Louisiana State University LSU Digital Commons
LSU Historical Dissertations and Theses Graduate School
1977 Morphology, Grain Size Characteristics, and Fluvial Processes of Two Bars, Colville River Delta, Alaska. Donald Franklin Nemeth Louisiana State University and Agricultural & Mechanical College
Follow this and additional works at: https://digitalcommons.lsu.edu/gradschool_disstheses
Recommended Citation Nemeth, Donald Franklin, "Morphology, Grain Size Characteristics, and Fluvial Processes of Two Bars, Colville River Delta, Alaska." (1977). LSU Historical Dissertations and Theses. 3075. https://digitalcommons.lsu.edu/gradschool_disstheses/3075
This Dissertation is brought to you for free and open access by the Graduate School at LSU Digital Commons. It has been accepted for inclusion in LSU Historical Dissertations and Theses by an authorized administrator of LSU Digital Commons. For more information, please contact [email protected]. INFORMATION TO USERS
This material was produced from a microfilm copy of the original document. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the original submitted.
The following explanation of techniques is provided to help you understand markings or patterns which may appear on this reproduction.
1. The sign or "target" for pages apparently lacking from the document photographed is "Missing Page(s)". If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting thru an image and duplicating adjacent pages to insure you complete continuity.
2. When an image on the film is obliterated with a large round black mark, it is an indication that the photographer suspected that the copy may have moved during exposure and thus cause a blurred image. You will find a good image of the page in the adjacent frame.
3. When a map, drawing or chart, etc., was part of the material being photographed the photographer followed a definite method in "sectioning" the material. It is customary to begin photoing at the upper left hand corner of a large sheet and to continue photoing from left to right in equal sections with a small overlap. If necessary, sectioning is continued again — beginning below the first row and continuing on until complete.
4. The majority of users indicate that the textual content is of greatest value, however, a somewhat higher quality reproduction could be made from "photographs" if essential to the understanding of the dissertation. Silver prints of "photographs" may be ordered at additional charge by writing the Order Department, giving the catalog number, title, author and specific pages you wish reproduced.
5. PLEASE NOTE: Some pages may have indistinct print. Filmed as received.
University Microfilms International 300 North Zeeb Road Ann Arbor, Michigan 48106 USA St. John's Road, Tyler's G reen High Wycombe, Bucks, England HP10 8HR 77-25,392 NEMETH, Donald Franklin, 1938- MORPHOLOGY, GRAIN SIZE CHARACTERISTICS, AND FLUVIAL PROCESSES OF TWO BARS, COLVILLE RIVER DELTA, ALASKA. The Louisiana State University and Agricultural and Mechanical College, Ph.D., 1977 Geology
Xerox University Microfilms,Ann Arbor, Michigan 48106
@ 1977
DONALD FRANKLIN NEMETH
ALL RIGHTS RESERVED MORPHOLOGY, GRAIN SIZE CHARACTERISTICS, AND FLUVIAL PROCESSES OF TWO BARS, COLVILLE RIVER DELTA, ALASKA
A D issertation
Submitted to the Graduate Faculty of the Louisiana State University and A gricultural and Mechanical College in partial fulfillm ent of the requirements for the degree of Doctor of Philosophy
i n
The Department of Geology
by Donald F. Nemeth B.S., University of Southern California, 1962 M.S., University of Southern California, 1969 May, 1977 ACKNOWLEDGMENTS
In order to conduct this study of a remote and somewhat inhospi
table area it was necessary to enlist the aid of many individuals and
organizations.
The author wishes to thank Dr. Harley J. Walker for proposing the expedition; offering guidance, assistance, and suggestions throughout the project; working out the logistics of the field study; sharing his vast knowledge of the arctic; and acting as co-major professor. Without his assistance this project would not have been possible.
Dr. Donald R. Lowe, co-major professor, provided assistance, many valuable suggestions, and took time from his busy schedule to take on the burden of co-major professor at a late date. The author would also like to thank his minor professor, Dr. Robert A. Muller, and the other members of the doctoral committee, Dr. Donald H. Kupfer and Dr.
