ASSESSMENT AND PREDICTION OF STREAMBANK ...

ASSESSMENT AND PREDICTION OF STREAMBANK EROSION RATES IN A SOUTHEASTERN PLAINS ECOREGION WATERSHED IN MISSISSIPPI.

By John Jairo Ramirez Avila

A Dissertation Submitted to the Faculty of Mississippi State University in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in Water Resources in the Department of Civil and Environmental Engineering Mississippi State, Mississippi April 2011

Copyright by John Jairo Ramirez Avila 2011

ASSESSMENT AND PREDICTION OF STREAMBANK EROSION RATES IN A SOUTHEASTERN PLAINS ECOREGION WATERSHED IN MISSISSIPPI

By John Jairo Ramirez Avila

Approved: _________________________________ William H. McAnally Research Professor of Civil and Environmental Engineering (Major Professor)

__________________________________ James L. Martin Professor of Civil and Environmental Engineering (Co-Major Professor)

__________________________________ Eddy J. Langendoen Adjunct Professor of Civil and Environmental Engineering (Committee Member)

_________________________________ Ronald L. Bingner Adjunct Professor of Department of Plant and Soil Sciences (Committee Member)

__________________________________ James L. Martin Professor of Civil and Environmental Engineering Graduate Coordinator

__________________________________ Sarah A. Rajala Dean of Bagley College of Engineering

Name: John Jairo Ramirez Avila Date of Degree: April 29, 2011 Institution: Mississippi State University Major Field: Civil Engineering Major Professor: Dr. William H. McAnally and Dr. James L. Martin Title of study: ASSESSMENT AND PREDICTION OF STREAMBANK EROSION RATES IN A SOUTHEASTERN PLAINS ECOREGION WATERSHED IN MISSISSIPPI Pages in study: 297 Candidate for Degree of Doctor of Philosophy

The Town Creek Watershed (TCW) is a representative area of the Tombigbee River Basin and the Southeastern Plains Ecoregion in Mississippi. The principal channel and four main tributaries have been included for several years within the MS Section 303(d) list of waterbodies biologically impaired due to sediment. The TMDL developed for TCW recommended that streams located near cultivated lands, road crossings and construction activities are a priority for streambank and riparian buffer zone restoration and sediment loads reduction. Development of remedial measures and future BMPs within TCW for reducing water quality impairment and downstream dredging costs requires identification of sediment sources and loads currently transported within TCW. Streambank erosion processes were hypothesized to be an important mechanism driving sediment supply from TCW. The overall goal of this research was to identify mechanisms and the potential effects of streambank erosion processes and to quantify and model the magnitude and rates of these processes within TCW. Research goal and

specific aims were addressed in four substudies combining field reconnaissance and detailed data collection, laboratory analysis and computational modeling techniques. The first substudy involved a temporal and spatial analysis of observed suspended sediment transport rates, and determined the stage of channel evolution and identified streambank erosion as an important source of sediment supply for reaches in TCW. Streambank erosion contributions of up to 28.5 Mg per m of streambank were quantified in a second substudy monitoring and determining streambank erosion processes and factors within TCW. Results from a third substudy assessed predictions of the computational model CONCEPTS for time of occurrence and magnitude of streambank failures and top width retreat along a 270-m modeling reach. Empirical and analytical approaches used to estimate rates and depths of fluvial erosion were developed in a final substudy. The rate and depth of fluvial erosion were estimated as a function of hydraulic and hydrologic properties of flow events, vegetation on streambanks, flow induced forces and streambank geometry and soil properties. Reduction of suspended sediment loads should focus on attenuation of geomorphic processes and stabilization of reaches and agricultural lands adjacent to streambanks along incised headwaters within TCW.

Key words: streambank erosion, sediment transport, water quality, modeling, CONCEPTS model, Town Creek watershed, Southeastern Plains Ecoregion, Mississippi

ii

DEDICATION

I wish to dedicate this work with the biggest feeling of gratitude to my loving son Santiago “Santi”, to my beloved wife Sandra, to my adored parents Hernan y Amparo, to my siblings Wily and Judy and to my nephews Migue and Juancho. Sandra, my precious wife. I will be eternally grateful for your love, your support, your understanding and your presence in my life, and for giving me the greatest joy: Santi! My little child, I will always be in amazement and grateful of what God has created. Father and mother, your words of love, encouragement and motivation are always with me. Thanks for everything you have done for me, for everyone of us. Wily, my best friend. I will always feel proud to be the brother of such an incredible human being. Judy, my little sister, I love you. You have so many beautiful qualities, just believe in yourself. Migue and Juancho, Santi will be so happy playing with both of you all the time.

ii

ACKNOWLEDGEMENTS

This research was funded by the Northern Gulf Institute, a National Oceanic and Atmospheric Administration Cooperative Institute, under projects 07-MSU-05, Sediment Modeling in Mobile Basin, and 09-NGI-05, Sediment and Mercury Modeling. This research was also awarded the 2008 Kenneth E. Grant Research Scholarship from the Soil and Water Conservation Society and a 2010 Graduate Research Grant from the Geological Society of America. Graduate Research Assistantships were given by the Kelly Gene Cook Sr. Endowment from the Civil and Environmental Engineering Department (CEE) and the Geo-Resources Institute (GRI). A Graduate Research Fellowship was also given by the Bagley College of Engineering in 2009 and 2010. I would like to express my great appreciation to my advisors, Dr. William H. McAnally and Dr. James L. Martin, for all their expertise, guidance, confidence and assistance throughout the time I have performed my doctoral studies at Mississippi State University. Under their advisement, I have learned academic and research knowledge, which I consider to have improved my skills as a student, researcher and writer. I also have learned from their qualities and virtues as human beings, which have made the time I have been at MSU an unforgettable instance for my family and me, and have taught me invaluable life lessons.

iii

I am grateful for the expertise, guidance, training and assistance provided by my committee members, Dr. Eddy J. Langendoen and Dr. Ronald Bingner, researchers from the USDA-ARS NSL in Oxford, MS. I am truly appreciative for the support, direction and knowledge received from Dr. Langendoen during all the times I interacted with him. I have sincere appreciation for the assistance provided by numerous students, staff and faculty members from different departments of Mississippi State University and researchers from the USDA-ARS NSL. I want to thank Matt Moran, David Bassi, Carlos Moreno, Chris Hall, Justin Shaw, Trey Davis, Carter Pevey, Alina Young, Katherine Sloan and Kimberly Pevey for their collaboration during the strenuous field work and the laboratory analysis procedures; also thanks goes out to Allison Bates for her technical writing tutoring. I would also like to recognize the collaboration received from Mr. Joe Ivy, Mrs. Sandra Ortega, Mrs. Charlsie Halford, Mrs. Mary Box and Mrs. Jodie Womack, staff members from the Civil and Environmental Engineering Department. I want to thank Dr. Robert Wells, Dr. Glenn Wilson, Dr. Greg Hanson, Dr. Robert Thomas, Dr. Andew Simon and Mr. Tianyu Zhang from the USDA-ARS NSL, for their contributions to this research by training, teaching or by performing laboratory analyses. Special thanks to Dr. Dennis Truax for his support and collaboration as the Head Department of the Civil and Environmental Engineering Department. Thanks also to Dr. Chris Saucier, Dr. Jairo Diaz, Mr. H. King and Dr. Isaac Howard from the Civil and Environmental Engineering Department and to Dr. Michael Cox from the Department of Plant and Soil Sciences. Lastly, I would like to acknowledge the love and support of my family and friends. I am grateful to my wife Sandra, who worked with me on the field and performed iv

laboratory analysis to the collected samples. Without her support, love and understanding it would have been really hard to reach this goal. Thanks to my parents, siblings and nephews for their love and support even from the distance. Very special thanks to Todd and Sandra Hawblitzel and their family for their help and motivation. Thanks to Eng. Carlos Gonzalez and Eng. Edgar Almansa for teaching me basic fundamentals of soil and water sciences. Thanks to Dr. Gustavo Martinez, Dr. Fernando Vega, Dr. David Sotomayor and Dr. Miguel Muñoz for encouraging Sandra and me to keep climbing. Thanks to each person who has believed in me and has befriended me.

v

TABLE OF CONTENTS

DEDICATION……………………………………………………………………

ii

ACKNOWLEDGEMENTS……………………………………………………..

iii

LIST OF TABLES………………………………………………………………..

ix

LIST OF FIGURES………………………………………………………………

xii

LIST OF SYMBOLS……………………………………………………………..

xxiii

CHAPTER I.

II.

III.

IV.

INTRODUCTION………………………………………………...........

1

Goals and Objectives…………………………………………………...

5

STREAMBANK EROSION……………………………………………

7

Streambank Erosion Processes………………………………………… Hydraulic Action Process……………………………………………… Gravitational Mass Failure Process…………………………………….

8 11 14

STUDY AREA…………………………………………………………

25

Tombigbee River Basin………………………………………………... Town Creek Watershed………………………………………………... Physiography and Geology………………………………………. Watershed Physiographic Characterization……………………… Land Use…………………………………………………………. Climate…………………………………………………………… Streamflow……………………………………………………….. Water Quality Conditions………………………………………...

25 29 32 34 39 41 43 46

SUSPENDED SEDIMENT TRANSPORT WITHIN THE TOWN CREEK WATERSHED……………………………………….........

49

vi

V.

VI.

Introduction……………………………………………………………. Methods………………………………………………………………... Spatial Variation of Suspended Sediment within the Town Creek Watershed……………………………………………....... Suspended Sediment Load Estimations by Using the Suspended Sediment Transport Rating Relation……………………... Analysis of Suspended Sediment Transport Rating Relations…... Results and Discussion………………………………………………… Spatial Variation of Suspended Sediment within the Town Creek Watershed………………………………………………... Suspended Sediment Load Estimations by Using the Suspended Sediment Transport Rating Relation…………..………… Analysis of Suspended Sediment Transport Rating Relations…... Conclusions…………………………………………………………….. Recommendations………………………………………………………

50 53

ASSESSMENT OF STREAMBANK EROSION RATES…………….

86

Introduction…………………………………………………………….. Intermediate Timescale Techniques……………………………... Short-Time Scales Techniques…………………………………... Cross Section Survey at the Yonaba Creek……………………………. Methods………………………………………………………….. Results……………………………………………………………. Transects 0 m and 20 m (0 m to 75 m)…………………... Transects 90 m, 135 m, 140m and 160 m (75 m to 185 m) Transects 210 m and 270 m (185 m to 270 m)…………... Streambank Erosion Rates Assessment………………….. Erosion Pins at the Town Creek……………………………………….. Methods………………………………………………………….. Results……………………………………………………………. Temporal Variability in Streambank Response………….. Mass Movements………………………...………………. Effects of Vegetation on Streambank Erosion Response... Conclusions……………………………………………………………..

86 88 89 92 92 98 107 111 118 121 128 128 134 142 148 149 154

MODELING STREAMBANK EROSION RATES……………………

155

Introduction…………………………………………………………….. CONCEPTS Model Description……………………………………….. CONCEPTS Java Graphic User Interface (GUI)………………... Methods………………………………………………………………... Model Setup……………………………………………………… Channel Geometry……………………………………….. Streambank Material Physical Properties………………... Discharge………………………………………………… vii

155 159 163 172 172 174 177 185

53 55 56 57 57 65 70 82 85

Modeling Assumptions, Calibration, Validation and Sensitivity Analysis…..………………………….. Results………………………………………………………………….. Factor of Safety and Streambank Instability……………………... Streambank Top Width Retreat………………………………….. Streambank Sediment Yield……………………………………... Sensitivity Analysis……………………………………………… Conclusions……………………………………………………………. VII.

PREDICTION OF FLUVIAL EROSION ON STREAMBANKS…….. Introduction…………………………………………………………….. Erosion Susceptibility of Streambank Soils……………………… Determination of Critical Shear Stress and Soil Erodibility……... Near Streambank Tractive Force………………………………… Development of an Approach Based on Field Observations…………... Development of an Analytical Approach to Estimate Fluvial Erosion... Analysis of Forces Acting on a Sediment Particle on a Streambank………………………………………………. Conceptual Analysis……………………………………………... Analytical Approach Application………………………………... Results…………………………………………………… Sensitivity Analysis……………………………………… Conclusions……………………………………………………………..

VIII.

186 187 188 191 207 211 218 220 220 225 229 234 237 241 241 244 251 258 262 267

SUMMARY AND CONCLUSIONS………………………………….. Research Contributions………………………………………………… Recommendations for Future Research………………………………...

BIBLIOGRAPHY………………………………………………………………...

270 272 273 275

APPENDIX A. GEOMORPHOLOGIC CHANGE OF THE STUDIED REACH ON THE YONABA CREEK IN BLUE SPRINGS, MS………………….

288

B. STEPWISE ANALYSIS TO DETERMINE STREAMBANK SOIL CRITICAL SHEAR STRESS BY CORRELATION OF 20 STREAMBANK SOIL PARAMETERS………………...……….

296

viii

LIST OF TABLES

2.1

Streambank erosion mechanisms observed along the Town Creek watershed…………………………………………………………

10

3.1

Land use distribution at Town Creek watershed…………………………

40

4.1

Mean flow discharge, suspended sediment concentration, load and yield from biweekly grab sampling between May 2008 and May 2009 at 7 stations along the principal channel in Town Creek and 4 stations on tributaries………………………………......................

59

Comparison of suspended sediment transport relations at the USGS gauging station 02436500 (Town Creek near Nettleton, MS) determined by different authors showing calculated load and yield estimated Q1.5=636 m3s-1 and drainage area A=1606 km2….

72

Annual suspended sediment transport relations at the USGS gauging station 02436500 (Town Creek near Nettleton, MS) showing estimated load and yield at the determined Q1.5=636 m3s-1 and drainage area A=1606 km2………...…………………..…………

74

Suspended sediment transport relations at USGS gauging station 02436500 (Town Creek near Nettleton, MS) determined for different periods between 1981 and 1995 and May 2008 to April 2009, showing calculated load and yield at the Q1.5=636 m3s-1 and drainage area A=1606 km2……..………………....................

81

Thalweg, streambanks height and streambank slope of transects along a 270 m reach on the Yonaba Creek near the bridge at the MS Road 9…………………………………………………………….

98

Comparison of streambank area change in transects along a 270 m reach on the Yonaba Creek near the bridge at the Road 9…..………….

126

Comparison of streambank widening and streambed erosion depth in transects along a 270 m reach on the Yonaba Creek between January 29, 2009 and March 18, 2010……………........................

127

4.2

4.3

4.4

5.1

5.2 5.3

ix

5.4

Mean streambank material bulk density and streambank material erosion or deposition along a 270 m reach on the Yonaba Creek between January 31, 2009 and March 18, 2010……………………..……

127

5.5

Dates of monitoring of pin erosion depths for two sites along the Town Creek…..…………………………………………………………

134

5.6

Erosion or deposition depth (mm) after 1 year at two monitoring sites along the Town Creek……..…………………………………….

136

6.1

Sediment size classes and sediment transport equations used in the CONCEPTS model…..…………………………………………..

160

6.2

Streambank material stratigraphy of seven representative streambank profiles and streambed of the Yonaba Creek study site……..……

183

6.3

Roughness values assigned to streambed and streambank sections of each cross section along the simulated reach…..…………………

184

6.4

Observed and simulated changes in top width occurred at the left streambank on two sections along the Yonaba Creek between February and July 2009……..……………………………………

192

Observed and simulated streambank erosion yield along a 270 m reach on the Yonaba Creek in Blue Springs, MS between February 22, 2009 and March 18, 2010…………….…………………………

210

6.6

Sensitivity analysis results for six CONCEPTS model parameters for simulation of the Yonaba Creek in Blue Springs, MS…..………

212

6.7

Sensitivity analysis results for CONCEPTS model parameters for simulation of total sediment yield on six sections along the Yonaba Creek in Blue Springs, MS……..………………………

216

Relative efficiencies (Er) for CONCEPTS model parameters for simulation of total sediment yield on six sections along the Yonaba Creek in Blue Springs, MS…………..…………………

217

7.1

Critical shear stress coefficients to account for vegetation………………

228

7.2

Streambank characteristics and streambank soil properties used to simulate fluvial streambank erosion depths on two streambanks along the Town Creek……………………………………………

253

Observed and simulated depths of fluvial erosion in two streambanks along the Town Creek……..……………………………………

261

6.5

6.8

7.3

x

7.4 7.5 7.6 7.7

Sensitivity analysis results for six parameters for simulation of fluvial erosion on the streambank #1……..…………………………….

263

Sensitivity analysis results for six parameters for simulation of fluvial erosion on the streambank #2…..……………………………….

264

Relative efficiencies (Er) for six parameters for simulation of fluvial erosion on the streambank #1…………………..……………….

265

Relative efficiencies (Er) for six parameters for simulation of fluvial erosion on the streambank #2…..……………………………….

266

xi

LIST OF FIGURES

2.1

Hydraulic or fluvial entrainment notch examples along the Town Creek..

12

2.2

Undercutting or scouring along tributaries within the Town Creek watershed ……………………………………………………….

13

Basal cleanout process after a streambank failure in a northern headwater of the Town Creek watershed…………………………

14

2.4

Planar failure on an incised headwater of the Town Creek………………

16

2.5

Rotational failure or slump failure on tributaries and principal channel of the Town Creek…………………………………………………...

17

2.6

Slab failure or block failure on incised headwaters of the Town Creek….

18

2.7

Cantilever failure on tributaries of the Town Creek……………………...

20

2.8

Wet earthflow failure and path of runoff flow on streambanks on the Yonaba Creek…………..………………………………………...

21

2.9

Popout failure on incised headwaters of the Town Creek………………..

22

2.10

Piping failure on incised headwaters of the Town Creek………………...

23

2.11

Soil fall on incised headwaters of the Town Creek………………………

24

3.1

Location of the Tombigbee River basin, the Mobile River basin and the Ecoregion 65 in Mississippi and Alabama……………...………..

27

3.2

Location of Town Creek watershed (Aerial Picture: MS GIS Council)….

31

3.3

(a) Physiographic region and (b) geological formations within the Town Creek watershed in Mississippi…..………………………………

35

3.4

Town Creek watershed elevation…………………………………………

37

2.3

xii

3.5

Profile and slope of the principal channel in Town Creek watershed (from Yonaba Creek at Road MS-9 to Town Creek outlet)……

37

(a) Gully erosion and (b) streambank erosion processes near agricultural areas at northern headwaters, and (c) transitional areas upstream of the Natchez Trace in the Town Creek watershed…………….

38

3.7

Channel morphology and streambank sediment deposition along the middle 20 km of Town Creek……………………………………

39

3.8

Land use distribution within Town Creek Watershed, Mississippi………

40

3.9

Gully erosion near streambanks along agricultural lands with no riparian vegetation…………………………………………………………

41

Average precipitation amount for different areas within the Town Creek watershed (Data source: The Weather Channel)…………………

42

Spatial distribution of the average annual precipitation within Town Creek watershed……………..……………………………………

42

Average Temperature for different areas within the Town Creek watershed (Source: The Weather Channel)………………………

43

Annual instantaneous peak discharges at two stations along Town Creek watershed. Data obtained from USGS and AHPS - NOAA Web site………………………………………………………………...

44

Recurrence Intervals for peak discharges at two stations along Town Creek watershed. Data obtained from USGS and AHPS - NOAA Web site………………..…………………………………………

45

Mean monthly discharge for Town Creek watershed at two stations along the principal channel. Data obtained from USGS and AHPS - NOAA Web site…………………………………………

45

Impaired waterbodies within Town Creek watershed (Section 303(d) List Fact Sheet for Watershed. Effective Listing Cycle: 2010)…..

47

Variation in Suspended Sediment Transport Efficiency during the course of Channel Evolution (Adapted from Simon, 1989)……………...

53

Monitoring stations along principal channel within Town Creek watershed……………………..…………………………………..

55

Variation of relative flow discharge, relative suspended sediment loads and relative suspended sediment yields along the Town Creek….

61

3.6

3.10 3.11 3.12 3.13

3.14

3.15

3.16 4.1 4.2 4.3

xiii

4.4

(a) Gully erosion and (b) streambank erosion processes near agricultural areas at northern headwaters, and (c) transitional areas upstream of the Natchez Trace in the Town Creek watershed……………...

62

(a) Natural pool at Mud Creek and (b) Sediment trap established at Town Creek both about 20 m downstream of the Main St. stations. (c) Natural pool 50 m upstream of the Eason Blvd station……………………………………………………………..

62

4.6

Channel morphology and streambank sediment deposition along the middle 20 km of Town Creek……………………..……………...

63

4.7

Western tributaries of the Town Creek watershed evidencing widening and degradation processes………………..………………………

64

4.8

Western tributary of the Town Creek watershed evidencing widening and aggradation processes…………………..……………………

64

4.9

Established A-Jax structures to stabilize streambank toe along the Chiwapa Creek………………..…………………………………..

65

4.10

Established A-Jax structures favoring fluvial streambank erosion and gravitational failures along the Chiwapa Creek…………………..

65

4.11

Mean daily flow and suspended sediment load at the USGS station on Town Creek near Nettleton, MS………………………………….

66

4.12

Relation between sediment load (Qs) and instantaneous flow discharge (Q) for 128 gauging station records at the USGS gauging station 02436500…………………………………………………………

68

(a) Cumulative suspended sediment load and (b) daily suspended sediment load at the USGS gauging station 02436500 on Town Creek estimated by integrating the suspended sediment transport curve from May 1, 2008 to May 15, 2009………………………..

69

Relation between sediment load (Qs) and instantaneous flow discharge (Q) for 1401 gauging station records at the USGS gauging station 02436500 (Town Creek near Nettleton, MS)……………..

76

Annual suspended sediment loads at USGS gauging station 02436500 (Town Creek near Nettleton, MS) determined by integrating two different suspended sediment rating curve: (1) suspended sediment rating curve based on the entire period of record (Equation 4.3), and (2) annual suspended sediment rating curves.

