Louisiana State University LSU Digital Commons

LSU Master's Theses Graduate School

2008 Late quaternary geochronologic, stratigraphic, and sedimentologic framework of the Trinity River incised valley: east Texas coast Matthew Gene Garvin Louisiana State University and Agricultural and Mechanical College, [email protected]

Follow this and additional works at: https://digitalcommons.lsu.edu/gradschool_theses Part of the Earth Sciences Commons

Recommended Citation Garvin, Matthew Gene, "Late quaternary geochronologic, stratigraphic, and sedimentologic framework of the Trinity River incised valley: east Texas coast" (2008). LSU Master's Theses. 1122. https://digitalcommons.lsu.edu/gradschool_theses/1122

This Thesis is brought to you for free and open access by the Graduate School at LSU Digital Commons. It has been accepted for inclusion in LSU Master's Theses by an authorized graduate school editor of LSU Digital Commons. For more information, please contact [email protected]. LATE QUATERNARY GEOCHRONOLOGIC, STRATIGRAPHIC, AND SEDIMENTOLOGIC FRAMEWORK OF THE TRINITY RIVER INCISED VALLEY: EAST TEXAS COAST

A Thesis

Submitted to the Graduate Faculty of the Louisiana State University and Agricultural and Mechanical College in partial fulfillment of the requirements for the degree of Master of Science in

The Department of Geology and Geophysics

by Matthew Gene Garvin B.S., Texas Tech University, 2005 A.S., Eastfield College, 2002 December 2008

ACKNOWLEDGEMENTS

I would like to thank my advisor and committee for their thoughtful reviews of the manuscript and many helpful suggestions. I also thank the Department of Geology and Geophysics at Louisiana State University and the Applied Depositional Geosystems

Program and its supporters for the opportunity to focus on research in earth sciences. I would like to extend my thanks to those who have provided assistance with field work and many thoughtful discussions back in the lab. To name just a few: Clint Edrington,

Eric Prococki, Jill Womack, Jenny Garvin, and Alan Hidy, without their help this work would not have been possible.

The list of names that have provided moral support throughout the entirety of my time here at LSU, and throughout my educational career would be endless, but I would like to extend my thanks to my family and friends, previous professors who inspired my love for geology, and of course, my wife, Jenny.

This research was supported by grants from the South Texas Geological Society

(Jones-Amsbury Grant) (grant to M. Garvin), the American Association of Petroleum

Geologists (Edward B. Picou Jr. Student Grant-in-Aid) (grant to M. Garvin), the LSU

Department of Geology and Geophysics (Baker-Hughes and Halliburton Funds) (grant to

M. Garvin), ConocoPhillips Inc. (grant to M. Blum), and The U.S. National Science

Foundation (#115405192) (grant to M. Blum).

ii TABLE OF CONTENTS

ACKNOWLEDGEMENTS...... ii

LIST OF TABLES...... v

LIST OF FIGURES ...... vi

ABSTRACT...... x

INTRODUCTION ...... 1

BACKGROUND ...... 5 Present Day River System and Setting and Study Area ...... 5 Geologic Setting: Gulf of Mexico Basin ...... 9 Climate and Sea–level Change ...... 12 Sea Level History...... 12 Glacial Period River System and Setting...... 14 Older Alluvial Plains of the Trinity River System: The Beaumont Formation ...... 18 The Post–Beaumont Lower Trinity Valley...... 21 “Deweyville” Deposits...... 21 Correlation to Offshore Record ...... 23 Recent Deposits of the Trinity River ...... 25 Paleohydrology ...... 25 Sediment Supply Controls ...... 26 Stored and Excavated Sediment Volumes ...... 27

METHODS ...... 28 Stratigraphic and Sedimentologic Framework...... 28 Chronologic Framework ...... 30 Paleohydrology ...... 30 Calculation of Sediment Supply ...... 35 Calculation of Stored and Excavated Sediment Volumes and Mass ...... 35

RESULTS ...... 38 Deweyville Deposits: General Framework...... 38 High Deweyville (HD) Unit...... 41 Middle Deweyville (MD) Unit ...... 44 Low Deweyville (LD) Unit...... 49 Deweyville Geochronology ...... 55 Post–Deweyville Deposits ...... 55 Channelbelt and Floodplain ...... 58 Bayhead Delta and Estuary...... 63 Valley Slopes, Longitudinal Profiles, and Channel Slopes ...... 67 Paleohydrology ...... 71 Sediment Supply ...... 76

iii Sediment Export During Valley Evolution...... 76

DISCUSSION...... 82 Valley–Fill Architecture ...... 82 Geochronology...... 90 Valley–Fill Evolution...... 92 Paleohydrology ...... 93 Sediment Mass Balance ...... 95 Allogenic Controls...... 98 Fluvial Response to Allogenic Controls ...... 102

CONCLUSIONS...... 108

REFERENCES ...... 111

APPENDIX: TRINITY RIVER GUAGING STATION DATA...... 122

VITA...... 124

iv LIST OF TABLES

Table 1. Planform properties and empirical equations used to assess paleohydrological estimates and estimates of modern discharge...... 34

Table 2. Optically stimulated luminescence results...... 56

Table 3. Valley longitudinal profiles, sinuosity, and channel slopes...... 71

Table 4. Estimates of mean annual discharge (Qm) using different parameters. General equations based upon a large number of rivers from different geographic settings...... 73

Table 5. Estimates of Qm, Q2, and Qmax. General equations based upon a large number of rivers from different geographic settings and equations derived from rivers grouped by the sediment caliber of the channel are noted...... 74

Table 6. Volumetric calculations of stored and excavated sediment...... 79

Table 7. Calculations of sediment mass stored in channelbelts and rates of excavated sediment mass...... 80

Table 8. Mass balance calculations of extra sediment due to excavated sediments per unit valley length (m), per unit channel length (m), and modeled plus excavated sediment delivered past the modern shoreline per year ...... 81

v LIST OF FIGURES

Figure 1. Digital Elevation Model, (DEM) of east Texas and northwestern Gulf of Mexico Bathymetry showing elevation range in the Trinity drainage basin...... 6

Figure 2. Satellite image of southeast Texas ...... 7

Figure 3. Average annual temperature (°C) and average annual precipitation (cm) over east Texas and the Trinity River drainage basin...... 8

Figure 4. (A). Generalized geologic map of east Texas, showing units present within the Trinity River drainage basin. Modified from USGS (2007), www.usgs.gov...... 10

Figure 4. (B). Generalized lithology of east Texas geologicic units...... 11

Figure 5. Combined sea level curves for the last 120 ka, modified from Waelbrock et al. (2002) and Peltier and Fairbanks (2006)...... 13

Figure 6. Temperature reconstruction of mid OIS 3–OIS 1 (40 ka–present) ...... 16

Figure 7. Stratigraphic column of the Late Quaternary in East Texas and correlative oxygen isotope stages...... 19

Figure 8. (A). Satellite image of Pleistocene Beaumont alluvial deltaic plains, (B) Interpreted geologic map of Beaumont alluvial deltaic plains ...... 20

Figure 9. Geologic map of northwestern GOM showing the OIS 4–2 fluvial valleys as they extended towards the shelf edge and merged to deposit the Brazos–Trinity shelf edge delta and filled the Brazos–Trinity slope minibasins ...... 24

Figure 10. Collection of cores in the field. (A). Vibrating core into the ground, (B). Core retrieval with the help of a winch...... 29

Figure 11. Collection of optically stimulated luminescence samples in the field ...... 31

Figure 12. Measurements of channel planform properties. (A). Satellite image depicting the measurements of meander wavelength, Lm, (B). Satellite image depicting the measurement of radius of curvature, rc...... 32

Figure 12. Measurements of channel planform properties. (C). Satellite image and, (D). Digital elevation model depicting the measurement of bankfull width, Wb...... 33

Figure 13. Cartoon depicting the removal and export of sediment to downdip reaches . 36

Figure 14. (A). Satellite image of the Coastal Plain incised valley ...... 39

vi Figure 14. (B). Detailed geologic map of the Coastal Plain with OSL sample locations and overall valley–fill cross section locations ...... 40

Figure 15. (A) Measured section where OSL sample “HD_Liberty” was collected. Ten meter thick sandy point bar. Basal erosion surface is just below water level. (Flow direction is from left to right), (B) upper bar with typical cross bar channel deposits ~15– 39 cm thick., (C) trough cross beds 11–20 cm thick., (D) trough cross beds 14–20cm thick containing large clay rip–up clasts...... 42

Figure 16. (A) Sandy point bar covered by a thin A and E horizons, thick Bt horizon with subangular to angular peds. Cox horizon exhibits diminishing coloration with abundant lamellae. OSL sample MD_Romayor_railroad was collected in fine to medium grained cross–bedded sands in the upper part of the bar deposits. The lower bar contains trough cross bedded sands and gravels (A, D). Blue lettering marks location of pictures for the following figures. (B) ripple cross–laminated sands and small scale trough cross–beds 3– 5cm thick. (C) Upper bar deposits where large scale cross–beds (ca. 30 cm) typical of cross–channel chute bar facies, (D) basal erosional unconformity of Middle Deweyville unit on top of pre–Deweyville channel and channel–fill sands and muds and fine–grained topstratum...... 45

Figure 17. (A) Middle Deweyville outcrop sandy point bar and channel fill mud and sand where OSL sample MD_Kenefick was collected. Grain size ranges from very coarse sand with some pebble–sized gravels fining upwards to alternating layers silt and clay and very fine–grained ripple laminated sand and silt interpreted as channel–fill deposits., (B) channel fill silts and very fine– grained sand interbedded with clay, overlain by well developed Bt horizon and a ca. 80 cm thick A horizon,...... 47

Figure 17. (C) channel fill mud and sand containing ripple–laminated silts and very fine– grained sand, (D) 30 cm thick cross–bedded very coarse sand with some gravel grading upwards into 3–10 cm thick cross–bedded sands with alternating with ripple cross– laminated silts and sands...... 48

Figure 18. (A) Low Deweyville outcrop where OSL sample LD_Kenefick was collected. Sandy point bar grading upwards into channel fill mud and sand overlain by post– Deweyville clays and silts coarsening upwards into silts and very fine–grained sands interpreted as post–Deweyville floodplain deposits coarsening upward into modern levee deposits. (B) modern levee containing ripple laminated silts and very fine–grained sand interbedded with clay., (C) post–Deweyville distal levee and floodplain muds and silts.51

Figure 18. (D) soil profile developed in Low Deweyville channel fill muds and sands., (E) carbonate precipitation due to burial of soil profile and ground water reflux, commonly found in Low Deweyville buried terraces., (F) dark grey root mottles developed in channel sands consisting of trough cross beds comprised of medium grained sand, cross–beds 10–25 cm thick, (G) cross–beds 11–50cm thick in fining upwards to very coarse–grained sand and pebble gravels...... 52

vii Figure 19. (A) Site where OSL sample LD_Liberty_South was collected. Post– Deweyville channel fill muds and fine sands lying unconformably on Low Deweyville channel sands. (B) coarsening upward modern floodplain and levee deposits resting unconformably on top of Low Deweyville paleosol., (C) photograph of OSL sampling in medium–grained cross–bedded sands...... 53

Figure 20. Cedar Valley site located downstream of the Lake Livingston Dam...... 54

Figure 21. Graphical representation of OSL dates with error bars...... 57

Figure 22. Geologic map showing the locations of post–Deweyville channelbelts and figure locations 23, 18, and 19...... 59

Figure 23. Photos of modern point bar core...... 61

Figure 24. Core locations across I–10 (C–C’), the modern bayhead delta, and downdip cross section locations...... 62

Figure 25. Core photos illustrating older marsh surface and overlying central basin estuarine muds containing oyster and Rangia shells (11.01.06 BD1 and BD8). Core photo of modern marsh surface (BD 10)...... 64

Figure 26. Core photos of 9.30 BD1 showing prodelta muds coarsening upwards to bayhead delta sands typical of progradational deltaic deposits ...... 65

Figure 27. Core photos of BD 12. Moving vertically from base of core, prodelta clays and silts, coarsening upward to channel deposits containing laminated sands, silts, and muds, alternating between channel silts, sands, and muds. Top of core represents channel levee and aggradational marsh surface ...... 66

Figure 28. Digital elevation model of the Trinity River Coastal Plain displaying the incised valley axis...... 68

Figure 29. Longitudinal profiles of stratigraphic units. (A) Onshore Coastal Plain flood plain longitudinal profiles...... 69

Figure 29. (B) Data points and flood plain longitudinal profiles projected to the shelf edge...... 70

Figure 30. Calculated sediment discharge for the Trinity River over the last glacial– interglacial cycle using the Syvitski and Milliman (2007) BQART model...... 77

Figure 31. Cross–sections across the Coastal Plain Trinity River incised valley...... 83

Figure 32. 3–D schematic of valley–fill components and their stratigraphic architecture...... 85

viii Figure 33. Cross–sections in downdip reaches of the Coastal Plain valley. (A) Cross– section D–D' depicting recent valley–fill sediments and interpreted depositional environments...... 88

Figure 33. Cross–sections in downdip reaches of the Coastal Plain valley. (B) Dip oriented cross–section E–E’...... 89

Figure 34. Composite graph showing ages of stratigraphic units, modeled (BQART) and modeled plus excavated sediment discharge over the last 120 kyrs...... 97

ix ABSTRACT

Optically stimulated luminescence dates from the last glacial period falling–stage and lowstand Deweyville deposits along the Trinity River valley in Texas provided insight to their timing of deposition and allogenic controls on fluvial processes. Three distinguishable periods of incision and lateral channelbelt migration were the effect of both lowered sea levels and climate controlling factors within the drainage basin. Valley widening and deposition of the High, Middle, and Low Deweyville units were constructed and subsequently preserved as fluvial terraces during oxygen isotope stages

(OIS) 3 through 2, from 65–32.5 ka, 32.5–23 ka, and 23–16 ka respectively.

Although numerous workers using discharge retrodiction equations have inferred much greater discharges as the primary cause of the larger paleochannels of the glacial age Deweyville units, the reconstructed paleohydrology of the Trinity River using measurable parameters such as channel width, sediment caliber, and other geomorphic planform properties yields contrasting results to earlier studies.

Sediment mass stored and exported during valley creation was assessed and calculated using a mass balance approach. Results clearly show that falling stage and lowstand fluvial deposits account for a large volumetric portion of the Trinity valley–fill sediments. Results also show that excavated sediment from the Coastal Plain represents only ca. 13% of the total hinterland derived sediment flux delivered by the Trinity system to downstream point sources over the last glacial period: periods of channel incision and valley deepening contribute little additional export of sediment, rather the process of lateral migration and channelbelt formation is contemporaneous with enhanced sediment

x flux to downslope systems. These results contrast long held concepts of incision and bypass during valley creation.

An alternative hypothesis using a process based framework is presented as the primary cause of the larger Deweyville paleochannels. The resultant alluvial morphology of the Deweyville units was the result of floodplain longitudinal profile adjustment to sea–level change and the preexisting boundary conditions of the drainage basin and the emerging topography of the shelf, and an attempt of the fluvial system to attain the minimum channel slope required to transport upstream controls on water and sediment loads.

xi INTRODUCTION

It is well–known that continental margin fluvial systems responded to fluctuating

Late Quaternary glacial–interglacial climate and sea–level change (reviewed in Blum and

Törnqvist, 2000). However, interpretation of causal factors is difficult because allogenic forcing mechanisms as well as pre–existing boundary conditions can produce differing signatures in the morphology of alluvial channels through convergence and divergence of system response (Schumm, 1991). Because of the complicated nature of these systems, careful interpretation relies on the identification of multiple forcing mechanisms and accounting for all possible system responses.

In light of new views on the importance of source–to–sink studies, Late

Quaternary valley fills are integral to linking drainage basin source terrains to the sinks in the sedimentary basin, and act as a buffer zone between the two that determines sediment dispersal to the shelf, slope, and deep basin (Blum and Törnqvist, 2000; Blum, 2003;

Blum and Aslan, 2006; Blum, 2007). Long held concepts of excavation and bypass of sediment due to valley creation have been viewed as the mechanism for increased sediment flux to downslope systems. Recent investigations utilizing mass balance approaches have shown that volumes of sediment removal from valley systems are typically an order of magnitude smaller than background rates in continental margin systems with large hinterland drainage areas (Blum and Törnqvist, 2000). Due to recent advances in the understanding of alluvial morphology and the temporal and spatial resolution of Quaternary research, valley systems can now be revisited to enhance understanding of morphologic response to various forcing mechanisms and to begin to approach first order approximations of sediment storage and excavation rates within the

1 fluvial system and how sediment storage and export impacts sediment delivery to point sources farther downdip.

The lower Trinity River, Gulf Coastal Plain of east Texas, is an ideal location to study these issues. Previous work has shown the Trinity River responded to climate and sea–level changes of the last glacial period through incision and channel extension across the now submerged shelf, and the valley–fill records contrasting alluvial morphologies between the glacial period system and the modern counterpart (Barton, 1930a; Bernard,

1950; Gagliano and Thom, 1967; Thomas and Anderson, 1989; Thomas, 1990; Thomas and Anderson, 1991; Thomas and Anderson, 1994; Morton et. al, 1996; Blum and

Törnqvist, 2000; Blum, 2003; Anderson et al., 2004; Blum and Aslan, 2006). Important distinctions between the glacial period and modern river are the “Deweyville” (see discussion below) terraces and deposits, which include the very large meanders, and which are characterized by larger radii of curvature, bankfull widths, and meander wavelengths. These contrasts are well exposed because, in contrast to some of the other valleys along the Texas Coastal Plain, the Trinity valley has not yet filled with sediment during the recent transgression and highstand. Hence, relict channel morphologies are subaerially exposed and can be studied in great detail.

The large relict channels on Deweyville surfaces appear to be ubiquitous around the edges of Gulf of Mexico (GOM) Valleys, with the exception of the Lower Mississippi

Valley, and has led to the typical classification of modern rivers as “underfit” (e.g. Dury,

1964; Dury, 1965; Rotnicki, 1983; Alford and Holmes, 1985; Patton, 1987; Sylvia and

Galloway, 2006). Moreover, relict meander scars have been used to attempt retrodictions of paleodischarge values, have almost always led to the interpretation of formative

2 discharges that were significantly higher than present values, as well as overall wetter climatic conditions (Barton, 1930a; Bernard, 1950; Gagliano and Thom, 1967; Alford and Holmes, 1985; Gagliano, 1991; Blum et al., 1995, Durbin et al., 1997; Durbin, 1999;

Sylvia and Galloway, 2006). However, there is no consensus on whether or not wetter conditions reflect increased precipitation or if cooler climatic conditions resulted in increased effective moisture (Gagliano and Thom, 1967; Alford and Holmes, 1985; Blum et al., 1995; Durbin, 1999; Blum and Törnqvist, 2000; Blum and Aslan, 2006; Sylvia and

Galloway, 2006). On the other hand, other workers (e.g. Thomas, 1990; Thomas and

Anderson, 1991; Thomas and Anderson, 1994; Anderson et al., 1996) have attributed the

Deweyville terraces to sea-level changes, without addressing the issue of relict meander geometries.

It is reasonably well–known that the three major controls on incised valley–fill

(IVF) deposits are: 1) fluctuations in base level occurring during periods of falling and rising sea level, 2) climatic fluctuations present throughout the drainage basin during valley creation and filling causing unsteadiness in discharge (Q) and sediment supply

(qs), and 3) the pre–existing boundary conditions of the system (Blum and Törnqvist,

2000; Blum, 2003; Anderson et al., 2004; Blum, 2007). Critical assessment of controls on changing fluvial morphology in the Trinity and other river valleys of the Texas coast has been limited by the lack of geochronological data, critical examination of facies, and development of a process–based physical framework for channel adjustments to changes in forcing mechanisms. Therefore, the purpose of this study is to:

1) Refine the mapping and stratigraphic framework of the IVF using outcrop and

core data, satellite imagery, and digital elevation models to clearly define the

3 architecture of the falling stage to lowstand deposits, and geometry of the onlapping wedge of Holocene alluvial sediments and modern bayhead delta,

2) Develop a chronologic framework for valley–fill sediments with optically stimulated luminescence (OSL) dating to correlate events to climatic and sea level conditions that prevailed during their respective time periods of deposition,

3) Model the paleohydrology of the glacial period Trinity River using published discharge retrodiction methods, and to compare results with observed and calculated results from the modern river,

4) Model sediment supply during the glacial period and quantify sediment storage and export during initial valley incision and subsequent valley–fill evolution, and

5) Test how unsteadiness in discharge and sediment supply, and sea–level change, produce different signatures within the valley–fill deposits, and evaluate the relative importance of different controls on alluvial architecture.

4 BACKGROUND

Present Day River System and Setting and Study Area

The Trinity River drainage basin is approximately 46,000 km2, with a maximum relief of 300–400 m (Fig. 1; Ulery et al., 1994; USGS, 2008), and a total channel length

of 885 km (BEG, 1996). The climate is subtropical humid with warm summers, and a

dominant onshore flow of maritime–tropical air from the GOM (Larkin and Bomar,

1983; Ulery et al., 1994). Mean annual temperature ranges from ca.18 C in the northwest

to ca. 21 C in the southeast, and mean annual precipitation varies from less than 90 cm in

the northwest to greater than 122 cm in the southeast (Fig. 3; Larkin and Bomar, 1983).

