accommodation provides both increased elevation and erodible sediment supply to essentially anchor the island despite high rates of sea-level rise, and thus promotes barrier resilience. Conversely, increased back-barrier accommodation may lead to rapid landward migration.
To investigate the role of antecedent topography in the evolution of an ancient barrier island system, we focus on the Trinity incised valley offshore modern Galveston Bay in the central Gulf of Mexico (Figure 1; Anderson et al., 2016). The regional estuarine stratigraphic framework has been previously studied, identifying fluvial and deltaic units overlain by estuarine units including lower bay and tidal-inlet sands (Thomas and Anderson, 1994;Anderson et al., 2004;Rodriguez at al., , 2005Swartz, 2019). Rates of RSLR in the northern Gulf of Mexico over the last 10 kyr are well-constrained, displaying a decrease from 5 mm yr -1 to 3 mm yr -1 (Milliken et al., 2008). However, previous authors have focused on identifying the adjustment and evolution of the fluvial-deltaic system during RSLR (Swartz, 2019), or have lacked the geophysical resolution to accurately identify fine-scale stratigraphic architecture necessary to capture the complex nature of the estuarine section (Thomas and Anderson, 1994;Anderson et al., 2004;Rodriguez at al., , 2005. In this study, we utilize a dense (~250 m line spacing) grid of high-resolution chirp data, in combination with micropaleontological, sedimentological, and radiocarbon dating methods to investigate the role of the antecedent geology on the initiation and evolution of a paleo-barrier

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
island system. This study provides a methodology to study the evolution of a paleo-barrier system where little of the barrier itself is preserved in the stratigraphy. Using this methodology, we revise the established Holocene paleoshoreline model for the Trinity incised valley.  Thomas and Anderson, 1994). Brown tracklines represent available geophysical data for the area. Yellow lines represent chirp geophysical data used in this study. Grey coherency extraction represent spatial extent of publicly available, former industry collected 3D seismic data used in Swartz, 2019 and this study.
Red circles represent gravity and piston cores used in Standring et al., 2019 and this study. Green circle represents Boring 4 from Thomas, 1991.

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
The northern Gulf of Mexico experiences highly variable regional basin subsidence, exhibiting relatively low rates in coastal regions (0.03 mm/yr), yet high rates along the shelf edge (> 1.0 mm/yr) (Anderson et al., 2016). The central Texas region is currently classified as a dry subhumid region (Thornthwaite, 1948). However, multiple studies focused on the post-glacial climatic history of central Texas observe marked changes from cool and wet ~18 -7.5 ka, to warm and dry ~7.5 to 3.5 ka, and finally variable, millennial scale shifts between cool-wet and warm-dry conditions since 3.5 ka (Humphrey and Ferring, 1994;Nordt et al., 1994Nordt et al., , 2002Toomey et al., 1993). The oceanographic circulation along the inner shelf is dominated by a strong westward current called the Louisiana Coastal Current (Cochane and Kelly, 1986;Oey 1995;Jarosz and Murray, 2005).
The Trinity incised valley formed during Marine Isotope Stages (MIS) 5-3, approximately 119 to 22 ka. During this time, the region experienced ~120 m in episodic sealevel fall, prompting a stepped downcutting of the fluvial system into the continental shelf. This resulted in an incised valley with a terraced morphology extending from the modern Galveston Bay to the shelf margin ( Fig. 1; Thomas and Anderson, 1994;Anderson et al., 2016). The incised valley reached its most basinward location during MIS 2 approximately 22 to 17 ka, depositing a large shelf edge delta and basin floor fan complex (Pirmez et al., 2012;Anderson et al., 2016). Thomas and Anderson (1994) measured the valley surface to have 30-40 m of total relief; however, upper portions of the valley have since been infilled or removed by the transgressive ravinement during subsequent sea-level rise, and only lower portions of the valley exist offshore overlain by the seafloor.
Following the MIS2 lowstand, rapid sea-level rise (~4.2 mm/yr) between ~17 -10 ka prompted a shift from progradation to aggradation, retrogradation, and valley filling for much of

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
the fluvial systems across the Gulf of Mexico (Milliken et al., 2008;Anderson et al., 2016). By 10 ka, the rate of sea level rise slowed progressively from 4.2 mm/yr to 1.4 mm/yr, leading to slower rates of transgression, better preservation of thick transgressive deposits, and, as a result, a more complete sedimentological record (Milliken et al., 2008;Anderson et al., 2014;. After 10 ka, the Trinity valley was infilled by a series of discontinuous, landward stepping depositional packages consisting of fluvial, deltaic, bay and tidal packages (Thomas and Anderson, 1994;Rodriguez et al., 2004;Anderson et al., 2016).
Previous authors have interpreted large shelf sand bodies on the inner shelf as remnants of coastal barriers formed during sea-level stillstands, and subsequently overstepped and reworked during transgression (Rodriguez et al., 1999;Anderson et al., 2016). Rodriguez et al. (2004) identified estuarine muds overlain by shoreface and/or tidally influenced sands within cores beneath Heald Bank and Sabine Bank (Fig. 2). Ages acquired from the estuarine units at both locations suggest that estuaries existed at Heald Bank and Sabine Bank from 8.4 to 7.7 ka and 7.4 to 4.7 ka, respectively. Dates taken from estuarine packages in West and East Bay ( Fig. 2) suggest that a lagoon existed in those areas since at least 7.5 and 7.7 ka, respectively. Thus, the shoreface and/or tidally influenced unit below Heald Bank was interpreted as the 7.7 ka barrier island location, fronting a 50-km long estuary extending from Heald Bank to West and East Bays . These authors suggested that the western portion of the barrier shoreline rapidly retreated during RSLR until Galveston Island began prograding approximately 5.3 ka.
Conversely, according to Rodriguez et al. (2004), the eastern portion of the barrier shoreline remained relatively stable, reaching Sabine Bank around 5.3 ka, leading to an irregular, perpendicular shoreline extending from Sabine Bank to East Bay (Fig. 2). The shoreline