77

4.5

4.13

4.14

4.15

xiv

4.16

4.17

4.18

4.19

4.20

5.1 5.2 5.3

5.4

5.5 5.6

Rate of increase in load (rating curve slope parameter) and loads at low/base flow (rating curve intercept parameter) obtained for each year with recorded data at the USGS gauging station 02436500 (Town Creek near Nettleton, MS)…………………….

77

Relation between the annual suspended sediment load at low or base flow and the annual rate of increase in load at USGS gauging station 02436500 (Town Creek near Nettleton, MS)…………….

78

Clustered relation between annual average instantaneous flow (m3s-1) and the rate of increase in load obtained for each year with recorded data at the USGS gauging station 02436500 (Town Creek near Nettleton, MS)……………………………………….

78

Variation in suspended sediment transport efficiency (rating curve exponent) for different periods between 1981 and 1995 and May 2008 to April 2009 in the Town Creek watershed………………..

79

Relations between suspended sediment load (Qs) and instantaneous flow (Q) for different time periods (clusters) at USGS gauging station 02436500 (Town Creek near Nettleton, MS)…………………….

82

Aerial view of the reach on the Yonaba Creek near the bridge at the Road 9………………………………………………………..…...

95

Aerial view from upstream to downstream of the reach on the Yonaba Creek near the bridge at the Road 9……………………………....

96

View of the different transects along the 270 m reach on the Yonaba Creek near the bridge at the Road 9 in Blue Springs, MS. (a) Transects 0 m and 20 m, (b and c) transects 90 m to 165 m and (d) transects 210 m and 270 m…………………………………...

96

(a) Streambank erosion along incised unstable streambanks and (b) gully erosion processes from agricultural lands near streambanks with limited riparian vegetation at the northern area of the Town Creek watershed………………………………….……………..

97

Cross section surveying by using (a and b) Total Station Positioning System and (c and d) Real Time Kinematic (RTK) GPS System.

97

Channel cross sectional surveys at transect 0m (0-5 m). The orientation of the survey is left streambank to right streambank looking upstream………………………….………………………………

99

xv

5.7

5.8

5.9

5.10

5.11

5.12

5.13

5.14

5.15

Channel cross sectional surveys at transect 20m (5-75 m). The orientation of the survey is left streambank to right streambank looking upstream………………………………………………...

100

Channel cross sectional surveys at transect 90m (75-105 m). The orientation of the survey is left streambank to right streambank looking upstream……………………………………………...

101

Channel cross sectional surveys at transect 135m (105-135 m). The orientation of the survey is left streambank to right streambank looking upstream…………………………………………………

102

Channel cross sectional surveys at transect 140m (135-150 m). The orientation of the survey is left streambank to right streambank looking upstream………………………………………………….

103

Channel cross sectional surveys at transect 160m (150-185 m). The orientation of the survey is left streambank to right streambank looking upstream………………………...………………………..

104

Channel cross sectional surveys at transect 210m (185-235 m). The orientation of the survey is left streambank to right streambank looking upstream………………………………………………….

105

Channel cross sectional surveys at transect 270m (235-270 m). The orientation of the survey is left streambank to right streambank looking upstream……………………….…………………………

106

(a, b and c) Downstream view of transects 0 m and 20 m; (d) streambank deposition on transect 20 m (between abscises 5 m and 75 m) and (e and f) confluence of the Spout Spring Branch and the Hall Creek upstream of the 0 m segment along the 270 m reach on the Yonaba Creek near the bridge at the Road 9……………………..

109

(a, b, c) Downstream and (d) upstream view of the bendway segment (transects 90 m, 130 m, 140 and 165 m); (e) Debris accumulation at the beginning of the bendway segment; (f) Deep water streamflow near the left streambank along transect 90 m; and (g) Deposited loose sandy material and streambed sandy bar caused by winter stormflow events during December 2008 and January 2009 along the bendway segment ………………………………..

114

xvi

5.16

(a) Snow remains on streambanks surface on May 3, 2009; (b and c) apparently action of subaerial processes on unvegetated streambanks surface; and (d) Effect of raindrops impact, surface runoff on streambank (small rills), seepage and streamflow on a streambank previously affected by subaerial processes (freezethaw cycling)……………………………………………………...

115

Difference of erosion degree for two different transects (130 m and 145 m) along an earth levee segment armored with grass…………….

115

5.18

Occurrence of seepage erosion on the bendway segment streambanks…..

117

5.19

(a) Effect of the metal pipe on the across right side streambank; and (b) sandy loose material deposition on streambank toe………………

117

View of transects 210 and 270 m (between abscises 185 m and 235 m) along the 270 m reach on the Yonaba Creek near the bridge at the Road 9……………………….…….………………………...

120

5.21

Lower streambank undercutting along the right side of the straight segment after the bend…………………….……………………..

121

5.22

(a) Streambed outcrop of Selma Shale (gray shale) along the Yonaba Creek (b) Transition between sandy alluvial streambed and streambed with outcrop along the Yonaba Creek………………..

125

Aerial view of the pin erosion plots location along the Town Creek near the bridge at Brewer Rd. (MS 521) (Source : www.bing.com/maps).....................................................................

130

Aerial view of the pin erosion plots location on the Yonaba/Town Creek near the bridge at the Natchez Trace Parkway (Source : www.bing.com/maps).....................................................................

131

Initial condition of the different plots located along the Town Creek. a), b) and c) plots 1, 2 and 3 along the left streambank in site #1; d) and e) plots 4, and 5 along the right streambank in site #1; f) plot 7 along the left streambank in site #2; g) and h) plot 8 along the right streambank in site #2…………………………………..……

132

Cumulative mean depth of erosion (+) or deposition (-) after stormflow events for plots #1 (up) and #4 (down) along 13 monitoring dates in Town Creek……………………………………………………

137

5.17

5.20

5.23

5.24

5.25

5.26

xvii

5.27

Cumulative mean depth of erosion (+) or deposition (-) after stormflow events for plots #2 (up) and #5 (down) along 13 monitoring dates in Town Creek……………………..……………………………..

138

Cumulative mean depth of erosion (+) or deposition (-) after stormflow events for plots #3 (up) and #6 (down) along 13 monitoring dates in Town Creek……………..……………………………………..

139

Cumulative mean depth of erosion (+) or deposition (-) after stormflow events for plots #7 (up) and #8 (down) along 13 monitoring dates in Town Creek…………………………………………….……...

140

Changes on streambank surface conditions (vegetation and surface material) induced by a peak flow event which occurred in December 2008 at both sites……………………………….……..

141

Failure at the top of the right streambanks on site #1 induced after the occurrence of a bankfull event and successive high stormflow events in December 2008 and January 2009…………………..…

142

5.32

Annual change on plots #1 (up) and #4 (down) on the Town Creek (July 2008 to June 2009)…………………………………..……………

144

5.33

Annual change on plots #2 (up) and #5 (down) on the Town Creek (July 2008 to June 2009)………………………..………………………

145

5.34

Annual change on plots #3 (up) and #6 (down) on the Town Creek (July 2008 to June 2009)………………………..………………………

146

5.35

Annual change on plots #7 (up) and #8 (down) on the Town Creek (July 2008 to June 2009)………………………………..………………

147

5.36

Soil fall occurred at the lower part of the right streambanks in site #1 during Event #5 (up) and during evidence of a freeze thaw cycle during stormflow events between December 10, 2008 and January 16, 2009 (down)…………………………………………

151

5.37

Seepage erosion occurring at the low part of the streambank in plots #5 and #6………………………..….………………………………..

152

5.38

Mass movement along plots #5 and #6 (up) and difference in morphology at the lower part of plot #4 (down)………………….

152

5.28

5.29

5.30

5.31

xviii

5.39

Deposition and new distribution on streambank vegetation observed in the different plots during the monitoring event #5……..………...

153

Channel changes, loss of agricultural land and water quality impact caused by erosive processes along the Yonaba Creek in Blue Springs, MS…………………………………..…………………..

158

Jet test device installed on toe of streambank to determine the streambank soil critical shear stress and erodibility……………...

161

Summary of forces on a failure streambank used in CONCEPTS streambank instability analysis…………………………………...

163

6.4

View of the Java interface for the CONCEPTS model…………………..

164

6.5

View of the material sub-tab in the CONCEPTS GUI, showing the characterization and the plot of the particle size distribution for a specific streambank soil type……………………………………..

165

View of the cross section sub-tab in the CONCEPTS GUI, showing the characteristics and the plot of the geometry and the defined internal boundaries for a specific cross section…………………..

166

View of the reaches sub-tab in the CONCEPTS GUI, showing the cross sections included in a specific reach……………………………..

167

6.8

View of the channel models tab in the CONCEPTS GUI……………….

168

6.9

View of the run control data sub-tab in the CONCEPTS GUI…………..

169

6.10

View of the output option sub-tab in the CONCEPTS GUI……………..

170

6.11

View of the scenarios sub-tab in the CONCEPTS GUI, showing the appearance of the console while running a specific simulation scenario…………………………………………………………..

171

Yonaba Creek study site. Plan view showing the location of surveyed cross sections……………………………………………..………

173

a) Streambank gully erosion process and b) berm on streambank toe along the third segment of the modeling reach…………………..

173

6.14

Cross sections geometry input for the CONCEPTS model………………

175

6.15

Representative streambank soil profiles observed along the modeling reach at Yonaba Creek in Blue Springs, MS……………………..

179

6.1

6.2 6.3

6.6

6.7

6.12 6.13

xix

6.16

Determination of friction angle (’) based on the occurrence of a streambank failure………………………………………………..

180

6.17

View of the a) jet test device setup, b) jet of water applied on the streambank material, and c) scour generated on the streambank material after a test………………………………………………..

182

Relationship between erodibility coefficient (kd) and critical shear stress (c) for streambank fine grained materials based on jet tests from the Yonaba in Mississippi. Red dots represent vary sandy loose material deposited on streambanks toes and were not included in the correlation analysis…………………………………………...

182

6.19

Discharge observed at the upstream cross section (X=0 m) of the 270 m modeling reach along the Yonaba Creek in Blue Springs, MS…..

185

6.20

Observed and simulated streamflow discharge for a 270 m modeling reach along the Yonaba Creek in Blue Springs, MS……………..

188

6.21

Predicted Factor of Safety (FS) for the left streambanks on a bendway along a 270 m reach on the Yonaba Creek, MS, between February 22, 2009, and March 18, 2010………………………….

190

6.22

Predicted changes on the Factor of Safety (FS) due to changes in streamflow depth…………………………..……………………..

191

6.23

Comparison between simulated and observed changes in section #1……

193

6.24

Comparison between simulated and observed changes in section #2……

194

6.25

Comparison between simulated and observed changes in section #3 between February 2009 and June 2009…………………..……….

195

6.26

Comparison between simulated and observed changes in section #3 between July 2009 and March 2010…………………………..….

196

6.27

Comparison between simulated and observed changes in section #4 between February 2009 and June 2009…………………………..

197

6.28

Comparison between simulated and observed changes on section #4 between July 2009 and March 2010……………………………..

198

6.29

Comparison between simulated and observed changes in section #5 between February 2009 and June 2009…………………………..

199

6.30

Comparison between simulated and observed changes on section #5 between July 2009 and March 2010……………………...............

200

6.18

xx

6.31 6.32 6.33 6.34 6.35 6.36 6.37

6.38 7.1

7.2 7.3

7.4

7.5

7.6

Comparison between simulated and observed changes in section #6 between February 2009 and June 2009………………..………….

201

Comparison between simulated and observed changes on section #6 between July 2009 and March 2010………………………..…….

202

Comparison between simulated and observed changes in section #7 between February 2009 and June 2009………………………..….

203

Comparison between simulated and observed changes on section #7 between July 2009 and March 2010……………………..……….

204

Comparison between simulated and observed changes in section #8 between February 2009 and June 2009………………………..….

205

Comparison between simulated and observed changes on section #8 between July 2009 and March 2010…………………………..….

206

Streambank erosion yield simulated by CONCEPTS and observed by cross section survey along 8 sections on the Yonaba Creek in Blue Springs, MS…………………………..……………………..

207

Comparison between actual and simulated planar failure, streambed deposition and erosion of deposited material on streambank toe..

211

Jet test device used to determine erosion parameters and sketches of the jet test device and the jet testing device nozzle illustrating jet scour parameters (adapted from Hanson and Simon (2001))….…

231

Erosion rate of a cohesive streambank as a function of applied shear stress () determined by jet testing………………..………………

231

Effect of soil chemistry (Sodium Adsorption Ratio (SAR), salt concentration and dielectric dispersion) on the magnitude of critical shear stress (Adapted from Langendoen (2000))………...

232

Wetting time under specific peak stormflow events for different elevations on streambanks along the middle 20 km of the principal channel of Town Creek……………..………………….

238

Effect of the elevation of streambanks with trees and roots on parameters a and b included in the relationships between peak stormflow and wetting time…………………………..…………………………..

239

Relationship between fluvial streambank erosion depth and wetting time for three specific elevations on streambanks on the principal channel along Town Creek……………………………………….

240

xxi

7.7

Forces acting on a soil particle on a streambank…………………………

243

7.8

Gravitational force diagram on a streambank soil particle……………….

249

7.9

Flow depth observed for streambanks #1 and #2 on the principal channel of Town Creek……………………..……………………………..

254

Distribution of average velocity for different flow depths used to simulate fluvial erosion on streambanks #1 and #2………………

254

7.11

Temporal change of the average shear stress () and the shear stress acting at 0.3 m of elevation on the streambank (sb)……………..

255

7.12

Temporal change of the average shear stress () and the shear stress acting at 1.15 m of elevation on the streambank (sb)………..…...

255

7.13

Temporal change of the average shear stress () and the shear stress acting at 3.3 m of elevation on the streambank (sb)…………...…

256

7.14

Initial results for simulated vs observed streambank fluvial erosion depths (mm) on two streambanks along Town Creek, MS…….....

258

7.15

Plot of simulated vs observed streambank fluvial erosion depths (mm) on two streambanks along Town Creek, MS……………..………

259

7.16

Cumulative depth of fluvial erosion along thirteen flow events for three different elevations on two streambanks………………..………..

260

7.10

xxii

LIST OF SYMBOLS

Symbol Description

Units

a

Regression coefficient (Load at a low flow or base flow discharge)

A

Area

4.1 m2 or km2

AC

Areal coverage

aF

Coefficient used to determine the cohesive force between soil particles

a

Empirical coefficient

As

Area of the streambank soil particle diameter

alpha

Used in Equation

%

7.24

1.98 Pa m3 kg-1 m2

7.31

7.22, 7.30

Statistical significance level

b

Regression coefficient (Rate of increase in load for an increase in discharge)

B

Stream bottom width

bF

Coefficient used to determine the

4.1

m

xxiii

7.17 7.24

cohesive force between soil particles 9.6 Pa mm-1

b

Empirical coefficient

C

Chezy’s roughness coefficient

cF

Coefficient used to determine the cohesive force between soil particles

CL

Percentage of clay content

CL

Coefficient of lift force

c’

Streambank soil cohesion

c

Empirical coefficient

cc

Critical shear stress coefficient accounting for vegetation

7.30, 7.38, 7.39, 7.40, 7.42, 7.43

d

Empirical exponent

7.1

D

Interparticle distance

Dr

Dispersion ratio

7.6

D50

Mean particle size

7.7, 7.25, 7.28, 7.30, 7.31, 7.35, 7.38, 7.39, 7.40, 7.41, 7.42, 7.43

d

Empirical coefficient

m1/2 s-1

7.31 7.27, 7.28, 7.29

7.24

%

7.8, 7.31 7.22, 7.25, 7.38, 7.39, 7.40

Pa 0.0727 Pa

m

m

0.121 Pa

xxiv

7.31

7.24

7.31

-1

E

Fluvial streambank erosion rate

Ed

Fluvial streambank erosion depth

Er

Relative Efficiency

e

Empirical coefficient

FB

-1

m s or m min

mm

6.1, 7.1, 7.2, 7.41, 7.42, 7.43, 7.44, 7.45, 7.46 7.20, 7.21, 7.44, 7.45, 7.46 6.4

4.0012 Pa

7.31

Buoyancy force

N

7.32

FC

Streambank soil cohesive force

N

7.24

FE

Erosive force acting on a streambank soil particle

N

7.36, 7.38, 7.39, 7.40, 7.41

FH

Hydrodynamic force acting on a streambank soil particle

N

7.25, 7.36, 7.37

FL

Lift force on a streambank sediment particle

N

7.22

FR

Resistance force of a streambank soil particle

N

7.30, 7.36, 7.37

Fs

Seepage forces

N

FS

Factor of safety

N N-1

FSW

Submerged weight of a streambank soil particle

xxv

N

7.23, 7.32, 7.33, 7.34, 7.35, 7.36, 7.37

Ft

Hydrostatic force exerted by water in tension crack

N

FW

Hydrostatic force exerted by the surface water on the vertical slip face

N

FWS

Streambank soil particle weight

N m s-2

7.32

g

Gravity acceleration

h

Flow depth

m

6.3, 7.15, 7.17, 7.18

IW

Plasticity index

%

7.5, 7.9

kd

Streambank soil erodibility coefficient

Ki

the initial rate of soil streambank erosion

ks

correction factor for points on the streambank determined as a function of channel shape

n

Manning’s roughness coefficient

N

Normal force at the base of a failure block

O, Or

output (prediction obtained under each change in parameter for sensitivity analysis)

OM

Organic matter content

m3 N-1 s-1 m min-1

6.1, 7.1, 7.10, 7.41

7.2, 7.3

7.16, 7.17

s m-1/3

7.28

N

6.4

%

xxvi

7.23, 7.27, 7.29

7.31

P, Pr

Pt

input (the modified value to account for sensitivity analysis)

6.4

Wetted perimeter on the channel

m

7.12

Pbed

Wetted perimeter on the streambed

m

7.12, 7.13, 7.14, 7.15

Psb

Wetted perimeter on the streambank

m

7.12, 7.13, 7.14

Pleak flow depth

m

7.19

Peak Q

Flow discharge

m3 s-1

Q1.5

Effective flow

m3 s-1

Qs

Daily suspended sediment load

Mg d-1

R, R1, R2

4.1, 4.2, 4.3, 6.3

4.1, 4.2, 4.3

Radius of streambank soil particles

m

7.24

Rh

Hydraulic radius

m

7.28

S

Shear force at the base of a failure block

N

Percentage of Silt+Clay content

%

7.4

SFbed

Shear force acting on the streambed

N

7.11

SFsb

Shear force acting on the streambanks

N

7.11, 7.13, 7.14, 7.15

SC

xxvii

SFt

Total shear force acting on a channel

N

7.11

tF

Time the total erosive force acted on a streambank soil particle located on a specific elevation

s

7.44, 7.45, 7.46

tw

Wetting time of the stormflow event on a specific elevation

s

7.19, 7.20, 7.21

u

Velocity near streambank surface

m s-1

7.22, 7.27, 7.29

u*

Friction velocity at the streambank surface

m s-1

7.26, 7.27

Vs

Volume of a streambank soil particle

m3

7.23, 7.33, 7.34

W

Top width of the stream surface

m

7.14, 7.15

Ws

Weight of the failure block

N

Ww

Weight of the surface water on the failure block

N

X

Transect (Distance)

m

y

Specific vertical distance from the streambank surface

m



Streambank angle

o

(Degrees)

6.2, 7.15, 7.23, 7.34, 7.35, 7.38, 7.39, 7.40, 7.42, 7.43

β

Streambank failure angle

o

(Degrees)

6.2

xxviii

7.18, 7.19, 7.20

(Degrees)

7.23, 7.34, 7.35, 7.38, 7.39, 7.40, 7.42, 7.43 6.2



Channel slope

o

’

Streambank soil friction angle

o

(Degrees)

b

Streambank soil suction angle

o

(Degrees)



Pi number

b 

Streambank soil particle bulk density

kg m-3

7.31

s 

Streambank soil particle density

kg m-3

7.23

w

Stream water density

kg m-3

7.22, 7.23, 7.26, 7.29



Surface tension of a soil particle

7.24

Average shear stress

Pa

6.1, 7.9, 7.14, 7.15, 7.16, 7.18, 7.26, 7.29

Boundary shear stress applied by the flow on the streambed

Pa

7.12

6.1, 7.1, 7.4, 7.5, 7.6, 7.7, 7.8, 7.9, 7.10, 7.30, 7.31



bed

7.24, 7.25, 7.30, 7.35, 7.38, 7.39, 7.40, 7.41, 7.42

c

Streambank soil critical shear stress

Pa

c

Streambank soil critical shear stress

dyne cm-2

7.2, 7.3

sb

Boundary shear stress applied by the

dyne cm-2

7.2

xxix

flow on the streambank surface

sb

Boundary shear stress applied by the flow on the streambank surface

Pa

7.1, 7.12, 7.14, 7.15, 7.16, 7.18, 7.25, 7.38, 7.39, 7.40, 7.42, 7.43

t

Total boundary shear stress applied by the flow on the channel

Pa

7.12

s

Unit weight of the streambank material

dyne cm-3

7.3

s

Unit weight of the streambank material

kg m

7.33, 7.34, 7.35, 7.38, 7.39, 7.40, 7.42, 7.43

w

Unit weight of the stream water

kg m-3

7.33, 7.34, 7.35, 7.38, 7.39, 7.40, 7.42, 7.43

-3

xxx

CHAPTER I INTRODUCTION Streambank, in a geomorphic context, is the landform distinguished by the topographic gradient from the streambed of a channel along the lateral land-water margin up to the highest stage of flow or up to the topographic edge where water begins to spread laterally over the floodplain surface (Florsheim et al., 2008). Streambank erosion refers to the erosion of sediment from this distinct landform. Streambanks are often characterized by bare sediment, live vegetation or snags (Roy et al., 2003). Streambank erosion is integral to the functioning of stream ecosystems. It is a geomorphic process that promotes riparian vegetation succession and creates dynamic habitats crucial for aquatic and riparian plants and animals (Florsheim et al., 2008). All streams are dynamic, gradually changing shape as they erode, transport, and deposit sediment. Natural streams will have slowly eroding streambanks, developing sandbars, migrating meanders, and channels reshaped by flood flows in a state of dynamic equilibrium. Under this condition, the stream maintains a stable shape (dimension, pattern, and profile) over time without excessive erosion or sedimentation even as natural or artificial changes occur in the watershed. A stream system maintains this dynamic equilibrium when its natural flexibility and a functional connection to the floodplain are preserved (MDNR, 2010). Extreme streambank erosion indicates an unstable stream. The instability is caused by a change of shape, flow or connectivity in the stream. Changes could be the result of direct 1

factors such as ditching, dredging, straightening or dams establishment, or indirectly caused by land use changes within the watershed such urbanization, logging, degradation of natural riparian vegetation or agricultural production. Jaeger et al. (2010) reported that at least 39 studies have documented streambank widening rates on perennial streams on different physiographic provinces within the southeastern area of the United States (e. g. Beidenharn, 1989; Simon and Hupp, 1992; Simon and Darby, 2002; Thomas, 2000). The mean widening rates for streams on Mississippi and Tennessee were summarized and reported by Jaeger et al. (2010) as 2.46 m y-1 and 10.3 m yr-1, respectively. A preliminary study performed by researchers from the USDA ARS-NSL (Simon et al., 2002) determined that the dominant source of sediments in the James Creek in Mississippi, a watershed within the Southern Plains Ecoregion or Ecoregion 65, is the channel network. Simon et al. (2002) identified that up to 89% of the eroded sediment leaving the system comes from streambeds and streambanks with the balance coming from upland sources, and up to 78% of all eroded materials in the studied watershed comes from unstable streambanks. These results have important implications for managing similar streams in the Ecoregion that have undergone the same type of channel maintenance over the past century. The Town Creek watershed is located in the northeastern part of Mississippi within the Southeastern Plains Ecoregion or Ecoregion 65. Its total area covers 1,769 km2 and represents approximately 50% of the upper Tombigbee River basin area contributing to the Aberdeen Pool on the Tennessee-Tombigbee Waterway. The 950,000 Mg yr-1 of sediment yield from Town Creek watershed estimated by USACE (1988) greatly attributes to the 570,000 Mg yr-1 of deposition in Aberdeen pool estimated (USACE, 2

