W&M ScholarWorks
Dissertations, Theses, and Masters Projects Theses, Dissertations, & Master Projects
2007
Shelf sediment dispersal mechanisms and deposition on the Waiapu River shelf, New Zealand
Tara A. Kniskern College of William and Mary - Virginia Institute of Marine Science
Follow this and additional works at: https://scholarworks.wm.edu/etd
Part of the Geology Commons, and the Oceanography Commons
Recommended Citation Kniskern, Tara A., "Shelf sediment dispersal mechanisms and deposition on the Waiapu River shelf, New Zealand" (2007). Dissertations, Theses, and Masters Projects. Paper 1539616720. https://dx.doi.org/doi:10.25773/v5-05pd-dq04
This Dissertation is brought to you for free and open access by the Theses, Dissertations, & Master Projects at W&M ScholarWorks. It has been accepted for inclusion in Dissertations, Theses, and Masters Projects by an authorized administrator of W&M ScholarWorks. For more information, please contact [email protected]. SHELF SEDIMENT DISPERSAL MECHANISMS AND DEPOSITION
ON THE WAIAPU RIVER SHELF, NEW ZEALAND.
A Dissertation
Presented to
The Faculty of the School of Marine Science
The College of William and Mary in Virginia
In Partial Fulfullment
Of the Requirements for the Degree of
Doctor of Philosophy
by
Tara A. Kniskem 2007
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPROVAL SHEET
This dissertation is submitted in partial fulfillment of
The requirements for the degree of
Doctor of Philosophy
Tara A. Kniskem
Approved May, 2007
Courtney K. Harris, Ph.D Committee Chairman/ Advisor
ittee Chairman/ Advisor
L i cv\i_£-_ Lionel Carter, Ph.D Victoria University of Wellington Wellington, New Zealand
'v jL Jesse MeNinch, Ph.D
A a Linda C.:. Scfiaffner, (Pn.D
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. DEDICATION
To my husband, Mark Kram.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TABLE OF CONTENTS
ACKNOWLEDGEMENTS...... x
LISTOF TABLES...... xii
LIST OF FIGURES...... xiii
ABSTRACT...... xvi
INTRODUCTION...... 2
CHAPTER 1:
Sediment accumulation and preservation potential on the Waiapu River shelf, East
Cape, New Zealand ...... 9
ABSTRACT...... 9
INTRODUCTION...... 11
BACKGROUND...... 14
OBJECTIVES...... 16
METHODS...... 17
RESULTS...... 19
A. Sedimentary structure and texture ...... 19
B. Radiochemical data ...... 21
C. Stable Isotope data ...... 23
DISCUSSION...... 24
A. Calculating 210Pb accumulation rates ...... 24
B. Sediment budget ...... 30
C. Event preservation ...... 32
D. Identifying low excess 210Pb activity layers ...... 36
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. E. Long- versus short-term accumulation ...... 39
CONCLUSIONS...... 42
REFERENCES...... 44
TABLE...... 55
FIGURE CAPTIONS...... 56
FIGURES...... 60
CHAPTER 2:
Flood dispersal and deposition on the Waiapu River shelf, New Zealand ...... 76
ABSTRACT...... 76
INTRODUCTION...... 78
OBJECTIVES...... 84
BACKGROUND...... 84
A. The river and shelf environment ...... 84
B. The deployment period ...... 87
METHODS...... 89
APPROACH...... 91
RESULTS...... 94
A. Simulation estimates compared to observations ...... 94
1. Bed shear stresses ...... 94
2. Sediment concentrations ...... 95
3. Currents ...... 95
4. Summary ...... 97
B. Sediment transport and deposition ...... 98
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1. Shelf sediment budget ...... 98
2. Sediment export ...... 100
3. Sediment depositional patterns ...... 101
C. Floods with large wave energy compared to floods with low
wave energy...... 103
DISCUSSION...... 103
A. Sediment transport mechanisms ...... 104
1. Time-averaged flux patterns ...... 104
2. Flux patterns during high discharge ...... 106
B. Flux convergence and divergence across the shelf...... 108
C. Temporal and spatial variability in sediment transport and
depositional patterns ...... I l l
D. Sensitivity to settling velocity...... 112
E. Dispersal processes not represented by the model ...... 113
1. Currents ...... 113
2. River-initiated hyperpycnal sediment transport ...... 116
CONCLUSIONS...... 118
REFERENCES...... 120
TABLES...... 130
FIGURE CAPTIONS...... 140
FIGURES...... 146
CHAPTER 3:
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Sediment transport pathway sensitivity and strata-formation on the Waiapu River Shelf,
New Zealand ...... 163
ABSTRACT...... 163
INTRODUCTION...... 165
OBJECTIVES...... 168
BACKGROUND...... 169
A. Regional setting...... 169
B. Shelf sediment transport mechanisms and depositional patterns 171
METHODS...... 175
APPROACH...... 176
A. Baseline case for comparison with sensitivity experiments ...... 177
B. Sensitivity experimental design ...... 178
1. Wave energy...... 178
2. Wind-driven currents ...... 178
3. Sediment settling velocity ...... 178
4. Fluvial sediment concentration and load ...... 179
5. Historical terrestrial sediment concentration ...... 179
6. Experimental comparison ...... 179
RESULTS and DISCUSSION ...... 180
A. Sediment mass distribution for each experiment ...... 180
1. Baseline case ...... 181
2. Wave sensitivity: wave height, steady waves, and wave timing
experiments ...... 181
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. viii
3. Wind sensitivity: wind speed and direction ...... 183
4. Sensitivity to flocculation fraction ...... 185
5. Sediment load sensitivity ...... 186
6. Sediment concentration sensitivity ...... 186
B. General sediment mass distribution for all experiments ...... 187
C. Time-averaged along- and across-shelf flux ...... 189
1. General patterns of along- and across-shelf flux ...... 190
2. Along- and across-shelf flux sensitivities ...... 192
3. Sediment export...... 197 D. Flux convergence and divergence ...... 199 E. Sediment depositional patterns across the shelf...... 203 F. Observed and simulated deposition and preservation ...... 205 1. Short- and long-term depositional patterns ...... 205 2. Long-term shelf preservation patterns ...... 208 3. Transport mechanisms ...... 210 G. Historical changes in sediment concentration and load ...... 212 CONCLUSIONS...... 215 REFERENCES...... 217
TABLES...... 226
FIGURE CAPTIONS...... 235
FIGURES...... 241
CONCLUSIONS...... 272
FUTURE WORK...... 273
REFERENCES...... 275
APPENDIX A ...... 281
APPENDIX B...... 282
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPENDIX C
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ACKNOWLEDGEMENTS
They say it takes a village to raise a child. So too does it require the efforts of a
scientific community to create a scientist. First, I would like to thank Drs. Duncan
FitzGerald and Ilya Buynevich who showed me how to successfully navigate my first
undergraduate research project and were always supportive. Secondly, I would like to
thank Dr. Steve Kuehl, who taught me how to think like a scientist, how to write like a
scientist, and responded with encouragement when my research interests diverged into
the realm of numerical modeling. Thirdly, I would like to thank Dr. Courtney Harris,
who agreed to work with me even though I lacked experience in numerical modeling.
From Courtney I learned how to recognize the value of my work in the context of the
larger picture and how to network with other scientists. Both Steve and Courtney have
acted as excellent role models, professionally and personally. I would also like to thank
my committee: Drs. Jesse McNinch and Linda Schaffner at VIMS, Dr. Lionel Carter
from Victoria University at Wellington, New Zealand, and Dr. Carl Friedrichs from
VIMS, who is an honorary committee member. Each contributed their invaluable
insight to the Waiapu River shelf project and to my career. I owe a special thanks to
Dr. Linda Schaffner, who agreed to participate in both my masters and doctoral degree
committees. I admire her as a scientist and as a person.
Many people worked on the Waiapu River shelf project to whom I owe thanks
for hard labor, data, and great B.S. sessions: Lisa Addington, Hannah Brackley, Grace
Browder, Lionel Carter, Carl Friedrichs, Bob Gammisch, Lila Gerald, Brian Hasty,
Karla Hunt, Steve Kuehl, Peony Ma, Jesse McNinch, Linda Meneghini, Jennifer
Miselis, Keith Moodie, Lisa Northcote, Wayne Reisner, Si Schiavone, Malcolm Scully,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Liz Tennant, Heidi Wadman, Don Wright, and the crews of the RV Tangaroa (NIWA)
and the RV Kilo Moana (Univ. of Hawaii). Additionally, I owe many thanks to our
colleagues in New Zealand: Hannah Brackley, Lionel Carter, Murray Hicks, Keith
Lewis, Lisa and Peter Northcote, Alan Orpin, Alan Palmer, Mike Page, David Peacock,
and Noel Trustrum for data, for help, and most of all for friendship. I am especially
grateful to the Ngati Porou, who generously shared their rich culture with a bunch of
Yankees scientists.
The impact of all those seemingly small day-to-day interactions within the
VIMS community looms large as I contemplate the ending of my graduate program.
Linda Meneghini, a.k.a. “the Mud Lady”, kept the lab going through thick and thin, kept
us all well fed, and is an excellent source of information and advice on just about any
topic. I wish to acknowledge and thank Michael Cooke, Cynthia Harris, Cindy
Hornsby, Paulette Lewis, Janette Millen, Beth Marshall, and Phyllis Spriggs for all their
help and for great conversations. I also thank the VIMS faculty for all the good and the
not-so-good that I have observed and learned from over the years. I would like to thank
my friends and fellow students, especially Lisa Addington, Aaron Bever, Lorraine
Brasseur, Grace Cartwright, Pat Dickhudt, Lila Gerald, Lindsey Kraatz, Lynsey Lemay,
Peony Ma, Annie Miller, Jen Miselis, J. Paul Rinehimer, Cielomar Rodriguez-Calderon,
and Heidi Wadman. I especially thank my closest friends Katie Farnsworth, Mindy
Gensler, Liz Mountz, and Liz Wall for more than can be written.
Finally, I thank my extended family for making me who I am, Mark for giving
me the world, and Nathan for giving me joy.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF TABLES
CHAPTER 1
Table 1.1 Kasten core statistics ...... 55
CHAPTER 2
Table 2.1. Tripod instruments ...... 130
Table 2.2. Bed shear stresses ...... 131
Table 2.3. Depth-averaged currents ...... 132
Table 2.4. Wave boundary layer currents ...... 133
Table 2.5. Currents 1 m above the bed ...... 134
Table 2.6. Time-averaged sediment mass distribution ...... 135
Table 2.7. Sediment flux budgets ...... 136
Table 2.8. Flood statistics ...... 137
Table 2.9. Sediment deposition across the shelf ...... 138
Table 2.10. Relative importance of the sediment transport mechanisms 139
CHAPTER 3
Table 3.1. Sensitivity experimental design ...... 221
Table 3.2. Shelf sediment budget ...... 223
Table 3.3. Bed shear stress statistics ...... 225
Table 3.4. Mean along-shelf fluxes ...... 226
Table 3.5. Mean across-shelf fluxes ...... 228
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF FIGURES
CHAPTER 1
Figure 1.1. Study location map ...... 60
Figure 1.2. Waiapu River discharge ...... 61
Figure 1.3. Average surface grain-size...... 62
Figure 1.4. Transect D x-radiographs ...... 63
Figure 1.5. Excess 210Pb activity profiles...... 64
Figure 1.6. Excess 010 Pb accumulation rate map ...... 65
Figure 1.7. Pulsed event layer thickness map...... 66
Figure 1.8. Sedimentary facies map ...... 67
n Figure 1.9. Be activity map...... 68
Figure 1.10. Kasten Core 18 excess 210Pb, bulk carbon, and 8 I3C profiles 69
Figure 1.11. Carbon sources ...... 70
Figure 1.12A. Kasten Core 18 excess 210Pb and grain-size correlations ...... 71
Figure 1.12B. Kasten Core 18 excess
210Pb profile normalized by grain-size ...... 72
o i n Figure 1.13. Kasten core 18 excess Pb and carbon correlations ...... 73
Figure 1.14. Waiapu River Shelf sediment budget ...... 74
Figure 1.15. Kasten Core 18 fraction less than 20pm ...... 75
CHAPTER 2
Figure 2.1. Study location map ...... 143
Figure 2.2. model inputs and Waiapu rating curve ...... 144
Figure 2.3. model grid ...... 145
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. xiv
Figure 2.4. wind rose diagrams ...... 146
Figure 2.5. depth averaged currents ...... 147
Figure 2.6. model and tripod comparison ...... 148
Figure 2.7. depth averaged and near bed currents ...... 149
Figure 2.8 sediment mass distribution ...... 150
Figure 2.9. planview of buoyant plume ...... 151
Figure 2.10. cumulative sediment export and along shelf flux ...... 152
Figure 2.11. wave timing with peak discharge planview...... 153
Figure 2.12. time-averaged flux patterns and vertical sediment profiles ...... 154
Figure 2.13. correlation of flux with wave and current velocities ...... 155
Figure 2.14. flux patterns and vertical sediment profiles for floods ...... 156
Figure 2.15. sediment distribution with depth across the shelf ...... 157
Figure 2.16. time-averaged flux convergence and divergence patterns ...... 158
Figure 2.17. sediment distribution and export for various floe fractions ...... 159
CHAPTER 3
Figure 3.1. Study location map...... 237
Figure 3.2. Waiapu River discharge record ...... 238
Figure 3.3. model grid and forcings ...... 239
Figure 3.4. model versus measured outputs ...... 240
Figure 3.5. baseline case plume and sediment distribution ...... 241
Figure 3.6. sensitivity experiment sediment distribution ...... 242
Figure 3.7. planview of steady wave and wave height deposition ...... 244
Figure 3.8. planview and sediment distribution
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. XV
of wave timing experiment ...... 245
Figure 3.9. export versus mean bed shear stress ...... 246
Figure 3.10. planview of wind experiment deposition ...... 247
Figure 3.11. planview of floe fraction experiment
and sediment distribution ...... 248
Figure 3.12. planview of load experiment and sediment distribution ...... 249
Figure 3.13. planview of concentration experiment
and sediment distribution ...... 250
Figure 3.14. baseline case flux patterns ...... 251
Figure 3.15. sediment load and concentration experiment flux patterns ...... 252
Figure 3.16. wave experiment flux patterns ...... 253
Figure 3.17. wind experiment flux patterns ...... 255
Figure 3.18. mean bed shear stress versus mean along shelf flux ...... 256
Figure 3.19. along shelf flux time-series for wind experiments ...... 257
Figure 3.20. export for floe fraction experiment and along shelf flux for
floe fraction, load, and concentration experiments ...... 258
Figure 3.21. load and concentration experiment export ...... 259
Figure 3.22. wave and wind experiment export ...... 260
Figure 3.23. flux convergence patterns ...... 261
Figure 3.24. flux convergence patterns ...... 262
Figure 3.25. flux convergence patterns ...... 263
Figure 3.26. sediment distribution across the shelf for the baseline case ...... 264
Figure 3.27. Sediment depocenters on the shelf...... 265
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ABSTRACT
The Waiapu River, located on the North Island of New Zealand, is characterized
by a small catchment composed of fissile and easily eroded material. Deforestation
over the last 150 years combined with high rainfall rates has resulted in one of the
highest sediment yields in the world. The river delivers most of its annual sediment
load during short flood pulses into coastal waters subject to energetic waves and
currents. These conditions are favorable for producing multiple sediment transport
mechanisms, including transport in dilute suspension, wave- and current-supported
gravity flows, hyperpycnal plumes, and resuspension.
Analyses of seabed data collected in 2003 and 2004 showed that multiple
sediment transport mechanisms influence strata formation on the shelf. The presence of
short-lived 7Be exclusively in surface sediments sampled from locations shallower than
80 m indicated that sediments were ephemerally deposited on the inner shelf before
being resuspended and transported further offshore. High, long-term excess 210Pb
accumulation rates, ranging between 0.2 -3.3 cm/yr, showed that sediments were
retained mainly on the mid- to outer-shelf region at water depths between 60 and 190
m. These high shelf accumulation rates are sufficient to bury and preserve physically
dominated structures at depth in the sediment column. Pulsed event layers, evidenced
by layers of low excess 210Pb and terrestrial 8 13C, were most frequently seen in areas of
high sediment accumulation. Analyses of these data were used to estimate that 24% of
fluvial sediments were retained on the shelf over the last 100 years.
To address the mechanisms by which sediment escaped the shelf and to assess
the relative importance of the various transport mechanisms, a three-dimensional
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. numerical model, ECOMSED, was used to simulate fine sediment transport and
deposition during a field experiment in 2004. The simulation was able to reproduce
time-averaged currents, near-bed sediment concentrations, and bed shear stresses at a
tripod deployed off the river mouth at 60 m depth. However, the model did not
represent large velocity excursions seen in the tripod data, because it neglected large-
scale current systems offshore that may occasionally interact with shallower shelf
waters.
Simulation and sensitivity experiments showed that the relative influence of the
dilute suspension, resuspension by waves and currents, and gravitationally-driven flow
varied across- and along-shelf. Sediment settling velocities largely determined
sediment transport pathway and fate. Gravity-driven transport was most important on
the inner and mid-shelf, whereas dilute transport became more important beyond 60 m
depth. Dilute suspension exported sediments to the north of the shelf area. Two
depositional patterns were observed depending on the timing of energetic waves and
floods. Sediments bypassed the inner shelf and were deposited in deeper waters when
waves were energetic during the peak fluvial discharge. When wave energy was low
during a flood, sediments were deposited on the inner shelf before being resuspended
and transported further offshore. The relative importance of these multiple sediment
transport mechanisms varied according to oceanographic/meteorologic conditions and
influenced fine-scale strata formation and sediment retention on the shelf.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. SHELF SEDIMENT DISPERSAL MECHANISMS AND DEPOSITION
ON THE WAIAPU RIVER SHELF, NEW ZEALAND.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. INTRODUCTION
Continental shelf systems serve as a link between land and the deep ocean, and
are influenced by fluvial delivery of terrestrial materials. Once these terrestrially-
derived materials are delivered to coastal waters, they undergo modification by
chemical, physical, and biological processes (Guinasso and Schink, 1975; Nittrouer and
Sternberg, 1981; Hedges and Kiel, 1995; Wheatcroft, 2006). The relative importance
of these processes is influenced by shelf morphology, the timing and amount of fluvial
delivery, and the offshore wave and current regime (e.g., Mulder and Syvitski, 1995;
Kineke et al., 1996; Wright et al., 2001, Friedrichs and Wright, 2004). Seabed
deposition and burial on the shelf are observable products of these physical and
biological processes, but provide only an imperfect record because they tend to
obliterate prior signals. Process-oriented studies over a variety of time-scales are
therefore fundamental to understanding the link between sediment input from rivers,
reworking processes, and the resulting sedimentary strata.
Both sediment transport processes and biological processes influence sediment
depositional patterns, strata formation, and reworking. Multiple sediment transport
mechanisms distribute terrestrially-derived sediments and nutrients variably across and
along shelf systems. These include:
1. buoyant, or hypopycnal, fluvial plumes (Geyer et al., 2000; Geyer et al.,
2004)
2. negatively buoyant, or hyperpycnal, fluvial plumes (Mulder and Syvitski,
1995)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3. wave-and current-supported gravity flows (Kineke and Sternberg, 1995;
Cacchione et al., 1995; Traykovski et al., 2000; Wright et al., 2002;
Friedrichs and Wright 2004)
4. wave- and current-resuspension (Sternberg and Larsen, 1976; Grant and
Madsen, 1986; Ogston and Sternberg, 1999; Palanques et al., 2002)
All of these transport mechanisms have been observed on shelf systems, and their
relative contributions can vary with fluvial and oceanographic conditions (Cacchione et
al., 1995; Kineke et al., 2000; Traykovski et al., 2000; Friedrichs and Scully, 2007).
Generally, river-derived sediments are delivered to coastal waters in buoyant plumes,
but these may become hyperpycnal when sediment concentrations exceed a threshold
gravity-driven transport, e.g. -40 g/1 (Mulder and Syvitski, 1995). Sediment
resuspension into the water column by energetic hydrodynamic conditions has been
believed to dominate transport on shelves (Sternberg and Larsen, 1976; Grant and
Madsen, 1986). Wave resuspension tend to dominate (Drake and Cacchione, 1985).
Recent studies have shown, however, that gravity flows, where the weight of suspended
sediment is strong enough to generate offshore flux, may be important on some river-
shelf systems (Traykovski et al., 2000; Harris et al., 2005; Friedrichs and Wright,
2004). Energetic waves, currents, and steep slopes can sustain near-bed turbid mud
flows that are driven across-shelf by gravitational processes (Ogston et al., 2000;
Traykovski et al., 2000; Kineke et al., 1996; Wright et al., 2001, 2002; Scully et al.,
2002, 2003, Friedrichs and Wright, 2004). Gravity flows sustained by wave energy
may be depth-limited by wave attenuation, whereas the current- and slope-sustained
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. gravity flows are not depth-limited (Friedrichs and Wright, 2004; Wright and
Friedrichs, 2006; Ma et al., in prep).
Additionally, shelf waters serve to promote rich and diverse aquatic and benthic
communities that influence sediment bioturbation patterns. Biological processes that
can affect shelf sediments range from phytoplankton and bacteria in the water column,
responsible for altering terrestrially-derived nutrients, to benthic communities capable
of altering both nutrient signals and sedimentary strata. Activities such as burrowing,
boring, and feeding mix the substrate and delay burial of sediment particles (Sanford,
1992). The spatial and temporal variability of benthic community composition
influences bioturbation rates and the depth to which sediments are disturbed (Officer
and Lynch, 1989; Sanford, 1992; Dellapenna et al., 1998; Kirchner et al., 1998,
Schaffner et al., 2001). Most animals live near the sediment-water interface, disturbing
the top 5 cm of the sediment column, but some animals burrow to depths greater than 30
cm (Hines and Comtois, 1985; Schaffner et al., 1987; Wheatcroft et al., 1994). Both
physical and biological processes influence nutrient and contaminant cycling and the
fate of terrestrially-derived materials (Wheatcroft and Drake, 2003).
Although accumulation rates on continental shelves adjacent to rivers generally
range between 0.1-1 cm/y, short dissipation times and long transit times of sediments
through the surface mixed layer tend to preclude preservation of most pulsed event
layers (Wheatcroft and Drake, 2003). For example, accumulation on the Eel River
shelf is dominated by episodic floods that can result in 10 cm-thick flood layers, yet
bioturbation rates are sufficient to obliterate the signals left by most flood deposits
(Wheatcroft and Drake, 2003; Sommerfield and Nittrouer, 1999). Strong tidal currents
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5
on the Amazon River margin, however, produce 1 m-thick fluid muds, exerting strong
physical controls on sedimentary structures (Cacchione et al., 1995; Kineke and
Sternberg, 1995; Dukat and Kuehl, 1995). An understanding of these physical and
biological processes and their resultant depositional products is therefore requisite to
deciphering the stratigraphic record.
This study investigated both sediment transport mechanisms and their resultant
depositional signatures in the Waiapu River Shelf, in the East Cape region of New
Zealand. This area was chosen because the river experiences discharge pulses each
year, during which sediment concentrations exceed the accepted threshold for
hyperpycnal plumes of 40 g/L (Mulder and Syvitski, 1995). Energetic waves and
currents measured during a benthic tripod deployment (Wright et al., 2006), indicated
that several sediment transport mechanisms, including gravity flows, buoyant plumes,
and resuspension, were important for redistribution of sediments on the shelf.
Potentially high accumulation rates due to deforestation over the last 150 years
combined with a shelf basin that contains ~ 100 m of Holocene sediment deposition
indicated that Waiapu River Shelf sediments might preserve records of these transport
processes. The Waiapu River Shelf therefore provided an ideal setting to investigate
sediment transport mechanisms and their resultant depositional products.
The main objectives of this study included:
1. assessing the relative importance of various sediment transport mechanisms
on the shelf,
2. evaluating the sediment budget
3. finding the transport pathways of sediment escape from the shelf
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6
4. investigating sedimentary strata formation
5. assessing the potential impacts of increased terrestrial delivery over the last
100 years.
To accomplish these goals, both numerical and observational approaches were used.
The observational data consisted of a suite of 3 m long gravity cores collected in
August 2003 and 1 m long box cores collected in August 2003 and May 2004. Water
column observations of current velocities, wave properties, and suspended sediment
concentrations were collected during a field experiment in 2004 (Wright et al., 2006;
Ma et al., in prep). These were used as input and validation for a three-dimensional
numerical model (see Harris et al., 2004) that simulated sediment transport and
deposition. This three-dimensional sediment transport model, which accounted for
both suspended and gravity-driven transport, was also used to assess the sensitivity of
transport and deposition to environmental and model parameters.
The main findings of this study are:
1. The shelf has retained 24% of fine sediments delivered by the Waiapu River
over the last 100 years between 30 - 200 m water depths.
2. Sediment deposition and accumulation rates on the Waiapu River continental
shelf exceed the ability of the extant benthic community to effectively
bioturbate the sediments.
3. Sediments are transported across the shelf and buried rapidly enough that
sediments retain a terrestrial carbon signal.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7
4. The relative importance of dilute suspension and wave boundary layer
gravitational flows varies across the shelf, which results in distinct transport and
depositional characteristics for each shelf region.
5. Substantial export of sediments from the proximal shelf can be accounted for by
dispersal of dilute suspension.
6. Sediments transported within gravitationally-driven, wave boundary layer flows
tend to be retained on the shelf and may form preservable deposits.
7. Sediment transport mechanisms and shelf deposition are variably sensitive to
currents, waves, floe fraction, sediment load, and suspended sediment
concentration.
The Waiapu River is a physically dominated system, and provided a unique
opportunity to investigate physical processes. Accumulation rates on the Waiapu River
shelf were up to an order of magnitude higher than the world-wide average
accumulation rate range for continental shelves. As a result, bioturbation rates were
low, and multiple transport and reworking signals were preserved in the shelf (e.g.
Nittrouer and Sternberg, 1981). The approach used in this study allowed an evaluation
of sediment transport processes and their depositional products. The model was able to
successfully estimate where gravity-driven flows, dilute suspensions, and resuspension
would be important on the shelf and the conditions which would result in deposition.
The results of this study are presented in three chapters. Chapter 1 addresses
short- and long-term accumulation patterns, sedimentary structure, and sediment
retention over the last 100 years. Sediment transport and depositional patterns were
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8
simulated and analyzed for a two-month period and compared with oceanographic data
collected during a field experiment in 2004, which are presented in Chapter 2. Chapter
3 uses the numerical model to investigate sediment transport and depositional
variability in response to changes in wave energy, wind-driven currents, sediment
settling velocities, sediment load, and fluvial suspended sediment concentration.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9
CHAPTER 1
Sediment accumulation and preservation potential on the Waiapu River shelf, East
Cape, New Zealand
I. Abstract
The Waiapu River has a small, mountainous catchment that delivers
approximately 35 MT/y of sediment, and disperses a significant fraction of its annual
fluvial sediment load, onto the adjacent shelf during floods. Estimated fluvial sediment
concentrations may exceed the threshold for gravity-driven plume formation several
times per year, indicating that both positively and negatively buoyant plumes disperse
flood sediments on the shelf. Measured energetic waves and currents indicate that
gravity flows and resuspension may be important processes on the shelf as well. These
various sediment transport and physical reworking mechanisms dominantly influence
sediment structure on the Waiapu River shelf.
Sedimentary structure and geochemical signatures from a suite of gravity cores
collected from the Waiapu River shelf revealed that multiple sediment transport and
reworking processes influenced fine-scale strata formation and preservation. The
majority of cores collected from the shelf exhibited non-steady state excess 210Pb
profiles, characterized by layers of low excess activity. These layers correlated with
relatively terrestrial 8 13C and C/N values, indicating that the fluvial sediments
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. deposited rapidly. The sediments inter-bedded with these pulsed event layers were
either not deposited rapidly or were a product of bioturbation. The pattern of pulsed
event layers preserved in the sediment record extended along the shelf in either
direction of the river mouth, primarily between 50 -100 m depth. This isobathymetric
pattern suggests that, no matter the fluvial dispersal mechanism, resuspension by
energetic waves and currents influenced shelf strata formation.
Sedimentary structures associated with pulsed fluvial sediment delivery, high
accumulation rates, and reworking by energetic waves and currents were preserved in
the sediment record because bioturbation rates on the Waiapu River shelf were low.
Accumulation rates on the Waiapu River shelf were extremely high and averaged 1.5
+0.9 cm/y, ranging between 0.2 - 3.5 cm/y. Pulsed event beds preserved in the
sediment column ranged between 10 -20 cm thick. A gradient of physical and
biological signatures extended radially from the Waiapu River mouth, suggesting that
the high fluvial delivery negatively influenced benthic communities.
The Waiapu River retained about 24% of the fluvial load on the shelf between
30 - 200 m over the last 100 years. Excess 210Pb and 7Be activities indicated that
muddy sediments were only temporarily deposited at the boundary of the inner and mid
shelf (30 -50 m), whereas portions of the mid- and outer shelf (50 - 130 m) acted as
fine sediment repositories at least over the last 100 years. Cores collected from the
shelf break and slope (130 - 700 m), indicated that some Waiapu River sediment was
delivered to deeper waters. Those sediments not retained on the shelf were likely
transported to the slope and along the shelf beyond the sampling area.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. II. Introduction
Collision margin rivers account for a significant portion (> 50%) of the world’s
terrigenous input to the coastal ocean (Milliman et al., 1999; Farnsworth and Milliman,
2003). These rivers tend to be small, with mountainous catchments composed of fissile
and highly erodible materials that produce some of the highest sediment yields in the
world (Griffiths and Glasby, 1985; Milliman et al., 1999). Sediment delivery tends to
be episodically driven, with a significant portion of a river’s load delivered during
floods (Farnsworth and Milliman, 2003, Milliman and Syvitski, 1992, Hicks et al.,
2004).
Recent studies show that multiple sediment transport mechanisms influence
sediment dispersal, accumulation, and fine-scale strata formation on collision margin
shelves, potentially resulting in terrigenous material delivery to the nearby adjacent
slope (Mulder and Syvitski, 1995; Morehead and Syvitski, 1999; Kineke et al., 2000,
Friedrichs and Wright, 2004, Mullenbach et al., 2004). An understanding of the
sediment dispersal mechanisms and the burial history of sediments on collision margins
may enhance our understanding of the exchange and burial of important chemical
constituents, and the interplay between biological and physical processes in forming
fine-scale strata.
Although fluvial sediments are typically dispersed in buoyant plumes, these
small collision margin rivers can produce suspended sediment concentrations that
exceed the threshold for negatively buoyant, or hyperpycnal plumes (Mulder and
Syvitski, 1995). Hyperpycnal plumes initiate at the mouth of the river when fluvial
suspended sediment concentrations reach between 20 - 45 g/1 depending on the salinity
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 12
and temperature of the receiving coastal waters (Wright et al., 1990; Mulder and
Syvitski, 1995). Gravity-driven sediment flows not initiated by high fluvial suspended
sediment concentrations can form when sediment convergence creates highly turbid
layers or when the slope of the shelf exceeds 0.01 (Friedrichs and Wright, 2004). Since
continental shelf slopes rarely exceed the critical slope of 0.01, energetic waves and
currents capable of suspending turbid layers necessarily support across-shelf gravity-
driven flows on most shelves (Ogston et al., 2000, Traykovski et al., 2000, Wright et
al., 2001, Wright et al., 2002, Friedrichs and Wright, 2004; Wheatcroft and
Sommerfield, 2005).
Gravity-driven sediment transport within a near-bed fluid mud layer has been
observed on several river-shelf systems. Most large rivers typically distribute sediment
onto wide continental shelves via positively buoyant hypopycnal plumes (e.g., Nittrouer
and Wright, 1994; Nittrouer et al., 1995; Wright and Nittrouer, 1995) and do not
generate hyperpycnal plumes at their mouths in part because so much sediment is
stored within coastal flood plains (Kineke et al., 2000), resulting in low suspended
matter concentration delivery to the ocean. When the characteristically high sediment
delivery rate of these larger rivers is accompanied by rapid settling, estuarine-like
circulation (Kineke and Sternberg, 1995), or some other mechanism that concentrates
sediments on the adjacent shelf, gravity-driven transport can be initiated (Mulder and
Syviski, 1995). Near-bed fluid mud layers have been observed on shelves adjacent to
some of the world’s largest rivers, such as the Amazon (Kineke and Sternberg, 1995),
Yellow (Wright et al., 1988 and 1990), and Zaire (Eisma and Kalf, 1984) Rivers.
Gravity-driven hyperpycnal river plumes as well as wave-initiated, near-bed fluid mud
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. layers have been observed on shelves off of small rivers such as the Eel (Ogston et al.,
2000; Traykovski, et al., 2000), Chosui (Milliman and Kao, 2005), and Tsengwen
Rivers (Mulder and Syvitski, 1995; Liu et al., 1998). Because small, mountainous
rivers do not effectively trap sediment within their watersheds and their catchments are
highly erodible, they deliver high concentrations of fluvial suspended sediment that
potentially produce gravity-driven plumes. These various combinations of sediment
transport pathways, in addition to the particular shelf geometry, accumulation patterns,
and oceanographic/meteorologic conditions primarily control the formation of
stratigraphic sequences on continental shelves (Jervey, 1988; Morehead and Syvitski,
1999; Walsh and Nittrouer, 2003).
Proximal submarine canyons and bar-restricted river mouths, in conjunction
with episodic delivery of high fluvial sediment concentrations, can also significantly
affect sediment dispersal pathways. Although sediment sequences on the Eel River
shelf are affected by tectonically-imposed geometries, the nearby submarine canyon
also plays a role in diverting sediments from the Eel River shelf (Orange, 1999;
Mullenbach et al., 2004). Nearly 88% of the Eel sediment budget is lost either over the
shelf-break via gravity flows or down the nearby submarine canyon (Nittrouer et al.,
1979; Sommerfield and Nittrouer, 1999; Walsh and Nittrouer, 2003; Mullenbach et al.,
2004). The bar at the mouth of the Sepik River alters circulation, temporarily trapping
sediment and initiating gravity-driven sediment transport into the adjacent canyon
(Kineke et al., 2000), which acts as a conduit for these sediments to deeper waters.
Concurrently, a positively buoyant, or hypopycnal, plume carries sediment east and
west of the Sepik River mouth (Kineke et al., 2000; Kuehl et al., 2004). Characteristic
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 14
of these systems, a belt of muddy sediment is deposited on the mid-shelf, with a
significant portion of the river sediment budget lost to the slope or a nearby submarine
canyon (Sommerfield and Nittrouer, 1999; Kineke et al., 2000; Kuehl et al., 2004;
Harris et al., 2005). Investigating sediment transport pathways and accumulation
patterns will foster insight into stratigraphic formation, nutrient and contaminant burial,
and their effects on benthic community structure.
The Waiapu River, North Island, New Zealand (Fig 1) disperses an estimated 35
MT of sediment annually to the adjacent continental shelf despite having a relatively
small catchment area of ~1700 km2 (Hicks et al, 2004). Most of this sediment is
debouched onto the shelf during flood events that occur several times per year, with
maximum fluvial concentration often exceeding the theoretical river-initiated gravity-
flow threshold (Mulder and Syvitski, 1995; Hicks et al., 2004; Gisborne District
Council, GDC, data). The narrow shelf is approximately 20 km wide, and is dominated
by a recurving shelf-edge formed by the head scarp of a mega-submarine landslide
(Lewis et al., 1998 and Lewis et al., 2004). The high frequency of storm events makes
this an ideal location to assess sediment dispersal and preservation patterns for a
collision margin system.
III. Background
The geometry of the North Island, New Zealand collision margin has played a
significant role in sediment delivery and deposition on the shelf. The East Cape region
is characterized by a series of actively uplifting ranges and basins that comprise the
backstop and forearc basin of the Hikurangi subduction zone (Walcott, 1987). The East
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. coast rivers, including the Waipaoa, Uawa, and the Waiapu, drain this area, delivering
crushed and easily eroded Triassic to Pleistocene materials (Field and Uruski, 1997) to
adjacent continental shelf waters.
The shelf is narrow, only about 20 km wide, and bounded on the seaward side
by the Hikurangi Trough (Lewis et al., 1998; Collot et al., 2001; Lewis et al., 2004).
The Ruatoria Indentation, the head scarp of a vast submarine landslide initiated by a
subducting seamount 2 to 0.2 Ma and subsequent debris and avalanche flows between
40 to 170 ka define the shelf break (Lewis et al., 1998; Collot et al., 2001).
Transpression from the obliquely converging Australian and Pacific plates contributed
to the formation of a synclinal basin on the shelf (Lewis et al., 2004). Faulting within
the basin and shelf subsidence rates approaching 4 m/ka created enough space to
accommodate 1 km-thick Quaternary deposits and Holocene deposits about 100 m thick
(Lewis et al., 2004). Neogene and Quaternary faulting within the basin constrained the
main loci of deposition near the shelf break (Lewis et al., 2004).
The Waiapu River, New Zealand (Fig 1), has one of the highest sediment yields
in the world, 17,800 tons/km yr, due to high rainfall rates averaging 2.4 m/y, and the
erodible quality of the basin rocks (Hicks et al., 2004). Like other small rivers, the
basin area is too small to modulate storms (Milliman and Syvitski, 1992), resulting in
high-load floods every time there is a rainfall event. Although the average river
sediment concentration is 10 g/L with an average freshwater discharge of 86 m3/s,
fluvial concentrations associated with floods exceed 40 g/L, which is the estimated
threshold for high-density flows, several times a year (Fig. 2) (Mulder and Syvitski,
1995; GDC). The annual sediment load is 35 MT/y, which is large considering that the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 16
drainage basin is only 1734 km2. For comparison, the sediment load of all California
Rivers is 42 MT/y with a cumulative catchment area of a little less than 100,000 km2
(Farnsworth and Warrick, 2005). Most of the Waiapu River’s annual load is delivered
to the shelf during pulses of high discharge. The largest flood to affect the Waiapu
River area in the last 100 years was Cyclone Bola in 1988, which resulted in an
estimated maximum freshwater discharge of 4600m3/s, a suspended sediment
concentration of 60 g/L, and a total event yield of 73-93 MT, over twice the average
annual sediment load in only a few days (Hicks et al., 2004).
Seasonal climate changes and large-scale shelf currents may influence sediment
transport pathways and deposition once the material reaches the coastal ocean. For
example, the annual flood season runs from June-September; whereas the highest wave
energy is generally between April-July. Along the shelf break, the East Cape Current
(ECC) flows south along the eastern coast of the North Island (Chiswell, 2000) (Fig. 1).
Low temperature and lower salinity waters inshore of the ECC flow to the north
defined by the Wairarapa Coastal Current (Chiswell, 2000). A number of studies
indicate that these currents and the Wairarapa and Hikurangi eddies produced by these
currents may ephemerally affect shelf water circulation (Chiswell, 2000, 2003, 2005).
Waves are energetic during, or shortly after floods, with r.m.s. (root mean squared)
wave heights of 2.5 to 3.5 m and significant wave heights of 3.5 to 5 m (Ma et al., in
prep). These waves are capable of suspending fine-grained material to water depths of
-100 m.
IV. Objectives
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. This study seeks to identify sediment accumulation patterns on multiple time
scales, determine whether sediment transport history can be inferred from the sediment
record, and produce a sediment budget for the last 100 years for the Waiapu River
shelf. Comparison of this observational seabed data with the numerical modeling
results are presented in Chapter 3.
V. Methods
Kasten cores, measuring a maximum 3 m in length, and multi-cores, measuring
a maximum 1 m length, were collected during August 2003, near the end of the annual
flood season aboard the National Institute for Water and Atmospheric Research’s
(NIWA) RV Tangaroa. Maximum 0.5 m long box cores were also collected in May
2004, during the annual high wave energy period aboard the RV Kilo Moana from the
National Oceanic and Atmospheric Administration (NOAA) and run by the University
of Hawaii. A total of 33 Kasten cores, 28 multi-cores, and 63 box cores comprise the
data set.
Analyses included x-radiographs, radioisotope activity profiles, organic carbon
and nitrogen content, and grain-size. Kasten cores were first sub-sampled using
plexiglass trays (30 cm long by 3 cm thick) in order to preserve sediment structure. The
slabs of sediment were then exposed to X-rays using a Medison portable x-ray unit at
20 mA and 60 kV for an average 16 s and then developed on Kodak Industrex
Redipack™ film onboard. The resulting images reflect changes in bulk density (e.g.
grain-size, water content, sediment composition) such that dark and light features
represent relatively transparent and opaque sediments, respectively. All cores were sub
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. sampled in 2-cm thick sections at regular intervals along the length of the cores for
grain-size, organic carbon and nitrogen, and radiochemical analyses.
X-radiographs, radioisotopes, and grain size were used to characterize sediment
structure as well as indicate mixing and accumulation patterns on the shelf (Olsen et al.,
1981; Nittrouer et al., 1983; Kuehl et al., 1996; Hirschberg et al., 1996; Dellapenna et
al., 1998). The x-radiographs were used to identify physical and biological mixing in
the sediment column and to help interpret the radioisotope and bulk carbon and 8 13C
results. Multiple radioisotope signatures were investigated in order to assess mixing
processes and accumulation on multiple time scales. For example, 7Be, with a X\n of 53
days, was used to infer mixing and accumulation on time scales of several months, 910 whereas excess Pb activities, with a Xm of 22.3 years, was used to asses processes on
time scales ranging from a few decades up to 100 years. Bomb-produced 137Cs first
appeared in sediments in 1954, and was used to assess sediment mixing and excess
210Pb accumulation rates (Nittrouer, et al., 1983). Atmospheric precipitation is the
primary source of these radioisotopes, but scavenging from coastal and fluvial waters as
well as in-situ decay of 226Ra (for 210Pb) are also important sources (Bruland et al,
1974; Dukat and Kuehl, 1995; Srisuksawad et al., 1997).
Excess 210Pb (ti/2 =22.1 years) activities were determined following a modified
version of Flynn’s (1968) methods. The sample was spiked with a known amount of
209Po, the yield determinant, and partially digested in 16 N F1N03 and 6 N HC1,
resulting in the release of 210Po from the fine fraction. An estimation of 226Ra-supported
activities was allowed when excess 2I0Pb activities dropped to negligible levels at depth
in the core. Both 137Cs and 7Be activities were measured using a planar high purity
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. germanium detector coupled with a multi-channel analyzer. Approximately 70 -90 g of
wet sediment was packed into a petri dish, and then gamma radiations were measured
for 24-48 hours.
The elemental and isotopic composition of organic matter was determined using
an Isotope Ratio Mass Spectrometer (IRMS) at the University of California at Davis
Stable Isotope Facility. Samples from a single core, KC18, were acidified with 10% 11 HC1 to remove inorganic carbon and dried. Results included 8 C, and bulk carbon and
nitrogen. The results were used to identify terrestrial signals preserved in the sediment
record.
Waiapu River shelf surface grain size was analyzed in order to determine where
muddy sediments are concentrated on the shelf. Sediments were wet sieved to separate
the >63 pm sand fraction and then standard pipette methodology was used to determine
the relative amounts of silt and clay for surface samples from both cruises. Higher- 91 n resolution analyses were performed on KC18 for comparison with Pb and carbon
data using a Sedigraph 5100 at the Skidaway Institute for Oceanography
Sedimentology Lab.
VI. Results
The following sections describe sedimentary and geochemical structures of
Waiapu River shelf sediments.
A. Sedimentary texture and structure
Seabed surface grain size analyses revealed that the mid- to outer shelf was
dominated by muddy sediments, whereas slightly sandier sediments were found on the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0
inner to mid-shelf boundary, between 30-40 m water depth, and along the shelf edge,
between 140 - 200 m for both cruises (Fig.3). Most samples were primarily a mix of
clay and silt. Surface samples averaged 9.3 + 14.9 % and 7.5+9.7% sand, 53.8+_9.4%
and 58.0+8.5% silt, and 36.9+12.2% and 34.5+9.3% clay for the August and May
cruises, respectively (Appendix A). The small differences in surface sample grain size
between the two cruises can be explained by spatial heterogeneity, and may not
necessarily represent temporal changes.
Surface grain size, Kasten core length, and x-radiographs of the sediment cores
showed that sedimentary structure and texture/consolidation varied both along- and
across the shelf (Fig. 4). The spatial distribution of muddy sediments on the mid-shelf
was patchy. Several times the Kasten corer failed to recover sediment, suggesting that
the seabed in these areas was not muddy. Those cores collected between 30 - 45 m
water depth, with a seaward limb extending to 80 m water depth near the Waiapu River
mouth, were shorter than 40 cm, suggesting that surface sediments in these areas were
underlain by coarse materials that limited Kasten core penetration (Appendix A). The
mid-shelf was characterized by laminations from a few mm up to 10 cm-thick.
Seaward of the mid-shelf region, the number of preserved laminations within the
sediment profile decreased radially from the Waiapu River mouth, and increasingly,
sediments showed signs of bioturbation. Kasten core sediment penetration was greatest
on the middle to outer shelf between 60 - 150 m water depth, gradually decreasing to
the north and south, and suggesting that shelf sediments were predominantly muddy.