The passage of midlatitude cyclonic storms in Spring and Fall dominate modern flood

regimes (Blum and Aslan, 2006). Compared to potential evapotranspiration, the drainage

basin receives a summer deficiency of precipitation and a winter surplus (Ulery et al.,

1994). Runoff ranges from less than 10.2 cm in the northwestern edge of the drainage

basin to greater than 41cm in the southeast (Ulery et al., 1994). Environmental settings

vary across the drainage basin due to variations in geology, location, topography,

physiography, streams, aquifers, natural vegetation, soils, and climate (Land et al., 1998).

The study area is located in the lower Trinity valley of southeast Texas along the

northwestern GOM Coastal Plain and encompasses ca. 3,500 km2 (Figs. 1, 2). Average annual runoff for the period 1936–1954 was 7.02 km3/year (LeBlanc and Hodgson,

3 1959). Mean annual discharge, Qm, at Liberty, Texas is 730 m /s, with a contributing

2 drainage of 45,242 km (USGS, 2007). The bankfull discharge, Qbf, at Liberty, Texas is

3 989 m /s (Phillips et al., 2005; 2006). Mean annual sediment discharge, (qs), for the period 1936–1954 was 5.52 million tons (MT) per year (LeBlanc and Hodgson, 1959).

5 99 °0'0"W 98 °0 '0 "W 97°0'0"W 96°0'0"W 95°0'0"W 94°0'0"W

Legend Trinity_River_Drainage_Basin . Elevation_(m) Valu e

33°0'0"N 33°0'0"N High : 667

Low : 0 Gulf_of_Mexico_Bathymetry_(m) Valu e High : 0

32°0'0"N 32°0'0"N Low : - 20

102° 0'0"W 100° 0'0"W 98 °0 '0 "W 96 °0 '0 "W 94 °0 '0 "W 92°0'0"W 90 °0 '0 "W 88°0'0"W 86°0'0"W

43 °0 '0 "N 43°0'0"N

42 °0 '0 "N 42°0'0"N 41 °0 '0 "N ­ 41°0'0"N 40 °0 '0 "N 40°0'0"N

39 °0 '0 "N 39°0'0"N

31°0'0"N 31°0'0"N 38 °0 '0 "N 38°0'0"N 37 °0 '0 "N 37°0'0"N

36 °0 '0 "N 36°0'0"N

35 °0 '0 "N 35°0'0"N

34 °0 '0 "N 34°0'0"N

33 °0 '0 "N 33°0'0"N

32 °0 '0 "N 32°0'0"N

31 °0 '0 "N 31°0'0"N

30 °0 '0 "N 30°0'0"N

30°0'0"N 30°0'0"N 29 °0 '0 "N 29°0'0"N

28 °0 '0 "N 28°0'0"N

27 °0 '0 "N 27°0'0"N

26 °0 '0 "N 26°0'0"N

25 °0 '0 "N Gulf of Mexico 25°0'0"N 0525 0100150200 24 °0 '0 "N 24°0'0"N 23 °0 '0 "N 23°0'0"N 0 125 250 500 750 1,000 Kilometers 22 °0 '0 "N Kilometers 22°0'0"N 102° 0'0"W 100° 0'0"W 98 °0 '0 "W 96 °0 '0 "W 94 °0 '0 "W 92°0'0"W 90 °0 '0 "W 88°0'0"W 86°0'0"W 99 °0 '0 "W 98 °0 '0 "W 97°0'0"W 96°0'0"W 95°0'0"W 94°0'0"W Figure 1. Digital Elevation Model, (DEM) of east Texas and northwestern Gulf of Mexico Bathymetry showing elevation range in the Trinity drainage basin. Location of figure on left displayed on regional image (bottom right). DEM data from National Geophysical Data Center (www.ngdc.noaa.gov).

6 95 °0 '0 "W 94 °5 0'0 "W 94°4 0'0 "W 94°3 0'0 "W 94°20'0"W 94°10'0"W 94 °0'0"W Lake Livingston Dam

30°3 0'0 "N 30°30'0"N Romayor .

30°2 0'0 "N 30°20'0"N

30°1 0'0 "N 30°10'0"N

Liberty

30°0 '0"N 30°0'0 "N Moss Bluff

29°5 0'0 "N 29°50'0"N

Trinity Bay 29°4 0'0 "N 29°40'0"N

Galveston Bay

29°3 0'0 "N 29°30'0"N

29°2 0'0 "N 29°20'0"N 010203040 Kilometers

95 °0 '0 "W 94 °5 0'0 "W 94°4 0'0 "W 94°3 0'0 "W 94°20'0"W 94°10'0"W 94 °0'0"W

Figure 2. Satellite image of southeast Texas. Study is focused on the lower Trinity valley, between Lake Livingston Dam and Trinity Bay. Imagery from www.zulu.gov

7 99°0'0"W 98°0'0"W 97°0'0"W 96°0'0"W 95°0'0"W 94°0'0"W

17 Legend 17 34°0'0"N 70 90 110 34°0'0"N precipitation, (cm) 80 100 ­ temperature, (degrees C) Texas counties Trinity River drainage basin 33°0'0"N 33°0'0"N 19

18 120 32°0'0"N 32°0'0"N 130

31°0'0"N 31°0'0"N 19

30°0'0"N 20 30°0'0"N 140

120 130 130

29°0'0"N 29°0'0"N 21 110 100

28°0'0"N 90 28°0'0"N 22 80 0 30 60 120 180 240 70 Kilometers

99°0'0"W 60 98°0'0"W 97°0'0"W 96°0'0"W 95°0'0"W 94°0'0"W

Figure 3. Average annual temperature (°C) and average annual precipitation (cm) over east Texas and the Trinity River drainage basin. Modified from Larkin and Bomar (1983).

8 LeBlanc and Hodgson (1959), Winker (1979), Galloway (1981), and Blum and Aslan

(2006) describe the Trinity valley as an unfilled valley due to low sediment supply from the contributing drainage basin relative to other valleys of the Texas Coast.

Geologic Setting: Gulf of Mexico Basin

The Trinity River discharges to the passive margin GOM basin, which initiated with the late Jurassic break up of Pangea (Buffler, 1991). The Trinity drainage basin

(TDB) is mostly composed of clastic rocks ranging in age from the Pennsylvanian through the Holocene (Fig. 4). The far northwestern edge of the TDB headwaters originate in Pennsylvanian and Permian age sandstones and shales. Channels then traverse rocks of Cretaceous age that primarily consist of terrigenous or siliciclastic sand, limestone, mudstone, shale and clay or mud. The lower half of the drainage basin is comprised of Cenozoic siliciclastics with lithologies that range from clays or muds, fine– grained mixed clastics, shales, sandstones, siltstones, mudstones, and unconsolidated sands. An increase in siliciclastic deposition began in the Paleocene (ca.60 Ma) because of increasing tectonic activity in the western United States and has dominated the progradational nature of the margin through the present (Winker, 1991; Galloway et al.,

2000). As a result, limestone covers only ca.13% of the drainage basin, whereas softer siliciclastic lithologies account for the rest.

The Inner Coastal Plain is comprised of siliciclastic rocks that dip more steeply and below the Outer Coastal Plain, whereas the Outer Coastal Plain consists of the flat lying Late Quaternary Lissie and Beaumont Formations, a succession of coalescing alluvial–deltaic plains (Deussen, 1914; Blum and Aslan, 2006). In the Trinity River system, the post–Beaumont valley is distinctively entrenched into the Beaumont alluvial

9 99°0'0"W 98°0'0"W 97°0'0"W 96°0'0"W 95°0'0"W 94°0'0"W

Legend 34°0'0"N 34°0'0"N Trinity_River_drainage_basin ­ older_undifferentiated Holocene_Alluvium Late_Holocene-Present_Barrier_System Late_Pleistocene_Deweyville_Terraces 33°0'0"N 33°0'0"N Pleistocene_Ingleside_Barrier_OIS5e_shoreline Quaternary_Beaumont Late_Pleistocene_Fluviatile_Terraces Quaternary_Middle Pleistocene_Lissie Pliocene_Willis 32°0'0"N 32°0'0"N Miocene_Fleming Oligocene Eo-Oligocene Middle Eocene

31°0'0"N 31°0'0"N Eocene Paleocene to Eocene Paleocene Late Cretaceous; Gulfian Series Late Cretaceous 30°0'0"N 30°0'0"N Early Late Cretaceous and Late Early Cretaceous Early Cretaceous; Comanchean Series Early Cretaceous Permian; Wolfcamp Series Pennsylvanian to Early Permian 29°0'0"N 29°0'0"N ico Late Pennsylvanian; Virgil Series ex Pennsylvanian; Virgil Series f M Late Pennsylvanian; Missouri Series lf o Gu Pennsylvanian; Missouri Series 28°0'0"N 28°0'0"N 0 30 60 120 180 240 Kilometers

99°0'0"W 98°0'0"W 97°0'0"W 96°0'0"W 95°0'0"W 94°0'0"W

Figure 4. (A). Generalized geologic map of east Texas, showing units present within the Trinity River drainage basin. Modified from USGS (2007), www.usgs.gov.

10 99°0'0"W 98°0'0"W 97°0'0"W 96°0'0"W 95°0'0"W 94°0'0"W

Legend 34°0'0"N ­ 34°0'0"N _Trinity_River_drainage_basin Lithology limestone sandstone 33°0'0"N 33°0'0"N siltstone mudstone shale gravel sand 32°0'0"N 32°0'0"N sandy_terrace silt fine-grained mixed clastic clay or mud

31°0'0"N 31°0'0"N water

30°0'0"N 30°0'0"N

29°0'0"N ico 29°0'0"N ex f M lf o Gu 28°0'0"N 28°0'0"N 0 30 60 120 180 240 Kilometers

99°0'0"W 98°0'0"W 97°0'0"W 96°0'0"W 95°0'0"W 94°0'0"W

Figure 4 (cont.) (B). Generalized lithology of east Texas geologic units. Modified from USGS (2007), www.usgs.gov.

11 plain surface and is comprised of flanking fluvial terraces and more recent sediments.

The modern river is confined within the valley and discharges into an estuary creating a

bayhead delta.

Climate and Sea–level Change

Sea Level History

High frequency Late Quaternary sea–level fluctuations are due to the waxing and

waning of glaciers because of 100 kyr Milankovitch cycles. Sea–level reconstructions

were chosen based upon their applicability to represent the global eustatic signal over the

last 120 kyrs, the last interglacial period to the present. Waelbrock et al. (2002) use the

oxygen isotope ratios of benthic foraminifera to reconstruct sea level and water

temperature over the past 430 ka, (Fig. 5), based upon changes in the δ18O due to fractionation processes which trap and store isotopes of Oxygen that accumulate in ice during glacial periods. Measured samples are then calibrated to the temperature of the water in which the calcite shells of the benthic foraminifera formed as well as the isotopic composition of the water to interpret the global ice volume during their respective time periods of deposition (Waelbrock et al., 2002). Peltier and Fairbanks (2006) reconstruct sea level history using two methods. In Barbados, sea level is reconstructed by the dating of corals and identifying their morphology and the respective depth ranges in which they form. They also use the ICE–5G model that incorporates the most recent ice sheet growth data, the rotation of the earth data, the position of the coastline with respect to the shelf width and present day bathymetry, and the global glacio–isostatic components (Fig.

5).

12 OIS 1 OIS 2 OIS 3 OIS 4 OIS 5 ) l (m e v e a L Se alent v i e–Equ c I TTii mmee (ka( k a) ) Waelbrock et al. (2002) Peltier and Fairbanks (2006)

Figure 5. Combined sea level curves for the last 120 ka, modified from Waelbrock et al. (2002) and Peltier and Fairbanks (2006).

13 During oxygen isotope stage (OIS) 5, from ca. 120-75 ka, sea level fluctuated between ca. 7–8 meters above present sea level and ca. 35 meters below present. This period is inferred to be the previous interglacial highstand, which marks the final period of deposition for the Beaumont Formation (Durbin et al., 1997; Blum and Price, 1998;

Blum and Aslan, 2006). During OIS 4, (ca. 75–63 ka), sea level fell ca. 70 to 80 meters below present before a subsequent rise into OIS 3. During OIS 3, (ca. 63–26 ka), sea level experienced an initial rise to ca. 60 to 80 meters below present then a slow fall to ca.

88 meters below present around 30 ka, followed by a rapid fall to 120 meters below present, (ca. 30–26 ka), that correlates to the rapid expansion of the Laurentide Ice Sheet

(LIS) (Peltier and Fairbanks, 2006). During OIS 2, (ca. 26–14 ka), sea level reached its lowest position at ca. 120 meters below present, (ca. 26–21 ka), which was during the last glacial maximum (LGM). Sea level rose gradually to 110 meters below present around 16

ka, followed by the rapid rise associated with overall transgression and deglaciation (ca.

16–7 ka). Following the rapid transgression out of the LGM, sea level slowed its overall rise ca. 7 ka.

Glacial Period River System and Setting

During glacial periods, many aspects of the river system are likely to change due to climatic forcing, including changes in temperature, precipitation, and vegetation, which have the most pronounced effects on the river system due to changes in effective moisture, surface runoff, and sediment supply. Glacial–period climate conditions have been examined from empirical studies and through simulations from general circulation climate models.

14 Stute et al. (1992) used dissolved atmospheric noble gases located within the

Carrizo aquifer of south–central Texas to obtain mean annual temperatures from the present back to approximately 40 ka (Fig. 6). The results suggested regional temperatures were approximately 2.5 C cooler during OIS 3 (ca. 33.5 ka to ca. 22.5 ka), and approximately 5.2 C cooler during OIS 2 (ca. 18 ka). Toomey et al. (1993) concluded that average summer temperatures in east Texas during the period from 20 ka to 14 ka were approximately 5 C cooler than present based upon the presence of vertebrate faunas and pollen records not present in the region today. Bryant and

Holloway (1985) used the presence of spruce pollen, Picea glauca, in Boriack Bog to infer a decrease in July mean temperature of 5 C for the full glacial period. Ganopolski et al. (1998) suggest seasonal temperature changes at the LGM of -4 C to -8 C during summer months and -3 C to -6 C during winter, relative to modern mean temperatures.

Roche et al. (2007) show a mean surface air temperature change of -3 C to -6 C over the

Texas region during the LGM, whereas Koch et al. (2004) model mean annual temperature over the Trinity River drainage basin ranging from 10 C in the north to 14 C in the south at 18 ka, a change of -8 C to -6 C, respectively.

Delcourt and Delcourt (1981) and Bryant and Holloway (1985) used the analysis of fossil pollen records in west Texas and south–central Missouri to infer that herbaceous plants and grasses with a relatively small population of conifers were present in north and central Texas during late OIS 3 to OIS 2. Most authors agree the easternmost portion of

Texas was forested and represents a boundary between the deciduous forests of the southeastern United States and the grassland savanna and scrub grassland located within northern and central Texas during OIS 3 (Bryant and Holloway, 1985, Delcourt and

15 C) º e Gas Temperature ( Nobl

Time (ka)

Figure 6. Temperature reconstruction of mid OIS 3–OIS 1 (40 ka–present). Modified from Stute et al. (1992).

16 Delcourt 1985; Hall and Valastro, 1995; Nordt et al., 2002; Koch et al., 2004). Early

work done by Bryant and Holloway (1985) proposed that central and north–central Texas

was heavily forested during the full glacial period from ca.22–14 ka. Toomey et al.

(1993) subsequently examined fossil faunas in caves, Hall and Valastro (1995) examined

pollen assemblages, and Koch et al. (2004) examined tooth enamel from fossil mammals.

The general consensus was that during the LGM, central Texas was covered by

grassland, with deciduous trees and pinyon pines restricted to riparian habitats and

canyons, and C4 plants comprised a large proportion of the vegetation in the region.

Earlier interpretations of widespread vegetation are attributed to differential preservation

of pollen grains due to post–depositional processes (Hall and Valastro, 1995).

In addition to vegetation differences, upland areas were also covered by thick,

deeply weathered soils during the full glacial period (Toomey et al., 1993; Blum and

Valastro, 1994; Cooke et al., 2003). These interpretations are based on mammalian fossil

presence in cave fill deposits in central Texas and erosional soil remnants throughout the

region. Recent dating and interpretation of soil erosion rates has confirmed that during the glacial to interglacial transition, these deep soils were removed due to shifts in climate as aridity and storm intensity increased into the Holocene (Toomey et al., 1993;

Blum and Valastro, 1994; Cooke et al., 2003).

Glacial period moisture regimes are less well resolved because empirical studies are rare, and distinctions between changes in precipitation and changes in evapotranspiration have proven difficult. The presence of mesic vegetation and species not present in the region today, and growth rates of speleothems have all been used as indicators of a cooler and moister glacial period than its modern counterpart (Bryant and

17 Holloway, 1985; Toomey et al., 1993; Hall and Valastro, 1995; Mock and Bartlein, 1995;

Musgrove et al., 2001). Musgrove et al. (2001) identified three distinct periods of greatly

increased speleothem growth rates in central Texas caves (71–60 ka, 39–33 ka, and 24–

12 ka) and concluded that increased growth rates were the result of increased

precipitation and/or effective moisture during these intervals. Toomey et al. (1993),

Blum and Valastro (1994), and Cooke et al. (2003) agree that a decrease in effective

moisture and increased aridity occurred around the same interval, ca. 14–12 ka. Most

explanations for increased effective moisture during the LGM include the southward deflection of the polar jet stream, which resulted in cooler conditions and less evaporation. These interpretations are consistent with model simulations of LGM climate as well (e.g. (Kutzbach and Guetter, 1986; COHMAP, 1988; Toomey et al., 1993;

Musgrove et al., 2001).

Older Alluvial Plains of the Trinity River System: The Beaumont Formation

Previous investigations into the Quaternary geology (Fig. 7) of the Texas Coastal

Plain identified alluvial deltaic plains and barriers associated with previous interglacial periods. Following Hayes and Kennedy (1903), Deussen (1914) formally named these deposits that were older than the most recent deposits the Beaumont formation. Barton

(1930a;1930b) identified the distributary nature of an ancestral Trinity River that extends updip and merges on both sides of the present alluvial valley. Numerous subsequent authors noted the alluvial ridges and the extensive alluvial–deltaic plain of an ancient

Trinity River system when compared to its modern counterpart, as well as a presence of a previous barrier system (Ingleside) that the river transected during progradation (Fig. 8;

Doering, 1935; Bernard, 1950; Bernard et al., 1962; Winker, 1982; Blum and Aslan,

18 Isotope Period Age Stratigraphic (ka) Stage Epoch (OIS) Unit

Holocene Recent 10 1 undifferrentiated

2 Low 3 Deweyville Middle ? High Late

y 4 r a

n 5

~132 ater Qu e

n Beaumont

te 6 –

a 9 or 10 stoce ei l P ~300 –400 ddle to L i M

~11– Lissie ddle

i 22 ? M ~780

Figure 7. Stratigraphic column of the Late Quaternary in East Texas and correlative oxygen isotope stages. Modified from Durbin (1999).

19 95°0'0"W 94 °5 0'0 "W 94 °4 0'0 "W 94°30'0"W 94°20'0 "W 94 °1 0'0 "W 94 °0 '0 "W 95°0'0"W 94°50'0"W 94°40'0"W 94°30'0"W 94°20'0"W 94°10'0"W 94°0'0"W 30°40'0"N . .

30°30'0"N

30°20'0"N

30°10'0"N

30°0'0"N

29°50'0"N

29°40'0"N

29°30'0"N 0 5 10 15 20 A 0 5 10 15 20 B Kilometers Kilometers Legend Figure 8. (A). Satellite image of Pleistocene Beaumont alluvial deltaic plains, post-Beaumont valleys www.zulu.gov. (B). Interpreted geologic map of Beaumont alluvial deltaic plains, depicting pre-Wisconsin shoreline relict channel courses, preserved barriers, and post–Beaumont valleys. Beaumont alluvial deltaic Modified from BEG (1982; 1992), USGS (2007), www.usgs.gov. Beaumont shoreface (Ingleside) . Beaumont pre-Beaumont undifferentiated

20 2006). Blum and Price (1998) obtained thermoluminescense (TL) ages of the Beaumont formation for the nearby Colorado River system showing that the representative time period of deposition spanned from OIS 10 or 9–OIS 5a, (ca. 400 to 300 ka–85 ka). These deposits are now considered to consist of multiple cross–cutting glacial–to–interglacial scale cycles of valley incision and filling (Blum and Price, 1998).

The Post–Beaumont Lower Trinity Valley

The lower Trinity River valley is an incised valley formed by incision and abandonment of Beaumont alluvial plains, and has been interpreted to have occurred during sea–level fall during the last glacial period (OIS 4–2). This incision was step– wise, and resulted in a downward–stepping flight of fluvial terraces that are lower than the Beaumont alluvial plain surface (Blum et al., 1995; Morton et al., 1996; Blum and

Törnqvist, 2000; Blum and Aslan, 2006).

“Deweyville” Deposits

Deussen (1914) first recognized terraced alluvial deposits along the valleys of major rivers within the Texas Gulf Coastal Plain. Barton (1930a, 1930b) identified terraces present along the San Jacinto River, where Oyster Creek flows along the Brazos

River Valley and within the Trinity and Brazos valleys, and that were inset against, and younger than, Beaumont alluvial deltaic plains, and were a record of a complex and long series of events. He also pointed out their occurrence in other valleys of east Texas like the Neches and Sabine. Bernard (1950) coined the term “Deweyville” for these units, based on terraces that flank the west bank of the Sabine River at Deweyville, Texas.