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
gradually retreated to the southwestern Louisiana chenier plain by 2.8 ka, and again to Bolivar Peninsula by 1.5 ka as it began to form by spit creation, washover, and ephemeral flood-tidal delta deposition (Fig. 2;Gould and McFarlan 1959). The stabilization, and subsequent rapid retreat of the paleoshoreline has been attributed to the presence of fluvial deposits directly beneath Heald and Sabine Banks within the Trinity-Sabine incised valley . As the shoreline retreated, the fluvial deposits provided both an anchor point and a local sediment supply to remain relatively stable; however, once these sources were depleted and/or a sediment supply-to-sea-level threshold was surpassed, rapid shoreline retreat ensued.
Alternatively, Thomas and Anderson, 1994 earlier proposed that these sand banks do not have origins as submerged, overstepped barriers, but rather as modern marine sand bodies. They observe Sabine and Heald Banks to overlie a ravinement eroding underlying estuarine and tidal deposits from relict barrier island systems. Therefore, it is possible that these sand bodies are marine sand bodies reworking relict barrier island material, and established at shoreline positions that pre-date the formation of the bank.  Thomas and Anderson, 1994). Red lines represent paleoshorelines modified from Rodriguez et al., 2004.

Chirp and 3D Seismic Data
This study utilizes high-resolution chirp sub-bottom reflection data and archived 3D seismic datasets. Over ~ 1000 km of 2D chirp data have been collected by the University of  Swartz (2019) retrieved this dataset and conducted further processing steps in Petrel 2016. A crossline filter of 90 m was applied to the data volume in order to smooth striping artifacts resulting from the survey's acquisition footprint. A shallow, regionally conformable horizon located between 40-60 ms interpreted as the base of the Trinity incised valley was mapped and agrees with previous interpretations (Thomas and Anderson, 1994). A 20 ms envelope median coherency attribute was calculated and extracted from this horizon. The result of this median coherency extraction is an amalgamated, time-transgressive representation of the Trinity incised valley fill, highlighting the most anomalous amplitudes within the stratigraphic section.

Core Data
This study utilizes 2 piston cores, 4 gravity cores, and 1 platform boring to ground-truth seismic datasets. The platform boring was originally collected for foundation studies for drilling rigs and production platforms by Fugro-McClelland Engineers in Houston, which was subsequently reexamined by for lithological and environmental interpretation by Thomas (1991). Course. Core sites were selected based on shallow seismic stratigraphic structures observed within seismic data to characterize key transitions within the incised valley fill, evaluate paleoenvironmetal settings, date the transitions from key paleoenvironments, and interpret depositional processes. Thorough detail on the acquisition, processing, and analysis of core data is provided by Standring et al. (2019).  Thomas, 1991.

Results and Interpretation
Interpretation of chirp data is divided into 5 seismic facies analyzed from full-waveform data (Fig. 4), and 8 seismic horizons that bound 7 seismic units  analyzed on both envelope and full-waveform data. Seismic horizons and units are correlated to stratigraphic (Fig.   11), foraminiferal, and radiocarbon data within cores  in order to interpret the depositional environment and age of each unit.

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
Eight seismic horizons have been mapped bounding characteristic components of the Trinity incised valley fill . These surfaces have been defined based on stratigraphic superposition, termination style, and when necessary, associated seismic facies both above and below a given horizon. Time-structure maps were created using a combination of Python and GMT 6 programming scripts, in which horizons were gridded and interpolated onto 50-meter cells (Fig. 21). When appropriate, horizons were converted from two-way travel time in milliseconds to depth in meters below sea level (mbsl) using an average velocity of 1525 m/s (Abdulah et al., 2004).

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
H1 is the modern seafloor, which displays a general southwestward dip ( Fig. 21 -H1).
H1 is shallowest within the northeastern portion of the study area at 19 ms depth (~14.5 mbsl) and deepest within the southwestward portion at 25 ms depth (~19 mbsl). This horizon is often erosional in nature, and frequently amalgamates with the underlying H2 horizon 12,14,16,17).

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
Based on foraminiferal analyses (Standring et al., 2019), H3 is interpreted to be the base of (Figs. 14 -18) or lie within (Fig. 13) outer bay sediments, a high-energy environment just landward of ancient barriers formed by tidal, storm, and wave processes (Dalrymple et al., 1992;Zaitlin et al., 1994).