Judith A. Schiebout, for their assistance and suggestions. Valuable suggestions were also given by Dr. Clarence 0. Durham, Jr. and Dr.
James P. Morgan, who served as co-major professor until his retirem ent.
The author gratefully acknowledges Dr. Prentiss E. Schilling who designed the statistical analyses and Dr. Penelope Hale who con ducted the analyses and offered valuable consultation along the way.
M essrs. Jeffrey Peake, Lawrence McKenzie, and Donald White, and
Dr. William Richie aided with the field work.
Laboratory facilities and supplies were provided by the Depart ment of Geology, Louisiana State University. Travel expenses were financed through a Penrose Grant from the Geological Society of America.
ii Logistic support in Alaska was supplied by the Naval Arctic Research
Laboratory, Barrow, Alaska. The project was supported through the
Office of Naval Research, under contract N00014-69-A-0211-0003,
Project NR 388 02. The manuscript was diligently typed by Mrs. Mary
M e v e rs.
This study would not have become a reality without the support of the author’s family and especially his wife, Darlyne, who persevered throughout the project and assisted in ways too numerous to mention.
iii TABLE OF CONTENTS
Page
TITLE PAGE...... i A '
ACKNOWLEDGMENTS...... i i
LIST OF FIGURES...... v i i
LIST OF SYMBOLS ...... x
LIST OF TABLES...... x i i
ABSTRACT...... x i i i
INTRODUCTION...... 1
THE STUDY AREA...... 2
Nomenclature ...... 2
C lim a te ...... 5
H y d ro lo g y ...... 6
Environmental Subareas ...... 8
PROCEDURES. . . ‘...... 15
Samples Collected ...... 15
Method Employed ...... 16
MORPHOLOGIC CHANGES ...... 17
Before Spring Flooding ...... 17
Thickness of Snow Cover ...... 17
Ablation of Snow Cover ...... 23
Thickness of River Ice ...... 24
Spring Flooding ...... 25
Summer Flow Regim e ...... 39 Page
SEDIMENTS...... 43
Grain Size S tatistical Parameters...... 47
Change with Elevation and Location ...... 47
Group I —Submerged River Sediments and a ll Nongravel Bar Sediments ...... 48
Group II—Gravel Deposits of the Longitudinal Bar. . . 52
Environmental Subareas and Grain Size S tatistical P a r a m e t e r s ...... 55
Scatter Plots ...... 57
D is c u s s io n ...... 61
Group 1 ...... 61
Group I I ...... 63
Effects of Ice Rafted Sediments ...... 64
Sediment Thaw ...... 65
PROCESSES INTERPRETED FROM GRAIN SIZE DATA...... 69
Fluvial Processes Interpreted from C-M Diagram ...... 69
Truncation Points and Hydrodynamic Principles ...... 71
Aeolian Transport ...... 73
Fluvial Transport of Sands, S ilts, and Clays ...... 74
Fluvial Transport of Gravels ...... 80
ORIGIN AND DEVELOPMENT OF THE BARS...... 85
SUMMARY AND CONCLUSIONS...... 88
REFERENCES...... 95
APPENDICES...... 99
Appendix I - Derivation of Fluvial Bar Terminology ...... 100
Appendix II - Field Procedures ...... 102
v Page
Appendix III - Laboratory Procedures ...... 105
Appendix IV - S tatistical Methods ...... 107
Appendix V - Orthogonal Comparisons ...... 109
Appendix VI - Methods used for Determination of Fluvial Processes...... 115
Appendix V II - Longitudinal Profiles with Sediment Thaw. . . 125
VITA...... 129
vi LIST OF FIGURES
Figure Page
1. Location Map of Northern Alaska and the Colville River D elta ...... 3
2. Map of the Colville River Delta with the Location of the Study Area ...... 3
3. Map of the Study Area and its Surroundings ...... 4
4. River Stages of the Colville River, Spring and Summer, 1962, 1964, and 1971 ...... 8
5. Vertical Aerial View of the Two B ars ...... 9
6. Topographic Map of the Study Area with Environmental Subareas, Sample Locations, and Sites of Grounded Ice Blocks ...... 10