Bernard (1950) mapped portions of the Deweyville terraces along the Sabine, Neches,

21 and Trinity, noted their presence in the San Jacinto, and Nueces Rivers in Texas, as well as the Pearl River in Louisiana and Mississippi, and inferred the deposits were present in the subsurface along the Colorado and Brazos Rivers of Texas.

Gagliano and Thom (1967), Blum et al. (1995), Morton et al. (1996), Blum and

Aslan (2006), and Heinrch (2006) identified several anomalous distinctions within the

Deweyville units when compared to the modern river: 1) Deweyville surfaces exhibit steeper gradients than the modern river, and 2) Deweyville fluvial deposits are coarser than the modern river deposits and 3) the terraces were stepped and often unpaired.

Lastly, when compared to the modern river, Deweyville surfaces are characterized by larger wavelength meanders, radii of curvature, and bankfull widths (Barton, 1930a;

Matthes, 1941; Bernard, 1950; Gagliano and Thom, 1967; Alford and Holmes, 1985;

Blum et al. 1995; Durbin et al., 1997; Blum and Aslan, 2006; Heinrich, 2006; Sylvia and

Galloway, 2006). These observations led many of the authors mentioned above to suggest that the terraces were graded to a lowered sea level and there were discontinuities in the Deweyville sequence.

Blum et al. (1995) noted that the numerous studies of Deweyville landforms and deposits had become confusing, and it was not clear that all investigators were referring to the same deposits or features, noting that multiple terraces with characteristics normally attributable to Deweyville can be identified along all major river systems of the

Gulf Coastal Plain, with the exception of the Mississippi River. Moreover, terraces with

Deweyville characteristics can be mapped in each valley, and in many cases terrace surfaces are onlapped and buried by younger deposits, such that they are no longer terraces in the classic sense. Blum et al. (1995) suggested that Deweyville units should

22 be viewed as a succession of at least three unconformity bounded allostratigraphic units,

with each unit representing a period of valley incision followed by lateral migration and

aggradation, then renewed valley incision during the long and complex sea-level fall and

lowstand of the last glacial period.

Correlation to Offshore Record

During sea–level lowering, the Trinity and Sabine valleys merged on the

continental shelf to produce a single cross–shelf valley (Fig. 9; Thomas and Anderson,

1989; Thomas, 1990; Anderson et al., 2004). Thomas (1990), Thomas and Anderson

(1994), and Rodriguez et al. (2005) identified Deweyville deposits in Galveston Bay and

on the shelf from the Trinity, San Jacinto, Neches, and Sabine rivers using geophysical

methods and core data. The merging of the Trinity and San Jacinto Rivers in Galveston

Estuary increases the drainage basin size to approximately 60,000 km2, and the merging

of the Neches and Sabine with the Trinity and San Jacinto valley on the shelf enlarges the

overall drainage to approximately 90,000 km2 (Anderson et. al, 2004; Blum and Aslan,

2006; Mattheus et al., 2007). At lowstand, the Brazos River also merged with the system to provide an overall contributing drainage area of 208,000 km2 (Anderson et al., 2004).

The downslope basins for the Trinity River system have been identified using seismic investigations along with core analysis. Beauboeuf et al. (2003) termed these downslope minibasins the Brazos–Tinity slope system due to the merging of the Brazos and Trinity lowstand valleys on the shelf and the various types of deposits within the basin fill units (Fig. 9). Sedimentary fill was correlated between basins 1 through 4, but the only chronostratigraphic constraints on timing of basin fill has been documented in

Basin 4 (Mallorino et al., 2006).

23 95°0'0"W 94°0'0"W Legend Brazos-Trinity slope system

. 30°0'0"N Brazos-Trinity lowstand delta OIS 4-2 valleys Beaumont shoreface (Ingleside) Beaumont pre-Beaumont undifferentiated Bathymetry, (m) 29°0'0"N

0 0 -2 20 - 2 -

29°0'0"N

-20

-40

28°0'0"N -60 -80

-160 -200 -300 -120 -400 -400 -100 0 0 - 0 1 1 -6 0 - 0 40 -4 0 28°0'0"N - 00 0 6 -

-6 - - 0 6 0 8 0 00 0 -8 - 0 0 4 1 0 - - 6 03060 4 0 0 0 -16 -160 0 0 -800 0 0 -100 -6 0 Kilometers 00 -30 -2 95°0'0"W 94°0'0"W 93°0'0"W Figure 9. Geologic map of northwestern GOM showing the OIS 4–2 fluvial valleys as they extended towards the shelf edge and merged to deposit the Brazos–Trinity shelf edge delta and filled the Brazos–Trinity slope minibasins. Modified from Thomas (1990), Anderson et al. (2004), Mallorino et al. (2006), and USGS (2007), www.usgs.gov. Bathymetric data from the National Geophysical Data Center, www.ngdc.noaa.gov

24 Recent Deposits of the Trinity River

Although this thesis is focused on the nature and significance of the Deweyville

deposits, post–Deweyville deposits in the lower Trinity incised valley are characterized

by the following environments of deposition: barrier island, tidal, middle and lower bay,

upper bay and bayhead delta, and channel sands with laterally extensive floodplain

deposits (Thomas, 1990; Thomas and Anderson, 1994; Blum and Aslan, 2006). The

basinward portion of the Trinity River valley would be classified as a wave–dominated

estuary (sensu Boyd et al., 1992; Dalrymple et al., 1992).

Paleohydrology

Numerous authors have used the presence of the large Deweyville paleochannels

and their respective planform properties to interpret increased precipitation, runoff,

effective moisture, or any combination of these. As used in this thesis, Qm is the

3 -1 statistical daily mean discharge (m s ) for the period of record, Qmax is the mean flood

3 -1 discharge (m s ), defined as the mean for the annual maximum flood series, and Q2 is the flood discharge (m3s-1) with a recurrence interval of two years (Dunne and Leopold,

1978). The mean flood discharge typically has a recurrence interval of between 1–2

years, or a recurrence interval of 1.58 years is the most probable in many river systems

(Wolman and Leopold, 1956; Bridge, 2003), and is therefore similar to Q2: these measures are, in turn, similar to Qbf, defined as the bankfull discharge, which has a recurrence interval of ca. 1.6 years.

Many previous investigations have estimated mean annual discharge from channel planform properties. However, it is widely acknowledged that bankfull

25 discharge, also known as channel forming discharge, is the primary control on channel

shape and size (Wolman and Miller, 1959; Bridge, 2003). As noted by Rotnicki (1991),

mean annual discharge can only be estimated from the channel forming discharge and

retrodiction of low stages and discharges, as well as the magnitude of overbank flows are

impossible.

Sediment Supply Controls

Modern pre–dam sediment loads for the Trinity River have been measured at 5.5

MT/yr. To date, there have been no direct measurements of sediment supply for the

glacial period. However, a recent empirical model (BQART) published by Syvitski and

Milliman (2007) provides an opportunity to estimate glacial period sediment loads

(Womack, 2007; Blum and Womack, 2008–in press). Through analysis of a dataset consisting of 488 modern rivers that discharge to coastal oceans, they derived an empirical equation that accounts for 96% of between river variance in sediment loads.

The model includes five different variables that explain the variance, with four of these

important for this study: drainage area, relief, lithology, and climate.

During the high frequency oscillations of sea–level change during the Late

Quaternary, major changes in the drainage basin area, relief, and lithology are not likely

to occur due to the longer temporal scales of landscape evolution. Variables within the

model that have no bearing on glacial period sediment loads within the Trinity River

system include anthropogenic effects and glaciation. Drainage area, relief, and lithology

are the most important variables, but do not change over the time scales of interest for

this study (100–105 years). Climate is a second order control on sediment supply when

compared to drainage area, relief, and lithology, but does change over time scales of

26 interest here (Syvitski et al., 2003; Syvitski and Milliman, 2007). In the Syvitski and

Milliman (2007) model, sediment supply (qs) increases with increases of temperature and water discharge (Q).

Stored and Excavated Sediment Volumes

Sediment bypass and delivery to downslope depocenters during base–level lowering has long been thought to be a process that accounts for high sediment volumes sequestered in shelf phase/margin deltas, slope minibasins and deep water deposits (see review in Blum and Törnqvist, 2000). Recent research by Blum and Törnqvist (2000) has shown that sediment volumes produced by valley incision are approximately an order of magnitude less than inherited background rates from the contributing drainage basin.

However, this issue has not been examined in great detail, and recent research continues to imply that excavated sediment volumes account for large portions of the sediment supplied to shelf deltas and deepwater systems (Anderson et al., 2004).

27 METHODS

The goals of this thesis revolve around establishment of a more detailed

stratigraphic, sedimentologic, and geochronologic framework for Deweyville units,

followed by examination of paleohydrologic implications, and the role of valley incision

in sediment mass balance (storage and export).

Stratigraphic and Sedimentologic Framework

Deweyville units are again treated as unconformity–bounded stratigraphic units,

following Blum et al. (1995), and are mapped and described in greater detail, and over a

larger spatial extent than in previous studies. Revised mapping of the lower Trinity River

incised valley was accomplished using satellite imagery, DEM data, and data collected in

the field and plotted into ArcGIS software to aid in preparing georeferenced geologic

maps.

Deweyville and post–Deweyville units are well exposed along active river cutbanks, which provided the means to evaluate facies typical of each valley–fill component. Facies were described and interpreted using nomenclature from Bridge

(2003), and color was described using the Geological Society of America Rock Color

Chart (1991). Post–Deweyville fluvial deposits were studied from available outcrops along cutbanks, modern point bars, and a single core through a recent point bar.

Sediment cores were obtained from the bayhead delta region as well, using a vibracore capable of extracting continuous and discontinuous 3” diameter samples up to ca. 9 m in depth (Fig. 10), which was transported to coring locations by boat. Strike–and dip–oriented coring transects were located to obtain an accurate depiction of the valley fill stratigraphic sequence.

28 B

A

Figure 10. Collection of cores in the field. (A). Vibrating core into the ground, (B). Core retrieval with the help of a winch.

29 Chronologic Framework

Samples were collected for optically stimulated luminescence (OSL) at various

Deweyville outcrops (Fig. 11). OSL samples were pre–processed in a laboratory at LSU,

then shipped to the Risø National Laboratoryand Nordic Laboratory for Luminescence

Dating, located ate the University of Aarhus in Denmark. Preprocessing followed the

steps outlined in Stokes et al. (2003) and by the Risø National Laboratory (Professor

Andrew Murray, personal communication, 2006). Samples were weighed wet, then dry

(ca. 300 g) to determine field water saturation after drying overnight at 40° C overnight.

Carbonate present in the samples were removed with 10% HCl, then wet sieved to a grain

size fraction of 180–250 µm. Samples were dried again at 40° C overnight prior to

shipping to the Risø Lab for single–aliquot–regenerative–dose (SAR) protocol, as

outlined in Wintle and Murray (2000) and Murray and Wintle (2003).

Paleohydrology

Morphological characteristics of the modern river channel and its Late Quaternary

counterpart were measured from topographic maps, satellite images, and DEM data using

ENVI and ArcGIS software. Parameters measured include meander wavelength, Lm, radius of curvature, rc, and bankfull width, Wb, following procedures outlined in Brice

(1974), Brice (1984), Alford and Holmes (1985), and Sylvia and Galloway (2006) (Fig.

12). Table 1 displays the equations tested over the course of this study and the sources

from which they were acquired. Direct retrodiction of mean annual discharge was

impossible, however, mean annual discharge estimates are included to test the

applicability of various equations derived using different variables. Results from various

30

Figure 11. Collection of optically stimulated luminescence samples in the field. Sample is collected within one cross–bed set and where preserved sedimentary structures can be observed.

31

Figure 12. Measurements of channel planform properties. (A). Satellite image depicting the measurements of meander wavelength, Lm, (B). Satellite image depicting the measurement of radius of curvature, rc.

32

Figure 12. (C). Satellite image and, (D). Digital elevation model depicting the measurement of bankfull width, Wb.

33 Table 1. Planform properties and empirical equations used to assess paleohydrological estimates and estimates of modern discharge.

Planform Equation Reference Equation measurement number 2.03 rc Qm= 0.0021rc Alford and 1 (General Equation) Holmes (1985) 1.58 rc Qm= 0.025rc (General Equation) Williams (1984) 2 2.15 Lm Qm= 0.000047Lm Alford and 3 (General Equation) Holmes (1985) 1.66 Wb Qm= 0.06Wb (General Equation) Williams (1984) 4 1.71 Wb Qm= 0.027Wb Osterkamp and 5 (General Equation) Hedman (1982) 2.12 Wb Qm= 0.031Wb Osterkamp and 6 (for rivers with a high silt clay bed) Hedman (1982) 1.76 Wb Qm= 0.033Wb Osterkamp and 7 (for rivers with medium silt clay bed) Hedman (1982) 1.62 Wb Qm= 0.029Wb Osterkamp and 8 (for rivers with sand bed and banks) Hedman (1982) 1.22 Wb Q2= 1.9Wb Osterkamp and 9 (General Equation) Hedman (1982) 1.32 Wb Q2= 0.96Wb Osterkamp and 10 (for rivers with sand bed and banks) Hedman (1982) 1.86 Wb Q2= 2.0Wb Osterkamp and 11 (for rivers with high silt clay bed) Hedman (1982) 1.27 Wb Q2= 2.6Wb Osterkamp and 12 (for rivers with medium silt clay bed) Hedman (1982) 1.38 rc Qmax= 0.28rc (General Equation) Williams (1984) 13 1.16 Wb Qmax= 1.0Wb (General Equation) Williams (1984) 14

34 equations were tested against actual values in the modern system to determine clearly

which equations are the most appropriate for this study.

Calculation of Sediment Supply

The BQART model from Syvitski and Milliman (2007) was used to estimate

sediment supply during the last glacio–eustatic cycle. Variables required by the BQART

model, drainage basin area, relief, basin–averaged temperature, and mean annual

discharge, were obtained from published reports. Results pertain to sediment discharges

for the lower Trinity River, and do not include channel extension on the shelf or merging

with other systems that ultimately deliver sediment to point sources on the shelf and in

the deep basin.

Calculation of Stored and Excavated Sediment Volumes and Mass

Volumes of sediment stored in each stratigraphic unit were calculated by first defining the surface area of each mappable unit with ArcGIS and thickness measurements

from measured sections. In locations where entire unit surface areas are not preserved

due to cannibalization by younger units, extrapolations were drawn within the valley

margins by connecting either paired or unpaired terraces to give reasonable first order

estimates. Volume of sediment was then converted to mass by assuming a density of 2.7

T / m3, and then multiplied by 0.6, which corrects for a porosity value of 40%.

Volumes of sediment exported by each period of valley incision and lateral

migration were calculated by using the mapped surface areas of stratigraphic units, and

the difference in elevations between their correlative depositional surface and the

depositional surface of the previous older unit (Figure 13). This approach assumes that

35

Figure 13. Cartoon depicting the removal and export of sediment to downdip reaches. (A). Time 1: Schematic of typical valley cross section, showing that during incision and lateral migration of the oldest unit, there is a net removal of sediment (denoted by the red x’s) that is removed over the length of Time 1. (B). Time 2: Schematic illustrating that during a subsequent period of renewed incision followed by lateral migration results in a net removal of sediment (denoted by red x’s) that is removed over the duration of Time 2. (C). Time 3: Schematic of another period of incision and lateral migration. Shaded regions in B and C represent portions of the valley that were previously removed.

36 each period of incision is followed by lateral migration of the channel and channelbelt, and leaves behind a deposit: the net volume of sediment excavated is that which is completely removed and not stored within the newer and younger unit. Volumes were converted to mass, and rates of sediment export were calculated using available geochronological data.

37 RESULTS

Results are reported below here in terms of project goals. The revised

stratigraphic, sedimentologic, and geochronologic framework for Deweyville deposits are

discussed first, followed by the stratigraphic and sedimentologic features of post–

Deweyville units. Results of paleohydrological analyses are then discussed, followed by

estimates of sediment supply and sediment mass balance calculations.

Deweyville Deposits: General Framework

This research makes slight modifications to previously mapped areas, extends the

existing geologic mapping and stratigraphic framework upstream to the Lake Livingston

dam, and thoroughly describes the deposits. In general, as noted by previous workers,

and confirmed in this study, all Deweyville allostratigraphic units examined consist of

fining upward, sand–dominated successions, with local lenticular deposits of silt and

clay. Deweyville surfaces exhibit ridge and swale scroll topography that can be readily identified on terrace surfaces in satellite imagery and digital elevation models. Hence, the Deweyville units are interpreted to represent channelbelts dominated by sandy point bar and muddy channel fill successions, and each channelbelt, or Deweyville unit, represents lateral migration of the river during its respective time period of deposition.

Figure 14b presents a revised map of the Trinity valley where three Deweyville units can be mapped and correlated through the study area: Blum et al. (1995) used the terms High, Middle, and Low Deweyville units to refer to their topographic relationships.

The High Deweyville channelbelt is inset into and against the older Beaumont deposits, the Middle Deweyville channelbelt inset into both High Deweyville and Beaumont deposits, and the Low Deweyville is inset and against the Beaumont, and High, and

38

95°0'0"W 94°50'0"W . 30°40'0"N

30°30'0"N

30°30'0"N

30°20'0"N

30°20'0"N

30°10'0"N

30°10'0"N

30°0'0"N

30°0'0"N

29°50'0"N

29°50'0"N

036912 Kilometers 94°50'0"W 94°40'0"W

Figure 14. (A). Satellite image of the Coastal Plain incised valley. Imagery from www.universityofmaryland.edu

39 95°0'0"W 94°50'0"W Legend 30°40'0"N . post-Deweyville Alluvium post-Deweyville Channelbelt Low Deweyville Terrace Veneer Cedar_Valley_site Middle Deweyville Terrace 30°30'0"N High Deweyville Terrace A Beaumont A pre-Beaumont undifferentiated A’ 30°30'0"N A B LD_Romayor_North LD_Cypress_Lakes _North

30°20'0"N

B

30°20'0"N MD_Romayor_RR

MD_Moss_Hill 30°10'0"N

30°10'0"N C HD_Sandune C LD_Kenefick

D B B’ MD_Kenefick 30°0'0"N

D

30°0'0"N LD_Liberty MD_Liberty

HD_Liberty

29°50'0"N C C’

29°50'0"N LD_Port_of_Liberty _South 036912 Kilometers 94°50'0"W 94°40'0"W Figure 14. (cont.) (B). Detailed geologic map of the coastal plain with OSL sample locations and overall valley–fill cross section locations. Modified from BEG, (1992), Blum et al., (1995), Morton et al., (1996).

40 Middle Deweyville deposits. Each Deweyville unit consists of a basal erosion surface which lies on older deposits and bound above by a well–developed paleosol. In downdip reaches, the paleosol developed in Low Deweyville deposits is buried by a wedge of post–Deweyville strata due to recent sea–level rise (Blum and Törnqvist, 2000; Blum and

Aslan, 2006). Farther updip, post–Deweyville deposits (floodplain) are relatively thin

(<1 m), and the constructional ridge and swale topography of the Low Deweyville surface is easily distinguished on satellite imagery (Fig. 14A). Figure 14B shows the

Low Deweyville map unit in this part of the valley as a terrace veneer. As was the case in previous investigations, appreciable thicknesses of continuous fine–grained topstratum have not been identified within the Deweyville deposits, and paleosols are generally developed on sandy point bar or muddy channel–fill deposits.

High Deweyville (HD) Unit

High Deweyville deposits are exposed where the modern channel intersects terraces and depositional remnants within the valley fill, and in sand quarries. They are capped by paleosols, ranging from ca. 60–150 cm thick, with a dark yellowish brown

(10YR 4/2) colored A horizon and underlain by a light brown (5YR 5/6) to pale yellowish brown (10YR 6/2) E horizon. Below the E horizon, Bt horizons consist of moderate reddish–brown (10R 4/6) fine– to very fine–grained sandy clay with subangular to angular blocky pedogenic structures. Oxidized lamellae characterize the transition to the Cox horizon and unweathered parent fluvial deposits (Fig. 15A). The High

Deweyville terrace surface dips below the modern delta plain surface in Trinity Bay.

Thickness of the High Deweyville unit ranges from ca. 5.5 m updip to ca. 11 m downdip. All High Deweyville successions fine upwards, which is typical of point–bar

41 A

A E Bt Cox B Cu C 10 m D y n b b ul

vf mvca a Lithology ilt fc cl s gr pe co bo

Figure 15. (A) Measured section where OSL sample “HD_Liberty” was collected. Ten meter thick sandy point bar. Basal erosion surface is just below water level. (Flow direction is from left to right), (B) upper bar with typical cross bar channel deposits ~15–39 cm thick., (C) trough cross beds 11–20 cm thick., (D) trough cross beds 14–20cm thick containing large clay rip–up clasts.

42 B D

C

Figure 15 (cont.)