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
area, but frequently truncated by H4, H4a, and H4b within the seaward portion of the study area 15,19,20), resulting in holes within the H5 horizon time-structure map ( Fig. 21 -H5). This horizon is deepest within the seaward portion of the study area, reaching its maximum depth of 37 ms (~28 mbsl) within a 2 km wide channelized feature ( Fig. 21 -H5). H5 then rapidly shallows toward the valley margins and the landward portion of the study area ( Fig. 21 -H5). H5 has been previously interpreted as the boundary between underlying deltaic or upperbay sediments and overlying bay sediments (Thomas, 1991;Swartz, 2019).

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
U1 is has an upper boundary of H1, a lower boundary of H2, and is composed of SF1 . This unit ranges in thickness from 0 to 2.8 ms (~0 to 2 m) ( Fig. 22 -U1), and is unique in that it is the only unit in this study to have been deposited beyond the valley walls. Sediments cored from this unit vary from uncompacted, "soupy" sediments (Figs. 14,(16)(17)(18), to coarsergrained sediments (Figs. 13, 15). U1 is generally thinnest along the western portion of the study area, where its lower bounding horizon U2 amalgamates near or directly with the seafloor, and corresponds to disturbed core samples due to uncompacted sediment (Figs. 17, 18). Where U1 is better preserved in flat-lying regions, foraminiferal data, which display an overall increase in diversity (Standring et al., 2019), reveal U1 is likely derived from an inner shelf environment (Figs. 14,15).
This unit reaches its greatest thickness along the eastern portion of the study area where it infills deep channel-like features formed by the horizon H2 (Figs. 13, 21, 22). Sediment cored from the margin of the largest of such channel-like features shows U1 to be composed of an ~0.8 m package of sandy sediments (Fig. 13). This coarse-grained package overlies horizon H2, which is observed to truncate the underlying H3 horizon (Fig. 13). Foraminiferal analysis of this isolated package reveal U1 is likely composed of one or multiple washover deposits overlying an erosive transgressive lag deposit ( Fig. 13; Standring et al., 2019).
Dating this unit is difficult due to the removal of material above the transgressive ravinement, apart from a single radiocarbon date taken from a mollusk shell dated to ~4.3 ka (Figure 13; Standring et al., 2019). However, age models performed on environmental transitions indicate that the paleo-estuary likely transitioned to an inner shelf environment by ~6.9 ka (Standring et al., 2019). Therefore, U1 is interpreted as a time-transgressive, patchy veneer of sediment composed of either uncompacted fine-grained material derived from an inner shelf environment,

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
or ~2m thick coarse-grained package composed of washover deposits and transgressive lags derived from a relict barrier, forming at least by ~6.9 ka, and persisting through ~4.3 ka. These unconsolidated muds, sands, and shells are interpreted to be associated with the transgressive ravinement, reworking underlying material over a span of ~2.6 kyr (Nummendal and Swift, 1987;Swift and Thorn, 1991;Cattaneo and Steel, 2003).
These observations, along with high percentages of the benthic foraminiferal assemblage Elphidium (Standring et al., 2019), suggest Unit 2 formed in an outer bay environment-a highenergy environment just landward of the paleo-barriers subject to frequent variations in energy

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
source, ranging from tidal, storm, and wave processes (Dalrymple et al., 1992;Zaitlin et al., 1994). Radiocarbon dates within this unit suggest it began forming no later than ~8.5 ka, lasted until ~6.9 ka (Standring et al., 2019).
This unit ranges in thickness from ~0 to 16 ms (~0 to 12 m) ( Fig. 22 -U3), and is thickest within the seaward portion of the study area where it infills an ~11 km long, 3.5 to 5 km wide channellike feature along horizon H4/H4a that rapidly shallows landward and to the west (Figs. 5,9,19).
The fill of this large channel-like feature is composed of SF4, but rapidly changes to SF3 in the landward portion of the study area (Figs. 12, 17).
Based on this unit's stratigraphic position, the highly erosional and channelized nature of its basal horizons H4, H4a, and H4b, this unit is interpreted as tidal fill encompassing both tidal inlet and tidal delta sediments. This is a high tidal energy environment proximal to the barriers (Dalrymple et al., 1992;Zaitlin et al., 1994). Unit 3 is divided into two sub-units U3a and U3b based on their associated with SF4 and SF5, respectively.
This unit is composed of SF4 within the seaward portion of the study area (Figs. 7,20), and SF3 within the landward portions of the study area (Figs. 12, 17). This unit varies in thickness from ~0 to 16 ms (~0 to 12 m) (Fig. 22 -U3a). U3a is thickest toward the seaward portion of the study area within an 11 km long, 1 to 2 km wide, 16 ms (~12 m) deep channel along horizon H4a that curves toward the northwest, interpreted as the tidal inlet ( Fig. 22 -U3a). Although this unit is present within an 11 km long channelized feature, the highly erosive nature of its basal horizon H4a is most evident within the first 5 km of the seaward portion of the study area ( Fig. 21 -H4a). Within this erosive channel-form, U3a is composed of SF4, characterized by internal reflections that onlap adjacent erosional surfaces within the unit, and are generally arranged in a draping pattern that reflects its deeply incised, channelized lower boundary. This cut-and-fill reflector configuration is indicative of multiple aggradational packages separated by periods of erosive channel formation, which is commonly observed within paleo-tidal inlets (Siringan and Anderson, 1993;Anderson et al., 2008;Storms et al., 2008;Ronchi et al., 2018Ronchi et al., , 2019. U3a transitions to SF3 toward the landward portion of the study area, consisting of lowamplitude, parallel, laminated to transparent reflections (Figs. 12,17). Coring through the landward portion of this unit reveals it is mostly composed of mud with some sandy intervals containing foraminiferal assemblages indicative of a central bay environment (Fig. 17). A radiocarbon date found within PC-2 returns a date of ~8.5 ka for this unit ( Fig. 17; Standring et al., 2019). The landward transition in seismic facies from draping, onlapping reflections (SF4) to parallel, laminated reflections (SF3) filling the tidal inlet has been interpreted as the stratigraphic
Seismic characteristics of this nature have been commonly associated with tidal deltas (Siringan and Anderson, 1993;Rodriguez et al., 1998Rodriguez et al., , 2005Anderson et al., 2008). SF1 is also common within the base of these channels, observed as chaotic to discontinuous reflections with internal truncations (Figs. 10, 11, 15). Commonly, SF5 lies adjacent to SF1 (Figs. 11, 15), where unidirectional dipping reflections (SF5) transition into chaotic to discontinuous reflections (SF1) within a short distance. It is possible that stratigraphically these facies both represent dipping, laterally accreting strata composing the tidal delta unit, but is at times not well-imaged either due to large grain sizes or unfavorable dip directions, resulting in SF1.
Coring of this unit by GC-2 further lends support for the interpretation that U3b is a tidal delta ( Fig. 15), in which its unidirectional dipping, laterally accreting reflections are observed to be composed of muds with numerous layers of silty sand. Carbon dates taken from the transition