Sediments furthest from the Waiapu River mouth contained a few laminations
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 1
interspersed with mottled and bioturbated sediments. Kasten cores from the shelf break
tended to be shallow, suggesting that muddy sediments were patchy in this area.
B. Radiochemical data
Multiple radioisotopes signatures were investigated in order to assess transport
processes and accumulation patterns on a range of time scales. For example, naturally
produced 210Pb (ti /2 = 22.3 years) was used to interpret mixing and accumulation for
time scales of a few decades up to 100 years, whereas the presence of 7Be (ti /2 = 53.3
days) indicated sediment accumulation only within the last few months. o i n Three patterns of deposition were identified by excess Pb sediment profiles:
cores shorter than 50 cm that had uniform activity profiles; long and short cores,
between 50 - 200 cm long, showed logarithmic excess 210Pb decay; and non-steady-
state long cores, up to 250 cm, had variable 210Pb activity with depth (Fig. 5). Each
profile was assessed to determine the supported 210Pb level. Where this supported
activity could not be determined, an average value from the shelf was substituted. The 0*)(\ estimated supported level for KC18 was verified by measuring the Ra activity using
the methods employed by Cutshall (1983) and Moore (1984).
The characteristics of the excess 2I0Pb profiles changed both along-shelf and
across-shelf. Adjacent to the river mouth, the inner to mid-shelf cores, KC 7, 8, 25, and
6, tended to be short, at less than 50 cm, with uniform (i.e. mixed) excess 210Pb profiles
and activities ~ ldpm/g (Fig. 5 and 6). Cores from the mid- to outer-shelf were longer
and excess 210Pb activities were found at greater depths, indicating a zone of muddy
accumulation (Appendix A). Two types of excess 210Pb activity profiles were identified
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. on the mid- to outer shelf: steady-state profiles that showed logarithmic decay with
sediment depth, and non-steady-state profiles that showed variable excess 210Pb
activities with depth (Fig. 5, 6, and 7). Excess 210Pb signatures farthest from the
Waiapu River mouth, Kasten cores 1, 9,22,23, 31, 32, and 34, exhibited steady-state
geochemical signatures. Of these, KC 32, and 23 were shorter cores, less than 50 cm,
collected along the shelf break. The steady-state sediment core collected from the slope,
KC34 at 693 m, indicated that some sediment is bypassing the shelf and accumulating
on the slope (Fig 5, 6, and 7).
The non-steady state cores were characterized by a general decrease in excess
activity with core depth, interrupted by sediment layers that had significantly lower
activities, dubbed low activity layers, than the activities above and below the sample.
91 fl These layers, defined by low excess Pb activities and compared with x-radiographs,
varied in thickness from a few centimeters up to 41 cm thick. Event layer frequencies
and the total thicknesses of event layers in the sediment column varied along and across
shelf (Fig. 7; Table 1). Total event layer thicknesses exceeded 80 cm, and were thickest
on the outer shelf between 80-100 m depth, extending to the south of the Waiapu River
mouth. Event layer thickness decreased to the north and south of this region, also
decreasing towards the shelf break save for a few cores in transect D. These non-steady
state cores correlated with x-radiographs dominated by laminated structures, whereas
those cores with steady-state profiles tended to have preserved bioturbation signals and
fewer physically-produced laminations (Fig 8). The potential mechanisms that
produced low excess activity layers will be discussed below.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 3
Surface sediments from the Waiapu River shelf reveal that the presence of 137Cs
was patchy, with low activities. Therefore, one core (KC 9) from the Waiapu River
71 n shelf with steady-state excess Pb activities was chosen for an in-depth analysis of
i 17 7 1 n Cs in order to assess its potential use in corroborating Pb accumulation rates. The 117 Cs activities within the core were too low to identify either the first appearance of
137Cs or the 1963 peak, presumably as a result of low southern hemisphere bomb fallout
and/or dilution of the signal by the extremely high Waiapu River catchment yield.
Further examination of the cores for 137Cs was not attempted based on these results.
Recent depositional patterns were assessed by measuring 7Be activities in
surface samples from both the August and May cruises (Fig. 9). The highest activities
were found at shallower depths, less than 80 m, although some samples collected from
the northern part of the shelf also contained 7Be. Cores collected from deeper waters
had 7Be activities below detectable limits. The activities ranged from 0.3-1.8 dpm/g
with an average of 0.8 + 0.5 dpm/g. The May, 2004 cruise revealed a larger area of
recent deposition than the August cruise, possibly due to a larger sampling area (Fig. 9).
The presence of 7Be in shallower waters and high accumulation of excess 210Pb in the
mid- to outer-shelf region illustrated a potential disconnection between short-term and
long-term processes on the shelf during the observation period.
C. Stable isotope data
In order to better establish the provenance of the low 210Pb activity layers, bulk
carbon and nitrogen, and 8 13C were measured in a single core in the outer shelf region,
KC18 (Fig. 10). The average 8 13C and C/N values for the core were -24.8 + 0.4°/oo and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 4
10.7 + 0.9 0/00, respectively, indicating that the sediments were a mixture of mostly
terrestrial and marine carbon signals with a bias towards terrestrial sources (Fig 11).
Down-core variations in C/N ranged from 9.6 - 13.8, and del 13C ranges from -25.87
to -24.35, reflecting the variability in 2I0Pb activities observed in KC18.
VII. Discussion
This section investigates how accumulation rates are estimated, the Waiapu
River shelf sediment budget, flood event preservation in continental shelf sediments,
pulsed event layer sediments, and long- versus short-term accumulation patterns. 0 A. Calculating Pb accumulation rates
There are multiple approaches available to assess accumulation rates and
patterns on a shelf dominated by storm deposition (Sommerfield and Nittrouer, 1999;
Wheatcroft and Drake, 2003, Mullenbach et al., 2004). Those sediments with layers of
uniform excess 210Pb activity profiles were interpreted to have been emplaced within a
short period of time or to have been mixed by physical and/or biological processes
(Nittrouer and Sternberg, 1981). Accumulation rates for those sediments identified as
steady-state were calculated by applying the constant flux/constant sedimentation rate
model (Appleby and Oldfield, 1992) with a surface mixed layer defined by the
observed depth of bioturbation (Nittrouer et al., 1983). Evaluating the depositional
history of sediments with non-steady-state profiles was more complicated because there
are various mechanisms that can affect adsorption of Pb010 (Appleby and Oldfield,
1992; Dukat and Kuehl, 1995, Sommerfield and Nittrouer, 1999).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Multiple theoretical scenarios have been proposed to explain non-steady-state
profiles that are characterized by layers of low excess 210Pb activity. Assuming that the
residence time of particles in the water column is constant, excess 210Pb adsorption by
sediment particles may be influenced by sediment grain-size (Nittrouer et al., 1979;
Kuehl et al., 1989), mineralogy (Dukat and Kuehl, 1995), and organic carbon content
(Davis, 1984; Moore and Dymond, 1988). Seabed profiles can also be affected by
small-scale slumping (Sanford et al., 1990; Kniskem and Kuehl., 2003). Additionally,
if particle affinity is constant and residence time in the water column is variable, the
^ |A resulting excess Pb activity profile may be non-steady state (Dukat and Kuehl, 1995).
For example, if sediments are resuspended and moved across the shelf or delivered to
the shelf such that there is little interaction with 9 Pb-rich1 n shelf waters, then the 91 a resulting initial Pb concentrations will be low (Dukat and Kuehl, 1995; Sommerfield
and Nittrouer, 1999). This occurred on the Amazon River Shelf when remobilization of
a fluid mud bed resulted in non-steady state profiles (Dukat and Kuehl, 1995, Kuehl et
al., 1996). Another explanation assumes constant but limited Pb910 availability in the
water column, where fluctuations in suspended sediment concentration in the water
column cause varying initial activity in the seabed (Dukat and Kuehl., 1995,
Sommerfield and Nittrouer, 1999). Sediments delivered to the Eel River shelf during
large flood events overwhelmed the 210Pb available in the water column and were
rapidly deposited on the inner shelf (Sommerfield and Nittrouer, 1999). Sediments
were then resuspended and moved to the mid-shelf via gravity-driven flows
(Traykovski et al., 2000; Scully et al., 2003; Harris et al., 2005) resulting in sediment
layers with low excess 210Pb (Sommerfield and Nittrouer, 1999). The resultant deposits
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 6
had low excess 210Pb activities and were relatively clay rich (Sommerfield and
Nittrouer, 1999; Leithold et al., 2005, Wheatcroft and Drake, 2003, Mullenbach et al.,
2004). Whichever of these processes predominate, the resulting excess 210Pb activity
profiles are non-steady state, complicating interpretation of mixing and accumulation
rates.
Down core changes in sediment grain size on the Waiapu River shelf were
evaluated for cores from transect D that exhibited non-steady state profiles to make sure
that the low activity layers were not an artifact of grain-size. Specifically, differences
in grain-size for sediments from low spike layers were compared with grain-size from
91H layers that did not exhibit low excess Pb activities (Fig. 12 A and B). Although there
appeared to be a weak relationship between grain size and excess 210Pb activity (Fig.
12A), it was not enough to account for the low activity layers (Fig. 12B). 910 Previous work has identified increased absorption of Pb onto particles with
high organic carbon content (Moore and Dymond, 1988; Biscaye and Anderson, 1994). 910 When a negative correlation is observed, it can be an indication of limited Pb
availability (Radakovitch and Heussner, 1995). However, the organic carbon weight
percent for KC18 varied between 0.35-0.73%, revealing a weakly positive correlation 910 (r-squared = 0.23) between organic carbon and excess Pb activities that was
controlled by the low excess 210Pb activity layers (Fig. 13). A positive correlation (r-
squared= 0.51) was found among the samples associated with low activity layers; there
was no correlation for those samples not associated with low activity layers. Although
there was a weak correlation between organic• carbon content and excess Pb210 activity, • •
it was caused by a covariant rather than a causal relationship.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A comparison of 8 13C and atomic C/N with excess 210Pb activities revealed two
populations driving the correlations. Low excess 210Pb activities associated with the
pulsed event layers displayed relatively terrestrial carbon signatures compared to
• •• 710 • • sediments with higher excess Pb activities. It is likely that the observed high
•••• 9in « variability in excess Pb values was produced by a combination of transport
mechanism prior to sediment burial. The relatively terrestrial signal represented by the
low excess 210Pb activity layers suggests that these sediments were deposited rapidly
and were not profoundly reworked by the benthos.
Accumulation rates were calculated by removing the low spikes from the data
and using a least squares linear regression of the log for excess 210Pb, the approach used
by Sommerfield and Nittrouer (1999). On the Eel River shelf, the low excess 210Pb
activity layers were presumed to represent instantaneous deposition and thus the
samples with low activity were removed from the accumulation calculations. In this
paper, samples with low excess 210Pb activity were identified by evaluating whether the
sample excess 210Pb activity was lower than the samples both above and below. A
maximum 10% overlap of activity error was allowed for those samples identified near
or within a low activity layer. The rates calculated using this method were different by
5 - 46% from the rates calculated where low activity samples were included, far greater
than the 10% difference observed on the Eel River shelf (Table 1) (Sommerfield and
Nittrouer, 1999). The probable reason for this difference was that the low excess 210Pb
values from the Waiapu River shelf were much lower than that observed on the Eel
River shelf. These lower values could be caused by either lower residence times in the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 8
water column prior to deposition, lower bioturbation rates post-deposition relative to
the Eel River shelf, or both.
The approach used by Sommerfield and Nittrouer (1999) assumes that depth in
the sediment bed is a proxy for time:
accumulation rate = A*Z/(ln (C0/Cz))
(Appleby and Oldfield, 1992; Nittrouer et al., 1983).
In this equation, A is the decay constant, Z is depth in the sediment bed, and C0
and Cz are the excess 210Pb activities at the surface and at depth, respectively.
However, when the seabed record is dominated by pulsed event layers, this equation
may not adequately represent the burial history. A background accumulation rate was
calculated and used to assess the impact of pulsed event delivery to the seabed. In this
method, not only were the low excess 210Pb activities removed from the profile, but the
thickness of the pulsed event layer was also removed (Fig 7; Table 1). The boundaries
of the pulsed event layer were constrained by evaluating the excess 210Pb profile and
the associated laminated structures within the x-radiographs. For those few instances
where no visible boundary could be identified in the x-radiograph, then the boundary
was defined as the mid-point between a low excess Pb activity sample and the
nearest high activity sample provided the intervening distance was no larger than 10
cm. The identified pulse event layer thicknesses ( Z e v ) as well as their associated low
2I0Pb activities for the length of the sediment core were removed from the calculations
such that:
Accumulation rate = A *(Z -Zev)/ (In (C0/Cz)).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 9
Although there was little difference in the average accumulation rates produced by each
method, 1.5+0.9 cm/y for the Sommerfield and Nittrouer (1999) approach and 1.2 +0.6
cm/y for the background accumulation approach, the difference in the two approaches
provides an idea of how much the sediment record was influenced by pulsed
contributions and the preservation of this type of signal on the shelf (Fig. 7). Pulsed
event layer preservation increased with increasing accumulation rate, suggesting that
deposition from dilute suspension was important for preservation on the shelf as well as
deposition by gravity flows. Additionally, identification of the pulsed event layers
showed where gravity-driven flows were important on the shelf
Three zones of accumulation were identified. There was a mid-shelf region 30 -
45 m water depth where there is no long-term accumulation of muddy sediment. These
cores exhibited low excess activities ~1 dpm/g that were uniform over the short depths
of the profiles. Long-term accumulation rates were not calculated for cores (KC 28, 30,
3, and 26) on the shelf when the majority of the sediment profile was dominated by low
excess Pb activity layers. Shelf excess Pb accumulation rates for the shelf ranged
from 0.2-3.3 cm/y with an average of 1.5 + 0.9 cm/y"(Fig. 6). The overall pattern of
accumulation trended SW to NE across shelf bathymetry, roughly overlapping the
Quaternary shelf basin identified by Lewis et al. (2004; their figure 8). The area of
highest accumulation for both methods was on the mid-to outer-shelf. Those sediments
most affected by pulsed event layers were spread out along the shelf between 80-100 m
depth. Pulsed event layers identified in transect D sediments indicated that this area of
the shelf was acting as a conduit for sediments to deeper waters. A recent investigation
of sediment accumulation along the shelf break supported this hypothesis (Addington et
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 0
al., in press). The high accumulation rates and the presence of pulsed event layers in
the cores at the northern extent of the study area suggest that Waiapu River sediments
may have been transported beyond the study area. The southern extent of Waiapu
River sediment dispersal does not likely extend much further than the study area based
on the low accumulation rates observed here and the bioturbation signals preserved in
the sediments. Numerical modeling results indicated that most sediments exported
from the system were transported to the north of the river mouth (Kniskem, Chapters 2
and 3).
B. Sediment budget
A portion of the sediments delivered by Waiapu River, New Zealand were
deposited on the adjacent narrow shelf and slope (Fig. 14). For the most part, muddy
sediments were only ephemerally deposited on the mid-shelf, between 30-50 m,
because energetic waves and currents (Ma et al., in prep; Wright et al., 2006) moved the
sediments offshore (Kniskem, Chapters 2 and 3). Therefore, the inner shelf was likely
acting as a bypass zone for most muddy sediments. Seismic data and vibracores
collected from this region, however, revealed inter-bedded sands and muds to the south
of the river mouth and massively emplaced sands and muds to the north of the river
mouth, accounting for an estimated 8-15% of the annual fluvial load (Wadman and
McNinch, in review). This estimate is commensurate with estimates from the Eel River
shelf system o f-13% retention (Crockett and Nittrouer, 2004). Accumulation rates of
muddy sediments on the mid- to outer shelf, ranging between 1.2 - 3.5 cm/y, were up
to three times higher than usually observed on continental shelf settings (Wheatcroft
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and Drake, 2003). Rates observed on the shelf break and slope, greater than 130 m
depth, ranged between 0.2 and 1 cm/y, confirming that terrestrially derived sediments
reach deep waters.
A fine-sediment budget for the shelf was calculated using the apparent 210Pb
accumulation rates (Figs 5, 6, and 14). The offshore extent of the shelf was defined by
the 200-m isobath on the east. The northern and southern boundaries were determined
using a combination of facies observations from this study and the observed shelf basin
sediment thicknesses from Lewis et al. (2004). The southern boundary was well
defined by low accumulation rates and bioturbation structures. The northern boundary
was less well defined because accumulation rates were still significantly high at the
northern-most extent of the study area. In addition, numerical modeling results
indicated that sediments were primarily exported from the shelf to the north of the
Waiapu River mouth (Kniskem, Chapters 2 and 3), making it difficult to define the
extent of the shelf deposit. The northern boundary was therefore based on the basin
dimensions defined by Lewis et al. (2004) and the location of sandy sediments just to
the north of East Cape (Fig. 1).
The shelf was divided into sub-regions based on 210Pb accumulation rates. The
apparent 210Pb accumulation rates for each sub-region were averaged and applied over
the whole. Porosity was held at 60 % based on average values obtained from several
cores. Annual fluvial delivery was assumed to be 35 MT per year (Hicks et al., 2004).
Sediment retention on the shelf ranged between 32 - 39% dependent on the
accumulation rate estimated for the inner shelf. The inner shelf (< 30 m) accounted for
between 8 - 15% of the total budget; the remainder of the shelf retains ~ 25% of the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 2
annual fluvial budget. Errors associated with this estimate of shelf sediment retention
included: spatial resolution of the sampling scheme, unrealistic uniform porosity, only
apparent accumulation rates were calculated for the non-steady state cores, 137Cs could
not be used to corroborate the steady state core accumulation rates, and the inner shelf
budget was estimated using a different approach. Additionally, this approach assumes
that there was no variability of annual river input.
C. Event preservation
The preservation potential of physical transport and mixing processes in the
sediment record varied both along and across shelf. Sediments collected from depths
o 1 n shallower than -50 m were physically dominated. Their mixed excess Pb profiles
with supported levels at less than 50 cm depth, were characterized by thick, up to 15
cm, packets of laminations with little to no bioturbation structures. Laminations
decreased and bioturbation increased with increasing distance from the Waiapu River
mouth. The influence of the river on sediment structure was evidenced by the semi
circular pattern of these deeper water facies (Figs. 4 and 8). If wave energy was the
primary influence on sedimentary structure, then the pattern would have paralleled the
bathymetry.
Generally, as accumulation rate increases on a continental shelf, the increase in
organic carbon delivery results in a prolific benthic community that destroys storm
event layers (Wheatcroft and Drake, 2003). This is because, on most continental
shelves, accumulation rates are too low to adversely affect benthic community structure
(Boudreau, 1994; Wheatcroft and Drake, 2003). World-wide shelf accumulation rates
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. range between 0.1 and 1.0 cm/y (Wheatcroft and Drake, 2003), although higher
accumulation rates have been found on large river-shelf systems such as the Amazon
(Kuehl et al., 1996), Ganges-Brahmaputra, and the Changjiang Rivers (Nittrouer et al.,
1985). Biodiffusivity rates typically range between 10 and 100 cm /y (Wheatcroft and
Drake, 2003). Waiapu River shelf accumulation rates exceeded 1.0 cm/y and sediment
delivery was largely controlled by floods capable of depositing layers ~ 20 cm thick
(Kniskem, Chapters 2 and 3). These high deposition rates, combined with wave- and
current-supported gravity flows allowed physical rather than biological processes to
dominate sedimentary structure on the shelf. Previous studies have indicated that
benthic communities may be negatively impacted by high accumulation rates and
physical mixing in estuarine systems, where accumulation rates are of the order of
those observed on the Waiapu River shelf (Schaffner et al., 1987; Dellapenna, et al.,
1998). Strong shelf currents on the Amazon River shelf result in a 1 m-thick near bed
layer (Kuehl et al., 1995; Kineke et al., 1996) that can negatively impact benthos (Aller
and Stupakoff, 1996).
Theoretically, a freshly deposited event layer must transit through the surface
mixed layer (Goldberg and Koide, 1962; Guinasso and Schink, 1975), commensurate
on the shelf with the zone of bioturbation between 5-30 cm, quickly enough that the
physical characteristics of the event bed are not dissipated (Nittrouer and Sternberg,
1981; Wheatcroft, 1990; Wheatcroft and Drake, 2003). Once the event bed is advected
through the surface mixed layer, it may be preserved in the sediment record (Nittrouer
and Sternberg, 1981; Nittrouer et al., 1983). In their study of the sediments deposited
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 4
on the Eel River Shelf, Wheatcroft and Drake (2003) showed that event preservation on
most continental shelves is unlikely.
The predominance of lamination structures in the sediments on the Waiapu
River shelf, however, showed that flood events were preserved in the sediment record
(Figs. 4 and 8). Using the assumptions in Wheatcroft and Drake (2003) for average
shelf values of mixed layer thickness, and accumulation rates from the Waiapu River
shelf (Table 1), an event bed of only a few centimeters would be preserved. It is
unlikely, however, that the surface mixed layer on the Waiapu River shelf was defined
by bioturbation activities. The dearth of observed bioturbation structures, especially in
the high accumulation zone, suggests that physical processes dominate the mixed layer
thickness in the mid- shelf region rather than bioturbation. High accumulation rates
combined with frequent disturbances either through floods or wave events may have
hindered the ability of benthic macrofauna to survive in this environment. Evidence of
this is borne out by the increased frequency of bioturbation structures and decreased
frequency of laminations away from the river mouth. The uniform activity observed in
sediments at water depths less than 50 m implied that the surface mixed layer may have
been as thick as 30 cm. Resuspension by waves and currents likely influenced the
surface mixed layer thickness and decreased with depth as waves were attenuated.
The high number of event beds preserved in the shelf sediments (Fig. 7), when
taken into account with the accumulation rates (Fig. 6), indicated that the frequency of
flood and storm events was high enough that event layers were quickly transited
through the mixed layer. Several event layers observed in the x-radiographs were on
the order of 5-10 cm (Fig. 4). Additionally, transport and deposition simulated during a
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2004 benthic tripod deployment indicated that, days after a flood, sediment deposits
could be as thick as 20 cm (Kniskem, Chapter 2). The paradigm proposed by
Wheatcroft and Drake (2003), Tm=[Lb-Ls/2]/S, where the transit time of the event layer
is Tm, Lb is the mixed layer thickness, Ls is the event bed thickness, and S is the
accumulation rate, used a surface mixed layer ranging between 10-15 cm. Application
of this model to the Waiapu River shelf using a surface mixed layer of 15 cm and an
average shelf deposition of 1.5 cm/y would require an event bed of only 5 cm to result
in bed preservation. Applying the equation to steady-state profiles with accumulation
rates between 0.2 -1.2 cm/y, however, revealed that only very thick event beds might be
preserved in the sediment record in areas with lower accumulation rates. The steady
state profiles are therefore likely a product of mixing by bioturbation and decreased
pulsed sediment delivery.
High load floods that deliver a significant fraction of the annual load to the
shelf, or floods such as Cyclone Bola that delivered an estimated 70 -90 MT in a few
days, may produce event beds exceeding the surface mixed layer thickness proposed by
Wheatcroft and Drake (2003). The equation proposed by Wheatcroft and Drake (2003)
does not account for event beds that exceed the surface mixed layer. Such deposits
would result in a negative number for the transit time. Therefore, this equation needs to
be revised to account for deposition and accumulation on shelf systems characterized
by frequent floods and pulsed event layers whose thickness can exceed that of the
surface mixed layer. Furthermore, this study indicated that bioturbation may not
always be the dominant mixing process, requiring refinement of Wheatcroft and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 6
Drake’s (2003) assessment of the surface mixed layer thickness on shelf systems with
high accumulation and energetic waves and currents.
D. Identifying low excess 210Pb activity layers
Those sediments that were dominated by pulsed event layers, identified by non- i n steady state Pb profiles, were found along the extent of the shelf from East Cape in
the north to Tolaga Bay in the south (Fig. 7). Most of the cores are located on the mid-
and outer shelf from 50 to 120 m depth. Furthermore, all of the cores from transect D,
extending to the shelf break, were non-steady state, indicating that this region may act
as a conduit for terrestrial materials. Sediments from basins along the shelf break
contained pulsed event layers and were located just seaward of transect D (Addington
et al., in press), supporting this hypothesis. As shown by KC34 and other cores
collected on the shelf break and slope, terrigenous sediments accumulated at rates of up
to ~1.4 cm/y (Addington et al., in press), higher than usually observed in these
environments (Wheatcroft and Drake, 2003). The along-bathymetry orientation of the
non-steady-state cores suggested that wave-initiated gravity flows played a role in
terrigenous sediment distribution on the shelf. Alternatively, strong along shelf
currents may have caused dilute sediment suspension convergence along the shelf,
potentially producing a near-bed fluid mud layer resulting in pulsed event layers (Ma et
al., in prep). In either case, energetic waves (Kniskem, Chapters 2 and 3) prevented
muddy sediments from accumulating on most of the mid-shelf between 30-50 m water
depth.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Several potential scenarios may produce non-steady state signatures (Dukat and
Kuehl, 1995, Sommerfield and Nittrouer, 1999). Rapid deposition of sediments during
flood events, hyperpycnal plumes, and wave-initiated gravity-driven flows are potential
9 1 n transport mechanisms that can produce low excess Pb layers on shelves influenced
by small mountainous river systems (Sommerfield and Nittrouer, 1999; Kineke et al.,
2000, Milliman and Kao, 2005). Sommerfield and Nittrouer (1999) observed a
radioisotope signature on the Eel River shelf similar to that on the Waiapu River shelf.
Sediments delivered to the shelf during large flood events overwhelmed the 210Pb
available in the water column and were rapidly deposited on the inner shelf
(Sommerfield and Nittrouer, 1999). Sediments were then resuspended and moved to
the mid-shelf via gravity-driven flows resulting in clay- and organic carbon-rich flood
layers with low excess 210Pb (Sommerfield and Nittrouer, 1999; Leithold et al., 2005,
Wheatcroft and Drake, 2003; Harris et al., 2004, 2005).
The low excess 210Pb activities found in KC18 were weakly correlated with
grain-size, carbon content, and x-radiographs (Figs. 4, 10, 12A, and 13). High 910 resolution grain-size data showed, however, that several of the low Pb spikes were
coarser than the mean grain size, whereas other pulsed event layers were finer (Fig. 12
A and B). The lack of coarsening at 45 cm and 84-89 cm may be explained by sampling
resolution, the relatively higher bioturbation signature found in those sections of the
Kasten core, or difference in sediment transport mechanism (Fig. 4 and 12B). Each
pulsed event layer identified by excess 210Pb activity was associated with a relatively
terrestrial signal in the carbon data (Figs 10 and 13). The variability in the carbon
weight percent response for the identified pulsed events suggested variations in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. sediment delivery mechanisms or of Waiapu River catchment sediment sources (Fig.
11) (i.e. percent vascular plant material; Leithold and Blair, 2001). The low 2I0Pb
activity spikes did not always correlate with laminations in the x-radiographs, and not
all laminations correlated with low 210Pb spikes. Sediments at 21 cm and 84-89 cm
were associated with darker or finer laminations, whereas the other pulsed event layers
were associated with lighter, or coarser, laminations. It is important to note, however,
that there was about 5 cm of error in correlating the seabed data with x-radiographs due
to the sampling method.
The trends identified in KC18 suggested that not all of the low 210Pb activity
layers were the result of the same transport mechanism. Wheatcroft and Drake (2003)
suggested that by comparing the peak weight percent of the <20 pm fraction to the
mean weight percent of that fraction for the core, that it is possible to differentiate
between wave-supported flow deposits, during storms not associated with a river flood,
and flood deposits. When the <20 pm fraction was finer than two standard deviations
of the mean, it was considered a flood deposit; wave resuspension deposits were
coarser than two standard deviations. The application of this method to sediments from
the Eel River Shelf indicated that floods largely influenced strata formation (Wheatcroft
and Drake, 2003; their figure 8). When this same approach was applied to KC18, some
of the low 210Pb spikes were associated with finer sediments, and some were associated
with coarser sediments (Fig. 15). The finer layers may represent a hyperpycnal plume
deposit or wave- or current-initiated gravity-driven flow during floods. The coarser
sediments may represent gravity-driven flows or simply represent wave or current
resuspension into the water column following a river flood. There is also the possibility
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 9
that the sampling resolution of these sediments was too low to effectively pick up
changes in grain size within a flood layer. Waves and currents on the shelf were strong
enough to keep sediments in suspension up to 3 g/L in the wave-current boundary layer
(Wright et al., 2006), and numerical modeling efforts indicated that sediments can be
transported in wave-supported gravity flows, supporting the hypothesis that gravity
flows may be preserved in the sediment column (Kniskem, Chapter 2).
Analyses of Waiapu River shelf sediments indicated that multiple sediment
transport mechanisms may influence fine-scale strata formation on the shelf. Buoyant
delivery of sediments likely resulted in deposition of those sediments not associated
with pulsed event layers. Multiple sediment transport mechanisms may be associated
with the pulsed event layer sediments. In summary, four possible transport pathways
are suggested by the seabed data for the pulsed event layers: 1) high sediment 710 concentration buoyant river plume that quickly overwhelmed available excess Pb in 010 the water column such that sediment deposited during this event had low excess Pb
activities; 2) high river sediment concentrations that resulted in a negatively buoyant,
hypopycnal, gravity driven flow which quickly deposited sediments such that not much
excess 210Pb was scavenged onto sediment particles; 3) wave- or current-supported
gravity-driven flow; 4) either gravity flow or hyperpycnal plume delivery followed by
wave resuspension and subsequent winnowing of the flood layer deposit.
E. Long- vs. short-term accumulation
Radioisotope activities on the Waiapu River shelf indicated that sediment
^ 1 ft depositional patterns varied over long- and short-term time-scales. Excess Pb
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. activities showed that modem sediments were deposited on the mid-shelf to beyond the
shelf break, ranging from 50 m - >600 m water depth (Figs. 6 and 8). The maximum
depths of 210Pb indicated that up to 2.5 m of sediment were deposited on the mid-shelf
and the outer shelf in the last 80-100 years. The modem deposit thickness decreased
towards the shelf edge and along-shelf from the river mouth. The location of the 210Pb
depocenter and the spatial extent of the modem deposit are similar to the area defined
for overall Holocene deposition (Lewis et al., 2004). Muddy sediments were
ephemerally deposited on the mid-shelf, between 30-50 m, as indicated by the presence
of 7Be in surface sediments and the mixed, low excess 210Pb activity profiles (Fig. 9).
Furthermore, the presence of 7Be only in the mid-shelf region for both the August 2003
and May 2004 cruises suggests that sediments were first deposited on the inner and
mid-shelf and subsequently transported to deeper waters. Preliminary model results
indicated that this occurred when along shelf currents were strong enough to trap the
fresh water and sediment plumes on the inner shelf. However, when wave energy was
sufficiently stronger than along-shelf wind-forced currents, sediments were rapidly
moved offshore (Kniskem, Chapters 2 and 3).
The Waiapu River shelf system may be subject to seasonal signals: a high wave
energy period (April-July) and a high sediment delivery period (June -September). The
two cruises attempted to capture these seasonal differences in recent depositional
patterns, but similar short-term depositional patterns were observed during both seasons
(Fig, 9). Seasonal variability may be difficult to discern because local wind patterns
during storms may be more important. For example, subtropical cyclones usually pass
to the east of North Island, producing NE to SE winds during the beginning of the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 41
storm, switching to NW winds after peak fluvial discharge (Orpin, A.R. pers. comm.).
A similar wind pattern was observed during the winter storms of 2004, indicating that
initial sediment deposition was controlled by local wind-driven shelf currents. When
storms are associated with atypical wind patterns, sediments may be transported
quickly to deep waters rather than being stored temporarily in the inner and mid-shelf
areas (Orpin et al., 2006; Kniskem, Chapters 2 and 3).
Although there is a lack of long-term, local wave data, there may be seasonal
differences in the characteristics and preservation potential of flood deposits. For
example, sediments delivered to the shelf by storm events during a high wave energy
period will be deposited further offshore than those delivered during a low wave energy
period (Traykovski et al., 2000; Kniskem, Chapter 2). The seabed data was not
sufficient to resolve this issue; modeling results suggested that although there may be
differences in deposition on the scale of months due to differences in wave conditions,
large wave events were capable of reworking sediments and moving them to deeper
waters (Kniskem, Chapters 2 and 3). Seasonal depositional patterns may therefore be
masked by reworking of sediments by large wave events.
Fluvial delivery to the Waiapu River shelf was dominated by storm frequency
(Fig. 2), and initial sediment deposition was confined to shallow waters when along-
shelf wind-driven currents were strong (Kniskem, Chapters 2 and 3). Sediments were
not retained on the mid-shelf over time-scales longer than a few months. By the time
sediments were redistributed to deeper waters, Be activities may have been below
detectable limits. As a result, there is an apparent disconnect between short- and long
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 2
term depositional patterns. Seasonal patterns of deposition could not be resolved with
the sampling scheme.
Although other coastal systems temporarily store sediments in shallower waters
before sediments are resuspended and transported to deeper waters, the presence of 7Be
is not confined to shallow waters (Gerald and Kuehl, 2006; Sommerfield et al., 1999).
The presence of 7Be only in water depths less than 80 m was likely a function of the sea
conditions during the flood events just prior to sampling. If along-shelf currents were
weak during a flood event, terrigenous sediments may have been distributed to deeper
waters quickly such that 7Be activities would be detectable (Kniskem, Chapters 2 and
3).
VIII. Conclusions
1) Accumulation rates on the Waiapu River shelf were high, with an average of 1.5
cm/y :L0.9; ranging from 0.2 -3.5 cm/yr, which exceeded the global range estimated for
continental shelf environments (Wheatcroft and Drake, 2003). Muddy sediments
tended to bypass the shelf down to 50 m water depth. The highest accumulation rates
were found on the mid-outer shelf, with rates ranging from 1.2 -3.5 cm/yr, and
decreased towards the shelf break.
2) Long-term accumulation patterns differed significantly from short-term patterns.
Potential reasons included: seasonal disconnect and differences in sediment transport
mechanisms. X-radiographs, mixed 2I0Pb, and 7Be indicated that the inner shelf and
mid-shelf up to ~50 m was an area of ephemeral, short-term deposition. Longer-term
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 43
accumulation rates peaked on the middle shelf and decreased towards the shelf break.
Sediment profiles from Transect D suggested that this part of the shelf is a conduit for
terrestrially-derived sediment transport off-shelf.
3) Physical processes controlled sediment structure on most of the shelf. Laminations
were as thick as 20 cm and were most frequent on the mid-shelf. Bioturbation
increased radially from the Waiapu River mouth, suggesting that riverine input (versus
wave energy) controlled sediment mixing patterns
91 n 4) The non-steady-state nature of the Pb signal in several cores on the mid- to outer
shelf versus the steady state cores found on the shelf break and slope suggests multiple
sediment delivery mechanisms influenced strata formation on the shelf including dilute
suspensions, gravity-driven flows, and water column resuspension. Pulsed event layers
exhibited a distinctly terrestrial signal versus the non-event sediments. The number of
pulsed event layers decreased from the mid-shelf to the outer-shelf. Steady-state
profiles are likely a product of bioturbation and decreased pulsed event sediment
delivery.
5) Grain-size parameters indicated that the terrestrial pulse event layers identified by
low excess 210Pb values and carbon data were potentially the product of multiple
sediment transport pathways: river-initiated hyperpycnal plume, wave-or current
supported gravity flows, and either of these two scenarios followed by a wave or
current resuspension event that produces slightly coarser deposits.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. References
Addington, L.D., Kuehl, S.A., and McNinch, J.E., in press. Distinguishing sediment
transport modes to the outer-shelf off the Waiapu River, New Zealand. Marine
Geology.
Aller, J.Y., and Stupakoff, I., 1996. The distribution and seasonal characteristics of
benthic communities on the Amazon shelf as indicators of physical processes.
Continental Shelf Research, 16: 717-751.
Appleby, P.G., Oldfield, F., 1992. Application of lead-210 to sedimentation studies. In:
Ivanovich, M., Harmon, R.S. (Eds.), Uranium Series Disequilibrium,
Applications to Earth, Marine and Environmental Sciences. Clarendon Press,
Oxford, pp. 731-778.
Biscaye, P.E., and Anderson, R.F., 1994. Fluxes of particulate matter on the slope of
the southern Middle Atlantic Bight: SEEP-II. Deep Sea Research Pat II: Topical
Studies in Oceanography, 41:459-509.
Boudreau, B.P., 1994. Is burial velocity a master parameter for bioturbation?
Geochimica et Cosmochimica Acta 58, 1243-1249.
Bruland, K.W., Koide, M., and Goldberg, E.D., 1974. The comparative marine
geochemistries of Lead 210 and Radium 226. Journal of Geophysical Research
79: 3083-3086.
Chiswell, S.M., 2000. The Wairarapa Coastal Current. New Zealand Journal of Marine
and Freshwater Research, 34: 303-315.
Chiswell, S.M., 2005. Mean and variability in the Wairarapa and Hikurangi Eddies,
New Zealand. New Zealand Journal of Marine and Freshwater Research,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 39:121-134.
Chiswell, S.M., 2003. Circulation within theWaiararapa Eddy, New Zealand. New
Zealand Journal of Marine and Freshwater Research, 37:691-704.
Collot, J-Y., Lewis, K., Lemarche, G., and Lallemand, S., 2001. The giant Ruatoria
debris avalanche on the northern Hikurangi margin, New Zealand: Result of
oblique seamount subduction. Journal of Geophysical Research, 106
(B9): 19271-19,297.
Crockett, J., and Nittrouer, C. A., 2004. The sandy inner shelf as a repository for muddy
sediment: an example from Northern California. Continental Shelf Research,
24: 55-73.
Cutshall, N.H., Larsen, I.L., Olsen, C.R., 1983. Direct analysis of 210Pb in sediment
samples: self-absorption corrections. Nuclear Instruments and Methods 206:
309-312.
Davis, J.A., 1984. Complexation of trace metals by adsorbed natural organic matter.
Geochimica et Cosmochimica Acta, 48: 679-692.
Dellapenna, T.M., Kuehl, S.A., and Schaffiier, L.C., 1998. Sea-bed mixing and particle
residence times in biologically and physically dominated estuarine systems: a
comparison of lower Chesapeake Bay and the York River Subestuary.
Estuarine, Coastal, and Shelf Science, 46:777-795.
Dukat, D.A., and Kuehl, S.A., 1995. Non-steady-state 210Pb flux and the use of
^ '1/' Ra/ Ra as a geochronometer on the Amazon continental shelf. Marine
Geology. 125: 329-350.
Eisma, D. and Kalf, J., 1984. Dispersal of Zaire River suspended matter in the estuary
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 46
and the Angola Basin. Netherlands Journal of Sea Research. 17: 385-411.
Farnsworth, K.L and J.A. Warrick, 2005. Sources of Fine Sediment to the California
Coast: Computation of a 'Mud Budget', Interim Report to the California
Sediment Management Working Group, California Dept, of Boating and
Waterways, 128 pages.
Farnsworth, K.L., and Milliman, J.D., 2003. Effects of climatic and anthropogenic
change on small mountainous rivers: the Salinas River example. Global and
Planetary Change, 39:53-64.
Field, B.D., Uruski, C.I., 1997. Cretaceous-Cenozoic Geology and Petroleum Systems
of the East Coast Region, New Zealand. Monograph, vol. 19. Institute of
Geological and Nuclear Sciences, Lower Hutt, NZ.
Flynn, W.W., 1968. The Determination of Low Levels of Polonium-210 in
Environmental Materials. Analytica Chimica Acta. 43:221-227.
Friedrichs C.T. and L.D. Wright, 2004.Gravity-driven sediment transport on the
continental shelf: implications for equilibrium profiles near river mouths.
Coastal Engineering, 51:795-811.
Gerald, L.E., and Kuehl., S.A., 2006. Recent sediment dispersal and preservation of
storm events on the continental shelf off the Waipaoa River, NZ. Ocean
Sciences Meeting, Honolulu, Hawaii. Abstract OS16A-29.
Goldberg, E.D., Koide, M., 1962. Geochronological studies of deep sea sediments by
the ionium/thorium method. Geochimica et Cosmochimica Acta 26, 417-450.
Griffiths, G. A.; Glasby, G. P., 1985. Input of river-derived sediment to the New
Zealand continental shelf. I. Mass. Estuarine, coastal and shelf science 21:773-
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 47
787.
Guinasso, N.L., and Schink, D.R., 1975. Quantitative estimates of biological mixing
rates in abyssal sediments. Journal of Geophysical Research 80, 3032-3043.
Harris, C.K., P.A. Traykovski, and R.W. Geyer. 2005. Flood dispersal and deposition
by near-bed gravitational flows and oceanographic transport: a numerical
modeling study of the Eel River shelf, northern California. Journal of
Geophysical Research, 110, C09025, doi:10.1029/2004JC002727.
Harris, C.K., P. A. Traykovski, and R.W. Geyer. 2004. Including a Near-Bed Turbid
Layer in a Three Dimensional Sediment Transport Model With Application to
the Eel River Shelf, Northern California. Proceedings of the Eighth Conference
on Estuarine and Coastal Modeling, ASCE, M. Spaulding (editor). 784-803.
Hicks, D.M., Gomez, B., and Trustrum, N.A., 2004. Event suspended sediment
characteristics and the generation of hyperpycnal plumes at river mouths: east
coast continental margin, North Island, New Zealand. The Journal of Geology,
112:471-485.
Hirschberg, D.J., Chin, P., Feng, H., and Cochran, J.K., 1996. Dynamics of sediment
and contaminant transport in the Hudson River Estuary: Evidence from
sediment distributions of naturally occurring radionuclides. Estuaries. 19:931-
949.
Jervey, M.R., 1988. Quantitative geological modeling of siliciclastic rock sequences
and their seismic expression. SEPM Spec. Publ. 42: 47-69.
Kineke, G.C. and Sternberg, R.W. 1995. Distribution of fluid muds on the Amazon
continental shelf. Marine Geology, 125: 193-233.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Kineke, G.C., Sternberg, R.W., Trowbridge, J.H., Geyer, W.R., 1996. Fluid mud
processes on the Amazon continental shelf. Continental Shelf Research 16,
667-696.
Kineke, G.C., Woolfe, K.J., Kuehl, S.A.; Milliman, J.D., Dellapenna, T.M., and
Purdon, R.G., 2000. Sediment export from the Sepik River, Papua New Guinea;
evidence for a divergent sediment plume. Continental Shelf Research, 20:2239-
2266.
Kniskem, T. A. and Kuehl, S.A., 2003. Spatial and Temporal Variability of Seabed
Disturbance in the York River Subestuary. Estuarine, Coastal and Shelf Science,
58 (1): 37-55.
Kuehl, S. A., Haim, T.M., and Moore, W.S., 1989. Shelf sedimentation off of the
Ganges-Brahmaputra river system: Evidence for sediment bypassing to the
Bengal Fan. Geology, 17: 1132-1135.
Kuehl, S.A., Nittrouer, C.A., Allison, M.A., Faria, L.E.C., Dukat, D.A., Jaeger, J.M.,
Pacioni, T.D., Figueiriedo, A.G., Underkoffler, E.C., 1996. Sediment
deposition, accumulation, and seabed dynamics in an energetic fine-grained
coastal environment. Continental Shelf Research 16: 787-815.
Kuehl, S.A., Pacioni, T.D., and Rine, J.M., 1995. Seabed dynamics of the inner
continental shelf: temporal and spatial variability of surficial strata. Marine
Geology, 125:283-302
Kuehl, S.A., G.J. Brunskill, K. Bumes, D. Fugate, T. Kniskem and L. Meneghini, 2004.
Nature of sediment dispersal off the Sepik River, Papua New Guinea:
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 49
preliminary sediment budget and implications for margin processes. Continental
Shelf Research, 24(19):2417-2429.
Leithold, E.L., and Blair, N.E., 2001. Watershed control on the carbon loading of
marine sedimentary particles. Geochimica et Cosmochimica Acta, 65:2231-
2240.
Leithold, E.L., Perkey, D.W., Blair, N.E., and Creamer, T.N., 2005. Sedimentation and
carbon burial on the northern California continental shelf: the signatures of
land-use change. Continental Shelf Research, 25:349-371.
Lewis, K.B., Collot, J-Y., and Lallemand, S.E., 1998. The damned Hikurangi Trough: a
channel-fed trench blocked by subducting seamounts and their wake
avalanches (New Zealand- France GeodyNZ Project). Basin Research, 10:
441-468.
Lewis, K.B., Lallemand, S.E., and Carter, L., 2004. Collapse in a Quaternary shelf
basin off East Cape, New Zealand: evidence for passage of a subducted
seamount inboard of the Ruatoria giant avalanche. New Zealand Journal of
Geology and Geophysics, 47:415-429.
Liu, J.T., Yuan, P.B., and Hung, J.-J., 1998. The coastal transition at the mouth of a
small mountainous river in Taiwan. Sedimentology, 45:803-816.
Ma, Y., Wright, L.D., and Friedrichs, C.T., in prep. Observations of sediment transport
on the continental shelf off the mouth of the Waiapu River, New Zealand:
evidence for current-supported gravity flows. To be submitted to Continental
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Shelf Research
Milliman, J. D. and Syvitski, 1992. Geomorphic/tectonic control of sediment
discharge to the ocean; the importance of small mountainous rivers. Journal of
Geology, Vol. 100, no. 5, p. 525-544.