2000), where annual dredging between 1985 and 2006 averages 280,000 Mg yr-1 (USACE Records presented by Sharp (2007)). The principal channel of the Town Creek watershed was listed by the Mississippi Department of Environmental Quality (MDEQ) in the Mississippi 2010 Section 303(d) List of Impaired Water Bodies from headwaters to mouth at the Tombigbee River. Four other streams within the watershed area were also considered biologically impaired due to sediments. The proposed Total Maximum Daily Load (TMDL) for this watershed (MDEQ, 2006) recommended that streams within the Town Creek watershed especially located near cultivated lands, road crossings and construction activities should be considered a priority for streambank and riparian buffer zone restoration and sediment reduction Best Management Practices (BMPs). To develop remedial measures and future BMPs within the Town Creek watershed for reducing water quality impairment and dredging costs (expressed in terms of a percent reduction of sediment loads), it is necessary to identify the sediment sources and loads currently transported within the watershed. A study focused on determining a sediment budget for the Town Creek watershed is performed by researchers at the Mississippi State University (MSU) and the Northern Gulf Institute (NGI), supported by researchers from the USDA Agricultural Research Service - National Sedimentation Laboratory (NSL) in Oxford Mississippi. The entire approach to determine the sediment budget for the Town Creek watershed is developed by means of flow, streambank, streambed and suspended sediment monitoring, field reconnaissance, GIS applications, and modeling. The data collection plan for the test watershed relied on four monitoring strategies: (a) spatially distributed grab sampling at 24 stations, 7 of them along the principal channel; (b) stream 3

gauging and water sampling at 3 automatic monitoring stations; (c) streambank erosion monitoring using erosion pin arrays, cross section surveys, and jet testing; and (d) streambed sediment sampling. The modeling methods involves the temporal analysis of suspended sediment transport rates and behavior for the Town Creek watershed, and determines different stages of channel evolution of reaches or areas inside the watershed; the development of empirical relationships between streambank erosion rates and streambank and streamflow properties and the application of the one dimensional channel evolution model CONCEPTS (Conservational Channel Evolution and Pollutant Transport System), developed by the USDA Agricultural Research Service (ARS); and finally, the application of the SIAM model (Sediment Impact Analysis Methods) included in the latest version of the HEC-RAS model. Hypothesizing streambank erosion an important mechanism driving sediment supply into the streams and an important portion of the sediment budget for Town Creek watershed, this dissertation research focused on the identification, assessment, evaluation and prediction of streambank erosion processes within the Town Creek watershed, a representative area within the Tombigbee River Basin and the Southeastern Plains Ecoregion. The preliminary study performed by Simon et al. (2002) has brought important results and insights about the channel contribution on overall sediment loads from Ecoregion 65 watersheds. However, the present study is the first research within the Southeastern Plains Ecoregion in Mississippi combining streambank erosion monitoring in situ and modeling techniques to assess the actual contribution of streambank erosion within a watershed in this Ecoregion.

4

Goals and Objectives The overall goal of this research was to identify mechanisms and the potential effects of streambank erosion processes and to quantify and model the magnitude and rates of these processes within a watershed in the Southeastern Plains Ecoregion in Mississippi (also identified as Ecoregion 65). The present research was intended to evaluate and quantify streambank erosion rates, and to generate empirical or analytical correlations for estimating streambank erosion from physical and geomorphic variables for streams in the Ecoregion 65 in Mississippi. Focusing on the Town Creek watershed as a representative watershed for the studied Ecoregion, the research used a combination of methods including field reconnaissance and detailed data collection, laboratory analysis, and channel modeling. Specific objectives include the following: 1. Evaluate spatial and temporal variation of suspended sediment yields and loads, as well as to determine the relation between the sediment rating parameters, suspended sediment load trends, channel evolution and watershed characteristics, in order to identify trends and possible mechanisms driving sediment supply and exportation. 2. Quantify streambank erosion rates and determine the processes and factors affecting those rates on different areas within the Town Creek watershed. 3. Assess the application of the computational model CONCEPTS to predict streambank erosion in Town Creek.

5

4. Develop empirical and/or analytical expressions for predicting streambank erosion rates by fluvial erosion within the Town Creek watershed as a representative watershed within the Ecoregion 65. To address the research objectives, four substudies were undertaken. The first substudy involved a temporal analysis of suspended sediment transport rates and behavior for Town Creek, and determined different stages of channel evolution of reaches or areas inside the watershed. The second substudy assessed streambank erosion (or deposition) rates on pin erosion plots established at two locations along the principal channel on the Town Creek and on eight periodically surveyed cross sections on incised streambanks along a 270 m channel reach at the northern headwaters on the Yonaba Creek. The third substudy performed computational modeling by applying the CONCEPTS model. The fourth substudy evaluated the effects of streambank soil properties and streambank and flow characteristics on streambank erosion rates by using results from in situ jet testing, streambank erosion monitoring and streambank soil and stream water sampling and laboratory analysis. The methodology, results, and discussion for each of the studies is presented in separate chapters.

6

CHAPTER II STREAMBANK EROSION One of the most interesting characteristics of open systems such as streams is their capacity for self-regulation (Gordon et al., 2008). Streams continuously shape and adjust their channels through erosion of the channel boundary (streambed and streambanks) and the reworking and deposition of sediments (Charlton, 2008). However, channel width adjustment due to streambank erosion is a common mode of channel form adjustment as streams respond to changes in runoff and sediment supply from the surrounding landscape. Several variables represent channel adjustment such as change in local slopes and velocities, distribution of streambed materials, sediment transport rates, streambank morphology (e.g. channel width, incision) and channel pattern. From a geomorphological point of view, streambank erosion is a natural process, which contributes to the overall physical functioning of the stream and its counteraction may involve cascading changes in channel geometry (Kondolf and Piegay, 2003). Sustained streambank erosion and channel width shift can help to maintain high biological diversity on floodplains and continually create new opportunities for some species of flora (Salo et al., 1986). However, accelerated streambank erosion is a major cause of non-point source pollution associated with increased sediment supply (Rosgen et al., 2001). Streambank erosion can present serious problems to river engineers, environmental managers and farmers through loss of agricultural land, danger to riparian 7

and floodplain structures, increased downstream sedimentation, and occasional riverine boundary dispute (Lawler et al., 1997). Excessive erosion and the transport and deposition of sediment in surface waters are major water-quality problems. Physical, chemical and biological damage associated with excess sediment costs about $16 billion annually in North America (ARS, 2003). Studies have shown that sediment from streambanks can account up to 90% of watershed sediment yields (Simon et al., 2002; Capello, 2008). During the past decades, around 575,000 streambank miles were estimated as actively eroding in the United States, requiring an average annual treatment cost of $1.1 billion (USACE, 1981). This estimation was corroborated by Bernhardt et al. (2005) who determined that over one billion dollars have been spent annually since 1990 on stream restoration. Increased streambank erosion results not only in accelerated sediment yield, but also destabilizes streams with associated changes in stream type.

Streambank Erosion Processes There are two main streambank erosion processes (also called streambank retreat), namely hydraulic action (also called hydraulic erosion, fluvial erosion or tractive erosion) and gravitational mass failure. The two process groups are frequently linked, with hydraulic processes often a precursor to gravitational failures (e. g., a hydraulic induced mechanism, such as streambank undercutting, can cause gravitationally induced collapse such as a cantilever failure). Two major factors contribute to streambank erosion: streambank characteristics (erodibility potential) and hydraulic/gravitational forces (erosive potential). Other processes contributing to streambank erosion include 8

surface and gully erosion, liquefaction, development of positive pore water pressure or wetting of streambank soils, subaerial processes (freeze, thaw), soil piping and soil cracking (Watson and Basher, 2006). Table 2.1 provides a guide for recognition of the typical flow conditions, the type of streambank material and the streambank soil moisture conditions related to the different types of hydraulic and gravitational failures observed within the Town Creek watershed. Each process and mechanism is fully described in the subsequent sections.

9

Table 2.1

Streambank erosion mechanisms observed along the Town Creek watershed

Process

Hydraulic

Hydraulic/Gravitational (seepage effect)

Mechanism Hydraulic or fluvial entrainment Streambank undercutting or scouring Basal cleanout Pop-out failure

Streambank soil characteristics

Streambank soil moisture condition

Varies (stormflows)

Cohesive and non-cohesive

Varies

Varies (stormflows)

Cohesive and non-cohesive

Varies

Varies

Cohesive and non-cohesive

Varies

Low

Cohesive

Saturated

Low

Interbedded (mixed layers of cohesive soils)

Saturated

Low cohesive

Saturated

Cohesive

Saturated

Low

Cohesive

Varies

Low

Mixed Cohesive and low cohesive or noncohesive

Varies

Low

Cohesive

Saturated

Low

Low cohesive

Dry

Piping failure

Planar failure Rotational or slump failure Gravitational

Typical Flow Conditions

Low or after high stormflow Low or after high stormflow

Slab or block failure Cantilever failure Wet earthflow Soil/rock fall

10

Hydraulic Action Process Hydraulic action refers to the detachment, entrainment and removal of soil particles and aggregates from the streambank face by hydraulic forces, mainly occurring during rising flows or flood events. Removal of streambank material by hydraulic action is closely related to near-streambank flow energy conditions, especially the velocity gradient close to the streambank and local turbulence conditions, as these determine the magnitude of hydraulic shear stress (Brierley and Fryirs, 2005). Where the shear stress exceeds streambank erosion resistance sediment transport will be initiated; shear stress usually increases as flow increases, whereas streambank erosion resistance typically decreases with increasing flow (e.g. under gradual streambank saturation). Hydraulic action process mechanisms can be classified as: hydraulic or fluvial entrainment, undercutting or scouring and basal clean out. The effectiveness of hydraulic action depends on the balance between motivating forces that include the downslope component of submerged weight and fluid forces of lift and drag, and resisting forces that include the interparticle forces of friction and interlocking. Cohesive, fine-grained streambank material is usually eroded by entrainment of aggregates or crumbs of soil which are bound tightly together by electrochemical forces, rather than as individual particles. Different authors (e. g. Grissinger, 1982; Osman and Thorne, 1988, Julian and Torres, 2006) consider that these conditions strongly depend upon physical properties of the streambank soil (mineralogy, dispersivity, particle size distribution, moisture content, plasticity index) and on properties of the pore and entraining fluid (temperature, pH, electrical conductivity).

11

Hydraulic or fluvial entrainment (also called as fluvial or hydraulic erosion) occurs when individual grains are removed or shallow slips occur along almost planar surfaces (surface erosion) as the motivating forces overcome the resisting forces of friction and cohesion. In some instances, a distinct notch may be left in the streambank following rising flow or a flood event, indicating the peak stage achieved (Figure 2.1).

Figure 2.1

Hydraulic or fluvial entrainment notch examples along the Town Creek

Undercutting or scouring is the direct removal of streambank material at or below water level by the physical action of flowing water and the sediment it carries (Figure 2.2). This mechanism is observed when velocity and boundary shear stress maxima occur in the lower streambank region or the presence of more erodible material near the streambank toe. High rates of streambank retreat at bend apices are explained by large velocity gradients and high shear stresses generated within large-scale eddies against outer streambank. This process can be the result of the redirection and acceleration of flow around obstructions such as debris and vegetation within the channel, or streambank soil characteristics such as poor drainage and/or seams of readily erodible non-cohesive material within the streambank profile. During this process, flow not only entrains 12

material directly from the streambank face, but also scours the base of the bank. This process could lead to oversteepening, and eventually gravitational failure.

Figure 2.2

Undercutting or scouring along tributaries within the Town Creek watershed

Basal cleanout is the removal of supportive or protective material as well as collapsed streambank material at the streambank base during the stream flow or flooding events (Figure 2.3). Fluid entrainment of basal material following collapse is vital to the effectiveness of the streambank erosion hydraulic action process. Streambank erosion slows down if basal clean-out proceeds slowly, since the mass failure debris temporarily decrease toe erosion (Crosato, 2007). Failed, cohesive material can act as a form of natural streambank toe protection by consuming and diverting flow energy that may otherwise be used to further scour the basal zone of incising channels (Wood et al., 2001).

13

Mar 09

Figure 2.3

Jun 09

Sep 09

Nov 09

Basal cleanout process after a streambank failure in a northern headwater of the Town Creek watershed

Gravitational Mass Failure Process Gravitational streambank failure is also known as mass wasting or mechanical failure and includes mass movement failures and individual grain failures. This process refers to the detachment of sediment primarily from cohesive streambanks and makes it available for fluvial transport. It occurs when the weight of the streambank is greater than the shear strength of the soil. Increasing streambank height or streambank angle due to the fluvial erosion mechanism (channel incision) and the presence of tension cracks are frequently precursors for gravitational mass failure processes (ASCE, 1988). Streambanks susceptibility to mass failure depends on their geometry, structure or stratigraphy, material properties and presence and type of riparian vegetation (Thorne, 1990). Mass failures are uncommon in non cohesive streambanks, where basal scour and oversteepening are more common. Occurrence of weakening and weathering processes (pre-wetting, desiccation and freeze-thaw) on streambanks reduce the strength of streambank material, thereby decreasing streambank stability. The effectiveness of these processes is related to 14

streambank soil moisture conditions. Mass failures often occur following flood or stormflow events. Precipitation and a rising stream stage increase the moisture content and weight of streambank soils. At the same time, apparent soil cohesion is decreased through the reduction of matric suction. Prolonged rainfall often develops positive pore pressures, resulting in a decrease in frictional soil strength (Watson and Basher, 2006). The stability of the streambank with regard to mass failure depends on the balance between gravitational forces of friction and cohesion that resist movement (Langendoen, 2000). Streambanks can be brought to failure by over-deepening due to bed scour, or undercutting which increases the streambank angle. An increase in degree of streambank soil saturation can occur as result of recharge into the streambank during stormflow or flooding events, or snowmelt. This substantially increases the mass of the streambank soil above a potential failure surface and can be enough to instigate failure, even in the absence of over-deepening or over-steepening (Lawler et al., 1997). Planar failure is a failure mechanism frequently observed in low cohesive and cohesionless materials along steep or vertical streambank angles (steep to incised channels), where the water table and/or the channel water level is usually low relative to the streambank height (Lawler et al., 1997). Water seepage from the streambank can substantially reduce the stable streambank angle (Lawler et al., 1997). This mechanism usually preceds rotational and/or slab failures (Thorne, 1998) contributing large amounts of sediment to the base of the streambank where it can be removed by a basal clean out (Figure 2.4). Streambank vegetation normally helps to stabilize against this type of failure. Planar failure could be differenced from the shallow failure because the layer of material moves along a non-parallel plane to the streambank surface. 15

Figure 2.4

Planar failure on an incised headwater of the Town Creek

Rotational failure or slump failure usually occurs on moderately high or steep and cohesive streambanks. This mechanism is identified by deep-seated movements of material downward and outward along a curved slip surface (Figure 2.5). After failure the upper slope of the slipped block is typically tilted inward toward the streambank (Watson and Basher, 2006). The formation of vertical tension cracks within the streambank structure reduces stability, particularly when water-filled, resulting in a high pore water

16

pressure within the streambank or as a result of scour at the streambank base. The water table position significantly affects the occurrence of this type of failure.

Figure 2.5

Rotational failure or slump failure on tributaries and principal channel of the Town Creek

Slab failure or block failure is associated with steep or near vertical, low height, fine-grained cohesive streambanks and tend to occur during low flow conditions. Under this mechanism, slides produce vertical faced slab-type features which either does not rotate at all, or which rotate away from the streambank slightly (Figure 2.6). Slab failures result from the combination of scour at the streambank toe, the development of deep 17

tension cracks at the top of the streambank and elevated pore water pressure in the upper streambank soil profile. This type of failure is not significantly affected by the groundwater table level, hence streambank stability primarily depends on the tensile strength of the streambank soil.

Figure 2.6

Slab failure or block failure on incised headwaters of the Town Creek

Cantilever failure is the collapse of an overhanging block often observed when streambank erosion at the lower portion of the streambank is controlled by fluvial undercutting and the upper portion of the streambank is controlled by gravitational failure. The failed block typically comes to rest on the lower streambank (or directly in the stream) immediately adjacent to the streambank and with the grass root matting intact and the above ground biomass on top (Figure 2.7). The stability and dimension of cantilevers, as well as the mode of the cantilever failure (shear, beam, or tensile failure) is dependent upon the thickness of the upper streambank material and its engineering properties (Thorne and Lewin, 1979). Crack developments in a cantilever are often the critical factor that causes failure. Roots and rhizomes of grasses enhance the stability of 18

the upper streambanks by reinforcing the soil, thereby inhibiting development of cracks. Shear failures occur by downward displacement of the overhanging block initiated by tension cracks that form at some distance back from the streambank. Failure comes about because the shear stress due to the weight of the block overcomes the shear strength of the soil (Thorne and Tovey, 1981). In tensile failures the tensile stress due to the weight of the lower part of the hanging block exceeds the tensile strength of the soil causing detachment of the lower portion. The failed block often has no vegetation on it. Tension failures often leave root bound remnant blocks as overhangs which eventually fail by the beam mechanism. In beam failures, the block rotates during failure and ends up on the lower streambank or in the stream with the vegetated mat still in place but lying in a vertical plane and pointing towards the stream (Thorne and Tovey, 1981). Generally, beam failure is the most common mechanism of cantilever collapse.

19

Figure 2.7

Cantilever failure on tributaries of the Town Creek

Wet earthflow failure frequently occurs in streambanks subject to strong seepage and poor drainage and is typically caused by waterlogging due to high rainfall, snowmelt or rapid drawdown of water in the channel. This type of failure occurs because the increased soil water content of a saturated streambank increases streambank weight and decreases streambank soil strength, to a point where the soil slides down the streambank in a semi-liquid state to form lobes at the toe. Those lobes of eroded soil are easily

20

removed by basal cleanout even at low stream flows (Figure 2.8). Failures commonly occur on slopes that are gentle to moderate.

Figure 2.8

Wet earthflow failure and path of runoff flow on streambanks on the Yonaba Creek

Popout failure can occur in the lower half of steep, wet streambanks with strong seepage and/or positive pore water pressure forces acting within its structure. Popout failure occurs when small to medium sized blocks are forced out at the base of the streambank due to excessive pore pressure and overburden (O’Neill and Kuhns, 1994). A slab of material in the lower half of the steep streambank face falls out, leaving an 21

alcove-shaped cavity (Figure 2.9). The over-hanging roof of the alcove subsequently collapses as a cantilever failure. Evidence includes steep streambank face with seepage area low in the streambank and alcove shaped cavities in streambank face.

Figure 2.9

Popout failure on incised headwaters of the Town Creek

Piping failure is the collapse of part of the streambank due to high groundwater seepage pressures and rates of flow causing selective removal of sections of the streambank (Biedenharn et al., 1997). This mechanism is due to preferential groundwater flow along interbeded saturated layers contained within stratified streambanks, with lenses of sand and coarser material inserted between layers of finer cohesive material (Watson and Basher, 2006). Sections of the streambank disintegrate and are entrained by the seepage flow (sapping). They may be transported away from the streambank face by surface run-off generated by the seepage, if there is sufficient volume of flow. Evidence includes pronounced seep lines especially along sand layers or lenses in the streambank; pipe shaped cavities in the streambank; notches in the streambank associated with 22

seepage zones; and run-out deposits of eroded material on the lower streambank (Watson et al., 1999) (Figure 2.10).

Figure 2.10

Piping failure on incised headwaters of the Town Creek

Soil/rock fall occurs only on steep low cohesion streambanks where grains, grain assemblages or blocks fall into the channel. Soil and rock falls often occur when a stream undercuts the toe of sand, gravel or deeply weathered rock streambank (Thorne, 1998). This process could be identified by the presence of debris falling into the channel; failure

23

masses broken into small blocks; and the absence of rotation or sliding failures on the eroding streambank (Figure 2.11).