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
from tidal delta to outer bay sediments in GC-2 (Fig. 15) provided ages of ~8.5 ka, confirming this tidal delta was active during this time (Standring et al., 2019).
SF2 is also common within this unit: a facies composed of high-amplitude, U-shaped, aggrading reflections commonly associated with an outer bay environment proximal to the barriers (Figs. 8,   10, 13, 15). This marked change from laterally accreting to aggrading reflections is likely due to a change in tidal regime associated with the infilling of the tidal inlet by ~8.5 ka, as observed in PC-2 (Fig. 17). Thus, as the tidal inlet unit U3a infilled by ~8.5 ka, the adjacent tidal delta unit U3b reflected this change in tidal regime by switching from a laterally accreting (SF1/SF5) to an aggrading (SF2) seismic facies.
Coring of this unit indicates U4 is composed of mostly fine-grained mud, with some interbedded muddy sand intervals (Figs. 14, 16 -18). This fine-grained composition, along with foraminiferal evidence of a spike in Ammonia (Strandring et al., in prep), support the interpretation that these sediments were deposited in a low-energy, quiescent central bay environment landward of the barriers (e.g., Dalrymple et al., 1992;Zaitlin et al., 1994;Rodriguez et al., 2005;Anderson et al., 2008). Carbon dates from this unit indicate that this environment

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
initiated as early as ~9.6 ka (Fig. 17) and lasted until ~8.0 ka, when it transitioned to an outer bay environment (Fig. 17, 18; Standring et al., 2019).
Although U4 is generally thickest in the western portion of the study area, U4 reaches it maximum thickness in the form of three mound-like features measuring ~500 m in width, ~8 m in height (Figs. 5,7,9,19,20,22). Goff et al. (2016) have interpreted similar mound-like features in Corpus Christi Bay, Texas as oyster reefs, described as having a high-amplitude, draping reflection as its upper boundary, a chaotic to acoustically blank interior, a varying amplitude, planar reflection at its base, and adjacent onlapping reflections. A similar configuration is observed within the mound-like features in this study, having a high-amplitude upper boundary defined by H4, H4a, and H4b, a chaotic to acoustically blank interior associated with SF3, a varying amplitude lower boundary defined by H5, and adjacent onlapping reflections by U3a and U3b (Fig. 7, 20). Unfortunately, these mound-like features were not groundtruthed, and their depositional environment cannot be confirmed without further investigation.

Unit 5 -Bayhead Delta (U5)
U4 has an upper boundary of H2, H4, and H5 and a lower boundary of H6 , and varies in thickness between ~0 to 23 ms (~0 to 17.5 m) (Fig. 22 -U5). This is the thickest unit observed in this study, composing over half of the total incised valley fill. This unit is thinnest along the flanks of the valley (Fig. 5, 6, 8, 16, 20), above or adjacent to underlying Pleistocene fluvial terraces (Fig. 9, 17 -19), or when it is heavily reworked by tidal-related horizons units H4, H4a, and H4b within the seaward portion of the study area (Figs. 7,8,11,15,19,20). U5 is thickest when it is filling in accommodation provided by the underlying fluvial geology (Figs. 5,9,12,19). This is clearly observed as the topography of its lower bounding
This unit is primarily composed of high-amplitude, laterally accreting to aggrading reflections (Figs. 5 -18;Swartz, 2019;Thomas, 1991). Coring of this unit reveals it is composed of very fine to fine sands, organic rich muds and clays, and clays interbedded with find sandy layers with organic wood fragments and plant debris (Fig. 12, 17, 18; Thomas, 1991). Lack of foraminifera in these sediments suggest a terrestrial depositional environment. Therefore, sediments in U5 are interpreted as deriving from an upper bay or bayhead delta environment (Swartz, 2019).
Age models from cores PC-2 (Fig. 17) and PC-4 ( Fig. 18) indicate that the environment transitioned from fluvial to upper bay ~9.8 ka (Standring et al., 2019). This upper bay environment lasted until at least ~8.8 ka as the system was inundated and transitioned into a central bay environment (Fig. 18). The location for this chirp line can be found in Figure 3.