43 deposits described from the literature (Bridge, 2003). Lower bar deposits range from

medium to very coarse sand and sometimes pebble sized grains (D50 = 0.25 to 2 mm,

sometimes 16 mm, where D50 corresponds to the median grain size on a cumulative weight percent curve), and are dominated by trough cross–beds up to 30 cm thick with occasional clay rip–up clasts, whereas upper bar deposits range from fine to medium sand

(D50 = 0.125 to 2 mm), and consist of ripple cross–laminations and planar beds from 4 to

13 cm thick (Figs. 15A–D). Large–scale trough cross beds up to 39 cm thick are common

in the upper part of many High Deweyville exposures. Following Bridge (2003), the

beds are interpreted to represent cross–channel bar deposits (Figs. 15B).

Middle Deweyville (MD) Unit

Middle Deweyville deposits are exposed along the modern river. Terrace surfaces

are characterized by paleosols ranging from ca. 160–240 cm thick and consist of dark

yellowish brown (10YR 4/2) A horizons, ranging from ca. 40–80 cm thick, and light

brown (5YR 5/6) to pale yellowish brown (10YR 6/2) E horizons, ca. 30 cm thick (Figs.

16A and 17). The A and E horizons overlie well developed dark yellowish orange (10YR

6/6) and moderate reddish–brown (10YR 4/6) fine– to very fine–grained sandy clay Bt

horizons ranging from 60–150 cm thick, with subangular to angular blocky pedogenic

structures, which are underlain by a transition to very pale orange (10YR 8/2) grading

into pale yellowish orange (10YR 8/6) Cox horizons that also contain light grey (N7) to

very light grey (N8) root mottles. Below the soil profiles the deposits consists of

unweathered parent fluvial deposits. The Middle Deweyville terrace surface dips below

the modern delta plain surface just below where I–10 crosses the valley (cross section C–

C’), at the updip limits of the modern bayhead delta.

44 A

A E

Bt B Cox Cu

C 5.8m

D l n y b b u a t vf mvc l Lithology a si pe co cl fc gr bo

Figure 16. (A) Sandy point bar covered by a thin A and E horizons, thick Bt horizon with subangular to angular peds. Cox horizon exhibits diminishing coloration with abundant lamellae. OSL sample MD_Romayor_railroad was collected in fine to medium grained cross–bedded sands in the upper part of the bar deposits. The lower bar contains trough cross bedded sands and gravels (A, D). Blue lettering marks location of pictures for the following figures. (B) ripple cross–laminated sands and small scale trough cross–beds 3–5cm thick. (C) Upper bar deposits where large scale cross–beds (ca. 30 cm) typical of cross–channel chute bar facies, (D) basal erosional unconformity of Middle Deweyville unit on top of pre– Deweyville channel and channel–fill sands and muds and fine–grained topstratum.

45 B D

C Basal erosion surface, (unconformity)

Figure 16 (cont.)

46 A

-1 m B Bt -2 m

Co x

-3 m C Cu

-4 m

-5 m

D

A B b y

t vf mvc l an Lithology a si peb co cl fc gr boul Figure 17. (A) Middle Deweyville outcrop sandy point bar and channel fill mud and sand where OSL sample MD_Kenefick was collected. Grain size ranges from very coarse sand with some pebble–sized gravels fining upwards to alternating layers silt and clay and very fine–grained ripple laminated sand and silt interpreted as channel–fill deposits., (B) channel fill silts and very fine– grained sand interbedded with clay, overlain by well developed Bt horizon and a ca. 80 cm thick A horizon,

47 C

D

Figure 17 (cont.), (C) channel fill mud and sand containing ripple–laminated silts and very fine–grained sand, (D) 30 cm thick cross–bedded very coarse sand with some gravel grading upwards into 3–10 cm thick cross–bedded sands with alternating with ripple cross–laminated silts and sands.

48 Middle Deweyville deposit thicknesses range from ca. 6 m updip to ca. 9–11 m

downdip. All Middle Deweyville successions fine upwards, which is typical of point–bar

deposits described from the literature (Bridge, 2003). Lower bar deposits observed in

this study range from coarse to very coarse sand (D50 = 0.5 to 2 mm), and are dominated by trough cross–beds 17 to 45 cm thick and contain occasional clay rip–up clasts (Fig.

16D). In outcrops where sandy channel bar deposits are continuous to the top of the section, upper bar deposits range from fine to coarse sand (D50 = 0.125 to 1 mm), and

consist of trough cross–beds and ripple cross–laminations, and planar beds ranging from

ca. 3 to 15 cm thick (Fig. 16B). Common features in upper bar deposits are large scale

cross strata ranging from ca. 25–65 cm thick (Fig. 16C). Following Bridge (2003), the

beds are interpreted to represent cross–channel bar deposits. Occasionally upper channel deposits in outcrops are comprised of medium gray (N5) to medium dark grey (N4) mud layers 4–20 cm thick containing no sedimentary structures, light grey (N7) silt beds and moderate yellow (5Y 7/6) very fine–grained ripple laminated sand and silt beds ca. 7–8 cm thick (Fig 17A–D). Where these characteristic features are observed, following

Bridge (2003), they are interpreted as channel–fill deposits. Common features within distinct sandy and silty beds are dark yellowish orange (10YR 6/6) mottles and dark grey

(N3) clay nodules.

Low Deweyville (LD) Unit

The Low Deweyville unit is buried by a wedge of younger sediments in the downdip portions of the valley, whereas farther updip, terrace surfaces are characterized by ridge and swale topography capped by a thin veneer (<1–2.5 m thick) of overlying post–Deweyville floodplain deposits (Fig. 14A). The Low Deweyville unit is widely

49 exposed along active cut banks because modern channels flow through this older course; but, entire thicknesses are only observable in far updip reaches because the surface of this unit dips below the modern floodplain ca. 65 km from the modern bayhead delta, and ca.115 km from the modern shoreline along the valley axis (refer to figure 28). The farthest downdip exposures occur near Moss Bluff. Soil profiles below the onlap point consist of slightly thicker (ca. 40–50 cm thick) dark yellowish brown (10YR 4/2) to moderate yellowish brown (10YR 5/4) A horizons underlain by ca. 100–200 cm thick moderate yellowish brown (10YR 5/4) to moderate reddish brown (10YR 4/6) fine to very fine–grained sandy clay Bt horizons (Figs. 18–20). Evidence of gleying below soil profiles occurs commonly below channel fill muds (Fig. 18F). Where there is an appreciative thickness of overlying post–Deweyville sediments, nodules of secondary carbonate and carbonate cements occur in thin (ca. 14 cm thick) white (N9) to very pale orange (10YR 8/2) Btk horizons (Fig. 18E). Terrace surfaces are interpreted to be characterized by vertisols or cumulic soil profiles where the Low Deweyville unit is presently buried by recent sediments. Between the more recent sediments and the top of the Low Deweyville paleosol, there is an unconformity which reflects a significant period of little to no deposition and soil development.

Thickness of the Low Deweyville unit ranges from ca. 7 m updip to ca. 11–12 m downdip. All Low Deweyville successions fine upwards, which is typical of point–bar deposits described from the literature (Bridge, 2003). Lower bar deposits range from medium to pebble sized grains (D50 = 0.25 to 4mm) and are dominated by trough cross– beds, ranging from ca. 8–50 cm thick, whereas upper bar deposits range from fine to coarse sand (D50 = 0.25 to 1 mm) and consist of trough cross–beds ca. 4–14 cm thick

50 B

B A 0 m C A il o

S Bt e v i t

D la u m

E Cu Btk Cg Cox F Cu C G 5 m b y vf mvc b t ul l an A Lithology a si fc pe co cl gr bo

Figure 18. (A) Low Deweyville outcrop where OSL sample LD_Kenefick was collected. Sandy point bar grading upwards into channel fill mud and sand overlain by post–Deweyville clays and silts coarsening upwards into silts and very fine–grained sands interpreted as post–Deweyville floodplain deposits coarsening upward into modern levee deposits. (B) modern levee containing ripple laminated silts and very fine–grained sand interbedded with clay., (C) post–Deweyville distal levee and floodplain muds and silts.

51 D E

F G

Figure 18 (cont.), (D) soil profile developed in Low Deweyville channel fill muds and sands., (E) carbonate precipitation due to burial of soil profile and ground water reflux, commonly found in Low Deweyville buried terraces., (F) dark grey root mottles developed in channel sands consisting of trough cross beds comprised of medium grained sand, cross–beds 10–25 cm thick, (G) cross–beds 11–50cm thick in fining upwards to very coarse–grained sand and pebble gravels.

52 A Modern levee B

LD paleosol

Post–Deweyville slope wash slope wash C

Figure 19. (A) Site where OSL sample LD_Liberty_South was collected. Post–Deweyville channel fill muds and fine sands lying unconformably on Low Deweyville channel sands. (B) coarsening upward modern floodplain and levee deposits resting unconformably on top of Low Deweyville paleosol., (C) photograph of OSL sampling in medium–grained cross–bedded sands.

53

Figure 20. Cedar Valley site located downstream of the Lake Livingston Dam. Basal erosional surface of the Low Deweyville unit is just below water level. Post–Deweyville channelbelt sands and overlying laterally extensive floodplain muds, which coarsen upwards into levee deposits, which unconformably overly Low Deweyville channel sands and channel–fill sands and muds.

54 along with ripple cross–laminations and planar beds ranging from ca. 5–18 cm thick

(Figs. 17–20). Upper channel deposits in outcrops also consist of medium grey (N5) to medium dark grey (N4) mud layers 4–20 cm thick containing no sedimentary structures, light grey (N7) silt beds and moderate yellow (5Y 7/6) very fine–grained ripple laminated sand and silt beds ca. 7–8 cm thick (Fig. 18A, D–F). Where these characteristic features are observed, following Bridge (2003), they are interpreted as channel–fill deposits.

Common features within the channel–fill deposits are dark yellowish orange (10YR 6/6) mottles and dark grey (N3) clay nodules.

Deweyville Geochronology

Deweyville allostratigraphic units have proven difficult to date, but have been interpreted by previous workers to have been deposited during OIS 4–2 (Blum et al.,

1995; Morton et al., 1996; Blum and Törnqvist, 2000; Blum and Aslan, 2006). Of the 11 samples analyzed, 8 yielded age estimates that agree with cross–cutting relationships, whereas 3 yielded dates that violate cross–cutting relationships. The age ranges reported below do not include the 3 samples awaiting additional analysis but are included in Table

2 and Figure 21. One sample taken from the High Deweyville unit yielded an age representing the timing of channelbelt activity from 35–31 ka. Four samples from the

Middle Deweyville unit yielded ages ranges from 34–23 ka. Three samples from the

Low Deweyville unit yielded age ranges from 23.2–18.8 ka.

Post–Deweyville Deposits

Post–Deweyville units are not the focus of this study, but display a markedly different surface morphology with distinctly different facies when compared to

Deweyville units. Post–Deweyville environments of deposition and deposits consist of

55 Table 2. Optically stimulated luminescence results. HD = High Deweyville unit, MD = Middle Deweyville unit, LD = Low Deweyville unit. The age estimates of samples HD_Sandune, LD_Cypress_Lakes_North, and LD_Kenefick violate cross–cutting relationships and are shown in italics whereas samples in bold are considered robust.

Sample Name Latitude Longitude Burial H2O Dose Rate Dc (Gy) Number OSL Age (N) (W) depth Content (Gy/ka) of Estimate (cm) (%) Aliquots (ka) HD_Liberty 30°3’18.4” 94°49’08.2” 213 15 0.77 ± 0.05 25..8 ± 0.7 27 33 ± 2 HD_Sandune 30°7’13.1” 94°46’23.4” 119 5 1.56 ± 0.09 20.1 ± 0.4 27 12.9 ± 0.8 MD_Kenefick 30°06’24.0” 94°48’52.9” 307 15 1.05 ± 0.05 33.6 ± 1.0 25 32 ± 2.0 MD_Romayor_railroad 30°27’3.3” 94°50’54.2” 266 3 0.74 ± 0.05 20.6 ± 0.6 20 28 ± 2 MD_Moss_Hill 30°16’17.5” 94°47’13.3” 376 2 0.69 ± 0.04 17.7 ± 0.3 27 25.8 ± 1.8 MD_Liberty 30°03’20.1” 94°49’06.2” 249 4 0.80 ± 0.05 19.8 ± 0.4 27 24.8 ± 1.8 LD_Romayor_North 30°28’24.9” 94°52’27.4” 232 1 1.17 ± 0.06 25.4 ± 0.7 20 21.8 ± 1.4 LD_Liberty_South 30°02’51.4” 94°49’59.6” 153 18 0.68 ± 0.04 14.3 ± 0.2 27 20.9 ± 1.4 LD_Port_of_Liberty_South 30°01’37.2” 94°49’39.4” 135 10 0.88 ± 0.05 17.9 ± 0.7 26 20.3 ± 1.5 LD_Kenefick 30°06’47.1” 94°48’53.5” 141 4 0.6 ± 0.04 29.2 ± 2.0 25 49 ± 5 LD_Cypress_Lakes_North 30°19’31.7” 94°47’57.4” 202 3 0.67 ± 0.04 20.4 ± 0.7 17 30 ± 2

56 Age (ka) 10 15 20 25 30 35 40 45 50 55

HD_Liberty

HD_Sandune

MD_Kenefick

MD_Romayor _railroad MD_Moss_Hill

MD_Liberty

LD_ Kenefick mple names LD_ Cypress_Lakes Sa _North LD_Romayor _North LD_Liberty_South

LD_Port_of_Liberty _South

Figure 21. Graphical representation of OSL dates with error bars. Hollow symbols represent the dates in italics from Table 1.

57 channlbelt sand and mud in updip reaches of the study area, which grade downdip to bayhead delta sand and mud, central basin estuarine mud and shell hash, and barrier island sand. In addition, overbank floodplain strata cover Low Deweyville units, with thicknesses decreasing in the upstream direction: in Trinity Bay, the mean thickness of post–Deweyville sediments is ca. 8 m whereas ca. 100 km farther updip, post–

Deweyville terrace veneers taper to <1 m in thickness, but up to 2.5 m, with thickness greatest proximal to the modern channel. All post Deweyville units are thought to be much younger than the Low Deweyville unit because of the degree of soil development on the Low Deweyville surface, and are hereafter interpreted to be Holocene in age.

Channelbelt and Floodplain

Post–Deweyville channelbelts can be mapped by using DEM’s and satellite imagery (Fig. 22). Two distinct channelbelts were differentiated from cross–cutting relationships. Both channelbelts display widths ranging from 1.5–2.5 km and contain numerous individual channel–neck cutoffs (oxbows), and, in many instances, modern creeks flow through these older channel courses.

The modern Trinity River channel exhibits three different plan–view morphologies. From Lake Livingston dam to Romayor, the modern channel is currently confined within the relatively narrow Low Deweyville channelbelt, appears to be flowing through Low Deweyville paleochannels, and the channel base is actively incising into older undifferentiated indurated sands and muds: it seems likely this condition is a result of clear–water erosion due to the construction of Lake Livingston Dam (Fig. 20: see

Phillips et al. 2004 for discussion). In this sense, the channel is not strictly alluvial in its

58 94°50'0"W

Legend . Post-Deweyville Alluvium Y 30°20'0"N Younger Post-Deweyville Channelbelt Older Post-Deweyville Channelbelt Low Deweyville Terrace Veneer Middle Deweyville Terrace High Deweyville Terrace Beaumont Pre-Beaumont Undifferentiated

30°10'0"N

Figure 23

Figure 18

Figure 19

30°0'0"N

0246

29°50'0"N Kilometers

Figure 22. Geologic map showing the locations of post–Deweyville channelbelts and figure locations 23, 18, and 19. As mentioned in text, the channelbelt outlined in grey is interpreted to be older than the channelbelt outlined in green based on cross cutting relationships. Post–Deweyville alluvium refers to the onlapping wedge of sediment sequestered due to post–glacial sea–level rise.

59 origin, as its morphology is controlled by the lithologies of the sediments through which the channel is moving.

Between Romayor and Moss Bluff, the modern channel is highly sinuous with many cutoffs and abandoned channels. The channel mostly flows through, and remains confined within, the Low Deweyville channelbelt, although thick fine–grained post–

Deweyville floodplain deposits are ubiquitous (Figs. 18B, C, 19A, B). Along some cutbanks, the modern channel is migrating laterally into the Middle and High Deweyville units. Median grain size of the modern channel bed ranges from fine to medium sand

(0.125–0.5 mm), but is greater on point bars that occur directly downstream from cut banks that expose coarser Deweyville deposits. Moreover, a significant portion of the fine to medium–sized sand fraction actually consists of sand sized aggregates of clay.

Figure 23 displays photographs of a core taken in a modern point bar just upstream of

Liberty, Texas. Modern lower point bars consist of medium– to fine–grained trough cross–bedded sand, which fines towards the upper point bar and downstream. Upper bar surfaces consist of very fine–grained sand to silt and are topped by laterally continuous fine–grained sediments interpreted as post–Deweyville floodplain deposits (Figs. 22A–C,

23A, B, 24). Outside of the modern channelbelt, floodplain strata rest unconformably on older Deweyville surfaces, and typically consist of fine–grained clay and silt, that coarsens upwards to ripple cross–laminated silt and fine–grained sand and is interpreted as transitioning from distal floodplain to proximal channel–levee deposition (Figs. 18–

20).

60

Figure 23. Photos of modern point bar core. Fine– to medium–grained cross–stratified sand and silt fining upwards to ripple cross–laminated sand, silts and clays. Darker areas are comprised of sand–sized aggregates of clay as well as silt and clay, whereas lighter areas are comprised predominantly of sand. See figure 22 for core location.

61 94°40'0"W . MB_38 E MB_34 MB_35 MB_36 MB_37 C’

29°50'0"N 29°50'0"N MB_31 C

BD 10 Cotton lake 1 8.06 CL1 < BD 11 D < << < Cotton Lake 2

012345 Kilometers 94°40'0"W

Figure 24. Core locations across I–10 (C–C’), the modern bayhead delta, and downdip cross section locations.

62 Downstream from Moss Bluff, the modern river channel is less sinuous, and flows

primarily through fine–grained Holocene flood plain and delta plain muds, although

locally, the river still flows through Deweyville cut–off channels and paleomeanders (see

Aslan and Blum, 1999). Just downstream of I–10, the river flows into Trinity Bay and becomes distributary with many crevasse channels.

Bayhead Delta and Estuary

The modern Trinity River discharges into Trinity Bay creating a bayhead delta.

Fourteen cores were collected from the bayhead delta region (Fig. 24): seven distinct

facies were identified in cores, and occur in predictable successions that can be linked to

specific depositional environments (Figs. 25–27). These are (a) massive to weakly

laminated medium dark grey (N4) clays, with shell hash comprised mostly by Rangia with occasional oyster shells, interpreted as central basin estuarine (CBE) environments of deposition (Fig. 25); (b) shell–rich (Rangia) laminated silts interbedded with dusky yellowish brown (10YR 2/2) to brownish black clays (5YR 2/1), which coarsen upwards to shell–rich laminated fine sands and silts, and which are interpreted as the transition from prodelta to prograding delta front, (PD–DF) depositional environments, (Fig. 26), which are overlain by; (c) sharp–based fine– to medium–grained sands with shell hash, which are interpreted to represent distributary mouth bar (DMB) depositional environments (Fig. 26), which are overlain by; (d) ripple–laminated fine sands and silts, interpreted as aggradational middle ground bar (MGB) depositional environments (Figs.

26, 27); (e) alternating laminated silts and fine– to very fine–grained sand to muds, containing occasional shell fragments and wood, which are interpreted as channel facies

(CH) and, which are often overlain by ripple laminated silts and muds, interpreted as

63 11.01.06 BD1 BD 8 BD 10 Concentrated shell material comprised primarily of Rangia and msl oyster Modern marsh surface with laminated muds and organics, slightly oxidized at surface Outer bay gray clays -6.6 mbpsl Flooding surface

-6.58 mbpsl

Marsh surface with laminated muds and organics

Figure 25. Core photos illustrating older marsh surface and overlying central basin estuarine muds containing oyster and Rangia shells (11.01.06 BD1 and BD8). Core photo of modern marsh surface (BD 10).

64

Figure 26. Core photos of 9.30 BD1 showing prodelta muds coarsening upwards to bayhead delta sands typical of progradational deltaic deposits. Sharp contact of coarser sand overlying prodelta deposits is interpreted as distributary mouth bar deposits, overlain by ripple laminated sands, silts and clays interpreted as middle ground bar deposits. Base of distributary mouth bar sands would commonly be referred to as the bayhead delta diastem. See figure 30 for core location.

65

Figure 27. Core photos of BD 12. Moving vertically from base of core, prodelta clays and silts, coarsening upward to channel deposits containing laminated sands, silts, and muds, alternating between channel silts, sands, and muds. Top of core represents channel levee and aggradational marsh surface. See figure 30 for core location.

66 natural channel levee deposits (NL) (Fig. 27); (f) medium grey (N5) to brownish gray

(5YR 4/1) laminated muds, silts, containing a large percentage of laminated organics, and

commonly containing numerous burrows, interpreted as aggradational marsh (M)

depositional environments (Fig. 25), and; (g) brownish gray (5YR 4/1) clay, with plant

burrows and organics and occasionally silty streaks grading into levee sand in proximal

channel locations, interpreted as interdistributary bay (IB) depositional environments

(Fig. 25).