Milliman, J.D., Farnsworth, K.L., and Albertin, C.S., 1999. Flux and fate of fluvial
sediments leaving large islands in the East Indies. Journal of Sea Research, 41:
97-107.
Milliman, J.D., Kao, S.-J., 2005. Hyperpycnal Discharge of Fluvial Sediment to the
Ocean: Impact of Super-Typhoon Herb (1996) on Taiwanese Rivers. The
Journal of Geology, 113:503-516
Moore, W.S. and Dymond, J.,1988. Correlation of Pb removal with organic carbon
fluxes in the Pacific Ocean. Nature, 331:339-341.
Moore, W.S., 1984. Radium isotope measurements using germanium detectors. Nuclear
Instruments and Methods in Physical Research, 223: 407-411.
Morehead, M.D., Syvitski, J.P., 1999. River-plume sedimentation modeling for
sequence stratigraphy; application to the Eel margin, Northern California.
Marine Geology, 154:29-41.
Mulder, T. and Syvitski, J. P. M. (1995) Turbidity currents generated at river mouths
during exceptional discharge to the world's oceans. Journal of Geology 103:
285-299.
Mullenbach, B.L., Nittrouer, C.S., Puig, P., and Orange, D.L., 2004. Sediment
deposition in a modem submarine canyon: Eel Canyon, northern California.
Marine Geology, 211:101-119.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Nittrouer, C.A., Sternberg, R.W., Carpenter, R., Bennett, J.T., 1979. The use of Pb-210
geochronology as a sedimentological tool: application to the Washington
continental shelf. Marine Geology 31:297-316.
Nittrouer, C.A. and Sternberg, R.W., 1981. The formation of sedimentary strata in an
allochthonous shelf environment: the Washington continental shelf. Marine
Geology, 31:297-316.
Nittrouer, C.A., DeMaster, D J., McKee, B.A., Cutshall, N.H., and Larsen, I.L.,
1983. The Effects of sediment mixing on 210Pb accumulation rates for the
Washington continental shelf. Marine Geolology, 54:201-221.
Nittrouer, C.A., DeMaster, D.J., Kuehl, S.A., McKee, Thorbjamarson, K.W., 1985.
Some questions and answers about accumulation of fine-grained sediment in
continental margin environments. Geo-Marine Letters, 4: 211-213.
Nittrouer, C.A. and L.D. Wright, 1994. Transport of particles across continental
shelves. Reviews of Geophysics, 32: 85-113.
Nittrouer, C.A., S.A. Kuehl, R.W. Sternberg, A.G. Figueiredo and L.E.C. Faria, 1995.
An introduction to the geological significance of sediment transport and
accumulation on the Amazon continental shelf. Marine Geology, 125: 177-192.
Ogston, A.S., D.A. Cacchione, D.A., R.W. Sternberg and G.C. Kineke, 2000.
Observations of storm and river flood-driven sediment transport on the northern
California continental shelf. Continental Shelf Research, 20: 2141-2162.
Olsen, C.R., Simpson, H.J., Peng, T.H., Bopp, R.F., and Trier, R.M., 1981. Sediment
mixing and accumulation rate effects on radionuclide depth profiles in Hudson
Estuary sediments. Journal of Geophysical Research, 86:11020-11028.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 52
Orange, D.L., 1999. Tectonics, sedimentation and erosion in northern California:
submarine geomorphology and sediment preservation potential as a result of
three competing processes. Marine Geology, 154: 369-382.
Orpin, A.R., Alexander, C., Carter, L., Kuehl, S., and Walsh, J.P., 2006. Temporal and
spatial complexity in post-glacial sedimentation on the tectonically active,
Poverty Bay continental margin of New Zealand. Continental Shelf Research,
26: 2205-2224.
Radakovitch, O., and Heussner, S., 1995. Fluxes and budget of 210Pb on the
continental margin of the Bay of Biscay (northeastern Atlantic) Deep-Sea
Research II, 46:2175-2203.
Sanford M.W., Kuehl, S.A., and Nittrouer, C.A., 1990. Modem sedimentary processes
in the Wilmington Canyon area, US east coast. Marine Geology, 92:205-226.
Schaffner, L.C., Diaz, R.J., Olsen, C.R., and Larsen, I.L., 1987. Faunal characteristics
and sediment accumulation processes in the James River Estuary, Virginia.
Estuarine, Coastal, and Shelf Science, 25: 211-226.
Scully, M.E., C.T. Friedrichs and L.D. Wright, 2003. Numerical modeling results of
gravity-driven sediment transport and deposition on an energetic continental
shelf: Eel River, Northern California. Journal of Geophysical Research, (C4):
17-4-17-14.
Sommerfield, C.K. and C.A. Nittrouer, 1999. Modem accumulation rates and a
sediment budget for the Eel shelf: a flood-dominated depositional environment.
Marine Geology, 154: 227-242.
Sommerfield, C.K., Nittrouer, C.A., and Alexander, C.R., 1999. 7Be as a tracer of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 53
flood sedimentation on the northern California continental margin. Continental
Shelf Research, 19:335-361.
Srisuksawad, K., Pomtepkasemsan, B., Nouchpramool, S., Yamkate, P., Carpenter, R.,
Peterson, M.L., Hamilton, T., 1997. Radionuclide activities, geochemistry, and
accumulation rates in the Gulf of Thailand. Continental Shelf Research, 17:
925-965.
Traykovski, P., W.R. Geyer, J.D. Irish and J.F. Lynch, 2000. The role of wave-induced
density-driven fluid mud flows for cross-shelf transport on the Eel River
continental shelf. Continental Shelf Research, 20: 2113-2140.
Wadman, H.M. and McNinch, J.E., in review. Spatial variation on the inner shelf of a
high yield river, Waiapu River, New Zealand: Implications for fine sediment
dispersal and preservation. Continental Shelf Research.
Walcott, R.I., 1987. Geodetic strain and the deformational history of the North Island of
New Zealand during the late Cainozoic. Philosophical Transactions Royal
Society of London, A321:163-181.
Walsh, J.P. and Nittrouer, C.A., 2003. Contrasting styles of off-shelf sediment
accumulation in New Guinea. Marine Geology 196:105-125.
Wheatcroft, R.A., 1990. Preservation potential of sedimentary event layers. Geology
18, 843-845.
Wheatcroft, R.A., and Drake, D.E., 2003. Post-depositional alteration and preservation
of sedimentary event layers on continental margins,I. The role of episodic
sedimentation Marine Geology 199: 123-137.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Wheatcroft, R.A., Sommerfield, C. K., 2005. River sediment flux and shelf sediment
accumulation rates on the Pacific Northwest margin. Continental Shelf
Research, 25:311:332.
Wright, L.D., Wiseman, W.J., Prior, D.B., Suhayda, J.N., Keller, G.H., Yang, Z.-S. and
Fan, Y.B., 1988. Marine dispersal and deposition of Yellow River silts by
gravity driven underflows. Nature, 332: 629-632.
Wright, L.D., Wiseman, W.J., Jr., Yang, Z. S., Bomhold, B.D., Keller, G.H., Prior,
D.B., Suhayda, J.N., 1990. Processes of marine dispersal and deposition of
suspended silts off the modem mouth of the Huanghe (Yellow River).
Continental Shelf Research, 10:1-40.
Wright, L.D. and C.A. Nittrouer, 1995. Dispersal of river sediments in coastal seas: six
contrasting cases. Estuaries, 18: 494-508.
Wright, L D; Friedrichs, C T; Kim, S C; Scully, M E, 2001. Effects of ambient currents
and waves on gravity-driven sediment transport on continental shelves. Marine
Geology, 175:25-45.
Wright, L.D., C.T. Friedrichs and M.E. Scully, 2002. Pulsational gravity-driven
sediment transport on two energetic shelves. Continental Shelf Research,
22:2443-2460.
Wright, L.D., Ma ,Y., Scully, M, Friedrichs, C.T., 2006. Observations of Across Shelf
Sediment Transport During High Energy Flood Events off the Mouth of the
Waiapu River, New Zealand Eos Trans. AGU, 87(36), Ocean Sci. Meet. Suppl.,
Abstract OS11L-03.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 55
Kasten Accum. Event Accum. Maximum Core Rate (cm/y) Thickness Rate Depth of (Sommerfield (cm) (cm/y) Pbxs and Nittrouer, (Normalized to 1999) event thickness) 1 0.8 0 0.8 85 2 3.3 62 1.8 135 3 N/A 81.5 N/A 250 4 0.9 14.5 0.8 80 5 1.1 15 1 130 6 N/A 0 N/A 20 7 N/A 0 N/A 25 8 N/A 0 N/A 25 9 1.2 0 1.2 130 10 1.8 32 1.4 165 11 3.5 2 3.3 100 12 1.8 43.5 1 105 13 1.6 15 1.4 110 14 N/A N/A N/A N/A 15 1.2 0 1.2 90 16 2.6 37.5 1.8 165 17 N/A 0 N/A 0 18 2.4 47.5 1.9 150 19 1.8 10 1.5 150 20 1.1 19 0.9 105 21 2.1 17 1.9 170 22 1.1 0 1.1 95 23 0.3 0 0.3 25 25 N/A 0 0 15 26 N/A 36 0 120 27 1 5.5 0.9 95 28 N/A 33 0 130 29 0.9 3.5 0.8 75 30 N/A 41 0 110 31 0.5 0 0.5 65 32 0.2 0 0.2 25 33 1 18 0.9 125 34 0.6 0 0.6 75 36 1.7 21 1.3 109 37 2.2 27.5 1.7 114 ation rates wereTable 1. Kasten core statistics. Non-steady state cores in bold. Accumul ation rates wereTable calculated using the approach outlined by Sommerfield and Nittrouer (1999) and the background accumulation equation outlined in this paper. Event layer thickness, in the third column, is the total thickness of event layers in each core. The last column shows o 1 n the maximum depth of excess Pb.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 56
Figure Captions
Figure 1. The Waiapu River is located on the North Island of New Zealand. Thirty-five
Kasten cores were collected from the shelf between 30 - 700 m water depth during an
August, 2003 cruise aboard NIWA’s RV Tangaroa (black circles). Box cores collected
during the Kilo Moana cruise in May, 2004 reoccupied the August sites and extended
coverage (black triangles). Only the cores collected during the August cruise were
analyzed for excess Pb; Be activities were measured for both cruises. Sandy
sediments collected to the north of East Cape identified by black squares.
Figure 2. The top panel shows the rating curve produced by Hicks et al. (2004).A
discharge record for the last several years (courtesy of Dave Peacock at the GDC)
shows that sediment concentrations have exceeded this threshold several times.
Figure 3. Average surface grain-size for the two cruises. A belt of muddy sediment
dominates the shelf (shaded area).
Figure 4. X-radiographs from Transect D illustrate the change in sediment fabric across
the shelf from physically dominated to predominantly bioturbated.
Figure 5. Excess 210Pb profiles for the Waiapu River shelf (excluding KC 6 , 25, 27, 36,
and 4). Figure position reflects map orientation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 57
91 ft Figure 6 . Apparent excess Pb accumulation rates. Rates were not calculated for
short cores with surface mixed layers (circles) or cores dominated by pulsed event
layers such that an accumulation rate could not be modeled (squares). Rates were
calculated using the approach by Sommerfield and Nittrouer (1999).
Figure 7. Pulsed event layer isopach map. The layers were identified using excess 210Pb
profiles and x-radiographs.
Figure 8. A facies map is produced by combining accumulation patterns and
sedimentary structures revealed by x-radiographs.
Figure 9. Significant Be activities were generally observed in sediments collected at
depths shallower than 80 m. Quantitative data from 2004 are shown. Data from 2003
was qualitative, so the shaded area represents where Be was found for the two cruises.
Figure 10. Profiles of atomic C/N and 8 13C versus excess 210Pb for KC18. Each sample
that has a low excess Pb activity corresponds with a relatively terrestrial signal in the 9i n carbon data. The exception is at 159 cm, where the excess Pb activity was not
assessed for that sample.
1 3 Figure 11. A plot of 8 C versus N/C ratio indicates that the sediments on the shelf
preserve a terrestrial signal. The pulsed event layers (white triangles) display a
significantly more terrestrial signal than non-event layer samples from KC18 (black
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 58
triangles). The pulsed event layers also contain more vascular plant material than the
rest of the core, as observed during core cutting.
Figure 12. A. Plot of excess 210Pb versus sand, silt, and clay fractions for KC18. Red
triangles represent samples from pulsed event layers. B. Excess 210Pb activities are
decay corrected and normalized to mean grain size and clay fraction. The mean grain-
size profile is shown to the right for comparison.
Figure 13. Plots of carbon percent, del 13C, and C/N ratio versus excess 210Pb
activities for KC18. Red triangles represent samples from pulsed event layers.
Figure 14. The fine sediment budget over the last 100 years was estimated using
apparent 210Pb accumulation rates and observations of the extent of the Waiapu Shelf
Basin by Lewis et al. (2004). Sediments were assumed to have a porosity of 60%.
About 24% of the fluvial load has been retained over the last 100 years.
Figure 15. The fraction finer than 20 microns is plotted relative to mean grain-size for
KC18. The corresponding excess 210Pb activities are plotted for easy identification of
the pulsed event layers. This approach has been used by Wheatcroft and Drake (2003)
to identify the relevant sediment transport mechanism. When the samples plot lower
than 2 standard deviations below the mean (outlined by the box), the sediments were
likely deposited by a wave- or current-supported gravity flow. If sediments are finer
than the mean, then the sediments were deposited by a hyperpycnal plume. According
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. to this scheme, at least two layers were deposited by gravity flows, but the majority of
layers were deposited by negatively buoyant plumes. It is likely that the importance of
these two mechanisms vary across the shelf.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced
178°30'00"EI 178°40'00"E I 178°50'00"E I 60 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced
100000 (T/ 8 i u 3 ) ® o o o o PQ (q/Sui) uopBJiuaauoa juaimpas [BiAny juaimpas uopBJiuaauoa (q/Sui) rr> O' J*5 B 59 ® © Tf «s ® ON v> ® CJ O N G < 2 5 VO JS 00 © 00 ® 62 s.aofoie S.ftOSoLZ S ,a 0 o 8 € i 1_ w p o bJ) 0 fed Pe Orr opv 00 178°3 s.ftofro^e s.aoso^e StQ;0o8£ Fig. 3 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 63 sJ& 8 *2S U L. g S3 | 3 * 9 § Fig. 4 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 64 KC30 150 - 180 —» KC11 KC12 KC31 10 0.1 1 10 0.1 0.1 10 0.110 1 i-i.mm) o i i i unit i i 111iJl|| 120 - 1 5 0 - KC7 KC8 KC10 KC21 KC9 KC32 c. 1 10 0.1 1 10 0.1 1 10 0.1 1 10 0.1 1 10 0.1 1 10 3 0 - 3 0 - i - j f H 3 0 - 3 0 - yv 3° - 6 0 - 6 0 - 6 0 - / jN 6 0 - J 'l 6 0 - Jh 6 0 - 9 0 - 9 0 - 9 0 - 9 0 - 9 0 - W 90 “ 1 2 0 - 1 2 0 - 1 2 0 - hT 120p 1 2 0 - 120 " 1 5 0 - 1 5 0 - 1 5 0 - ,Jzp 1 5 0 - , / 150- 150- 180 180 J 180 J 18flM 180 J 180 J KC28 KC18 KC19 KC33 KC20 KC34 D. 0.1 1 10 0.1 1 10 0.1 1 10 0.1 1 10 0.1 1 10 0.1 1 10 ■tiuiml luu I 0 — 3 0 -3 0 - 3 0 - 3 0 - 3 0 - 3 0 - 6 0 - 90 - 9 0 - 9 0 - 9 0 - 9 0 - 120- 120 - 1 2 0 - 1120 — 1 20 - 1 2 0 - 1 5 0 - 1 5 0 - 1 5 0 - 1 5 0 - 1 5 0 - 1 5 0 - 1 8 0 - 1 8 0 - 1 8 0 - 1 8 0 - 1 8 0 - KC37 KC23 0.1 1 10 0.1 10 0.1 10 0.1 1 Excess 210Pb Activity (dpm/g) 1 8 0 - oa I 1 sample error 3 KC2o KC16 oa , 1 10 0.1 10 0.1 10 0.1 Q 1 2 0 - 1 5 0 - 1 5 0 - 1 8 0 - 1 8 0 - Fig. 5 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ON <_n 178°50’00'E cm/y cm/y 2 . 0 1 . 0 - - 1 . 0 2.0 - 3.5 cm/y 0 high event preservation low mud accumulation Excess 210Pb Accumulation Excess 210Pb 178 178 40’00'E 178 50'00"E 178°40'00"E _L 178°30’00"E 178°30'00"E A Kilometers 0 2.5 5 10 o\ 51 crq Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Event layer thickness (cm) 45 - 81 cm ■|* 25-45 cm A 1- 25 cm % no 210Pb event layers Kilometers 0 2.5 5 10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced 178°30’0"E 178°40T)"E 178°50T)"E s„aot' if O 'Z £SuOtO£o££ SiiftOt'oZ.£ 0 z.£ — S»»(X09o££ I l M Dl Q_ ca o ra ro S„a0o8£ Su(X0o8£ u Fig. 8 Fig. 67 178°30X)"E 178°40T)"E 178°50T)"E 68 S.^ot’o I t SuftOSo Lt S»ft0o8£ 178°30T)”E 178°30T)”E 178°40X)’’E 178°50X)"E S.&OPoLt S«aOSoAS S„ft0 o8 £ Fig. 9 Reproduced with permission of the copyright ownerFnrthp ■ r reproduction prohibited without permission. 69 Excess 210pb Activity (dpm/g) ^ I ' I 1 I ' I 9 10 11 12 13 14 A(C/N)a 0 .1Excess 210pb Activity1 j (dpm/g) v f &/ 1Q 5 0 - -a I 100 'w ' £ S” 150 200 250 1 I 1 I 1 I T I 1 I -26 -25 -24 -23 -22 -21 ▲ del 13C (PDB) Fig. 10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced (N/C) 0.025 - 0.075 0.125- 05- 5 .0 0 0.15 0.1 -21 I I I I | I I I I I I I I | I I I I | I I I I | I I I I event layer A A After Leitholdand Hope, 1999. MarinePlankton -22 non-event sediments -23 del 13C del (PDB) -24 -25 A A I I I I | I I I I | I I I I 4 ^ A a a Vascular Plant Vascular SoilOrganic Matter Debris -26 i. 11 Fig. -27 70 71 8 — I o> E Q. V >> o ~ ® •-** 4 O (0 “ -Q o ° ® w (0 oo> x ® R-sauared = 0.49069 1 I 1 I 1 I 1 I 1 I 0 20 40 60 80 100 silt % TO aE ■o 6 — T 3 ® > O ~ ® > Sm 4dl O re «2 - 8 ° ■8 ww (0 oa> x ® 20 50 60 70 clay % TO A £ a T 3 A T 5 ® > O ~ ® 2. 4 — i- & O re A A «2 - a)o ? ■8 " (A (0 ® O X CD -2 1 1 I 1 2 4 sand % Fig. 12A Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced Depth (cm) 5 - 150 0.1 ——-V V— decay corrected excess 210Pbdecay corrected excess (dpm/g) L-L.L-LLI 210Pb mean grain-size mean clay% 1 L-L-LLU I 10 □ 120 5 Mean Grain size (phi) size GrainMean I 6 I I I 7 I I I i. 12B Fig. 8 I I I 72 9 I 73 0.8 2 0.7 g> "53 3 c 0.6 X!o nj “ 0.5 c a> H < D q . 0.4 0.3 ■ n ■ ■ 'I- 1 i 1 i 1 i -2 0 2 4 6 8 -24 -24.4 — O -24.8 — co ■° -25.2 normal samples 210Pb spike samples -25.6 -26 ■2 0 2 4 6 8 14 —i 13 — 12 — Z o 10 — ■2 0 2 4 6 8 decay corrected excess 210Pb activity (dpm/g) F ig.13 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced 178°30’3"E 178°40’3"E 178°50’3"E : ^ 1 , u ^ S I Su0t09o8€ SuO^OSoIC i. 14 Fig. 178°30,3"E 178°40’3"E 178°50,3ME 74 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced Depth (cm) 8—- 180— ISO— 120 90 — 90 60 — 60 30 — 30 0 - — - 0.1 1 .6 .2 .8 .4 . 07 07 06 06 0.6 0.64 0.68 0.72 0.76 0.8 0.84 0.88 0.92 0.96 Decay corrected excess 210Pb(dpm/g) correctedactivities excess Decay % <20 microns 210Pb % < 20 microns 20 %< i. 15 Fig. 5 7 7 6 Chapter 2 Flood dispersal and deposition on the Waiapu River shelf, New Zealand. I. Abstract The Waiapu River has a small, mountainous catchment, which delivers an estimated 35 MT/y of suspended load to a narrow collision margin shelf subject to energetic waves and currents. Measurements made by tripods during a winter storm season show that the shelf adjacent to the Waiapu River was subjected to wave heights up to 5 m and depth-averaged, along-shelf currents above 80 cm/s. Waiapu River discharge and suspended sediment records indicate that 24 MT of terrestrial material were delivered to the coastal ocean from May - July, 2004. Several floods that occurred during this period had suspended sediment concentrations that exceeded the ~40 g/1 threshold for gravity-driven transport. Seabed accumulation rates, which averaged 1.5 + 0.9 cm/y and peaked on the mid- to outer shelf, indicated that the shelf retained about 24% of the fluvial budget over the last 100 years. Geochronology and carbon data further suggested that multiple transport mechanisms, including dilute suspensions and gravity-driven transport, affect the sediment record. These observations do not identify mechanisms by which approximately 76% of the fluvial load escapes the shelf, nor do they discern the relative importance of the various sediment transport pathways. To address these questions, a three-dimensional numerical Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 7 model was used to simulate fine sediment transport and deposition on the Waiapu River shelf. The numerical model accounted for hydrodynamics driven by winds and buoyancy, sediment resuspension by energetic waves and currents, and gravitationally- driven transport within a thin, near-bed fluid mud layer. The simulated currents match the observed time-averaged currents. However, the model fails to represent the large velocity fluctuations seen in the tripod data, perhaps because it neglects large-scale shelf waves that may be responsible for energetic currents. Calculated sediment concentrations were commensurate with observations that reached 3 g/L at 1 m above the bed. Simulated bed shear stresses agree with observed values during most of the record. The model results show that depositional patterns on the Waiapu River shelf are influenced by the settling speeds of sediment aggregates, the fluvial discharge, relative importance of suspended and gravitationally-driven transport, and wave properties. Particles with lower settling velocities tend to be widely dispersed and removed from the shelf. Sediments transported across the shelf within a near-bed fluid mud layer are largely composed of particles with settling velocities representative of flocculated material. A significant portion of material with higher settling velocities is also resuspended and exported during times of energetic waves and wind-driven currents. Generally, gravity-driven transport accounts for about one-third of the total flux, though sediment flux within the wave boundary layer can exceed that of dilute suspension during floods. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 78 Sediment depositional patterns vary across the shelf, with muddy sediments ephemerally deposited for up to several weeks on the inner shelf (0 - 30 m) and mid shelf (30 -70 m). The model estimates that an average 21 % of the fine sediments delivered to the shelf from May- July, 2004 are retained on the outer shelf (70 - 130 m depth). This compares favorably with seabed geochemical observations of retention of -24% (Kniskem, Chapter 1). II. Introduction Small river systems with mountainous, highly erodible catchments deliver a significant portion of the global terrestrial budget to the coastal ocean and are typically found along collision margins (Milliman and Syvitski, 1992). Most of the annual yield of these rivers is dispersed to shelf waters during floods (Milliman and Syvitski, 1992; Farnsworth and Milliman, 2003). Flood sediments are often delivered to shelves within buoyant, or hypopycnal, plumes, but negatively buoyant, or hyperpycnal, plumes may form at river mouths when fluvial sediment concentrations exceed —40g/l (Mulder and Syvitski, 1995). Sediment resuspension by energetic waves and currents has long been believed to dominate transport on shelves (Sternberg and Larsen, 1976; Drake and Cacchione, 1985; Grant and Madsen, 1986). Recent studies have shown, however, that gravity flows may be important on some river-shelf systems (Traykovski et al., 2000; Harris et ah, 2005; Friedrichs and Wright, 2004). When waves and/or currents are energetic or the shelf slope is sufficient to facilitate auto-suspension, suspended sediments may be concentrated sufficiently to form a near-bed mud layer (Friedrichs and Wright, 2004). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 9 The pressure gradients formed by this fluid mud may be large enough to cause down- slope flux (Friedrichs and Wright, 2004). The relative importance of dilute suspension, resuspension, and gravity-driven transport can influence the exchange and burial of important contaminants and nutrients, benthic community structure, and sedimentary strata formation -Kineke et al, 2000; Leithold and Blair, 2001; Wheatcroft and Drake, 2003; Kuehl et al., 2004; Goni et al., 2006). Quantifying these combinations of sediment transport, resuspension, and burial may contribute to our understanding of sedimentary stratigraphic formation and reworking on continental shelves. Several mechanisms can create turbid flows whose density is large enough to initiate significant gravitational forces. Other than hyperpycnal plume formation at the river mouth due to high fluvial sediment concentrations, energetic waves or currents, estuarine-like circulation on the shelf, earthquakes resulting in mass failures, or, in general, mechanisms that result in the convergence of sediments on the shelf may promote gravity-driven transport (Kineke et al., 1996; Mulder and Syvitski, 1995; Friedrichs and Wright, 2004). Once initiated, gravity-driven transport requires either a relatively steep slope or energetic waves and currents to sustain transport to deeper waters (Scully et al., 2003; Friedrichs and Wright, 2004). Since most continental shelves exhibit gradients below the critical slope of 0.01 required for auto-suspension, energetic waves and currents are important mechanisms for sediment transport within a near-bed fluid mud layer (Friedrichs and Wright, 2004). As a result, a mud-belt commonly develops on the mid-shelf where the influences of waves is attenuated (Traykovski et al., 2000; Friedrichs and Wright, 2004; Harris et al., 2005). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Gravity-driven sediment transport has been observed on continental shelves adjacent to many rivers including: the Amazon (Kineke and Sternberg, 1995), Eel (Ogston et al., 2000 and Traykovski, et al., 2000), Fly (Walsh and Nittrouer, 2003; Walsh et al., 2004), Sepik (Kuehl et al., 2004, Kineke et al., 2000), and Yellow Rivers (Wright et al., 1998). The combination of transport pathways for each river and shelf system produces diverse accumulation patterns. The Amazon River, with an annual load of 1100-1300 MT/y (Milliman and Meade, 1983), delivers sediment to a wide shelf system via a buoyant river plume (Curtin and Legeckis, 1986; Kineke and Sternberg, 1995). Once deposited, the sediments are resuspended by strong tidal currents and often form a fluid mud layer up to 1 m thick (Kineke et al., 1996). The Fly River and other rivers in the Gulf of Papua deliver 300 MT of sediment to the adjacent shelf area annually (Walsh and Nittrouer, 2003). The sediments are temporarily stored on the inner shelf and are eventually transported to the mid-shelf within fluid-mud flows induced by sediment flux convergence (Walsh and Nittrouer, 2003; Walsh et al, 2004). Very little of the Gulf of Papua sediment budget is transported off the shelf (Walsh and Nittrouer, 2003). Two main sediment transport pathways were observed on the Sepik River shelf. Convergence at the mouth of the Sepik River forms a hyperpycnal plume and diverts the majority of the river’s sediment budget into a nearby submarine canyon, acting as a conduit for terrestrially-derived sediments to deep water (Harris et al., 2005). Only a fraction of the Sepik River sediment load, 7-15% of 85 MT/y, is dispersed on the adjacent narrow shelf (Kuehl et al., 2004; Kineke et al., 1996). Although a submarine canyon is located just south of the Eel River mouth, local circulation patterns deliver only an estimated 12% of the 15 MT/y fluvial load to the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. canyon (Mullenbach et al., 2004). Fine sediments are primarily dispersed to the north of the river mouth, with the inner shelf and the mid-shelf retaining 6-13% and -20%, respectively (Sommerfield and Nittrouer, 1999; Crockett and Nittrouer, 2003, Harris et al., 2005). An estimated 40-50% of the fluvial budget is apparently transported to the north of the proximal shelf (Harris et al., 2005). Flood simulations indicate that fluvial sediments are initially deposited on the Eel River inner shelf, but are quickly resuspended by waves and transported to the mid-shelf within a near bed fluidized mud layer (Scully et al., 2003; Harris et al., 2005). Observational data from the Eel River Shelf suggested that 80% of across-shelf flux was accounted for by gravity-driven transport within the wave boundary layer (Traykovski et al., 2000). Deposits associated with wave-initiated gravity-driven transport have been identified in the seabed record on the mid-shelf and within the canyon on the Eel River Shelf (Sommerfield and Nittrouer, 1999; Mullenbach et al., 2004). Variable fluvial input, hydrodynamic regime, and accommodation space influence sediment transport on each of these continental shelf systems. An understanding of the complex controls of these sediment transport pathways is requisite to characterizing nutrient cycling, sedimentary structure, and strata formation on margins. This study focuses on the continental shelf adjacent to the Waiapu River, New Zealand, an area that experiences energetic waves, strong currents, and relatively frequent floods (Hicks et al., 2004; Wright et al., 2006), (Figs. 1 and 2). It is an ideal setting within which to investigate depositional products of multiple sediment transport mechanisms. Fluvial delivery is largely episodic with little discharge between floods (Fig. 2). Furthermore, flood sediment concentrations are expected to exceed the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 2 threshold for gravity-driven transport several times per year (Hicks et al., 2004). A field study conducted in 2004 showed that both waves and currents on the mid-shelf were energetic, with near-bed current speeds up to 60 cm/s and storm waves exceeding 5m (Wright et al., 2006). Currents on the shelf were sufficient to support gravity-flows several meters thick within the wave-current boundary layer (Ma et al., in prep.). Seabed observations from the Waiapu shelf revealed high apparent accumulation rates (average: 1.5 + 0.9 ; range 0.2 -3.2 cm/y) of fine sediments on the mid- to outer-shelf, with an estimated 24% of the fluvial budget retained on the shelf over the last 100 years (Kniskem, Chapter 1). The estimated budget was calculated for fine sediments, assuming that sands and coarser sediments account for only 2-5% of the fluvial load (Griffiths and Glasby, 1985). Massive sandy deposits and interbedded muds and sands on the inner shelf may account for between 8-15% of the Waiapu River sediment budget (Wadman and McNinch, 2006; Wadman, pers. comm.), leaving between 58-64% of the budget unaccounted. Analysis of the muddy deposits indicated that flood sediments are rapidly transported across the shelf and to the shelf break and form event beds that can be distinguished from non-flood or reworked sediments (Addington et al., in press, Kniskem, Chapter 1). These observations, however, do not identify the mechanisms by which sediment escapes the shelf, nor do they discern the relative importance of various sediment transport mechanisms. To address these questions, a three-dimensional numerical model was used to simulate fine sediment transport and deposition on the Waiapu River shelf. Although one- and two-dimensional models are capable of simulating sediment transport on shelves, and can account for gravity-driven flow (Wiberg and Smith, 1983; Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Wiberg et al., 1994; Zhang et al., 1999; Scully et al., 2002; Fan et al., 2004), they cannot adequately represent the spatial flux patterns inherent in river plume dispersal systems. Many of these 2D models attempted to circumvent this issue by beginning with sediment already deposited on the seabed and available for resuspension by waves (Harris and Wiberg, 2001; Scully et al., 2002). Only a few three-dimensional models have attempted to account for both dilute suspensions and gravity-driven transport. Sediment was included in plume density calculations in a modified version of the Princeton Ocean Model for the Tseng-wen River, Taiwan (Liu et al., 2002). However, the maximum sediment concentration used was 1000 mg/1, far lower than values observed on the Tseng-wen River shelf, and far lower than those observed in the Eel River or the Waiapu River (Syvitski and Morehead, 1999; Liu et al., 1998, 2000; Hicks et al., 2004). The sediment concentration values used were also lower than needed for river-initiated hyperpycnal transport (Mulder and Syvitski, 1995; Liu et al., 2002). Another approach included transport of both suspended sediment and material within a near-bed fluid layer that scaled in thickness with the wave boundary layer (Harris et al., 2004, 2005). This model parameterized sediment entrainment from the near-bed fluid mud layer to the overlying water column as a function of the gradient Richardson number and turbulence above the wave boundary layer (Harris et al., 2004). They did not resolve suspended sediment profiles or velocities within the fluid mud layer, but instead assumed that it was as thick as the wave boundary layer, and fairly well-mixed. This is the only three-dimensional model that I know of that simulates wave-initiated gravity-driven transport and allows sediment exchange between the wave boundary layer and the overlying water column (Harris et al., 2004). This modified Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 4 version of ECOM-SED (the Estuarine Coastal Ocean Model -SEDiment; see Blumberg and Mellor, 1987; HydroQual, 2002) accounts for hydrodynamics driven by winds and buoyancy, sediment resuspension by energetic waves and currents, and transport within a thin, near-bed fluid mud layer (Harris et al., 2004). Here, it is used to investigate sediment transport and deposition on the Waiapu River shelf during May - July, 2004, a stormy period that coincided with a field experiment (Fig. 2). III. Objectives: The simulation of sediment delivery and dispersal on the Waiapu River Shelf was used to address the following objectives: 1. Investigate the temporal and spatial variability of sediment transport in dilute suspension and in the wave boundary layer on the Waiapu River shelf. 2. Investigate the transport mechanisms and pathways whereby sediments escape the proximal shelf. 3. Investigate the parameters that influenced sediment dispersal and deposition on the Waiapu River shelf during the 2004 field season. IV. Background The Waiapu River and shelf system and the environmental conditions during the 2004 field experiment are described in this section. A. The river and shelf environment Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Although the Waiapu River in New Zealand is small, with a catchment area of only 1700 km , it has one of the highest sediment yields in the world at 17,800 -22,520 tons/km2/yr (Walling and Webb, 1996; Hicks et al., 2000, 2004) (Fig. 1). Its annual suspended sediment load of 35 MT (Hicks et al., 2000; 2004) is comparable to the estimated fine-grained load of the state of California, which is an estimated 34 MT/y (Farnsworth and Warrick, 2005). Observed Waiapu River suspended sediment concentrations during floods range from 42 mg/1 to over 60, 000 mg/1, with an estimated discharge weighted average sediment concentration of 10,000 mg/L, corresponding to freshwater discharge of 86 m /sec (Hicks et al., 2000; 2004). Little is known about the grain-size composition of the sediments dispersed by the river. Griffiths and Glasby (1985) indicate from their limited data that bed-load generally comprises 2-5% of a New Zealand river's total yield, but no estimate is provided about the relative amounts of sands, silts, and clays. Similar to other small rivers with mountainous catchments, floods dominate Waiapu River sediment delivery. During the annual wet season, from June through September, sediment concentrations are expected to exceed the threshold for gravity- driven transport during high discharge events at least once per year (Mulder and Syvitski, 1995; Hicks et al., 2004; National Institute of Water and Atmospheric, NIWA, unpublished data). Cyclones also produce high fluvial suspended sediment concentrations capable of forming hyperpycnal plumes and have a recurrence interval of 2.6 years (Hicks, 1995; Hicks et al., 2004, Kasai et al., 2005). The largest such cyclone in the historical record was Cyclone Bola, in March, 1988. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A combination of hydrodynamic forcings influences sediment dispersal patterns on the Waiapu River shelf. Along the shelf break, the East Cape Current (ECC) flows south-ward along the eastern margin of the North Island with an estimated volume transport of 10-25 Sv and current speeds up to 25 cm/s at 100 m depth (Chiswell, 2000, Chiswell and Roemmich, 1998). Low temperature and lower salinity waters are found inshore of the ECC, where the north-ward flowing Wairarapa Coastal Current (WCC), has been identified (Chiswell, 2000). The WCC is significantly smaller than the ECC, with a volume transport of only 1.6 Sv, and mean surface currents at ~20 cm/s (Chiswell, 2000). Temporally and spatially variable eddies, the East Cape Eddy and the Wairarapa Eddy, form off the East Cape coast ephemerally influencing shelf circulation patterns (Chiswell, 2002; Chiswell 2005). A wave and current record collected during a field experiment in 2004 indicates that the mid-continental shelf (40-60 m depth) may be subjected to very strong currents, up to 100 cm/s, and waves that often exceeded 5 m (Wright et al., 2006) (Fig. 2). The observed wave and current conditions during the deployment period in 2004 indicated that both dilute suspensions and gravity-driven transport may be important and that bed shear stresses exceeded the critical shear stress for sediment transport at water depths up to 100 m (Ma et al., in prep.). Both waves and currents were energetic enough to resuspend sediments into dilute suspension and to support near-bed (< 1 m above the bed) fluid mud layers (Wright et al., 2006). Strong across-shelf currents, in conjunction with sediment concentrations on the order of 3 g/L at 1 m above the bed, support this interpretation (Wright et al., 2006). Seabed geochronology, carbon, and textural data also implied that multiple sediment transport mechanisms influence Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 7 sedimentary structure and that flood events are preserved in the sediment record over 100 year time-scales (Kniskem, Chapter 1). The Waiapu River shelf is relatively narrow (~20 km), and provides limited accommodation space. The inner shelf may act as a repository for modem inter-bedded sands and clays, accounting for between 12-18% of the fluvial budget over the last few hundred years (Wadman and McNinch, 2006). High excess Pb910 accumulation rates suggest that 25% of the fluvial budget over the last 100 years is retained on the mid outer shelf (Kniskem, Chapter 1). The shelf break is defined by the cuspate Ruatoria reentrant (Fig. 1), a scar created by a giant submarine landslide triggered by earthquakes associated with subducting seamounts and subsequent debris flows (Collot et al., 2001). Retention of sediments in this area may be influenced by smaller-scale bathymetry. Small basins on the edge of the Ruatoria reentrant at water depth of ~200m trap an estimated 3% of the fluvial budget with accumulation rates ranging between 0.6 to 1.8 cm/yr (Addington et al., in press). B. The deployment period During a field experiment, the Virginia Institute of Marine Science (VIMS) deployed two tripods directly offshore of the Waiapu River mouth at 40 m and 60 m water depths from May - October 2004 (Fig. 1). Instrumentation included an Acoustic Doppler Velocimeter, Pulse-Coherent Acoustic Doppler Profiler, Acoustic Doppler Current Profiler, Acoustic Backscatter, YSI, and sediment traps (Table 1) (Wright et al., 2006; Ma et al., in prep). Waiapu River discharge records and wind data from nearby Hicks Bay were also collected during this interval by the Gisborne District Council Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 88 (GDC). From these records, a period, May 15 -July 10, 2004, of high fluvial delivery and energetic waves and currents was identified. Records of Waiapu River discharge and the observed wave height and current record reveal three main stormy periods during May, 15 - July 10, 2004 (Fig. 2). Two of the Waiapu River floods produced estimated sediment concentrations above 40 g/L, the threshold for hyperpycnal plume transport (Mulder and Syvitski, 1995). During this period, large excursions in current strength were also seen (Fig. 2) (Wright et al., 2006). Measured depth averaged along-shelf current velocities at times exceeded 80 cm/s, and across-shelf velocities exceeded 40 cm/s. When depth-averaged along-shelf currents flowed to the north-east, depth-averaged, across-shelf currents flowed seaward. The opposite pattern was observed when depth-averaged, along-shelf currents flow to the southwest, when across-shelf currents were oriented in the landward direction (Fig. 2). Each storm period was associated with a characteristic wave and current record (Fig. 2, Table 2). The first stormy period, from May 20 - Jun 7, was characterized by several relatively smaller discharge events grouped together, and produced river sediment concentrations below 35 g/1. Wave height did not exceed 4 m until after the last peak in river discharge on 6/8/2004. During this period, depth-averaged, along-shelf currents varied from being oriented towards the north-east, to the south-west, and back to the north-east. Along-shelf currents during the second storm event (6/18 - 6/24/ 2004) generally flowed to the south-west throughout the storm. Wave heights exceeded 2 m before and after the peak storm discharge. The third storm (6/29 - 7/5/2004) quickly followed after the second, concurrent with high wave energy (up to 4 m), and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 9 extremely strong depth-averaged along shelf currents to the north-east just after the peak of the storm (Fig. 2). The Waiapu River discharge record, a wind record from nearby Hicks Bay, and wave height data collected during the 2004 field experiment were used to run the three- dimensional model (Figs. 1 and 2). The following sections (Y and VI) describe how the model was configured and how the tripod, discharge, and wind data are incorporated into the simulations. V. Methods This study used a modified version of ECOMSED (Blumberg and Mellor, 1987) described by Harris et al. (2004, 2005) to evaluate the relative importance of suspension and gravity flows on the Waiapu River shelf. Sediment transport and deposition for periods of up to 80 days were successfully simulated without apparent instabilities. The model was used to investigate sediment transport and depositional sensitivities to settling velocities, the timing of fluvial delivery, wave height, and current strength from May 15 through July 10, 2004. Simulation of longer time periods seemed unrealistic because the model does not account for consolidation, and overestimates the remobilization of sediment. The model grid contains 100 x 120 horizontal grid cells, with an average resolution of 400m x 500 m (Fig 3). Horizontal grid resolution was highest near the river mouth. Grid cells along the model’s open boundaries were expanded to ameliorate instabilities caused by river plume interaction with the boundary edges. The southern boundary was elevated to create ~ 6 cm/s geostrophic current, so that modeled currents Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 0 matched observed time-averaged values. The bathymetric data provided by NIWA was smoothed with a low-pass filter (Shapiro, 1970) to avoid model instabilities. Depths greater than 200 m were neglected because the model cannot realistically simulate deep water circulation, in effect, assuming that the surface layer (<200m) does not interact with deeper waters. Nine sigma layers with high resolution at the top and the bottom of the vertical grid were used. Time steps of 45 seconds were used for model stability. Wave, wind, and discharge data served as input to the model. Wave propagation was assumed to be perpendicular to shelf bathymetry at 129 degrees, and uniform, not accounting for shoaling or diffraction. Observational data and NIWA’s TIDE2D model indicate that waves and non-tidal currents are much stronger than tidal currents. Therefore, tides were neglected. Initial sea salinity was set at 32 psu based on satellite estimates (from SeaWiFS images). Oceanic and Waiapu River temperatures were based on monthly average temperatures for the region and were held at 16°C (SeaWiFS images) and 10°C (historical averages from weatheronline.co.uk), respectively. Hourly wind data from Hicks Bay, to the north of the study area, were used to force wind- driven currents (Figs. 1 and 2). Hourly Waiapu River discharge and suspended sediment concentration data (Fig. 2) were available and used to specify sediment delivery. The model assumed that the Waiapu River was the only source of sediments and that, initially, there were no sediments on the seafloor. The model assumed likely partitioning of fine-grained sediment into flocculated and unflocculated populations. Flocculated sediment was assumed to settle at 0.1 cm/s (Hill and McCave, 2001), but a fraction of the fluvial load was assumed to be unflocculated and settle more slowly at Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 0.01 cm/s. Aggregation and disaggregation of these sediment types were neglected. Transport of coarser sands was not addressed because we were concerned only with material that makes up the observed muddy deposit. Furthermore, ECOMSED becomes unreliable when simulating coarser fractions because the model was designed for sediments no coarser than fine sand (van Rijn, 1984). Resolving high settling velocities would require a prohibitive decrease in time-step, violating the Courant-Friedrichs-Levy (CFL) stability condition (ElydroQual, 2002). Therefore, the model was used to estimate transport and deposition of muds that make up the flood deposit. VI. Approach Most of this chapter discusses model results made using the combined suspended sediment transport and gravity flow model of Harris et al. (2004). Oceanographic conditions and sediment transport for May- July 2004 were simulated using this model. Tripod data from 2004, including near-bed sediment concentrations, smoothed wave data, and current profile velocities, were compared with model results and used to select model parameters. Currents and sediment input were specified based on the Waiapu River hourly discharge record and winds from Hicks Bay (GDC data). Suspended sediment concentrations were estimated using the rating curve of Hicks et al. (2004). After demonstrating that this simulation represented conditions during the field experiment, model calculations were used to evaluate the relative importance of sediment transport mechanisms on the shelf. The quality of the model forcings were evaluated and selected based on how well the model estimates fit the observed tripod bed shear stress, currents, and sediment Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. concentrations. The two closest wind stations to the field site were located at Hicks Bay and the Gisborne Airport, and each record had potential problems (Fig 1). The Hicks Bay hourly record provided a higher resolution record than the Gisborne record, which provided data every 2 -3 hours. The Gisborne coastline is oriented similarly to the Waiapu River coast, however, potentially providing more accurate wind direction data than the Hicks Bay coastline (Fig. 1). Winds were primarily from the north-west and south for the Hicks Bay record, with winds from the NW during peak fluvial discharge (Fig. 4). Wind velocities at Gisborne were slightly lower than the Hicks Bay record with winds blowing primarily from the north-west. A comparison of depth-averaged current magnitude for model runs that used the two wind records versus observed conditions indicated that the Hicks Bay winds better represent the period of the field experiment than the Gisborne winds (Fig. 5, Table 3). A lack of sediment grain-size composition data and knowledge of aggregate packaging for the Waiapu River also contributed to uncertainty in the model calculations. Aggregate properties define the degree to which fine sediments are flocculated or unflocculated, thereby altering the effective settling velocity of the sediments. The transport calculations are sensitive to settling velocity which is difficult to specify for fine silts and clays (Hill and McCave, 2001). Therefore, several combinations of floe fraction were used as fluvial input, and then compared to tripod data. Sediment concentration observations were derived by calibrating YSI turbidity data from 99 cm above the seabed at the 60 m tripod site (Table 1) (Ma, unpub. data). The maximum thickness of the wave boundary layer, as estimated from the wave height data, was 12 cm at this location, indicating that the YSI measurements were collected Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 93 well above the wave boundary layer. Therefore, sediment concentrations from the sigma layer above the wave boundary layer, at about 1 m above the bed, were compared with the tripod data. The majority of this paper describes the set of calculations that best fit the observed sediment concentrations: a partitioning of material as 75% flocculated and 25% unflocculated. Wave time-series were measured by both tripods, but there were distinct differences in the two wave height records (Fig. 2). The mean wave height at the 40 m tripod was 2.3 m with a significant wave height of 3.5 m, higher than the mean wave height of 1.5 m and significant wave height of 2.4 m observed at the 60 m tripod. There were also periods of high waves observed at each location that were not seen in the record from the other location. The model was run using the smoothed wave height record from both tripods (Fig. 2). The wave record from the 60 m tripod was chosen as preferable to the 40 m record primarily because sediment concentration data were collected only at this station. Modeled bed shear stresses at the 60 m site were comparable with those modeled by Ma et al. (in prep) using the Wiberg et al. (1994) 1 D model. Bed shear stresses at the 40 m site calculated using the Wiberg ID model were much lower than estimated by the ECOMSED model (Table 4). To simulate river-initiated hyperpycnal plumes, a second set of calculations was also attempted. The model was altered to include sediment concentration into calculations of density. Average sea and fluvial winter temperatures were used to test realistic thresholds for hyperpycnal sediment transport. Floods of various peak sediment concentrations from 0.042 g/1 to 60 g/1 for up to 20 days were simulated. A discharge and sediment concentration curve was developed to slowly increase sediment Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 4 concentration from 0.042 g/1 to 60g/l to identify the stability limits of the model. Unfortunately, this approach resulted in violations of the vertical stability criteria in the model, which are further described in the Discussion, section VIII d. VII. Results Model results were compared to bed shear stresses, suspended sediment concentrations, and current profiles collected during 2004 by a bottom boundary layer tripod (Table 1) (Wright et al., 2006). This section describes the model results for these parameters calculated at the site of the 60 m tripod as well as the general transport and depositional patterns observed on the shelf. Three periods of interest were identified in the simulated period (Fig. 2, Table 2). Several relatively smaller discharge peaks comprise the first identified flood period during the latter part of May, 2004, with a maximum estimated fluvial suspended sediment concentration of 35 g/L. This stormy period was followed by two floods, from 6/18 - 6/24 and 6/29-7/05, when estimated sediment concentrations reached about 50 g/L. A. Simulation estimates compared to observations In this section, model estimates of currents, near-bed sediment concentration, and bed shear stresses are compared to values obtained from the 60 m tripod measurements. 