Figure 2.11

Soil fall on incised headwaters of the Town Creek

24

CHAPTER III STUDY AREA Tombigbee River Basin The Tombigbee River Basin provides one of the principal routes of commercial navigation in the southern United States, connecting at its upper end to the Tennessee River by the Tennessee-Tombigbee Waterway and the Gulf of Mexico and Gulf Intracoastal Waterway at its lower end. The Tombigbee River, located at the western edge of the Mobile Bay basin represents an area of approximately 35,656 km2 within the states of Alabama and Mississippi (Figure 3.1). The Tombigbee River basin includes the Upper and Lower Tombigbee River sub basins. The major tributaries for the Tombigbee River are the Buttahatchee River, the Noxubee River and the Sucarnooche River from the west, and the Sipsey River from the northeast. The main system of the river joins with the Black Warrior River at Demopolis, AL and the Alabama River south of Jackson, AL to finally drain its waters to the Mobile River, before entering the Mobile Bay and the Gulf of Mexico. The Tombigbee River basin encompasses three physiographic provinces that are characterized by marked differences in underlying geology. These are the East Gulf Coastal Plain, the Cumberland Plateau, and the Valley and Ridge. The area separating the uplands of the Valley and Ridge and Cumberland from the Coastal Plain lowlands is known as the Fall Line Hills and is characterized by broad, flat ridges separated by 25

narrow stream valleys. Adjacent to the Fall Line Hills lies the Black Belt. The physiography of the Black Belt is markedly different from the surrounding region, with low relief and mature weathering (Futato et al., 1989). The Tombigbee River Basin is located within Ecoregion 65, also known as the Southeastern Plains Ecoregion. An Ecoregion has similar of climate, geology, topography and ecology. Land use along these irregular plains is mixed among croplands, pastures, woodlands, and forest (Figure 3.1). The streams at this area are relatively low gradient and generally have sand beds (Simon et al., 2002). The Upper Tombigbee River subbasin from its source to Columbus, MS flows through an area of hilly to mountainous terrain. The Upper Tombigbee River lies entirely within the coastal plain physiographic province with elevations ranging from about 300 m above mean sea level (amsl) at the highest point to 57.9 m amsl at the lowest point, located at the lower end of the Waterway Canal section (USACE, 1980). The tributaries at this subbasin present higher gradient headwater streams (>1m km-1) and narrow floodplains making them very susceptible to flash flooding. The East Fork of the Upper Tombigbee River is formed by the junction of Browns Creek and Mackeys Creek and flows about 100 km almost due south to the conjunction with the Town Creek in the Aberdeen Lock and Dam Pool. The Upper Tombigbee River and its subbasin run mostly in a North-South direction in West-Central Alabama and Northeastern Mississippi. The Lower Tombigbee River and its subbasin run from the confluence of the Tombigbee and Black Warrior Rivers in a southerly direction to confluence with the Alabama River in southeastern Alabama (Bankhead et al., 2008).

26

The Tombigbee River basin involves natural and built systems that evidence different associated sediment issues. Haydel and McAnally (2002) have identified that sedimentation of the navigation channel and the ports on the Tennessee-Tombigbee Waterway (TTW) in Mississippi has averaged over 611,644 m3 per year since completion of the Waterway.

Figure 3.1

Location of the Tombigbee River basin, the Mobile River basin and the Ecoregion 65 in Mississippi and Alabama

For at least the last 10 years have been collected in the Tombigbee River Basin no measurements of suspended sediment and discharge. Previous reports have determined that changes in hydraulic effects on the natural river system caused by the construction of 27

the TTW were greatest at their junction, and the effects attenuate upstream to a minimal background level (USACE, 1988; USACE 1991). Sediment supply to the streams and Tombigbee River section is sufficient to cause ports and waterway sedimentation problems at any location inside the watershed, and especially along the TTW where the capacity to transport sediment is smaller than the sediment supply. Haydel and McAnally (2002) reported that the five public ports on the TTW located within the Tombigbee River basin have experienced sedimentation problems. Based on the limited available data, Haydel and McAnally (2002) categorized the sedimentation sources and mechanisms as: vessel-resuspended bed sediment from the waterway and port is transported into the port and deposits; sediment transported into and through the port during flow events deposits; and other mechanisms contribute sediment (slumping, piping, etc.). From available data, observations by port officials, and experience, Haydel and McAnally (2002) estimated that vessel resuspension contributed the highest relative sedimentation rate at the two upstream ports on the waterway, while the sediment transport from upstream was considered the most important cause of sediment deposition, increasing along the three downstream ports. Their estimations also showed that other mechanisms contributed up to 20% of the sediment deposition quantified at the entire set of ports. Transport from upstream occurs primarily during the high flow period of January through April, but can occur during any high flow event caused by local hydrology or upstream releases that exceed the threshold for sediment transport.

28

While sediment deposition is observed upstream of the regulating waterway structures in their pools, the structures could significantly affect downstream discharge and channel morphology in different forms. Generally, maximum streambed degradation occurs downstream of the dam, where the energy available for sediment transport and sediment supply is most out of balance with pre-construction system. Degradation progressively decreases downstream of the dam. Changes in channel width could occur further downstream as its magnitude exceeding streambed degradation. Changes in width depend on different factors such as the streambed and streambank materials, flow regime, vessel traffic and type and density of riparian vegetation. Bankhead et al. (2008) estimated that the Tombigbee River has widened up to 85 m from 1974 to 2003. Average widening rate is approximately 1.2 m yr-1, but is as high as 2 to 3 m yr-1 in some reaches just downstream from some of the dams. Their analysis shows no significant difference between the observed average widening rates for the pre and post-waterway periods of time (1.4 and 1.0 m yr-1, respectively). Both conditions exceed widening rates that normally occur in stable alluvial streams, where streambank erosion and channel widening would be minimal. Bankhead et al. (2008) found that areas of high streambank erosion were mainly located between the dams. They therefore relate the widening to the dams, because these structures alter the transfer of water and sediment downstream.

Town Creek Watershed The Town Creek watershed is located in the northeast area of Mississippi (Figure 3.2). Its total area covers 1,769 km2 and represents approximately 50% of the upper Tombigbee River basin area contributing to the Aberdeen Pool on the TTW, and about 29

5% of the entire Tombigbee River basin area. Although the major extent of the Town Creek watershed lies in Lee and Pontotoc counties in Mississippi, this watershed also drains a small proportion of Chickasaw, Itawamba, Monroe, Prentiss and Union counties in Mississippi. The watershed is bordered to the west by the Little Tallahatchie River watershed, to the southwest by the Yalobusha river watershed and to the north by the Upper Hatchie River watershed.

30

Figure 3.2

Location of Town Creek watershed (Aerial Picture: MS GIS Council) 31

Physiography and Geology The Town Creek watershed encompasses three physiographic regions oriented in the North to South direction along the watershed (Figure 3.3). These physiographic regions are generally dominated by formations of the Cretaceous age typical of the Selma Group and the Eutaw formation, with a minimal presence of the Clayton and Porters Creek formations at the western area of the watershed, which correspond to the Midway Group formed at the Paleocene age. The most representative physiographic region is the Black Prairie or Blackbelt, which geology contains chalks or marls of the Selma Group (especially the Demopolis Formation). This region presents a flat to gently undulated topography. Its soils are mostly dark-colored alkaline vertisols, few entisols and inceptisols, rare mollisols, with clays prominent, often calcareous, some well-drained acidic overburden over chalk. The original vegetation is not well known, but probably was scattered trees with prairie grasses and wild flowers. Some common tree species are bur oak, durand oak, chinkapin oak, nutmeg hickory, and carolina buckthorn. The Tombigbee Hills are also known as the Tennessee River Hills, an extension of the Fall Line Hills which are formed by the innermost coastal plain deposits extending across Alabama into Georgia. Inside the Town Creek watershed, this physiographic region lies into the eastern watershed area. This zone contains sands, clays, and gravels from the Selma Group and the Eutaw geological formations. The landscape generally presents ravines and ridges, and numerous streams. Very old ultisols, some few alfisols, and observed entisols in stream drainages are the representative soils inside of this physiographic region. Those are highly weathered and acid. Hardwoods and pine (shortleaf and loblolly pine) are the most common vegetation along this zone. This 32

physiographic region lies into the east and northeast of Tupelo, Corinth, Baldwyn and Nettleton in the direction of Amory. The ridges and valleys geology from the North Central Hills region within Town Creek watershed present sands and clays from the Clayton and the Porters Creek Formations, which belong to the Midway Group (from the Paleocene age), and the Ripley Formations which belong to the Selma Group. Mostly acidic ultisols and entisols in drainages are the representative soils for this area. Hardwoods and pine (loblolly and shortleaf) are the most frequent vegetation observed in this area. The Selma Group geological formation covers practically the entire watershed. The group is composed of, in ascending order, the Mooreville Chalk Formation, Demopolis Chalk Formation, Ripley Formation, and Prairie Bluff Chalk Formation (Figure 3.3). Formations from the Selma Group are considered to greatly affect the topography with marked lithologic characters. This Formation underlies a belt from Mississippi extending eastward to Alabama with an average width of 36 to 45 km. The material of the formation is a semi-indurated limestone of rather constant characteristics throughout its crop, though showing three distinct phases from bottom to top. The basal portions are very sandy, the sand having been worn by the waves and streams from the older lands of the sandy Eutaw and spread upon the bottom of the Selma sea. These sands however, are highly calcareous and become more so upward until the sands disappear, indicating that the adjacent lands had been worn down and the streams no longer deposited sands on the sea floor (Lowe, 1915). The uniform composition of the Formations into the Selma Group has caused it to be more deeply and evenly wasted by erosion and solution than the more sandy formations north and south of it. 33

The Eutaw Formation outcropping southeast of the Prairie or Selma Formation belt into Town Creek watershed shows no marked topographic features. This is prevailingly a sand formation, with occasional intercalated clay beds, which may pass laterally within a few rods into sand beds. The basal parts of the formation consist of strikingly variegated sands of red, yellow, blue and other colors (Lowe, 1915). These sands, like those of the underlying formation, are distinctly micaceous and conspicuously cross-bedded. Beds of the formation, both sands and clays, are discontinuous and variable over short distances. The upper beds become gluconitic and calcareous toward the top, merging into the overlying Selma Chalk of the Prairie Region (Lowe, 1915).

Watershed Physiographic Characterization The physiographic characterization of the Town Creek watershed yields an asymmetric index (Ia) of 1.42, which is representative of a watershed with a principal channel located towards the eastern part of the watershed. This asymmetry could cause a possible reduction in contributing runoff by this part of the watershed. The watershed shape factor (Kf) of 0.37 and the compactness coefficient (Kc) of 1.5 are representative of an oblong oval area. The watershed mean elevation is around 31.8 m amsl, and the maximum and minimum elevations are 72.3 m amsl and 17.4m, respectively. The length of the principal channel is 73.2 km (Figure 3.4). The estimated average channel slope is 0.57 m km-1 and the mean downstream slope is 0.25 m km-1 (Figure 3.5).

34

a

b

Figure 3.3

(a) Physiographic region and (b) geological formations within the Town Creek watershed in Mississippi 35

The principal tributaries of Town Creek are Tishomingo Creek, Mud Creek and Yonaba Creek in the northern area, Chiwapa Creek, Coonewah Creek, Tallabinella Creek and Tubbalubba Creek in the western area, and Tulip Creek in the eastern area of the watershed (Figure 3.4). The northern headwaters and the western tributaries in Town Creek, located within the Black Prairie Subecoregion, exhibit incised streams with unstable active streambank profiles near agricultural lands. The most common gravitational failure mechanisms are slab failure, soil fall, and cantilever failure, accompanied by a basal clean out process when stormflow events occur. Streambank failure events at these channels tend to be periodic, most frequently occurring during stormflow events. Unstable conditions and resulting failures are frequently caused by saturation of the streambank, which decreases the strength of streambank soils and increases soil unit weight. Field observations also evidenced seepage from intermediate soil layers along the incised streambanks causing piping and sapping failures. Agricultural practices near streambanks with limited or no riparian vegetation destabilize streambanks and favor gully erosion (Figure 3.6). Channel morphology changes from incised V-shaped channels to wide U-shaped channels with an increase in riparian vegetation density upstream of near the Natchez Trace (Figures 3.7 and 3.8).

36

Figure 3.4

Town Creek watershed elevation 150 Profile Gross Channel Slope

Elevation (ft amsl)

125

Average Channel Slope

North Central Hills

Black Prairie Tombigbee Hills

100 Rd.

75 Natchez Trace

Yonaba Creek

Eason Blvd.

Town Creek

Brewer Rd.

50

USGS Station

25 0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

Tombigbee River

40.00

45.00

Distance (mile)

Figure 3.5

Profile and slope of the principal channel in Town Creek watershed (from Yonaba Creek at Road MS-9 to Town Creek outlet) 37

A

B

C Figure 3.6

(a) Gully erosion and (b) streambank erosion processes near agricultural areas at northern headwaters, and (c) transitional areas upstream of the Natchez Trace in the Town Creek watershed

The middle 20 km of the principal channel system are located within the transitional zone between the Tombigbee Hills and the Black Prairie subecoregions. Wide stable channels showing evidence of streambank erosion induced by fluvial erosion, shallow slides, and rotational failures are mixed with natural, vegetated zones and regions with sediment deposition on streambed and streambanks (Figure 3.7). Especially along this section of the principal channel, sediment bed deposition and erosion vary seasonally with flow conditions. Low flow velocities and sediment deposition occur on the inside of incipient meander bends in the sinuous reach, along the downstream most 10 km before the outlet at the Tombigbee River.

38

Figure 3.7

Channel morphology and streambank sediment deposition along the middle 20 km of Town Creek

Land Use The land use distribution of the Town Creek watershed is presented in the Figure 3.8 and Table 3.1. About 22% of the total watershed area is cultivated, the remaining lands are mainly pasture/hay lands (19%) and forest lands (29%). The predominant crops established within the area include soybean, cotton, wheat and corn. Although some farming practices have been recommended to reduce soil erosion, agricultural practices significantly affect streambank gully erosion and streambank erosion processes where riparian vegetation is limited (Figure 3.9). The city of Tupelo, with an area approximately of 133.2 km2 is located in the central east area of the Town Creek watershed, and

39

represents about 75% of all urban area. Other towns within the watershed are Verona, Nettleton, Shannon, Blue Springs, Plantersville and Saltillo (See figure 3.2). Table 3.1

Figure 3.8

Land use distribution at Town Creek watershed

Land Use

% Area

Forest Crops Pastures Shrublands Wetlands Urban Water Grassland Bare rock

29.10 21.50 19.00 11.80 6.20 10.20 1.55 0.60 0.05

Land use distribution within Town Creek Watershed, Mississippi 40

Figure 3.9

Gully erosion near streambanks along agricultural lands with no riparian vegetation

Climate The climate for the Town Creek watershed is typical of the Southeast region of the United States. Generally, the highest rainfall occurs in the winter and spring months. March is generally the wettest month, while August is the driest on average (Figure 3.10). The summer months are characterized by afternoon thunderstorms that can produce locally heavy rainfall. Average annual precipitation ranges from 1,410 mm to 1,472 mm (Figure 3.11). The higher average annual precipitation amounts within the Town Creek watershed are observed across the western area (near Pontotoc, MS), where the North Central Hills physiographic region extends. The lower average annual precipitation depths are observed north to south along the extension of the Black Prairie physiographic region. Monthly average air temperatures are highest in July and August and lowest in January (Figure 3.12). The average annual low and high temperatures are 9.8oC and 23oC respectively, with an annual average temperature of 16.3oC for the entire watershed.

41

180 Blue Springs Tupelo

160

Shannon

Average Precipitation (mm)

140

120

100

80

60

40

20

0 Jan

Feb

Mar

Apr

May

Jun Jul Month

Aug

Sep

Oct

Nov

Dec

Figure 3.10

Average precipitation amount for different areas within the Town Creek watershed (Data source: The Weather Channel)

Figure 3.11

Spatial distribution of the average annual precipitation within Town Creek watershed 42

30 Blue Springs Tupelo Shannon 25

Temperature (oC)

20

15

10

5

0 Jan

Figure 3.12

Feb

Mar

Apr

May

Jun

Jul Month

Aug

Sep

Oct

Nov

Dec

Average Temperature for different areas within the Town Creek watershed (Source: The Weather Channel)

Streamflow Daily flow discharge in Town Creek watershed has been monitored by the U. S. Geological Service (USGS) from 1939 at the USGS gaging station number 02436500 (Town Creek Near Nettleton, MS) and from 1970 at the USGS gaging station number 02435020 (Town Creek at Eason Boulevard at Tupelo, MS), which is since 2003 administered by the NOAA -Advanced Hydrologic Prediction Service (AHPS). Maximum peak instantaneous discharge for the station near Nettleton, MS is 4,275.8 m3 s-1 and occurred on March 22, 1955. After 1970, the peak instantaneous discharge for this station is 2055.8 m3 s-1 and occurred on March 16, 1973. Peak instantaneous discharge for the station at Tupelo, MS is 1,073.2 m3 s-1 and occurred on May 27, 1991. At this date, the station near Nettleton registered the highest peak discharge for the last 25 years of 1,761.3 m3 s-1 (Figure 3.13). The recurrence intervals for peak discharges at the two stations along Town Creek watershed is shown in Figure 3.14. 43

10000 USGS 02436500 TOWN CREEK NR NETTLETON, MS USGS 02435020 TOWN CREEK AT EASON BOULEVARD AT TUPELO, MS

4275.8

Discharge (m3s-1)

2055.8 1761.3

1073.2

1000

100 1940

1945

1950

1955

1960

1965

1970

1975

1980

1985

1990

1995

2000

2005

2010

Water Year (October 1 to September 30)

Figure 3.13

Annual instantaneous peak discharges at two stations along Town Creek watershed. Data obtained from USGS and AHPS - NOAA Web site

The station at Tupelo has a drainage area of 603.5 km2, representing 34% of the entire watershed area. The station near Nettleton, MS, has a drainage area of 1605.8 km2, which represents 37.6% of the watershed area. Mean monthly discharge at the station at Tupelo varies between 23.8% and 55.5% of the mean monthly discharge at the station near Nettleton, with an average of 42%. The mean monthly discharge for both stations at Town Creek follows the precipitation regime, with the highest values during the winter and spring seasons, and the lowest amounts during the summer and fall seasons. March is the month with the highest mean monthly discharge with 58 m3 s-1 and 24 m3 s-1 at the station near Nettleton and at the station at Tupelo, respectively. The lowest mean monthly discharge is observed in August with 5.5 m3 s-1 and 2.1 m3 s-1 respectively for the same stations (Figure 3.15).

44

Discharge (m3s-1)

10000

1000

Town Creek nr Nettleton, MS Town Creek at Eason Boulevard at Tupelo, MS 100 1

10

100

1000

Recurrence Interval (yr)

Figure 3.14

Recurrence Intervals for peak discharges at two stations along Town Creek watershed. Data obtained from USGS and AHPS - NOAA Web site

65 USGS 02436500 TOWN CREEK NR NETTLETON, MS

60 USGS 02435020 TOWN CREEK AT EASON BOULEVARD AT TUPELO, MS

55

Mean Monthly Discharge (m3 s-1)

50 45 40 35 30

Mean Monthly Discharge 28 m3s-1

25 20 15

Mean Monthly Discharge 12 m3s-1

10 5 0 Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

Month

Figure 3.15

Mean monthly discharge for Town Creek watershed at two stations along the principal channel. Data obtained from USGS and AHPS – NOAA Web site

45

Water Quality Conditions Four tributary streams and the principal channel within the Town Creek watershed have been included on the Mississippi Section 303(d) List of Water Bodies as impaired due to sediment in the Effective Listing Cycle of 2008 (Figure 3.16). Cowpeena Creek and the majority of Town Creek (before receiving waters from Chiwapa Creek near Nettleton, MS) are considered impaired due to sediment and siltation. Tallabinella Creek, Chiwapa Creek, Mud Creek and Kings Creek are considered biologically impaired due to sediment (Figure 3.16). These streams were listed due to failure to meet minimum water quality criteria for biological use support based on biological sampling (MDEQ, 2010). For these streams, a detailed assessment of the watershed and potential point sources, called stressor identification, showed that sediment was a probable primary stressor in these streams. The water use classification for all the streams and waterbodies considered impaired within the Town Creek watershed, as established by the State of Mississippi Water Quality Criteria for Intrastate, Interstate, and Coastal Waters regulation, is “Fish and Wildlife Support” (MDEQ, 2010). Waters with this classification are intended for fishing and propagation of fish, aquatic life, and wildlife. Waters that meet the Fish and Wildlife Support criteria should also be suitable for secondary contact, which is defined as incidental contact with water including wading and occasional swimming (MDEQ, 2006).

46

Figure 3.16

Impaired waterbodies within Town Creek watershed (Section 303(d) List Fact Sheet for Watershed. Effective Listing Cycle: 2010)

For the streams considered impaired within Town Creek watershed, there is an acceptable range of sediment loadings at the effective discharge of the water body. These ranges were developed from suspended sediment concentration (SSC) data measured at stable streams in the Southern Plains Ecoregion. The target range for the streams within the Town Creek watershed and for the watersheds within the Southern Plains Ecoregion, also known as Ecoregion 65, is 0.0896 to 0.4032 Mg km-2 per day at the effective discharge. The effective discharge is the discharge which is the channel-forming flow. The effective discharge used to determine this range was assumed as the flow with a return period of 1.5 yr (Q1.5), which is commonly considered as the bankfull discharge 47

(Julien, 2002). This discharge has been selected as the critical condition for the development of TMDLs for this watershed (Simon et al., 2002b). If the sediment target applicable for sediment in the streams is not exceeded during critical conditions, then the health of the stream should improve. The estimated existing range of sediment yields for the streams and waterbodies within the Town Creek watershed is 0.448 to 12.096 Mg km2

per day at the effective discharge. The estimated existing range is larger than the

reference range. Therefore, it is recommended that streams within the Town Creek watershed should be considered a priority for streambank and riparian buffer zone restoration and sediment reduction Best Management Practices (BMPs), especially for road crossings, agricultural activities, and construction activities. The implementation of these BMP activities should reduce the sediment load to water bodies within the Town Creek watershed. Reduction of the sediment loads to the streams within the Town Creek watershed to that of a similar relatively stable stream will provide improved habitat for the support of aquatic life in the streams and waterbodies and will result in the attainment of the applicable water quality standards (MDEQ, 2006).