Valley Slopes, Longitudinal Profiles, and Channel Slopes

The mapping and stratigraphic framework above permitted construction of flood

plain longitudinal profiles, channel sinuosities, and channel slopes for post–Beaumont

stratigraphic units. Flood plain long profiles (Sv), channel sinuosities (P), and channel

slopes (Sc) for the High, Middle, and Low Deweyville units, and the modern flood plain

are reported in Table 3. The updip Beaumont long profile was constructed from

mappable units surrounding the Trinity valley, whereas downdip the long profile extends

to the inferred shoreline via the alluvial–deltaic plains that crossed the Ingleside barrier

system (Fig. 8). The Deweyville units and modern floodplain long profiles were

constructed along the valley axis both onshore and offshore (Fig. 28). The reconstructed

longitudinal profiles show that the High, Middle, and Low Deweyville terraces each

maintained steeper valley gradients compared to the Beaumont or modern floodplain

(Fig. 29). Reconstruction of channel slopes (Sc) for the Deweyville units were based

upon reconstruction of channel sinuosities (P) from relict channels and channel courses,

where sinuosity (P) is defined as channel length (Lc) divided by valley length (Lv).

Calculated channel slopes (Sc) are defined as valley slope (Sv) divided by sinuosity (P).

67 95°0'0"W 94°50'0"W

30°40'0"N Legend . Incised valley axis Elevation, (m) High : 70 30°30'0"N 50 1 Low : 0

30°30'0"N

30°20'0"N

30°20'0"N

30°10'0"N 100

30°10'0"N

30°0'0"N

30°0'0"N

29°50'0"N

50 29°50'0"N

036912 Kilometers 94°50'0"W 94°40'0"W

Figure 28. Digital elevation model of the Trinity River Coastal Plain displaying the incised valley axis. Numbers in white text denote distance in kilometers from the modern shoreline (Galveston and Bolivar barriers). DEM data from www.usgs.gov

68 60

40 Onlap point 20

A 0 150 100 50 0

Figure 29. Longitudinal profiles of stratigraphic units. (A) Onshore Coastal Plain flood plain longitudinal profiles. Arrow denotes the onlap point of recent sediments onto the surface of the Low Deweyville unit. Orange = Beaumont, Brown = High Deweyville, Red = Middle Deweyville, Blue = Low Deweyville, Green = Modern floodplain

69 60 Onshore + bathymetry Offshore Valley_Axis_bathymetry Thomas, 1990, LD 40 MF_allu vial Thomas, 1990, MD Low_Deweyville Thomas, 1990, HD Mid_Deweyville 20 LD_Rod0 6 High_Deweyville MD_Rod06 Beaumont_Alluvial HD_ Ro d06 0 m , n

o -20 i t a v

e -40 l E -60

-80

-100 B -120 150 100 50 0 -50 -100 -150 -200 Distance from modern shoreline, km

Figure 29. (B) Data points and flood plain longitudinal profiles projected to the shelf edge. MF = Modern Floodplain. Offshore data from Thomas (1990), Trinity Bay data from Rodriguez et al. (2006), and this study. Refer to Table 3 for slope data.

70 The reconstruction of channel slopes (Sc) for the Deweyville units show that Sc is slightly

higher than the modern system for the High Deweyville unit, whereas Sc is actually less for the Middle Deweyville and Low Deweyville units when compared to the modern Sc.

Table 3. Valley longitudinal profiles, sinuosity, and channel slopes Slope of Channel Floodplain Profile Sinuosity (P) 2 Allostratigraphic Unit R Slope (Sc) Valley Slope(Sv) Lc/Lv (m/m) (m/m)

Beaumont 0.000300 0.992 N/A N/A

High Deweyville 0.000269 0.986 1.85 0.000146

Middle Deweyville 0.000267 0.972 2.92 0.0000913

Low Deweyville 0.000307 0.986 2.89 0.000106

Modern River 0.000238 0.980 1.76 0.000135

Paleohydrology

Estimates of paleodischarge were calculated for each Deweyville stratigraphic

unit, and the modern river at Liberty, Texas. The equations that proved to be the most

reliable for estimating actual values for the modern river system were the equations that

used the planform parameter of bankfull width and that were derived from rivers grouped

by their bed and bank sediment caliber parameters.

The modern river at Liberty, Texas has a mean radius of curvature of 243 m from

39 measurements, a mean meander wavelength of 902 m from three measurements, and a

mean bankfull width of 105 m from eighteen measurements (Table 4). Mean annual

71 3 -1 3 -1 discharge is 730 m s , whereas Q2 is ca. 1269 m s , which is the flood with a recurrence

3 -1 interval of 2 years, and Qmax is ca. 1329 m s , which is the mean annual flood over the

period of record (Appendix 1). Tables 4 and 5 summarize empirical discharge estimates

and deviations from actual values.

Percentage difference = ((actual – modeled)/actual)*100 (15)

The only empirical equations that yielded estimates within 25% of actual values

were the calculations of mean annual discharge using the equation derived from channels

with high silt clay beds, and estimates for Q2 using the equation derived from channels with medium silt clay beds, both from Osterkamp and Hedman (1982).

Relict channels of the High Deweyville unit have a mean radius of curvature of

761 m from seven measurements, and a mean bankfull width of 202 m from 21 relict channels (Table 4). Estimates of mean annual discharge (Qm) calculated from general

equations based upon a large number of rivers from different geographic settings range

from ca. 36% lower than the actual modern to ca. 22% increase relative to the modern

using radius of curvature. Using bankfull width (Wb), estimates of Qm range from 45–

68% less than the modern (Williams (1984) and Osterkamp and Hedman (1982)

respectively). The sand bed and banks equation of Osterkamp and Hedman (1982)

estimates mean annual discharge to be 78% lower than modern values. Estimates of Q2

are 3% and 17% lower from Osterkamp and Hedman’s (1982) general equations based

upon a large number of rivers from different geographic settings and sand bed and banks

respectively. Estimates of Qmax were 100% higher than modern values using radius of

curvature and 25% less than modern values using bankfull width equations from

Williams (1984).

72 Table 4. Estimates of mean annual discharge (Qm) using different parameters. General equations based upon a large number of rivers from different geographic settings.

Stratigraphic Radius of Meander Bankfull Qm = Qm = Qm = Qm = Qm = Unit 2.03 1.58 2.15 1.66 1.71 Curvature Wavelength Width 0.0021rc 0.025rc .000047Lm 0.06Wb 0.027Wb

rc Lm Wb Alford and Williams Alford and Williams (1984) Osterkamp and Holmes (1985) (1984) Holmes (1985) Hedman (1982)

m m m m3/s m3/s m3/s m3/s m3/s High 761 N/A 202 469 893 N/A 403 236 Deweyville Middle 1318 5362 267 1430 2126 1785 640 381 Deweyville Low 1428 4191 255 1681 2411 1051 593 352 Deweyville Modern 243 902 105 46 146 39 136 77 Liberty, ca. (-94) ca. (-80) ca. (-95) ca. (-81) ca. (-89) (% ± actual)

73 Table 5. Estimates of Qm, Q2, and Qmax. General equations based upon a large number of rivers from different geographic settings and equations derived from rivers grouped by the sediment caliber of the channel are noted.

1.38 1.3 Stratigraphic Qm = Q2 = Q2 = Qmax = 0.28rc Qmax = 1.0Wb

Unit 1.62 1.22 1.32 0.029Wb 1.9Wb 0.96Wb

sand bed/banks General Equation sand bed/banks General Equation General Equation Osterkamp and Osterkamp and Osterkamp and Williams (1984) Williams (1984) Hedman (1982) Hedman (1982) Hedman (1982) m3/s m3/s m3/s m3/s m3/s m3/s m3/s High 157 1234 1060 2653 993 Deweyville Middle 247 1734 1532 5660 1427 Deweyville Low 230 1640 1442 6317 1344 Deweyville 1.38 1.3 Qm = Qm = Q2 = Q2 = Q2 = Qmax = 0.28rc Qmax = 1.0Wb

2.12 1.76 1.22 1.86 1.27 0.031Wb 0.033Wb 1.9Wb 2.0Wb 2.6Wb

High silt/clay bed Med silt/clay bed General Equation High silt/clay bed Med silt/clay bed General Equation General Equation Osterkamp and Osterkamp and Osterkamp and Osterkamp and Osterkamp and Williams (1984) Williams (1984)

Hedman (1982) Hedman (1982) Hedman (1982) Hedman (1982) Hedman (1982) Modern, 597 119 555 11493 959 172 424 Liberty ca. (-18) ca. (-84) ca. (-56) ca. (+806) ca. (-25) ca. (-87) ca. (-68) (% ± actual)

74 Relict channels of the Middle Deweyville unit have a mean radius of curvature of

1318 m from 7 measurements, a mean meander wavelength of 5362 m from 4

measurements, and a mean bankfull width of 267 m from 29 measurements (Table 4).

Mean annual discharge estimates using radius of curvature are higher by 96% (Alford

and Holmes, 1985) and 191% (Williams, 1984). Using bankfull width, estimated values

for Qm range from 12% less (Williams, 1984) to just 48% less (Osterkamp and Hedman,

1982) than the modern using general equations based upon a large number of rivers from

different geographic settings. Mean annual discharge calculated using meander

wavelength from Alford and Holmes (1985) yielded an estimated increase of 145%. In

contrast, Qm calculated using the equation from Osterkamp and Hedman (1982) based

upon rivers with sand beds and banks is 66% less than modern values. Estimates of Q2 were 37% and 21% higher from general equations based upon a large number of rivers from different geographic settings and sand bed and banks respectively. Williams (1984) equations retrodict that Qmax is 326% higher using radius of curvature and 7% higher

using bankfull width.

Relict channels of the Low Deweyville unit have a mean radius of curvature of

1428 m from 20 measurements, a mean meander wavelength of 812 m from 11

measurements, and a mean bankfull width of 255 m from 10 measurements (Table 4).

Mean annual discharge estimates using equations from Alford and Holmes (1985) and

Williams (1984) based upon radius of curvature yielded higher than modern values at

130% and 230% respectively. Alford and Holmes (1985) equation using meander

wavelength yields a Qm estimate 44% larger than the modern. Using equations from

Williams (1984) and Osterkamp and Hedman (1982) based upon bankfull width,

75 estimates for Qm are 19% and 52% less than the modern using general equations based upon a large number of rivers from different geographic settings. By incorporating sand bed and banks, the Qm estimate from Osterkamp and Hedman’s equation using bankfull width is 68% less than the modern. Osterkamp and Hedman’s (1982) equations using general equations based upon a large number of rivers from different geographic settings and sand banks yielded increases in Q2 by 29% and 14%, respectively. Estimates of Qmax using radius of curvature are 375% larger than the actual modern values and only 1% larger than the modern values using bankfull width (Williams, 1984).

Sediment Supply

Estimates of sediment discharges for the Trinity River over the course of the last interglacial–glacial cycle, as calculated from the BQART model, are shown in Figure 30.

Estimates for water discharge (Q) were obtained from the BQART model using drainage area, but were multiplied by a factor of two to rectify discharge to actual modern values and the regional subtropical humid climate. The two different curves represent the inclusion of increased effective moisture based upon Musgrove et al. (2001). Results suggest ca. 12% decrease in sediment discharge during the latest portion of OIS 3, and ca.

25% decrease in sediment discharge during OIS 2 when compared to modern calculated values (Fig. 30).

Sediment Export During Valley Evolution

Total sediment excavated from the post–Beaumont Coastal Plain valley is calculated to be ca. 38,626 MT. The High Deweyville incises older pre–Beaumont clay and cemented very fine to fine–grained sandstone in updip reaches of the study area, and uncemented sands and muds of the Pleistocene Beaumont alluvial–deltaic plains farther

76 Sediment Discharge (BQART) Sediment Discharge (BQART) (including 25% i ncrease in Q from Musgrove et al. (2001) OIS 1 OIS 2 OIS 3 OIS 4 OIS 5 5.2 a)

/ 5.0 4.8 4.6

harge (MT 4.4 4.2 4.0 3.8

Sediment disc 3.6 3.4

Time (ka)

Figure 30. Calculated sediment discharge for the Trinity River over the last glacial–interglacial cycle using the Syvitski and Milliman (2007) BQART model. The blue curve incorporates an increase in discharge (Q) of 25%, according to time periods of more effective moisture from Musgrove et al. (2001).

77 downstream. The phase of valley incision and lateral migration recorded by the High

Deweyville unit resulted in net export of 24,726 MT (Table 7), or 64% of the total.

Dividing over the period of time represented by the High Deweyville unit (ca. 30 kyrs) yields an increase in sediment discharge of ca. 0.82 MT/yr. The Middle and Low

Deweyville units primarily record incision into the older channelbelts within the post–

Beaumont valley, but also Beaumont and older undifferentiated sediments. Incision and formation of the Middle Deweyville channelbelt (ca. 9.5 kyrs) represents net removal of sediment mass of ca. 4,860 MT (13% of the total) or ca. 0.51 MT/yr, whereas incision and formation of the Low Deweyville channelbelt (ca. 7 kyrs) represents a net removal of sediment mass of ca. 9,040 MT (23% of the total) or ca. 1.29 MT/yr.

Estimates of exported sediment mass were combined with sediment discharges from farther upstream, (estimated from the BQART model) to produce estimates of total sediment discharge from the lower Coastal Plain during formation of the Trinity valley over the course of the last glacial period falling stage and lowstand, as well as during specific periods of channelbelt activity (Table 8). When averaged over the entire glacial period falling stage and lowstand, the total mass of sediment exported is estimated to be ca. 12% of the total discharge from farther upstream, or ca. 10% of the overall sediment budget from combined exported sediment and upstream supply. However, mean annual sediment discharge was ca. 10% greater than modeled modern values during incision and formation of the High Deweyville unit, ca. 2% less than modeled modern values during incision and formation of the Middle Deweyville unit, and ca. 5% greater than modeled modern values during incision and formation of the Low Deweyville unit.

78 Table 6. Volumetric calculations of stored and excavated sediment.

Unit and areas Surface area, Deposit Volume sediment Thickness Volume of sediment removed calculated, (x106 m2) thickness, stored in channelbelt, removed (x106m3) (km updip from modern (m) (x106m3) (m) shoreline, ref. Fig. 32) High Deweyville 475 6 2,850 12–18 7,125 (km 102–155) High Deweyville 775 11 8,525 9–12 8,138 (km 0–102) Mid Deweyville 350 7 2,450 3 1,050 (km 102–155) Middle Deweyville 650 10 6,500 3 1,950 (km 0–102) Low Deweyville 260 7–8 1,950 2–4 780 (km 102–155) Low Deweyville 800 9–11 8,000 5–7 4,800 (km 0–102)

79 Table 7. Calculations of sediment mass stored in channelbelts and rates of excavated sediment mass.

Unit and areas Stored sediment mass Total excavated Inferred Rate of sediment calculated, in channelbelt, (MT) sediment mass, time period, excavated, (MT/yr) (km updip from (MT) (yrs) modern shoreline, ref. Fig. 26) High Deweyville 18,428 24,726 30,000 0.82 (km 0–155) Mid Deweyville 14,499 4,860 9,500 0.51 (km 0–155) Low Deweyville 16,119 9,040 7,000 1.3 (km 0–155)

80 Table 8. Mass balance calculations of extra sediment due to excavated sediments per unit valley length (m), per unit channel length (m), and modeled plus excavated sediment delivered past the modern shoreline per year. (These values do not include the confluence of the San Jacinto River in Galveston Bay.) Modern modeled and actual values (in bold) represent sediment delivered to Trinity Bay.

Unit and areas Extra qs, Extra qs, Modeled + qs modeled + calculated, (T/Lv/a) Lc=P x Lv, (T/Lc/a) increase excavated

(km updip from (Lv in m) (m) (Lc in m) (T/Lv/a) sediment modern actual values (MT/a) shoreline, ref. in bold Fig. 26) High 5.5 277,500 3.0 35.2 5.02 Deweyville (km 0–155) Mid Deweyville 3.4 438,000 1.2 31.4 4.71 (km 0–155) Low 8.6 433,500 3.0 33.6 5.04 Deweyville (km 0–155) Modern N/A 176,000 N/A 48.0. (55.0) 4.8, (5.5)

The excavated mass of sediment within the Lower Coastal Plain can also be used as a proxy for the accommodation that remains to be filled during the present transgression and highstand. At present sediment discharge rates of 5.5 MT/yr, the valley would require ca. 7,000 years to fill. If the Coastal Plain valley were to fill completely before the next major sea–level fall, then the Deweyville falling–stage to lowstand deposits would comprise ca. 44 % of the post–Beaumont valley fill, whereas post–

Deweyville sediments would comprise the remaining 55%.

81 DISCUSSION

Valley–Fill Architecture

Post–Beaumont valleys of the Texas Coastal Plain display contrasting facies between glacial and interglacial deposits and contain multiple internal sequence boundaries that exhibit characteristics of compound incised–valley fills (Zaitlin et al.,

1994). During the last glacial period, channel incision and lateral migration resulted in a flight of three downward–stepping and cross–cutting terraces (Deweyville), which are inset into, and therefore younger, than Beaumont strata. Each Deweyville terrace is underlain by mostly channelbelt sands that rest unconformably on Beaumont strata and are bounded at the top by a paleosol, indicating an extended period of subaerial exposure and weathering. The erosional base of each Deweyville channelbelt represents distinctive time transgressive unconformities (Blum and Aslan, 2006) that trace from their respective channelbelt margins and amalgamate on top of the soil profiles developed in older

Deweyville units and the surfaces of Beaumont and pre–Beaumont deposits outside of the valley (Figs. 31, 32). Furthermore, the Deweyville units lack appreciative thickness of fine–grained topstratum, which would be interpreted as overbank deposits, and paleosols are developed in fine– to very fine–grained sand.

The recent transgression and highstand has resulted in the formation of Trinity

Bay above the now submerged Trinity valley. The High and Middle Deweyville units remain as terraces in the modern incised valley updip from the modern bayhead delta, where the recent sediments rest unconformably on top of the paleosols. The Low

Deweyville paleosol has been buried ca. 65 km updip from the modern bayhead delta by

82 A Post–Deweyville Flo odplain A’ Post–Deweyville C hannelbelt 45 Low D eweyville Middle D eweyville 40 High D eweyville 35 Beaumont n, m

o 30

ati 25 20

elev 15 10 5 k m

Figure 31. Cross–sections across the Coastal Plain Trinity River incised valley. See figures 14b and 24 for cross–section locations.

A–A’) At this location, the High Deweyville is a terrace flanking the eastern side of the valley, the Middle Deweyville unit is a terrace flanking the western side of the valley and the Low Deweyville unit is referred to as a terrace veneer, due to the overlying thin veneer of post–Deweyville floodplain sediments.

B–B’) At this site, the High and Middle Deweyville units are terraces preserved within the eastern side of the valley while the Low Deweyville unit is not a terrace in the classic sense as it is now buried by a wedge of post–Deweyville sediments.

C–C’) The High and Middle Deweyville are still present as terraces along the eastern valley wall, while the modern river is currently flowing through an old Middle Deweyville paleochannel. The Low Deweyville unit is completely buried by Holocene alluvial, bayhead delta, and estuarine sediments. Modified from Morton et al. (1996).

83 Post–Dew eyville Floodplain B Post–Dew eyville C hannelbelt B’ 30 Low Deweyville 25 Middle D eweyville High Deweyville 20 Beaumont 15 10 n, m

o 5

ati 0 -5

elev - 10 - 1 5 5 k m C C’ 15 10 5 0

n, m -5 o i -1 0 at -1 5 -2 0 elev -2 5 -3 0 5 k m

Figure 31. (cont.)

84 Figure 32. 3–D schematic of valley–fill components and their stratigraphic architecture. Glacial period channelbelts are inset into and against older deposits while their respective basal surfaces of erosion represent time transgressive unconformities that are formed during incision and subsequent lateral migration of the active channel, which trace upwards at channelbelt margins onto the terrace surfaces of older units within the valley– fill as well as the previous highstand depositional surface (Beaumont). In updip reaches, above the onlap point, the channelbelts occur as terraces, whereas in downdip reaches, they are onlapped by more recent sediments that are sequestered due to post–glacial sea– level rise.

85 Glacial period terraces Post–Deweyville Flood plain Post–Deweyville Channelbelt Holocene channelbelt Low D eweyville Middle Deweyville High Deweyville Beaumont

Falling stage fluvial deposits

Onlap

Onlap

Onlap surface

Base of coastal–plain incised valley Transgressive channelbelt Lowstand fluvial deposits

86 younger fluvial and fluvial–deltaic facies associated with the recent transgression and highstand (Fig. 29). The cause of this onlap relationship is apparent in the flood plain longitudinal profiles of the Deweyville and post–Deweyville sediments. The steeper flood plain profiles of the Deweyville units dip below the less steep modern floodplain and alluvial–deltaic profiles (Figs. 29, 32). This relationship is primarily due to the present sea–level position, which the modern system is graded to when compared to the

Deweyville units. In the updip portion of the valley, the post–Deweyville valley–fill is comprised primarily of alluvial sediments. Post–Deweyville channelbelts currently flow through the Deweyville deposits and sediments of their own origin. These consist of sandy and muddy point bar and channel fill deposits, as well as laterally persistent vertically accreted fine–grained flood plain sediments that thicken downdip due to increasing accommodation provided by the contrasting gradient of the Low Deweyville flood plain when compared to the modern. These sediments unconformably onlap the paleosols developed on the steeper gradient flood plain profiles of Deweyville surfaces, according to their respective elevation and gradient relationships (Fig. 32).