1. Bed shear stresses Estimated bed shear stresses from ECOMSED correlated (correlation coefficient = 0.83) well with low-pass filtered values obtained using data from the 60 m tripod (Ma et al., in prep). The model tended to over-estimate bed shear stress values, however, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 5 when wave heights exceeded 3m (Fig. 6). The ECOMSED model may have over estimated bed shear stresses at times with high waves because the model neglected suspended sediment stratification that would dampen bed shear stresses (Traykovski et al., in press). The model also assumed a constant bed roughness of 0.03 cm instead of a time-varying bed roughness. The current component of the simulated bed shear stresses also was under-estimated on June 2 and Julyl, during periods of strong along-shelf depth averaged currents (Figs 5 and 6). 2. Sediment concentrations Model estimates of sediment concentration at the 60 m tripod location were similar in magnitude to observed values and timing was correct for 3 out the 4 resuspension events. The YSI turbidity record was low-pass filtered to make the resolution of the measurements comparable to the simulation record frequency of two hours. The observed filtered record showed four peaks in sediment concentration ranging from -0.5-1.8 g/L (Fig. 6). The model predicted these four peaks, but also exhibited an extra peak in sediment concentration above 1 g/L not seen in the observed record during a flood between 6/18 and 6/23. Furthermore, the two simulated peaks on 6/7 and 7/1 lagged behind the observed peaks, and were associated with a period of high waves and a flood, respectively. One potential reason for this discrepancy may be that the model continued to resuspend and transport sediment that may actually have been consolidated on the inner shelf. Spatially variable suspended sediment concentrations and suspended sediment stratification also may have contributed to the differences in the suspended sediment records, and will be discussed further in this paper. 3. Currents Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Observed time- and depth-averaged current speeds, measured by the ADCP, compared well with the simulated time-and depth-averaged currents, but the model did not produce the large fluctuations seen in the observed depth-averaged currents (Fig. 5; Tables 1 and 3). Shelf-wide, depth-averaged simulated current speeds were 13 cm/s with a range of 0 to 30 cm/s, and generally flowed to the north-east. Time- and depth- averaged simulated along-shelf currents at the 60 m tripod site were 4.4 cm/s to the north-east. This compared favorably with the observed mean along-shelf velocity of 6.3 cm/s to the north-east. The tripod data displayed periods of strong depth-averaged along-shelf currents to the north-east, concurrent with relatively strong across-shelf velocities in the offshore direction (Figs. 2 and 5). When observed depth-averaged, along-shelf currents flowed to the southwest, across-shelf currents flowed in the landward direction. Simulated across-shelf depth-averaged currents matched both the timing and the direction of the observed depth-averaged across-shelf currents during times of high fluvial discharge (Fig. 5). Simulated across-shelf depth averaged currents tended to be slightly stronger than observed and in the opposite direction when fluvial discharge was low, however. Estimated across-shelf mean velocities within the wave boundary layer compared favorably with observed across-shelf velocities (Table 5). Simulated mean along-shelf velocities within the wave boundary layer were slightly lower than observed. The component of the across-shelf velocity due to gravity can be estimated by subtracting the across-shelf current velocity at ~1 mab from the across-shelf current velocity in the wave boundary layer (Fig. 7; Table 6). Generally, in both the observed and simulated cases, gravity-driven currents were strong when depth-averaged across- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 7 shelf currents flowed landward (Fig. 7). A comparison of observed across-shelf velocity due to gravity and observed sediment concentration ~1 mab records showed that the presence of suspended sediment did not necessarily correlate to strong gravity- driven transport (Fig. 6). The model accounted for wind forcing and plume dynamics, and represented well for time- and depth-averaged current velocities and the average current velocities in the wave boundary layer (Tables 3,5, and 6). The simulation seemed to capture the timing of the depth-averaged currents in the along-shelf direction, with a correlation coefficient of 34%, and with the across-shelf currents at 52% (Table 3). Simulated wave boundary layer currents correlated well with observed currents (Fig. 5, Table 5). Since the model under-estimated the depth-averaged, along-shelf current velocities when current fluctuations were high, the model results may have under-estimated the influence of currents on the wave-current boundary layer and gravity-driven transport. Possible reasons that the model did not account for the magnitude of the depth-averaged currents include larger scale shelf currents not accounted for in the forcing and wind forcing data that did not accurately depict the conditions at the experiment site. Discrepancies between the model and data will be discussed later in the paper. 4. Summary The model compared favorably with observed conditions at the 60 m tripod using as input the waves recorded by the 60 m tripod, winds measured at Hick’s Bay, historical Waiapu River discharge data and a sediment aggregate composition of 75 % flocculated and 25% unflocculated. Although the model matched the observed time- and depth-averaged current speeds, the strong excursions exhibited in the observed Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 8 currents were not reflected in the model results. Simulated suspended sediment concentrations 1 m above the bed at the 60 m tripod site matched the magnitude and timing of most of the observed record, but the model predicted periods of high sediment concentration that were not observed. Calculated bed shear stresses generally agreed with those calculated using the 60 m tripod data, indicating that the wave component of the bed shear stress dominated the shear velocities near the bed. B. Sediment transport and deposition In this section, sediment mass distribution and export from the shelf area were used to evaluate shelf-wide transport and depositional patterns. 1. Shelf sediment budget Most Waiapu River sediments delivered to the shelf during floods were dispersed on the shelf in dilute suspension (Fig. 8). Generally, the freshwater plume was disconnected from the suspended sediment plume (Fig. 9). This disconnect was in part due to low fresh-water input for such high sediment delivery so that the freshwater plume dissipated before sediments settled out of the water column, and also in part due to energetic waves and currents that kept sediments in suspension and resuspended freshly deposited sediments from the sea-bed. A suspended sediment plume remained on the shelf up to several days after Waiapu River discharge had decreased to non-flood levels and the fresh-water plume had disappeared. When wind-driven currents were strong along the shelf or towards the coast, the suspended sediment plume was confined to the inner shelf. Strong along-shelf winds to the north-east resulted in down-welling favorable conditions, whereas strong Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 9 winds to the south-west resulted in up-welling favorable conditions (Fig. 9). When the winds relaxed, and along-shelf currents were weak, the suspended sediment plume spread out over the shelf. During the simulated floods, depth-averaged along-shelf wind-forced currents tended to flow to the south under up-welling favorable conditions, reversing direction after the peak discharge to more down-welling favorable conditions (Fig. 2). The distribution record of sediment mass on the shelf indicated that an average of 46% of the fluvial sediment budget, at any one time, was available for transport or deposition on the seabed during the simulation (Fig. 8, Table 7). This shelf budget was calculated by dividing the total mass of sediment in suspension, in the wave boundary layer, and on the seabed by the cumulative terrestrial sediment input for each time step and taking the time average. Sediments deposited on the seabed accounted for the majority of the sediment on the shelf, accounting for an average of 24.3% of the fluvial budget. Sediments in dilute suspension generally accounted for 15% of the fluvial budget, with periods of high suspended sediment mass during floods and periods of energetic waves. Times of relatively high suspended sediment mass corresponded with reduction of the sediment mass on the bed, indicating resuspension. Although the sediment mass in suspension was generally greater than that in the wave boundary layer, there were periods when the mass of sediment transported in the wave boundary layer exceeded dilute suspension (Fig. 8). This was likely to happen when energetic waves occurred during floods (6/28-7/04) (Fig. 2). By dividing the total mass of sediment in the wave boundary layer or in dilute suspension by the total amount of sediment in both for each time step, the relative Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. importance of the sediment transport mechanisms could be evaluated (Table 7). Although there were periods when the sediment mass in the wave boundary layer was greater than that in suspension, the sediments in dilute suspension accounted for an average of 70% of the sediments being transported on the shelf. The two settling velocity values (0.01 and 0.1 cm/s) that were used resulted in the segregation of the sediment types by the two transport mechanisms (Fig. 8). Even though material with low settling velocities (0.01 cm/s) only accounted for 25% of the sediment delivered to the coastal ocean, these unflocculated sediments accounted, on average, for 33% of sediment in suspension during the simulation (Table 7, Fig. 8). Flocculated sediments quickly settled through the water column and were preferentially transported within the wave boundary layer (Table 7). An average of 98% of the sediment transported within the wave boundary layer was flocculated, with the unflocculated sediment contribution increasing when wave energy and fluvial discharge were high (7/01/2004). This difference in transport mode between the two sediment types resulted in the preferential retention of flocculated sediments, which accounted for >98% of the deposited sediment (Fig. 8). 2. Sediment export Those sediments not accounted for in dilute suspension, within the wave boundary layer, or on the seabed at the end of the simulated period were exported out of the modeled area. About 62.4 % (~11.1 MT) of the 17.8 Mt of fluvial sediment delivered to the shelf was exported from the area of the model grid during the study period (Fig. 10). Almost all of the exported sediment was transported to the north of the shelf area and only about 0.2% of the fluvial budget was transported across the shelf Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 101 break (Fig. 10, Table 8). The mean geostrophic current to the north only allowed sediments to temporarily escape across the southern boundary. Therefore, all the sediments transported to the south of the shelf during southwest flowing currents were redelivered into the shelf area when the wind forced currents reversed and flowed to the north. These findings are consistent with the findings of Ma et al. (in prep), who indicated that bulk transport was to the north during the 2004 field deployment. Sediment transport out of the Waiapu Shelf area was sensitive to the assumed sediment settling velocities. Approximately 98% of available sediments with low settling velocities (0.01 cm/s) were exported, accounting for 47% of all sediments exported from the shelf area (Table 8). Flocculated sediments accounted for the remaining 53% of sediment exported from the shelf. Although about an equal proportion of flocculated and unflocculated sediments were transported across the northern boundary, sediment transport across the eastern boundary was dominated by unflocculated sediments, whereas flocculated sediments comprised the majority of transport across the southern boundary. Upwelling conditions, distance from the river mouth, attenuation of the wave boundary layer with depth, and settling velocity of the sediments likely account for the differences in sediment export records for the boundaries. Almost all of the sediment exported from the shelf was transported in dilute suspension, with transport in the wave boundary layer only accounting for a fraction of export across each boundary (Table 8). 3. Sediment depositional patterns At the end of the simulated period, approximately 38% (6700 MT) of the Waiapu River budget was retained on the shelf, 24% of the budget had been deposited Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. on the seabed, 13% remained in suspension, and the remainder had been exported from the shelf (Tables 7 and 8). Material modeled as flocculated (0.1 cm/s) comprised the majority of the sediment deposited, accounting for 99% of the deposit averaged over the study period. Most of the final deposit was located between 70 and 130 m depth, with the remaining fraction dispersed over the inner shelf, shelf break, and slope (Fig. 11 A and B). Sediments were only ephemerally deposited on the inner and mid-shelf. Most of the sediments deposited on the inner shelf were either resuspended and moved off the shelf or transported across the shelf within a few days. Sediments were retained on the mid-shelf for longer periods of time before energetic waves resuspended them and facilitated their removal. The main locus of deposition was directly offshore and to the north of the river mouth at between 60-70 m depth, reaching ephemeral deposit thicknesses of greater than 20 cm (Fig 11 A). When simulated wind-driven, along-shelf currents were strong and coincided with high river discharge, sediments were ephemerally deposited along the length of the inner shelf (Fig. 1 IB). This elongated pattern of deposition was maintained as sediments were transported off-shore primarily within the wave boundary layer. For example, the mid- to outer shelf depositional footprint extended to the south when strong along-shelf currents to the south-west coincided with high river discharge. Final deposit thicknesses were generally between 10 to 20 cm, but in some locations, ephemeral deposits reached up to 60 cm when river sediment delivery was high and sediments quickly settled to the seabed. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 103 C. Floods with large wave energy compared to floods with low wave energy Two patterns of sediment transport were identified in the simulated period (Fig. 11A and B). Floods were either accompanied by energetic waves (6/29 -7/4) or had low waves during peak Waiapu River discharge (5/20-6/7 and 6/18-6/25). When waves were low during high river discharge, sediments rapidly settled out of the river plume and deposited on the inner shelf (Fig. 1 IB). These sediments were not resuspended and moved to deeper waters until waves became more energetic. The longer suspended sediments were retained in shallow waters, the more likely they were to be dispersed along the length of the shelf, forming an elongate deposit. Furthermore, sediments were more likely to be exported from the shelf when they were not rapidly transported across the shelf to deeper waters. This occurred because sediments retained in shallower waters were more likely to be resuspended by energetic waves and currents subsequent to the flood. When wave heights were high during floods, then sediments were rapidly moved across the inner shelf to the mid- and eventually, outer-shelf, where waves were attenuated (Fig. 11 A). Therefore, sediments transported to the outer shelf were removed from the influence of all but the most energetic waves. When wave energy was high during floods, the rapid transport of sediments across the shelf resulted in a more localized deposit shape and thicker deposits than created by floods that occurred during low wave energy. VIII. Discussion Spatial and temporal variability in sediment transport and deposition and the environmental conditions that influence flux are discussed in the following sections. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 104 Uncertainties in settling velocities and model forcings are also examined. Finally, we address attempts to account for suspended sediment density in the hydrodynamic calculations. A. Sediment transport mechanisms 1. Time-averaged flux patterns Patterns of time-averaged flux with depth across the shelf indicated that the inner and mid-shelf were dominated by wave-induced, gravity-driven transport, the mid-shelf was a zone of resuspension and export, and across-shelf suspended flux increased in importance with depth across the shelf (Fig. 12A). The time-averaged across-shelf flux for each along shelf grid row of the model indicated that across-shelf transport within the wave boundary layer was highest on the inner and middle shelves to about ~60 m depth, beyond which wave bed shear stresses attenuated. Time-averaged across-shelf suspended sediment flux increased in importance with distance across- shelf, up to its peak at between 60-70 m depth. Transport of sediments in dilute suspension accounted for the majority of along-shelf flux, with only a fraction of sediment transported within the wave boundary layer (Fig. 12A, Table 8). Almost all of the sediments exported out of the shelf study area were transported to the north of the river mouth, with peak time-averaged suspended flux occurring between 65-85 m depth (Fig. 12 A). The time-averaged difference in the vertical distribution of flocculated (ws=0.1 cm/s) and unflocculated (ws=.01 cm/s) sediments resulted in different transport records (Fig.l2B). Unflocculated sediments, with high concentrations at the surface and bottom of the water column, were generally transported seaward and to the north within the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 105 buoyant river plume during floods (Fig. 12A). The relatively small amount of unflocculated sediment transported in the wave boundary layer was eventually resuspended and did not contribute to deposition. Most unflocculated sediments were transported in dilute suspension, whereas a little more than half of the available flocculated material was transported in dilute suspension (Figs. 8 and 12A). Flocculated concentrations increased towards the seabed and were greatest in the wave boundary layer (Fig. 12B). Therefore, flocculated sediments comprised most of the sediments flux within the wave boundary layer. The difference in settling velocity between unflocculated (0.01 cm/s) and flocculated (0.1 cm/s) materials resulted in very different along- and across-shelf flux patterns (Figs. 8 and 12). Although Waiapu River sediments were assumed to have a 75% flocculated to 25% unflocculated ratio, across-shelf flux was dominated by unflocculated sediments in waters shallower than -60 m and greater than 80 m in the along-shelf direction. Across-shelf flocculated suspended flux increased with depth across the shelf, up to its peak at between 55 - 80 m depth. This peak indicates that the mid-shelf zone was subject to either sediment resuspension from the seabed or wave boundary layer and/or removal of unflocculated sediments (Fig. 12A). Time-averaged, along-shelf flux patterns show that unflocculated sediment dominated flux of sediments from the shelf in water depths shallower than ~25m depth and greater than ~90m depth, with a peak at 60-80 m depth. Flocculated suspended flux accounted for the majority of flux out of the shelf area between -30 - 90m depth. Peak along- and across-shelf suspended flux occurred at 30-80 m depth, supporting the idea that some of the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 106 sediments resuspended from this region were both re-deposited and exported out of the shelf area. Waves and currents accounted for some of the variability in sediment transport and resuspension with depth across the shelf and with distance from the river. A correlation of across-shelf flux with the wave record indicated that wave-resuspension into the wave boundary layer and the overlying water column was more important than currents at water depths between 20-80 m (Fig. 13). While wave correlations were just as high at depths greater than 130 m, flux within the wave boundary layer and the water column was very low there (Figs. 12A and 13). During a period of high waves and discharge (6/30 - 7/05/2004), sediments were transported within the wave boundary layer at depths greater than 100 m. Currents did not correlate well with flux within the wave boundary layer, but, not surprisingly, influenced both along- and across-shelf suspended sediment flux (Fig. 13). Across-shelf suspended sediment transport and resuspension by currents became increasingly important relative to gravity-driven flux with depth across the shelf. Currents most strongly influenced along-shelf flux up to ~110 m depth, beyond which wave resuspension dominated, reflecting conditions during a flood that occurred during a time of high waves. 2. Flux patterns during high discharge Two patterns of sediment dispersal were identified during the simulated period, both centered around periods of high flood discharge (Fig 11 A and B). Floods were either accompanied by energetic waves (6/29 - 7/4) or had low waves during peak Waiapu River discharge (6/18 - 6/24). When waves were low during high river discharge, sediments were immediately lost from the freshwater plume on the inner Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 107 shelf (Fig. 1 IB), a pattern that has been observed offshore of the Eel River (Geyer et al., 2000). These sediments were not resuspended and moved to deeper waters until waves became more energetic after the flood. The longer suspended sediments were retained in shallow waters, the more likely they were to be dispersed by strong along-shelf currents, forming an elongate deposit on the inner shelf. Furthermore, sediments were more likely to be exported from the shelf when they were initially deposited on the inner shelf. This occurred because sediments in shallow water were more susceptible to resuspension into the water column by waves and subsequent removal by strong along- shelf currents than sediments deposited initially in deeper waters. When wave heights were high during floods, sediments were moved across the inner shelf by gravity flows to the mid- and, eventually, outer shelf within hours (Fig. 11 A). This rapid transport to deeper waters contributed to greater sediment retention on the shelf because the sediments were removed quickly from the area of high wave resuspension on the inner shelf. The resultant shape of the deposit was more localized, and created thicker deposits on the mid- and outer shelf areas than deposits formed during floods accompanied by low wave energy. A comparison of the time-averaged across-shelf flux for the simulation and the two floods described above showed that energetic waves during peak flood discharge resulted in greater sediment flux within the wave boundary layer and a seaward shift in total peak across-shelf flux than the other two cases (Fig. 14). Although the simulation indicated that sediments deposited on the inner shelf were eventually resuspended, it is important to note that because the model does not account for bed consolidation or burial by inner shelf sands, more sediment was eroded from the inner shelf than was Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 108 realistic. Actually, some sediment was likely retained on the inner shelf during floods and periods of low wave energy. It is also important to remember that differences in along-shelf currents during the floods played a significant role in dilute flux patterns. B. Flux convergence and divergence across the shelf Sediment transport mechanisms and depositional patterns varied significantly with bathymetry across the shelf. The inner, middle, and outer shelf, and shelf break zones can be distinguished by the relative importance of suspended sediment and gravity-driven transport and depositional history (Fig. 15, Table 9). Wave boundary layer fluxes were highest on the inner and mid-shelves (Fig. 12A). Sediment transport in dilute suspension became more important compared with wave boundary layer flux with depth as waves were attenuated (Fig. 12A; Table 9). Pulses in sediment transport both in dilute suspension and in the wave boundary layer increasingly lagged behind fluvial delivery with depth across the shelf (Fig. 15). Across-shelf convergences and divergences in sediment flux were used to assess the contribution of wave-induced gravity flow and dilute transport to sediment deposition and removal. The Exner equation, _sv=j_['s?i+a?, + s>i' dt cb dx dy dt ^ was used to infer the change in seabed elevation, 8t|, from the changes in flux across and along the shelf and the change in sediment volume over time. Changes in flux (5q) in the across- (x) and along-shelf (y) directions, divided by the area (8x and 8y) and the density of sediment (Cb) were used to evaluate depositional and erosional Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 109 thicknesses (5rj) over time (St). The sediment volume parameter (Vs) was not used in this approach. Along- and across-shelf flux gradients were calculated for each along- shelf row of the model grid. To estimate flux convergence and divergence in the across shelf direction, the time-averaged flux of sediment for each along shelf row was calculated in both the along and across-shelf direction and corrected for the area of the model grid row (Fig. 3). By comparing the spatial gradients in the along-and across-shelf flux, I was able to assess the relative contribution of each sediment transport mechanism to flux convergence and divergence (Fig. 16). It is important to keep in mind that a transport mechanism was assumed to be erosive if the flux out of an area increased respective to what was supplied to the area, causing flux divergence. Deposition was indicated when the flux out of an area was less than the flux into the area, causing flux convergence. The flux convergence and divergence values for the transport mechanisms may indicate multiple processes, however. If time-averaged across-shelf suspended sediment transport is negative, or flux convergent, values indicate that sediments were either deposited on the seabed or settled into the wave boundary layer. Positive, or flux divergent, values indicate that time-averaged across-shelf flux increased. The increase in flux out of an area may be indicative of sediment erosion from the seabed or sediment resuspension from the wave boundary layer. If time-averaged across-shelf flux within the wave boundary layer is negative, or flux convergent, sediments may have been either deposited or resuspended back into the water column. Positive, or flux convergent, values may indicate resuspension from the seabed or sediments settling into Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 110 the wave boundary layer from the water column. These results must therefore be compared with sediment mass distribution to correctly interpret the results. Time-averaged changes in across-shelf flux indicated flux convergence, whereas differences in flux in the along-shelf direction implied flux divergence (Fig. 16A). Across-shelf flux gradients were highest on the inner shelf between 25 - 40 m and on the mid-shelf between 55 - 75 m water depths. Along-shelf suspended sediment flux gradients to the north of the river mouth were divergent, and peaked on the mid-shelf. The time-averaged across-shelf flux gradients for each transport mechanism varied with depth across the shelf. Sediment flux convergence within the wave boundary layer indicated deposition across the shelf and/or resuspension back into the water column. At around 100 m depth, sediment flux convergence within the wave boundary layer decreased to zero. Flux divergence in the time-averaged suspended flux within the water column indicated erosion or entrainment up to ~65m water depth, beyond which the suspended flux gradients were convergent. This divergent suspended sediment flux gradient peak, between water depths of 40-75 m, corresponded with the convergent wave boundary layer flux gradient peak, indicating that about half of the flux convergence in the wave boundary layer may be accounted for by resuspension of sediments into the water column (Fig. 16B). These patterns in along- and across-shelf flux gradients suggest that sediments transported within the wave boundary layer were retained on the shelf whereas sediments in dilute suspension tended to be exported from the shelf area. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Ill C. Temporal and spatial variability in sediment transport and depositional patterns During the field experiment, fluvial delivery, frequency of floods, and wave energy impacted sediment preservation on the shelf. Sediments deposited on the shelf during floods were subsequently resuspended by energetic waves and currents (Figs 2, 8, and 15). Initial deposit location depended on wave energy during a flood, as discussed earlier. Frequent floods temporarily increased sediment retention on the shelf during the simulation by supplying more sediment than could be resuspended. Deposition on the inner shelf was largely ephemeral and associated with floods, whereas the rest of the shelf showed a steady increase in deposition over the period of the experiment (Table 9) (Fig. 15). Deposition on the mid and outer shelf significantly lagged behind floods, but deposition and accumulation peaked there. Very little sediment reached beyond ~100 m, and material did not reach those depths until about a month and a half into the simulation. At the end of the simulation, 23.6% of the fluvial budget was retained on the shelf, with most deposited on the outer shelf. These modeled depositional patterns were consistent with radioisotope geochronology data from the Waiapu River shelf (Kniskem, Chapter 1). Short-lived ?Be data suggested that fluvially-delivered fine sediments were deposited on the inner shelf temporarily. Excess 210Pb sediment profiles suggested that longer-term accumulation was largely constrained to the mid- to outer-shelf (Kniskem, Chapter 1). However, analyses of inner shelf seismic profiles suggest that sands and some fine fluvial sediments were retained on the inner shelf over longer time-scales (Wadman and McNinch, 2006). The simulation did not account for sediment consolidation, burial by Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 112 inner shelf sands, or non-linear wave transport, and may have therefore underestimated sediment retention on the inner shelf. Although this model was able to predict how sediments are delivered to the inner shelf, adding the aforementioned processes to the model would make it more robust for estimating sediment retention and accumulation. D. Sensitivity to settling velocity Flocculated and unflocculated sediment types were partitioned various ways to simulate sediment transport and depositional patterns on the Waiapu River shelf. The amount of sediment delivered to the shelf was estimated using the rating curve of Hicks et al. (2004), but information about sediment grain sizes or aggregation properties was not available. Four sediment cases were used: 100% flocculated sediment (ws = 0.1 cm/s), 100% unflocculated sediment (ws = 0.01 cm/s); 75% flocculated and 25% unflocculated; and 25% flocculated and 75% unflocculated. The resulting transport and depositional patterns indicated that sediment preservation on the Waiapu River shelf may be strongly influenced by sediment settling properties of silts and clays. It is important to keep in mind that this ratio can change with time and sediment concentration (Hill et al., 2000), and that the model did not account for processes of aggregation and disaggregation. The proportion of unflocculated material significantly influenced how much sediment was transported in dilute suspension and, therefore, impacted the export of terrestrially-derived sediment from the shelf (Figs. 17, Tables 7 and 10). When more sediment was assumed to be unflocculated (ws=0.01 cm/s), more sediment was held in suspension, and, subsequently, the majority of the terrestrial contribution was exported. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 113 As a result, those model simulations with a high proportion of unflocculated sediment showed that less sediment was retained on the shelf during the study period. When more sediment was assumed to be flocculated (ws=0.1 cm/s), less sediment was carried in suspension, and more sediment was transported within the wave boundary layer. Simulations with a higher proportion of flocculated sediments retained more sediment on the shelf. Between 48% and 97% of the available fluvial sediment was exported from the shelf depending on the assumed partitioning between flocculated and unflocculated sediment. A comparison of the sediment budgets produced by the sediment aggregate properties experiment with the observed geochronology sediment retention estimate of -24% (Kniskem, Chapter 1) suggests that between 75 to 100% of the Waiapu River sediment budget was delivered to the shelf in flocculated form. The suspended sediment concentration record from the 60 m tripod at 1 mab also correlates well with the 75% flocculated and 25% unflocculated case (Fig. 6). E. Dispersal processes not represented by the model This section discusses the potential causes that contributed to the discrepancy between the observed and the simulated depth-averaged currents. It also describes another model run which attempted to simulate hyperpycnal plume formation. 1. Currents The model represents well the time- and depth-averaged currents, but fails to capture the large fluctuations observed in the along-shelf currents (Fig. 5). There are two possible reasons the model did not reproduce the variability of the currents: (1) the high observed current velocities were forced by mechanisms that were not simulated in Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 114 the model, or (2) the winds used in the simulations do not represent the wind conditions on the Waiapu River shelf. The fluctuations in current were subtidal, and the pattern did not match the spring-neap cycle. The winds used in the model simulation are from Hick’s Bay, around East Cape from the Waiapu River mouth (Figs. 1, 2, 4, and 5). The modeled currents were wind-driven, and the non-local winds used might not be representative of the field site. Specifically, the Hicks Bay winds might underestimate the southern wind component because of the high topography of East Cape. In order to explore this issue, the model was run using Gisborne, New Zealand winds from -100 km to the south of the Waiapu River mouth (Fig. 1). The currents produced were even smaller than the currents predicted using the Hick’s Bay winds (Fig. 5A, Table 3). An analysis of observed depth-averaged currents and Hick’s Bay winds suggests that winds explained 16% of the variance in the observed currents and Gisborne winds accounted for 18% of the variance in current velocities. It therefore seems unlikely that the provenance of the wind data caused the discrepancies in currents between the model simulations and the observational data, but this cannot be confirmed without using winds from the Waiapu River shelf. Alternatively, a larger current system, unaccounted for by the boundary conditions, could cause the extremely strong currents observed on the Waiapu River shelf. Interactions between the southward flowing deep water current, the East Cape Current, and the northward flowing Wairarapa Coastal Current have created eddy systems that migrate along the East Cape coast (see Chiswell et al., 2002 their Fig 1). Although there have been few long-term moorings in the area and fewer measurements from the East Cape shelf waters, some studies have indicated that the East Cape Eddy Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 115 and the Wairarapa Eddy may influence circulation in shallower East Cape shelf waters (Chiswell et al., 2000 2003, 2005). Exchange of slope waters and shelf waters during periods of up-welling may also produce shelf circulation patterns unaccounted for in the model. Further analysis of the tripod data may explain the mechanisms that created the large excursions in current velocities. Attempts to simulate the observed depth-averaged, along-shelf current velocities by increasing geostrophic currents at the open boundary was only able to increase the mean current, and did not cause large fluctuations. Large scale currents could be accounted for by nesting model grid within a larger scale model (Pullen et al., 2003) or by data assimilation as done by Zavatarelli and Pinardi (2003), but that was beyond the scope of this study. The simulation reasonably predicted bed shear stresses, suspended sediment concentration magnitude, and the overall pattern and budget of sediment deposition in spite of underestimating current fluctuations. If the stronger currents were added to the simulation, it is likely that more sediment would be exported from the shelf to the north rather than potentially re-settling onto the seabed (Fig. 2). The observed wave and current record indicated that currents were to the north-east during times of relatively high wave energy and during the flood in early July and during the two smaller flood peaks in May and early June. Currents flowed weakly to the south-west during the mid- June flood and during the largest peak in May. The observed currents would also likely enhance gravity-driven transport on the shelf (Ma et al., in prep.). The ECOMSED model, however, confined gravity-driven transport to the thickness of the wave boundary layer, (Harris et al., 2004, 2005). This Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 116 may have limited the ability of the model to simulate gravity-driven transport in response to energetic currents in the thicker wave-current boundary layer. It is evident from the model results that the wave component of the bed shear stresses was more influential than the current component at the 60 m tripod site during the field experiment as has been seen at many other locations world-wide (Sternberg and Larsen, 1976; Sternberg 1986, Cacchione et al., 1987; Sherwood et al., 1994). 2. River-initiated hyperpycnal sediment transport Several studies have provided evidence that rivers whose sediment concentrations exceed ~40 g/L may form hyperpycnal plumes and that these may be important for stratigraphic development on continental margins (Mulder and Syvitski, 1995; Morehead and Syvitski, 1999; Hicks et al., 2004; Milliman and Kao, 2005). The Waiapu River discharge record indicates that the river can reach this value several times per year, even occurring twice in the space of two weeks in 2004 (Fig. 2). The ECOMSED model was modified by Harris et al. (2004) to account for gravity-driven transport within the thin wave boundary layer, accounting only for wave-initiated gravity-driven transport (Harris et al., 2004). If waves were neglected, gravitationally- driven transport would not develop in this model. On shelves where the fluvial sediment concentrations exceed the threshold for river-initiated gravity-driven transport, the Harris et al. (2004) model would fail to represent river-initiated hyperpycnal flows. Waiapu River suspended sediment concentrations were estimated to exceed the threshold for river-initiated hyperpycnal sediment transport at least three times over the period of the field experiment. To investigate transport during floods with high Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 117 sediment concentrations, models need to be developed that account for river-initiated gravity-driven flows by adding sediment to the density calculation. The ECOMSED model (Harris et al., 2004) was able to estimate sediment transport using this approach for fluvial suspended sediment concentrations less than 20 g/1. When sediment concentrations exceeded 20 g/L, however, simulated vertical velocities violated stability criteria of -0.2 cm/s. Sediment transport in simulations using sediment concentrations lower than 20 g/L did not vary significantly from model runs that did not include sediment concentration in the density calculations. The fluvial suspended sediment concentrations were too low for river-initiated hyperpycnal plumes to form. A model is needed that does not develop instabilities with high settling velocities. Only one other study has attempted to add sediment concentration to density calculations in a 3D model (Liu et al., 2002). Their modified Princeton Ocean Model (POM) held river sediment concentrations at 1 g/L, far below the fluvial suspended sediment concentrations observed on their study site, the Tseng-wen River, Taiwan (Liu etal., 1998,2000). Traditional vertical sigma coordinate systems might be incapable of representing the vertical structure of hyperpycnal plumes. The sigma coordinate system used by ECOMSED defines each vertical layer as a fraction of the water column. This means that the layers increase in thickness with depth, losing resolution. Furthermore, without very high vertical resolution, the model is unable to resolve sharp interfaces, reducing its capacity to simulate a lutocline. Two-dimensional models have been able to simulate river-initiated hyperpycnal plumes, but use a depth-averaged vertical configuration (Khan et al., 2005). ROMS, a Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 118 3D model, uses a stretched vertical grid, which allows those layers at the surface and near the bed to retain vertical resolution with depth. Since the ROMS model is not subject to as strict a stability criteria as ECOMSED and POM (Shchepetkin and McWilliams, 2003), it may be capable of simulating hyperpycnal plumes with riverine sediment concentrations exceeding -40 g/L. However, the vertical resolution of typical ocean models would likely still be too coarse to accurately represent the lutocline. Increasing the resolution, however, would increase the calculation time. More work needs to be done to resolve issues of model grid construction and river-initiated hyperpycnal plumes. IX. Conclusions 1) Approximately 62.4% of the fluvial budget for this simulation was exported from the shelf, mostly transported as dilute suspension to the north of East Cape. Most of the sediments retained on the shelf were deposited on the middle and outer shelf regions, accounting for 23.6% of the fluvial budget. A further 13.3% of the fluvial budget remained in suspension at the end of the simulation. 2) Sediment deposition varied both along-shelf and across-shelf. Sediments were temporarily deposited on the inner (0-30m) and mid-shelf (30-70m) before accumulating on the outer shelf (70-130m). Deposition was highest on the outer shelf over the length of the modeled period. Sediments were temporarily deposited to the south of the Waiapu River mouth when along-shelf winds to the SW were strong. Sediments were deposited to the north of the river mouth when along shelf winds to the north-east were dominant. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 119 3) Sediment transport mechanisms varied in importance with depth across the shelf. A near-bed fluid mud layer was active mostly on the inner and mid-shelf during the deployment period. Sediments were transported in the wave boundary layer on the outer shelf when wave energy was high. Suspended sediment transport increased in importance relative to wave boundary layer transport with depth across the shelf. Only a small percentage of sediment was transported to the shelf break and slope. 4) Sediments were transported along the shelf primarily in dilute suspension to the north of the river mouth and is consistent with the findings of Ma et al. (in prep). 5) Model estimates of bed shear stresses, sediment concentrations, and average currents were comparable to observations from the 60 m tripod from May - July, 2004. Although the model consistently under-estimated fluctuations in the along-shelf current velocities, the model readily estimated bed shear stresses and sediment concentrations at the tripod site. This is because wave resuspension dominated bed shear stresses most of the time. 6) Flocculation characteristics of flood material influenced sediment transport pathways and accumulation patterns. Sediments with high settling velocities, representing flocculated material, tended to be transported within the wave boundary layer and were more likely to be retained on the shelf as event deposits. Sediments with lower settling velocities tended to be removed from the shelf within dilute suspensions. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 120 References Addington, L.D., Kuehl, S.A., and McNinch, J.E., in press. Distinguishing sediment transport modes to the outer-shelf off the Waiapu River, New Zealand. Marine Geology. Blumberg, A.F. and G.L. Mellor, 1987. "A Description of a Three-Dimensional Coastal Ocean Circulation Model," In: Three-Dimensional Coastal Ocean Models, N. Heaps, Ed., 1-16, American Geophysical Union. Cacchione, D.A., Grant, W.D., Drake, D.E. and Glenn, S., 1987. Storm-dominated bottom boundary layer dynamics on the northern California continental shelf: measurements and predictions. Journal of Geophysical Research, 92: 1817— 1827. Chiswell, S.M., 2005. Mean and variability in the Wairarapa and Hikurangi Eddies, New Zealand. New Zealand Journal of Marine and Freshwater Research, 39:121-134. Chiswell, S.M., 2003. Circulation within the Waiararapa Eddy, New Zealand. New Zealand Journal of Marine and Freshwater Research, 37:691-704. Chiswell, S.M., 2002. Wairarapa Coastal Current influence on sea surface temperature in Hawke Bay, New Zealand. New Zealand Journal of Marine and Freshwater Research, 36:267-279. Chiswell, S.M., 2000. The Wairarapa Coastal Current. New Zealand Journal of Marine and Freshwater Research, 34: 303-315. Chiswell, S.M. and Roemmich, D., 1998. The East Cape Current and two eddies: a mechanism for larval retention?. New Zealand Journal of Marine and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Freshwater Research, 32:385-397. Collot, J-Y., Lewis, K., Lemarche, G., and Lallemand, S., 2001. The giant Ruatoria debris avalanche on the northern Hikurangi margin, New Zealand: Result of oblique seamount subduction. Journal of Geophysical Research, 106 (B9): 19271-19,297. Crockett, J.S., and Nittrouer, C.A., 2003. The sandy inner shelf as a repository for muddy sediment: an example from Northern California. Continental Shelf Research, 24:55-73. Curtin, T. B. and Legeckis, R. V., 1986. Physical observations in the plume region of the Amazon River during peak discharge: I. Surface variability. Continental Shelf Research, 6:31-51. Drake, D. E., and Cacchione, D. A., 1985. Seasonal variation in sediment transport on the Russian River shelf, California. Continental Shelf Research, 4: 495-514. Fan, S., Swift, D.J.P., Traykovski, P., Bentley, S., Borgeld, J.C., Reed, C.W., and Niedoroda, A.W., 2004. River flooding, storm resuspension, and event stratigraphy on the northern California shelf: observations compared with simulations. Marine Geology, 210:17-41. Farnsworth, K.L and J.A. Warrick, 2005. Sources of Fine Sediment to the California Coast: Computation of a 'Mud Budget', Interim Report to the California Sediment Management Working Group, California Dept, of Boating and Waterways, 128 pages. Farnsworth, K.L., Milliman, J.D., 2003. Effects of climatic and anthropogenic change Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 122 on small mountain rivers: the Salinas River example. Global and Planetary Change, 39:53-64. Friedrichs C.T. and L.D. Wright, 2004.Gravity-driven sediment transport on the continental shelf: implications for equilibrium profiles near river mouths. Coastal Engineering, 51:795-811. Geyer, W. R., Hill, P. S., Milligan, T. G., and Traykovski, P. A., 2000. The structure of the Eel River plume during floods, Cont. Shelf Res., 20: 2095-2111. Goni, M.A., Monacci, N., Gisewhite, R., Ogston, A., Crockett, J., and Nittrouer, C., 2006. Distribution and sources of particulate organic matter in the water column and sediments of the Fly River Delta, Gulf of Papua (Papua New Guinea). Estuarine, Coastal, and Shelf Sciences, 69:225-245. Grant, W. D. and Madsen, O.S.,1986. The continental shelf bottom boundary layer. Annual Review of Fluid Mechanics, 18. 