48

CHAPTER IV SUSPENDED SEDIMENT TRANSPORT WITHIN THE TOWN CREEK WATERSHED The overall goal of this research is to evaluate streambank erosion rates and generate empirical correlations for estimating streambank erosion involving physical and geomorphic variables influencing this process for streams in the Southeastern Plains Ecoregion in Mississippi. This chapter presents the first of four substudies that address the contribution of sediment from streambank erosion processes to total sediment supply within the Town Creek watershed. The substudy’s objective is to evaluate spatial and temporal variation of the suspended sediment yields and loads, as well as to determine the relations among the sediment rating parameters, suspended sediment load trends, channel evolution and watershed characteristics, in order to identify trends and possible mechanisms driving sediment supply and exportation. To reach the proposed objectives, an initial analysis describes the variation of suspended sediment loads and yields on 7 stations along the principal channel and stations on 4 tributaries which describe the contribution from the northeastern and western tributaries of the watershed, considered the most significant areas affected by streambank erosion processes. A second analysis determines annual sediment loads and yields by routing a determined suspended sediment transport rating relation throughout the flow dataset record obtained during the studied period. A final analysis involves the evaluation of the entire dataset of flow and 49

suspended sediment records from the USGS station 02436500 (station #8) from 1981 to 1995, to determine time trends and possible relationships between parameters in the different suspended sediment transport rating relations.

Introduction Information about sediment loads is useful for the evaluation of sediment yield and erosion rates, to describe sediment dynamics during floods and to assess downstream geomorphic effects of the sediment transport. From a technical perspective, this information can be also valuable for planning as well as for various civil engineering purposes (Batalla and Sala, 1994). In many lowland streams, the majority of the sediment is transported in suspension. Suspended sediment is that portion of the total sediment load of rivers and streams that is carried in the water column. It contains the portion termed "wash load", or that portion of the suspended load not represented in the streambed material, and the suspended bed material load. In practice, wash load usually comprises the silt/clay or colloidal fraction that is more controlled by supply than energy, but depending on stream transport capacity, any given sediment class may be classified as either wash load or bed material load in a particular river reach. The behavior of suspended sediment in streams is a function of energy conditions, i.e. sediment is stored at low flow and transported under high discharge conditions. However sediment transport rates are also a function of sediment availability. Sediment sources vary within a watershed in relation to physical setting and anthropogenic activities. Variability of suspended sediment concentration and loads can 50

be observed along streams and watersheds due to seasonal effects, exhaustion processes and to sediment behavior during storm events. Suspended sediment loads can be characterized by marked fluctuations in time, being highly sensitive to changes in climatic conditions (i.e. storm events) which are by no means time constant (Batalla and Sala, 1994). Empirical relationships between sediment load and water discharge are often applied. The most commonly used relationship is a rating curve in the form of a power function (Bača, 2002). Most of the suspended sediment transport and wash load relations are derived from measured sediment rating curves and flow-duration curves. Suspended sediment ratings represent the relationship between suspended sediment load and water discharge in a stream. The most commonly used sediment rating curve is a power function: Q s  aQ b

Eq. 4.1

where Qs is suspended sediment load (Mg d-1), Q is water discharge (m3 s-1), and a and b are regression coefficients. This expression implicitly incorporates the effect of stream power at increased discharge and the availability of additional sources of sediment during weather conditions that cause high discharge. The physical meaning of the regression coefficients is varied. Due to the inverse correlation between a and b, the combination of a and b can be a measure of soil erodibility and stream erosivity (Thomas, 1988; Yang et al., 2007). Interpretations of the sediment rating curve by Asselman (2000) did not establish a physical meaning for the regression coefficients. Morgan (2006) considered the intercept (a) a stream erosion index. Yang et al. (2007) considered large values of the slope (b), representing an index of the erosive power of the river, indicative of streams 51

with a strong increase in entrainment and transport with increasing flow discharge. Simon et al. (2002) considered a to represent the load at a low flow or base flow discharge and indicative of background levels from the channel system. Kuhnle and Simon (2000) suggested that b, which is the rate of increase in load for an increase in discharge, is indicative of sediment availability in the watershed and channel system with increasing flow discharge. Simon (1989) related changes in the levels of b to different stages of channel evolution (Figure 4.1) and describes possible scenarios in terms of the processes that influence the contribution of suspended sediment to streams: (1) In Stage I, yields are relatively low and are a function of contributions from sheet erosion and gully erosion in the fields. (2) Following construction (Stage II), migrations of knickpoints up tributary streams cause a significant increase in yields during Stage III. (3) Streambank failures by mass wasting during Stage IV serve to further increase yields. (4) During Stage V, mass wasting slows, main channel exhibits much lower energy conditions, and yields decrease. (5) Suspended sediment emanating from tributary streams and gullies continues to be delivered to main channel, thereby maintaining yields during net aggradational phases (Stage VI).

52

3.0

Slope of Rating Curve

2.5

2.0

1.5

1.0

0.5

0.0 I

II

III

IV

V

VI

Stage of Channel Evolution

Figure 4.1

Variation in Suspended Sediment Transport Efficiency during the course of Channel Evolution (Adapted from Simon, 1989)

Hence, the parameters of a suspended sediment rating curve can be generally, but not precisely associated with river or stream channel morphology and processes.

Methods The present chapter includes an analysis of a data collection plan for the test watershed which relied on two of the four monitoring strategies previously described in chapter I: a) spatially distributed grab sampling at 24 stations and b) automatic stream gauging and water sampling at 1 of the 3 automatic monitoring stations (see Figure 1.1).

Spatial Variation of Suspended Sediment within the Town Creek Watershed Water samples were collected at bi-weekly intervals from May 2008 to May 2009 at 24 stations (7 along the principal channel) throughout the entire watershed (Figure 53

4.2). Stream velocity and depth profiles were measured in situ by using a Son Tek flow tracker. One water quality monitoring station (ISCO model 2700, station #8, see Figure 4.2,) was established at the 1606 km2 subwatershed outlet. In that location, the established real time USGS station 02436500 was monitored by telemetry and stagedischarge data posted on the internet. Automatic sampling started on February 1, 2009 to sample daily water quality and during increases in stage-discharge. The monitoring site was visited weekly or after a significant runoff event. Water samples collected were placed in 500 ml bottles. Samples were filtered for analysis of Suspended Sediment Concentration (SSC) using binder free glass microfiber filters (GF/F, Whatman) with a nominal pore size of 0.7 m (Guy, 1969). Spatial variation of suspended sediment concentration, loads and yields for each station were evaluated using a normalization procedure. The relative flow for a specific station was determined as the ratio between the flow at that station and the flow at the USGS Station near Nettleton, MS (Station #8). Relative suspended sediment yield and load for each station were similarly determined.

54

Town Creek Watershed Mobile River Basin

1 Mississippi

Alabama

2 3

4 Grab sampling stations

5 6 7 8

Figure 4.2

1. Yonaba Ck at Rd 9 2. Town Ck at Natchez Trace 3. Town Ck at Main St. Tupelo, MS 4. Mud Ck at Main St. Tupelo, MS 5. Town Ck at Eason Blvd Tupelo, MS 6. Town Ck at Brewer Rd 7. Town Ck at Hwy 278 near Nettleton, MS 8. Town Ck at USGS Station near Nettleton, MS

Monitoring stations along principal channel within Town Creek watershed

Suspended Sediment Load Estimations by Using the Suspended Sediment Transport Rating Relation A suspended-sediment transport rating relation was obtained for station #8 from the flow discharge and suspended sediment dataset composed of grab samples and 128 automatic sampling records obtained between May 1, 2008 and April 30, 2009. The suspended sediment load at the effective discharge was used as an indicator of suspended sediment transport conditions for Town Creek watershed. The effective discharge in the present analysis was assumed to be the flow discharge with a return period of 1.5 yr (Q1.5), which can be representative of the bankfull discharge (Julien, 2002). The effective discharge was determined by using the Log Pearson Type III method in USGS Bulletin 17-B (USDI, 1982) through the application of the computer model PeakFQ Win 5.2. 55

Standard WATSTORE formatted peak discharge data were obtained from the web page of USGS station 02436500 (station #8) to run the model. Suspended sediment yield was defined as the suspended sediment load at the effective discharge divided by the 1606 km2 of drainage area at the study site. Daily and cumulative suspended sediment loads were determined by applying the transport curve for the studied period, by summarizing calculated daily sediment loads using the average daily flow data of the USGS station 02436500.

Analysis of Suspended Sediment Transport Rating Relations A suspended sediment transport rating curve was determined from the daily average flow dataset and suspended sediment concentration records from January 1, 1981 to December 31, 2007, and May 1, 2008 to April 30, 2009. The derived suspended sediment transport rating curve parameters (slope and intercept), the suspended sediment load, and the suspended sediment yield estimated at the effective discharge annual and by summarizing the daily sediment loads were compared (confidence intervals at alpha=0.05) with parameters and results generated by suspended sediment transport relations from Simon et al. (2002) and Sharp (2007). During a second analysis, an annual suspended sediment transport rating relation, suspended sediment load and suspended sediment yield amounts for each year of records at the station #8 from January 1, 1981 to December 31, 2007 and May 1, 2008 to April 30, 2009 were estimated. Estimations from the annual relations were compared with annual estimations obtained by integrating the suspended sediment rating curve based on the full data set for each year with recorded data at the studied site. Trend, significant 56

correlation (alpha=0.05) and a power regression (if correlation was significant) between annual rate of increase in load (rating curve slope parameter) and annual load at low or base flow conditions (rating curve intercept parameter) were evaluated. To determine the correlation between annual average instantaneous discharge and the rate of increase in load, a clustering procedure was used. Three clusters grouped into five consecutive years (1981 to 1985, 1986 to 1990, and 1991 to 1995) were selected. The last cluster also included the year of the CEE-MSU research between May 1, 2008 and April 30, 2009. Significant correlation (alpha=0.05) and power regressions were obtained for each cluster. A final analysis was performed to determine if suspended sediment loads from Town Creek watershed have changed over time. Initially, the records from May 1, 2008 to April 30, 2009 were excluded from the third cluster and analyzed as an individual group. A new transport curve was determined for each cluster. The confidence interval (alpha =0.05) for the slope and the coefficient obtained from each regression in each cluster were used as indicators of possible temporal changes in sedimentation processes within the studied area.

Results and Discussion Spatial Variation of Suspended Sediment within the Town Creek Watershed Mean flow discharge and suspended sediment concentrations, loads and yields from grab-sample monitoring at 7 stations along the principal channel and 4 stations along some tributaries of Town Creek watershed are presented in Table 4.1. The relative flow discharge for a specific station was determined as the ratio between the flow 57

discharge at that station and the flow discharge at the USGS Station near Nettleton, MS. Relative suspended sediment yield and load for each station were similarly determined. Figure 4.3 shows the downstream variation of relative flow discharge, and relative suspended sediment loads and yields. Higher rates of increase in suspended sediment load and yield were observed along the upper 20% of study area (350 km2), representing the Northern headwaters of Town Creek watershed and the transitional zone upstream of the City of Tupelo, MS (stations #1, #2 and #3). At this drainage area flow discharge was about 25% of that at the USGS station near Nettleton, MS (station #8) and suspended sediment discharge was about 65% of that at the same station #8 (Figure 4.3). Suspended sediment yield was 2.4 times larger than that at the outlet. Northern and northeastern headwaters evidenced streambank erosion processes from incised unstable streambanks and gully erosion processes from agricultural lands near streambanks with limited riparian vegetation (Figure 4.4). Agricultural lands near incised channels were the most significant contributors of suspended sediment loads during stormflow events in the area. Clearly, incised headwaters channels are major sediment producers. A similar condition was observed by Patton (1973), who describes that the 20% of the Rio Puerco arroyo watershed which channels are predominantly incisive produces only 8% of the total runoff, but delivers almost the 50% of the sediment that reaches the outlet of the entire basin of the Rio Grande in Colorado. Channel morphology within the Town Creek watershed changes from incised Vshaped channels to wide trapezoidal channels with an increase in riparian vegetation density upstream of the Natchez Trace station. 58

Table 4.1

Mean flow discharge, suspended sediment concentration, load and yield from biweekly grab sampling between May 2008 and May 2009 at 7 stations along the principal channel in Town Creek

Station

59

1 - Yonaba Ck at Rd 9 2 - Yonaba/Town Ck at Natchez Trace 3 - Town Creek at Main St.- Tupelo, MS 4 - Mud Creek at Main St. - Tupelo, MS 5- Town Creek at Eason Blvd - Tupelo, MS 6 - Town Creek at Brewer Rd 7 - Town Creek at Hwy 278 near Nettleton, MS 8 - Town Creek at USGS Station near Nettleton, MS

Subbasin Area (km2)

Instantaneous discharge* (m3s-1)

Suspended Sediment Conc. (mg L-1)

Suspended Sediment Load (Mg d-1)

Suspended Sediment Yield (Mg d-1km2)

20 225 340 260 604 725 1125

5.2 37.9 47.0 38.0 89.6 115.8 140.3

22.5b 52.3a 26.3b 27.5b 27.4b 24.6b 18.8c

2.3f 102.1d 169.0c 94.2d 208.5b 192.9b 187.3b

0.12e 0.45a 0.48a 0.36b 0.35b 0.27c 0.18d

1606

225.8

21.3c

263.7a

0.20d

Significance grouping (letters), based on ANOVA indicated a significant variation (alpha=0.05) among stations along the principal channel. *Instantaneous discharge measured in situ at the grab sampling date

The Eason Blvd station (#5) with a drainage area of 604 km2 receives the combined contribution from the upper Town Creek reaches and the Mud Creek subbasin, with a discharge of about 40% of that near Nettleton, MS. Although suspended sediment loads and flow from the stations on these creeks at Main Street were statistically different (alpha=0.05), they were relatively close in amount. From the Main Street stations to the confluence of Mud and Town Creeks (3.2 km downstream of the Main Street stations) suspended sediment load was reduced about 25%. The suspended sediment load reduction was probably caused by the presence of natural and constructed sediment traps in these channel reaches (Figure 4.5). The confluence of Town Creek and Mud Creek is 300 m upstream of the Eason Blvd station. Wide stable channels with a wide riparian vegetation buffer are common between the confluence and the station location, favoring reduction in flow velocity and energy. Although the relative suspended sediment yield was reduced when compared with the station at Main Street, it was still about 1.75 times larger than that at the station #8. Suspended sediment concentration and yields at the station located on Town Creek at Hwy 278 near Nettleton, MS (#7) were significantly lower than those at the immediate upstream stations Brewer Rd (#6) and Eason Blvd (#5). Suspended sediment load was reduced by 10%, suspended sediment yield was reduced by 50% and suspended sediment concentration was reduced by 25%. Suspended sediment yield at this station was around 90% of that at the subwatershed outlet, while relative suspended sediment discharge was around 70%. Streambank erosion monitoring performed along this specific section showed low streambank erosion rates dominated by fluvial erosion and high streambank deposition rates especially during the winter and the spring seasons. 60

Sediment deposition on streambanks (Figure 4.6) was principally due to peak discharge attenuation during high stages on the Tombigbee River. Rising water levels on the Tombigbee River were affected by weather conditions, as well as by the presence of the Aberdeen lock, dam and pool on the Tenn-Tom Waterway only a few kilometers away from the Town Creek watershed mouth. Sediment streambed deposition and erosion were significantly modified by flow conditions.

2.5 Relative Suspended Sediment Yield Relative Suspended Sediment Discharge Relative Flow Discharge

Relative fraction

2.0

1.5

1.0

0.5

0.0 0

Figure 4.3

200

400

600

800 1000 2 Area (km )

1200

1400

1600

Variation of relative flow discharge, relative suspended sediment loads and relative suspended sediment yields along the Town Creek

61

a

b

c Figure 4.4

(a) Gully erosion and (b) streambank erosion processes near agricultural areas at northern headwaters, and (c) transitional areas upstream of the Natchez Trace in the Town Creek watershed

a

b

c Figure 4.5

(a) Natural pool at Mud Creek near station #4, (b) Sediment trap established at Town Creek about 20 m downstream of the station #3 and (c) Natural pool 50 m upstream of the station #5 62

Deposition

Figure 4.6

Channel morphology and streambank sediment deposition along the middle 20 km of Town Creek.

The western tributaries of Town Creek contributed up to 30% of the total amount of flow discharge and suspended sediment exported by the entire 1606 km2 subwatershed. This result was obtained after analyzing the monitoring records for the western tributaries stations, and by comparing suspended sediment loads and yields, and flow discharges between the stations #7 and #8 on the principal channel near Nettleton, MS. The western tributaries (Tallabinella and Cooneewah Creeks) discharge into Chiwapa Creek a few meters before Chiwapa Creek enters Town Creek downstream of the station on Hwy 278 (#7). The western area within Town Creek watershed presents very similar characteristics in land use conditions to those observed at the northern headwaters, with significant absence of riparian vegetation zone. Channels in this zone present stages of evolution evidencing degradation and widening processes (Stages III and IV) (Figure 4.7), and very particularly along the middle and lowest length of the Chiwapa Creek, streambank instability causing widening and streambed aggradation is observed (Stage V) (Figure 4.8). Along the Chiwapa Creek, between the Road MS-178 and approximately a mile after the intersection of the creek with the Highway 45, A-Jax 63

structures which are supposed to stabilize the streambank toe have been established. Favorable effects appear to be observed upstream of the highway 45 where streambanks with A-Jax structures look stable and vegetated (Figure 4.9); downstream of the highway, the structures appear to be affecting the stream course, favoring fluvial streambank erosion and gravitational failure processes (Figure 4.10).

Figure 4.7

Western tributaries of the Town Creek watershed evidencing widening and degradation processes

Figure 4.8

Western tributary of the Town Creek watershed evidencing widening and aggradation processes

64

Figure 4.9

Established A-Jax structures to stabilize streambank toe along the Chiwapa Creek

Figure 4.10

Established A-Jax structures favoring fluvial streambank erosion and gravitational failures along the Chiwapa Creek

Load and Yield Estimations by Using the Suspended Sediment Transport Rating Relation Figure 4.11 shows the daily flow discharge and the suspended sediment records obtained from May 1, 2008 to May 15, 2009 at the USGS station on Town Creek near Nettleton, MS (station #8). Flow discharges ranged from 0.5 to 626 m3s-1 with mean daily discharge of 29.3 m3s-1. The lower flow discharges were observed during summer 2008 with levels below 1 m3 s-1 and the higher flow discharges during winter and fall 65

stormflow events. Suspended sediment concentrations at this station ranged from 4.8 to 1,578 mg L-1. A peak flow event on December 10, 2008 occurred after several small successive events carrying a suspended sediment concentration of about 900 mg L-1. However, higher suspended sediment concentrations (above 1,000 mg L-1) occurred during stormflow events at the end of the spring season (April and May 2009). Grab-sample records from summer 2008 showed higher suspended sediment concentrations than those obtained during low flow conditions in the fall season by up to one order of magnitude. The higher concentrations were due to the discharges from the Publically Owned Treatment Works of the City of Tupelo, MS. Field observations in the summer showed very low flow velocities and green color in the water caused by algae growth which increased the water turbidity and suspended sediment concentration.

10000 Average Daily Flow (cms) SSC (mg/L) -1

Flow (cms) - SSC (mgL )

1000

100

10

1

0.1 5/1/2008

Figure 4.11

7/9/2008

9/16/2008

11/24/2008 Date

2/1/2009

4/11/2009

Mean daily flow and suspended sediment load at the USGS station on Town Creek near Nettleton, MS

66

Continuous automatic sampling after February 1, 2009 showed that high suspended sediment concentrations occurred during runoff events. The following suspended sediment transport relation (R2=0.94) was developed for Town Creek near Nettleton, MS encompassing the 12 months of grab sampling monitoring and 4 months of continuous automatic sampling at the USGS 02436500 gauging station (Figure 4.12):

Q s  0.408Q 1.864

Eq. 4.2

where Qs is the suspended sediment load (Mg d-1) and Q is the flow discharge (m3 s-1). The coefficient in Equation 4.2 representing the load at a low flow or base flow is an indication of background levels of the channel system (Simon et al., 2002), and the exponent representing the rate of increase in load is an indication of sediment availability in the watershed and channel system with increasing flow (Kunhle and Simon, 2000). The coefficient value of 0.408 suggests low availability of suspended sediment when low flow or base flow conditions occur in the stream. The exponent value of 1.864 is representative of degrading Stage III toward Stage IV channels in the channel evolution conceptual model of Simon (1989) (see Figure 4.1). The previous analysis within Town Creek watershed showed wash load contributions from agricultural lands near streambanks, degrading tributaries and streambank failures. These sources represent different agents of erosion that could be present during the degradation phase. Dominant channel processes, especially at the northern headwaters, are characterized by gravitational failures, basal erosion and pop-out failures on streambanks. Streambed material transport is dominant during Stage III of channel evolution because streambed degradation rates are greatest.

67

Using the estimated effective discharge Q1.5= 636 m3 s-1 for the 1606 km2 of drainage area at the site of study, average suspended sediment load and suspended sediment yield are 68,547 Mg d-1 and 40 Mg d-1 km2, respectively. Daily and cumulative suspended sediment loads were determined by integrating the transport curve for the studied period (Figure 4.13).

1000000 100000

-1

Qs (T d )

10000 1000 100 10 Qs = 0.408Q1.864

1

R2 = 0.94 0.1 0.1

1

10

100

1000

Q (cms)

Figure 4.12

Relation between sediment load (Qs) and instantaneous flow discharge (Q) for 128 gauging station records at the USGS gauging station 02436500

A cumulative suspended sediment load of 387,600 Mg was estimated from May 1, 2008 to May 15, 2009. Approximately 30% of the total suspended sediment discharge was produced in two consecutive days during the stormflow event which peaked on December 10, 2008. Significant stormflow events also occurred on March 26 and May 6, 2009 contributing approximately 25% of the total discharge. Peak flows were a major factor in exporting suspended sediment from Town Creek watershed. Flow conditions in this study showed that the relative rapid rise to peak 68

flow resulted in corresponding high sediment yields during the rising stage and generally a higher drop in suspended sediment during the slower falling stage. The presented suspended sediment data may be indicative of relative suspended sediment discharge and

Daily Suspended Sediment Load (T))

Thousands

yields typical for Southeastern Plains Ecoregion watersheds.