7. View of Snow Cover in the Willows of the Longitudinal Bar, Early May, 1971 ...... 18
8. Deflation of Snowdrift ...... 18
9. Deflation of Snow from Northeast Facing Slope, Longitudinal Bar, Mid-May, 1971 ...... 19
10. Snow Cover in Mid-May, 1971, Profile A-A1 ...... 19
11. Snow Cover in Mid-May, 1971, Profile B-B1 ...... 20 I 12. Snow Cover in Mid-May, 1971, Profile C-C' ...... 20
13. Snow Cover in Mid-May, 1971, Profile D-D1 ...... 21
14. Snow Cover in Mid-May, 1971, Profile E-E' ...... 21
15. Small Delta-Shaped Lobes formed by Meltwater in Mid-May on the Longitudinal Bar ...... 25
16. Prebreakup Aerial View of the Almost Totally Submerged Longitudinal B a r ...... 27
17. Prebreakup Aerial View of Spring Flooding of the Study Area ...... 27
18. Prebreakup Aerial View, Point of Junction of Floating Ice Bands ...... 28
v i i Figure Page
19. Prebreakup Aerial View of Bank of Floating Ice in Side Channel ...... 28
20. Snow Cover on the Longitudinal Bar, June 2 and 7, 1971, Profile F -F ' ...... 31
21. Postbreakup Aerial View, June 6th, of the Longitudinal B ar ...... 33
22. Postbreakup Aerial View, June 11th, of the Longitudinal Bar Looking Downstream ...... 33
23. Postbreakup Aerial View, June 11th, of the Longitudinal Bar Looking Upstream ...... 34
24. The Downstream Tip of the Longitudinal Bar, Ju n e 1 1 th ...... 34
25. Trail Formed by Ice Block Temporarily Grounded in the G ravels ...... 36
2 6 . Trial Left in Ripple Marked Sand and Silt...... 36
2 7 . Grounded Ice Block on Longitudinal Bar with Ice Shoved Sediment...... 37
2 8 . Ice Shoved Sediment after Melting of Grounded I c e B lo c k ...... 37
29. Wave Built Gravel Ridges on the Gravel Sheet, S u b area 3 ...... 40
30. View in July of Wave Cut Terraces on the Side Bar ...... 40
31. Sand Shadow Dunes on the Longitudinal Bar...... 41
32. Linear Bands of Sand on the Northeast Facing Slope of the Longitudinal Bar ...... 41
33. Postbreakup State, Screen on the Gravel Sheet, Subarea 3 ...... 44
3 4 . Postbreakup State, Screen on the Sands, Silts, and Clays, of Subarea 5 ...... 45
35. Close-Up of Sediment Deposited on Screen of Figure 34. . . 45
36. Columnar Sections of the Longitudinal Bar, P r o f i l e M-M1 ...... 46
viii Figure Page
37. Group I — Predicted Curve for Mean Diameter vs. E levation ...... 49
38. Group I -- Predicted Curve for Standard D e v ia tio n v s . E l e v a t i o n ...... 49
39. Group I — Predicted Curve for Skewness vs. E le v a tio n ...... 50
40. Group I — Predicted Curve for Kurtosis vs. E l e v a ti o n ...... 50
41. Predicted Curves for Groups I and II; Mean Diameter vs. Longitudinal P osition ...... 53
42. Group I — Predicted Curve for Kurtosis vs. Longitudinal Position ...... 53
43. Group II — Predicted Curve for Standard Deviation vs. Longitudinal Position ...... 54
44. Group II — Predicted Curve for Skewness vs. Longitudinal Position ...... 54
45. Scatter Plot -- Mean Diameter vs. Standard D e v ia tio n ...... 58
46. Scatter Plot — Mean Diameter vs. Skewness ...... 58
47. Scatter Plot — Mean Diameter vs. K urtosis ...... 58
48. Scatter Plot — Skewness vs. Kurtosis ...... 59
49. Cumulative Probability Curves of the Ice Rafted and Nonrafted Sediments ...... 66
50. C-M Diagram of Sediment Samples According to Environmental Subareas ...... 70
51. Truncation Points in Subareas 3 and 4 ...... 72
52. Log-Normal Subpopulations of Aeolian Sediment ...... 75
53. Log-Normal Subpopulations of Suspended Sediment and Surface Sediment Deposited in 1971 ...... 76
54. Suspended Sediment Deposited Upon Willow Branches ...... 78
ix LIST OF SYMBOLS
Symbol
A = channel cross-sectional area
= a function of the channel shape