265-305. Griffiths, G. A.; Glasby, G. P., 1985. Input of river-derived sediment to the New Zealand continental shelf. I. Mass. Estuarine, Coastal and Shelf Science, 21:773- 787. Harris, C.K., P.A. Traykovski, and R.W. Geyer. 2005. Flood dispersal and deposition by near-bed gravitational flows and oceanographic transport: a numerical modeling study of the Eel River shelf, northern California. Journal of Geophysical Research, 110, C09025, doi:10.1029/2004JC002727. Harris, C.K., P.A. Traykovski, and R.W. Geyer. 2004. Including a Near-Bed Turbid Layer in a Three Dimensional Sediment Transport Model with Application to the Eel River Shelf, Northern California. Proceedings of the Eighth Conference Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 123 on Estuarine and Coastal Modeling, ASCE, M. Spaulding (editor). 784-803. Harris, C.K. and P.L. Wiberg, 2001. A two-dimensional, time-dependent model of suspended sediment transport and bed reworking for continental shelves. Computers and Geosciences, 27: 675-690. Hicks, D.M., 1995. A way to estimate the frequency of rainfall-induced mass movements (note). Journal of Hydrology, New Zealand, 33, 59- 67. Hicks, D.M., B. Gomez and N.A. Trustrum, 2000. Erosion thresholds and suspended sediment yields, Waipaoa River basin, New Zealand. Water Resources Research, 36:1129-1142. Hicks, D.M., Gomez, B., and Trustrum, N.A., 2004. Event suspended sediment characteristics and the generation of hyperpycnal plumes at river mouths: east coast continental margin, North Island, New Zealand. The Journal of Geology, 112:471-485. Hill, P. S., and McCave, I. N., 2001. Suspended particle transport in benthic boundary layers. In The Benthic Boundary Layer: Transport Processes and Biogeochemistry, edited by B. P. Boudreau and B. B. Jorgensen. Oxford University Press, pp. 78-103. Hill, P.S., Milligan, T.G., Geyer, W.R., 2000. Controls on effective settling velocity of suspended sediment in the Eel River Plume. Continental Shelf Research, 20:2095-2111. HydroQual (2002). A primer for ECOMSED User’s manual version 1.3. HydroQual Inc, One Lethbridge Plaza, Mahwah, N.J., 07430, 188 p. Kasai, M., Brierley, G.J., Page, M.J., Marutani, T., and Trustrun, N.A., 2005. Impacts of Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 124 land use change on patterns of sediment flux in Weraamaia catchment, New Zealand. Catena, 64:27-60. Khan, S.M., Imran, J., Bradford, S., and Syvitski, J.,2005. Numerical modeling of hyperpycnal plume. Marine Geology, 222-223:193-211. Kineke, G.C., R.W. Sternberg, J.H. Trowbridge and W.R. Geyer, 1996. Fluid Mud Processes on the Amazon Continental Shelf. Continental Shelf Research, 16: 667-696. Kineke, G.C. and R.W. Sternberg, 1995. Distribution of fluid muds on the Amazon continental shelf. Marine Geology, 125: 193-233. Kineke, G.C., Woolfe, K.J., Kuehl, S.A.; Milliman, J.D., Dellapenna, T.M., and Purdon, R.G., 2000. Sediment export from the Sepik River, Papua New Guinea; evidence for a divergent sediment plume. Continental Shelf Research, 20:2239- 2266. Kuehl, S.A., G.J. Brunskill, K. Bumes, D. Fugate, T. Kniskem and L. Meneghini, 2004. Nature of sediment dispersal off the Sepik River, Papua New Guinea: preliminary sediment budget and implications for margin processes. Continental Shelf Research, 24(19):2417-2429. Leithold, E.L., and Blair, N.E., 2001. Watershed control on the carbon loading of marine sedimentary particles. Geochimica et Comsmochimica Acta, 65:2231- 2240. Liu, J.T., Yuan, P.B., and Hung, J.-J., 1998. The coastal transition at the mouth of a small mountainous river in Taiwan. Sedimentology, 45:803-816. Lui, J.T., Huang, J.-S., Hsu, R.T., and Chyan, J.M., 2000. The coastal depositional Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 125 system of a small mountainous river: a perspective from grain-size distributions. Marine Geology, 165:63-86. Liu, J.T., Chao, S.-y., and Hsu, R.T., 2002. Numerical modeling study of sediment dispersal by a river plume. Continental Shelf Research, 22: 1745-1773. Ma, Y., Wright, L.D., and Friedrichs, C.T., in prep. Observations of sediment transport on the continental shelf off the mouth of the Waiapu River, New Zealand: evidence for current-supported gravity flows. To be submitted to Continental Shelf Research. Milliman J. D. and Meade, R. H., 1983. World-wide delivery of river sediment to the oceans. Journal of Geology, 91:1-21. Milliman, J. D. and Syvitski, J.P.M., 1992. Geomorphic/tectonic control of sediment discharge to the ocean; the importance of small mountainous rivers. Journal of Geology, 100(5): 525-544. Milliman, J.D., Kao, S.-J., 2005. Hyperpycnal Discharge of Fluvial Sediment to the Ocean: Impact of Super-Typhoon Herb (1996) on Taiwanese Rivers.The Journal of Geology, 113:503-516 Morehead, M.D., Syvitski, J.P., 1999. River-plume sedimentation modeling for sequence stratigraphy; application to the Eel margin, Northern California. Marine Geology, 154:29-41. Mulder, T. and Syvitski, J. P. M., 1995. Turbidity currents generated at river mouths during exceptional discharge to the world's oceans. Journal of Geology 103: 285-299. Mullenbach, B.L., Nittrouer, C.S., Puig, P., and Orange, D.L., 2004. Sediment Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 126 deposition in a modem submarine canyon: Eel Canyon, northern California. Marine Geology, 211:101-119. Ogston, A.S., D.A. Cacchione, D.A., R.W. Sternberg and G.C. Kineke, 2000. Observations of storm and river flood-driven sediment transport on the northern California continental shelf. Continental Shelf Research, 20: 2141-2162. Pullen, J., Doyle, J.D., Hodur, R., Ogston, A., Book, J.W., Perkins, H., and Signed., R., 2003. Coupled ocean-atmosphere nested modeling of the Adriatic Sea during winter and spring 2001. Journal of Geophysical Research, 108: CIO, 3320. Scully, M.E., C.T. Friedrichs and L.D. Wright, 2003. Numerical modeling results of gravity-driven sediment transport and deposition on an energetic continental shelf: Eel River, Northern California. Journal of Geophysical Research (C4): 17-4-17-14. Scully, M.E., Friedrichs, C.T., and Wright, L.D., 2002. Application of an analytical model of critically stratified gravity-driven sediment transport and deposition from the Eel River continental shelf, northern California. Continental Shelf Research, 22:1951-1974. Shapiro, R., 1970. Smoothing, filtering, and boundary effects. Review of Geophysics and Space Physics, 8:359-387. Shchepetkin, A. F., and McWilliams, J. C., 2003. A method for computing horizontal pressure-gradient force in an oceanic model with a nonaligned vertical coordinate, Journal of Geophysical Research, 108(C3): 3090. Sherwood, C.R., Butman, B, Cacchione, D.A., Drake, D.E., Gross, T.F., Sternberg, R.W., Wiberg, P.L., and Williams, A.J., III., 1994. Sediment-transport events on Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 127 the northern California continental shelf during the 1990-1991 STRESS experiment. Continental Shelf Research, 14:1063-1099. Sommerfield, C.K. and Nittrouer, C.A., 1999. Modem accumulation rates and a Sediment budget for the Eel shelf: a flood-dominated depositional environment. Marine Geology, 154: 227-242. Sternberg, R.W., and Larsen, L.H., 1976. Frequency of sediment movement on the Washington continental shelf: a note. Marine Geology, 21:M37-M47. Sternberg, R.W., 1986. Transport and accumulation of river-derived sediment on the Washington continental shelf, USA. Journal of Geological Society, London 143: 945-956. Syvitski, J.P., and Morehead, M.D., 1999. Estimating river-sediment discharge to the ocean: Application of the Eel margin, northern California. Marine Geology, 154:13-28. Traykovski, P., Wiberg, P.L., and Geyer, W.R., in press. Observations and modeling of wave-supported sediment gravity flows on the Po prodelta and comparison to prior observations from the Eel shelf. Continental Shelf Research. Traykovski, P., W.R. Geyer, J.D. Irish and J.F. Lynch, 2000. The role of wave-induced density-driven fluid mud flows for cross-shelf transport on the Eel River continental shelf. Continental Shelf Research, 20: 2113-2140. van Rijn, L.C., 1984. Sediment transport, part II: suspended load transport, ASCE J. Hydr. Engr., 110(11):1613-1638. Wadman, H.M., and McNinch, J.E., 2006. Inner shelf processes affecting shelf-wide sedimentation, Waiapu River, New Zealand. Ocean Sciences Meeting, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Honolulu, Hawaii. Abstract OS16A-33. Walling, D.E., Webb, B.W., 1996. Erosion and sediment yield: a global overview. IAHS Publication 236, 3 - 19. Walsh, J.P., Nittrouer, C.A., Palinkas, C.M., Ogston, A.S., Sternberg, R.W., and Brunskill, G.J., 2004. Clinoform mechanics in the Gluf of Papua, New Guinea. Continental Shelf Research, 24:2487-2510. Walsh, J.P. and Nittrouer, C.A., 2003. Contrasting styles of off-shelf sediment accumulation in New Guinea. Marine Geology 196:105-125. Wheatcroft, R.A., and Drake, D.E., 2003. Post-depositional alteration and preservation of sedimentary event layers on continental margins, I. The role of episodic sedimentation. Marine Geology, 199:123-137. Wiberg, P.L., Smith, J.D., 1983. A comparison of field data and theoretical models for wave-current interactions at the bed on the continental shelf. Continental Shelf Research, 2:147-162. Wiberg, P.L., Drake, D.E., Cacchione, D.A., 1994. Sediment resuspension and bed armoring during high bottom stress events on the northern California inner continental shelf: measurements and predictions. Continental Shelf Research, 14:1191-1219. Wright, L.D., Wiseman, W.J., Bomholdt, B.D., Proir, D.B., Suhayda, J.N., Keller, G.H., Yang, Z.-S., and Fan, Y.B., 1998. Marine dispersal and deposition of Yellow River silts by gravity-driven underflows. Nature, 332:692-632 Wright, L.D., Ma ,Y., Scully, M, Friedrichs, C.T., 2006.0bservations of Across Shelf Sediment Transport During High Energy Flood Events off the Mouth of the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 129 Waiapu River, New Zealand Eos Trans. AGU, 87(36), Ocean Sci. Meet. Suppl., Abstract OS 11L-03. Zavatarelli, M., and Pinardi, N., 2003. The Adriatic Sea Modeling System: A nested approach. Annales Geophysicae. Mediterranean Forecasting System Pilot Project, 21:345-364. Zhang, Y., Swift, D.J.P., Fan, S., Niedoroda, A.W., and Reed, C.W., 1999. Two- dimensional numerical modeling of storm deposition on the northern California shelf. Marine Geology, 154:155-167. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 130 Tables Instrument Instrument Function 40 m tripod 60 m tripod Name Acronym Acoustic ADV Current velocities 35.5 cmab 50.5 cmab Doppler to 59 cmab to 66.5 Velocimeter cmab Pulse-Coherent PCADCP Velocity profiles near 69 cmab 78 cmab Acoustic the bed Doppler Current Profiler Acoustic ADCP Upward looking, 232 cmab 225 cmab Doppler Current velocity profile from Profiler the surface to 5 m above the bed Acoustic ABS Near-bed suspended 87 cmab 70 cmab Backscatter sediment concentration YSI meter Turbidity, 43 cmab 99 cmab temperature, and salinity Sediment Trap Sediment collection 89 cmab N/A for instrument 90 cmab calibration Table 1. Instruments on the 40 m and 60 m tripods. Data was provided by Carl Friedrichs, Peony Ma, and Don Wright (Wright et al., 2006, Ma et al., in prep.). Centimeters above the bed is abbreviated as cmab. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 131 Flood Statistics Fluvial Mean Along- Mean Across- Mean wave Sediment shelf current shelf current height (MT) (cm/s) (cm/s) (m) Stormy Period A 6.13 6.0 -4.2 1.5 Stormy Period B 5.57 -4.8 1.2 1.6 Stormy Period C 5.83 2.5 -1.8 2.1 Simulation 17.83 6.7 -4.6 1.7 (5/16-7/10/2004) Table 2 Mean current and wave conditions during each stormy period, and the amount of sediment delivered to the shelf during these times. Mean simulated currents were calculated from the depth-averaged currents shown in Figure 2. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 132 Depth A .veraged Hicks Bay Winds Gisborne Winds Tripod Mean 6.3 -2.2 4.4 7.7 -0.2 -0.4 Std. dev. 22.5 4.8 5.3 10.7 8.5 1.2 Minimum -36.4 -16.8 -9.2 -22.3 -17.9 -3.0 Maximum 101.2 6.3 12.8 24.6 19.9 3.5 Correlation 0. 16 0. 18 winds: Correlation 0.34 0.52 0.35 -0.13 observed Currents: Table 3. Mean, standard deviation, and range of the depth- and time-averaged currents, in cm/s, for the tripod and the simulations using Hicks Bay and Gisborne winds. Low- pass filtered data was used from the ADCP for ~ 8 to 56 m depth in the water column. The currents from the two simulated wind cases used sigma layers 3 -7, corresponding with ~6 m to ~54 m. Negative current values in the minimum current velocity row indicate currents velocities in the southwest and northwest, or landward, direction in the along- and across-shelf directions, respectively. All correlation coefficients were determined using the low-pass filtered observed and simulated data. The correlation coefficient for the winds was determined for the observed current data compared with the observed Hicks Bay winds and Gisborne. In the bottom row, the simulated depth- averaged currents in the along- and across-shelf directions were correlated with the observed depth averaged currents. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 133 Bed shear stresses 40 m site 60 m site Tripod Observations 0.49 ±0.41 0.51 ±0.46 Simulation 60m tripod waves; 1.11 ± 1.17 0.52 ±0.55 Hicks Bay winds Simulation 60 m tripod waves; 1.05 ± 1.16 0.48 ±0.54 Gisborne winds Simulation 40 m tripod waves; 1.55 ±1.28 0.63 ± 0.48 Hicks Bay winds Simulation 40 m tripod waves; 1.49 ± 1.25 0.59 ± 0.47 Gisborne winds Table 4. Mean and standard deviation bed shear stresses in Pascals (Pa) for each wind case compared to the tripod observations at each location. The 40 m and 60 m wave records were used to simulate bed shear stresses at each location and for each wind case. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 134 Observed Tripod Currents Model Currents Wbl wbl Along Shelf Across Shelf Along Shelf Across Shelf Mean 6.03 1.2 -1.2 1.8 Range -18.0 to 36.5 -15.0 to 22.7 -10.1 to 1.3 -4.1 to 8.6 Correlation with observed -0.59 0.59 wbl currents Correlation of wbl currents 0.64 0.60 -0.77 0.68 with 'tbed U gravity -0.53 -0.91 Correlation of UgravWith Cs 1 -0.06 0.17 mab Correlation of -0.20 0.26 UgravWith Tbed Table 5. Mean, standard deviation, and range of the time-averaged wave boundary layer (WBL) currents in cm/s for the 60 m tripod and the simulation. Observed values were taken from the PCADCP instrument at 10 cm above the bed and were low-pass filtered. The currents from the model were taken from the wave boundary layer, whose thickness is determined by the wave height record. Negative current velocities indicate southwest and northwest (landward) currents in the along- and across-shelf directions, respectively. The simulated wave boundary layer currents were compared with the observed low-pass filtered currents. Both the observed and simulated currents in the wave boundary layer were compared with the bed shear stress record. Calculated gravity currents, estimated by subtracting the currents at 1 mab from the currents in the wave boundary layer, were compared with the bed shear stress and sediment concentration (1 mab at the 60 m tripod) records. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 135 Tripod Currents Model Currents 1 mab 1 mab Along Shelf Across Shelf Along Shelf Across Shelf Mean 2.9 0.8 1.4 4.4 Std. dev. 7.1 6.0 1.6 6.4 O 1 Range -15.6-20.9 -13.1-19.9 1 -12.9-14.2 Table 6. Comparison of observed and estimated time-averaged currents 1 m above the bed. Tripod data came from the ADV and was low-pass filtered. Negative current velocities indicate southwest and northwest, or landward, currents in the along- and across-shelf directions, respectively. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 136 Sediment Budget % in terms of mass (kg)/sediment input up to time x Time averaged and final 100% floe 75% floe, 25% floe, 100% unfloc 25% unfloc 75% unfloc Mean Suspended 12.2 15.4 21.1 23 Sediment Mean Wave 8 6.7 3.4 0.7 Boundary Layer Mean Bed 35.8 24.3 5.5 0.5 Sediment Mean Sediment on 56 46.4 30 24.2 the Shelf Range Suspended 0.1-79.4 0.1-80.9 0.1-83.5 0.02-84.7 Range WBL 0.3-28.9 0.2-23.6 0.1-12.5 0-3.9 Range Bed 1.9-56.1 0.7-40.4 0.1-11.9 0-3.4 Final Suspended 18.2 13.3 4.8 0.06 Sediment Final Wave 1.1 0.7 0.3 0.01 Boundary Layer Final Bed 32.9 23.6 4.4 0.2 Sediment Final Sediment on 52.2 37.6 9.5 0.3 the Shelf Table 7 Time-averaged sediment budgets for each aggregate composition case expressed in percent. Time-averaged values were calculated by taking the mean of the ratio of the mass (in suspension, in the wave boundary layer, or deposited on the seabed) and the sediment delivered to the shelf up until that time. The final ratios of mass in suspension, within the wave boundary layer, and on the seabed were taken from the final time-step of the simulation. The final sediment deposit percent of the total fluvial input was calculated by dividing the amount of sediment deposited on the bed at the end of the simulated period by the 24 Mtons of sediment that was delivered by the river during the simulation. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 137 Sediment Flux Budget Transport Mechanism Aggregate Properties % % Total WBL% Water Flocculate Unfloccula Transport Column% d % ted% (kg) Into the 1.5 x 101U 24.5 75.5 78.7 21.3 Grid Out of the 9.2 x 10y 0.7 99.2 52.9 47.1 Grid North 9.3 x 10y 0.8 99.2 53.3 46.7 Boundary South -7.3 x 107 3.3 96.7 96.5 3.5 Boundary East 1.5 x 107 0.03 99.9 0.9 99.1 Boundary Table 8. Flux of sediment across the open boundaries of the model. Transport mechanism percents were calculated by taking the amount of sediment crossing a boundary for each transport mechanism and dividing by the total mass transported across the boundary. The proportion of flocculated and unflocculated sediment transported across each boundary was similarly calculated. The negative value for the southern boundary indicates net transport to the north. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 138 Sediment Budget Across-shelf % WBL Suspended Bed Mean Range Mean Range Mean Range Final Inner 4.8 0.04-23.2 8.6 0.02-80.1 5.4 0.08-27.2 0.08 Shelf Mid- 1.4 0-9.6 4.3 0-12.0 11.3 0-25.0 9.92 Shelf Outer 0.4 0-4.9 2.6 0-9.6 7.7 0-19.7 13.60 Shelf Slope 5.2e'v 0-0.7e'3 0.02 0-0.4 0.01 0-0.05 0.05 Break Table 9. Vlean percent and range of sediment mass in suspension, in the wave boundary layer, and deposited on the bed for each shelf zone. Values were calculated by dividing the mass by the sediment delivered by the Waiapu River at each time step and then taking the mean and range. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 139 Relative Importance of sediment transport mechanism Transport 100% floe 75% floe, 25% floe, 100% unfloc Mechanism % 25% unfloc 75% unfloc Mean Suspended 59.9 70 82.3 92 Sediment Mean WBL 40.2 30 17.4 7.6 Sediment Range Suspended 6.4-99.8 11.5 - 99.8 33.1 -99.8 53.5-99.9 Sediment Range WBL 0.2-93.6 0.2 - 88.5 0.2-66.9 0.1 -46.6 Sediment Mean Suspended 100 67.1 36 0 Sediment, Flocculated Mean WBL 100 98.3 88.2 0 Sediment, Flocculated Table 10. The relative importance of the two sediment transport mechanisms on the shelf in terms of the time-averaged mass transported in suspension or within the wave boundary layer. Values were calculated by dividing the amount of sediment either in suspension or within the wave boundary layer by the total amount of sediment in both the water column and the wave boundary layer for each time-step and then taking the average. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 140 Figure Captions Figure 1. East Cape, North Island, New Zealand. Inset shows the North Island and a regional map of East Cape. The regional map shows the locations of the Waiapu River and catchment, Hicks Bay, Gisborne, and the Tatapouri NIWA buoy. The Waiapu River shelf map shows the river mouth location and shelf bathymetry. Two VIMS tripods were deployed at the 40 m and 60 m isobaths in 2004, as described in Wright et al. (2006) and Ma et al. (in prep). Figure 2. A. The Waiapu River discharge record, observed wave heights, Hicks Bay winds, and the observed depth-averaged along and across shelf currents during the field experiment. The Waiapu River discharge record and Hicks Bay winds are hourly measurements from the Gisborne District Council (unpublished data). Hourly wave height records are for both the 40 m and 60 m tripod and were smoothed using a low- pass filter. The depth averaged currents were collected by the ADCP on the 60 m tripod (Wright et al., 2006; Ma et al., in prep). Positive is to the north-east and negative is to the south-west for along-shelf currents. Positive across-shelf currents flow offshore, whereas negative across-shelf currents flow landward. Periods characterized by high fluvial discharge are delineated by the gray boxes. B. The sediment rating curve for the Waiapu River (Hicks et al., 2004). Figure 3. A. Bathymetry of the Waiapu River shelf and the area to be simulated. B. The model grid. The white grid delineates the boundaries used to analyze sediment flux on the shelf. The A line corresponds to about 28 m depth, the B line with 47 m, the C line Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 141 with 70 m, the D line with 106 m and the E line with 182 m depth. Each of these sections is delineated as the inner shelf, middle shelf, outer shelf, and shelf slope/break. All calculations presented in this paper refer to the area outlined by the white grid. Model bathymetry was smoothed and maximum depths were held to 200 m. Figure 4. Average wind direction and speed for both Hicks Bay and Gisborne during the deployment period and during floods, which occur when discharge is greater than base flow. Wind direction shows from where winds are blowing. Figure 5. A. Current magnitude for the Hicks Bay (red) and Gisborne (blue) winds versus observed (black). B. Observed depth averaged currents (black) from the ADCP at the 60 m tripod and simulated depth averaged currents (red) using Hicks Bay winds. Surface currents and currents near the bed were not included in the calculations. Positive values indicate northeast flowing currents and offshore currents in the along- and across-shelf directions, respectively. Figure 6. The first panel shows observed and simulated bed shear stresses. Observed bed shear stresses were calculated using a 1-dimensional model approach by Wiberg et al., (2001), using wave orbital velocities measured by ADV sensors and a bed roughness of 0.15 cm (Ma, pers. comm.). In this figure, the observed bed shear stresses were smoothed with a low-pass filter. The ECOMSED model used the Grant-Madsen wave-current model to calculate the bed shear stresses due to combined waves and currents (Grant and Madsen, 1979). Measured sediment concentrations at 1 m above Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 142 the bed (black) versus the simulated estimates (red) at the 60 m tripod site in the second panel. Calibrated YSI measurements were used to calculate the observed sediment concentrations, which were then low-pass filtered (Ma, unpublished data). Model estimates of sediment concentration are from the sigma layer directly above the wave boundary layer (~ lm). Gravity-driven currents are shown in panel three. The fourth and fifth panels show the smoothed 60 m tripod wave height record and the Waiapu River discharge record. Figure 7. A. Depth averaged currents (red), wave boundary layer currents (blue), and calculated gravity-driven velocities (red) from the ADCP, ADV, and PCADP observations, respectively. B. The second panel shows velocities from the simulation. Across-shelf gravity-driven velocities were calculated by subtracting the current velocities at ~1 mab from the current velocities near the bed. C. The observed and simulated currents within the wave boundary layer. Positive values indicate seaward current velocities, whereas negative values indicate landward current velocities. Figure 8. A. The mass of sediment deposited, in the wave boundary layer, and in suspension calculated for each time step. The top panel shows the total sediment in the model domain (black dash), the amount of sediment in suspension (red), the wave boundary layer (blue), and on the bed (green), compared to the cumulative fluvial input (black). B. Suspended sediment, wave boundary layer sediment transport and deposition for high settling velocity sediment (flocculated, black) and low settling Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 143 velocity sediment (unflocculated, black dash). The mass of flocculated and unflocculated sediment in each grid cell was calculated for each time step. Figure 9. Planview of sediment deposition, sediment transport in the wave boundary layer, suspended sediments, and freshwater plume estimated for varying strong along- shelf current directions. Figure 10. A. The cumulative amount of sediment exported from the shelf during the simulation. The top panel shows the amount of sediment that was exported in dilute suspension (blue) versus within the wave boundary layer (red). Cumulative Waiapu River input is delineated by the black line. The bottom panel shows the cumulative export of flocculated (black line) and unflocculated sediment (dashed black line). Export was calculated by summing the amount of sediment that crossed the North, South, and East boundaries (white lines in Fig. 1). B. Flux across the each of the model boundaries during the simulation in the top panel. The second panel shows flux out of the model domain for unflocculated and flocculated sediments. Positive values indicate flux out of the model grid across each boundary, and negative flux values indicate flux into the model grid area. Figure 11. Planview of the Waiapu River shelf with sediments deposited on the bed and sediments in the wave boundary layer and the depth averaged suspended sediment concentrations. Sediments in the wave boundary layer and in suspension are expressed as the thickness of the resulting sediment deposit if the sediments were deposited on the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 144 bed. The A panel shows the typical sediment depositional and transport patterns during a flood with energetic waves. The B panel shows typical sediment depositional and transport patterns during a flood with low energy waves. The discharge and wave record are referenced below. Figure 12. A. The time-averaged along- and across-shelf flux for each along-shelf row (see Fig. 3) of the model grid. The results are plotted versus depth across the shelf for total flux, flux within the wave boundary layer, and suspended flux. Flocculated and unflocculated flux components are shown only for the suspended flux because flux within the wave boundary layer was about 99% flocculated sediment. B. The time- averaged vertical distribution of flocculated (black) and unflocculated (black dash) sediments at the 60 m tripod. Figure 13. The correlation coefficient of along- and across-shelf flux in each along- shelf row with the smoothed wave and simulated depth-averaged current records were calculated for across-shelf suspended and gravity-driven transport mechanisms. The bottom panel shows the along-shelf suspended sediment correlation with winds and currents with depth across the shelf. Figure 14. A. Time-averaged across-shelf suspended and gravity-driven flux with depth across the shelf for the model simulation, the flood with low waves (flood B) and the flood with high waves (flood C). The bottom panel shows the time-averaged suspended along-shelf flux to the north of the Waiapu River mouth with depth across the shelf for Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 145 each period. Along-shelf flux was calculated for each along shelf row from the river mouth to the North Boundary identified in Fig. 3. B. The time-averaged vertical distribution of flocculated (solid) and unflocculated (dashed) sediments at the 60 m tripod for the simulation and the two floods. Figure 15. Sediments in suspension (red), in the wave boundary layer (blue), and deposited on the bed (green) on the inner shelf, mid-shelf, outer-shelf, and shelf break area. The values were calculated for the grid areas outlined in figure 3 for each time step. Figure 16. Using a modified version of the Exner equation, we can show the time- averaged flux gradients along-and across-shelf (A), across-shelf suspended and wave boundary layer transport (B), and across-shelf flocculated and unflocculated flux convergence. Negative numbers indicate flux convergence and deposition, whereas positive numbers indicate flux divergence and erosion. Figure 17. A. Time-series record of sediment in suspension, the wave boundary layer, and deposited on the seabed for various combinations of aggregate composition: 100% flocculated (black), 100% unflocculated (black dash), 75 % flocculated and 25 % unflocculated (red), and 25% flocculated and 75% unflocculated (blue). The mass of sediment in each grid cell was calculated for each time step. B. The total amount of sediment exported from the shelf area for each combination of flocculated and unflocculated sediment during the simulation. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced Hicks Bay I 178°30'00"E 178°40'00"E 178°50'00"E _____ SuOO.Ot'oZX I ____ 147 2000n Waiapu River Discharge Waiapu Shelf Waves - - 40m tripod 60 m tripod JO*/, Hicks Bav Winds Shelf Currents 80 d - - across shelf 60 along shelf to 40 £ 20 5/16 5/24 6/01 6/09 6/17 6/25 7/03 Date 2004 100000 10000 1000 cn 100 10 1 10 100 10001000 00.001 0.01 0.1 1 10 100 1000100000.001 Q (m3/s) Fig. 2 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. \45 % X ,S\°' lVoW• ^ edxH' ^ GV tf**0' fu',<\PeV ^et ;V^S‘,S \° ° ,e^ ^ep*0'$£& ■t*. © 0-5 5-10 I>15 - 20 - I>15 ■> ■> I Wind Speed (m/s) 90 90 * 0 0 180 180 Gisborne Storm Winds Hicks Bay Storm Winds -I'm 180 180 Gisborne Winds Hicks Bay Winds Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 150 Depth-averaged current magnitude 120 Hicks Bay Winds Tripod Gisborne Winds 100 £ 60 40 20 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Date 2004 Depth-averaged along-shelf currents Tripod Model 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Depth-averaged across-shelf currents 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Date 2004 Fig. 5 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 151 Bed Shear Stress i i | 11 Model co 2 Tripod A Q. O' 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Sediment Concentration ~1mab 60 m Tripod Site I ..... I1 I I ".... ""T1-—■""" 2 "8>1 ... Laa 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Gravity-Driven Currents 20 ■ ■ ■ ■ ■ E0 iv j-Jtei J f r \ P f U V A ^ a / V o i g>--- tk/VN nr -20' 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Wave Height , i,1——■■■ T...... I" E2 ■ ■ A A A Vh\ivr Al //V V- u ■ ■ N 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Fluvial Suspended Sediment Concentration 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Date 2004 Fig. 6 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 152 1 mab Depth-averaged A. Observed Depth-Averaged, 1 mab, and Near Bed Currents 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 B. Simulated Depth-Averaged, 1 mab, and Near Bed Currents 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 i a i a a a i a ■ 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Date 2004 Fig. 7 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 153 A. Sediment Distribution on the Shelf Suspended Wbl Deposited - - Total Sediment Fluvial 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 B. Suspended Sediment Flocculated • 2 Unflocculated 4-Jo 2 1 O' 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Wave Boundary Layer Sediment 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Bed Sediment o4 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Date 2004 Fig. 8 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 154 178.4 178.4 178.5 178.6 178.4 178.5 178.6 178.4 178.5 178.6 w Fig. 9 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 155 A. Cumulative Sediment Export and Transport Meachanism (0 — River Input oc — Suspended S 10 WBL 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Cumulative Sediment Export and Sediment Aggregate Properties (0 — Flocculated c o - - Unflocculated 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 B. North Shelf Sediment Export 20000 East South 10000 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Export Rates by Sediment Aggregate Properties 20000------1------1— c—i 1 i * —i ia ? i i *i ------Flocculated 10000 Unflocculated 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Fig. 10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 156 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced Flux(kg/s) Flux (kg/s) ■1000 2000 3000 1000 200 100 ■100 150 50 0 C B. B. A. 20 0 0 0 0 0 10 140 120 100 80 60 40 20 Vertical Sediment Distribution of Sedimentsof DistributionVerticalSediment 40 o 0 2 - £ - -40 8- Time-averaged AcrossShelf Flux 0 0 0 10 140 120 100 80 60 ' T Time-averaged AlongShelf Flux -60 Depth Across-shelf (m) 0 DepthAcrossshelf (m) 5 100 50 0 Sediment (mg/L) Sediment .... unfl suspended suspended — total wbl — ~~ floesuspended total wbl floe suspended suspended suspended unfl suspended 6 180 160 160 8 200 180 i. 12 Fig. 200 157 158 Across-shelf Suspended Flux Correlation with Waves and Currents " 1 I I " *l~ ...... I------T — ------1------1 I I---- 0.6 0.4 0.2 - 0.2 ■*-»c ) 40 100 120 140 160 180 200 - 0.2 40 100 120 140 160 180 200 O Along-shelf Flux Correlation with Waves and Currents O I I "—T...... I------1------1 I- 0.6 waves 0.4 currents 0.2 - 0.2 100 120 140 160 180 200 Depth Across-shelf (m) F ig.13 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 159 A. Time-averaged Across Shelf Suspended Flux 2000 i 1 *1 i i r i 1000 o> -1000 20 406 0 8 0 100 120 140 160 1 8 0 200 Time-averaqed Across Shelf WBL Flux 8 0 0 0 TTTTI T I I ' I.. T to 6 0 0 0 S’ * 4 0 0 0 2000 0 20 4 0 6 0 8 0 100 120 1 4 0 1 6 0 1 8 0 200 Time-averaqed ...... Alonq ** Shelf p - Flux Simulation FloodB FloodC 60 80 100 120 140 200 Depth Across-shelf (m) B. Vertical Sediment Distribution of Sediments Flocculated Sediments Unflocculated Sediments 0 11 -20 -20 < / ■ i \ a> -40 £ -40 Q Q -60 -60 A. 0 50 100 0 50 100 Sediment (mg/L) Sediment (mg/L) F ig .14 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ON o i Date 2004 Date Inner Shelf Inner Outer Shelf Shelf Shelf Break Middle Middle Shelf bed sediment bed suspended suspended sediment wave sedimentboundary layer ■ 05/16 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 05/16 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 05/16 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 05/16 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 o2 w o2 02 0.02 00.01 t—‘ 21 U\ CTQ* Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 200 200 200 180 180 gravitydriven total suspended - alongshelf — — acrossshelf — flocculated total unflocculated 140 160 140 160 180 120 120 100 100 80 100 120 140 160 Shelf Flux Convergence Depth Acrossshelf (m) 40 60 80 20 0 0.2 0.2 0.2 0.2 0.2 0.2 -0.4- - - - -0.4 -0.4 - x a> o > x 3 c ® o> W o 8 ■o 1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 162 A. Sediments in Suspension 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Wave Boundary Layer Sediments — 75fl25unfl 25fl75unfl 100floc - - 100unfl 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Deposited Sediments 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Date 2004 Cummulative Sediment Export 18 16 75% floe 25% unfl 25% floe 75% unfl 100% floe 14 - - 100 %unfl 0)c o 05/16 05/21 05/26 05/31 06/05 06/10 06/15 06/20 06/25 06/30 07/05 07/10 Date 2004 Fig. 17 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 163 Chapter 3: Sediment transport pathway sensitivity and strata-formation on the Waiapu River Shelf, New Zealand I. Abstract The Waiapu River has a small, mountainous catchment that delivers approximately 35 MT/y of suspended load to a narrow collision margin shelf that is subject to energetic waves and currents. Recently, observational and theoretically- based data from the shelf indicated that multiple sediment transport mechanisms contribute to sediment dispersal and fine-scale strata formation. These transport mechanisms include dilute suspensions, gravity-driven transport, and resuspension into the water column. High apparent accumulation rates, up to 3 cm/y, enable preservations of the depositional products of multiple sediment transport mechanisms in the sediment record. Geochemical data revealed pulsed event layers with terrestrial signatures inter bedded with layers of marine pelagic origin. A two-month long simulation of the Waiapu River shelf using observed winds, river discharge, and wave records from a field experiment in 2004 showed that depositional patterns on the Waiapu River shelf were influenced by aggregation of fluvial material, fluvial discharge, wind-forced currents, and resuspension and gravity flow by energetic waves. In this paper, a series of experiments was conducted to assess transport mechanisms and depositional sensitivity to these parameters. Three factors were considered in the analysis: average Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 164 sediment mass distribution between water column, wave boundary layer, and seabed; average sediment flux patterns; and flux convergence and divergence patterns. Sediment transport and depositional patterns were sensitive to wind-forced currents, waves, sediment load, sediment rating curve, and flocculation to varying degrees. Strong along-shelf wind-forced currents restricted sediments to the inner shelf, whereas strong along shelf currents coupled with energetic waves resulted in sediment export from the shelf area. Wind patterns that did not produce sustained along-shelf currents resulted in greater across-shelf flux, especially within the wave boundary layer. Although increasing the mean bed shear stress due to currents reduced retention on the shelf, increasing mean bed shear stress due to waves resulted in greater across-shelf flux within the wave boundary layer and deposited sediments at greater water depths. The fraction of the fluvial load that was flocculated greatly influenced sediment retention on the shelf. Unflocculated sediments, with low settling velocities, tended to be exported from the shelf. Flocculated sediments, with higher settling velocities, tended to be deposited on the shelf. Almost all of the sediments transported within the wave boundary layer were flocculated. Both flocculated and unflocculated materials were transported in dilute suspension. A zone of high across- and along-shelf flocculated flux was caused by resuspension on the mid-shelf. Increasing the sediment load and the sediment concentration for a given freshwater discharge increased the percentage of retention on the shelf after a flood. Simulations that assumed lower sediment loads represented fluvial conditions before massive deforestation, about 150 years ago. For these pre-deforestation cases, a higher percentage of the sediment load was transported within the wave boundary layer, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. suggesting that the flux within the wave boundary layer reaches capacity before transport within the overlying water column. This also suggests that wave-induced gravity-driven transport may have been even more important for across-shelf flux before deforestation than present day. All of the simulations indicated that although short-term deposition rates were high on the inner shelf, most of the sediments deposited there were resuspended and deposited on the mid- to outer-shelf where they were retained. Overall, sediment transport was most sensitive to the partitioning of sediment into flocculated and unflocculated types, and sediment deposition was most sensitive to current and wave energy. II. Introduction: Small river systems with mountainous, highly erodible catchments deliver a significant portion of the global terrestrial budget to the coastal ocean and are typically found along collision margins (Milliman and Syvitski, 1992). Most of the fluvial load of these rivers is dispersed to shelf waters during floods (Milliman and Syvitski, 1992; Farnsworth and Milliman, 2003). When flood loads from these rivers debouche onto shelf waters that are subjected to energetic waves and currents, the fluvial sediments are dispersed by multiple sediment transport mechanisms (Mulder and Syvitski, 1995, Kinkeke et al., 2000; Traykovski et al., 2000, Friedrichs and Wright, 2004). Several sediment dispersal and transport mechanisms have been identified: buoyant, or hypopycnal, river plumes, negatively buoyant, or hyperpycnal, plumes, wave- and current supported gravity-driven flows, and wave- and current-generated dilute suspensions (Mulder and Syvitski, 1992, Kineke et al., 1996; Kineke et al., 2000; Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 166 Traykovski et al., 2000; Milliman and Kao, 2005). The relative importance of these transport mechanisms can influence residence times of sediment in various parts of the marine sedimentary system and thereby impact the exchange and burial of important contaminants and nutrients, benthic community structure, and sedimentary strata formation (Schaffner et al., 1987; Sommerfield and Nittrouer, 1999; Dukat and Kuehl, 1995; Leithold and Blair, 2001; Goni et al., 2006). Fluvial sediments are often delivered to shelves within buoyant plumes, but may be transported by negatively buoyant plumes when fresh water sediment concentrations exceed ~40g/L (Mulder and Syvitski, 1992). Sediments in dilute suspensions may be dispersed by waves and currents or may be concentrated sufficiently to form a near-bed mud layer that is driven down-slope by gravitational forces (Friedrichs and Wright, 2004). After initial deposition, sediments may be resuspended into the water column and subsequently redistributed by energetic waves and currents. Although resuspended sediments are diffused throughout the bottom boundary layer, stratification at the top of the wave boundary layer can trap sediment below it, so that sediment concentrations there are sufficient to form a near-bed, fluid mud layer (Harris et al., 2005). Quantifying the various mechanisms that contribute to sediment transport, resuspension, and burial will improve to our understanding of fine-scale strata formation and reworking on continental shelves. This study focuses on the continental shelf adjacent to the Waiapu River, New Zealand, an area that experiences energetic waves, strong currents, and high fluvial input (Fig. 1). It is an ideal setting to investigate multiple sediment transport mechanisms including plume delivery, resuspension, and gravitationally-driven flux. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Although fluvial delivery is episodic and flood-driven, high rainfall rates ensure flooding several times a year (Fig. 2). A field study conducted in 2004 revealed that both waves and currents were energetic during storms, with near-bed current velocities up to 60 cm/s during periods of current- and wave-supported gravity transport (Wright et al., 2006, Ma et al., in press). The presence of pulsed event layers in shelf sediments, 0 1 ft n defined by low excess Pb activities, terrestrial 5 C values, and terrestrial C/N ratios, indicated that flood sediments are rapidly transported across the shelf and to the shelf break and are distinctly different than non-flood or reworked sediments (Addington et al., in press, Kniskem, Chapter 1). High apparent accumulation rates, averaging 1.5 ± 0.9 cm/y with a range of 0.2 -3.2 cm/y, on the mid- to outer-shelf contributed to preservation of these pulsed event layers and inhibited bioturbation (Kniskem, Chapter 1; Kniskem et al., 2006). From these data, a shelf sediment budget was constmcted, indicating that an estimated 24% of the fluvial budget over the last 100 years has been preserved on the shelf (Kniskem, Chapter 1). A two-month simulation, using wave data from a 2004 field deployment and wind data from nearby Hicks Bay (Fig. 1), showed that although sediments were rapidly deposited on the shelf over short time-scales, most of the fluvial load was resuspended and transported out of the shelf area (Kniskem, Chapter 2). The simulation indicated that most of the sediment that escaped the shelf was transported to the north of the Waiapu River mouth in dilute suspension (Kniskem, Chapter 2). Sediments transported as a gravitationally-driven flow within the wave boundary layer, however, tended to be retained on the shelf. Sediment flux and patterns of erosion and deposition was used to characterize bathymetric shelf zones: inner (0-30 m), middle (30-70 m), outer (70 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 168 130 m), and shelf break (> 130 m water depth), based on unique patterns of sediment transport, deposition, and erosion. The simulation also indicated sediment transport and depositional patterns were influenced by sediment flocculation, sediment load, and wave and current conditions during peak flood. To assess the importance of these environmental forcings, several numerical experiments were conducted to explore how sediment transport and depositional patterns respond to variations in flood discharge, wave energy, wind-driven current speed and direction, and sediment aggregate properties. These sensitivity experiments allowed us to evaluate longer time-scales than the field season and to relate the observed deposition to the full range of environmental conditions. III. Objectives The sensitivity experiments were used to evaluate the following objectives: a. Assess sediment transport and depositional sensitivities to variations in sediment load, concentration, waves, wind-forced currents, and aggregation of sediments. b. Identify the combinations of conditions that are conducive to sediment preservation on the shelf. c. Assess whether anthropogenic changes in the catchment over the last 100 years resulted in changes in sediment transport pathways and accumulation patterns. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 169 IV. Background The following discusses the regional geology of the Waiapu River shelf, and shelf sediment transport mechanisms. A. Regional setting: The North Island, New Zealand, is located to the west of the collision zone between the Australian plate and the obliquely subducting Pacific plate, and defined by the Hikurangi Trough (Walcott, 1987; Lewis and Pettinga, 1993). As a result of the tectonics associated with the subduction zone, a zone of ranges and basins developed on the North Island. The Taupo Volcanic Zone lies to the west of the axial and coastal ranges that comprise the backstop and the partially uplifted forearc basin of the subduction zone (Lewis and Pettinga, 1993). Triassic to Cretaceous greywacke and argillite and Cretaceous to Pleistocene sedimentary basement rocks are found in the axial and coastal ranges, respectively (Field and Uruski, 1997). Subduction associated with the Hikurangi Trough strongly influences deposition on the Waiapu River shelf (Lewis et al., 1998; Lewis et al., 2004). Along the shelf edge, the curved Ruatoria margin re-entrant contains massive debris flows and mass failures up to 3000 km3 in volume as a result of collision with seamounts carried by the subducting Pacific Plate (Collot et al., 2001). Transverse faults forming a graben along the edge of the Ruatoria re-entrant contain up to 1 km-thick deposits of Quaternary sediments (Lewis et al., 2004). Faulting during the Tertiary and Quaternary contributed to the formation of a synclinal shelf basin with subsidence rates up to 4m/ka (Lewis et al., 2004). The Last Glacial Maximum (LGM) surface was identified at a maximum of ~220m depth on the shelf (Lewis et al., 2004), indicating that the shelf has subsided Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 170 about 100 m since that time (Gibb, 1986; Lewis et al., 2004). Post-glacial deposition was contained within this subsiding mid-shelf basin and maximum sediment thicknesses exceeded 100 m (Brown, 1995; Lewis et al., 2004). The Waiapu River, located on the East Cape of the North Island, New Zealand, drains a mountainous catchment only about 1700 km (Fig. 1) (Hicks et al., 2004). Fill and cut fluvial terraces identified within the river valley record a history of sea-level and climate change (Litchfield and Berryman, 2005). The crushed nature of the rocks in the Waiapu River basin, frequent seismic activity, high rainfall of 2400 mm/yr, and increased erosion due to anthropogenic deforestation (Page et al., 2001) has resulted in an estimated suspended sediment load of 35 MT/y. The suspended sediment yield of this catchment, 20,520 t/km2/yr, is among the highest on Earth (Griffiths, and Glasby, 1985; Hicks et al., 2000; Hicks et al., 2004). Anthropogenic deforestation began with Polynesian settlers, who started to clear the landscape of thick temperate rain forest 800-500 14C yr BP. European colonists cleared the remaining native forest for pasture between 1880 and 1920 (Harmsworth et al., 2002; Kasai et al,. 