400 350 300

388

Cumulative Suspended Sediment Load (T) May 2008 - Apr 2009 Cumulative Suspended Sediment Load (T) May 2009

279 250 219

200 150

142 129

100

90

50 23 0 5/1/2008

7/9/2008

9/16/2008

11/24/2008 Date

2/1/2009

4/11/2009

1000000 Daily Suspended Sediment Load (T) Daily Suspended Sediment Load (T))

100000

45,876 52,208 38,607

10000 1000 100 10 1 0.1 5/1/2008

Figure 4.13

66,561

7/9/2008

9/16/2008

11/24/2008 Date

2/1/2009

4/11/2009

(a) Cumulative suspended sediment load and (b) daily suspended sediment load at the USGS gauging station 02436500 on Town Creek estimated by integrating the suspended sediment transport curve from May 1, 2008 to May 15, 2009 69

Analysis of Suspended Sediment Transport Rating Relations A suspended sediment transport rating curve was developed for Town Creek near Nettleton, MS encompassing a continuous 16-yr period of records from the USGS gauging station 02436500 and a 1-yr period of automatic sampling at the same location (Figure 4.2):

Q s  2.968Q 1.6536

Eq. 4.3

where Qs is the suspended sediment load (Mg d-1) and Q is the flow discharge (m3s-1). The rating curve has an R2=0.93. According to the conceptual channel evolution model of Simon (1989) the obtained slope coefficient is representative of a degrading Stage III channel. Suspended load contributions could come from agricultural lands near streambanks, degrading tributaries and streambank failures. Those sources represent different agents of erosion that could be present during this phase. Dominant channel processes are illustrated by basal erosion and pop-out failures on streambanks, and riparian vegetation may lean towards the channel. It is expected that Stage III reaches manifest greatest streambed-material discharges because this is the stage when maximum rates of streambed degradation occur. Using the estimated effective discharge Q1.5 = 636 m3s-1 for the 1606 km2 of drainage area at the site of study, an average annual suspended sediment load and annual suspended sediment yield of 127,700 Mg d-1 and 80 Mg d-1 km-2 were calculated respectively. Table 4.2 shows the comparison among suspended sediment transport relations obtained by previous studies at the same location and the present study results. A significant difference (alpha=0.05) was not observed between the slopes and 70

coefficients included in the expression proposed by Simon et al. (2002) and obtained in the present analysis. The estimated suspended sediment load and suspended sediment yield calculated using both expressions were very similar. This was expected although in the present analysis a new dataset of 128 records was included. The regional model of Sharp (2007) estimated suspended sediment load and suspended sediment yield values about six times smaller. Sharp’s regional suspended sediment rating curve is based on all gauging stations within the Tombigbee River Basin, and suspended sediment load rates for a similar flow discharge can vary up to 4 orders of magnitude. The reference yield for Southeastern Plains Ecoregion watersheds including Town Creek at the Q1.5 was determined by Simon et al. (2002) as 3.9 Mg d-1 km-2. For the drainage area at the USGS station location, a suspended sediment load of 6270 Mg yr-1 was previously estimated. The estimated suspended sediment yield from Town Creek for the analyzed period was about 20 times (one order of magnitude) greater than the reference condition at the effective discharge (see Table 4.1). An annual suspended sediment load of 1,073,500 Mg yr-1 and annual suspended sediment yield of 668 Mg yr-1 km-2 were obtained after integrating the suspended sediment rating curve using average daily flow discharges for Town Creek near Nettleton, MS. The estimated values using the regional model are about four times smaller in this case.

71

Table 4.2

Comparison of suspended sediment transport relations at the USGS gauging station 02436500 (Town Creek near Nettleton, MS) determined by different authors showing calculated load and yield estimated at Q1.5=636 m3 s-1 and drainage area A=1606 km2

Model 72 Entire dataset (1401 records) Regional model (Sharp, 2007) NSL (Simon et al., 2002)

Tier 1 Effective flow (Q1.5) Yield Load -1 (Mg d ) (Mg d-1 km2)

Tier 2 Daily Flow Load** Yield -1 (Mg yr ) (Mg yr-1 km2)

0.93

127,728

80

1,073,433

668

1.259

0.84

21,923

14

257,051

160

1.630

*

124,731

77

1,054,556

657

Rating curve coefficient a

Rating curve exponent b

R2

2.968

1.653

6.485 3.360

*Not reported by authors **Average annual value for continuous estimation from January 1 1981 to March 30 2008 and May 1 2008 to April 30 2009

Annual suspended sediment transport relations, suspended sediment loads and suspended sediment yield for the USGS station 02436500 are shown in Table 4.3. The average daily suspended sediment load and sediment yield at Q1.5 determined from each annual expression were 166,700 Mg d-1 and 104 Mg d-1 km-2, respectively. Estimated suspended sediment load and yield at effective flow for 1988 were the lowest for the period of analysis. The lowest values observed could be caused by significant low flow conditions observed during that entire year and the small range in instantaneous flow discharge (< 12 m3s-1) and suspended sediment loads (67 Mg d-1) recorded in the database. The estimated rating curve slope for this year was the lowest (close to 1) (but not the highest intercept) of all annual suspended sediment transport curves. The observed slope could be considered an indication of low sediment availability in the watershed and channel system with increasing flow. For that specific year, the suspended sediment load and suspended sediment yield at the effective discharge were lower than the “reference” rate of suspended sediment transport suggested by Simon et al. (2002) (Figure 4.14). Figure 4.15 compares the annual suspended sediment loads calculated by integrating the suspended sediment rating curve based on (1) the entire period of record (Equation 4.3) and (2) annual records. The annual suspended sediment loads using the general transport curve (Equation 4.3) are similar to those estimated by using each annual transport curve. For years with a reduced amount of cumulative annual flow discharge the suspended sediment transport rates estimated by the general transport equation were about 60% to 90% of those obtained by using the respective annual transport curve.

73

Table 4.3

Annual suspended sediment transport relations at the USGS gauging station 02436500 (Town Creek near Nettleton, MS) showing estimated load and yield at the determined Q1.5=636 m3s-1 and drainage area A=1606 km2

Year

74

1981* 1982 1983 1984 1985 1986 1987 1988** 1989 1990 1991 1992 1993 1994 1995 2008 - 2009

Rating Rating curve curve coefficient exponent a b 7.389 4.639 11.320 5.102 7.818 3.182 2.635 4.920 3.776 3.052 3.405 2.730 1.041 2.804 1.678 0.408

1.633 1.592 1.423 1.568 1.532 1.676 1.807 1.018 1.509 1.686 1.609 1.853 1.898 1.690 1.748 1.864

R2

0.98 0.95 0.96 0.96 0.92 0.94 0.96 0.82 0.93 0.96 0.92 0.88 0.99 0.87 0.96 0.94

Effective discharge (Q1.5) Yield Load (Mg d-1 (Mg d-1) 2 km ) 280,000 174 135,100 84 110,100 69 126,700 79 154,500 96 158,900 99 307,000 191 3,500 2 64,400 40 162,400 101 110,400 69 426,700 266 218,300 136 152,800 95 133,700 83 68,600 40

Daily flow discharge (Q) Yield Load (Mg yr-1 (Mg yr-1) km2) 358,600 223 2,102,400 1,310 2,649,100 1,650 752,800 469 600,500 374 703,000 438 791,900 493 24,900 16 733,100 456 1,785,100 1,112 3,549,300 2,210 1,065,800 664 419,700 261 1,094,000 681 759,400 473 262,400 163

* Instantaneous flow discharge for suspended sediment records always lower than 26 m3 s-1. ** Instantaneous flow discharge for suspended sediment records always lower than 12 m3 s-1

The annual suspended sediment loads varied between 24,900 Mg yr-1 and 3,549,300 Mg yr-1 when the annual cumulative suspended sediment load was estimated. Also in this analysis the lowest load and yield occurred in 1988.

1000000 2008-2009 Early 90's (1990-1994)

1.6536

Qs = 2.968Q 2 R = 0.93

Late 80's (1985-1989)

100000

Early 80s (1980-1984) Entire dataset

Qs (tons/yr)

10000

1000

100

10

1

0.1 0.1

1

10

100

1000

Q (cms)

Figure 4.14

Relation between sediment load (Qs) and instantaneous flow discharge (Q) for 1401 gauging station records at the USGS gauging station 02436500 (Town Creek near Nettleton, MS)

The values for the slope and the intercept obtained for each annual suspended sediment transport relation at the USGS gauging station 02436500 are shown in the Figure 4.16. Over the study period an increasing trend in the rate of increase in load was observed in combination with a trend towards reduction in the load at low or base flow conditions. The increase in the rate of increase in load reflects stream channels that transition from Stage III toward the Stage IV in the conceptual channel evolution model of Simon (1989). This condition indicates streambank erosion and streambed instability processes that result from changes in channel gradient and channel morphology. Incised channels (V-shaped) are observed upstream transition to broad U-shaped channels. 75

10,000,000

-1

Suspended Sediment Load (Tyr-1)

Qs=1,102,687 T yr

1,000,000

Qs=1,073,433 T yr

-1

100,000

Annual suspended sediment load estimated from entire dataset relation (Q) (T/yr) Average Annual suspended sediment load estimated from entire dataset relation (Q) (T/yr) Annual suspended sediment load estimated from individual annual relations (Q) (T/yr) Average Annual suspended sediment load estimated from individual annual relations (Q) (T/yr)

10,000 1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

Year

Annual suspended sediment loads at USGS gauging station 02436500 (Town Creek near Nettleton, MS) determined by integrating two different suspended sediment rating curve: (1) suspended sediment rating curve based on the entire period of record (Equation 4.3), and (2) annual suspended sediment rating curves

2.0

20

1.9

19

1.8

18

1.7

17

1.6

16

1.5

15

1.4

14

1.3

13

1.2

12

1.1

11 Rate of increase in load (b) Load at low/base flow (a)

1.0 0.9

10 9

0.8

8

0.7

7

0.6

6

0.5

5

0.4

4

0.3

3

0.2

2

0.1

1

0.0 1980

Load at a low/base flow (a)

Rate of increase in load (b)

Figure 4.15

0

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

Year

Figure 4.16

Rate of increase in load (rating curve slope parameter) and loads at low/base flow (rating curve intercept parameter) obtained for each year with recorded data at the USGS gauging station 02436500 (Town Creek near Nettleton, MS) 76

A significant correlation (alpha=0.05) and an exponential expression with a negative slope were determined between the suspended sediment load at low or base flow and the rate of increase in load (Figure 4.17). A similar condition was found by Asselman (2000). The pair of parameters for the year 1988 was determined as an outlier and was not included in the determination of the correlation. The correlation between annual average instantaneous flow discharge and the rate of increase in load was only observed with cluster analysis. For each individual cluster, as the magnitude of the annual average instantaneous flow discharge (Q) increased, the rate of increase in load (b) was reduced. At the same annual average instantaneous flow discharge, the relative efficiency of the stream to transport increasing amounts of suspended sediment with increasing flow rates increased in time. Although the cluster including the later records had a higher reduction rate of b with increasing Q (slope), this parameter did not show significant difference (alpha=0.05) among clusters (Figure 4.18). The cluster analysis showed more clearly an increase in the rate of increase in load (slope or exponent of the suspended sediment rating curve) in time, which reflects stream channels in transition from Stage III toward the Stage IV in the conceptual channel evolution model of Simon (1989) (Figure 4.19).

77

12 11

Load at low or base flow (Intercept)

10 9 8 -7.4581

a = 148.39b 2 R = 0.77

7 Outlier (1988)

6 5 4 3 2 1 0 1.0

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.9

2.0

Rate of increase in load (Slope)

Figure 4.17

Relation between the annual suspended sediment load at low or base flow and the annual rate of increase in load at USGS gauging station 02436500 (Town Creek near Nettleton, MS)

2.0 Early 90's (1991-1995) and 2008-2009 Late 80's (1986-1990)

1.9

Early 80's (1981-1985)

Rate of increase in load (b)

1.8 1.7 1.6 1.5 1.4 1.3

b = 1.8544Q-0.0554 R2 = 0.54

1.2

b = 2.5514Q-0.1304 R2 = 0.61

1.1

b = 2.7278Q-0.1277 R2 = 0.85

Outlier (1988)

1.0 0

10

20

30

40

50

60

70

80

Annual average intsantaneous flow (m3s-1)

Figure 4.18

Clustered relation between annual average instantaneous flow (m3s-1) and the rate of increase in load obtained for each year with recorded data at the USGS gauging station 02436500 (Town Creek near Nettleton, MS)

78

3.0

Rate of increase in load (b)

2.5

2.0

1.5

1.0

Early 80's

0.5

Late 80'2 Early 90's 2008 - 2009

0.0 I

II

III

IV

V

VI

Stage of Channel Evolution

Figure 4.19

Variation in suspended sediment transport efficiency (rating curve exponent) for different periods between 1981 and 1995 and May 2008 to April 2009 in the Town Creek watershed

Because there is a lack of data on suspended sediment loads from Town Creek watershed for the last 14 years, it is difficult to make a statement about future suspended sediment yield trends. The estimation of suspended sediment loads using Equation 4.3 showed a trend toward reducing the annual loads. Transport curves determined for each of the four clusters used in the determination of change in time of suspended sediment loads from Town Creek watershed are shown in Figure 4.20. The comparison of slopes among clusters (alpha=0.05) confirmed that a significant increase in the rate of increase in load for Town Creek watershed occurred over the time (Table 4.4). This behavior would suggest an increase in time of the sediment availability, transport capacity and probable suspended sediment loads in the watershed and channel system with increasing flow. However, at the same time the suspended sediment load amount under low or base flow was reduced. Both conditions can be explained by the difference in stages of 79

channel evolution presented in different sections along the Town Creek. The rising relative efficiency of the stream to transport increasing amounts of suspended sediment with increasing flow rates within Town Creek could be attributed to the increased contribution of significant amounts of suspended sediment loads (up to 70%) from agricultural lands, with high gully erosion and streambank erosion activities from the incised unstable (Stage III and IV) channels in the northern and western headwaters. The middle 20 km and the downstream meandering 10 km along the wide vegetated stable principal reach favor stream energy reduction, suspended sediment deposition on streambanks during and after storm flow events, seasonal streambed mobility, and suspended sediment concentration reduction (by increase in flow). This could explain the temporal reduction of suspended sediment loads at low flow conditions. Also, a trend toward the reduction of the annual average instantaneous flow was observed from the analysis of those records (not shown). That should be considered as another factor affecting the significant reduction in time of suspended sediment contribution at a specific flow rate between 1981 and 2009 from Town Creek watershed. The observed relation between the magnitude of the suspended sediment transport curve parameters and the trend toward reducing suspended sediment loads was similarly observed by Yang et al. (2007).

80

Table 4.4

Suspended sediment transport relations at USGS gauging station 02436500 (Town Creek near Nettleton, MS) determined for different periods between 1981 and 1995 and May 2008 to April 2009, showing calculated load and yield at the Q1.5=636 m3s-1 and drainage area A=1606 km2

Rating curve exponent b 1.545

R

Load (Mg d-1)

Yield (Mg d-1km2)

Load* (Mg yr-1)

Yield (Mg yr-1km2)

Early 80's (1981-1985)

Rating curve coefficient a 6.375

0.97

106,433

66

1,360,106

847

Late 80's (1986-1980)

3.543

1.597

0.93

136,274

85

682,922

425

Early 90's (1991-1995)

2.563

1.709

0.94

158,336

99

1,612,386

1004

May 2008-April 2009

0.408

1.864

0.94

68,597

40

262,405

163

Entire dataset (1401 records)

2.968

1.654

0.93

127,728

80

1,073,433**

668

Model

Effective Flow (Q1.5) 2

Daily Flow

81

*Average annual value for the years in the cluster. **Average annual value for continuous estimation from January 1 1981 to March 30 2008 and May 1 2008 to April 30 2009.

1000000 Early 80's Late 80's Early 90's

100000

2008 - 2009

Qs (tons/yr)

10000

1000

Qs = 6.375Q1.5319 R2 = 0.97

100

Qs = 3.543Q1.5967 R2 = 0.93 10

Qs = 2.563Q1.7089 R2 = 0.94 Qs = 0.408Q

1

1.864

R2 = 0.94

0.1 0.1

1

10

100

1000

Q (cms)

Figure 4.20

Relations between suspended sediment load (Qs) and instantaneous flow (Q) for different time periods (clusters) at USGS gauging station 02436500 (Town Creek near Nettleton, MS)

Conclusions Data collected in the Town Creek watershed within the Southeastern Plains Ecoregion in Mississippi showed high spatial and temporal variability of flow and suspended sediment concentrations, discharges and yields in the watershed. Clearly, incised channels in the northern and western area are the major producers of sediment within the Town Creek watershed. The suspended sediment yield was significantly reduced at the watershed outlet when compared with sediment production from headwaters. Reduction could be attributed to the presence of natural and established sediment control structures, as well as for the presence of wide stable vegetated channels and significant increase in flow (suspended sediment dilution) along the middle and lower area of the watershed. Mass balance analysis showed that erosive processes 82

dominate suspended sediment transport in northern headwater channels, while deposition tends to control the sediment transport process in downstream channels. Average suspended sediment loads at the USGS gauging station 02436500 (Town Creek near Nettleton, MS) over this 29-yr period are about 1,000,000 Mg yr-1. Estimated suspended sediment yield at the effective flow (Q1.5) is 80 Mg d-1 km-2. Suspended sediment yield at the last studied year 2008-2009 is reduced to about 40 Mg d-1 km-2. Annual suspended sediment loads near the outlet of Town Creek watershed, Mississippi, between 1981 and 2009 are trending downward. However, streambank erosion processes at the headwaters appear to be acting as the most significant sediment supplier and need to be reduced under a stream restoration process. Adoption and implementation of management practices such as streambank and riparian buffer zone restoration and establishment of other BMPs are necessary within Town Creek watershed to have measurable improvements in water quality. Reduction of suspended sediment loads should focus on the attenuation of geomorphic processes, and stabilization of reaches and agricultural lands near streambanks at the northern headwaters within Town Creek watershed. Regional models based on data from Southeastern Plains Ecoregion watersheds significantly underestimate suspended sediment loads and suspended sediment yields in the Town Creek watershed. Annual suspended sediment transport relations are significantly affected by the occurrence of low flow periods. The relative efficiency of Town Creek to transport increasing amounts of suspended sediment with increasing flow rates has increased in time. However, relative efficiency reduces with increasing instantaneous flow. 83

Temporal analysis of suspended sediment transport relations from 1981 to 2009 show a reduction in the amount of suspended sediment loads contributed by Town Creek watershed at a specific instantaneous flow. The reduction is higher under low flow conditions (< 10 m3s-1). Rising relative efficiency of Town Creek to transport increasing amounts of suspended sediment with increasing flow rates show the high erosion potential of important geomorphic processes in a specific area of the watershed. This condition can be inferred from considering that suspended sediment loads at low flow or base flow conditions have decreased, and considerably less sediment at a given discharge was transported in the last decade than in earlier years. Current relative efficiency of the Town Creek to transport increasing amounts of suspended sediment with increasing flow rates is attributed to active geomorphic processes at the northern 320 km2 of the watershed. Temporal reduction in suspended sediment loads at low flow or base flow conditions in Town Creek are attributed to the presence of wide stable vegetated channels and the significant increase in flow (suspended sediment dilution) along the middle and lower areas of the watershed. Over continuous periods of time (exceeding two decades), differences up to one order of magnitude can be achieved using a single suspended sediment transport relation based on the entire period of record instead of annual suspended sediment transport relations. Larger differences occur under low average annual instantaneous or daily flow conditions. The suspended sediment transport rating curves obtained for different periods at station #8 within the Town Creek watershed describes a sediment delivery directly produced by streambanks with a trend toward the increase of the sediment supply. 84

However, the reduction in time of the sediment loads evidences that this sediment delivery is predominantly supplied by specific areas from the headwaters reaches within the watershed, and after a specific zone, the variation in channel characteristics (vegetation, slope, area, shape) is favorably reducing the sediment export of the Town Creek watershed.

Recommendations Current and historic suspended sediment transport curve parameters for Town Creek watershed can be compared using annual average instantaneous or daily flow. The parameter estimation allows the establishment of a correlation between annual average instantaneous flow and the rate of increase in load and a correlation between the suspended sediment transport curve parameters. A joint analysis of field reconnaissance, suspended sediment transport relation parameters, flow and suspended sediment loads trend and changes in time can be used to determine different stages of channel evolution of reaches or areas inside a watershed.

85

CHAPTER V ASSESSMENT OF STREAMBANK EROSION RATES The overall goal of this research is to evaluate streambank erosion rates and generate empirical correlations for estimating streambank erosion involving physical and geomorphic variables influencing this process for streams in the Southeastern Plains Ecoregion in Mississippi. This chapter presents measured streambank erosion data from Town Creek at three different locations along the main channel. The main aim of this second study is to investigate the spatial and temporal pattern of streambank erosion processes over the monitored places. Complimentary, the study individually evaluates two field monitoring strategies to assess streambank erosion rates.