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
As rising sea levels act upon a barrier island system, features of the coastal lithosome are generally eroded by wave, storms, and tidal forces associated with the transgressive ravinement (e.g., Swift and Thorn, 1991;Cattaneo and Steel, 2003). Commonly referred to as a "wave ravinement surface" (e.g., Allen and Posamentier, 1993;Cattaneo and Steel, 2003) to highlight the erosive mechanism, the creation of this surface has also been attributed to downwelling coastal storm currents (Swift and Thorn, 1991) and major storms (Goff et al., 2015). Generally, this strong erosive action removes estuarine sediments, and transfers eroded sediment seaward from the barrier island system to the marine sediment sheet overlying the ravinement. Due to the erosive nature of the transgressive ravinement, subaerial and shallow features of the barrier island system, including the barriers themselves, are often removed, preserving only deep backbarrier sediments, such as bay and tidal facies, within the stratigraphic record (Ronchi et al., 2018;Ronchi et al., 2019).
There have been documented studies within the Adriatic Sea (Storms et al., 2008) and Tijucas Bay, Brazil (Cooper et al., 2016), investigating the preservation of the barriers throughout transgression through a process called "overstepping", in which the barrier is stranded on the continental shelf as a marine sand body (Carter, 1988). The overstepping and preservation of the barrier during transgression has been attributed to rapid sea-level rise, rapid increase in back-barrier accommodation, little wave action, or rapid burial by shoreface sediments (Storms et al., 2008;Cooper et al., 2016). This is one interpretation of paleoshoreline evolution across the northern Gulf of Mexico, wherein Sabine and Heald Banks are interpreted as relict, overstepped barriers preserved on the inner shelf as sand bodies ( Fig. 2; Rodriguez et al., 1999;Rodriguez et al., 2004). Sabine bank is located approximately 30 km offshore in water depths of 5 to 12 m, while Heald banks is

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
approximately 45 km offshore in 9 to 15 m water depth. Rodriguez et al., 2004 claim that these shorelines retreated over terraced fluvial deposits within the Trinity-Sabine incised valley, which served as local sand sources for barrier island stabilization and persistence throughout sea-level rise. These barriers were then overstepped and stranded on the inner shelf as isolated sand bodies once these fluvial sand sources were depleted, sea-level rise reached a critical threshold, or both .
Conversely, it has been proposed that the origin of modern sand banks is not as drowned, overstepped, detached barriers (i.e., Snedden et al., 1999), but rather as modern marine deposits that have reworked underlying remnant estuarine and tidal deposits (Thomas and Anderson, 1994;Dyer and Huntley, 1999). Thomas and Anderson (1994) observe that Sabine and Heald Banks overlie a ravinement reworking underlying estuarine and tidal delta facies. Thus, it is possible that these sand banks have a marine origin, which have formed from the reworking of remnant estuarine and tidal facies, rather than as downed, overstepped barriers.
In this study, erosion associated with the transgressive ravinement, interpreted as the H2 horizon, has significantly reworked underlying estuarine sediments , and completely eroded and removed the barrier within our study area. Observations of interpreted overstepped barriers include elongated, shore-parallel sand bodies with reflectors that prograde both seaward and landward, with a coarsening-upward stratigraphic composition (Storms et al., 2008;Cooper et al., 2016). These seismic stratigraphic and lithologic indicators are not observed within this study, supporting the interpretation that the paleo-barrier within this study area was not overstepped, but instead completely removed by transgressive erosion.
Despite significant transgressive erosion, considerable preservation of deltaic (U5) and back-barrier (U2 -U4) strata is observed in this study . The transgressive

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
ravinement (H2) is observed to rework shallow back-barrier units (U2, U3, U3a, U3b), while only reworking the distal flanks of the deeper central bay and bayhead delta units along the valley edges (U4, U5). Although preservation of the barrier can give excellent insights into the mode in which the estuarine system transgressed (i.e., erosional, rollover, overstepping -Carter, 1988), preserved deltaic, bay, and tidal facies can provide convincing spatial and temporal evidence for understanding the morphological evolution of the estuarine system throughout transgression (i.e., Ronchi et al., 2018;2019).

Estuarine Interaction with Underlying Geology
Heterogeneity in the underlying, antecedent geology in which the barrier island system transgresses has long been recognized as a major control on its evolution (e.g., Belknap and Kraft, 1985;Riggs et al., 1995, Mallinson et al., 2018Shawler et al., 2020). These are often paleodrainage systems that cut into the landscape by fluvial incision during subaerial exposure, and now filled with fluvial, coastal, and shelf facies as the system was flooded (e.g., Allen and Posamentier, 1993;Dalrymple et al., 1992;Zaitlin et al., 1994). Deep, topographic lows cut by the paleodrainage system form the incised valley and are important in defining the estuaries which transgress them, including controlling the location of the tidal inlet (Morton and Donaldson, 1973). Additionally, topographic highs in the form of fluvial interfluves and terraces provide both increased elevation in which barrier islands are pinned upon (e.g., Raff et al., 2018;Shawler et al., 2020), as well as sediment supply to the barrier via wave and tidal forces (e.g., Riggs et al., 1995;Anderson et al., 2016;Hollis et al., 2019). This study supports this growing body of literature, and finds that the estuarine system is intimately controlled by its antecedent geology.