The estuary, which resides between the seaward Galveston and Bolivar barriers and the updip subaerial components of the valley–fill, has and is providing the accommodation for post–Deweyville sediments to accumulate. In Trinity Bay, post– glacial sea–level rise resulted in backstepping depositional cycles. An older marsh surface rests above the presently drowned Deweyville stratigraphic units, and is interpreted to have formed when sea level was at or near 6 meters below present (Fig.

33). The slope of the mean bayline gradient from the offshore valley (Thomas, 1990), which appears to correlate to the older marsh surface identified in this study, is identical

87

Figure 33. Cross–sections in downdip reaches of the Coastal Plain valley. (A) Cross–section D–D’ depicting recent valley–fill sediments and interpreted depositional environments. See Figure 24 for cross–section and core locations. Distance between cores is not to scale a) Older marsh surface (ca. 6+ mbpsl) overlain by central basin estuarine (CBE), prodelta (PD), delta front (DF), updip alluvial and downdip channel (CH), natural levee (NL), interdistributary bay (IB), middle ground bar (MGB) and distributary mouth bar (DMB), and modern marsh (M) depositional environments.

88

Figure 33. (B) Dip oriented cross–section E–E’. See Figure 24 for cross–section location. Older marsh surface (ca. 6+ mbpsl) overlain by central basin estuarine (CBE), prodelta (PD), delta front (DF), updip alluvial and downdip channel (CH), natural levee (NL), interdistributary bay (IB), middle ground bar (MGB) and distributary mouth bar (DMB), and modern marsh (M) depositional environments.

89 to the Low Deweyville floodplain longitudinal profile. This relationship suggests that the accommodation for these sediments to accumulate, as well as their depositional profile, were controlled partly by this lower bounding surface. The marsh surface was then flooded as sea level continued to rise, which caused a landward translation of depositional environments and resulted in sediments deposited in a central basin estuarine environment residing overlying the older marsh surface. As sediment supply exceeded the rate of sea–level rise, the bayhead delta prograded into Trinity Bay, which resulted in prodelta, delta front, and other sediments associated with bayhead deltaic depositional environments overlying central basin estuarine sediments (Fig. 33).

Geochronology

Optically stimulated luminescence ages reported in Table 1 constrain the timing of channelbelt activity, abandonment due to incision and lateral migration, and the subsequent formation of terraces. All OSL age estimates place deposition of the

Deweyville units within the OIS 4–2 glacial period, when sea level was significantly lower than present and the glacial period Trinity River was extended across the shelf.

However, three samples produced age estimates that violate cross–cutting relationships, which can be due to several sources of error. The two samples that yielded age determinations that are too old for stratigraphic context may reflect partial bleaching during sediment transport, such that the luminescence signal is not completely reset

(Murray and Olley, 2002, Thomsen et al., 2005). By contrast, sample HD_Sandune yielded an age estimate that was too young given its stratigraphic context: it is from a

High Deweyville deposit, but gave an age estimate of 12 ka, which is younger than all other samples. Examination of data used to calculate this age estimate indicated a dose

90 rate that is approximately twice the mean of the other samples, which could be due to an unrepresentative presence of radioactive materials in the bulk sample, and/or error in lab measurement of dose rate (Aznar et al., 2003, Nathan et al., 2003, Ankjaergaard and

Murray, 2007).

From the stratigraphic framework and existing geochronological data, the incision and abandonment of the Beaumont depositional surface is attributed to the fall in sea level that occurred during OIS 4 (ca. 70–65 ka), which represents a maximum age for the

High Deweyville unit, although the OSL ages from this study suggest deposition occurred significantly later, ca. 33 ± 2 ka. The range of ages for channelbelt activity for the Middle Deweyville are from 34–23 ka. Because of the cross–cutting relationships between the High Deweyville and Middle Deweyville units, the timing of incision and abandonment of the High Deweyville channelbelt occurred sometime between 35 and 30 ka, which is the maximum age determination of sample OSL sample HD_Liberty and the minimum age determination of sample MD_Kenefick. Age ranges for the Low

Deweyville unit (23.2–18.8 ka) show that incision and abandonment of the Middle

Deweyville unit occurred some time between 26.6 and 20.4 ka. The maximum age determination of the Low Deweyville unit could be up to 26.6 ka, considering the maximum error from sample MD_Liberty. This maximum is excluded because of the cluster of Low Deweyville ages from samples LD_Romayor_North, LD_Liberty_South, and LD_Port_of_Liberty_South, as well as ages cited from other studies (see Blum and

Aslan, 2006). The youngest age for the Low Deweyville unit from these data is 18.8 ka, but it remains possible that the youngest active portions of the channelbelt may not have been sampled, and it seems likely that the latest phase of activity for the Low Deweyville

91 unit correlates with similar ages from nearby GOM age equivalent deposits at ca. 16 ka

(Blum and Aslan, 2006). From these data, the age ranges for each unit are interpreted as

follows:

• The High Deweyville unit was deposited between 65 – 32.5 ka during the

falling stage of sea level, providing a range of ca. 30 kyrs, whereas this

number used in calculations of sediment mass balance rates could be of a

smaller duration assuming that significant self cannibalization occurred

over the course of this time interval.

• The Middle Deweyville unit was deposited between ca. 32.5 –23 ka

during the falling stage and lowstand of sea level, representing a duration

of ca. 9.5 kyrs, and

• The Low Deweyville unit was deposited between ca. 23–16 ka during the

lowstand of sea level, representing ca. 7 kyrs of channelbelt activity

following incision and abandonment of the Middle Deweyville unit.

Valley–Fill Evolution

The stratigraphic and geochronologic sequences identified in the Trinity River incised valley is interpreted to reflect the following sequence of events: (a) initial channel incision (ca. 65 ka) resulting in the abandonment of the Beaumont depositional surface and began the formation of the Trinity River Incised Valley, which was followed by lateral migration of the active channel (30 kyrs) to form the High Deweyville channelbelt deposits; (b) renewed channel and valley incision (ca. 32.5 ka) with the subsequent lateral migration (ca. 9.5 kyrs) resulting in the formation of the High Deweyville terrace and the

Middle Deweyville channelbelt deposits; (c) renewed channel and valley incision (ca. 23

92 ka) with lateral migration (ca. 7 kyrs) forming the Middle Deweyville terrace and the

Low Deweyville channelbelt deposits; (d) renewed channel incision (ca. 16 ka) to form

the Low Deweyville terrace, followed by valley filling that continues today due to post–

glacial sea–level rise.

Paleohydrology

Paleohydrological analyses yielded mixed results. Some of the widely used

measures of channel geometry are not always related to discharge, a generalization

supported by analyses herein. For example, Leopold and Wolman (1956) and Rotnicki

(1983) note that discharge does not directly control meander wavelength (Lm), but rather wavelength is controlled more by channel width. This relationship is an important factor to consider because channel width is partly controlled by the grain size of sediment in the bed and banks (Schumm 1960; Schumm, 1968; Osterkamp and Hedman, 1982; Rotnicki,

1983; Church, 2006). Application of Williams’ (1984) empirical equations that use

radius of curvature (rc) illustrate these issues: estimates for the modern river greatly

underestimate the actual gauging station–derived values of Qmax, whereas estimates of

Qmax for the Deweyville units range from 100–375% higher than modern values. Because paleodischarge estimates using radius of curvature and other measures of plan–view geometry cannot predict modern discharge values, they should not be used to estimate formative discharge for the Deweyville units. Furthermore, using Williams’ (1984) empirical equations based upon bankfull width (Wb), Qmax estimates for the Deweyville units range from 25% less than modern values to just 7% greater than modern values, which are much lower than estimates using radius of curvature derived from the same dataset (Table 5).

93 By contrast, the empirical equations of Osterkamp and Hedman (1982), which

include a term for sediment caliber, predict Qm and Q2 to within 25% of modern values using equations derived from rivers with high silt/clay beds and medium silt/clay beds.

Given the apparent significance of sediment caliber, it follows that only equations that incorporate a term for sediment characteristics of the channel should be used to estimate paleodischarge values for the Deweyville units. Estimates of Q2 for Deweyville units,

which assumes sand beds and banks, suggest discharges were 17% less than modern

during deposition of the High Deweyville unit, and perhaps 21% and 14% greater during

deposition of the Middle and Low Deweyville units respectively. These values appear to

be more reasonable than previous interpretations of 4 to 5 fold increases in

paleodischarges relative to the modern: as noted by previous workers (Blum and

Valastro, 1994; Blum et. al., 1995; Morton et. al., 1996; Durbin, 1997; Blum and Price,

1998; Aslan and Blum, 1999; Blum and Törnqvist, 2000; Blum and Aslan, 2006; Sylvia

and Galloway, 2006) and this study, overbank fines are rare within the Deweyville units,

which suggests that extreme magnitude floods either did not occur, or were very rare,

during deposition of the Deweyville units.

Therefore, results of this study suggest previous estimates of paleodischarge are

inflated because they used measures of channel plan–view geometry and because they did

not incorporate sediment caliber into their methods. Although estimates of

paleodischarge are imprecise, there is no sound basis to interpret that channel formative

discharges during deposition of the Deweyville units were significantly greater than those

of the modern river.

94 Sediment Mass Balance

Estimates using the BQART model suggest a 12% decrease in sediment yield

during OIS 3, and a 25% decrease during the LGM, relative to the modeled modern

interglacial period, due to depression of temperatures (see also Womack, 2007, and Blum

and Womack, 2008–in press). Although treated as a step–wise steady–state value here, it

is important to recognize that modeling of drainage basin response to climate change

suggests that changes in sediment supply would occur gradually over time scales of ca.

103–104 years (Tucker and Slingerland, 1997). Furthermore, the transfer subsystem

(rivers) can buffer high–frequency sediment flux variations over time scales of 104–105 years (Castelltort and van den Driessche, 2003). Moreover, the BQART model only estimates suspended load. The percent bedload is difficult to estimate in modern systems, let alone systems of Deweyville age, so the discussion here assumes that changes in supply of bedload parallel changes in suspended load. Sediment mass balance calculations include significant uncertainty associated with estimating past sediment loads, with the assumptions made for duration of deposition of the Deweyville units, and are discussed here solely as first order approximations. For example, estimated net sediment export of 0.82 MT/yr from the incised valley during deposition for the High

Deweyville unit, which is ca. 20% of the isotope stage 3 sediment loads from upstream as estimated from the BQART model, results in an estimated total load delivered from the incised valley that would have been ca. 10% greater than modeled modern values (Fig.

34). Similarly, net sediment export from the incised valley during deposition of the

Middle Deweyville unit was 0.51 MT/yr, which is ca. 12% of the late OIS 3 sediment loads from upstream, resulted in a total load that was ca. 2% less than modeled modern

95 values. Finally, net sediment export during deposition of the Low Deweyville unit was

1.29 MT/yr, which is ca. 34% of the OIS 2 sediment loads from upstream, resulted in

total estimated sediment discharge rates that were 5% greater than modeled modern

values.

The key results that emerge from these calculations are that the total sediment

export from the incised valley over the course of an entire glacial period is a relatively

small sum, compared with the total sediment mass delivered from the drainage basin

(Blum and Törnqvist, 2000), and time periods dominated by channel incision and valley

deepening contributed little additional export of sediment. However, during time periods

of Deweyville deposition, the process of lateral migration and channelbelt formation

resulted in a net export of sediment that enhanced total loads by 12–34%, and offset

reductions in sediment yields that might result from glacial–period climate change.

Increased sediment delivery to downdip reaches of the fluvial–deltaic system therefore

reflect time periods of channelbelt deposition within the incised valley, and provide the

physical mass balance basis for coupling falling stage fluvial deposition and deltaic deposition on the shelf and shelf margin.

This interpretation is in sharp contrast to the traditional view that the period of incision within the incised valley correlates to falling stage and lowstand deltaic progradation on the shelf (e.g. Posamentier et al., 1992). In the Posamentier et al. (1992) model, incision with production of an unconformity represents a long period of sediment bypass that is necessary to provide the sediment needed to build deltas farther downdip.

However, Deweyville units represent a significant portion of the total falling stage and

96 Post– High LD MD Beaumont Deweyville Deweyville

s e e u l g +10% a r ha d v

c +5%

e ? s l ? i ? e d

d 0% t o n e

m -5% m n r di e -10% e d o s

n -15% m i o e t g -20% e n v a i t h - 25% a C l e r

Time (ka)

Sediment Discharge Range (BQART) Modeled Sediment Discharge (BQART) + Excavated Sedeimemnt .

Figure 34. Composite graph showing ages of stratigraphic units, modeled (BQART) and modeled plus excavated sediment discharge over the last 120 kyrs. Enhanced sediment flux to downdip reaches occurs during time periods of valley widening through channelbelt construction.

97 lowstand. The Trinity incised valley stratigraphic model, when coupled to the mass balance calculations illustrated above, indicates that: (a) the basal valley–fill surface was time–transgressive and was created in a step–wise manner by periods of channel incision followed by periods of lateral channel migration; (b) channelbelt deposits that rest on that surface were deposited contemporaneously with formation of the basal valley–fill surface, and correspond to periods of valley widening, (c) increased sediment flux from the incised valley, due to formation of the incised valley itself, is stepwise, and corresponded to time periods of channelbelt deposition within the incised valley, and (d) periods of channelbelt deposition in the incised valley therefore corresponded with periods of enhanced sediment loads delivered to a contemporaneous river mouth, and provide the mass–balance basis for coupling falling stage fluvial and deltaic deposition.

Allogenic Controls

Deciphering controls on changes in fluvial systems through time is problematic due to the numerous variables that can alter or change the system as a whole, and because different systems might show a convergence of response to different controls, or divergence of response to the same forcing mechanism (Schumm, 1991). Based on my results, I evaluate these upstream and downstream controls that influenced evolution of the Trinity incised valley system.

Temperature, precipitation, vegetation, soils, sediment supply, and precipitation intensity all interact and influence downstream responses within fluvial systems

(Rotnicki, 1983). First, as interpreted by previous workers, and confirmed by this study,

Deweyville units lack vertically accreted overbank deposits, which suggests that high magnitude floods were extremely rare during the last glacial period (Blum and Valastro,

98 1994; Blum et al., 1995; Morton et al., 1996). This is partly attributable to a decrease in onshore flow of GOM moisture during the glacial period due to cooling over land relative to the oceans, and the southward displacement of the jet stream due to presence of the

LIS (Kutzbach and Guetter, 1986; Toomey et al., 1993). Moreover, thick upland soils and C4 grasslands in upland source terrains have been interpreted from other drainage basins in Texas: if these same conditions prevailed in the Trinity system, formative discharges would have been longer in duration but with lower peaks (Blum and Valastro,

1994; Blum et al., 1995; Morton et al, 1996; Nordt et al., 2003), such that overbank flooding would have been very rare.

Second, Deweyville channel size has long been interpreted to reflect climate controls. Although there is a clear divergence of views on this issue, with some authors interpreting paleodischarges several orders of magnitude greater than today, it does seem likely that differences in channel morphology between the Deweyville and modern system reflect climate controls in some way. Paleohydrological analyses discussed above suggest that: (a) prior research used empirical equations that do not fit the modern Trinity

River, so are inappropriate for older deposits as well, (b) empirical equations, which include a term for grain size, provide the best fit for the modern system. When applied to the Deweyville units, these equations suggest only modest increases in the magnitude of formative discharges.

An important corollary is how such an interpretation links to existing paleoclimate data and model results. For example, climate reconstructions show that precipitation during the last glacial period was likely derived from mid latitude cyclonic storms (Toomey et al., 1993), and that rainfall events were less intense than at present,

99 especially during the summer months and tropical storm season. Moreover,

reconstructions of LGM precipitation anomalies show that the amount of annual

precipitation was not significantly greater than today, although there was more effective

moisture due to decreased temperatures (Mock and Bartlein, 1995; Shin et al., 2003). As a result, there is no clear basis to argue for significantly increased precipitation to produce significantly larger floods.

Downstream controls are important, as well. During the last glacial period, sea– level fall is interpreted to have affected the lower Trinity River system by forcing incision through the Beaumont highstand prism channel extension across the shelf. Longitudinal profiles for the High, Middle, and Low Deweyville units exhibit steeper gradients than the modern system, indicating they were graded to lowered sea level. The post–

Beaumont valley axis on the shelf (Fig. 9), mapped by Thomas (1990) and Anderson et al. (2004), shows that channels must have extended ca. 180 km to reach the -90 to -100 m bathymetric contour. A plot of valley longitudinal profiles and present day bathymetry show that projected High, Middle, and Low Deweyville surfaces would intersect the shelf edge at depths of 52, 57, and 72 m respectively (Fig. 29).

Average sea–level positions when the High and Middle Deweyville channelbelts were active are much lower (ca. -80 to -120 meters) than reconstructed and projected valley longitudinal profiles suggest, (ca. -55 to -75 m). Hence, their longitudinal profile

gradients may be primarily controlled by the gradient of the shelf, and did not reach

equilibrium with contemporaneous sea level. This relationship indicates that incision

through the highstand coastal prism is controlled primarily by channel extension and an

adjustment of the longitudinal profile to a preexisting boundary condition, being defined

100 by the shelf gradient. Moreover, longitudinal profiles for the High and Middle

Deweyville units are essentially the same within the coastal plain, although vertically offset by ca. 3 m. This relationship further suggests that incision in response to sea–level fall, other than that required to initially incise through the highstand prism and extend across the shelf, had not yet reached the updip coastal plain, and that the incision and abandonment of the High Deweyville unit, followed by formation of the Middle

Deweyville unit, was not triggered by sea–level fall. An alternative cause, or set of causes, would be climate–controlled decreases in qs and/or increases in Q.

By contrast, the Low Deweyville longitudinal profile is steeper than both the High and Middle Deweyville units, which suggests it was actually graded to a lower sea–level position. Van Heijst and Postma (2001) used a flume experiment to estimate relative rates of knickpoint migration due to base–level fall, and suggest that rates on Quaternary passive margin settings can range from 10–100 km/ 1000 yrs. If one assumes the steeper long profile for the Low Deweyville unit was actually graded to the lowstand sea–level position, then this would imply knickpoint migration up through the incised valley and through the coastal plain study area. Because the rapid fall of sea level into the LGM occurred at ca. 26 ka, and the abandonment of the Middle Deweyville unit occurred ca.

23 ka, this would imply that it took ca. 3 kyrs for the knickpoint to travel greater than 230 km, which would result in a rate of ca. 77 km/ 1000 yrs. Assuming the Low Deweyville unit was actually graded to a sea–level position of -72 m, and using sea level reconstructions to determine when sea level was at this position (ca. 43 ka), the resultant knickpoint migration rate would be ca. 11.5 km/ 1000 yrs.

101 Climate controls on the Low Deweyville unit should be considered as well: for

example, Musgrove et al. (2001) show that rapid speleothem growth rates around 24 ka

were either the result of an increase in precipitation or effective moisture. Incision and

abandonment of the Middle Deweyville depositional surface, and construction of the Low

Deweyville unit, is therefore interpreted to represent the convergence of an increase in

effective moisture around 24 ka coupled with response to the rapid fall in sea level to the

last glacial maximum position, which triggered incision and abandonment of the Middle

Deweyville unit and controlled the long profile evolution for the Low Deweyville unit.

The contrasts in post–Deweyville deposits are directly attributable to allogenic

controls. Sea–level rise during the recent transgression and highstand forced flattening of

the floodplain gradient within the lower valley and increased sediment storage in the

floodplain of the modern river. Moreover, climate change resulted in more intense

storms that stripped upland soils, introducing more fines into the main trunk channel, and

increased the flashiness of floods causing overbank deposition of a laterally extensive

fine grained floodplain in the lower valley. Moreover, at nearly every cutbank, the

resultant erosion of the fine grained sediments stored in the floodplain attribute more

suspended sediments into the system as well as sand sized aggregates of clay. The

landward translation of the shoreline, due to sea–level rise, flooded the lower valley, and resulted in heterolithic estuarine and alluvial–deltaic environments of deposition above the falling stage and lowstand fluvial deposits.

Fluvial Response to Allogenic Controls

Increased discharge was likely not the cause of the large Deweyville paleochannels, radii of curvature, and bankfull widths. To accurately assess the

102 morphological fluvial response to allogenic controls, there are a number of variables that

must be addressed. In this case, the initiation of valley incision can be due to several factors, including but not limited to a decrease in upstream sediment supply, base–level fall resulting in increased slopes, or an increase in discharge, all which would effectively increase stream power and the sediment transport capacity of the river (Blum and

Törnqvist, 2000; Bridge, 2003). The following discussion draws much of its basis from

Eaton et al. (2004) and Eaton and Church (2004). The governing variables that are most important include: 1) the preexisting shelf gradient, which has a large control on the resultant valley slope (Sv); 2) the downstream control on base–level which also affects

valley slope (Sv); 3) upstream controls on discharge (Q); 4) upstream controls on bedload

sediment supply/discharge (Qs) and suspended load sediment supply/discharge (qs); and

5) the upstream controls on sediment caliber (D) (Eaton and Church, 2004). Fluvial systems have been regarded as trying to achieve a graded state (equilibrium) that is adjusted to the imposed discharge and sediment loads. The concept of grade is often confusing, especially in the literature, due to the numerous definitions and implications of the concepts (Howard, 1982). The term equilibrium, used hereafter in this thesis, is applied in regards to the case of the Deweyville allostratigraphic units which display periods where lateral channel migration persisted with no net vertical incision or aggradation.