Introduction Several research efforts have demonstrated the important contribution that streambank erosion can have on total sediment loading in streams (e. g. Simon and Darby, 1999; Simon et al., 2002). Streambank erosion is one of the most dynamic geomorphological processes of significant interest for scientists, while the detailed understanding of this process has considerable economic value and engineering importance. However, the study of streambank erosion processes as well as the interest to determine and assess streambank erosion rates did not have relevance until the middle 1970s (Lawler, 1993; Thorne et al., 1997). The early pioneering work of Wolman (1959), 86

Schumm and Litchy (1963) and Twidale (1964) emphasized the rapidity of streambank erosion, and the complexity of its processes. Later and current contributions have demonstrated the wide range and interplay of processes responsible, as well as the significant environmental controls (Lawler, 1993). The assessment of streambank erosion rates and processes can have different applications. It can provide site specific information in a watershed or basin that could be used to target streambank rehabilitation or riparian restoration programs; it also can be used to evaluate and validate the consistency of a sediment budget model that predicts streambank erosion rates at watershed scales. The best way to quantify streambank erosion is to measure it directly in the field. Different techniques have been used at different places around the world to quantify streambank erosion and channel change. The diversity and proliferation of techniques is due to the worldwide large variety of stream environments; the differing interests, focus and spatial and temporal variation of performed research; and the professional background of researchers. No standard techniques have been established for the determination of streambank erosion rates. Although some innovative techniques have been recently adapted to assess in situ streambank erosion and channel widening rates (e. g. Photo-electronic erosion pin (PEEP) system, light detection and ranging (LIDAR) scanning), researchers still use basic techniques existent 50 years ago (e. g. Pin erosion, cross section survey, planimetric resurvey). A classification of techniques in terms of resolution and timescales of channel change for which they are appropriate was initially proposed by Lawler (1993). Lawler’s classification assigned each method to one of three overlapping temporal categories (long, intermediate, and short timescales). Most recent 87

methods not included in the original framework will be classified herein to allow their description. Lawler’s framework was intended to assist in the selection of appropriate technique(s), depending upon (a) the timescale needed about the streambank erosion rate and channel change information, and (b) the needed resolution of measurement. Only techniques for intermediate and short timescale are presented in this document.

Intermediate Timescale Techniques Two techniques are grouped on this timescale, which are considered appropriate for the investigation of streambank erosion and channel change over periods from one year to 30 years, although usually at a fairly low temporal resolution (Lawler et al., 1997). Planimetric resurvey. This technique involves a periodic survey of the position of the streambank edge using either (a) simple offsets from a fixed benchmarked baseline, (b) the compass and tape technique or (c) a theodolite, an electronic or laser distance meter, a total station or a real time kinematic (RTK) GPS system. Streambank line migration rates can be determined by the superimposition of the successive field surveys. This technique will only provide evidence of changes to the streambank line position, and not progressive undercutting of the streambank below the flood plain level (Lawler et al., 1997). Problems in the identification of the streambank top usually is the leading reason for errors by using this technique. Repeated cross section surveys. The repeated survey of a series of permanently marked cross sections or transects through the stream is a technique that reveals streambank recession and changes that may be occurring in other parts of the section 88

besides the streambanks, such as accretion of point bars, streambed form development or streambank undercutting. The profiles generated at the different surveys are superimposed to reveal the channel changes over time. Methods of surveying stream cross sections involve the use of levels, theodolites, total station, inclinometers, datum technique, electronic distance meters, airborne laser altimetry and real time kinematic GPS system among others. The more often surveys are done, the more data on seasonal or event changes to channel form and streambank retreat are generated. Temporal variations in rate are often crucial to correctly identify the processes leading to streambank erosion and the critical geometry for streambank failure (Lawler et al., 1997). When cross section resurvey induces the identification of those processes, then it becomes a technique reaching objectives from short time scale techniques.

Short-Time Scales Techniques Additionally to the estimation of streambank erosion rates or deposition, short time scale techniques also help to identify the causes, processes and mechanisms inducing that activity. These techniques also allow relating hydraulic, hydrological and meteorological events or event combinations to the erosional or depositional activity. Three techniques grouped on this timescale are considered appropriate for the investigation of streambank erosion and channel change over a time resolution of minutes, days, months or years. Erosion pins. Erosion pins are rods driven perpendicularly to the steep face of a streambank. The length of rod left exposed on the streambank surface after a specific period of time (weekly, monthly, yearly) or after a stormflow event, less the amount left 89

exposed when it was installed, defines the amount of surface erosion which has occurred locally. Early streambank erosion research using pin erosion was performed by Wolman (1959). This technique could be considered the most frequently used among the others in this group due to its simplicity, sensitivity, low cost and suitability for a wide range of fluvial environments. Limitations of this technique include its application on streambanks undergoing mass failure or severe erosion rates. Because this is a point specific technique, the derivation of areal, volumetric or gravimetric estimates of streambank erosion/deposition rates should be carefully taken due to the longstream, vertical and random spatial variability that could occur. Photo-electronic erosion pin (PEEP) system. This technique allows the automatic monitoring of streambank erosion and deposition processes, identifying clearly the timing, magnitude and frequency of the occurring process at the site. This method differs from other short timescale and intermediate timescale techniques like erosion pins and resurvey, which can only reveal net change to a site since the previous field visit, without therefore revealing the quasi-continuous temporal distribution of erosional and depositional activity. The PEEP sensor is an optoelectronic device composed of an array of photosensitive cells enclosed within a clear acrylic tube, inserted into the streambank face. The PEEP is connected to a datalogger by cable to provide a near-continuous record of streambank retreat or advance at one specific location. The analog millivolt signal emitted by the PEEP is a function (directly proportional) of its exposed length to the natural daylight. Erosion, therefore, increases voltage outputs, while deposition decreases outputs. The PEEP resolution was reported by Thomas and Ridd (2004) to be within ca. 2 mm, whilst accuracy was not specified. However, its resolution is not satisfactory for 90

very small-scale measurements, scouring could occur around the instrument due to hydrodynamics disturbance, and long deployments may be limited by fouling over the sensors. Distances between sites may vary from tens of meters to kilometers depending on the objectives of the survey, with the spatial resolution decreasing accordingly. Terrestrial light detection and ranging (LIDAR) scanning. The LIDAR scanner is an active remote sensor that emits a light pulse towards a target surface and then records the relative distance to objects from the scanner. A 3-D projection of the scanned area that includes light intensity and a RGB color range from a digital component, acting as a digital photographic camera, is produced (Resop and Hession, 2010). This technique has offered quick data collection, but is limited by the need for post processing algorithms (Hopkinson et al., 2004). Some effects of vegetation obscuring the ground surface have been observed. However, some LIDAR pulses have the ability to penetrate the vegetation and return a signal from the ground surface. The presence of water also poses significant problems for the interpretation of the laser data with partial penetration occurring in clear shallow water and high laser pulse incidence angles and beam refraction off the water surface at reduced incidence angles (Heritage and Hetherington, 2007). Two different methods were used to determine the rates of streambank erosion and channel change along three different locations on the Yonaba Creek and the Town Creek. The locations of the study site were based on a number of factors including (1) representativeness of the evaluated segment along the channel and (2) access to the streambank and landowner permission. For this study, it was not considered appropriate to compare the streambank erosion rates measured using repeated cross sections survey with those from pin erosion 91

plots, as both methods were employed measuring streambank erosion at different scales of time and space at different locations within the watershed.

Cross Section Survey at the Yonaba Creek Methods Eight transects (0, 20, 135, 140, 165, 210, 270 m) along a 270 m reach on the Yonaba Creek near the bridge at the Road 9 in Blue Spring, MS, a representative principal northern headwater tributary for the Town Creek (Figures 5.1 to 5.3), were periodically surveyed from late February 2009 to March 2010. The reach presented a bendway section about 50 m radius along ~100 m length. The reach evidenced streambank erosion processes from incised unstable streambanks and gully erosion processes from agricultural lands near streambanks with limited riparian vegetation (Figure 5.4), a typical scenario observed at the northern and western area of the entire watershed. The studied reach receives stream flow from Brown Creek and Hell Creek streams, which incised channels are actively widening due to processes of streambank erosion and streambank instability. A pond upstream of each tributary was observed at 1.6 km and 2.5 km from Brown Creek and Hill Creek, respectively. The stream at transect 0 m drains an approximated area of 32.6 km2. The streambank profile along the entire 270 m reach is very homogeneous, which very fine sandy loamy soils. Evidence of sandy loose material on the streambank surface was observed in some segments and more commonly along the right side streambanks. Initial cross sections were surveyed between left to right streambank looking upstream by using a Total Station Positioning System (Leica TSP 1100 Professional 92

Series) on February 27 and 28, 2009. A Real Time Kinematic (RTK) GPS system (TOPCON Hiperlite Plus) was used for the next surveys from May 19, 2009 (Figure 5.5). Surveying with the total station, unlike RTK-GPS surveying, was restricted to measurements between inter-visible points and the location of multiple control points among the survey area, a condition that made control propagation a time consuming task. Limitations to surveying with the RTK-GPS were related to the satellite signal reception by the equipment, which was not strong enough on very cloudy days, neither along areas where the communication between the station base and the rover base was limited by overhead obstructions (e. g. concrete materials (bridge) and trees), especially at the lower part of transects 0 m and 20 m. Transects were surveyed at smaller spacings along the streambank face and the streambed. Benchmarks (steel rods and flags) were set up upland at 1m and 3 m from the top of the left and right streambanks at each transect. After the first survey in late February, transects downstream of the bridge were surveyed 4 more times each (June 1, June 17 and October 8 in 2009, and March 18 in 2010). Two additional surveys were conducted at transects 90 m, 135 m and 165 m in June 1, 2009 and November 3, 2009. The first two transects located upstream of the bridge were only surveyed in late February 2009, June 1, 2009 and March 18, 2010 due to limitation in satellite signal restriction by the RTK-GPS. The change in area for each streambank side and the streambed at each transect and in each survey date was determined by using ARC VIEW 3.2 (ESRI) and the geoprocessing tools extension. The net change in area for each streambank side and streambed at each transect was determined by comparisons of the initial and the final surveys dates. The depth of streambank widening (streambank erosion) or contraction 93

(streambank deposition) was estimated for each survey by dividing the net change in area by the length of the streambank side slope from the top of the streambank. Assuming each transect was representative of a segment length, the total volume of sediment eroded or deposited on a specific streambank was estimated from the product of the net change in area at the streambank transect and its segment length. The volume was converted to a mass value by using the average streambank bulk density for each streambank side at each transect. Cross section resurveys merely reveal net change in the position of the streambank and streambed surface since the previous measurement without quantifying the temporal distribution of change between visits (Lawler, 1994). Therefore, field observations and photographic records were used to describe the temporal fluctuations or change that could occur between surveys dates.

94

1 2

ISCO autosampler and A-V device

3 4 5 6 7 8

95

N

Figure 5.1

Aerial view of the reach on the Yonaba Creek near the bridge at the Road 9 (Source: MS GIS Council)

8

7 6

5

4

3

2 1

Figure 5.2

Aerial view from upstream to downstream of the reach on the Yonaba Creek near the bridge at the Road 9. (Source : www.bing.com/maps)

a

b

c

d

Figure 5.3

ISCO autosampler and A-V device

View of the different transects along the 270 m reach on the Yonaba Creek near the bridge at the Road 9 in Blue Springs, MS. (a) Transects 0 m and 20 m, (b and c) transects 90 m to 165 m and (d) transects 210 m and 270 m 96

b

a

Figure 5.4

(a) Streambank erosion along incised unstable streambanks and (b) gully erosion processes from agricultural lands near streambanks with limited riparian vegetation at the northern area of the Town Creek watershed

a

c

Figure 5.5

b

d

Cross section surveying by using (a and b) Total Station Positioning System and (c and d) Real Time Kinematic (RTK) GPS System

97

Results As occurs with the major number of streams draining north and northeastern Mississippi, the principal headwater channel and most of the northern and western headwater tributaries draining the Town Creek watershed are incised channels experiencing streambank failures throughout their length. The heights of the streambanks along this reach ranged from 6.1 m to 7.1 m at the left side and from 5.0 m to 6.2 m at the right side of each transect (Table 5.1). The surveys of the eight transects are shown in the figures 5.6 to 5.13.

Table 5.1

Thalweg, streambanks height and streambank slope of transects along a 270 m reach on the Yonaba Creek near the bridge at the MS Road 9

Distance (m) (#1) (#2) (#3) (#4) (#5) (#6) (#7) (#8)

0 20 90 135 140 165 210 270

Segment length (m) 0-5 5 - 75 75 - 105 105 - 135 135 - 150 150 - 185 185 - 235 235 - 270

Streambank Height (m) Left Left 6.45 5.12 6.49 5.17 6.27 6.16 6.90 5.52 7.10 5.51 6.13 5.20 6.43 5.20 6.47 5.05

Thalweg (m) 89.56 89.29 88.98 89.16 89.33 88.98 89.27 89.29

* Slope at the net area of sediment deposition.

98

Streambank Slope (m m-1) Right Right 0.66 0.54 0.48 0.54 1.43 0.45* (0.17) 0.46 0.36* (0.16) 0.47 0.25* (0.12) 1.30 0.48* (0.13) 0.79 0.74 0.77 0.27* (0.37)

30.0

99

Elevation (m)

27.5

25.0

22.5

Feb 22 09 Jun 1 09 Mar 19 10

20.0 0

10

20

30

40

50

60

70

80

Distance (m)

Figure 5.6

Channel cross sectional surveys at transect 0m (0-5 m). The orientation of the survey is left streambank to right streambank looking upstream

30.0

100

Elevation (m)

27.5

25.0

22.5

Feb 22 09 Jun 1 09 Mar 19 10

20.0 0

10

20

30

40

50

60

70

80

Distance (m)

Figure 5.7

Channel cross sectional surveys at transect 20m (5-75 m). The orientation of the survey is left streambank to right streambank looking upstream

30.0

101

Elevation (m)

27.5

25.0

Feb 22 09 May 19 09 Jun 1 09 Jun 17 09 Aug 11 09 Oct 8 09 Nov 3 09 Mar 19 10

22.5

20.0 0

10

20

30

40

50

60

70

80

Distance (m)

Figure 5.8

Channel cross sectional surveys at transect 90m (75-105 m). The orientation of the survey is left streambank to right streambank looking upstream

30.0

102

Elevation (m)

27.5

25.0

Feb 22 09 May 19 09 Jun 1 09 Jun 17 09 Oct 8 09 Nov 3 09 Mar 19 10

22.5

20.0 0

Figure 5.9

10

20

30

40 Distance (m)

50

60

70

80

Channel cross sectional surveys at transect 135m (105-135 m). The orientation of the survey is left streambank to right streambank looking upstream

30.0

103

Elevation (m)

27.5

25.0

22.5 Feb 22 09 Jun 1 09 Jun 17 09 Oct 8 09 Nov 3 09 Mar19 10

20.0 0

Figure 5.10

10

20

30

40 Distance (m)

50

60

70

80

Channel cross sectional surveys at transect 140m (135-150 m). The orientation of the survey is left streambank to right streambank looking upstream.

30.0

104

Elevation (m)

27.5

25.0

Feb 22 09 May 19 09 Jun 1 09 Jun 17 09 Oct 8 09 Nov 3 09 Mar 19 10

22.5

20.0 0

10

20

30

40

50

60

70

80

Distance (m)

Figure 5.11

Channel cross sectional surveys at transect 160m (150-185 m). The orientation of the survey is left streambank to right streambank looking upstream

30.0

105

Elevation (m)

27.5

25.0

22.5

Feb 22 09 Jun 1 09 Jun 17 09 Oct 8 09 Nov 3 09 Mar 19 10

20.0 0

10

20

30

40

50

60

70

80

Distance (m)

Figure 5.12

Channel cross sectional surveys at transect 210m (185-235 m). The orientation of the survey is left streambank to right streambank looking upstream

30.0

106

Elevation (m)

27.5

25.0

22.5 Feb 22 09 Jun 1 09 Jun 17 09 Oct 8 09 Mar 19 10

20.0 0

10

20

30

40

50

60

70

80

Distance (m)

Figure 5.13

Channel cross sectional surveys at transect 270m (235-270 m). The orientation of the survey is left streambank to right streambank looking upstream

Transects 0 m and 20 m (0 m to 75 m) Transect 0 m is representative of the first 5 m segment and transect 20 m is representative of the next 70 m segment along the entire reach. From the total 75 m length of the segments, the first 50 m are located downstream of the bridge at the MS road 9 on Yonaba Creek and the last 25 m are located upstream. Both segments have trapezoidal channels with slopes 1V:2H. Some small trees, shrubs and grass were observed along the right side of the first 5 m segment, while the left side had only some grass patches and Kudzu vines. The last 70 m segment was revetted with rip-rap along both sides around 2003, when the bridge was constructed (Figure 5.14). A combined analysis of the limited number of cross section surveys for these transects, field observations and photographic records (see Appendix A) during the 12.5 month study suggested a cyclic change in streambed height. During the initial survey in late February 2009, loose sandy material was observed on the streambanks along both segments. Loose material have been deposited during the recession of the extreme high stormflows which occurred during the winter season (December 2008 and January 2009), and would be mainly supplied by the wasted streambank material from upstream incised channels (also by gully erosion and upland erosion), which would be induced through instability and mass failure processes by the hydrologic and hydraulic conditions occurring during the previous month. Loose sandy aggregates were observed not only at this location, but also along the streambank sides of the entire 270 m studied reach, the vegetated stable channels along the middle 20 km of the Town Creek and the meandering length before the watershed outlet.

107

Changes in streambed height in the first 5 m segment of the 270 m studied reach showed apparent streambed erosion process when the June 1, 2009 survey was compared with the initial survey in late February 2009 (Figure 5.6). The 70 m segment including transect 20 m evidenced apparent streambed deposition, while part of the small amount of sandy loose material previously deposited on and among the rocks on the revetted streambank segment was eroded during the same period of analysis (Figure 5.14). Streambed deposition along this segment would be favored by the formation of a pool along an approximated length of 35 m from the beginning of the revetted streambank, where a fallen tree in the stream was the upstream boundary, and rocks from the revetted streambank that moved and organized along the streambed formed a small dam at the downstream (Appendix A).

108

a

b

c

d

f

e

Figure 5.14

(a, b and c) Downstream view of transects 0 m and 20 m; (d) streambank deposition on transect 20 m (between abscises 5 m and 75 m) and (e and f) confluence of the Spout Spring Branch and the Hall Creek upstream of the 0 m segment along the 270 m reach on the Yonaba Creek near the bridge at the Road 9

Field observations and photographic records evidenced that the stormflow events occurring at the beginning of the spring season (March and April 2009), induced streambank retreat by fluvial erosion and planar failures of the loose sandy material deposited along this segment during the winter (Appendix A). As occurred during the

109

winter season, the continuous hydrologic activity during the first half of May 2009 and the occurrence of a significant high stormflow event on May 6, 2009 produced a significant amount of sediment supply from upstream and consequently, loose material deposition on the lower streambanks and the streambed along the studied segment during the stormflow recession. By the end of May and the beginning of June, the frequency, the duration and the magnitude of the stormflow events appeared to generate enough stream power to lower the streambed height below the level observed at the initial survey along the first segment. Neither the significant reduction in streamflow level, nor the reduced amount of stormflow events during the months of July and August induced a noticeable change in streambank or streambed height at this period. During the months of September and October the water depth and the effective flow area within the channel increased by the runoff. One more time, successive stormflow events induced streambank and streambed deposition, and only after the first stormflow event occurred in the month of December, streambed and loose deposited material on streambanks eroded. A fallen tree located a few meters from transect 0 m after the October 11, 2009 event initially induced streambed deposition along the initial segment, when it was located perpendicular to the streamflow direction. Later, after the first winter event on December, the stream repositioned the tree diagonally along the first segment. This reposition promoted the erosion of the right side of the streambed, producing a streambed bar behind the tree and eroded the right streambank, initially by fluvial erosion, later by undercutting. Finally, although recent stormflow events favored some deposition on the streambank toes and streambed, by March 2010 the streambed height on transect 0 m was lower than in June 2009 evidencing more degradation along the right side. The streambed on the second

110

transect was eroded from June 2009 to March 2010. This transect showed a minimum change between the initial survey on February 2009 and the last one done in March 2010.

Transects 90 m, 135 m, 140m and 160 m (75 m to 185 m) Transects 90 m, 135 m, 140m and 160 m were located along a 50 m radius and 110 m length bendway segment downstream of the bridge at Road 9 on the Yonaba Creek. Transects 135 m and 140 m were located along a grassed earth levee built in September 2008 by the National Resource Conservation Service (NRCS) to concentrate overland flow to the stream and control the formation of gullies on the upland field. Transect 140 m was surveyed on the location of a 0.5 m (20”) corrugated metal pipe that conducted the concentrated overland flow through the earth levee to the stream. The left streambank cross sections had slope 1V:1H at transects 90 m and 160 m and slope 1V:2H at transects located within the earth levee. No vegetation other than Kudzu and grass was observed along this side of the reach. The right streambanks slopes ranged between 1V:2H and 1V:4H where vegetation observed included small shrubs, some trees at the higher part of the streambank and Kudzu. Each transect was surveyed on seven dates except for transect 140, which was not surveyed on June 1, 09. On the first survey date, streambank faces at the left side of the reach showed evidence of recent mass wasting activity and basal clean out caused by streambank failures, while streambed at this side evidenced some channelized scour. That streambed scour could be caused by the association of two conditions: (1) the reduction of the transport capacity and the possible increase in stream velocity caused by the small dam created by the accommodation of rip rap along the upstream segment, and (2) by the

111

stormflow diversion generated by the debris accumulation at the right side of a pipe line transecting the beginning of the bendway segment (Figure 5.15). Loose sandy material was observed along the lower streambank areas in more significant amounts along the right side. Streambed sand bars along some segments evidenced an increase of the streambed height after the recession of the extreme high stormflows that occurred during the winter (December 2008 and January 2009) (Figure 5.15). The low temperatures that occurred normally during February and March appear to be the cause of an arid aspect visualized before the spring season on the vertical unvegetated streambank surfaces along the left side streambanks (Figure 5.16). This condition could imply that subaerial processes, which are climate related phenomena, could be acting toward the reduction of the streambank material strength. Small rills observed along the streambank height after the first stormflow events on April, implied the effect of the streambank material weakness, increasing the streambank material susceptibility to be eroded or affected by raindrops, streambank surface runoff (rills and small gullies formation), seepage (pop out and undercutting) and streamflow (fluvial erosion and undercutting) (Figure 5.16). Different authors (e.g. Thorne and Tovey, 1981; Prosser et al., 2000 and Wynn, 2006) consider that erosion of cohesive incised streambanks located at upper reaches of a stream system is dominated by subaerial processes, such as freeze-thaw cycling during winter and clay desiccation during summer. The survey performed in May 19, 2009, evidenced how aggressive were the stormflow events occurring at the beginning of spring (March and April 2009) along the bendway segment. The collected information does not permit a deep and detailed analysis

112

of the incidence of the subaerial processes on the streambank erosion along the studied reach. The observed freeze-thaw process caused by the climatic conditions during February and March 2009 could be considered a “preparatory process” that significantly favored the effect of the rainfall and stormflow events occurring during the spring season. The left side streambanks indicated an active period of instability and failure, except for transect 135 m, which evidenced some streambank toe deposition. It would indicate that the design characteristics (material density, slope and the armoring of the earth levee with grass) had been efficient in reducing the erosivity of the stormflows until this point (Figure 5.17). However, this efficiency was significantly reduced at transect 140 m, which showed an important morphologic change probably caused by the presence of the corrugated pipe, which could induce a difference in streambank material strength. It can be inferred that the armoring vegetation on the surface of the 135 m transect, limited the potential effect of the freeze-thaw cycling. Evidence of seepage erosion and piping erosion was observed along the entire bendway segment (Figure 5.18). The channelized scour on the streambed observed in the first survey along the left side was partially filled by wasted material from the streambanks at all transects. On the other hand, a streambed height decrease was observed along the left side of the segment. This condition could suggest that during the spring season the streambed erosion process occurred along the entire segment, but was not observed at the left side of the stream by the filling from the wasted material.