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
The underlying fluvial topography, interpreted as horizon H6 (Figs. 21, 23B-C), is observed to control the morphology of all overlying units by providing accommodation in which subsequent stratigraphic units fill. H6 is observed as a channelized, erosional surface within the seaward portion of the study area, reaching depths of ~50 ms (~38 mbsl), and that gradually shallows landward within the valley to ~22 ms depth (~29 mbsl) (Figs. 21, 23B-C ). Where H6 is relatively deep, accommodation is created in which overlying units preferentially fill. This is clearly observed as the deep topography of H6 along the eastern and western flanks of the seaward portion of the study area correlate well with the lateral variations in thickness of seismic units U5, U4, U3, U3a, U3b and U2 (Figs. 21,22). This is compounded by the observation that channels created by the underlying fluvial surface are propagated upward throughout the stratigraphic section, owing to the influence in which this surface has on the spatial variations in channelization of strata above (Figs. 5,9,11,18,19).
Beyond creating accommodation for estuarine sediments to fill, the morphology of the incised valley itself appears to promote barrier island initiation. Figure 23A displays a 20 ms envelope median coherency extracted for the approximate base of the Trinity incised valley (Swartz, 2019). This coherency extraction, although likely incorporating anomalous amplitudes throughout the entire incised valley fill, shows the presence of sinuous channel forms, shallow fluvial terraces, lateral accretion surfaces, and a well-defined overall valley extent (Swartz, 2019). In the seaward-most portion of the study area, the valley extent narrows to ~2.5 km in width (Fig. 21, 23A), which rapidly widens landward to ~10 km in width (Fig. 5, 23A) over a distance of ~1 km (Fig. 23A). This framework effectively "bottlenecks" the estuarine system, promoting tidal units (U3, U3a, U3b) to preferentially deposit along the seaward and eastern portions of the study area (Figs. 19,22,23D). The influence of the valley extent on determining the location of tidal deposits, and by proxy paleo-barrier island position (e.g., Thomas and Anderson, 1994;Rodriguez et al., 2004;Anderson et al., 2016), is best seen in Figure 11, where reflectors associated with the tidal delta unit (U3b) switch dip direction from landward-dipping to seaward dipping. Dip direction is a key indicator of flow direction, interpreted as landwarddipping, flood-tidal delta deposits, and seaward-dipping, ebb-tidal delta deposits (e.g., Rodriguez et al., 1998;Anderson et al., 2008). The presence of a fluvial terrace underlying this switch in tidal flow direction is thus interpreted as the underlying fluvial topography providing a direct "pinning point" for the barrier (e.g., Raff et al., 2018;Shawler et al., 2020). However, the underlying fluvial geology does not act as a sediment source for the paleo-barrier island system, which may be expected from previous studies (e.g., Riggs et al., 1995;Anderson et al., 2016;Hollis et al., 2019;Shawler et al., 2020). This is demonstrated by the top of fluvial sediments (H6) very rarely being truncated and reworked by overlying tidal (H4, H4a, H4b) or wave erosional surfaces (H2) (Figs. 5,(7)(8)(9)11,19). Instead, sediment sources reworked by tidal units U3, U3a, and U3b, and supplied to the barrier island system, are preferentially from the bayhead delta (U5) and central bay (U4) units (Figs. 5,(7)(8)(9)(10)15,19,20).
In addition, the relatively shallow fluvial topography within the landward portion of the study area provides decreased accommodation within the back-barrier, promoting the aggradation of barrier-proximal outer bay unit (U2) for ~2.5 ka (Figs. 5 -20;Standring et al., 2019). The transition from an active tidal inlet and tidal delta to an outer bay environment occurred ~8.5 ka (Figs. 17, 18;Standring et al., 2019), indicating the deactivation and infilling of the tidal inlet at this point. This evolution is manifested by the lower bounding surface of the outer bay unit (H3) truncating strata within tidal units, and depositing channelized, aggradational outer bay strata above (Figs. 7,8,10,11,15,19,20) until the onset of inner shelf sediments (U1)

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
~6.9 ka (Standring et al., 2019). The deactivation of a tidal inlet has been proposed to be followed by the complete overstepping and in-place drowning of the barrier island system, wherein an abrupt increase in back-barrier accommodation would increase tidal prism, drown low-lying marshes and tidal flats, and lead to the complete drowning of the estuary (FitzGerald et al., 2006;Fitzgerald et al., 2008;Storms et al., 2008;Ronchi et al., 2019). However, following the deactivation and closure of the tidal inlet, the landward increasing fluvial elevation helped to promote the vertical aggradation of outer bay strata, and possibly stability for the barrier island system, as opposed to retrogradation and barrier disintegration or overstepping associated with an increase in back-barrier accommodation. This finding therefore agrees with previous studies suggesting the reduced back-barrier accommodation provided by antecedent topographic highs can stabilize a barrier island system (e.g., Shawler et al., 2020).