2005). Lacustrine and marine sediment accumulation rates indicate that current Waiapu River sediment fluxes to the coastal ocean may be 4-7 times greater as a result (Page and Trustrum, 1997; Foster and Carter, 1997). This significant increase in sediment flux may be reflected in the sediment accumulation rates and patterns on the shelf (Carter et al., 2002). The sediments delivered to the coast post-deforestation are also thought to have relatively higher mud content than the pre deforestation load (Kasai, 2006). These anthropogenic changes in sediment load and texture may have modified modem sediment dispersal mechanisms and accumulation Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 171 patterns, so that modern-day conditions differ from those present throughout the majority of the Holocene. High-rainfall storms play an important role in sediment transfer to the coastal ocean here because the small river basin responds rapidly to precipitation (Milliman and Syvitski, 1992; Hicks et al., 2004). Storms trigger shallow landsliding and extensive gully erosion, particularly those storms associated with subtropical cyclones, because they can activate gullies (Kasai et al., 2005). Landslides and gullies may deliver to the river an estimated 64% of the sediments eroded during these events (Page et al., 2001). However, long-term estimates show that gullying and landsliding may only account for 15% of the fluvial sediment budget (Page et al., 2001). B. Shelf sediment transport mechanisms and depositional patterns The Waiapu River is expected to achieve sediment concentrations that exceed 40 g/L and are therefore sufficiently high to produce hyperpycnal plumes during the annual winter-spring floods (Fig. 2) (Mulder and Syvitski, 1995; Hicks et al., 2004). Highest discharge occurs from June through September, and wave energy peaks from April though July. Data collected during a field experiment in 2004 showed that both waves and currents were energetic on the shelf (Ma et al., in press). Between May - July, 2004, significant wave heights were over 2 m and along-shelf, depth-averaged currents reached speeds up to 1 m/s (Ma et al., in press). Waiapu River discharge records indicated that 24 MT of terrestrial material were delivered to the coastal ocean during the field experiment (Hicks et al., 2004). Fluvial suspended sediment concentrations may have exceeded 50 g/L during two periods of high discharge (Hicks Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 172 et al., 2004) and coincided with mean near bed current velocities exceeding 50 cm/s (Ma et al., in press). Water column observations showed the fluvial sediment was dispersed on the shelf by both buoyant river plumes and gravity-driven flows within a moderately turbid layer about 1 m thick (Ma et al., in press). Gravity flows on the shelf during this period may have been sustained by both energetic waves and currents, becoming auto-suspending where the shelf slope exceeds 0.01 (Friedrichs and Wright, 2004; Ma et al., in prep.). Wave, wind, and discharge data were used as input to a three-dimensional numerical model that accounted for hydrodynamics, sediment resuspension by energetic waves and currents, and transport within a thin, near-bed fluid mud layer (Kniskern, Chapter 2; Harris et al., 2004). The simulation was able to estimate time-averaged currents, the magnitude of near-bed sediment concentrations, and accounted for 83% of the observed bed shear stresses at a tripod deployed off the river mouth at 60 m depth. The model did not account for large velocity excursions seen in the tripod data, however, because it neglected large-scale current systems offshore that may interact with shallower shelf waters (Chiswell, 2005). Calculated sediment concentrations were commensurate with observations that reached 3 g/L at 1 m above the bed. Simulated bed shear stresses agreed with observed values during most of the record, but under estimated bed shear stresses when currents were strong. The model results also showed that transport and depositional patterns on the Waiapu River shelf were influenced by settling characteristics of silts and clays, fluvial discharge record, transport pathway, and energetic waves (Kniskem, Chapter 2; Addington et al., in press). Sediments transported across the shelf within a near-bed Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. fluid mud layer were largely composed of sediments with settling velocities representative of flocculated material. Sediments with lower settling velocities, representative of unflocculated material, tended to be removed from the shelf. A significant portion of material with higher settling velocities, representative of flocculated material, was also resuspended and exported during times of energetic waves and wind-driven currents. Generally, gravity-driven transport accounted for about one-third of the total flux, though sediment flux within the wave boundary layer did exceed that of dilute suspension during a flood. Sediments transported within the wave boundary layer contributed to seabed deposition mostly on the inner and mid shelf, decreasing in importance with depth across the shelf because wave boundary layer thickness attenuates with depth. Deposition beyond ~75 m depth was due mostly to settling of sediments in dilute suspension. Numerical simulation of the tripod deployment showed that sediment depositional patterns varied across the shelf, with muddy sediments ephemerally deposited for up to several weeks on the inner shelf, between 0 - 30 m, and the mid shelf, between 30 -70 m water depth (Kniskem, Chapter 2). The model estimated that 37% of the fine sediments delivered to the shelf from May - July 10, 2004 were retained on the outer shelf, between 70 - 130 m depth. The simulation results compared favorably with seabed geochemical observations (Kniskem, Chapter 1). Excess 9 1 Pbn and Be 7 activities indicated that muddy sediments were only temporarily deposited at the boundary of the inner and mid-shelf, whereas portions of the mid- and outer shelf, between 50 - 130 m water depths, have acted as fine sediment repositories at least over the last 100 years. An Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 174 estimated 24% of the Waiapu River load over the last 100 years has been retained on the shelf, between 30 - 200 m depth. Cores collected from the shelf break and slope, >130 m, indicate that some Waiapu River sediment is delivered to deeper waters, with accumulation rates ranging between 0.5 - 1 cm/y. A significant portion of the Waiapu River sediment budget, up to 76%, is not retained on the shelf, with most of the sediment likely transported along the shelf and exported beyond the sampling area (Kniskem, Chapter 2). The combination of sediment transport mechanisms that occur here influence sediment deposition and accumulation patterns, physical and biological mixing signatures, and sediment retention on the Waiapu River shelf. There are several motivations for investigating sediment transport sensitivities. Foremost, the simulation and field observations only evaluated sediment transport and deposition over a two month period, making it difficult to compare these results with seabed observational data that accounts for the last 100 years of deposition. Some of the sensitivity experiments therefore address floods of greater magnitude than those observed during the field deployment. The sediment rating curve was also adjusted to evaluate sediment transport and deposition under estimated pre-deforestation conditions (Page and Trustrum, 1997). Although the simulation was able to account for the time- and depth averaged currents, it did not account for the strong fluctuations in the observed along-shelf currents. Therefore, transport and depositional patterns under strong along-shelf currents needed to be investigated. Furthermore, the difference in sediment transport and depositional patterns during two stormy periods suggested that wave energy and flood timing may be important. Sensitivity to wave height and timing were therefore tested to better evaluate how transport and deposition is affected by Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 175 wave energy during floods. Finally, the lack of data concerning sediment grain size and settling velocities on the shelf necessitated an in-depth investigation into how sediment transport and deposition vary in response to the fraction of sediment that is flocculated. V. Methods This study used a modified version of ECOMSED described by Harris et al. (2004, 2005), that allowed an evaluation the relative importance of dilute suspension and wav boundary layer gravity flows on the Waiapu River shelf. The model grid contains 100 x 120 cells, with an average resolution of 400 m x 500 m (Fig 3A). Vertical velocity instabilities developed when the model relied directly on bathymetric data from NIWA (National Institute for Water and Atmospheric Research, New Zealand), therefore the bathymetry was smoothed using a Shapiro filter (Shapiro, 1970). Depths greater than 200 m were neglected, in effect assuming no interaction of shelf waters with deeper slope waters. Nine sigma layers with high resolution at top and bottom were used. Time steps of 45 seconds were used for model stability. Horizontal grid resolution was highest near the river mouth. Grid cells along the model’s open boundaries were expanded to ameliorate instabilities caused by river plume interaction with the boundary edges. The southern boundary was elevated to create a ~ 6 cm/s geostrophic current, so that modeled currents matched observed time-averaged values. Modified wave, wind, and discharge data were used to force the model, depending on the sensitivity experiment. Waves were assumed to be perpendicular to shelf bathymetry with a propagation angle of 129 degrees. Observational data and NIWA’s TIDE2D model indicate that waves and non-tidal currents are much stronger than tidal Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 176 currents. Therefore, tides were neglected. Initial sea salinity was set at 32 psu. Oceanic and Waiapu River temperatures were based on monthly average temperatures for the region and were held at 16°C (SeaWiFS images) and 10 °C (historical averages from www.weatheronline.co.uk), respectively. The model assumed that the Waiapu River was the only source of sediments and initially neglects sediments on the bed. The model assumed that fine-grained material would either travel as flocculated particles or single grains. Flocculated sediment was assumed to settle at 0.1 cm/s (Hill and McCave, 2001), but some material was assumed to be unflocculated and settle more slowly at 0.01 cm/s. Aggregation and disaggregation processes were neglected. Transport of sands and coarser sediments was not addressed because we were concerned only with material that makes up the observed muddy deposit (Kniskem, Chapter 1). Furthermore, ECOMSED becomes unreliable when simulating coarser fractions because the model was designed to model sediments no coarser than fine sands (van Rijn, 1984), and resolving high settling velocities requires a prohibitive decrease in time-step. VI. Approach The following section describes the setup of the baseline case that used the unaltered wave, wind, and discharge records from the latter half of July, 2004 (Fig. 3B). The second section describes how the forcings from the baseline case were altered for each experiment. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 177 A. Baseline case for comparison with sensitivity experiments The observed Waiapu River discharge, Hicks Bay winds, and tripod-measured wave heights from a stormy period in 2004 were chosen to provide the basis for comparison with observed conditions. Hourly wind data from Hicks Bay, to the north of the study area, were used to force currents (Figs. 1 and 3B). Hourly Waiapu River discharge and suspended sediment concentration data were used to simulate terrestrial sediment delivery. Waves from the tripod record (Ma et al., in press) were assumed to propagate from the east/offshore direction. This period from July 13 through July 31 was chosen because the wind pattern during the flood was characteristic of some of the other stormy periods in 2004 and because it was a single peak flood (Fig. 3B). The timing effects of wave energy with peak fluvial discharge were therefore investigated for a single peak, making analysis more consistent. At peak flood (2000 m /s) sediment concentrations reached 50 g/L, significant wave height was 3.2 m, and the winds shifted from the north-east to south after peak flood (Fig. 3B). Currents were driven by the wind record and matched observed depth- and time-averaged currents (Fig. 4). The model was not able to account for the period of strong along-shelf currents, which were likely large-scale currents or eddies (Kniskem, Chapter 2, Chiswell, 2005.). Despite this, the model correlated with 81% of the bed shear stresses and the timing and magnitude of the sediment concentration at 1 mab. These findings are consistent with simulation results of the 2004 field experiment (Kniskem, Chapter 2). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 178 B. Sensitivity experimental design Each experiment was designed to test sediment transport sensitivity to one of the model forcings: wave, winds, or fluvial delivery. Parameters that were tested included (Table 1 A): 1. Wave energy: Three experiments were conducted to examine the effects of wave-induced bed shear stresses on sediment transport and depositional patterns. The first experiment addressed mean wave energy by altering the wave heights from the smoothed 60 m tripod record. The second experiment addressed how much the variability of the wave record influences sediment transport and deposition. The wave record shows that wave heights vary between 0.15 and 4.22 m, with a mean wave height of 2.06 + 1.03 m. Three cases were run where the wave height was held at a steady value. The influence of the wave record during floods on sediment transport and deposition was investigated in the third experiment. 2. Wind-driven currents: The relative influences of wind direction and speed were investigated in two experiments. In the first experiment, wind direction was altered to increase or decrease the along-shelf component, but wind speed was not changed. Wind speeds were increased and decreased relative to the original record, but wind direction was not changed in the second experiment. 3. Sediment settling velocity: In order to evaluate how past changes in flocculation of fluvial material may have influenced sediment transport pathway and deposition, the proportion of material with high settling velocities was varied from 0 to 100% in seven experiments This experiment is especially valuable because settling velocity data from the study site is unavailable. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 179 4. Fluvial sediment concentration and load: Various sediment loads were used to investigate how suspended sediment concentrations may affect sediment transport in dilute suspension and in a wave-induced gravity flow. The size of the flood event was altered by decreasing the discharge record from July, 2004 to match the peak discharge for a given sediment concentration. This experiment was used to indicate what magnitude of flood may result in sediment retention on the shelf. Three cases were tried: one with peak sediment concentration at 20 g/L, the original sediment case where peak suspended sediment concentration was 49 g/L, and a third case, where peak concentration was 60 g/1. 5. Historical terrestrial sediment concentration: This experiment was used to investigate the potential impact of historical changes in the Waiapu River sediment load and rating curve. The discharge record from July, 2004 was altered to accommodate the maximum estimated suspended sediment concentration measured during Cyclone Bola in 1988 (Hicks et al., 2004). To investigate how historical changes in Waiapu River sediment delivery may have influenced sediment transport and deposition, the sediment concentration was decreased by a factor of 2 to 7 (Page and Trustrum, 1997). 6. Experimental comparison The results from these experiments were evaluated by analyzing flood deposit location, sediment mass distribution, and relative contributions by dilute suspension and wave boundary layer gravity flow. The results from the experiments were compared with each other using the parameters of mean bed shear stress, flood load, flood concentration, and percent flocculated sediment. Mean bed shear stress was Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 180 calculated at the 60 m tripod location so that the results could be placed in the context of observational data. VII. Results and Discussion The following sections describe the results from the numerical experiments that investigated the effects on sediment transport and depositional patterns and budgets of wind speed and direction, wave height and timing, sediment load, and historical changes in sediment concentration. Time-averaged sediment mass distribution in the water column, the wave boundary layer, and deposited on the seabed are used to quantify the general results of each experiment. These values were time-averaged to reduce the effects of temporal fluctuations on values at an arbitrary time. Patterns of mean along- and across-shelf flux and flux gradients are used to compare the results from the multiple experiments and evaluate the relative importance of gravity-driven flows, dilute suspensions, and resuspension. Across-shelf depositional patterns, observed versus simulated transport and deposition, and anthropogenic influences on sediment transport mechanisms and depositional patterns are discussed. A. Sediment mass distribution responses for each experiment This section presents the general sediment mass distribution in the water column, the wave boundary layer, and the seabed calculated for each sensitivity experiment (Table 1A and B). Calculating the sediment mass distribution allows assessment of the entire shelf area. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 181 1. Baseline case As a result of high wave energy during the flood, sediments bypassed the inner shelf rapidly, within a few days, and were deposited on the mid-shelf even though along-shelf currents flowed strongly to the south-west during the first half of the flood (Fig. 5A). As wind-forced currents switched to be towards the north-east after the peak discharge, the suspended sediment plume moved to the north of the river mouth (Fig. 5B). On average, there was more sediment in the water column than in the wave boundary layer or deposited on the seabed (Figs 5C, Appendix A). About half of the fluvial load was in dilute suspension at any one time in the simulation. Slightly more sediment mass was deposited than was carried in the wave boundary layer (Figs. 5C and 6A), but the final deposit only accounted for 1.3% of the fluvial load, with most sediment retained on the mid-shelf (Fig. 5B). 2. Wave sensitivity: wave height, steady waves, and wave timing experiments Three variables were tested in the wave experiments: wave height variability (steady-wave experiment), the timing of waves with respect to peak fluvial discharge (wave-timing experiment), and wave height (wave-height experiment) (Table IB). Although the greatest range in time-averaged bed shear stress at the 60 m tripod location was observed in the steady-wave experiment, with mean wave heights ranging between 2 to 5 m, the variability in bed shear stress was not reflected in the sediment transport and depositional budgets (Fig. 6B). The time-averaged ratio of fluvial sediment mass in dilute suspension to sediment mass in the wave boundary layer decreased marginally with increasing wave height. The mean amount of sediment in Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 182 suspension remained at 46% of the fluvial load with increasing mean bed shear stress while the amount of sediment in the wave boundary layer increased from 9% to 13% of the fluvial load (Fig. 6B). The increase in the mean amount of sediment in the wave boundary layer was accompanied by a decrease in the mean deposit size, indicating that sediments were resuspended into the wave boundary layer. It is important to note that although the original baseline case had a mean wave height of 2 m, the steady wave 2 m case had less mean wave boundary layer sediment and greater mean sediment deposit than the original case. This suggests that time-averaged wave heights may not be a representative value for sediment transport studies (Fig. 6B). Although the time-averaged sediment deposit mass only varied by 4% for the steady wave experiment, the location of sediment deposition shifted significantly with wave height (Fig. 7A). The greater the significant wave height, and therefore the mean bed shear stress, the deeper the main locus of deposition, consistent with Scully et al. (2003) and Harris et al. (2005). The case using steady 2 m waves, commensurate with the mean of the original case, resulted in a sediment deposit at a shallower depth than the other cases (Fig. 7A; Table 2). Sediment mass within the wave boundary layer increased as wave height increased in the wave-height experiment (Figs. 6B and 7B, Appendix A). The relative amount of sediment in dilute suspension decreased slightly with increasing mean wave height, corresponding with a relative increase of sediment mass in the wave boundary layer. Although the time-averaged amount of sediment in dilute suspension and in the wave boundary layer varied by only a few percent, the time-averaged sediment deposit size varied by 13% and the deposit location shifted from 70 to 10 m water depth (Figs. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 183 6B and 7B). Time-averaged sediment deposition for the wave-height experiment generally decreased with increasing mean wave height and the depocenter shifted seaward (Fig. 6B). The case with the highest mean bed shear stress and the baseline case resulted in sediments primarily deposited on the mid-shelf. Decreasing the wave heights resulted in sediments deposited and retained in shallower waters (Fig. 7B, Table 2). The wave-timing experiment produced the most variable time-series record of suspended sediment mass, wave boundary layer sediment mass, and sediment deposition. Gravity-driven sediment transport was greatest when sediment was available on the shelf and during times when wave heights were high (Figs. 6B and 8A). Although the wave-timing experiment, with mean wave heights varying between 1.2 and 2.06 m, had the smallest range of mean bed shear stress, it produced the second greatest range in final deposit mass and location (Fig. 8B). The case with low waves during peak fluvial discharge had the lowest mean bed shear stress, resulting 45% of the fluvial load retained on the shelf (Fig. 8B, Table 2). The wave-timing experiment indicated that when wave heights were low during peak fluvial discharge, sediments were rapidly deposited and retained on the inner shelf (Fig. 8, Table 2). When wave heights were high during peak fluvial discharge, sediments tended to rapidly bypass the inner shelf and were deposited on the mid-shelf within hours of the flood. 3. Wind Sensitivity: wind speed and direction The final and time-averaged sediment deposit size increased with decreasing mean bed shear stress for both the wind-speed and wind-direction experiments (Appendix A, Fig. 6B). Both wind speed and direction influenced the amount of the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 184 fluvial load that escaped the shelf. The wind speed experiment results correlated well with mean bed shear stress, whereas sediment export increased with increasing mean along-shelf wind stress for the wind direction experiment (Fig. 9). An average 40% of the Waiapu River sediment budget was deposited on the shelf for the lowest wind speed case compared to 1% in the highest wind speed case. Furthermore, the shape of the final deposit appeared to be influenced by the along-shelf component of the mean bed shear stress (Fig. 10). Strong along-shelf wind-forced currents resulted in an elongated depositional footprint. Strong along-shelf winds also resulted in a greater proportion of the sediment deposit retained in shallower waters, whereas weak winds allowed sediments to be transported and deposited at greater water depths (Table 2). The relative proportion of sediment in dilute suspension to sediments in the wave boundary layer increased significantly with increasing mean bed shear stress associated with increasing wind velocities. As mean bed shear stress increased with increasing wind velocity, the mean amount of sediment transported in the wave boundary layer decreased drastically. This contrasts with the trend seen in the wave experiments where increasing mean bed shear stress resulted in less sediment transported within dilute suspension and more sediment transported in the wave boundary layer (Fig. 6B, Appendix A). Mean sediment mass transport did not vary as strongly for the wind-direction experiment, because the range in mean bed shear stress was smaller than in the wind speed experiment (Fig. 6B). Although wind speeds were the same for the wind direction experiment, each case produced variable mean bed shear stresses as a result of varying along shelf wind stresses and different current patterns (Fig. 6B, Table 3). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Those wind cases that produced strong along-shelf currents resulted in greater mean bed shear stresses and, therefore, smaller sediment deposits. When winds were stronger in the across-shelf direction, along-shelf currents and mean bed shear stresses were lower, resulting in a larger final sediment deposit (Fig 10B, Tables 2 and 3). Mean bed deposition ranged between 4% and 17%, with depositional footprints temporarily confined to the inner shelf when along-shelf currents were strong (Fig 10B). Sediments tended to be retained on the shelf when along-shelf bed shear stresses were lowest, as for the NW wind case. 4. Sensitivity to flocculation fraction The partitioning between high settling velocity material (ws = 0.1 cm/s) to low settling velocity material (ws = 0.01 cm/s) affected sediment transport pathways more than any other experiment (Appendix A, Fig. 6A). Time-averaged seabed deposition and the amount of sediment transported within the wave boundary layer increased with increasing proportion of flocculated material (Figs. 6A and 11). The relative proportion of sediment transported in dilute suspension decreased with increasing proportion of high velocity settling material from nearly 100% to 60 %. The range in average deposition was 14%, and increased with increasing floe fraction. The range in final sediment deposit size was only a few percent, however (Figs. 6A and 11, Table 2), indicating that resuspension was important on the shelf. Settling velocities also controlled where sediments were deposited on the shelf. Sediments with high settling velocities were deposited in shallower waters, whereas sediments with low settling velocities were transported and deposited to deeper waters (Table 2). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 186 5. Sediment load sensitivity This experiment was used to evaluate sediment transport and deposition for various flood loads using the sediment rating curve by Hicks et al. (2004). Time- averaged sediment deposition and sediment transport pathways were significantly affected by the sediment load. Mean sediment deposition and the final deposit size increased with sediment load, whereas the ratio of sediment transported within dilute suspension to sediments transported in the wave boundary layer marginally increased with increasing sediment load. The high deposition for the high load cases suggests that the river delivered more sediment than could be transported in either dilute suspension or within the wave boundary layer (Fig. 6 A and 12). Increasing the sediment load also resulted in a landward shift in the location of peak deposition (Table 2). 6. Sediment concentration sensitivity To test how sediment transport and depositional patterns might have changed due to deforestation over the last 100 years, the river suspended sediment rating curve (Hicks et al., 2004) was adjusted. As sediment concentration was decreased with river discharge held steady, mean sediment deposition lowered from 24% to 7% (Appendix A and Fig. 6A). The proportion of time-averaged sediments in dilute suspension remained steady between 34-35% of the fluvial load, whereas proportion of sediments in the wave boundary layer slightly increased with decreasing sediment concentration. As the sediment concentration decreased, the main locus of deposition was shifted seaward (Fig. 13, Table 2). This suggests that floods with higher sediment concentrations deliver more sediment to the shelf than is capable of being transported. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 187 B. General sediment mass distribution for all experiments The amount of the fluvial load carried in dilute suspension, transported as a gravity-driven flow in the wave boundary layer, and deposited on the seabed varied for each experiment and indicated how changes in mean bed shear stress, flocculated fraction, and sediment load and concentration affected sediment distribution on the shelf (Fig. 6, Appendix A). The time-averaged amount of sediment in the water column always exceeded the amount of sediment in the wave boundary layer, and ranged between 60% and 98% of the total amount of sediment in transport (Table 2). The mean sediment mass within the wave boundary layer varied between 2% and 40% of the total amount of sediment in transport. The fraction of sediment deposited on the bed at the end of each experiment, 13 days after the peak fluvial discharge, accounted for between 0.08% and 54.2% of the flood load compared with 1-48% mean deposition (Table 2, Appendix A). The final deposit fraction of the flood load was generally smaller than the mean deposit fraction, indicating that sediment resuspension occurred. Overall, the results showed that increasing the bed shear stresses, whether through elevating wave height or increasing wind-forced currents, resulted in a decrease in the bed deposit size to varying degrees (Fig. 6B, Table 2). This occurred because increasing both wave- and current-induced bed shear stresses resulted in increased resuspension. This is consistent with previous work that suggests resuspension by waves and currents is an important process on shelves (Sternberg and Larsen, 1976, Grant and Madsen, 1986; Kineke and Sternberg, 1995; Ogston and Sternberg, 1999). Although results from Harris et al. (2005) produced similar shelf budgets for varying Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 188 mean wave heights, their simulations only investigated initial flood deposition and were not long enough to examine the potential effects of post-flood resuspension. While mean deposition varied by 15 % with variable floe fraction in this study, final deposit size only varied a few percent, indicating that resuspension was more important in determining deposit size than settling velocity (Fig. 6A). Larger sediment loads and concentrations resulted in greater deposition and larger final deposits. The ratio of the final deposit to mean deposition indicated that less sediment was resuspended for the larger flood loads. Therefore, flood loads and concentrations did affect resuspension. Sediment distribution between dilute suspension and the wave boundary layer varied in a more complex way. Analysis of the sediment mass distribution indicated that the fate of resuspended sediments varied according to the dominant source of bed shear stress (Fig. 6B). When current-induced bed shear was strong, sediments were preferentially resuspended into the water column. When waves were energetic, however, sediments were preferentially resuspended into the wave boundary layer. Although energetic waves did resuspend sediments into the water column above the wave boundary layer, the capacity of the wave boundary layer increased with wave energy, increasing the likelihood that sediments would be contained within it (Harris et al., 2005). Other studies have shown that currents enhance sediment resuspension and transport, but that current-induced bed shear stresses are not enough to resuspend sediments without waves (Syvitski et al., 2006). The measured currents on the Waiapu River shelf, however, exceeded critical bed shear stress for resuspension (Ma et al., in Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 189 prep; Ma, pers. comm.). The currents produced by the wind speed experiment reached the same magnitude of the observed currents and indicated that currents enhanced sediment resuspension. The wind-direction experiment, where wind speeds were held constant, showed that wind-induced current circulation patterns also influenced sediment transport. Strong along-shelf wind stresses were associated with a decrease in the amount of sediment transported in the wave boundary layer (Fig. 6B). The proportion of the fluvial load transported within the wave boundary layer increased with flocculated sediment percent (Hill et al., 2000; Geyer, et al, 2000; Harris et al., 2005), and decreasing sediment load and sediment concentration (Fig. 6A). The floe fraction experiment indicated that settling velocities influenced residence time in the water column, thereby influencing deposition and the relative dominance of each sediment transport mechanism (Geyer et al., 2000; Harris et al., 2005). Increasing the fluvial sediment load and concentration resulted in higher amounts of sediment transported within the wave boundary layer and within the water column, but resulted in a decrease in the fraction of the fluvial load transported in the wave boundary layer. This result indicated that more sediment was available than could be transported in the wave boundary layer, and is supported by the increase in deposition and in dilute transport with increasing sediment load (Figs. 6B, 12, and 13). C. Time-averaged along and across-shelf flux Time-averaged along- and across-shelf flux in the water column and in the wave boundary layer allowed evaluation of the relative importance of transport in dilute suspension, gravity-driven transport, and resuspension to the fate of the sediments. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 190 Time-averaged along- and across-shelf flux was therefore calculated for each along- shelf row in the model grid, allowing an evaluation of flux spatial variability with depth across the shelf. (Figs. 3 and 14-17). Additionally, flux across the boundaries of the model was used to calculate sediment export from the proximal shelf (Fig. 3). The following are discussed: general patterns of along-and across-shelf flux, flux sensitivities, and sediment export. 1. General patterns of along- and across-shelf flux Although mean across-shelf flux reached rates as high as mean along-shelf flux (Fig. 14A, Tables 4 and 5), there was significant spatial variability across the shelf. Mean across-shelf flux was highest on the inner shelf, decreasing with depth across the shelf (Table 4). Similarly, mean along-shelf flux was high on the inner shelf, decreasing with depth across the shelf (Fig. 14A). Several experiments showed a mid shelf peak in mean along- and across-shelf flux which represented a zone of resuspension. The various sediment transport mechanisms produced spatially variable flux patterns across the shelf (Fig. 14A). Gravity-driven transport was highest on the inner shelf, and decreased with depth across the shelf as waves were attenuated similar to results of other studies (Harris et al., 2005; Wright and Friedrichs, 2006; Scully et al., 2003; Traykovski et al., 2000). Mean across-shelf flux within the water column was also greatest on the inner shelf and generally decreased across the shelf. Mean flux seaward of the shelf break was minimal (Table 4). Several of the experiments, however, showed a peak in dilute suspension at water depths that ranged between 30-85 m depth, associated with resuspension. Across-shelf gravity-driven flux was generally most Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 191 important up to about 60 m, beyond which suspended sediment transport became relatively more important (Figs. 14-17). Sediment settling velocities largely determined transport pathway and dominant transport mechanism. Both flocculated and unflocculated sediments were transported in dilute suspension, whereas flocculated sediments comprised almost all of the sediments transported in the wave boundary layer (Fig. 14B, Appendix A). Unflocculated sediments with low settling velocities were preferentially transported along the shelf in shallow waters, whereas flocculated sediments were relatively more important with depth across the shelf. Similar sediment transport pathway sensitivities to settling velocities have been observed on other shelf systems including the Eel River Shelf (Geyer et al., 2000; Hill et al., 2000; Syvitski et al., 2006), the Po River Prodelta (Fox et al., 2004; Milligan et al., 2007), and the Amazon River Shelf (Gibbs and Konwar, 1986; Berhane et al., 1997). The mid-shelf peak in across-shelf total suspended flux was associated with a peak in flocculated suspended sediments, confirming that this zone was susceptible to resuspension by waves and currents (Figs. 14-17). The location of the resuspension zone was just seaward of where flux within the wave boundary layer began to decrease as waves were attenuated (Traykovski et al., 2000; Scully et al., 2003). Wright and Friedrichs (2006) suggested that when the wave boundary layer dissipates, sediments in the wave boundary layer may either be entrained into the overlying water column or transported further across the shelf within a current- or auto-suspending gravity flow. The version of ECOMSED used in this study only accounted for gravity-driven transport in the wave boundary layer (Harris et al., 2004), and, therefore, sediments Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 192 were either resuspended into the water column or deposited on the seabed. Adding auto-suspension to the ECOMSED model might increase gravity-driven flux on some parts of the Waiapu River Shelf. Along-shelf currents consistently transported sediments to the north of the river mouth in dilute suspension for most of the experiments (Table 5; Figs.14-17). The wind direction scenario with winds from the north-east was the only case that had sediment flux out of the model grid across the southern boundary (Figs. 10B and 17B). Although along shelf flux was greatest on the inner shelf, many of the experiments showed a smaller peak in along-shelf flux on the mid-shelf. This peak correlated with the across- shelf peak in suspended flocculated flux between 50 -80 m, indicating that some of the sediments resuspended from the mid-shelf were exported (Kniskem, Chapter 2). 2. Along- and across-shelf flux sensitivities Time-averaged along and across-shelf patterns of flux varied with bed shear stress, floe fraction, and sediment load and concentration (Figs. 14-17). Changes in mean along- and across-sediment flux rates and the relative location of peaks in along- and across-shelf time-averaged flux were used to assess sensitivities. Across-shelf flux varied the most with wind strength (Fig. 17), wind direction, and wave height (Fig. 16, Table 4). The location of the mid-shelf resuspension zone depended on sediment deposition on the shelf and the depth of wave attenuation. Specifically, the resuspension zone was located just seaward of where gravity-driven fluxes began to decrease, generally between 50 -80 m water depth. The experiments that addressed wind speed, sediment settling velocity, and wave timing produced the largest variance in the mean along-shelf flux (Table 5). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Time-averaged across-shelf flux responded variably to wind- and wave-induced bed shear stresses. Increasing wave-induced bed shear stresses typically resulted in increased time-averaged across-shelf flux, especially within the wave boundary layer, and decreased along-shelf flux (Fig. 16), consistent with work by Harris et al. (2005) on the Eel River Shelf. The increase in time-averaged gravity-driven transport resulted in a seaward shift in the resuspension zone. These trends were reversed for current-induced bed shear stresses (Fig. 17). Increased current shear stresses associated with strong winds resulted in decreased time-averaged across-shelf flux and increased time- averaged along-shelf flux. The resuspension zone shifted landward with the decrease in mean gravity-driven flux. Increased time-averaged along-shelf flux was also associated with strong along-shelf currents on the Eel shelf (Traykovski et al., 2000; Geyer et al., 2000) and the Amazon River Shelf (Kineke and Sternberg, 1995). These sensitivity patterns have multiple implications. For the wave experiments, mean flux to the north increased with bed shear stress up to about 0.2 - 0.4 Pa, beyond which time-averaged along shelf flux generally decreased (Fig. 18A). This decrease in mean sediment flux from the shelf with increasing bed shear stress due to waves (Table 5) suggests that across-shelf gravity-driven sediment transport initiated by waves acted as a mechanism for sediment retention on the shelf. Although energetic waves resulted in reduced along-shelf sediment escape and increased across-shelf flux, sediment deposition was negatively impacted by increased mean bed-shear stresses (Figs. 6B and 7, Tables 2 and 4). The apparent discrepancy is explained by a decrease in sediment resuspension with lower wave heights, reducing the amount of sediment available for either along- or across- shelf transport (Figs. 7B 16, Table 2). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 9 4 Additionally, the higher wave-height cases resulted in sediments transported to depths where they were not easily resuspended and exported. Sediment availability strongly influenced mean along-shelf flux and the total amount of sediment exported from the shelf area. Availability depended on how long sediments were retained on the inner shelf and the relative influences of the current, wave, and flood conditions. As the time sediments were retained on the inner shelf either due to strong currents or low wave energy increased, the chance that these sediments would be resuspended and transported increased. For example, the two experiments with lowest wave-induced mean bed shear stresses deposited sediments primarily on the inner shelf (Figs. 6B, 7, and 16). The deposited sediments were not easily resuspended, and the resultant time-averaged along-and across-shelf fluxes were relatively low. Mean flux was greater for the case representing half the original wave heights because wave-induced bed shear stresses were more likely to resuspend sediments from the inner and mid-shelf than the case representing 0.2 times the original wave heights (Tables 4 and 5). Sediment availability was also influenced by wave-timing with reference to the peak flood discharge (Fig. 16C). Although the mean bed shear stress was lower for the case where waves were small during peak fluvial discharge for the wave-timing experiment, a greater percentage of sediment was transported in the wave boundary layer (Table 2 and Appendix A). This occurred because sediments were allowed to settle and deposit in shallower waters during the period of non-energetic waves, and then were resuspended and preferentially transported within a near bed fluid mud layer by energetic waves that occurred after the flood (Fig. 8). This pattern of storm and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 195 flood coherence has been identified on the Eel River Shelf (Ogston et al., 2000). Although the Waiapu River shelf may experience times of low wave energy during floods, it may be rare for wave heights to remain low long enough for sediment consolidation on the shelf. Therefore, a significant fraction of sediments deposited on the inner shelf during low wave and current conditions may be remobilized within a week or two of deposition. Along-shelf current speed and direction also played a role in sediment availability (Figs 10, 17, 18B, and 19) (Geyer et al., 2000; Kineke and Sternberg, 1995). When along-shelf wind-driven currents were strong and directed north-ward, the suspended sediment plume was confined to the inner shelf, and sediments were rapidly transported to the north of the river mouth. A similar plume pattern was observed on the Eel River Shelf (Geyer et al., 2000; Traykovski et al., 2000). Cases with lower along-shelf current speeds had more sediment available for transport at later times (Figs. 19). Sediment flux timing and direction was strongly affected by flocculated fraction, with low settling velocity sediments rapidly transported out of the shelf area and high settling velocity sediments slowly transported from the shelf (Figs. 11, 20A and B) (Geyer et al., 2000). As a result unflocculated and flocculated sediments had very different transport, deposition, and resuspension pathways. Unflocculated sediments tended to be retained in dilute suspension, whereas flocculated sediments were recycled between the water column, wave boundary layer, and the seabed. Flocculated sediments were, therefore, available for transport, deposition, and resuspension for much longer time periods than unflocculated sediments. Furthermore, increasing the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 9 6 floe fraction of the fluvial load resulted in increased mean across-shelf flux, mostly through increased gravity-driven transport. Accordingly, the resuspension zone shifted from the inner shelf to the mid-shelf when floe fraction increased. Time-averaged across- and along-shelf flux sensitivity to fluvial load and concentration indicated that dilute transport and gravity-driven transport are limited in different ways. Increasing the fluvial load or the sediment concentration resulted in increased mean across- and along-shelf flux (Figs, 15 and 20B). The fraction of the fluvial load transported within the wave boundary layer (Fig. 6A), however, decreased, and the resuspension zone shifted landward. This result is counter-intuitive. The increase in fluvial load and sediment concentration should have caused an increase in sediments transported via gravity-driven flows (Harris et al., 2005). The sensitivity experiments indicated, however, that transport within the wave boundary layer is flow limited. In other words, the wave boundary layer can only transport so much sediment. This idea is supported by the relatively wider resuspension zone defined by the peak in mean flocculated suspended along-shelf flux that extends from 20 m to 75 m water depth for the highest load case (Fig. 15). Too much sediment in the water column resulted in high deposition followed by resuspension into the water column. Wave-induced gravity-driven flows are controlled by wave properties, flood magnitude, and fluvial sediment loads (Harris et al., 2005). A comparison of these parameters revealed that the Waiapu River experiments used similar wave heights, lower flood magnitudes, and similar sediment loads to those used in Eel River Shelf studies (Table IB) (Harris et al., 2005). The Eel River Shelf work showed that a critical amount of sediment was required to produce gravity-driven flows and that flux in the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. wave boundary layer increased with increasing load (Harris et al., 2005). Although sediment loads for the load and concentration experiments in this study were commensurate with the range used in the Eel River Shelf work, Waiapu River sediment concentrations were much higher than experienced on the Eel River Shelf (Harris et al., 2005). The higher sediment concentrations on the Waiapu River Shelf were above the critical threshold of 10 g/L identified by Harris et al.’s (2005) work on the Eel River Shelf, likely triggering gravity-driven transport within the wave boundary layer more quickly than on the Eel River Shelf. For example, gravity-driven transport on the Eel River Shelf occurred only after about 80% of the fluvial load had been delivered to the shelf, whereas gravity-driven transport on the Waiapu River Shelf was observed before peak flood when only 28% of the load had been delivered. The sediment load and concentration experiments in this study indicated that when flood delivery exceeded the capacity of the wave boundary layer to transport sediments across the shelf sediments were either deposited or transported in the water column. 