/3 = dimensionless critical shear stress
c = coarsest one percentile
Y s specific weight of fluid % s specific weight of solid
d = particle diameter
= d65 sieve size for which 65% of the mixture is finer D = mean depth of flow
g = gravitational constant
i = g r a d ie n t
= Kg graphic kurtosis
k = channel roughness factor
R0 = von Raman's coefficient for turbulent exchange
Mz = graphic mean particle size
= dynamic velocity of the fluid
'Y* S3 kinetic viscosity
= fluid density
/ ° s = density of solid
R = hydraulic radius
Re = critical boundary Reynolds number r 2 = amount of variability accounted for
x i Symbol
inclusive skewness SkI "
2 l - inclusive standard deviation
* c - critical shear stress
d o shear stress acting on the bottom
P particle diameter given by -log 2 ( d ia . i n
X mean
u* = shear velocity
u average current velocity
^ t o t a l total shear velocity acting on the bottom w = channel width
UJ = particle settling velocity LIST OF TABLES
Table Page
1. Snow Cover of the Longitudinal B ar ...... 22
2. Equations for Predicted Curves ...... 51
3. Comparisons between All Subareas ...... 56
4. Comparisons of Average Mean Diameter of A ll Subareas excluding Gravels ...... 56
5. Group I — Regression Curves vs. Scatter Plots ...... 60
6. Group II — Regression Curves vs. Scatter Plots ...... 60
7. Active Layer Correlations ...... 67
8. Truncation Points per Subarea ...... 73
9. Average Truncation Points for All Samples ...... 74
10. Calculated Shear Velocities at Section 1, 1962 ...... 76
11. Shear Stress and Shear Velocity for Gravel Movement. . . . 81
12. Average Current Velocities for Gravel Movement ...... 82
13. Grain Size when U*=»u j ...... 83
xii ABSTRACT
Hydrological and morphological changes occurring from late winter through summer, 1971, were studied on a longitudinal bar and a side bar at the apex of the Colville River delta in arctic Alaska.
During the brief period of spring flooding, the major changes on these bars were noted as follows: (1) complete removal of snow cover, (2) rise in river stage to a height of 17.8 ft (5.4 m) above the level of the river ice in approximately 4 days, (3) total submergence of the side bar, (4) near total submergence of the longitudinal bar, (5) grounding of 57 blocks of river ice on the longitudinal bar during breakup, (6) initiation of thaw of bar sediments, and (7) return of the river stage to near sea level at the end of flooding.
From grain size analyses of 138 surficial sediment samples, a determination was made of the character of fluvial action- during spring flooding. For purpose of analysis, the samples are classified into 9 environmental subareas, which encompass two major groups: Group I and
Group II. River bottom sediments and a ll nongravel bar sediments com prise Group I. These sediments were actively transported during spring flooding. Bar gravels comprise Group II. These sediments were not actively transported.
Regression analyses show that significant differences in mean diameter, standard deviation, skewness, and kurtosis occur for the deposits of Group I with elevation. Significant differences for mean diameter and kurtosis also occur with distance downstream. Group II deposits reveal no significant changes with elevation; rather, with xiii distance downstream significant changes do occur for mean diameter,
standard deviation, and skewness.
Mean diameter of river and all nongravel bar deposits (Group I)
decreases linearly with elevation above the bed. The decrease in grain
size is probably a product of decreased competency along the stream bed
with elevation and a decrease in suspended sediment size with elevation.