Paleo-barrier Island Evolution
The initiation of the estuarine system within this study area is proposed to occur ~9.8-9.6 ka (Figs. 24-B), indicated by the transition from a fluvial to upper bay/bayhead delta environment in cores PC-2 (Fig. 17) and PC-4 (Fig. 18;Standring et al., 2019). During this time, it has been proposed that a large, 75 km long estuary extended from approximately 20 km offshore modern Galveston Bay, to seaward of Thomas Bank (Fig. 24A; Thomas and Anderson, 1994;Anderson et al., 2008Anderson et al., , 2016. Thomas and Anderson (1994) identified a paired tidal inlet and upper bay/bayhead delta facies bound by upper and lower flooding surfaces, in which the tidal inlet facies was located just seaward of Thomas Bank (Fig. 24A). A flood tidal delta facies was identified extending landward ~20 km from this tidal inlet location. The upper bay/bayhead delta sequence was located depositing within and landward of the confluence of the Trinity-Sabine river valleys, and of which its most landward extent reached into this study's area of focus (Fig. 24A). Dating of this parasequence shows it initiated by ~10 ka (Thomas and  , agreeing well with dates within cores PC-2 (Fig. 17) and PC-4 ( Fig. 18) that an upper bay/bayhead delta environment existed within this study's area of focus during this time.

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
Thus, we suggest that the upper bay/bayhead delta deposits observed in this study, identified as seismic unit U5, initiated ~9.8 -9.6 ka, and represents the landward-most component of an estuary extending almost 75 km seaward (Figs. 24A-B; Thomas and Anderson, 1994).
At approximately 8.8 ka, the area transitioned from an upper bay/bayhead delta environment to a central bay environment (Figs. 24C-D), although PC-2 provides a date within this unit of ~9.6 ka (Figs. 17, 18;Standring et al., 2019). The deposition of central bay sediments (U4) draped on top of bayhead delta strata, passively filling accommodation, but was largely confined within the valley walls (Figs. 5,6,8,16,19,. Although the true extent of the bay may have extended beyond the valley extent, severe subsequent reworking by the tidal (H4, H4a, H4b) (Figs. 5,(7)(8)(9)(10)15,19,20) and transgressive (H2) ravinements (Fig. 8, 16) limits this interpretation. It is also likely that the initiation of the tidal inlet and delta (U3a, U3b) occurred during this time (Figs. 24C-D), as carbon dates provided in GC-2 (Fig. 15) and PC-2 ( Fig. 17) provide an age of at least 8.5 ka (Standring et al., 2019). The extreme narrowing of the incised valley within the seaward and eastern portion study area effectively controlled the spatial extent of the tidal inlet and delta, and thus the location of barrier island (Fig. 23D, 24C-D).
These data are inconsistent with previous interpretations which assert a barrier island existed adjacent to Shepard and Heald Banks from 8.7 to 7.7 ka (Thomas and Anderson, 1994;Rodriguez et al., 2004). Thomas and Anderson (1994) observe a large tidal delta deposit within the Trinity-Sabine valley extending seaward from Heald Bank to just landward of Thomas Bank (Fig. 24C). However, this tidally influenced facies lacks evidence of a tidal inlet deposit, which can be more precise in locating paleo-barrier position than solely tidal deltas due their smaller

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
and more discrete geographic extent (e.g., Thomas and Anderson, 1994). Additionally, Rodriguez et al., (2004) interprets estuarine strata beneath Heald Bank with dates ranging from 8.4 to 7.7 ka. They describe this facies as consisting of landward-dipping seismic reflectors, characteristic of a flood-tidal delta's landward transport direction, and indicates the paleoshoreline lies seaward of this location. However, the seismic data in which these cores were tied to were of low quality (Rodriguez et al., 1999), and the direction of dipping reflectors may be untrustworthy without reprocessed and/or new geophysical data. Therefore, this large tidal delta deposit, given its age constraints of 8.7 to 7.7 ka, possibly represents the ebb-tidal delta component of the barrier island system initiating ~8.8 ka within this study's area of focus (Fig.   24C). However, sparse geophysical data between these two areas hinder conclusive evidence for how this tidal facies relates to this study. By ~8.5 ka, the estuarine system experienced a major shift from central bay (Fig. 17, 18) and tidal inlet and delta environments (Figs. 13,17), to that of an outer bay environment (Figs. 24E-F; Standring et al., 2019). This is a high-energy environment just landward of the barrier inclined to a mixture of tidal, storm, and wave processes (Dalrymple et al., 1992;Zaitlin et al., 1994). This led to the closure and infilling of the tidal inlet and tidal delta units, providing a total lifespan of the tidal system of ~300 years (Figs. 24E-F). Rather than a complete overstepping and in place drowning of the estuary, which is expected following the deactivation of the tidal inlet (e.g., FitzGerald et al., 2006, FitzGerald et al., 2008, reduced back-barrier accommodation provided by increased landward antecedent topography (Figs. 23B-C, 24E-F) helped to promote the aggradation of outer bay sediments, and barrier island stability for the next ~1.6 kyr. By ~6.9 ka, cores GC-2 and GC-5 indicate a shift to an inner shelf environment, leading to the deposition of U1 (Figs. 14, 15, 24G-H; Standring et al,. 2019). During this time, the paleo-