Eaton et al. (2004) explain that stability, or equilibrium, can best be achieved by maximizing the resistance to flow in the system (fsys),

fsys = f’ + f’’ + f’’’ (16)

103 Where fsys is separated into the three components; f’ is grain resistance; f’’ is form

resistance from dune scale to bar features, and f’’’ is the reach–scale flow resistance due to changes in channel length (Lc), or equivalently sinuosity (P).

Recent studies have shown that fsys is maximized by slope minimization behavior of the system, which is the response of the system to obtain the minimum slope required to transport its water and sediment load (Eaton et al., 2004; Eaton and Church, 2004;

Eaton et al., 2006). Early observations of this effect in flumes and field examples were noted due to fluctuations and base–level and tectonic uplift or subsidence which influence valley slope (Schumm, 1963; Schumm and Khan 1972; Schumm, 1993; Wood et al.,

1994; Holbrook and Schumm, 1999; Schumm, 2005). In the flume studies of Eaton and

Church (2004), they observed that f’’’ was the most important component to the reach

scale flow resistance, that this was achieved by increases in sinuosity (P), and that f’ and

f’’ are both second order effects with f’ having no appreciable effect on the equilibrium

channel pattern. Eaton and Church (2004) concluded that higher sinuosities, (or lower

channel slopes), were the result of decreasing sediment concentration, which is the ratio

of bedload discharge (Qs) to discharge (Q).

Evidence for this behavior in the case of the Deweyville units can be related to a measure of stream competence. Bridge (2003) defines bed shear stress (τ), (or the spatially averaged fluid shear stress at the bed), as the balance of gravity, friction, and pressure forces in the flow direction and is shown in equation 17,

τ = ρgdS (17)

104 where ρ is the fluid density, g is gravity, d is the mean flow depth, and S is the slope of

the bed and water surface (channel gradient, Sc). This expression is further expanded to include channel dimensions and then becomes,

τ = ρgScR (18)

where R is the hydraulic radius which is defined as,

R = (dw/(2d+w)) (19)

where d and w are depth and width of the channel respectively, (assuming a rectangular channel), and whereas when channels widen, R approaches depth. Stream competence is the ability of a stream to transport its sediment, (of a particular size), and is defined by the Shields number (τ*):

* τ = τ/g(ρs-ρ)D (20)

where ρs is the sediment density and D is the grain diameter of the sediment.

The High Deweyville unit displays τ* values that are slightly higher than the

modern river. Using the τ* and the estimated sediment discharges as a measure of

transport capacity, these values appear to be in good agreement. The Middle Deweyville

and Low Deweyville units exhibit τ* values that are lower than the High Deweyville unit

and the modern river, and are primarily due to decreases in channel slope (Sc). The

Middle Deweyville unit exhibits the lowest channel slopes, and estimates of sediment

discharge during this time were also the lowest. This further supports the likely cause of

incision and abandonment of the High Deweyville unit is attributable to upstream climate

controls through a decrease in sediment supply, or an increase in discharge. Following

incision, the system readjusted through commencement of lateral migration and a

subsequent decrease in channel slope by an increase in sinuosity.

105 The Low Deweyville unit exhibits slightly increased Sc relative to the Middle

Deweyville. This suggests that sediment concentration (Qb/Q) is slightly increased relative to the Middle Deweyville unit, but is less than the High Deweyville unit. This is in contrast to BQART modeled sediment supply, but agrees with calculated net excavated plus modeled sediment discharge rates during the deposition of the Low Deweyville unit.

The convergence of response of increased valley slopes due to base–level fall and an increase in discharge, from Musgrove et al. (2001), resulted in incision and abandonment of the Middle Deweyville through degradation, but once the channel slope reached some critical point, lateral deformation was initiated and sinuosity increased until the resultant channel slope was suited to transport the imposed sediment and water delivered to the channel from both upstream controls and the net removal of sediment.

The implications of these studies are that when discharge, sediment supply, and valley slope are independent variables, the channel will adjust to deliver the sediment and fluid supplied to the system by modifying channel width, depth, and channel length

(slope). This is interpreted to be the primary cause of the larger radii of curvature,

meander wavelengths, and enlarged paleochannels of the Deweyville units. During base–

level fall associated with the last glacial period, the long profiles of the fluvial system

adjusted to higher slopes due to channel extension and the boundary conditions set in

place by the emerging topography of the previous highstand surface. Because an increase

in slope initiated a higher transport capacity due to valley longitudinal profile adjustment,

the system responded initially through vertical incision until a threshold was met at which

lateral migration commenced and flow resistance was maximized by decreasing channel

slope (increasing sinuosity). Because there was still a net excavation of sediment out of

106 the channel, the system is still supply limited, but instead of net degradation occurring vertically, this was done through lateral migration of the river, which resulted in channelbelt deposition and formation. This is further supported from the research by

Eaton et al. (2006) where at the apex (cut–bank) there is a net excavation of sediment so that slope is further reduced to maintain the reach averaged sediment throughput. The lack of cohesive banks due to the absence of laterally persistent fine–grained flood plain deposits created a situation in which lateral migration was not prohibited during deposition of the Deweyville units. This is more apparent in the case of the Middle and

Low Deweyville units, where they are migrating laterally primarily into unconsolidated

High Deweyville channelbelt sands and gravels as opposed to the High Deweyville unit, which is migrating laterally into the fine–grained sediments of the Beaumont Formation.

To classify the Deweyville channels, they plot between the mixed load and bedload channels of Dade and Friend (1998) and display the characteristics of transitional channels (see Church 2006 for discussion). Eaton and Church (2004) noted in their flume experiments that larger channel widths displayed a stronger correlation to increases in sinuosity than compared to sediment caliber. Comparisons of the sediment caliber and channel widths between the High, Middle, and Low Deweyville units show that sediment caliber is essentially the same for all three units, while there is a direct correlation between sinuosity and channel width. The High Deweyville unit has a sinuosity of 1.85, and average bankfull width of ca. 202 m while the Middle and Low Deweyville units have roughly the same sinuosities (ca. 2.92 and 2.89, respectively) and similar channel widths (ca. 267 and 255 m, respectively).

107 CONCLUSIONS

• Valley–fill deposits display contrasting facies between deposits of the last glacial

period and the more recent valley fill and can be attributed to changing allogenic

controls during their respective time periods of deposition.

• The falling stage and lowstand Deweyville allostratigraphic units represent

periods of lateral migration following brief periods of incision and abandonment

of an older depositional surface. Each unit is bound by a basal time transgressive

unconformity that represent ca. 30 kyrs, 9.5 kyrs, and 7 kyrs for the High

Deweyville, Middle Deweyville and Low Deweyville unit respectively and

capped by a paleosol that represent periods of little to no deposition and soil

formation. The basal unconformities trace from their respective channelbelt

margins and amalgamate on top of the soil profiles developed in older Deweyville

units and the surfaces of older deposits outside of the valley.

• The High, Middle, and Low Deweyville units were deposited during the falling

stage and lowstand of sea level during the last glacial period from ca. 65 – 32.5

ka, 32.5 –23 ka, and 23–16 ka respectively.

• The channelbelt deposits that rest on the basal valley–fill surface were deposited

contemporaneously with the formation of that surface, and correspond to periods

of valley widening through lateral migration of the active channel.

• Paleohydrological analyses from this study show that:

1) Previous interpretations of increased discharge during deposition of the

Deweyville units are inflated,

108 2) There is no basis to argue that channel forming discharges during deposition

of these units were significantly greater than those of the modern river

3) These results agree with existing paleoclimate data and model results

• Total sediment export from the incised valley over the course of an entire glacial

period is a relatively small sum, compared with the total sediment mass delivered

from the drainage basin. Periods of increased sediment flux from the incised

valley, due to formation of the incised valley itself, is stepwise, and correspond to

time periods of channelbelt deposition within the incised valley, and therefore

provide the mass–balance basis for coupling falling stage fluvial and deltaic

deposition

• The net export of sediment during valley creation can enhance total loads by 12–

34%, and offset reductions in sediment yields that might result from glacial–

period climate change.

• Using the net excavated amount of sediment removed during valley creation, and

the current sediment supply, the Trinity Valley would require ca. 7,000 years to

completely fill. Assuming that the present interglacial period will have a similar

duration to the previous interglacial (ca. >30 kyrs), the Trinity valley will fill

completely before the next major eustatic fall associated with the next glacial

period. Upon completion of valley–filling, the falling stage and lowstand

Deweyville deposits would account for ca. 44% of the overall valley–fill within

the Coastal Plain Trinity Valley.

• Sea–level fall during the last glacial period forced channel extension across the

shelf, yet the longitudinal profiles of the Deweyville units do not appear to have

109 reached equilibrium with contemporaneous sea level, which suggests that their

long profiles may be primarily controlled by the gradient of the shelf, and not

necessarily exact sea–level positions during their respective time periods of

deposition.

• The initial incision and abandonment of the Beaumont depositional surface,

followed by the formation of the High Deweyville deposits is directly linked sea–

level fall during OIS 4 to 3.

• The renewed incision that formed the High Deweyville terrace, and the

subsequent lateral migration that formed the Middle Deweyville deposits can not

be linked to sea–fall, and was likely caused by upstream climate–controls.

• Renewed incision, which formed the Middle Deweyville terrace followed by the

construction of the Low Deweyville deposits is the result of convergence of an

increase in effective moisture around 24 ka, coupled with response to the rapid

fall in sea level to the last glacial maximum position.

• Changing system boundaries due to the effects of lowered sea level and

interactions between longitudinal profile adjustment, coupled with upstream

controls on sediment supply and discharge, resulted in slope minimization

behavior of the glacial period Trinity system. An increase in sinuosity was the

attempt of the system to increase the reach scale flow resistance by decreasing

channel slope. The larger wavelength meanders, radii of curvature, and channel

widths associated with the Deweyville allostratigraphic units were the final result

of this process.

110 REFERENCES

Alford, J.K., and Holmes, J.C., 1985, Meander scars as evidence of major climate changes in southeast Louisiana: Annals of the Association of American Geographers, v. 75, p. 395–403.

Allen, G.P., Posamentier, H.W., 1993, Sequence stratigraphy and facies model of an incised valley fill: the Gironde Estuary, France: Journal of Sedimentary Petrology, 63, p. 378–391.

Anderson, J.B., Abdulah, K., Sarzalejo, S., Siringan, F., and Thomas, M.A., 1996, Late Quaternary sedimentation and high–resolution sequence stratigraphy of the east Texas shelf, in De Batist, M., and Jacobs, P. eds., Geology of Siliciclastic Shelf Seas: Geological Society Special Publication 117, p. 95–124.

Anderson, J.B., Rodriguez, A., Abdullah, K., Fillon, R.H., Banfield, L.A., McKeown, H.A., Wellner, J.S., 2004, Late Quaternary stratigraphic evolution of the northern Gulf of Mexico margin: a synthesis, in Anderson, J. B., Fillon, R. H., eds., Late Quaternary Stratigraphic Evolution of the Northern Gulf of Mexico Margin: SEPM, Special Publication 64, p. 1–23.

Ankjaergaard, C., and Murray, A.S., 2007, Total beta and gamma dose rates in trapped charge dating based on beta counting: Radiation Measurements, v. 42, p. 352– 359.

Aslan, A., and Blum, M.D., 1999, Contrasting styles of Holocene avulsion, Texas Gulf Coastal Plain, in Smith, N.D., Rogers, J.J. eds., Fluvial Sedimentology VI: International Association of Sedimentologists Special Publication, v. 28, p. 193– 209.

Aznar, M.C., Nathan, R., Murria, A.S., and Bøtter–Jensen, L., 2003, Determination of differential dose rates in a mixed beta and gamma field using shielded Al2O3: C: results of Monte Carlo modeling: Radiation Measurements, v. 37, p. 329–334.

Barton, D.C., 1930(a), Deltaic Coastal Plain of southeastern Texas: Geological Society of America, Bulletin, v. 41, i. 3, p. 359–382.

Barton, D.C., 1930(b), Surface geology of coastal southeast Texas: American Association of Petroleum Geologists Bulletin, v. 14, p. 1301–1320.

Beauboeuf, R.T., Abreu, V., Van Waggoner, J.C., 2003, Basin 4 of the Brazos–Trinity slope system, western Gulf of Mexico: The terminal portion of a Late Pleistocene lowstand system tract: in Roberts, H.H., Rosen, N.C., Fillon, R.H., and Anderson, J.B., eds., Shelf margin deltas and linked down slope petroleum systems: Global significance and future exploration potential: Proceedings of the

111 23rd Annual Research Conference, Gulf Coast Section SEPM Foundation, p. 45– 66.

Bureau of Economic Geology, 1992, Geologic Atlas of Texas, 1:250,000 Beaumont sheet.

Bureau of Economic Geology, 1982, Geologic Atlas of Texas, 1:250,000 Houston sheet.

Bureau of Economic Geology, 1996, River Basin Map of Texas.

Bernard, H.A., 1950, Quaternary geology of southeast Texas [unpublished PhD dissertation]: Louisiana State University, Baton Rouge, 164 p.

Bernard, H.A., LeBlanc, R.J., and Major, C.F., Jr., 1962, Recent and Pleistocene geology of southeast Texas, in Geology of the Gulf Coast and Central Texas, Guidebook of Excursion: Houston Geological Society, p. 175–224.

Blum, M.D., 1994, Genesis and architecture of incised valley fill sequences: a Late Quaternary example from the Colorado River, Gulf Coastal Plain of Texas, in Weimer, P., Posamentier, H.W., eds., Siliciclastic Sequence Stratigraphy: Recent Developments and Applications: American Association of Petroleum Geologists Memoir, 58, p. 259–283.

Blum, M.D., 2007, Large river systems and climate change, in Gupta, A., ed., Large Rivers: Geomorphology and Management, John Wiley and Sons, Ltd., p. 627– 659.

Blum, M.D., and Aslan, A., 2006, Signatures of climate vs. sea–level change within incised valley–fill successions: Quaternary examples from the Texas Gulf Coast: Sedimentary Geology, v. 190, p. 177–211.

Blum, M.D., and Price, D.M., 1998, Quaternary alluvial plain construction in response to interacting glacioeustatic and climatic controls, Texas Gulf Coastal Plain, in Shanley, K.W., McCabe, P.W., eds., Relative Role of Eustasy, Climate, and Tectonism in Continental Rocks: SEPM, Special Publication 59, p. 31–48.

Blum, M.D., and Törnqvist, T.E., 2000, Fluvial response to climate and sea–level change: a review and look forward: Sedimentology, v. 47, (Supplement): p. 1–48.

Blum, M.D., and Valastro, S., 1994, Late Quaternary sedimentation, lower Colorado River, Gulf Coastal Plain of Texas: Geological Society of America Bulletin, v. 106, p. 1002–1016.

Blum, M.D., Morton, R.A., and Durbin, J.M., 1995, "Deweyville" terraces and deposits of the Texas Gulf Coastal Plain: Gulf Coast Association of Geological Societies, Transactions, v. 45, p. 53–60.

112 Blum, M.D., 2003, Sediment supply to the shelf margin and beyond: Alluvial valley responses to sea–level change, in Roberts, H.H., Rosen, N.C., Fillon, R.H., and Anderson, J.B., eds., Shelf Margin Deltas and Linked Down Slope Petroleum Systems: Global significance and future exploration potential: Proceedings of the 23rd Annual Research Conference, Gulf Coast Section SEPM Foundation, p. 1– 26.

Blum, M.D., 2007, Glacial–interglacial scale fluvial responses, in Elias, S.A., ed., Encyclopedia of Quaternary Science, v. 2, p. 995–1010.

Blum, M., and Hattier-Womack, J., 2008–in press, Climate change, sea–level change, and fluvial sediment supply to deepwater depositional systems: a review. in Kneller, B., McCaffrey, W., and Martinsen, O. eds., External Controls on Deepwater Depositional Systems. SEPM, Special Publication.

Boyd, R., Dalrymple, R.W., Zaitlin, B.A., 1992, Classification of clastic coastal depostional environments, in Donoghue, J.F., Davis, R.A., Fletcher, C.H., Suter, J.R., eds., Quaternary Coastal Evolution: Sedimentary Geology, v. 80, p. 139– 150.

Brice, J.C., 1974, Evolution of meander loops: Geological Society of America Bulletin, v. 85, p. 581–586.

Brice, J.C., 1984, Planform properties of meandering rivers, in Elliott, C.M. ed., River Meandering: American Society of Civil Engineers, p. 1–15.

Bridge, J.S., 2003, Rivers and Floodplains: Forms, Processes, and Sedimentary Record: Oxford, UK, Blackwell Publishing, 491 p.

Bryant, V.M. Jr., and Holloway, R.G., 1985, A Late–Quaternary Paleoenvironmental Record of Texas: An overview of the pollen evidence, in Bryant, V.M., and Holloway, R.W., eds., Pollen records of Late Quaternary North American Sediments: Dallas, Texas: American Association of Stratigraphic Palynologists, p. 39–70.

Buffler, R.T., 1991, Introduction to central Gulf Coast geology: Early evolution of the Gulf of Mexico Basin, in Goldwaithe, D., ed., Introduction to Central Gulf Coast Geology: New Orleans Geological Society Publication 4, p. 1–15.

Castelltort, S., and van den Driessche, J., 2003, How plausible are high–frequency sediment supply–driven cycles in the stratigraphic record?: Sedimentary Geology, v. 157, p. 3–13.

Church, M., 2006, Bed material transport and the morphology of alluvial river Channels: Annual Review of Earth and Planetary Science, v. 34, p. 325–354.

113 COHMAP, 1988, Climatic changes of the last 18,000 years: observations and model simulations: Science, v. 241, p. 1043–1052.

Cooke, M.J., Stern, L.A., Banner, J.L., Mack, L.E., Stafford, T.W., Toomey, R.S., 2003, Precise timing and rate of massive late Quaternary soil denudation, Edwards Plateau, Texas: Geology, v. 31, p. 853–856.

Dade, W. B., and Friend, P.F., 1998, Grain–size, sediment–transport regime, and channel slope in alluvial rivers: The Journal of Geology, v. 106, p. 661–675.

Dalrymple, R.W., Zaitlin, B.A., and Boyd, R., 1992, Estuarine facies models: conceptual basis and stratigraphic implications: Journal of Sedimentary Petrology, v. 62, p. 1130–1146.

Delcourt, P.A., and Delcourt, H.R., 1981, Vegetation maps of Eastern North America: 40,000 yrs. B.P. to the present, in Romans, R., ed., Proceedings of the 1980 Geobotany Conference: Plenum, New York, p.123–166.

Delcourt, H.R., and Delcourt, P.A., 1985, Quaternary palynology and vegetational history of the southeastern United States, in Bryant, V.M., and Holloway, R.W., eds., Pollen records of Late Quaternary North American Sediments, Dallas, Texas: American Association of Stratigraphic Palynologists, p. 1–37.

Deussen, A., 1914, Geology and underground water resources of southeastern part of Texas Coastal Plain: U.S. Geological Survey, Water Supply Paper 335, 365 p.

Doering, J., 1935, Post–Fleming surface formations of coastal southeast Texas and south Louisiana: American Association of Petroleum Geologists, Bulletin, v. 19, p. 651–688.

Dunne, T., and Leopold, L.B., 1978, Water in environmental planning: New York, W. H. Freeman and Company, 818 p.

Durbin, J.M., 1999, Geomorphic response to Late Quaternary climate and sea–level change, lower Nueces River, Texas [unpublished PhD dissertation]: University of Nebraska–Lincoln, Lincoln, 242 p.

Durbin, J.M., Blum, M.D., and Price, D.M., 1997, Late Pleistocene Stratigraphy of the Lower Nueces River, Corpus Christi, Texas: Glacio–eustatic Influences on Valley–fill Architecture: Gulf Coast Association of Geological Societies,Transactions, v. 68, p. 119–130.

Dury, G.H., 1964, Principles of underfit streams: U.S. Geological Survey, Professional Paper 452-A, 66 p.

114 Dury, G.H., 1965, Theoretical implications of underfit streams: U.S. Geological Survey, Professional Paper 452-C, 40 p.

Eaton, B.C., and Church, M., 2004, A graded stream response relation for bed load– dominated streams: Journal of Geophysical research, 109(F03011). DOI:10.1029/2003JF000062.

Eaton, B.C., Church, M., and Davies, T.R.H., 2006, A conceptual model for meander initiation in bedload–dominated streams: Earth Surface Processes and Landforms, v. 31, p. 875–891.

Eaton, B.C., and Church, M., and Millar, R.G., 2004, Rational regime model of alluvial channel morphology and response: Earth Surface Processes and Landforms, v. 29, p. 511–529.

Gagliano, S.M., and Thom, B.G., 1967, Deweyville terrace, Gulf and Atlantic Coasts: Louisiana State University, Coastal Studies Bulletin, v. 1, p. 23–41.

Gagliano, S.M., 1991, Late Quaternary Deweyville interval superfloods. Gulf Coast Association of Geological Societies, Transactions, v. 41, p. 1524.

Galloway, W.E., 1981, Depositional architecture of Cenozoic Gulf Coastal Plain fluvial systems, in Ethridge, F.G., Flores, R.M., eds., Recent and Ancient Non–Marine Depositional Environments: SEPM, Special Publication, v. 31, p. 127–156.