Phi boundaries between modes of fluvial sediment transport were
established by determining the truncation points between log-normal
subpopulations of grain size distributions, comparing the results with
samples of known depositional origin,and interpreting the results in
light of hydrodynamic theory. For the spring floods of 1962 and 1971
the competency was 0.9 0. The boundary between particles transported
in contact with the bed and an intermittent suspension was 1.2 0. The
boundary between interm ittent suspension and continuous suspension
occurs at 3.0 0 in areas exposed to the prevailing current during
flooding and at 4.0 0 in areas somewhat protected from the prevailing
currents. These truncation points occurred in sediments deposited
during a single spring flood as well as in sediments deposited over
several years, indicating that hydrodynamic conditions do not vary appre
ciably from one spring flood to the next.
Based upon analyses of grain size statistical parameters and hydrodynamic theory, it is concluded that, for the study area, the dis
tribution of sediments according to grain size is determined predomi nantly by the conditions existing during spring flooding.
Finally, it is proposed that the gravels of the longitudinal bar were transported to their present location from an upstream source at a
xiv time of high stream velocity, possibly resulting from the breakup of an ice jam. The deposition of the gravels formed the nucleus for bar development. It is concluded that the presence of these gravels is primarily responsible for maintaining the stability of the bar.
xv INTRODUCTION
An opportunity to study the fluvial and deltaic processes active within a braided river in an area of continuous permafrost is afforded by two river bars which lie at the apex of the Colville River delta in arctic Alaska. Of these two bars, the one positioned in the river channel is unique to the delta in that its surface is partly composed of the largest outcropping of gravel to be found on a bar in the delta.
Comparisons of aerial photographs from 1949 and 1971 indicate that this bar has remained in essentially the same position over a 22 year period.
Therefore, this bar is unique from downstream bars which exhibit a definite migration downstream.
The present study includes (1) documentation of morphologic changes occurring from late winter to mid-summer, 1971, (2) descrip tion of sediment samples via four grain size statistical parameters: mean diameter, standard deviation, skewness, and kurtosis, and (3) determination of active processes from (a) grain size characteristics of deposits and (b) analysis of modes of sediment transport and forces necessary for movement of particles. These methods are used to deter mine the effectiveness of processes contributing to the development of th e b a r s .
1 THE STUDY AREA
The Colville River drains northward across the Arctic Coastal
Plain of northern Alaska into the Arctic Ocean where it has formed a
232 mi^ (600 km^) delta (Walker, 1976) (Fig. 1). The distributary system of this delta begins approximately 25 mi (40 km) south of the
Arctic coast. From the head of the Colville River to its terminus at the Arctic Ocean, the river flows through three physiographic provinces the Arctic Mountains, the Arctic Foothills, and the Arctic Coastal
Plain (Wahrhaftig, 1965). The proportion of the drainage basin found in each of the aforementioned physiographic provinces is respectively
26, 64, and 10 percent (Walker and McCloy, 1969).
The study area, composed of two bars and the adjoining river channels, is located at lat. 70° 11' N and long. 150° 55' W (Fig. 1).
It is a braided reach of the river at the apex of the delta (Fig. 2).
The study area, with a length of 1.6 mi (2.6 km) and a width of 0.5 mi (0.8 km), is located slightly north of the first delta distributary, the West Channel (Fig. 3). It is positioned in the East Channel approximately 1.5 mi (2.2 km) north of the last major tributary to the
Colville River, the Itk illik River.
Nomenclature
Of the two bars studied, one is positioned within the stream channel. The second, located downstream from the first, is attached to the river's right bank. Hereafter, the bar located within the stream channel w ill be referred to as a longitudinal bar and the bar attached to the bank w ill be referred to as a side bar. Ore (1964, p. 1) 3
'AREA
.Faitliankt COLVILLE I DELTA M ILES. 100 > 3 5 T S T WLKETT .
[ARCTIC \ COtfSML PLA^N
CAPE LISBURNE
Figure 1. Location map of northern Alaska and the Colville River delta.