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
barrier retreated and degraded, as evidenced by multiple washover deposits overlying an erosive transgressive lag in GC-6 ( Fig. 13). This barrier-related material filled multiple 2 to 4.5 km long, 1.3 to 1.5 km wide, 0.75 to 2.8 ms (~1 to 2 m) deep channel-like features focused on the eastern side of the valley (Fig. 22 -U1, 24G-H). It is possible that the preservation of this unit is focused on the eastern side of the valley (Figs. 24G-H) due to its heavy reworking of underlying mudrich outer bay and sand-rich tidal delta units (Figs. 5,8,13), as opposed to the western side of the valley where it predominantly reworks solely mud-rich outer bay and central bay units (Figs. 5,6,8,16). Barrier retreat and degradation along the eastern side of the valley persisted until at least ~4.3 ka, as observed by a dated transgressive lag shell in GC-6 ( Fig. 13), while the western side of the valley rapidly transitioned into an inner shelf environment. During this ~1.7 kyr span, the estuarine system was likely inundated by marine sediments, the barrier was degrading, and the shoreline was landward of the study area, possibly forming the Galveston Island by 5.3 ka and Bolivar Peninsula by 2.8 ka ( Fig. 24G; Rodriguez et al., 2004;Standring et al., 2019); however, the lack of data between this study's focus area and the modern shorelines hinder the ability to analyze the paleoshoreline evolution past ~6.9 ka as observed in cores GC-2 (Fig. 15) and GC-5 (Fig. 16). Figure 24: A) Paleoshoreline evolution map from 9.6 to 8.8 ka for the northern Gulf of Mexico inner shelf. The Trinity-Sabine incised valley is drawn as an opaque blue polygon, and is from Thomas and Anderson (1994). A revised interpretation of the Trinity incised valley from Swartz (2019) is displayed as a bold dark polygon.

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
Locations and extent for tidal inlet, tidal delta, and upper bay/bayhead delta are modified from Thomas and Anderson (1994). Fluvial extent is modified from Anderson et al. (2008), although the boundary between fluvial and upper bay/bayhead delta facies is not available, as denoted by a white question mark and dotted boundary. Black dotted box represents outline of 3D schematic in the following figure. B) Three-dimensional schematic of the portion of the Trinity incised valley within this study's area of focus. An upper bay/bayhead delta facies, identified in this study as seismic unit U5, deposits within the valley overlaying fluvial sediments. C). Paleoshoreline evolution map from 8.8 to 8.5 ka for the northern Gulf of Mexico inner shelf. The fluvial and bayhead delta locations and extent are modified from Anderson et al. (2008). Facies within the dotted black box (tidal inlet, tidal delta, central bay) are mapped and interpreted within this study. The tidal delta facies lying ~30 km seaward, and adjacent to Heald and Sabine Banks, has been modified from Thomas and Anderson (1994). This has been re-interpreted as a possible ebb-tidal delta. D) Three-dimensional schematic of the Trinity incised valley within this study's area of focus. The estuarine system is initiated, and controlled by the underlying incised valley's constricted geometry. This results in the development of stable barriers, the deposition of central bay sediments (U4), and preferential deposition of tidal inlet (U3a) and tidal delta (U3b) deposits along the eastern portion of the valley. E) Paleoshoreline evolution map from 8.5 to 6.9 ka for the northern Gulf of Mexico inner shelf. Locations and extent for fluvial, upper bay/bayhead delta, and central bay deposits are from Anderson et al. (2008). F) Three-dimensional schematic of the Trinity incised valley within this study's area of focus. Stable, yet degrading barriers led to the closure of the tidal inlet (U3a) and tidal delta (U3b) units, and the widespread deposition of outer bay (U2) sediments. Decreased back-barrier accommodation space resulting from increased fluvial topography promoted aggradation and resilience of outer bay sediments as opposed to retrogradation and marine incursion. G) Paleoshoreline evolution map from 6.9 to 5.3 ka for the northern Gulf of Mexico inner shelf. Location and extent of upper bay/bayhead delta, central bay, and tidal facies are from Anderson et al. (2008). H) Three-dimensional schematic of the Trinity incised valley within this study's area of focus. Degrading barriers migrated landward across the study area, leading to the deposition of inner shelf/transgressive lag deposits (U1). Large channels composed of washover deposits and transgressive lags are remnants of the now-eroded barriers. The paleoshoreline at this time likely transgressed to Galveston Island by 5.3 ka .

THIS MANUSCRIPT IS A NON-PEER REVIEWED EARTHARXIV PRE-PRINT
To investigate the role of antecedent topography in the evolution of an ancient barrier island system, we focus on the Trinity incised valley offshore modern Galveston Bay in the central Gulf of Mexico. We show that the underlying fluvial topography in which the barrier island system is transgressing plays a significant role in barrier island initiation and stability.
Although there is little to no tidal reworking and redistribution of the underlying fluvial sediments, this section provides higher elevation and reduced accommodation for back-barrier sediments. This increased fluvial elevation promotes vertical aggradation of back-barrier sediments and barrier island stability as opposed to transgression and barrier island disintegration. Additionally, this study provides a methodology to study the evolution of a paleo-barrier island system with sparse preservation of the barrier itself; instead, we investigate preserved back-barrier (deltaic, bay, and washover) and tidal (tidal inlet, tidal deltas) sediments to construct the depositional and erosional evolution of this paleo-barrier island system. Lastly, we revise the established Holocene paleoshoreline model proposed by Rodriguez et al., 2004 to suggest that the model includes a paleoshoreline position within our study are landward of Heald Bank, in agreement with previous authors (Thomas and Anderson 1994;Anderson et al., 2016).