113

a

b

c

e

Figure 5.15

d

g

f

(a, b, c) Downstream and (d) upstream view of the bendway segment (transects 90 m, 130 m, 140 and 165 m); (e) Debris accumulation at the beginning of the bendway segment; (f) Deep water streamflow near the left streambank along transect 90 m; and (g) Deposited loose sandy material and streambed sandy bar caused by winter stormflow events during December 2008 and January 2009 along the bendway segment

114

a

b

c

d

(a) Snow remains on streambanks surface on May 3, 2009; (b and c) apparently action of subaerial processes on unvegetated streambanks surface; and (d) Effect of raindrops impact, surface runoff on streambank (small rills), seepage and streamflow on a streambank previously affected by subaerial processes (freeze-thaw cycling)

Figure 5.17

Difference of erosion degree for two different transects (130 m and 145 m) along an earth levee segment armored with grass

115

Figure 5.16

The survey completed on June 1, 2009 showed that the stormflow events occurred in May 22 and May 23, 2009 induced streambanks to fail and favored the deposit of the failed material on the streambank basal zone (Appendix A). Along the left side of the bendway transects, sand bars previously formed during the spring season started being eroded by fluvial erosion and planar failures. This process was observed more along the earth levee transects, where the perpendicular flow coming out from the corrugated pipe could create a turbulent zone that pushed the streamflow against the material deposited on the low right streambank, cutting the formed sand bar and shaping it to conform a vertical face (Figure 5.19). The summer stormflow events occurring at the beginning of the season affected channel morphology of the bendway segment, followed by a significant reduction in streamflow level. Streambank material previously deposited at the toe streambank was cleaned out, but stormflow recession could induce the deposition of sandy loose material bars at the left side. Fall stormflow events induced some mass wasting from unstable streambanks, streambed erosion occurred more aggressively at the left side, and streambank deposition at the right side streambanks, except for transect 140 m. By the beginning of November, the corrugated pipe had an exposed length of about 5 m and the side of the levee has been cut to reach the same height and slope of the segment downstream the levee (Appendix A). Morphologic changes on the slope side of the levee included the presence of rills and gullies, and pipes induced by seepage erosion. Changes in March 2010 included the cyclic increase of the streambed height noticed one year before after the winter events, although at this time, the streambank retreat was not high when compared with the survey done in November 2009.

116

Figure 5.18

Occurrence of seepage erosion on the bendway segment streambanks

a

b

Figure 5.19

(a) Effect of the metal pipe on the across right side streambank; and (b) sandy loose material deposition on streambank toe 117

Transects 210 m and 270 m (185 m to 270 m) Transects 210 m and 270 m were located along a 85 m straight segment after the bendway segment along the Yonaba Creek. The left streambanks had slopes of between 1V:1.2H and 1V:1.3H. Vegetation observed along this side of the segment included small shrubs, grass and Kudzu. The right streambanks slopes ranged between 1V:1.4H and 1V:4H, where observed vegetation included small shrubs, some trees at the higher part of the streambanks and Kudzu. Streambank erosion activity along the extension of this segment was reduced when compared with the bendway segment. However, some sporadic mass wasting activity was observed for a couple of events at transects different to those surveyed within the segment, caused principally by (1) soil preparation and crop seeding procedures very close to the streambank border, and (2) formation of edges of gullies by concentrated flow on the top of the streambank. The streambank morphology and the observed vegetation reduced the contribution of streambank material into the stream at these locations (Figure 5.20). As observed in the two segments upstream, deposits of sandy loose material along the lower streambanks, and some sand bar formations on the streambed were present at different transects at the date of the first cross sections survey in late February 2009 (Figure 5.17). The streambed at transect 210 m was relatively flat and inclined to the right side, possibly because this transect was located at the outlet of the bendway and the streamflow moves along the bend with some preference along the right side. Streamflow and stormflow events observed during the beginning of the spring season washed out material deposited on sand bars along the streambed, and induced low fluvial 118

erosion from the streambanks at transect 210 m. A very small amount of loose material was deposited on the streambanks and streambed at the 270 m transect, because a sand bar with important accumulation of material located around 5 m downstream of transect induced velocity reduction of stormflow events favoring sediment deposition under their recession period. During summer, the level of water depth and the low supply of sediment from upland and streambanks favored the erosion of the streambed along the entire segment. Vegetation had growth on the deposited sand bars and even had been extended on the streambed surface when streamflow was very low at the end of July. Fall and winter stormflow events were the most important suppliers of loose material deposited along the streambanks and streambed at the last transect of the segment. Vegetation grown on sand bars and streambed in the previous season favored the deposition process. The bendway connection affected the streambed height at the right side of the channel favoring erosion and lower streambank undercutting during the last two seasons of the year (Figure 5.22). To conclude the period of analysis, streambanks and streambed suffered erosion under the conditions observed at the beginning of the spring (March 2010).

119

Figure 5.20

View of transects 210 and 270 m (between abscises 185 m and 235 m) along the 270 m reach on the Yonaba Creek near the bridge at the Road 9

120

Figure 5.21

Lower streambank undercutting along the right side of the straight segment after the bend

Streambank Erosion Rates Assessment The quantification of streambank erosion widening, rates and depths along the entire reach, confirmed the observations described by visual inspection of the surveys, and by the photographic record and field recognition. It evidenced more streambank activity by gravitational failures, channel erosion and channel expansion along the left streambanks of transects 90 m, 135 m, 140 m and 165 m, located along the bendway, than on transects 210 m and 270 m, located along the straight section after the bendway. Some sediment deposition was observed along the left streambanks, more noticeably along transects 165 m and 210 m. Although the distance between transects 135 m and 140 m, located on an earth levee constructed in September 2008 by the National Resource Conservation Service (NRCS) to control the concentrated overland flow and the formation of gullies on the upland field was only 5 meters, the streambank retreat was bigger at the second transect, probably due to the scouring effect caused by the stream flow around a 0.75 m diameter pipe, which collected the concentrated flow from upland (Appendix A). 121

Changes in channel width, retreated streambank area and streambed elevation were variable along the 270 m reach. Streambank widening ranged from 0.01 m to 2.67 m, while the depths of sediment deposition ranged from 0.01 m to 0.08 m. Over the studied period, there was a mean increase in channel width of 0.61 m reflected in a mean eroded streambanks area of 2.53 m2. As a reference, widening rates for Southeastern streams have been reported from 0.49 m yr-1 to 8.0 m yr-1 in Mississippi and from 2.1 m yr-1 to 18.2 m yr-1 in Tennessee (Jaeger et al., 2010). The left streambanks side of the reach evidenced an average rate of channel width and streambank retreat of 0.62 m and 5.70 m2, respectively. Channel width and retreated area from the left streambanks were higher on transects along the bendway segment (Table 5.2). The right streambank sides presented an average sediment deposited area of around 0.43 m2 and 0.04 m of mean contraction on channel width. The right streambank at transect 165 m presented the highest area of deposited sediment (3.26 m2) and width contraction (0.08 m). Over the 12.5 month study, approximately 1,930 Mg (1,350 m3) of streambank material were eroded from the streambanks, while 220 Mg (160 m3) of loose material (frequently fine sand) were deposited on the streambanks surface of the studied reach. The 95% of the eroded material (1,845 Mg or 1,300 m3) was supplied by the left side streambanks, principally those along the bendway segment, while 90% of the streambank deposition occurs along the right streambanks of the studied reach. The material eroded from the left streambank of the segments including transects 90 m, 140 m and 165 m represented the 25%, 16% and 52% of the total retreat from the entire reach, respectively. The right streambank of the segment including transect 165 m received 77% of the total material deposited on the streambanks surface. The differences in morphology and in the 122

type of process observed at each side of the streambanks at this segment, and in general along the entire reach length, could be evidencing a lateral migration process along the bendway segment of the studied reach. Degradation of the studied reach at Yonaba Creek and other tributaries at the northern and western area of the Town Creek watershed is directly related to destabilization of the streambanks, leading to channel widening by mass wasting processes, because streambanks heights and angles exceed the shear strength of the streambank material. A possible factor influencing channel degradation along the studied reach and some upstream segments could be the presence of ponds few kilometers upstream. In general, streams located downstream from dams, reservoirs and lakes initially adjust by channel degradation. Typically, a stream will scour, and thus lower its channel streambed as the sediment depleted water emerging from the dam or spillway lake attempts to replenish its sediment load (Juracek, 2001). At the studied reaches, preliminary channel incision could be caused by a reduction in sediment load along the streams induced by the presence of the upstream pond lakes. Considering that incision, which lowers the streambed of the channel, can cause channel to encounter bedrock and resistant alluvium, and this material would greatly reduce or halt incision (Simon and Darby, 1999). Streambed outcrop (gray shale) observed around 100 m downstream from the last surveyed transects (Figure 5.22) and the presence of a similar material between 1 to 2 feet below the streambed height along the studied reach, could be evidencing a geologic control affecting the channel deepening by incision. Once incision has commenced, it is unlikely that erosion will cease naturally until the channel has progressed through several stages described in an incised channel evolution model 123

(Schumm, 1999), as that proposed stages by Simon and Hupp (1992). In combination with the geologic changes, other hydrologic, climatic and human factors (e. g. land use change, elimination of riparian zones for agricultural production) would probably have contributed and changed the rate of channel evolution in the streams at the studied area. Following the channel evolution model, streambed degradation and channel incision (Stage II and Stage III) would reduce channel slope reducing the available stream power for given discharges with time, while streambanks height increased and fluvial undercutting and pore pressure processes would promote streambank steepness and induce streambank failures near the base of the streambank. Channel widening, observed along the studied reach at the Yonaba Creek, is the most common scenario occurring in the subsequent stage (Stage IV) of the channel evolution model. Streambed erosion still occurs along the studied reach, however streambed deposition is observed along the 110 m bendway segment favored possibly by two factors: (1) a reduction in the stream velocity and stream power when streamflow moves along this segment and (2) the significant high contribution of streambank material which exceeds the sediment transport capacity of the streamflow.

124

a

b

Figure 5.22

(a) Streambed outcrop of Selma Shale (gray shale) along the Yonaba Creek (b) Transition between sandy alluvial streambed and streambed with outcrop along the Yonaba Creek downstream transect #8

125

Table 5.2

Comparison of streambank area change in transects along a 270 m reach on the Yonaba Creek near the bridge at the Road 9

Left Streambank 126

Right Streambank

Distance (m) 0 20 90 135 140 165 210 270 0 20 90 135 140 165 210 270

Positive values = Deposition Negative values (-) = Erosion

Feb-09* -

May-09 -6.21 0.46 -7.35 -5.96

-0.55 0.44 0.00 1.61

Change in Cross Sections Area (m2) Jun 1 09 Jun 17 09 Oct-09 Nov-09 1.24 -0.37 -0.84 -2.43 -0.08 0.17 0.06 0.01 -1.05 -0.27 0.51 -2.52 -1.32 -0.04 -0.20 -3.96 -8.18 -0.01 -0.22 0.28 0.22 -0.51 1.04 0.18 0.14 0.03 0.18 0.16 -0.31 -0.52 -0.16 -0.20 0.05 -0.34 0.78 0.01 -0.07 -3.13 3.37 -0.14 -0.08 0.30 -0.09 0.24 0.38 0.06

Mar-10 Net change -1.00 0.24 0.27 -0.10 -1.87 -11.33 0.12 -0.68 -2.50 -13.18 -2.29 -20.63 -0.31 -0.24 -0.40 0.35 0.09 0.27 -0.58 -0.44 0.13 -0.37 0.00 -0.38 0.11 0.56 1.61 3.26 -0.04 0.10 -0.27 0.40

Table 5.3

Comparison of streambank widening and streambed erosion depth in transects along a 270 m reach on the Yonaba Creek between January 29, 2009 and March 18, 2010

Distance (m) 0 20 90 135 140 165 210 270 Mean

Net Streambank Change (mm) Left Right 20.5 24.9 -6.6* -4.3* -1,480.4 -11.0 -41.5 -10.9 -817.3 12.0 -2,672.5 81.4 -23.1 11.42 33.0 27.50 -623.5 42.4

Net Streambed Change (mm) -293.2 95.1 163.1 2.5 82.5 148.4 -146.6 -142.4 -11.3

Positive values = Contraction/Deposition Negative values (-) = Widening/Erosion * Erosion of material previously deposited along the 70 m rip-rapped segment Table 5.4

Distance (m) 0 20 90 135 140 165 210 270

Mean streambank material bulk density and streambank material erosion or deposition along a 270 m reach on the Yonaba Creek between January 31, 2009 and March 18, 2010

Mean Streambank Material Bulk Density (Mg m-3) Left Right 1.38 1.48 1.45* 1.45* 1.44 1.53 1.53 1.59 1.53 1.57 1.38 1.48 1.11 1.11 1.11 1.11 Total

Streambank Erosion per m of stream (Mg m-1 stream) Left 0.33 -0.15 -16.32 -1.04 -20.17 -28.47 -0.27 0.39 -65.68

Right 0.40 -0.64 -0.57 -0.60 0.88 4.82 0.11 0.44 4.85

Segment Streambank Erosion (Mg) Left 1.66 -10.15 -489.46 -31.21 -302.48 -996.43 -13.32 13.60 -1,827.79

* Bulk density of deposited material along the 70 m rip-rapped segment Positive values = Deposition Negative values (-) = Erosion

127

Right 2.00 -44.66 -16.98 -18.13 13.19 168.87 5.55 15.54 125.38

Erosion Pins at the Town Creek Methods Erosion pin arrays were used to obtain an estimate of streambank erosion or deposition depths at two sites along the principal channel of the Town Creek. There were a total of 4 pairs of pin erosion plots (left and right streambank). The first field site (site #1) with three pairs of pin erosion plots was located near the Brewer Road bridge crossing. The site can be characterized as a wide vegetated stable channel (Figure 5.23). Plot #1 presented cypress trees, cypress knees and roots along the low part and the middle part of the streambank. The plot also had shrubs and grassy vegetation that grows in spring and is absent during the winter. Plot #2 presented a band of deposited material from the toe up to a height of 2 m, where some shrubs and grassy vegetation grow in spring. Above this height some sparse trees, shrubs, and grassy vegetation were observed and to grow seasonally. Plot #3 was very similar to the previous plot at the lower part of the streambank, but trees were less dense at the middle height. The plots #4, #5 and #6 were located across the stream from plots #1, #2 and #3, respectively. These streambanks had a change in their morphology by presenting a ~1.7 m vertical streambank between the floodplain and the lower slope. They presented a dense amount of shrubs along the entire height of the streambank, and some trees were located at the top of the lower slope next to the vertical slope. The second site with one pair of pin erosion plots was located 16 Km upstream from site #1 at the edge of the northern headwaters near the bridge at the Natchez Trace with a drainage area of 280 km2 (Figure 5.24). The transition from incised stream channels with unstable active streambank profiles and agricultural lands near streambanks to wide stable vegetated channels is the most representative condition at this 128

site. The left streambank presented some grassy vegetation and shrubs from the toe to approximately ¾ of the entire streambank height and trees from this last height to the floodplain. The right streambank presented an initial slope where deposited and original material from the streambank conformed bars which allowed grass and shrubs to grow. The lower and higher parts of the streambank presented trees and shrubs. Initial conditions of the streambanks are presented in Figure 5.25. The studied streambanks have a continuous median angle of 42o and 48o along the entire streambank height at site #1 for the left and right side, respectively, and 47o at both sides of site #2. Channels at the pin erosion plots locations presented a trapezoidal shape with streambank heights ranging from 6.1 m to 6.7 m on the left side and from 7.1 m to 7.7 m on the right side. Channel bottom width ranged from 15 m to 30 m. Soils analysis found predominately loamy, very fine sand, with textures finer than loam (sandy clay loam, clay loam and clay) only at site #1. Soils at site #1 presented a higher content of organic matter ranging from 1.6% to 6.2%, while soils at site #2 ranged from 1.2 to 4.2%. Analysis of the critical shear stress and erodibility of the soils determined by Jet testing showed that soils from site #2 will became more erodible than soils from site #1 under very similar critical shear stress (between 0.7 and 0.9 Pa) generated by the streamflow.

129

1 130

4

2 5

3 6 N

Figure 5.23

Aerial view of the pin erosion plots location along the Town Creek near the bridge at Brewer Rd. (MS 521) (Source : www.bing.com/maps)

7 131

8

N Figure 5.24

Aerial view of the pin erosion plots location on the Yonaba/Town Creek near the bridge at the Natchez Trace Parkway (Source : www.bing.com/maps)

a

b

c

d

e

f

g

h

Figure 5.25

Initial condition of the different plots located along the Town Creek. a), b) and c) plots 1, 2 and 3 along the left streambank in site #1; d) and e) plots 4, and 5 along the right streambank in site #1; f) plot 7 along the left streambank in site #2; g) and h) plot 8 along the right streambank in site #2

132

Streambank erosion depths at each site were monitored using 760 mm (30”) long and 6 mm (1/4”) diameter steel-rod erosion pins inserted perpendicularly into the streambank face. Pins were inserted into the streambank without disturbing the surrounding soil and were left exposed by 125 mm (≈5”). To facilitate relocation, the top of each pin was painted and a flag was inserted next to each pin. Each erosion plot array had at least 24 columns (1 m apart) along the streamflow direction and four rows along the streambank height below the top of the streambank at both channel sides. Streambank side designation (left or right) was determined by looking upstream. The rows’ heights for the first location were 0.3 m, 1.15 m, 3.3 m, 4.45 m, while heights for the second location were 0.2 m, 1 m, 3 m and 4.5 m. One streambank at each side of the first site (plots #1 and #4) had an additional row at 5.45 m due to minimum changes in the morphological characteristics of these streambanks. During the monitoring period from late July 2008 to the end of June 2009, the exposed length of each pin was measured when possible after each storm flow event if peak flow height exceeded the first pin erosion line. Pins covered by deposited sediment were measured while trying to cause minimum disturbance on the streambank soil surface. When erosion pins were below the water line or they were not found after any erosion or deposition event, no exposition was assigned because of uncertainty regarding whether they had been completely eroded and buried, or were present but not found under the deposited sediment.

133

Results A total of 13 monitoring dates were generated by increase in streamflow during the period of study (Table 5.5).

Table 5.5

Dates of monitoring of pin erosion depths for two sites along the Town Creek Monitoring Event 1 2 3 4 5

6 7 8 9 10 11 12 13

Stormflow dates

Monitoring date

Oct 8, 2008 Oct 24, 2008 Nov 13, 2008 Dec 4, 2008 Dec 10, 2008 Dec 11, 2008 Dec 17, 2008 Dec 21, 2008 Dec 25, 2008 Dec 28, 2008 Jan 6, 2009 Feb 28, 2009 Mar 14, 2009 Mar 26, 2009 Apr 3, 2009 Apr 13, 2009 Apr 20, 2009 May 3, 2009 May 6, 2009 May 10, 2009 May 16, 2009 May 24, 2009 Jun 13, 2009

Oct 13, 2008 Oct 28, 2008 Nov 21, 2008 Dec 8, 2008 Jan 26, 2009

Peak flow depth at site #1 (m) 3.88 2.48 0.51 2.06 6.161

Mar 7, 2009 Mar 23, 2009 Apr 8, 2009

4.93 5.11 6.05

Apr 16, 2009 Apr 25, 2009 May 14, 2009

2.43 2.10 5.56

May 28, 2009

2.61

Jun 18, 2009

2.01

1

Peak height at December 10, 2008; 2 Peak height at Mar 26, 2009; 3 Peak height at May 6, 2009 Data are summarized as the depth of streambank erosion (+) or deposition (-) of each line for each plot for 13 monitored events, which occurred in approximately 1 year 134

(Table 5.6). The mean depth for all the pins in the studied period was 22.4 mm of sediment deposition. However, the upstream site (site #2) presented a mean depth of streambank erosion of 2.8 mm, while site #1 had a mean depth of sediment deposition of 30.1 mm. As observed in Figures 5.26 to 5.29, the major part of the monitored plots were affected by sediment deposition along the two higher lines of erosion pins and by erosion along the lower lines after the conclusion of a sequence of stormflow events that occurred between December 10, 2008, and January 6, 2009 (Event #5). Sediment deposition on the streambanks’ surfaces was as high as 300 mm, a condition that significantly modified the surface characteristics of the right streambanks at site #1 and the left streambank at site #2 by covering an important part of the initial dense vegetation and modifying the textural composition and physical characteristics of the streambank surface material (Figure 5.30). The sediment deposition along the higher line on site #1 was also favored by the failure of the vertical slope at the top of the streambank (Figure 5.31). Sediment deposited on the middle part of the streambank surface is mostly characterized as very fine sand with few amounts of clay (