Putu Channel
Colville Delta
Delta__ Apex K ilo io rte i Section 1
Figure 2. Map of Colville River delta with location of study area. 4
7777TT
STUDY AREA
soo i o o o M e n u s
ca
Figure 3. Map of the study area and its surroundings. defines a longitudinal bar as one which, "has .its long axis essentially aligned with stream flow." The side bar of this study has the following characteristics in common with the side bar defined by Collinson
(1970): (1) it is attached to a bank; (2) it is pointed in an upstream direction with the point corresponding to a headland; and (3) its downstream margin passes gradually into deeper water. It differs from
Collinson's side bar in that it gradually widens downstream until it reaches its maximum width and then gradually narrows until it rejoins the bank. Collinson's side bar gradually widens downstream and then abruptly rejoins the bank.
Because of the unequal division of the Colville River channel by the longitudinal bar, the larger western channel w ill be referred to 5
as the main channel and the smaller eastern channel w ill be referred to
as the side channel.
The terms applied to each of the major physiographic features
are meant to be purely descriptive and have no genetic im plications.
The sim ilarities between the bars of this study and those
described by other authors is presented in Appendix I.
C lim a te
The climate in the area of the Colville River delta is charac
terized by long, cold predominantly continental winters and short, cool
predominantly maritime summers (Walker and McCloy, 1969, p. 18).
Winter lasts for approximately 8 months and is a time of subfreezing
average monthly temperature and little precipitation. Much of this
precipitation falls as snow and is responsible ..for a thin snow cover
over almost all surfaces. Summer, which lasts for approximately 4
months, is a time of above-freezing average monthly temperature and
little precipitation which occurs as either snowfall or rainfall.
Although they are of short duration, the transition periods between
these two seasons are referred to as spring and fall.
Mean annual precipitation reported for Alaska (Searby and
Branton, 1975, p. 283) indicates that 5 in (12.7 cm) or less occur on
the Colville River delta. However, in parts of the drainage basin of
the Colville River, mean annual precipitation exceeds 10 in (25.4 cm)
and, in a few isolated areas in the Arctic Mountains, even exceeds 20
in (50.8 cm).
For the Colville delta, mean monthly temperature records for an
entire year are not available. The yearly fluctuations, however, may 6
be approximated from the records of Barrow, a maritime location, and
Umiat, a continental location. Mean monthly temperatures vary from a
February minimum of -2 7 .9 ° and -3 1 .5 °C at Barrow and Umiat, respec
tively, to a July maximum of 3 .9 ° and 11.8°C (McKenzie and Walker,
1974). During the 1971 field season at Putu (for location see Fig. 2),
the temperature fluctuated from a minimum of -27.5°C on April 21st to a maximum of 22.8°C on June 7th (McKenzie and Walker, 1974).
On the Colville River delta, two predominant wind directions, northeast and southwest, are apparent from the orientation of sand dunes and snowdrifts (Walker and McCloy, 1969, p. 26). This is consistent with information gathered at Pingok Island on the Arctic Ocean where
both winter and summer are characterized by predominantly northeasterly winds. During summer, however, cyclonic low pressure cells produce
southwesterly winds which occur approximately 30 to 40 percent of the
time (Short, 1973).
H ydrology
The discharge of the Colville River through the study area is
derived from a drainage basin of approximately 23,200 mi^ (60,000 km^).
Due to clim atic and environmental conditions, the discharge has an annual cycle which fluctuates from a spring maximum during flooding to a late winter minimum when no discharge has been detected (Walker,
1973a). The transition from active discharge to no discharge occurs in
the fall when air temperatures decrease to below 0°C and both surface water and groundwater freeze (Walker, 1973b).
At this time the stage of the Colville River is near sea level, 7
discharge becoriies insignificant, and the level of the distributaries
is controlled by the sea. It is at this level that freezeup occurs.
During winter, the river ice reaches a maximum thickness of about 6 ft
(1.8 m).
When water trapped below the river ice is connected to the sea,
the fresh water of the river is replaced by sea water. A salt water wedge has been found to penetrate as far upstream as 36 mi (58 km) from
the mouth of the river (Walker, 1973b).