Galloway, W.E., Ganey–Curray, P.E., Li, X., and Buffler, R.T., 2000, Cenozoic depositional history of the Gulf of Mexico basin: American Association of Petroleum Geologists, Bulletin, v. 84, no. 11, p. 1743–1774.

Ganopolski, A., Rahmstorf, S., Petoukhov, V., and Claussen, M., 1998, Simulation of modern and glacial climates with a coupled model of intermediate complexity: Nature, v. 391, p. 351–356.

Geological Society of America, 1991, Rock color chart, 8th printing, Boulder.

Hall, S.A., and Valastro, S., Jr., 1995, Grassland vegetation in the southern great plains during the last glacial maximum: Quaternary Research, v. 44, p. 237–245.

Hayes, C.W., and Kennedy, W., 1903, Oil fields of the Texas–Louisiana Gulf Coastal Plain: U.S. Geological Survey, Bulletin, 174 p.

Heinrich, P.V., 2006, Pleistocene and Holocene fluvial systems of the lower Pearl River, Mississippi and Louisiana, USA: Gulf Coast Association of Geological Societies, Transactions, v. 56, p. 267–278.

115 Holbrook, J., and Schumm, S.A., 1999, Geomorphic and sedimentary response of rivers to tectonic deformation; a brief review and critique of a tool for recognizing subtle epeirogenic deformation in modern and ancient settings, in Marshak, S., van der Pluijm, B.A., and Hamburger, M., eds., Tectonics of continental interiors: Tectonophysics, v. 305, i. 1–3, p. 287–306.

Howard, A.D., 1982, Equilibrium and time scales in geomorphology: Application to sand–bed alluvial streams: Earth Surface Processes and Landforms, v. 7, p. 303– 325.

Koch, P.L., Diffenbaugh, N.S., and Hoppe, K.A., 2004, The effects of Late Quaternary climate and ρCO2 change on C4 plant abundance in the south–central United States: Paleogeography, Paleoclimatology, Paleoecology, v. 207, p. 331–357.

Kutzbach, J.E., and Guetter, P.J., 1986, The influence of orbital parameters and surface boundary conditions on climate simulations for the past 18,000 years: Journal of the Atmospheric Sciences, v. 43, no. 16, p. 1726–1759.

Land, L.F., Moring, J.B., Van Metre, P.C., Reutter, D.C., Mahler, B.J., Shipp, A.A., and Ulery, R.L., 1998, Water quality in the Trinity River basin, Texas, 1992–95: U.S. Geological Survey, Circular 1171, p. 1–39.

Larkin, T.J., and Bomar, G.W., 1983, Climatic Atlas of Texas: Austin, Texas Department of Water Resources, 151 p.

LeBlanc, R.J., and Hodgson, W.D., 1959, Origin and development of the Texas shoreline, in Russell, R.J., ed., Proceedings 2nd Coastal Geography Conference: Coastal Studies Institute, Louisiana State University, p. 103–151.

Leopold, L.B., Wolman, M.G., 1956. River channel patterns: braided, meandering and straight: U.S. Geological Survey, Professional Paper 282-B, p. 39–84.

Mallarino, G., Beaubouef, R.T., Droxler, A.W., Abreu, V., and Labeyrie, L., 2006, Sea level influence on the nature and timing of a minibasin sedimentary fill (northwestern slope of the Gulf of Mexico): American Association of Petroleum Geologists, Bulletin, v. 90, p. 1089–1119.

Matthes, G.H., 1941, Basic aspects of stream meanders: American Geophysical Union, Transactions, p. 632–636.

Mattheus, C.R., Rodriguez, A.B., Greene, D.L., Jr., Simms, A.R., and Anderson, J.B., 2007, Control of upstream variables on incised–valley dimension: Journal of Sedimentary Research, v. 77, p. 213–224.

Mock, C.J., and Bartlein, P.J., 1995, Spatial variability of Late–Quaternary paleoclimates in the Western United States: Quaternary Research, v. 44, p. 425–433.

116 Morton, R.A., Blum, M.D., and White, W.A., 1996, Valley fills of incised coastal plain rivers, southeastern Texas: Gulf Coast Association of Geological Societies, Transactions, v. 46, p. 321–331.

Murray, A.S., and Olley, J.M., 2002, Precision and accuracy in the optically stimulated luminescence dating of sedimentary quartz: a status review: Geochronometria, v. 21, p. 1–16.

Murray, A.S., and Wintle, A.G., 2003, The single aliquot regenerative dose protocol: potential for improvements in reliability: Radiation Measurements, v. 37, p. 377– 381.

Musgrove, M., Banner, J.L., Mack, L.E., Combs, D.M., James, E.W., Cheng, H., and Edwards, R.L., 2001, Geochronology of late Pleistocene to Holocene speleothems from central Texas: implications for regional paleoclimate: Geological Society of America Bulletin, v. 113, p. 1532–1543.

Nathan, R.P., Thomas, P.J., Jain, M., Murray, A.S., and Rhodes, E.J., 2003, Environmental dose rate heterogeneity of beta radiation and its implications for luminescence dating: Monte Carlo modelling and experimental validation: Radiation Measurements, v. 37, p. 305–313.

Nordt, L.C., Boutton, T.W., Jacob, J.S., and Mandel, R.D., 2002, C4 plant productivity and climate–CO2 variations in south–central Texas during the Late Quaternary: Quaternary Research, v. 58, p. 182–188.

Nordt, L., 2003, Late Quaternary fluvial landscape evolution in desert grasslands of northern Chihuahua, Mexico: Geological Society of America, Bulletin, v. 115, i. 5, p. 596–606.

Osterkamp, W.R., Hedman, E.R., 1982, Perennial–streamflow characteristics related to channel geometry and sediment in Missouri River Basin: U.S. Geological Survey, Professional Paper 1242, p. 1–22.

Peltier, W.R., and Fairbanks, R.G., 2006, Global glacial ice volume and last glacial maximum duration from an extended Barbados sea level record: Quaternary Science Reviews, v. 25, p. 3322–3337.

Phillips, J.D., Slattery, M.C., and Musselman, Z.A., 2004, Dam–to–delta sediment inputs and storage in the lower Trinity river, Texas: Geomorphology, v. 62, p. 17–34.

Phillips, J.D., Slattery, M.C., and Musselman, Z.A., 2005, Channel adjustments of the lower Trinity River, Texas, downstream of Livingston Dam: Earth Surface Processes and Landforms v. 30, p. 1419–1439.

117 Phillips, J.D., and Slattery, M.C., 2006, Sediment storage, sea level, and sediment delivery to the ocean by coastal plain rivers: Progress in Physical Geography, v. 30, no. 4, p. 513–530.

Posamentier, H.W., Allen, G.P., James, D.P., and Tesson, M., 1992, Forced regressions in a sequence stratigraphic framework: examples, and exploration significance: American Association of Petroleum Geologists, Bulletin, v. 76, no. 11, p. 1687– 1709.

Roche, D.M., Dokken, T.M., Goosse, H., Renssen, H., and Weber, S.L., 2007, Climate of the last glacial maximum: sensitivity studies and model–data comparison with the LOVECLIM coupled model: Climate of the Past, v. 3, p. 205–224.

Rodriguez, A.B., Anderson, J.B., Simms, A.R., 2005, Terrace inundation as an autocyclic mechanism for parasequence formation: Galveston Estuary, Texas, U. S. A.: Journal of Sedimentary Research, v. 75, p. 608–620.

Rotnicki, K. 1983, Modelling past discharges of meandering rivers, in Gregory, K.J. ed., Background to Paleohydrology: A Perspective: Pittman Press, p. 321–346.

Rotnicki, K. 1991, Retrodiction of palaeodischarges of meandering and sinuos alluvial rivers and its palaeohydroclimate implications, in Starkel, L., Gregory, K.J., Thomas, J.B. eds., Temperate Paleohydrology, p. 431–471.

Schumm, S.A., 1960, The shape of alluvial channels in relation to sediment type: U.S. Geological Survey, Professional Paper 352-B, p. 17–30.

Schumm, S.A., 1963, Sinuosity of alluvial rivers on the great plains: Geological Society of America, Bulletin, v. 74, p. 1089–1100.

Schumm, S.A., 1968, River adjustment to altered hydrologic regimen–Murrumbidgee River and paleochannels, Australia: U.S. Geological Survey Professional Paper 598, p. 1–63.

Schumm, S.A., 1991, To interpret the Earth: Ten ways to be wrong: London, Cambridge University Press, 133 p.

Schumm, S.A., 1993, River response to base–level change: Implications for sequence stratigraphy: Geology, v. 101, p. 279–294.

Schumm, S.A., 2005, River variability and complexity: Cambridge, CambridgeUniversity Press, 220 p.

Schumm, S.A., and Khan, H.R., 1972, Experimental study of channel patterns: Geological Society of America, Bulletin, v. 83, i. 6, p. 1755–1770.

118 Shin, S.–I., Liu, Z., Otto–Bliesner, B., Brady, E.C., Kutzbach, J.E., and Harrison, S.P., 2003, A simulation of the last glacial maximum climate using the NCAR–CCSM: Climate Dynamics, v. 20, p. 127–151.

Simms, A.R., Anderson, J.B., Taha, Z.P., and Rodriguez, A.B., 2006, Overfilled versus underfilled incised valleys: examples from the Quaternary Gulf of Mexico, in Dalrymple, R., Leckie, D., and Tillman, R., eds., Incised Valleys in Time and Space: SEPM, Special Publication 85, p. 117–139.

Stokes, S., Ingram, S., Aitken, M.J., Sirocko, F., Anderson, R., and Leuschner, D., 2003, Alternative chronologies for Late Quaternary (last interglacial–Holocene) deep sea sediments via optical dating of silt–sized quartz: Quaternary Science Reviews, v. 22, p. 925–941.

Stute, M., Schlosser, P., Clark, J.F., and Broecker, W.S., 1992, Paleotemperatures in the southwestern United States derived from noble gases in ground water: Science, v. 256, p. 1000–1003.

Sylvia, D.A., and Galloway, W.E., 2006, Morphology and stratigraphy of the Late Quaternary lower Brazos valley: Implications for paleo–climate, discharge and sediment delivery: Sedimentary Geology, v. 190, p. 159–175.

Syvitski, J.P.M., and Milliman, J.D., 2007, Geology, geography, and humans battle for dominance over the delivery of fluvial sediment to the coastal ocean: Geology, v. 115, p. 1–19.

Syvitski, J.P.M., Peckham, S.D., Hilberman, R., and Mulder, T., 2003, Predicting the terrestrial flux of sediment to the global ocean: a planetary perspective: Sedimentary Geology, v. 162, p. 5–24.

Thomas, M.A., 1990, The impact of long–term and short–term sea level changes on the evolution of the Wisconsin–Holocene Trinity/Sabine incised valley system, Texas continental shelf [unpublished PhD dissertation]: Rice University, Houston, 247 p.

Thomas, M.A., Anderson, J.B., 1991, Late–Pleistocene sequence stratigraphy of the Texas continental shelf: relationship to δO18 curves: Gulf Coast Section Society of Economic Paleontologists and Mineralogists, Twelfth Annual Research Conference, Coastal Depositional Systems in the Gulf of Mexico, p. 265–270.

Thomas, M.A., Anderson, J.B., 1994, Sea–level controls on the facies architecture of the Trinity/Sabine incised–valley system, Texas continental shelf, in Dalrymple, R.W., Boyd, R., Zaitlin, B.A., eds., Incised-Valley Systems: Origin and Sedimentary Sequences: SEPM, Special Publication 51, p. 63–82.

119 Thomas, M.A., Anderson, J.B., 1989, Glacial eustatic controls on seismic sequences and parasequences of the Trinity/Sabine incised valley, Texas Continental Shelf: Gulf Coast Association of Geological Societies, Transactions, v. 39, p. 563–569.

Thomsen, K.J., Murray, A.S., and Bøtter–Jensen, L., 2005, Sources of variability in OSL dose measurements using single grains of quartz: Radiation Measurements, v. 39, p. 47–61.

Thomsen, K.J., Murray, A.S., Bøtter–Jensen, L., and Kinahan, J., 2007, Determination of burial dose in incompletely bleached fluvial samples using single grains of quartz: Radiation Measurements, v. 42, p. 370–379.

Toomey, R.S., Blum, M.D., Valastro, S., 1993, Late Quaternary climates and environments of the Edwards Plateau, Texas: Global and Planetary Change, v. 7, p. 299–320.

Tucker, G.E., and Slingerland, R., 1997, Drainage basin response to climate change: Water Resources Research, v. 33, i. 8, p. 2031–2047.

Ulery, R.L. et al, 1994, "Trinity River Basin Texas": Water Resources Bulletin, v. 39, p. 685–711.

United States Geological Survey, 2006. www.usgs.gov.

United States Geological Survey, 2007. www.usgs.gov.

United States Geological Survey, 2008. www.usgs.gov. van Hiejst, M.W.I.M., and Postma, G., 2001, Fluvial response to sea–level changes: a quantitative analogue, experimental approach: Basin Research, v. 13, p. 269–292.

Waelbroeck, C., Labeyrie, L., Michel, E., Duplessy, J.C., McManus, J.F., Lambeck, K., Balbona, E., Labracherie, M., 2002, Sea–level and deep water temperature changes derived from benthic foraminifera isotopic records: Quaternary Science Reviews v. 21, p. 295–305.

Williams, G.P., 1984, Paleohydrological methods and some examples from Swedish fluvial environments II-river meanders: Geografiska Annaler, v. 66-A 1-2, p. 89– 102.

Winker, C.D., 1979, Late Pleistocene fluvial–deltaic deposition Texas coastal plain and shelf [unpublished MA thesis], University of Texas at Austin, Austin, 186p.

Winker, C.D., 1991, Summary of the Quaternary framework, northern Gulf of Mexico, in Programs and Abstracts of the GCSSEPM Foundation, Twelfth Annual Research Conference, p. 280–284.

120 Winker, C.D., 1982, Cenozoic shelf margins, northwestern Gulf of Mexico: Gulf Coast Association of Geological Societies, Transactions 32, p. 427–448.

Wintle, A.G., and Murray, A.S., 2000, Quartz OSL: Effects of thermal treatment and their relevance to laboratory dating procedures: Radiation Measurements, v. 32, p. 387–400.

Wolman, M.G., Leopold, L.B., 1956, River flood plains: some observations on their formation: U.S. Geological Survey, Professional Paper 282-C, p. 87–109.

Wolman, M.G., Miller, J.P., 1959, Magnitude and frequency of forces in geomorphic processes: Geology, v. 68, no. 1, p. 54–74.

Womack, J.H., 2007, Empirical modeling of Late Quaternary sediment supply to the Gulf of Mexico shelf margin in a cool lowstand world [unpublished MS thesis] Louisiana State University, 95 p.

Wood, L.J., Ethridge, F.G., and Schumm, S.A., 1994, An experimental study of the influence of subaqueous shelf angles on Coastal Plain and shelf deposits, in Weimer, P., and Posamentier, H.W., eds., American Association of Petroleum Geologists, Memoir 58, Siliclastic Sequence Stratigraphy: Recent Developments and Applications, p. 381–391.

Zaitlin, B.A., Dalrymple, R.W., Boyd, R., 1994, The stratigraphic organization of incised-valley systems associated with relative sea–level change, in Incised Valley Systems: Origin and Sedimentary Sequences: SEPM, Special Publication v. 51, p. 45–60.

121 APPENDIX: TRINITY RIVER GUAGING STATION DATA

TRINITY RIVER AT LIBERTY Discharge rankings n=65 Qp Stage RI P Date (m3/s) (m) m (yrs) (%) 6/12/1940 852.43 7.77 48 1.38 72.73 12/17/1940 1741.68 8.26 16 4.13 24.24 5/12/1942 3228.48 8.96 2 33.00 3.03 6/24/1943 555.07 7.33 54 1.22 81.82 5/18/1944 1812.48 8.48 13 5.08 19.70 4/15/1945 2945.28 8.78 4 16.50 6.06 6/25/1946 1217.76 8.39 38 1.74 57.58 3/17/1947 1197.94 8.35 40 1.65 60.61 5/28/1948 756.14 7.96 51 1.29 77.27 4/24/1949 795.79 7.89 50 1.32 75.76 2/26/1950 1353.70 8.47 26 2.54 39.39 6/30/1951 419.14 6.18 62 1.06 93.94 4/25/1952 688.18 7.68 52 1.27 78.79 5/23/1953 1506.62 8.54 21 3.14 31.82 5/14/1954 523.92 6.90 57 1.16 86.36 4/15/1955 928.90 8.18 44 1.50 66.67 5/13/1956 402.14 6.36 64 1.03 96.97 5/12/1957 2494.99 8.92 6 11.00 9.09 5/21/1958 1665.22 8.64 19 3.47 28.79 4/20/1959 883.58 8.12 47 1.40 71.21 1/21/1960 906.24 8.13 46 1.43 69.70 1/17/1961 1483.97 8.62 22 3.00 33.33 5/4/1962 812.78 8.03 49 1.35 74.24 12/30/1962 566.40 7.30 53 1.25 80.30 4/28/1964 555.07 7.25 54 1.22 81.82 6/2/1965 1322.54 8.63 27 2.44 40.91 5/14/1966 1866.29 8.86 11 6.00 16.67 4/16/1967 305.86 5.61 65 1.02 98.48 4/14/1968 1322.54 8.72 27 2.44 40.91 5/27/1969 1483.97 8.74 22 3.00 33.33 3/21/1970 909.07 8.00 45 1.47 68.18 6/19/1973 1925.76 8.81 8 8.25 12.12 1/30/1974 1696.37 8.70 18 3.67 27.27 2/21/1975 1302.72 0.00 29 2.28 43.94 5/10/1976 1246.08 8.43 35 1.89 53.03 4/23/1977 1472.64 8.52 24 2.75 36.36 1/21/1978 495.60 6.28 58 1.14 87.88 4/24/1979 1699.20 8.69 17 3.88 25.76 5/21/1980 1274.40 8.37 32 2.06 48.48 6/20/1981 1260.24 8.46 34 1.94 51.52 5/20/1982 1240.42 8.44 36 1.83 54.55 5/26/1983 1268.74 8.46 33 2.00 50.00 3/21/1984 455.95 6.02 60 1.10 90.91 11/2/1984 1302.72 8.53 29 2.28 43.94 6/29/1986 1231.92 8.49 37 1.78 56.06

122

Qp Stage RI P Date (m3/s) (m) m (yrs) (%) 3/5/1987 931.73 8.03 43 1.53 65.15 1/4/1988 453.12 0.00 61 1.08 92.42 7/4/1989 1897.44 8.03 9 7.33 13.64 5/23/1990 3001.92 9.15 3 22.00 4.55 1/21/1991 2047.54 8.85 7 9.43 10.61 1/6/1992 2605.44 9.03 5 13.20 7.58 6/29/1993 1852.13 8.79 12 5.50 18.18 5/18/1994 957.22 8.01 42 1.57 63.64 10/18/1994 3823.20 9.45 1 66.00 1.52 12/19/1995 495.60 6.08 58 1.14 87.88 3/16/1997 1203.60 8.43 39 1.69 59.09 1/11/1998 1435.82 8.61 25 2.64 37.88 11/18/1998 1880.45 8.80 10 6.60 15.15 6/29/2000 413.47 5.93 63 1.05 95.45 3/18/2001 1755.84 8.73 15 4.40 22.73 12/31/2001 1022.35 8.07 41 1.61 62.12 11/10/2002 1803.98 8.77 14 4.71 21.21 7/3/2004 1566.10 8.60 20 3.30 30.30 11/27/2004 1302.72 8.35 29 2.28 43.94 4/1/2006 552.24 6.61 56 1.18 84.85

RI = (n+1)/m and P = (1/RI)*(100) where RI = Recurrence Interval, n = number of years of record, m = rank within time series with 1 being the largest and P = exceedence probability 3 -1 Qmax = the mean of the sum of the annual flood (86378.83 m s ) / the number of years of record (65 years) = 1,329 m3s-1 Q2= the flood that would statistically occur every 2 years (from curve below)= ca. 1269 m3s-1

Flood Frequency Curve

g 4500.00 ar 4000.00 y = 789.51Ln(x) + 561.23 2 sch i 3500.00 R = 0.9639

D 3000.00 ) m Q

s 2 u / 2500.00

3 Q m bf 2000.00 xi (m a 1500.00 M 1000.00 al

u 500.00 n

n 0.00 A 1.00 10.00 100.00 Recurrence Interval (yrs)

123 VITA

Matthew Gene Garvin was born in Dallas, Texas. He attended Garland High

School, graduating in 1998. He then attended Richland and Eastfield Colleges, earning the degree of Associate in Sciences in Business. During his final class at Eastfield, he met his wife, Jenny Garvin, on an earth sciences trip to the Grand Staircase. Following this, he attended Texas Tech University and majored in geology, receiving his Bachelor of Science in 2005. He then decided to continue at Louisiana State University under the support of the Applied Depositional Geosystems Fellowship, to pursue a Master of

Science degree in geology. His focus of study was in fluvial sedimentology and depositional systems in shallow marine settings. Upon graduation, he made his way back to Texas, to begin his career as a petroleum geoscientist at ExxonMobil Exploration

Company, in Houston, Texas.

124