3. Sediment export A significant portion of the sediment, mostly in suspension, was exported beyond the proximal Waiapu River Shelf, accounting for between 24% and 93% of the fluvial load. Waiapu River sediments were usually transported to the north of the river mouth in dilute suspension (Figs. 21 and 22, Table 5). Only a fraction of the fluvial load, between 0-0.63 MT, was transported on to the shelf break area for all of the experiments (Table 2). The wind direction case with strong north-east winds was the only case where sediments were transported out of the model grid across the southern boundary. Only a fraction of the fluvial load was transported seaward of the shelf break Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 19 8 (Table 2). The experiments addressing wind speed, sediment settling velocity, and wave timing produced the largest variance in the amount of the fluvial budget exported from the Waiapu River shelf system (Table 5). The highest export was observed when sediments were composed of 100% unflocculated material and when wind-forced currents were strongest. The lowest amount of sediment was exported for the low wind and low wave scenarios. The average mass exported from the shelf area accounted for an average 48 % of the available flocculated material and an average 92% of the available unflocculated material (Table 5). This pattern showed that unflocculated sediments were preferentially transported out of the shelf area. The low settling velocity of unflocculated sediments keep these particles in suspension, increasing the likelihood of export form the shelf. Flocculated sediments, on the other hand, comprise the bulk of sediments transported in the wave boundary layer, increasing the likelihood of retention on the shelf. This was because wave-initiated gravity-driven flow was limited to transporting sediments across the shelf. Therefore, less flocculated material was available for export out of the shelf area than suggested by the fluvial input ratio of flocculated to unflocculated sediment. The fraction of the fluvial load that was exported from the shelf varied with mean bed shear stress, floe fraction, and sediment load and concentration. Since most of the sediment was exported to the north of the river mouth, the export budget reflected the time-averaged along-shelf flux patterns (Figs. 18, 20, 21, and 22; Table 5). Increasing mean bed shear stress typically increased the amount of sediment exported no matter whether induced by currents or waves (Fig. 22). The steady wave and wave Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 9 9 timing experiment showed that wave timing and variability can also enhance sediment export. The wind direction experiment showed that along-shelf currents mattered as much as the total bed shear stress in determining the fraction of the load that was exported. Scenarios that created strong along-shelf currents enhanced sediment export out of the shelf area, resulting in a significant decrease in across-shelf sediment transport to less than half of the fluvial budget transported beyond the inner shelf. The cases with strong north-east and south-west winds are the prime examples of this (Figs. 17b, 22A). Increasing the sediment load, fluvial concentration, and unflocculated fraction increased the fraction of the fluvial load exported out of the shelf (Table 5). D. Flux convergence and divergence Across-shelf convergences and divergences in sediment flux were used to assess the contribution of wave-induced gravity flow and dilute transport to sediment deposition and removal. The Exner equation, drj _ 1 d^ + ^ J L + dVs dt cb 8x dy dt (1) was used to infer the change in seabed elevation, 8r|, from the changes in flux across and along the shelf and the change in sediment volume over time. Changes in flux (8q) in the across (x) and along (y) directions, divided by the area (8x and 8y) and the density of sediment (Cb) were used to evaluate depositional and erosional thicknesses (Sri) over time (8t). The sediment volume parameter (Vs) was not used in this approach. To estimate flux convergence and divergence in the across shelf direction, the time-averaged flux of sediment for each along shelf row was calculated in both the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0 0 along and across-shelf direction and corrected for the area of the model grid row (Fig. 3). By comparing the spatial gradients in the along-and across-shelf flux, we were able to assess the relative contribution of each sediment transport mechanism to flux convergence and divergence (Figs. 23-25). It is important to keep in mind that a transport mechanism was assumed to be erosive if the flux out of an area increased respective to what was supplied to the area, causing flux divergence. Deposition was favored when the flux out of an area was less than the flux into the area, causing flux convergence. The flux convergence and divergence estimates for the transport mechanisms may indicate multiple processes, however. If time-averaged across-shelf suspended sediment transport gradients are negative, or flux convergent, values indicate that sediments were either deposited on the seabed or settled into the wave boundary layer. Positive, or flux divergent, values indicate that time-averaged across-shelf flux gradients increased. The increase in flux out of an area may indicate sediment erosion from the seabed or sediment resuspension from the wave boundary layer. If time- averaged across-shelf flux gradients within the wave boundary layer are negative, or flux convergent, sediments may have been either deposited or entrained into the overlying water column. Positive, or flux convergent, gradients may indicate resuspension from the seabed or sediments settling into the wave boundary layer from the water column. These results must therefore be compared with sediment mass distribution to correctly interpret the results. Overall, time-averaged changes in across-shelf flux indicated flux convergence, whereas differences in flux in the along-shelf direction implied flux divergence Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 20 1 (Figs.23-25). Across-shelf flux gradients tended to be highest on the inner shelf, except for those cases where gravity-driven flux convergence on the mid-shelf was relatively higher than suspended flux convergence on the inner shelf. The divergent suspended sediment flux gradient peak, between 40-70 m water depths, generally corresponded with a convergent wave boundary layer flux gradient peak, indicating that about half of the flux convergence in the wave boundary layer may be accounted for by entrainment of wave boundary layer sediments into the overlying water column (Figs 23-25). At around 70-80 m depth, sediment flux convergence within the wave boundary layer decreased to zero, beyond which flux convergence from suspended sediments became more important. Along-shelf flux divergence tended to peak at around 30 m depth, due to stronger currents on the inner shelf. Overall, changes in flux along- and across-shelf suggest that sediments transported within the wave boundary layer were retained on the shelf whereas sediments in dilute suspension tended to be exported from the shelf area. Across- and along-shelf flux gradient patterns were influenced by sediment floe fraction, wave and wind record, and sediment load and concentration on sediment transport pathways (Fig. 23-25). As floe fraction increased, along-shelf flux gradients became weaker, and the peak in across-shelf convergence shifted from the inner shelf to the mid-shelf. In addition, suspended sediment flux gradients became increasing divergent and corresponded with increasingly convergent wave boundary layer flux gradients. As sediment load and concentration were increased, across-shelf flux gradients became more convergent, whereas along-shelf gradients became more divergent. Higher loads and concentrations also resulted in a more convergent gradient Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0 2 flux in both the wave boundary layer and in the water column, indicating that there was too much sediment to transport (Fig. 23). The relative strength of winds and waves, and the resulting mean bed shear stresses, created diverse flux gradient responses. Increasing the along-shelf wind-driven current strength (Fig. 19) resulted in stronger gradients on the inner shelf (Fig. 24). Decreasing mean bed shear stresses due to winds resulted in weaker along-shelf gradients, corresponding with stronger across-shelf convergent flux gradients, especially on the mid-shelf. Across-shelf flux gradients in the wave boundary layer became stronger and more convergent with decreasing wind-induced mean bed shear stress, but weaker in the water column. Decreasing mean bed shear stress due to waves resulted in a very different response (Fig. 25). For example, decreasing wave-induced mean bed shear stresses resulted in stronger along-shelf divergent flux gradients and weaker across-shelf convergent flux gradients. Across-shelf flux gradients in the wave boundary became weaker with decreasing wave-induced mean bed shear stress, and more strongly divergent in the water column, reflecting an increase in along-shelf export. Identification of zones of sediment convergence and divergence can contribute to understanding of fine-scale sedimentary strata formation, benthic community structure, contaminant cycling, and carbon sequestration. These sensitivity experiments suggested that small changes in currents, waves, sediment aggregate composition, and sediment load and concentration can affect the pattern of along and across-shelf sediment convergence and divergence by altering the relative importance of the various sediment transport mechanisms. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0 3 E. Sediment depositional patterns across the shelf Seabed observational data and model simulations indicated that muddy sediments were temporarily stored on the inner shelf before being resuspended and transported to deeper waters (Kniskem, Chapters 1 and 2). Surface sediments from gravity cores collected in August 2003 and May 2004 from depths shallower than 50 m were significantly coarser than sediments from deeper waters, and were characterized by a muddy surface layer underlain by sandy silts (Kniskem, Chapter 1). The ECOMSED model estimated little to no fine sediment retention on the inner shelf (Fig. 26) because the model did not account for sediment consolidation, burying of muds by sand lenses, or non-linear wave effects. Therefore, sediments deposited on the shelf were always available for resuspension and erosion if wave and current shear stresses were sufficiently high. Recent data from continental inner shelves, however, suggest that 8-13% of fluvial sediments can be retained there over time-scales of 100 years (Crockett and Nittrouer, 2003; Wadman et al., 2006). Although the model over-estimated sediment removal from the inner shelf, it does suggest the conditions under which fine sediments would be delivered to the inner shelf. Sediments were deposited on the inner shelf during river floods, and were subsequently eroded by energetic waves and currents. When waves or currents produced shear stresses too low to resuspend sediment, a significant portion of the fluvial budget was retained in shallow coastal waters over the simulated period (Figs. 7, 8, and 10). Low wave energy during peak fluvial discharge resulted in deposition on the inner shelf (Figs. 7). Similarly, when sediment loads were high, more sediment was retained on the inner shelf (Figs. 12 and 13). The model indicated that most of the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0 4 sediments deposited on the inner shelf are eventually resuspended and removed, but observations by Wadman and McNinch (in review) indicate that 8-13% of fluvial sediments are retained on the inner shelf. Sediment deposition on the mid-shelf, 30 - 70 m, was rapid when wave- induced shear stresses were high such that sediments bypassed the inner shelf (Figs.7 and 8). Otherwise, sediment deposition on the mid-shelf lagged a few days behind peak flood. Over the simulated period of a few weeks, a significant portion of the fluvial budget was deposited on the mid-shelf (Table 2). Strong along-shelf currents, however, resulted in increased sediment export from the shelf and reduced across-shelf flux to the mid-shelf and deeper waters (Figs. 17,18B, and 19). Sediment settling velocities affected mean sediment deposition. In most cases, the majority of sediment was deposited on the mid-shelf (Fig 11 and 23 A). When 100% of the sediment was flocculated, more sediment was deposited on the inner shelf, whereas the case with 100% unflocculated sediments resulted in more sediment deposition on the outer shelf (Table 2). At the end of the model simulations, a significant portion of the sediments retained on the shelf were found on the outer shelf between 70 - 130 m depth. (Table 2, Figs. 7, 8,10-13). Deposition on the outer shelf lagged behind the fluvial sediment delivery signal by a few weeks (Fig. 26), and was enhanced by weak along-shelf currents and high wave-induced shear stresses. Most of the sediment deposited on the outer shelf settled out of dilute suspension because wave-supported gravity flows don’t get there (Figs. 23-25). Only when waves were high (>3m) were sediments deposited on the outer shelf by gravity-driven transport. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0 5 Sediment delivery to deeper waters, greater than 130 m, was attributable primarily to sediments settling out of dilute suspension (Figs. 23-25, 26). This is because the wave orbital velocities, wave boundary layer height, and wave shear stresses all attenuate with depth at all depths, dissipating at about -120 m depth. Generally, very little of the fluvial load was transported to the shelf break, except when sediment loads were high (Figs. 12 and 13). In summary, sediment mass distribution and depositional patterns varied across the shelf. Each across-shelf zone had a unique record of sediments in suspension, in the wave boundary layer, and deposited on the bed (Fig. 26, Table 2). The shelf zones were defined by the relative importance of suspended sediment and gravity-driven transport as well as depositional history (Kniskem, Chapter 2). The record of sediment mass showed that sediments were transported from the inner shelf (0-30 m) to the shelf break (>130 m) over a period of several days (Fig. 26). Generally the mass of sediment in suspension was greater than that transported within the wave boundary layer (Appendix Al). There were a few scenarios in which the sediment mass in the wave boundary layer exceeded that in suspension on the inner- and mid-shelf, however. The amount of sediment in the wave boundary layer decreased significantly with depth across the shelf, with very little sediment present on the outer shelf (70-130 m) and the shelf break. F. Observed and simulated deposition and preservation 1. Short- and long-term depositional patterns Observational data from sediment cores, collected in August 2003 and May 2004, implied that there was a disconnect between short- and long-term depositional Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0 6 patterns (Kniskem, Chapter 1). The presence of 7Be from between 40 m to -80 m suggested that these terrestrial sediments were deposited within the previous 4 to 5 months (Kniskem, Chapter 1, Fig. 11). According to excess 210Pb profiles, however, longer-term deposition, over the last 100 years, was concentrated between 60 m to 130 m depth (Kniskem, Chapter 1, Fig. 8). These two depositional patterns are consistent with model results. Simulated initial deposition of fluvial sediments ranged between 0 m to -60 m depth depending on wave height during peak fluvial discharge and wind-driven current strength (Fig. 8). After the flood, sediments were resuspended and reworked and then transported across the shelf or out of the proximal shelf area. When waves were low during the peak fluvial discharge, sediments were initially deposited at shallower depths before being resuspended and transported either to the north, along the shelf, or to deeper waters (Fig. 8). Higher wave heights during peak fluvial discharge, however, resulted in initial deposition on the mid-shelf region within hours of peak flood. This indicated that 7Be may be found at greater water depths after a flood associated with energetic waves. The wind record from 2004 showed a characteristic wind pattern for two of the floods: winds generally blew to the south as the floods began, rotating to the north after peak flood (Fig. 3) (Kniskem, Chapter 3, Fig. 4). The resultant strong along-shelf currents confined initial flood deposition to the inner shelf (Fig. 5A). Sediments were repeatedly resuspended, transported, and deposited along the length of the shelf when strong along-shelf currents flowed to the south. When the currents switched direction due to the change in winds after peak flood, inner shelf sediments were resuspended and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0 7 either transported to the north of the proximal shelf or transported across the shelf (Fig. 5b, 10). The model data indicated two possible explanations for the differing short- and long-term depositional patterns observed in the geochronology data. Waiapu River discharge data showed that the Kasten cores were collected immediately after a flood (Fig. 2). This suggests that the recent flood sediments had not yet been resuspended and transported to deeper waters. Had samples been collected a few weeks later, surface sediments from the mid- to outer shelf may have had a 7Be signal. Alternatively, the 7Be signal may be diluted with older, 7Be-poor sediments, during resuspension and transport within the wave boundary layer, reducing our ability to identify recently deposited sediments on the mid- to outer-shelf. This second hypothesis is more likely given that 7Be activities in samples from the mid- and outer- shelf were below detectible limits despite the flood sediments delivered during July of 2003 that should have had time to be resuspended and transported to deeper waters (Kniskem Chap 1, Fig. 9). It should be noted, however, that if waves during floods in the austral winter of 2003 were not energetic, then sediments may not have been resuspended from the inner and mid-shelf region. Variable patterns of short- and longer-term deposition were not observed on similar shelf systems such as the Eel River shelf and the Waipaoa River shelf. Sediments collected up to a few months after a flood retained a significant 7Be signal on the mid and outer shelves of these two systems (Gerald and Kuehl, 2006; Sommerfield et al., 1999). There is no evidence of temporary storage on the inner shelf of the Waipaoa River shelf, but there is evidence of ephemeral deposition on the Eel River Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0 8 Shelf after a winter storm in 1997 (Geyer et al., 2000). The wave and river discharge records from the Eel River shelf show that waves were energetic, at greater than 3 m in height during peak fluvial discharge during the flood (Harris et al., 2005). Sediments were temporarily deposited on the inner shelf before being rapidly transported to the mid-shelf within a near-bed fluid mud layer (Traykovski et al., 2000, Wheatcroft and Borgeld, 2000). Sediments collected from the Eel River Shelf shortly after the flood were rich in Be on the mid-shelf, where longer-term accumulation rates were high (Sommerfield et al., 1999). If wave energy had been lower during the 1997 Eel River flood, 7Be-rich sediments might have been found on the inner shelf. 2. Long-term shelf preservation patterns Seabed data and model results indicated that sediments accumulate on the mid to outer shelf on time scales ranging from a few months to over the last 100 years (Kniskem, Chapter 1, Addington et al., in press; Lewis et al., 2004). The simulation results showed the greatest sediment retention on the mid- and outer shelf. The various simulation and sensitivity experiments created depositional footprints that roughly matched the observed 100-year depositional pattern. There is a discrepancy between the observed and simulated angle of the area of highest deposition with the coast, however. The observed depocenter was aligned with a synclinal basin (Lewis et al., 2004), whereas the model estimates of the shelf depocenter was aligned with the smoothed bathymetry used in the model (Fig. 27). There are several potential explanations for this discrepancy. First, it is important to keep in mind that the seabed data characterized deposition over the last 100 years, whereas the model was used to simulate transport and deposition for only a few Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0 9 weeks for the sensitivity tests to a few months for the simulation experiment (Kniskem, Chapter 2). Secondly, the model did not accurately simulate the strong observed along- shelf currents, which could have affected the deposit alignment. Thirdly, the smoothed bathymetry used in the simulations may produce a different depositional pattern than observed. There is evidence that the bathymetry may play a part in the depositional footprint. The Waiapu River shelf basin, identified by Lewis et al. (2004), is defined by a synclinal basin with high subsidence rates at > 4mm/year. This observed subsidence may have resulted in a bathymetric low that is not apparent in the low-resolution bathymetry of the New Zealand map. Potential discrepancies in fine-scale bathymetry and the differences in time-scales between the observed and theoretical data likely account for the difference in depocenter alignment with the bathymetry. There is also a discrepancy between simulated and observed patterns of sediment transport and deposition beyond the shelf break (Figs.7, 8,10-13, 27). Simulated sediment transport and deposition during the field deployment indicated that only a fraction of fluvial flood load was transported and deposited beyond 130 m depth (Kniskem, Chapter 2). Excess 2I0Pb accumulation rates between 0.2 -1 .0 cm/y showed, however, that sediments were deposited beyond this depth and that a region along the shelf break may act as a conduit for gravity-driven transport (Kniskem, Chapter 1, Addington et al., in press). The sediment load and concentration sensitivity experiments indicated that sediment transport and deposition in this region probably occurs during larger flood and wave events than observed during the 2004 field Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 1 0 experiment (Figs. 12 and 13). Large floods during storms such as Cyclone Bola in 1988 may, therefore, be the source of sediments deposited beyond the shelf break. 3. Transport mechanisms Geochronological data and 8 13C sediment profiles indicated that terrestrially- derived flood sediments were rapidly transported, deposited, and buried on the shelf such that the physical and terrestrial chemical constituents of the pulsed event layers were preserved in the sediment record (Kniskem, Chapter 1). Sediments with preserved pulsed event layers were generally found between about 50 to 140 m depth and were limited spatially in the along-shelf direction (Kniskem, Chapter 1, Fig. 16). Pulsed event layers were preferentially preserved to the south of the river mouth. All of the cores from transect D contained pulsed event layers, suggesting that this part of the shelf might act as a conduit for fluvial sediment deliver to the shelf break (Kniskem, Chapter 1, Addington et al., in press). The pulsed event layers on the Waiapu River Shelf were interbedded with high excess 210Pb activity layers with marine 8 13C signatures (Kniskem, Chapter 1, Fig. 12). The presence of these high activity layers indicated that buoyant plume delivery, resuspension into the water column, and/or quiescent periods between floods, thereby allowing bioturbation, influenced fine scale strata formation on the shelf. Signs of bioturbation were few on the inner and mid-shelf, and increased radially away from the river mouth (Kniskem, Chapter 1, Figs. 4 and 9). The simulation and sensitivity experiments indicated that buoyant delivery, gravity-driven flows, and resuspension occurred on the Waiapu River Shelf. The transport patterns identified in the model correspond with the observations in several Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ways. First, the frequency of observed pulsed event layers in the sediments decreases across the shelf. The model indicated that deposition from gravity-driven flows decreases beyond about 60 m depth, and deposition from dilute suspension increases. Secondly, both observations and model estimates indicated that accumulation peaks on the mid to outer shelf and decreases across the shelf. This decrease in accumulation and/or frequency of delivery beyond 130 m water depth allowed recovery of the benthic community and an increase in bioturbation signals. Thirdly, both the seabed data and the modeling data indicated that sediment transport within buoyant fluvial plumes, gravity-driven flows, and sediment resuspension into the wave boundary layer and the overlying water column contributed to fine-scale sedimentary strata formation on the shelf. The model simulations and sensitivity tests estimated that a significant fraction of the sediments not retained on the shelf were exported from the proximal shelf to the north of the river mouth in dilute suspension. This finding generally matches the observational data, which indicated that the shelf break was not the primary repository for sediments escaping the shelf (Kniskem, Chapter 1). Although the time-averaged across-shelf gravity-driven flux patterns generally matched the observed patterns of pulsed event layer preservation (Figs. 14-17) (Kniskem, Chapter 1), simulated gravity-driven flows were not typical beyond -100 m depth. The presence of observed pulsed event layers in sediments beyond 100 m depth suggests that greater period swell than observed in 2004 may contribute to sediment transport to deeper waters. The sensitivity experiments did indicate that larger load floods dispersed during periods of energetic waves would be deposited at greater depths. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. There are several other possible reasons why the model did not account for gravity-driven flows to the shelf break. For example, small scale bathymetry not accounted for in the model may act as a conduit for pulsed event sediments to the area sampled by transect D of the seabed observational data (Kniskern, Chapter 1, Fig 5). Additionally, we may have underestimated the coherence of hyperpycnal plumes, which Friedrichs and Wright (2004) suggested would be dissipated by high wave energy on the inner shelf. Relict channels identified on the inner shelf (Wadman and McNinch, 2006) may act to steer hyperpycnal plumes to deeper waters during times of low wave energy. Furthermore, since the model assumed gravity-driven transport was confined to the thickness of the wave boundary layer, it may have underestimated across-shelf gravity-driven flux to deeper waters. Recent work also suggests that sediment concentrations in the energetic current-wave boundary layer were sufficient to support gravity flows (Ma et al., in prep). The shelf slope exceeds the critical threshold of 0.01 on the Waiapu River Shelf (Friedrichs and Wright, 2004), suggesting that auto suspension may be responsible for transporting sediments to the shelf break. G. Historical changes in sediment concentration and load As previously mentioned, the East Cape region has experienced periods of deforestation due to volcanic activity, climate change, and/or human activities during the Holocene. During the latter part of the 19th and the early 20th centuries, European settlers deforested the region to increase pasture land. As a result, the already characteristically high erosion and terriginous sediment delivery rates increased an additional 2- to 7-fold (Page and Trustrum, 1997). This increase in sediment delivery Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 1 3 and fluvial sediment concentration over the last 100 years may have altered the relative importance of the various sediment transport mechanisms and sediment depositional patterns on the shelf (Carter et al., 2002). For example, an increase in the fluvial suspended sediment concentration and load may have resulted in the more frequent development of near-bed fluid mud layers, potentially increasing the accumulation rate and preservation of sedimentary strata (Milliman and Kao, 2005). In order to assess how sediment transport and depositional patterns might have changed in response to increased sediment delivery, the sediment rating curve was altered to reflect a 2- to 7- fold decrease in sediment delivery during a flood (Figs 6a, 13,15, and 23). The results suggested that the increase in sediment delivery over the last century resulted in changes in sediment transport mode and a possible landward shift in the shelf depocenter. As sediment concentration was decreased to simulate pre deforestation fluvial concentrations, a greater percent of the flood load was transported in the wave boundary layer (Fig. 12). The ratio of 75% to 25% flocculated to unflocculated sediment in the model resulted in rapid settling and sediments were preferentially transported within the wave boundary layer. The amount of sediment deposited and retained on the seabed was much less than observed for the higher sediment load cases, however. The higher sediment load cases simulating modem fluvial conditions delivered more sediment to shelf waters than could settle into the wave boundary layer. Therefore, a greater fraction of the fluvial load was transported in dilute suspension. The results from the sediment concentration experiment indicated that sediment Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 1 4 transport in dilute suspension was supply limited, whereas transport within the wave boundary layer was flow limited. If sediment aggregate properties have not remained the same over the last 100 years (Kasai et al., 2005; Kasai, 2006), however, sediment transport and depositional patterns may have changed much more dramatically. While the depocenter appears to have shifted landward over the last 100 years with respect to the location of the maximum thickness of Holocene deposition identified by Lewis, et al. (2004), we do not currently have a seabed depositional record of sufficient resolution to indicate how depositional patterns and rates may have changed as a result of deforestation over the last 150 years. It has been suggested that the sediments delivered to the shelf have fined as a result of deforestation (Page and Trustrum, 1997). There is no way to tell whether the fine sediments were transported across the shelf as aggregates or were primarily disaggregated, however. There is also no record as to what proportion of the fluvial budget was composed of sandy material before deforestation. The sediment aggregate properties experiment indicated that if sediments prior to deforestation were delivered to the shelf as unflocculated sediments, then a significant percent of the sediment load was probably not retained on the shelf (Figs. 6a, 11,15, and 21). This is likely the case, because lower sediment concentrations floods would not promote flocculation as well as floods with higher sediment concentrations (Hill et al., 2000). This does not take into account the potential variability in organic material and the potential effects of the organic material quality on sediment flocculation (Jackson, 1990; Riebesell, 1991). Further work needs to be done to determine the floe fraction of fine sediments delivered to the Waiapu River Shelf. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 1 5 VIII. Conclusions 1) Wave and wind records strongly influence sediment transport patterns. Increasing the wave induced mean bed shear stresses resulted in greater across- shelf transport within the wave boundary layer. Increasing the current-induced mean bed shear stress resulted in more sediments exported from the shelf in dilute suspension. 2) Changing the floe fraction of the fluvial load influenced the relative importance of sediment transport in dilute suspension and within the wave boundary layer. Increasing the unflocculated fraction resulted in more sediments transported in dilute suspension. Flocculated sediments were preferentially transported in the wave boundary layer. 3) The sediment transport and depositional records for each across-shelf zone was unique and sensitive to waves and currents. An increase in current-induced bed shear stress was likely to restrict sediment transport to the inner shelf. Increasing the wave-induced bed shear stress increased across-shelf flux and deposition to deeper waters. 4) The simulated transport and depositional patterns compared favorably with the observed seabed data. Both indicated that fine sediments were generally ephemerally deposited on the inner shelf, and were retained on the mid- to outer shelf. Both theoretically-based and observational data indicated that gravity driven flows were most important for fine-scale strata formation on the mid shelf. Finally, both suggested that multiple sediment transport mechanisms were important for fine-scale strata formation on the shelf. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5) Decreasing the flood sediment load or the sediment concentration resulted in a higher proportion of the flood load transported within the wave boundary layer and smaller final sediment deposits. Deforestation has therefore likely resulted in higher accumulation rates, greater preservation of pulsed event layers, and a decrease in benthic community impact on sedimentary structure. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. References Addington, L.D., Kuehl, S.A., and McNinch, J.E., in press. Distinguishing sediment transport modes to the outer-shelf off the Waiapu River, New Zealand. Marine Geology. Berhane, I., Sternberg, R.W., Kineke, G.C., Milligan, T.G., and Krank, K., 1997. The variability of suspended aggregates on the Amazon Continental Shelf. Continental Shelf Research, 17:267-285. Brown, L.J., 1995. Holocene shoreline depositional processes at Poverty Bay, a tectonically active area, northeastern North Island, New Zealand. Quaternary International, 26:21-33. Carter, L., Manighetti, B., Elliot, M., Trustrum, N., and Gomez, B., 2002. Source, sea level and circulation effects on the sediment flux to the deep ocean over the past 15 ka off eastern New Zealand. Global and Planetary Change, 33:339-355. Chiswell, S.M., 2005. Mean and variability in the Wairarapa and Hikurangi Eddies, New Zealand. New Zealand Journal of Marine and Freshwater Research, 39:121-134. Collot, J-Y., Lewis, K., Lemarche, G., and Lallemand, S., 2001. The giant Ruatoria debris avalanche on the northern Hikurangi margin, New Zealand: Result of oblique seamount subduction. Journal of Geophysical Research, 106 (B9):19271-19,297. Crockett, J.S., and Nittrouer, C.A., 2003. The sandy inner shelf as a repository for muddy sediment: an example from Northern California. Continental Shelf Research, 24:55-73. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Dukat, D.A., and Kuehl, S. A., 1995. Non-steady-state 210Pb flux and the use of Ra/ Ra as a geochronometer on the Amazon continantal shelf. Marine Geology. 125: 329-350. Farnsworth, K.L., Milliman, J.D., 2003. Effects of climatic and anthropogenic change on small mountain rivers: the Salinas River example. Global and Planetary Change, 39:53-64. Field, B.D., Uruski, C.I., 1997. Cretaceous-Cenozoic Geology and Petroleum Systems of the East Coast Region, New Zealand. Monograph, vol. 19. Institute of Geological and NuclearSciences, Lower Hutt, NZ. Foster, G., and Carter, L., 1997. Mud sedimentation on the continental shelf at an accretionary margin - Poverty Bay, New Zealand. New Zealand Journal of Geology and Geophysics, 40:157-173. Fox, J.M., Hill, P.S., Milligan, T.G., and Boldrin, A., 2004. Flocculation and sedimentation on the Po River Delta. Marine Geology, 203:95-107. Friedrichs C.T. and L.D. Wright, 2004.Gravity-driven sediment transport on the continental shelf: implications for equilibrium profiles near river mouths. Coastal Engineering, 51:795-811. Gerald, L.E. and Kuehl, S.A., 2006. Recent Sediment Dispersal and Preservation of Storm Events on the Continental Shelf off the Waipaoa River, NZ . AGU Geyer, W. R., Hill, P. S., Milligan, T. G., and Traykovski, P. A., 2000. The structure of the Eel River plume during floods, Cont. Shelf Res., 20: 2095-2111. Gibb, J.G., 1986. A New Zealand regional Holocene eustatic sea-level curve and its application to determination of vertical tectonic movements. A contribution to Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 1 9 IGCP-Project 200. Royal Society of New Zealand, 24:377-395. Gibbs, R.J., and Konwar, L., 1986. Coagulation and settling of Amazon River suspended sediment. Continental Shelf Research, 6:127-149. Goni, M.A., Monacci, N., Gisewhite, R., Ogston, A., Crockett, J., and Nittrouer, C., 2006. Distribution and sources of particulate organic matter in the water column and sediments of the Fly River Delta, Gulf of Papua (Papua New Guinea). Estuarine, Coastal, and Shelf Sciences, 69:225-245. Grant, W. D. and Madsen, O.S.,1986. The continental shelf bottom boundary layer. Annual Review of Fluid Mechanics, 18. 265-305. Griffiths, G. A.; Glasby, G. P., 1985. Input of river-derived sediment to the New Zealand continental shelf. I. Mass. Estuarine, Coastal and Shelf Science 21:773- 787. Harmsworth, G., Warmenhoven, T., Pohatu, P., Page, M., 2002. Waiapu catchment technical report: Maori community goals for enhancing ecosystem health. Technical Report. Landcare Research New Zealand Ltd., Palmerston North, N.Z. Harris, C.K., P.A. Traykovski, and R.W. Geyer. 2005. Flood dispersal and deposition by near-bed gravitational flows and oceanographic transport: a numerical modeling study of the Eel River shelf, northern California. Journal of Geophysical Research, 110, C09025, doi:10.1029/2004JC002727. Harris, C.K., P.A. Traykovski, and R.W. Geyer. 2004. Including a Near-Bed Turbid Layer in a Three Dimensional Sediment Transport Model with Application to the Eel River Shelf, Northern California. Proceedings of the Eighth Conference Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. on Estuarine and Coastal Modeling, ASCE, M. Spaulding (editor). 784-803. Hicks, D.M., B. Gomez andN.A. Trustrum, 2000. Erosion thresholds and suspended sediment yields, Waipaoa River basin, New Zealand. Water Resources Research, 36:1129-1142. Hicks, D.M., Gomez, B., and Trustrum, N.A., 2004. Event suspended sediment characteristics and the generation of hyperpycnal plumes at river mouths: east coast continental margin, North Island, New Zealand. The Journal of Geology, 112:471-485. Hill, P. S., and McCave, I. N., 2001. Suspended particle transport in benthic boundary layers. In The Benthic Boundary Layer: Transport Processes and Biogeochemistry, edited by B. P. Boudreau and B. B. Jorgensen. Oxford University Press, pp. 78-103. Hill, P.S., Milligan, T.G., Geyer, W.R., 2000. Controls on effective settling velocity of suspended sediment in the Eel River Plume. Continental Shelf Research, 20:2095-2111. Jackson, G. A. (1990). A model of the formation of marine algal floes by physical coagulation processes. Deep Sea Research 37 (8): 1197-1211 Kasai, M., Brierley, G.J., Page, M.J., Marutani, T., and Trustrun, N.A., 2005. Impacts of land use change on patterns of sediment flux in Weraamaia catchment, New Zealand. Catena, 64:27-60. Kasai, M., 2006. Channel processes following land use changes in a degrading steep headwater stream in North Island, New Zealand. Geomorphology, 81:421-439. Kineke, G.C. and R.W. Sternberg, 1995. Distribution of fluid muds on the Amazon Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 2 1 continental shelf. Marine Geology, 125: 193-233. Kineke, G.C., R.W. Sternberg, J.H. Trowbridge and W.R. Geyer, 1996. Fluid Mud Processes on the Amazon Continental Shelf. Continental Shelf Research, 16: 667-696. Kineke, G.C., Woolfe, K.J., Kuehl, S.A.; Milliman, J.D., Dellapenna, T.M., and Purdon, R.G., 2000. Sediment export from the Sepik River, Papua New Guinea; evidence for a divergent sediment plume. Continental Shelf Research, 20:2239- 2266. Leithold, E.L., and Blair, N.E., 2001. Watershed control on the carbon loading of marine sedimentary particles. Geochimica et Comsmochimica Acta, 65:2231- 2240. Lewis, K.B., Lallemand, S.E., and Carter, L., 2004. Collapse in a Quaternary shelf basin off East Cape, New Zealand: evidence for passage of a subducted seamount inboard of the Ruatoria giant avalanche. New Zealand Journal of Geology and Geophysics, 47:415-429. Lewis, K.B., Pettinga, J.R., 1993. The emerging, imbricate frontal wedge of the Hikurangi Margin. In: Balance, P.B. (Ed.), South Pacific Sedimentary Basins. Sedimentary Basins of the World, vol. 2. Elsevier, Amsterdam, pp. 225-250. Lewis, K.B., Collot, J-V., and Lallemand, S.E., 1998. The dammed Hikurangi Trough: a channel-fed trench blocked by subducting seamounts and their wake avalanches (New Zealand-France GeodyNZ Project. Basin Research, 10:441-468. Litchfield, N.J., and Berryman, K.R., 2005. Correlation of fluvial terraces within the Hikurangi Margin, New Zealand: implications for climate and baselevel • Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. controls. Geomorphology, 68:291-313. Ma, Y., Wright, L.D., and Friedrichs, C.T., in prep. Observations of sediment transport on the continental shelf off the mouth of the Waiapu River, New Zealand: evidence for current-supported gravity flows. To be submitted to Continental Shelf Research. Milligan, T.G., Hill, P.S., and Law, B.A., 2007. Flocculation and the loss of sediment from the Po River plume. Continental Shelf Research, 27:309-321. Milliman, J. D. and Syvitski, J.P.M., 1992. Geomorphic/tectonic control of sediment discharge to the ocean; the importance of small mountainous rivers. Journal of Geology, 100(5): 525-544. Milliman, J.D., Kao, S.-J., 2005.Hyperpycnal Discharge of Fluvial Sediment to the Ocean: Impact of Super-Typhoon Herb (1996) on Taiwanese Rivers. The Journal of Geology, 113:503-516 Mulder, T. and Syvitski, J. P. M., 1995. Turbidity currents generated at river mouths during exceptional discharge to the world's oceans. Journal of Geology 103: 285-299. Ogston, A.S., and Sternberg, R.W., 1999. Sediment-transport events on the northern California continental shelf. Marine Geology, 154:69-82. Ogston, A.S., D.A. Cacchione, D.A., R.W. Sternberg and G.C. Kineke, 2000. Observations of storm and river flood-driven sediment transport on the northern California continental shelf. Continental Shelf Research, 20: 2141-2162. Page, M.J., and Trustrum, N.A., 1997. A late Holocene lake sediment record of the erosional response to land use change in a steepland catchment, New Zealand. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Zeitschrift fur Geomorphologie N.F. 413, 369-392. Page, M., Trustrum, N., Brackely, H., Gomez, B., Kasai, M.,Marutani, T., 2001. Waipaoa River North Island, New Zealand. In: Marutani, T., Brierley, G., Trustrum, N., Page, M. (Eds.), Source-to-Sink Sedimentary Cascades in Pacific Rim Geo-Systems, pp. 86-100. Riebesell, U., 1991. Particle aggregation during a diatom bloom. II. Biological aspects. Marine Ecology Progress Series, 69:281-391. Schaffner,L. C., Diaz,R. J., Olsen,C. R., Larsen,I. L., 1987. Faunal characteristics and sediment accumulation processes in the James River estuary, Virginia Estuarine, Coastal and Shelf Science, 25: 211-226. Scully, M.E., C.T. Friedrichs and L.D. Wright, 2003. Numerical modeling results of gravity-driven sediment transport and deposition on an energetic continental shelf: Eel River, Northern California. Journal of Geophysical Research, 108(C4):3120, doi: 10.1029/2002JC001467. Shapiro, R., 1970. Smoothing, filtering, and boundary effects. Review of Geophysics and Space Physics, 8:359-387. Sommerfield, C.K. and C.A. Nittrouer, 1999. Modem accumulation rates and a Sediment budget for the Eel shelf: a flood-dominated depositional environment. Marine Geology, 154: 227-242. Sommerfield, C.K., Nittrouer, C.A., and Alexander, C.R., 1999.7Be as a tracer of flood sedimentation on the northern California continental margin. Continental Shelf Research, 19:335-361. Syvitski, J.P.M., L. Pratson, P. Wiberg, M. Steckler, M. Garcia, R. Geyer, C.K. Harris, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 2 4 E. Hutton, J. Imran, H. Lee, M. Morehead, and G. Parker. 2006. Prediction of Margin Stratigraphy. Chapter XI of Continental-Margin Sedimentation: Transport to Sequence, C. Nittrouer, et al. (eds), Blackwell Press, (in press). Traykovski, P., W.R. Geyer, J.D. Irish and J.F. Lynch, 2000. The role of wave-induced density-driven fluid mud flows for cross-shelf transport on the Eel River continental shelf. Continental Shelf Research, 20: 2113-2140. van Rijn, L.C., 1984. Sediment transport, part II: suspended load transport, ASCE J.Hydr. Engr., 110(11):1613-1638. Wadman, H.M., and McNinch, J.E., 2006. Inner shelf processes affecting shelf-wide sedimentation, Waiapu River, New Zealand. Ocean Sciences Meeting, Honolulu, Hawaii. Abstract OS16A-33. Wadman, H.M. and McNinch, J.E., in review. Spatial variation on the inner shelf of a high yield river, Waiapu River, New Zealand: Implications for fine sediment dispersal and preservation. Continental Shelf Research. Wadman, H.M., McNinch, J.E. and Kuehl, S.A., 2006. Fine sediment sequestration on an active inner shelf, Waiapu River, New Zealand. EOS, Transactions (AGU), 87(36): OS23B-1644. Walcott, R.I., 1987. Geodetic strain and the deformational history of the North Island of New Zealand during the late Cainozoic. Philosophical Transactions Royal Society of London, A321:163-181. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Wheatcroft, R.A., and Borgeld, J.C., 2000. Oceanic flood deposits on the northern California shelf: large scale distribution and small-scale physical properties. Continental Shelf Research, 20:2059-2066. Wright, L.D., Friedrichs, C.T., 2006. Gravity Driven Sediment Transport on Continental Shelves: A Status Report. Continental Shelf Research 26,2092- 2107. Wright, L.D., Ma ,Y., Scully, M, Friedrichs, C.T., 2006.Observations of Across Shelf Sediment Transport During High Energy Flood Events off the Mouth of the Waiapu River, New Zealand Eos Trans. AGU, 87(36), Ocean Sci. Meet. Suppl., Abstract OS11L-03. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 2 6 Tables Sensitivity Experiment Design Variable Action Range Varied flocculated to 0 -100 % flocculated to Flocculated Fraction unflocculated ration unflocculated sediment Multiplied wave height by factor 0.4 - 2.5 m mean wave Wave Height of 0 .2 - 1.2. height. Mean Wave Height Used steady wave height for run. 2 -5 m mean wave height. High or low wave height during 1.2 - 2.1 m mean wave Wave timing flood. height. 39 NE, SW219, 275 NW, Kept wind speed from record, Wind Direction and 180° times record altered direction. direction. Mean 0.1 Pa. Kept wind direction from record, Wind Speed multiplied wind speed by 0.1 -2. Mean 0.001 - 0.6 Pa. Varied the sediment load using Sediment Load 0.5 MT - 23 MT Hicks et al., (2004) rating curve. Varied the peak sediment Sediment concentration, kept river 8.5 g/L - 60 g/L Concentration discharge constant with record. Table la. Summary of sensitivity experiments. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 2 7 Sensitivity Experiment Design Variable Case Name Action 100 unfl 0% flocculated sediments 25fl75unfl 25% flocculated sediments Flocculated Fraction 50fl50unfl 50% flocculated sediments 75fl25unfl 75% flocculated sediments lOOfloc 0% flocculated sediments Low waves 0.2 x wave record Half waves 0.5 x wave record Wave Height Original Observed wave record High waves 1.2 x wave record 2 m Waves held at 2m 3 m Waves held at 3m Mean Wave Height Original Observed wave record 5 m Waves held at 5m On Waves high at peak flood Wave timing Off Waves low at peak flood Original Observed wave record NE39 Wind direction at NE 39° Oppwind Wind direction opposite observed Wind Direction Original Observed wind record NW275 Wind direction at NW 275° SW219 Wind direction at SW 219° Winds 1 tenth 0.1 x wind speed record Half winds 0.5 x wind speed record Wind Speed Original Observed wind record windsHigh 2 x wind record 20 g/L, 0.5 MT Reduced load, used Hicks’ curve Observed discharge, used Hicks’ Sediment Load Original: 50 g/L, 6.3 MT curve 60 g/L, 23.4 MT Increased load, used Hicks curve 60 g/L, 23.4 MT Increased load, used Hicks curve Sediment 30 g/L, 11.7 MT 1/2 x 60g/L case Concentration 15 g/L, 5.9 MT 1/4 x 60g/L case 8.5 g/L, 3.3 MT 1/7 x 60g/L case Table lb. Summary of sensitivity experiments and case nomenclature. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 2 8 Sediment Budget for Sediment Load and Concentration Experiments Case Inner Middle Outer Shelf Shelf Shelf Shelf Shelf Break Average Final % % fluvial of fluvial budget budget Original 2.9 3.9 1.2 0.0198 10 1.3 Case 20 g/L, 0.05 0.05 0.02 0.0004 4 0.2 0.5 MT 60 g/L, 49.0 20.7 6.9 0.01319 24 15 23.4 MT 30 g/L 10.4 9.9 2.4 0.0701 14 4 15 g/L 1.3 4.6 1.6 0.0383 10 2 8.5 g/L 0.3 1.8 1.0 0.242 7 1.7 Sediment Budget for Aggregate Composition Experiment 100 floe 6.2 5.4 1.4 0.0054 15 1.9 Original 2.9 3.9 1.2 0.0198 10 1.3 60fl, 1.5 3.1 1.1 0.0283 7 1.1 40unfl 50 fl 0.9 2.5 1.0 0.0339 5 1.0 50 unfl 40fl, 0.5 1.8 0.8 0.0394 4 0.96 60unfl 25fl, 0.2 0.8 0.6 0.5988 2 0.94 75unfl 100 unfl 0.0 0.01 0.2 0.0578 0.003 0.96 Table 2. Time averaged deposition on the inner, middle, and outer shelf and the shelf break zone, >130 m, compared with the time-averaged budget for the whole shelf and the final sediment deposit at the end of the model runs. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 2 9 Sediment Budget for Wave Experiments Case Inner Middle Outer Shelf Shelf Shelf Shelf Shelf Break Average Final % % fluvial of fluvial budget budget Original 2.9 3.9 1.2 0.0198 10 1.3 Case high 2.3 3.1 2.0 0.02 9 1.1 waves half 9.8 4.5 0.5 0.01 18 5.8 waves low waves 36.3 0.5 0.3 0.01 22 46.5 Off waves 3.5 3.5 1.0 0.02 41 3.2 On waves 35.3 4.7 1.0 0.01 48 45.1 5m waves 2.4 1.8 2.4 0.6 8 0.4 3 m 3.2 3.5 0.9 0.02 9 0.7 waves 2m waves 5.2 4.0 0.6 0.02 12 1.0 Sediment Budget for Wind Experiments high 0.6 0.01 0.002 0.0003 1 0.1 winds half 4.7 25.1 5.2 0.007 19 50.7 winds low winds 4.1 28.9 3.7 0.0007 40 54.2 SW219 2.1 0.06 0.02 0.0 4 0.1 NE39 3.8 0.9 0.01 0.0012 6 0.2 opposite 2.1 5.6 2.4 0.005 6 1.1 wind NW275 3.0 2.9 8.8 0.0 17 15.5 Table 2 con’t. Time averaged deposition on the inner, middle, and outer shel ' and the shelf break zone, >130 m, compared with the time-averaged budget for the whole shelf and the final sediment deposit at the end of the model runs. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 3 0 Bed Shear Stress Statistics Wave Cases Case Mean bed low bed High bed Standard Mean Mean shear shear shear deviation wave wind stress height stress Original 0.603 0.0112 1.672 0.419 2.06 0.991 Case high 0.789 0.015 2.239 0.556 2.47 0.991 waves half 0.230 0.004 0.596 0.151 1.03 0.991 waves low waves 0.079 0.001 0.197 0.051 0.41 0.991 Off waves 0.494 0 1.190 0.490 1.72 0.991 On waves 0.373 0.006 1.076 0.420 1.22 0.991 5m waves 2.082 0.018 2.425 0.291 4.9 0.991 3 m 0.951 0.009 1.190 1.181 2.9 0.991 waves Bed Shear Stress Statistics Wind Cases high 0.817 0.012 2.054 0.539 2.06 0.551 winds half 0.559 0.012 1.673 0.419 2.06 0.023 winds low winds 0.5542 0.012 1.675 0.399 2.06 0.001 SW219 0.674 0.012 1.902 0.447 2.06 0.996 NE39 0.627 0.012 1.810 0.429 2.06 0.996 opposite 0.613 0.012 1.659 0.423 2.06 0.991 wind NW275 0.591 0.012 1.783 0.422 2.06 0.996 Table 3. Bed shear stress statistics, mean wave heights, and mean wind stresses fir the wave and wind experiments. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 231 Mean Across-Shelf Flux Wave Experiment Case Inner Middle Outer Shelf Mean % wbl flux Shelf Shelf Break Original 2820 1660 417 45 38 Case high 2830 1740 590 54 42 waves half 2580 939 164 30 20 waves low waves 1510 299 125 25 3 Off waves 2760 1560 390 46 36 On waves 2500 620 137 21 76 5m waves 2940 1910 955 107 49 3 m 2780 1560 389 46 36 waves 2m waves 2730 1240 210 35 26 Mean Across-Shelf Flux Wind Cases high 1540 186 25 2 2 winds half 3.390 2170 379 34 45 winds low winds 3470 2120 346 5 44 SW219 2240 134 0.02 0 10 NE39 1210 874 728 402 6 opposite 3030 2140 665 94 34 wind NW275 3270 1760 331 0.07 26 Table 4. Time-averaged lux for each zone of the s lelf and the average portion of the across-shelf flux transported within the wave boundary layer for the whole shelf. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 3 2 Mean Across-Shelf Flux Sediment Load and Concentration Experiments Case Inner Middle Outer Shelf Mean % wbl flux Shelf Shelf Break Original 2820 1660 417 45 38 Case 20 g/L, 177 79 19 0.6 41 0.5 MT 60 g/L, 10,100 5230 1230 172 28 23.4 MT 30 g/L 5440 3060 794 96 33 15 g/L 2720 1760 512 56 36 8.5 g/L 1510 983 328 36 37 Mean Across-Shelf Flux Aggregate Composition Experiment 100 floe 2870 1760 345 16 45 Original 2820 1660 417 45 38 60fl, 2780 1580 454 61 33 40unfl 50 fl 2770 1510 474 72 29 50 unfl 40fl, 2760 1410 491 83 25 60unfl 25fl, 2730 1230 502 97 18 75unfl 100 unfl 2690 930 482 119 1 Table 4 con’t. Time-averaged flux for each zone o: ■ the shelf and the average portion of the across-shelf flux transported within the wave boundary layer for the whole shelf. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 3 3 Mean Along-Shelf Flux Wave Experiment Case Mean Mean % Mean % in % Fluvial Flocculated Suspension Budget Exported Original Case 2583 63 99 64 high waves 2520 62 99 63 half waves 2770 66 99 69 low waves 2110 55 99 52 Off waves 2660 64 99 66 On waves 1190 22 99 30 5m waves 2590 63 99 64 3 m waves 2650 64 99 66 2m waves 2750 65 99 68 Mean Along-Shelf Flux Wine Experiment high winds 3500 72 99 87 half winds 1300 30 99 32 low winds 972 5 99 24 SW219 3010 67 99 75 NE39 -5.21e-6 58 99 58 opposite 2860 67 99 71 wind NW275 2590 62 98 64 Table 5. Time-averaged along shel ' flux, mean % of along-shelf flux that is flocculated, mean percent of along-shelf flux in dilute suspension, and the percent of the fluvial budget exported from the shelf. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 3 4 Mean Along-Shelf Flux Sediment Load and Concentration Experiments Case Mean Mean % Mean % in % Fluvial Flocculated Suspension Budget Exported Original Case 2583 63 99 64 20g/L, 0.5 MT 183 57 99 57 60 g/L, 23.4 9440 63 99 63 MT 30 g/L 4960 65 99 66 15 g/L 2430 65 99 65 8.5 g/L 1320 63 99 62 Mean Across-Shelf Flux Sediment Aggregat e Composition Experiment 100 floe 2240 100 99 56 Original Case 2583 63 99 64 60fl, 40unfl 2790 46 99 69 50 fl 2940 36 99 73 50 unfl 40fl, 60unfl 3090 27 99 77 25fl, 75unfl 3340 15 99 83 100 unfl 3770 0 99 94 Table 5 con’t. Time-averaged along shelf flux, mean % of along-shelf flux that is flocculated, mean percent of along-shelf flux in dilute suspension, and the percent of the fluvial budget exported from the shelf. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 3 5 Figure Captions Figure 1. Location map of East Cape, New Zealand. Waiapu River basin, Hicks Bay wind station, and study area are indicated. Large-scale study area map shows gravity core sites in black circles and tripod locations in gray circles. Figure 2. The suspended sediment rating curve created by Hicks et al. (2004), and a discharge record from the Gisborne District Council. The red dashed lines indicate a 40 g/L threshold for gravity-driven transport (Mulder and Syvitski, 1995). Figure 3. A) The ECOMSED model grid. The areas outlined in white were used to calculate sediment mass distributions, flux, and flux convergence. Areas A through E denote the inner shelf (A and B), mid-shelf, outer shelf, and shelf break. B) The wind, discharge, and wave inputs for the baseline case. The winds are from Hicks Bay; the discharge data is from the GDC, and the waves are from the 60 m tripod site shown in Figure 1. Figure 4. Bed shear stresses, depth averaged currents, and sediment concentrations at 1 mab were compared for the baseline case and the calculations from the 60 m tripod. Figure 5. The panels in A and B show bed sediment, sediment in the wave boundary layer, and sediment in the water column, from left to right for two time periods in the baseline case. Wind direction and magnitude are shown in the bed sediment panels. The Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 3 6 red arrows in the middle and right hand panels show current magnitude and direction. The wave and discharge conditions at each time are shown in the bottom panel. Figure C shows sediment distribution for the baseline case in the wave boundary layer, water column, and seabed. Figure 6 A and B. Time-averaged sediment mass in suspension, in the wave boundary layer, ratio of sediments in suspension to sediments in the wave boundary layer, and sediments deposited on the seabed represented as a fraction of the total fluvial budget. Each graph shows the results of one experiment, each point represents the time- averaged results for a case. Each experiment is plotted against the variable that was tested. For example, the aggregate composition cases were plotted against the percent of the fluvial budget that was flocculated. Both wind and wave experiments were plotted against mean bed shear stresses. Figure 7. A) Plan view of sediment deposition for each steady wave experiment case on day 8 of the model run .Wind direction is indicated in the upper left hand comer. B) Plan view of sediment deposition for each wave height experiment case. The record of sediment mass in suspension, in the wave boundary layer, and deposited on the bed for the wave height case is shown in the bottom panel. Figure 8. A) The sediment mass in suspension, in the wave boundary layer, and deposited on the seabed for each model case. Wave heights were high during peak fluvial discharge for the “on” case, and low during peak discharge for the “off’ case. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 3 7 B) Plan view of each wave-timing case showing sediments deposited, in the wave boundary layer, and in suspension after the flood. Figure 9. A) The percent of the fluvial load exported from the shelf area is plotted against the mean bed shear stress (Pa) for the wave height, timing, and steady cases. B) The percent of the fluvial load exported from the shelf area is plotted against the mean bed shear stress and the mean along-shelf wind stress for the two wind experiments. The positive values for the wind stress plot correlate with winds blowing to the north east, whereas negative values indicate winds to the south-west. The regression lines show a correlation between wind direction and along-shelf wind stress in the along- shelf wind plot, and between both wind direction and speed in the mean bed shear stress plot. The red plus sign shows the baseline, or original case. Figure 10. The plan view of deposition for each wind speed (A) and wind direction (B) case at model day 8 Figure 11. A) The fraction of the fluvial load in suspension, in the wave boundary layer, and deposited on the seabed for each floe fraction experiment case. B) Plan view of selected aggregate composition cases showing sediments deposited. Figure 12. A) The fraction of the fluvial load in suspension, in the wave boundary layer, and deposited on the seabed for each sediment load case. B) Plan view of each load case showing seabed deposition. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 3 8 Figure 13. A) The sediment mass in suspension, in the wave boundary layer, and deposited on the seabed for each sediment concentration case. B) Plan view of each concentration case showing seabed deposition. Figure 14. A) The time-averaged flux for each along-shelf row was calculated to assess across- and along-shelf flux with depth across the shelf for the baseline case. The total (black), suspended (blue), and gravity-driven (red) flux are shown. The flocculated (blue dot) and unflocculated (blue dash) portions of the suspended flux are shown as well. B) Time-averaged sediment flux with depth across the shelf for across-shelf suspended sediment flux, across-shelf wave boundary layer flux, and along-shelf suspended sediment flux. Values were calculated by taking the time-averaged flux for each along-shelf model-grid row. Selected results of the floe fraction experiment are plotted. Figure 15. Time-averaged sediment flux with depth across the shelf for across-shelf suspended sediment flux, across-shelf wave boundary layer flux, and along-shelf suspended sediment flux. Values were calculated by taking the time-averaged flux for each along-shelf model-grid row. Sediment concentration and load experiment results in 15A and 15B, respectively. Figure 16. Time-averaged sediment flux with depth across the shelf for across-shelf suspended sediment flux, across-shelf wave boundary layer flux, and along-shelf Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 3 9 suspended sediment flux. A) The wave height experiment, B) the steady wave experiment, and C) the wave timing experiment. The bed-shear stress records for the wave timing experiment are also shown in 16C. Figure 17. Time-averaged sediment flux with depth across the shelf for across-shelf suspended sediment flux, across-shelf wave boundary layer flux, and along-shelf suspended sediment flux. A) The wind speed experiment and B) the wind speed experiment. Figure 18. Mean along shelf fluxes versus mean bed shear stress (Pa) for selected wind and wave experiments. Positive flux values denote sediment flux to the north. Figure 19. Along shelf fluxes and along shelf wind stresses (Pa) for the A) wind speed experiment and B) wind direction experiment. Positive flux indicates flux to the north of the river mouth. Positive wind stresses represent winds to the north, whereas negative wind stresses denote winds to the south. Figure 20 A) Mean along shelf fluxes versus % floe fraction for the sediment aggregation experiment. Along shelf fluxes and along shelf wind stresses (Pa) for the B) floe fraction experiment, sediment concentration, and sediment load case. Positive flux indicates flux to the north of the river mouth. C) Positive wind stresses represent winds to the north, whereas negative wind stresses denote winds to the south. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 4 0 Figure 21. Millions of tons exported from the shelf for the sediment load, concentration, and floe fraction experiments versus the variable tested. Figure 22. Millions of tons exported from the shelf for the wind and wave experiments versus mean bed shear stress (Pa). Figures 23 -25. Time-averaged flux convergence (negative values) and flux divergence (positive values) calculated using a version of the Exner equation. Changes in flux convergence were estimated in the along- and across-shelf directions for each along shelf row. The panels for each experiment show the across shelf patterns of across- shelf flux convergence, along shelf flux convergence, across-shelf suspended sediment flux convergence, and across-shelf wave boundary layer flux convergence. Figure 26. Time-series of the amount of sediment (kg) in suspension (blue), in the wave boundary layer (red), and deposited on the seabed (black) for each zone of the shelf for the original, baseline case. Figure 27. This figure was adapted after Lewis et al., (2004). Quaternary structural trends and main locus of deposition are shown in red. The red zone outlines between 800- 1000 m of deposition. Faults, mostly in Neogene strata, in orange. The main locus of Holocene deposition in green, outlining ~100m of thickness. The main locus of deposition over the last 100 years (>2.7 cm/y) is outlined in blue (Kniskern, Chapter 1). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ~ - ~ ~ 8 r ~ ¥-1 r C"') C"') 0 8 8 ¥-1 ~ lfl I I 10 A 5 N 178°50'00"E A • :::::...... , Kilometers • 2.5 e '\ 0 ,. .,. £ V l / 178°4(l'OO"E J ~I \ 178°3boOO"E I \ I -C Island •Tatapouri I Zealand 176Df: New North 50km '"I1 - qci" I CD CD CJl 0 c c. ;:::;: a· 3 :iE ..., u;· ::::r - ? CD CD 0 0 0"" c c ;:::). g c. a· ;:::;: :::J ..., ..., ..., ..., ::::r ::::r 11 CD CD 0 0 0 (') :"""' :::J :::J ..., :iE ::::r ::::r - - CD CD CD CJl 0 c c. (') c. ;:::;: a· 3 ..., :iE ..., u;· ::::r :::0 - Reproduced with permission of the copyright owner. Further reproduction prohibited"0 without permission. "0 "0 "0 '< 100 10000 100000 i io«« ^ 3/15/00 10/1/00 4/19/01 11/5/01 5/24/02 12/10/02 6/28/03 1/14/04 8/1/04 2/17/05 9/5/05 B. 80000 — 40000 d 3 a «S S- e 8 c © © 8 e © 120000 8 s o C "2 TS " N- • N- (JQ Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 243 Boundary - 37.75 37.85 140 - 37.9 37.95 East undary 38.05 38.15 South Boun 178.4 178.5 178.6 178.7 178.9 B . ,r...... f 10 | o Iff -10 1 1 Waiapu River Discharge 2000 i t 1 4 1 1 4 i I t 1 421000 *E 0 Waiapu River Shelf Waves 5 E 2.5 0 —I_____ !_ 1 * 1 --- -I * 1 ...... ( ---1------13 15 17 19 21 23 25 27 29 31 Date, duty 2004 Fig-3 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Bed Shear Stress 2 Model Tripod A 0 s*— 07/13 07/18 07/23 07/28 Depth averaged along-shelf currents 100 - 1001— 07/13 07/18 07/23 07/28 Depth averaged across-shelf currents -50*— 07/13 07/18 07/23 07/28 Sediment concentration ~1mab 60 m tripod site 07/13 07/18 07/23 07/28 Date 2004 2 4 5 Bed Sediment WBL Sediment Suspended Sediment -37.7 f 'A ? /' i i .2 m.'s siLit/i,\ -37.8 rM/ i' ) o ■/ .' -v;: / -■ ii. />»v • <-■ -37.9 •***, * ■ - ~- ///«. "l/ /* t V * yr1 . / 1 v -. -38.0 * *. M:. W . , '/ 'V . ;:r- -38.1 178.4 178.5 178.6 178.4 178.5 178.6 178.4 178.5 178.6 -37.7 B. •37.8 V/l! -37.9 -38.0 i N*1 W J/ V ' I m'v '* -38.1 178.4 178.5 178.6 178.4 178.5 178.6 178.4 178.5 178.6 Waiapu Discharge 2000 B 1000 S 0 0 2 4 6 8 10 12 14 16 18 20 Waiapu Waves 6 3 S "— ______' 0 10 12 14 16 18 20 X10 2.5 suspended mass wave boundary layer mass bed mass 1.5 0.5 07/13 07/15 07/17 07/19 07/21 07/23 07/25 07/27 07/29 07/31 D ate 2004 Fig. 5 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. to OS T r T T T Sediment Concentration (g/L) 0 20 40 60 - 0 80 — 60 40 — 20 100 0 mean suspended % Q mean % wbl ------□ Q ------mean % deposition SedimentLoad (MT) jlbk ^ ^ ^ ratio susp. to % wbl Q Q SedimentLoad Experiment Sediment Concentration Experiment 0 5 10 15 20 25 60 80 40 20 100 Floe Fraction Experiment 0 0 20 40 60 80 100 % % high settling velocity material 0 80 60 20 40 100 > s ec s > OS -C -o • mm • c era Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. I 1 T* A I w 1 I 1 ■ I 1 A I I I Wind Direction Experiment mean bed shearstress (Pa) mean bed shearstress (Pa) 0 0 0.2 0.4 0.6 0.8 0 0.4 0.8 1.2 1.6 2 _ _Steady Wave Experiment - 0 0 — 0 0 - 80 80 — 80 — 60 — 60 60 — 40 — 20 — 40 — 20 100 100 — jqq % % wbl % ♦ ratio susp. to wbl £ mean suspended ■ ■ I ■ I ■ I ’ I ----- Wave Timing Experiment -A o A — mean deposition% o mean ♦ mean bed shearstress (Pa) 0 0 0.2 0.4 0.6 0.8 0 ♦ - O # 80 60 20 40 100 0.2 0.2 0.4 0.6 0.8 0.2 0.2 0.4 0.6 0.8 Wave HeightExperiment Wind Speed Experiment mean mean bed shearstress (Pa) mean bed shearstress (Pa) 0 0 ^ 100 100 9 9 M) O > 9 9 V DC CD' o ON "O A "eS 53 so* • m m 53 SM Xi .2 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 248 A. 0.1 20 cm •37.7 2m waves 3m wavesoriginal 5m waves ■37.8 v/ir 37.9 38.0 130 150 .« 38.1 178.4 178.6 178.4 178.4 178.6 178.4 178.6178.6 B. low waves half waves original high waves ■37.7 , ' / . i l l j ■37.8 v.-ii '* ^FV P ■■ ' ' JP? ' •37.9 ff / • /""/ ■. : Y : '< : : ■ i •38.0 38.1 178.4 178.6 178.4 178.6 178.4 178.6 178.4 178.6 x 10 Suspended Sediment ______ Low 0.2 half-waves High-waves 2 original 0 07/13 9 07/15 07/17 07/19 07/21 07/23 07/25 07/27 07/29 07/31 . x 10 Wave Boundary Layer Sediment 07/13 9 07/15 07/17 07/19 07/21 07/23 07/25 07/27 07/29 07/31 _x10 Sediment Mass Deposited 07/13 07/15 07/17 07/19 07/21 07/23 07/25 07/27 07/29 07/31 Date 2004 Fig. 7 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 4 9 A. Suspended Sediment 0 ------1 1------L 07/13 07/15 07/17 07/19 07/21 07/23 07/25 07/27 07/29 07/31 x 10 Wave Boundary Laver Sediment original ±£.o> 07/13 07/15 07/17 07/19 07/21 07/23 07/25 07/27 07/29 07/31 x 10' 6 Seabed Sediment 4 o> 2 0 i------1------1------1— 07/13 07/15 07/17 07/19 07/21 07/23 07/25 07/27 07/29 07/31 B. off on original -37.7 -37.8 -37.9 -38.0 -38.1 178.6 178.4178.4 178.6 178.4 178.6 Fig. 8 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 250 wave wave height wave V wave-timing ± speed wind A wind direction + case baseline AAA A A A steady 60 80 100 40 R-squared = 0.803843 20 20 30 40 50 60 70 % fluvial budget% exported from shelf % fluvial budget% exported from shelf 0.6 0.5 0.9 0*5 0.4 0.6 0.65 0.35 2 0.7 — v s I v S I 0.55 & U 43 0.45 T R-squared = 0.855367 ▲ ▲ ▲ 20 40 60 80 100 52 56 60 64 68 72 fluvial budget exported from shelf % 1 0 2 % fluvial budget% exported from shelf 1.5 0.5 2.5 0.08 0.08 — 0.04 A -0.08 -0.04 V s Vt S X u CS 0> i/l S <8 ■M w 4= TJ -o v s £ s e a X X 2 X a X W> m m ■ w 'O 13 43 *c3 C . B. A. VO 3 qq" Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 251 s \ £ s * O 00 © 00 00 178.4 178.4 178.5 178.6 178-4 178.5 178.6 17S.4 178.5 178.6 178.4 178.5 178.6 178.4 178.5 178.6 < « Fig. 10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 252 A. Suspended Sediment O) 07/13 q 07/15 07/19 07/21 07/23 07/25 07/27 07/2907/17 x 10 Wave Boundary Layer Sediment 6 T —T———T — - | ' lf lffl| ■— f | O) ■AC 2 0 JU 07/13 0 07/15 07/17 07/19 07/21 07/23 07/25 07/27 07/29 07/31 x 10 Seabed Sediment 6i T - lOOfl - 100unfl 75fl25unfl 25fl75unfl 50fl50unfl 07/13 07/15 07/17 07/19 07/21 07/23 07/25 07/27 07/29 07/31 Date 2004 B. 100% unfl 25%floc 75% unfl original 100% floe -37.7 -37.8 Will win I -37.9 i -38.0 i -38.1 178.6178.4 178.6 178.4 178.6178.4 178.4 178.6 Fig. 11 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced B. fraction fluvial budget 0.4 0.2 0.6 0.8 0.2 0.4 0.6 0.8 71 0/5 71 0/9 72 0/3 72 0/7 72 07/31 07/29 07/27 07/25 07/23 07/21 07/19 07/17 07/15 07/13 71 0/5 71 0/9 72 0/3 72 0/7 72 07/31 07/29 07/27 07/25 07/23 07/21 07/19 07/17 07/15 07/13 07/13 -37.8 -37.9 -38.0 -38.1 S®: m 7. 178.6 178.4 71 0/7 71 0/1 72 0/5 72 0/9 07/31 07/29 07/27 07/25 07/23 07/21 07/19 07/17 07/15 J~/ / / / D I ( : _ \ : ; — WaveBoundarv Laver Sediment \ : / : ; \ ' ; P ' ' ' r' i i i Suspended Sediment Suspended * ' ;• * 7. 7. 178.4 178.6 178.4 .. > ‘* SeabedSediment ■ 1 v 1 r 1 < ■ ' ' ■■ ■ ■■ ■ ' i ' 1 1 ■ i 1 . i 1 ' /■ii U ' 1 V' ' fz/f ; \ _ \ ' ; f / z f / f i f c f . > .■: Date 2004 1 ■ , ^ 1 1 ' | II | 1 1 \ > jpt • t ’j p If t If K ■ y,r,r -, -/ ■ / , - ■ - - 1 / r X , , , f t , * ,CXr ■. ii * / p I m : / / / / : ; , '■ •/ \ •, ) ) \ * \ 178.6 original ; / ig.12 F 253 254 A Suspended Sediment 3 07/13 07/15 07/17 07/19 07/21 07/23 07/25 07/27 07/29 07/31 Wave Boundary Laver Sediment > 0.4 15 g/L 8.5 g/L ° 0.2 S 07/13 07/15 07/17 07/19 07/21 07/23 07/25 07/27 07/29 07/31 0.6 Seabed Sediment 07/13 07/15 07/17 07/19 07/21 07/23 07/25 07/27 07/29 07/31 Date 2004 B. 60 g/L, 23.4 MT 30 g/L, 17.1 MT 15 g/L, 5.9 MT 8.5 g/L, 3.3 MT -38.1 178.4 178.6 178.4 178.6 178.4 178.6 178.4 178.6 Fig. 13 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 255 Time-averaged Across Shelf Flux . 3000 total — suspended 422000 — unfl suspended floe suspended — wbl u-1000 4060100 120 140 160 180 200 Time-averaged Along Shelf Flux 300 total suspended ST 200 unfl suspended floe suspended wbl E 100 80 100 120 140 Depth Across-shelf (m) 3000 Time-averaqed r B i Across i Shelf "" Suspended i..•*"..—T" Flux ... ;= 1000 20 60 80 100 120 140 160 180 200 3000 T T —Time-averaaed T— f i Across i Shelf i— wbl ------Flux1 i . >1...... 5>2000 1000 40 60 100 120 140 160 180 200 Time-averaqed Alonq Shelf Suspended Flux 600 i ...... I a i------a i.... r"L------i i - 100unfl - 25fl75unfl S' 400 — 50fl50unfl — 75fl25unfl u_ 200 —- 100floc 60 80 100 120 140 160 180 200 Depth acrossshelf (m) Fig. 14 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 256 Time-averaged Across Shelf Suspended Flux 15000 M 510000 Ii 5000 40 60 80 100 120 140 160 200 Time-averaged Across Shelf wbl Flux 15000 I5*10000 rr 5000 40 60 80 100 120 140 160 200 Time-averaged Along Shelf Suspended Flux 2000 60g/l 60 80 100 120 140 200 B. Depth across-shelf (m) Time-averaged Across Shelf Suspended Flux 15000 •§,10000 3 5000 200 Time-averaaed Across Shelf wbl Flux 15000 i- r •■-r - ...... -!------1— •s,5 1 0 0 0 0 X E 5000 40 60 80 100 120 140 160 200 2000 Time-averaaed Along Shelf Suspended Flux 0.5MT O) original "JT1 0 0 0 23.4 MT 3 60 80 100 120 140 160 180 200 Depth across-shelf (m) F ig.15 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. poue wt pr sin fte oyih onr Frhrrpouto poiie ihu pr ission. perm without prohibited reproduction Further owner. copyright the of ission perm with eproduced R . A Flux (kg/s) Flux (kg/s) Flux (kg/s) • F|ux ^kg/s^ Flux (kg/s) Flux (kg/s) 50 ■ 1500 2000 2000 1000 1000 2000 2000 1000 1500 1000 1500 500 200 300 200■ 400 100 500 500 500 20 20 20 20 20 0 6 8 10 120 100 80 60 40I 08 10 2 140 120 100 80 40 40 06 8i 60 40 i i i i i i i r i i ------Time-averaqed Across Shelf Suspended Flux Suspended Shelf Across Time-averaqed Time-averaged Across Shelf Suspended Flux Suspended Shelf Across Time-averaged Time-averaqed Along Shelf Suspended Flux Suspended Shelf Along Time-averaqed Time-averaaed Alona Shelf Suspended Flux Suspended Shelf Alona Time-averaaed 1 Time-averaqed Across Shelf wbl Flux wbl Shelf Across Time-averaqed Time-averai 0 0 0 10 4 10 8 200 180 160 140 120 100 80 60 60 010 200 180 140 v 120 100 80 „ 60 60 1 ------Depth across-shelf (m) Depthacross-shelf Depth across-shelf across-shelf Depth 3—i O 0 10 4 160 140 1 120 Fluxwbl r Shelf 100 Across iged JO r ------0 120 100 1 ------1 ------1— 1- ) t i ( 1—E ------4 160 140 4 6 180 160 140 1 1 ------1 ------.... . 160 1— 1 ------original 3m 2m original 5m 180 180 1 ------200 200 200 200 i 16 Fig 7 5 2 poue wt pr sin fte oyih onr Frhrrpouto poiie ihu pr ission. perm without prohibited reproduction Further owner. copyright the of ission perm with eproduced R . c Flux (kg/s) Flux (kg/s) 3000 2000 2000 3000 1000 o> -500 -500 300 100 20 ■ 40 71 0/5 71 0/9 72 0/3 72 0/7 72 07/31 07/29 07/27 07/25 07/23 07/21 07/19 07/31 07/17 07/29 07/15 07/13 07/27 07/25 07/23 07/21 07/19 07/17 07/15 07/13 0* 2 1 ------t ft 20 20 40 06 8 60 40 Time-averaged Across Shelf Suspended Flux Suspended Shelf Across Time-averaged - r Time-averaged Along Shelf Suspended Flux Suspended Shelf Along Time-averaged I ■i ■«- ■ 1 i I ■ Time-averai Waiapu River Sediment Discharge RiverSediment Waiapu 0 0 0 10 4 160 140 120 100 80 60 Depth across-shelf (m) across-shelf Depth Bed Shear Stress Shear Bed 0 0 120 100 80 Date 2004 Date Across Shelf wbl Flux wbl Shelf Across 0 10 4 160 140 120 100 --- fS -W. -rW fcS. I 4 6 180 160 140 ______original on - — off on — off — original 180 180 ----- i. 16c Fig. 200 200 200 258 poue wt pr sin fte oyih onr Frhrrpouto poiie ihu pr ission. perm without prohibited reproduction Further owner. copyright the of ission perm with eproduced R A. Flux (kg/s) Flux (kg/s) Flux (kg/s) Flux (kg/s) Flux (kg/s) Flux (kg/s) -1000 2000 2000 3000 3000 2000- 2000 - 1000 1000 1000 - 200 400 600 800 0 0 0 0 0 10 4 10 180 160 140 120 100 80 60 40 Time-averaged Across Shelf Suspended Flux Suspended Shelf Across Time-averaged 0 0 80 60 40 Time-averaged Across Shelf Suspended Flux Suspended Shelf Across Time-averaged Time-avera Time-averaged Along Shelf Suspended FluxSuspendedShelf Along Time-averaged Time-averaaed Across Shelf wbl Flux wbl Shelf Across Time-averaaed —i FluxShelf Time-averaqedwbl Across 0 0 0 10 4 160 140 120 100 80 60 0 0 0 10 140 120 100 80 60 ------Depth across-shelf (m) across-shelf Depth Depth across-shelf (m)Depth across-shelf — a —■ i ■— i— 1— a— ged Along Shelf Suspended Flux Suspended Shelf Along ged 0 10 4 10 8 200 180 160 140 120 100 mp hiii W iw — — ------b—Xii— — 1 ------ 160 1— oppwind original — NW275 SW219 NE39 - - windsHIGH tenth indsl w original halfwinds halfwinds 8 200 180 i. 17 Fig. 200 200 200 200 259 I T ' i p m 11 r i 1 1 1 i I 2 Wave Timing Experiments Wind Direction Experiment 1 1 • mean bed shearstress (Pa) mean bed shearstress (Pa) 0 0 0.4 0.8 1.2 1.6 2 2.4 0.5 0.6 0.7 0.8 0.9 0 0 — 0 1000 1000 — 3000 — -1000 S a I x a 53 5 2000 mean North flux flocculated flux mean North a -A 1 H T T 1 I I 1 I I Wind Speed Experiment 1 I I 1 Wave HeightExperiment mean bed shearstress (Pa) mean bed shearstress (Pa) 0 0.2 0.4 0.6 0.8 0.5 0.6 0.7 0.8 0.9 2000 3000 — i 3 | 1000 x x | 1000 S3 J 2000 53 i—‘ oo 3 qp’ B. A . 3000 — Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 07/31 ■ 1 '( 07/29 07/31 07/29 — NE39 SW219 original NW275 --- — windsHIGH —tenth windsl — halfwinds — original 1 y« 07/27 07/27 07/29 07/31 07/27 07/27 —I 1 07/25 07/25 07/25 07/25 07/27 07/29 07/31 1— tress Date 2004 Date ______1 TT iSirrTsr^, Along Shelf Flux Shelf Along 07/21 07/21 07/23 Along Shelf Flux Shelf Along Alongshelf Wind Stress Wind Alongshelf — ______1 07/19 07/19 07/19 07/19 07/21 07/23 07/25 — I— ___ \\ k ---- 1 07/17 07/17 07/17 07/17 07/19 07/21 07/23 —I ---- 1 T 07/15 07/15 07/15 07/15 07/17 07/19 07/21 07/23 ------x 10 x 1 5 0 2 07/13 4 07/13 07/15 07/13 07/13 07/13 t o / 10 10 1 1 2 0 S. O) O) 0 ^ * a I) B. d S 3 1 ere' Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 6 2 4000 3000 ©ID X G 2000 A mean flux § a A mean flocculated flux 1000 0 0 20 40 60 80 100 %of aggregates that are flocculated x 10 Alona Shelf Flux 100fl 75fl25unfl 50fl50unfl - - 25fl75unfl 07/13 07/15 07/17 07/19 07/21 07/23 07/25 07/2 - - 100unfl x 10 Alona Shelf Flux 60 a/L 07/13 07/15 07/17 07/19 07/21 07/23 07/25 07/27 07/29 07/31 x 10 Along Shelf Flux 10 60g/L, 23.4 MT 5 49g/L, 6.3 MT 20g/L, 0.5 MT 07/13 07/15 07/17 07/19 07/21 07/23 07/25 07/27 07/29 07/31 Along Shelf Wind Stress (0 Q. A -ST“V ------ 07/13 07/15 07/17 07/19 07/21 07/23 07/25 07/27 07/29 07/31 Date 2004 Fig. 20 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. to OS uo load experiment concentration experiment fluvial load (MT) A A A floe fraction experiment 0 0 5 10 15 20 25 0 4 8 16 12 x o a © t 40 60 % flocculated% Aggregate Composition Experiment SedimentLoad and Concentration Experiments 20 0 0 20 40 60 80 100 3.5 5.5 0 o g 4 * 4.5 2 H 5 peak fluvial concentration (g/L) 4 — 16 16 — | 12 — © a S t? N> ‘n w • CTQ Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 264 i ■ i Wave TimingExperiments 111111111111 Wind Direction Experiment m mean bed shearstress (Pa) mean bed shearstress (Pa) 0 0.4 0.8 1.2 1.6 2 2.4 0.5 0.6 0.7 0.8 0.9 5 — 2 H 4 — 6 6 — i 2 — 4 el 3 — g a, 3 B H S. SedimentExported -A Wind Speed Experiment Wave Height Experiment mean bed shearstress (Pa) 0 0.2 0.4 0.6 0.8 mean bed shearstress (Pa) 0.5 0.6 0.7 0.8 0.9 1 6 5 4 2 2 — 6 — I 5 — p. p. 3 I o a B t? g H d to to " t—»• B 00 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 265 ■ 200 200 200 200 100floc - 100unfl - 40fl60unfl - 25fl75unfl - 50fl50unfl - 60fl40unfl - 75fl25unfl i t— „ Depth Across-Shelf (m) Across-Shelf Suspended Flux i i Across-Shelf Flux Convergence 60 80 100 120 140 Across-Shelf Wave Boundary Laver Rux 40 60 80 100 120 140 ^* 20 40 60 80 100 120 140 ■ E o A. ’ > 31 to to Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. poue wt pr sin fte oyih onr Frhrrpouto poiie ihu pr ission. perm without prohibited reproduction Further owner. copyright the of ission perm with eproduced R . C B. cm cm cm cm -2.5 -2.5 1 -5 . 5 -5 . -10 -10 -10 -10 -10 -10 -10 10 -5 -5 -5 5 0 0 0 0 2 0 5 5 I I I I ------...... / A V 20 20 20 20 0 0 0 0 0 10 4 10 180 160 140 120 100 80 60 40 20 140 120 100 80 60 40 20 20 20 1 TT T T 1 1 1------1 ...... ,1 1 '' "1 060 40 06 80 60 40 80 60 40 40 060 40 40 1 ------...... i i i i 1 Across-Shelf Converqence WBL Flux Across-Shelf . .. 0 0 0 10 140 120 100 80 60 60 I Converqence Flux Across-Shelf I I ia I i 'I Across-Shelf Flux Converqence Converqence Flux Across-Shelf 1 1 Alonq-Shelf Flux Converqence ^ Converqence Flux Alonq-Shelf Alonq-Shelf Flux Converqence Flux Alonq-Shelf ------...... — — ^ * i w i .. Depth Across-Shelf (m) Across-Shelf Depth ■ Depth Across-Shelf (m) Across-Shelf Depth ,1 "1 1 80 1 1 ------100 100 100 100 100 1 III 1 mm—m m ------. | | |— .... ,1 120 120 2 101010 200 180 160 140 120 120 120 ■1 ■■— 1------1 ---- ... a—■— ■■■■■In- 4 180 140 140 140 140 ,1 - ■ ■ ■ ■ r gence ------...... ------. ,1, original, 6 MT6 original, 0.5 MT 0.5 160 6 180 160 23 MT 23 160 160 6 180 160 6 10 200 180 160 ...... | 1 ------60g/l 8.5g/l 15g/l 30g/l 8 200 180 180 1 ■ 1 ------. 200 200 200 200 200 Fig. 23B and C and 23B Fig. . 266 267 ^Across-Shelf —— Flux — ■■ Convenience ■ 80 100 120 140 200 -Shelf Flux Convergence 200 -2 1 ..."'T-.... -1 ■ ■ 1 ° 1 ■ - 2 — L ------1------1— 1 1 -- 1------1------1-- ■ 0 20 40 60 80 100 120 140 i«n mn oi Across-Shelf WBL Flux Converqence ---- original -2 — oppwind 1 -— NE39 oE" — - SW219 — NW275 20 40 60 80 100 120 140 160 180 200 Depth Across-Shelf (m) B. 2 ftcross-^helf Flijx Convergence 1 E o 0 1 0 20 40 60 80 100 120 140 160 180 200 Alonq-Shelf Flux Convergence -2 1 "■»■ 1 ■" ■ **"■ 0 20 40 60 80 100 120 140160 180 200 Across-Shelf Suspended Flux Convergence 40 60 80 100 120 140 original Across-Sh^lf WBL Flux Convergence winds 1 tenth halfwinds windsHIGH 60 80 100 120 140 160 180 200 Depth Across-Shelf (m) Fig. 24 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A. 1 ' 1 ■'■■"p...... 1...■— 1 1 " ""1 .. 1.— ■ ■ ___ -■ - - - i 1...—... 20 40 60 80 100 120 140 160 180 200 ...... i ...... ~ i " i ...... r — ■■■■-...... i i " ~ : V ■ - ■ b ------1------1------1------1------1------1------■ i,IIM 200 Across-Shelf Suspended Flux Convergence i i "'~ 1 1 " i.....~ " I"11 r i 200 _— /TVs. ---- original E r S —— low o ---- half —— high ______1__------1------1------1------1------1------1------20 40 60 80 100 120 140 160 180 200 Depth Across-Shelf (m) Across-Shelf Flux Converqence B. -1 "I —1~ ~t ------r -0.5 o£ o 0.5 -I- 20 40 60 80 100 120 140 160 180 200 -1 I -0.5 " I o 0.5 1 ------1----- 1 ...... 1 In ---- 1------1---- « — -»---- 20 40 60 80 100 120 140 160 180 200 -1 -0.5 I o 0.5 1 ------1------1------1------1------1------1------1---- ■ I 20 40 60 80 100 120 140 160 180 200 Across-Shelf WBL Flux Convergence -1 -0.5 original 2m I o 3m 0.5 5m 1 0 20 40 60 80 100 120 140 160 180 200 Depth Across-Shelf (m) Fig. 25 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. poue wt pr sin fte oyih onr Frhrrpouto poiie ihu pr ission. perm without prohibited reproduction Further owner. copyright the of ission perm with eproduced R c. cm 0 o -0.5 g -1.5 -1.5 -0.5 0.5 0.5 0 0 20 20 06 80 60 40 060 40 160 140 120 100 80 60 40 Across-Shelf SuspendedFlux Converqence Across-Shelf WBLFlux Converqence Across-Shelf Ffux Converqence ■ i ■ i Alonq-Shelf Flux Converqence DepthAcross-Shelf (m) 80 100 100 ------120 120 r— T — 140 4 10 8 200 180 160 140 4 6 180 160 140 6 180 160 off original on . — Fig. 25C Fig. 200 200 200 269 270 ■ bed bed mass suspended mass wave boundary layer mass 07/25 07/25 07/27 07/29 07/31 i " " " ...... Date 2004 Date Inner Shelf Inner Shelf Shelf Break ' ■ ' ■ ' i Middle Middle Shelf 07/21 07/21 07/23 ■ x 10 x 10 x x 10 x 10 Outer Shelf 3 2 07/13 07/15 07/17 07/19 07/13 07/15 07/17 07/19 07/21 07/23 07/25 07/27 07/29 07/31 07/13 07/15 07/17 07/19 07/21 07/23 07/25 07/13 07/27 07/15 07/29 07/17 07/31 07/19 07/21 07/23 07/25 07/27 07/29 07/31 or o> 2 a 20 KJ ON crq Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 271 178°20'0"E 178°30'0"E 178o40'0"E 178°50'0"E 37 o40'0"SH -37°40,0"S 37°50'0"S- t37o50'0"S 38o0’0"S- 38°0'0"S 38°10'0"S 38°10,0"S 178°20'0"E 178°30'0"E 178°40'0"E 178°50,0"E After Lewis et al., 2004 Fig. 27 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 7 2 CONCLUSIONS The main findings of this study are: 1) Sediment deposition and accumulation rates on continental shelves can exceed the ability of the extant benthic community to effectively bioturbate the sediments. Sedimentary structure in such environments is characterized by laminations rather than bioturbated, mottled sediments. These sediments are therefore excellent recorders of sediment transport processes, nutrient and contaminant changes in terrestrial materials, and climatic changes. 2) Sediments can be transported and buried rapidly such that shelf sediments retain a terrestrial carbon signal. Not only are the preserved terrestrial sediments excellent climatic records, their presence indicates that multiple transport processes are important on the shelf. Furthermore, these processes result in distinctly different deposits. 3) Fine sediments deposited on the inner shelf tend to be remobilized and transported across the shelf to deeper waters. Fine sediments tend to be retained on the mid- and outer-shelf, within a subsiding shelf basin identified by Lewis et al. (2004). Some sediment escapes the shelf break area defined by the Ruatoria Re entrant. 4) The relative importance of sediment transport mechanisms including dilute suspension and gravity-flows varies across the shelf, resulting in distinct transport and depositional characteristics for each shelf region. Therefore, sediment deposits in each shelf zone created under similar wave and current conditions will have distinct depositional characteristics. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 7 3 5) On a shelf system characterized by strong along-shelf currents, sediments in dilute suspension tend to be removed from the area. Sediments transported in the wave boundary layer tend to be retained on the shelf. FUTURE WORK Several approaches may be employed in order to more accurately assess sediment transport and depositional patterns using ECOMSED. One of the main sensitivities of the model was settling velocity. Almost all of the sediments with low settling velocities were advected out of the study area in dilute suspension. Although sediments with higher settling velocities were also transported out of the proximal shelf area, flocculated sediments comprised almost all of the final sediment deposit. The ratio of flocculated to unflocculated sediments was chosen based on comparison with observed sediment concentrations at one location on the shelf and the observed pattern of deposition. An empirical assessment of sediment grain size and settling velocities would likely improve the predictive capabilities of the simulation. Although sediments were ephemerally deposited on the inner shelf in the simulation, most of the sediments were resuspended and removed from these shallow waters. Sedimentary data from the inner shelf indicates that between 8-15% of the fluvial budget is retained on the inner shelf, however (Wadman and McNinch, in review). In order to simulate longer periods of time, the model should account for sediment consolidation and burial by shelf sands. Strong observed along-shelf currents may be important on the shelf. The model sensitivity tests showed that increasing the strength of the wind-driven currents resulted in an increase in the current component of the bed shear stress. Although the simulation failed to account for the strong observed along-shelf currents, it did account Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 7 4 for 83% of the bed shear stresses, suggesting that waves influenced resuspension more than currents at the 60 m tripod. The strong currents and high sediment concentrations observed within the current-wave boundary layer indicated that gravity-driven transport may be supported by both waves and currents, however (Ma et al., in prep). To better approximate the bed shear stresses and gravity-driven transport processes, the ECOMSED model can be modified to account for gravity-driven transport in the wave-current boundary layer. The model grid would also have to be nested within a larger model to account for the strong shelf currents. A new method has recently been developed (Steve Kuehl, pers. comm.) to investigate accumulation and mixing over the last -50 years by analyzing bomb produced 239,240Pu (Kim et al., 2000, Kenna, 2002). This method could be used to confirm the excess 210Pb accumulation rates and to evaluate the pulsed event layers. Tephras, identified within several cores by high magnetic susceptibility values, can be dated and used in conjunction with seismic data to better constrain the shelf budget. The variability of the pulsed event layers across- and along-shelf can be assessed by analyzing the C/N ratio and 8 13C for more of the cores. This analysis may better constrain the spatial extent of the initial deposition of flood sediments. If the pulsed layers are increasingly marine in nature towards the shelf break, it could either be a function of decreased deposition or it could indicate resuspension and across-shelf transport of sediments. Finally, a time-series of cores should be collected after a flood to verify the deposition and reworking of freshly deposited sediments across the shelf by tracking 7Be activities. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. REFERENCES Cacchione, D.^., Drake, D.E., Kayen, R.W., Sternberg, R.W., Kineke, G.C., and Tate, G.B., 1995. Measurements in the bottom boundary layer on the Amazon subaqueous delta. Marine Geology, 125:235-257. Dellapenna, T.M., Kuehl, S.A., and Schaffner, L.C., 1998. Sea-bed mixing and particle residence times in biologically and physically dominated estuarine systems: a comparison of lower Chesapeake Bay and the York River Subestuary. Estuarine, Coastal, and Shelf Sci. 46:777-795. Dukat, D.A., and Kuehl, S.A., 1995. Non-steady-state 210Pb flux and the use of Ra/ Ra geochronometer on the Amazon continental shelf. Marine Geology, 125:329-350. Friedrichs, C.T., and Wright, L.D., 2004. Gravity-driven sediment transport on the continental shelf: implications for equilibrium profiles near river mouths. Coastal Engineering, 51:795-811. Friedrichs, C.T., and Scully, M.E., 2007. Modeling deposition by wave-supported gravity flows on the Po River prodelta: From seasonal floods to prograding clinoforms. Continental Shelf Research, 27:322-337. Geyer, W.R., Hill, P.S., and Kineke, G.C., 2004. The transport and dispersal of sediment by buoyant coastal plumes. Continental Shelf Research, 24:927-949. Geyer, W.R., Hill, P.S., Milligan, T.G., and Traykovski, P., 2000. The structure of the Eel River plume during floods. Continental Shelf Research, 20:2067-2093. Grant, W. D. and Madsen, O.S.,1986. The continental shelf bottom boundary layer. Annual Review of Fluid Mechanics, 18. 265-305. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 7 6 Guinasso, N.L., Jr. and D.R. Schink, 1975: Quantitative estimates of biological mixing rates in abyssal sediments. Journal of Geophysical Research. 80, 3032- 3043. Harris, C.K., P.A. Traykovski, and R.W. Geyer. 2005. Flood dispersal and deposition by near-bed gravitational flows and oceanographic transport: a numerical modeling study of the Eel River shelf, northern California. Journal of Geophysical Research, in press (accepted June, 2005). Harris, C.K., P.A. Traykovski, and R.W. Geyer. 2004. Including a Near-Bed Turbid Layer in a Three Dimensional Sediment Transport Model with Application to the Eel River Shelf, Northern California. Proceedings of the Eighth Conference on Estuarine and Coastal Modeling, ASCE, M. Spaulding (editor). 784-803. Hedges, J.I., Keil 19, R.G., 1995. Sedimentary organic matter preservation: an assessment and speculative synthesis. Marine Chemistry, 49: 81-115. Hines, A.H., Comtois, K.L., 1985. Vertical distribution of infauna in sediments of a subestuary of central Chesapeake bay. Estuaries 8, 296-304. Kenna, T.C., 2002. Determination of plutonium isotopes and meptunium-237 in environmental samples by inductively coupled mass spectrometry with total sample dissolution. Journal of Analytical Atomic Spectrometry, 17:1471-1479. Kim, C-S, Kim, C-K, Lee, J-I, Lee, K-J., 2000. Rapid determination of Pu isotopes and atom ratios in small amounts of environmental samples by an on-line sample pre-treatment system and isotope dilution high resolution inductively coupled plasma mass spectrometry. Journal of Analytical Atomic Spectrometry, 15:247- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 7 7 2 5 5 . Kineke, G.C., R.W. Sternberg, J.H. Trowbridge and W.R. Geyer, 1996. Fluid Mud Processes on the Amazon Continental Shelf. Continental Shelf Research, 16: 667-696. Kineke, G.C., and Sternberg, R.W., 1995. Distribution of fluid muds on the Amazon continental shelf. Marine Geology, 125:193-233. Kineke, G.C., Woolfe, K.J., Kuehl, S.A.; Milliman, J.D., Dellapenna, T.M., and Purdon, R.G., 2000.Sediment export from the Sepik River, Papua New Guinea; evidence for a divergent sediment plume. Continental Shelf Research, 20:2239- 2266. Kirchner, G., Ehlers, H., 1998. Sediment geochronology in changing coastal environments: Potentials and limitations of the 137Cs and 210Pb methods. J. Coast. Res. 14:483-492. Ma, Y., Wright, L.D., and Friedrichs, C.T., in prep. Observations of sediment transport on the continental shelf off the mouth of the Waiapu River, New Zealand: evidence for current-supported gravity flows. To be submitted to Continental Shelf Research Mulder, T. and Syvitski, J. P. M., 1995. Turbidity currents generated at river mouths during exceptional discharge to the world's oceans. Journal of Geology 103: 285-299. Nittrouer, C.A., and Sternberg, R.W., 1981. The formation of sedimentary strata in an allochthonous shelf environment: The Washington continental shelf. Marine Geology, 42:201-232. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Ogston, A.S. and Sternberg,R.W., 1999. Sediment-transport events on the northern California continental shelf. Marine Geology, 154:69-82. Ogston, A.S., Cacchione, D.A., Sternberg, R.W., and Kineke, G.C., 2000. Observations of storm and river flood-driven sediment transport on the northern California continental shelf. Continental Shelf Research, 20:2141-2162. Officer, C.B, and Lynch, D.R., 1989. Bioturbation, sedimentation, and sediment-water exchanges. Estuarine, Coastal, and Shelf Sci. 28:1-12. Palanques, A., Puig, A., Guillen, J., Jimenez, J., Gracia, V., Sanchez-Arcilla, A., and Madsen, O., 2002. Near-bottom suspended sediment fluxes on the microtidal low-energy Ebro continental shelf (NW Mediterranean). Continental Shelf Research, 22:285-303 Sanford, L.P., 1992. New sedimentation, resuspension, and burial. Limnol. Oceanography. 37:1164-1178. Schaffner L. C., Dellapenna, T. M., Hinchey, E. K., Friedrichs, C. T., Neubauer, M. T., Smith, M. E., and Kuehl, S. A., 2001. Physical energy regimes, seabed dynamics and organism-sediment interactions along an estuarine gradient. In: Aller, Woodin, and Aller, eds. Organism-Sediment Interactions Columbia, S.C., Baruch Press: 2000. Schaffner, L.C., Diaz, R.J., Olsen, C.R., and Larsen, I.L., 1987. Faunal characteristics and sediment accumulation processes in the James River Estuary, Virginia. Estuarine, Coastal, and Shelf Sci. 25: 211-226. Scully, M.E., C.T. Friedrichs and L.D. Wright, 2002. Numerical modeling results of gravity-driven sediment transport and deposition on an energetic continental Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 7 9 shelf: Eel River, Northern California. Journal of Geophysical Research. Scully, M.E., Friedrichs, C.T., Wright, L.D., 2003. Numerical modeling of gravity- driven sediment transport and deposition on an energetic continental shelf: Eel River, Northern California. Journal of Geophysical Research 108 (C4),(17-l- 17-14). Sternberg, R.W., and Larsen, L.H., 1976. Frequency of sediment movement on the Washington continental shelf: a note. Marine Geology, 21:M37-M47. Traykovski, P., W.R. Geyer, J.D. Irish and J.F. Lynch, 2000. The role of wave-induced density-driven fluid mud flows for cross-shelf transport on the Eel River continental shelf. Continental Shelf Research, 20: 2113-2140. Wheatcroft, R.A., Olmez, I., Pink, F.X., 1994. Particle bioturbation in Massachusetts Bay: Preliminary results using a new deliberate tracer technique. J. Mar. Res. 52, 1129-1150. Wheatcroft, R.A., and Drake, D.E., 2003. Post-depositional alteration and preservation of sedimentary event layers on continental margins,I. The role of episodic sedimentation Marine Geology 199: 123-137. Wright, L D; Friedrichs, C T; Kim, S C; Scully, M E, 2001. Effects of ambient currents and waves on gravity-driven sediment transport on continental shelves. Marine Geology, vol. 175, no. 1-4, pp.25-45. Wright, L.D., C.T. Friedrichs and M.E. Scully, 2002. Pulsational gravity-driven sediment transport on two energetic shelves. Continental Shelf Research, 22:2443-2460 Wright, L.D., Ma ,Y., Scully, M, Friedrichs, C.T., 2006.Observations of Across Shelf Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 280 Sediment Transport During High Energy Flood Events off the Mouth of the Waiapu River, New Zealand Eos Trans. AGU, 87(36), Ocean Sci. Meet. Suppl., Abstract OS11L-03. Wright, L.D. and Friedrichs, C.T., 2006. Gravity-driven sediment transport on continental shelves: A status report. Continental Shelf Research, 26:2092-2107. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 281 APPENDIX A The seabed observational data discussed in Chapter 1 is contained on the supplemental CD in the folder labeled Appendix A. The data are organized into files with the following names: Al.Core Locations ...... 284 A2. Excess 210Pb Activities...... 287 A3. 7Be Activities ...... 298 A4. Bulk Carbon and 5 13C ...... 300 A5. Grain Size ...... 301 A6. X-radiographs ...... 302 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 8 2 APPENDIX B The river, wind, and wave input data discussed in Chapters 2 and 3 is contained on the supplemental CD in the folder labeled Appendix B. The data are organized into files with the following names: Bl. Tripod and Wind Locations ...... 326 B2. Waiapu River Discharge Data...... 327 B3. Hicks Waiapu River Rating Curve ...... 507 B4. Hicks Bay Wind Data ...... 513 B5. Gisborne Wind Data ...... 554 B6. 40 m Tripod Wave Data ...... 573 B7. 60 m Tripod Wave Data ...... 589 B8. Simulation Animation The simulation animation file can be played with Windows Media Player. The left panel in the simulation shows sediment deposition and winds (red arrow); the middle panel shows sediments transported in the wave boundary layer; and the right panel shows sediments transported in the overlying water column. The red arrows in the middle and right panels denote current speed and direction in the wave boundary layer and the depth-averaged water column, respectively. The freshwater plume is outlines by the blue line in the right panel. The two bottom panels show the Waiapu River discharge record and the wave height record with time. During the mid-June flood, waves are low, resulting in initial deposition on the inner shelf. The early-July flood is accompanied by high wave heights, resulting in initial deposition on the mid shelf. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPENDIX C The sediment mass distribution discussed in Chapter 3 is contained on the supplemental CD in the folder labeled Appendix C. The data are organized into files with the following names: Cl. Time-averaged Sediment Mass Distribution ...... 606 C2. Sensitivity Animations The sensitivity animation files can be played with Windows Media Player. The left panel in the simulation shows sediment deposition and winds (red arrow); the middle panel shows sediments transported in the wave boundary layer; and the right panel shows sediments transported in the overlying water column. The red arrows in the middle and right panels denote current speed and direction in the wave boundary layer and the depth-averaged water column, respectively. The freshwater plume is outlines by the blue line in the right panel. The two bottom panels show the Waiapu River discharge record and the wave height record with time. During the mid-June flood, waves are low, resulting in initial deposition on the inner shelf. The early-July flood is accompanied by high wave heights, resulting in initial deposition on the mid shelf. Animations for the floe fraction, wave, wind, and sediment load experiments are identified by the folder names. Each case is identified in the file name. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.