The river stages as recorded and compiled by H. J. Walker at
Putu in 1961, 1964, and 1971 are shown in Figure 4. Soon after the
first development of meltwater in the spring, renewed discharge
commences. Shortly thereafter, a stage increase of several feet occurs
over a period of only a few days. This is exhibited best by the records
for 1964 and 1971. The rapid increase in stage can be attributed to melting occurring south of the delta where the snow cover is greater and temperatures increase to above freezing earlier than on the delta.
During the spring floods of 1962 and 1971, maximum discharge was respectively 212,370 ft^/s (6010 m-fys) on June 14, 1962 (Arnborg, et a l., 1966, p. 197), and 331,450 ft^/s (9380 m^/s) on June 3, 1971
(Walker, personal communication).
Spring flooding in both 1964 and 1971 lasted for approximately
16 days. In 1962, it lasted for 25 days (Amborn, et al., 1966, p.
208). Therefore if an entire year is considered, the time of spring
flooding embraces only 4 percent of 1964 and 1971 and 7 percent of 1962.
If only the time of active discharge is considered (i.e., approximately
4 months), spring flooding embraces 13 percent for 1964 and 1971 and
21 percent for 1962. 8
1971 j
\ 1964
3
1962
*
May June
Figure 4. Colville River stages recorded at Putu Channel, spring and summer 1962, 1964, and 1971. Values are given in meters.
At the time of maximum stage in the study area, the effects are as follows: the area from the sand dunes on the west to the area of tundra surface on the east is almost totally submerged (Fig. 3); the flood waters inundate the lower parts of the tundra surface; and the
West Channel and the East Channel are both active delta distributaries.
During most periods of low water, the West Channel becomes inactive and in some years, such as 1971, is sealed off from the East Channel by a bar. At this time, discharge is confined entirely to the East Channel.
With the cessation of spring flooding on June 16, 1971, the stage fell to near sea level. It fluctuated around this level through out the summer. Increases in stage of a few feet did, however, occur during this period, probably corresponding to periods of precipitation within the drainage basin.
Environmental Subareas
The study area can be divided into a number of subareas on the basis of physiography, surficial sediments, and/or vegetation (Figs. 5 9
* 4 1 1 tf 7 - d
Li its.
Figure 5. Vertical aerial view of longitudinal and side bars, June 16, 1971 •mT
• M l CD CD *
EXPLANATION SAMPLE SITE SITE OF OROUNDED ICE BLOCK WILLOWS
IOOM0 500 FEET 20J> 190 METERS
CONTOUR INTERVAL IN FEET DATUM IS MEAN SEA LEVEL ■•v
l« T 1 ‘ .IN O U tX O
■ -12
• MO Ml
’S'
1 JWILLOWS i*\T v * :,' w
Figure 6. Topographic map of i environmental subart location of groundei spring flood of 197 Figure 6. Topographic map of the study area with environmental subareas, sample sites, location of grounded ice blocks from spring flood of 1971, and profile locations. 11
& 6).
River Channels
Subarea 1. Main channel
The main channel is approximately 2000 ft (0.6 km) at
its widest point and extends to depths in excess of
-16 ft (-4.9 m) at normal stage. Water, which is
always present, is derived from the fresh water dis
charge of the Colville River, from salt water
intrusion by the Arctic Ocean, or from a combination
o f b o th .
Subarea 2. Side channel
In plan view, the arcuate-shaped side channel maintains
a width of approximately 350 ft (0.1 km). It includes
local areas with depths in excess of -12 ft (-3.6 m).
From its upstream end, the thalweg meanders back and
forth across the channel width. Like the main channel,
the side channel is always submerged.
Longitudinal Bar
Subarea 3. Gravel Sheet
The gravel sheet is located at the upstream end of the
longitudinal bar and extends to a maximum elevation of
9 ft (2.7 m). For convenience, the zero contour (sea
level) is taken as the lower boundary of the subarea.
It is apparent, however, that the gravel sheet extends
below zero contour into the adjacent river channels.
Topographically the gravel sheet consists of: (1) wave-built gravel ridges, (2) a complex pattern
of what appears to be large-scale ripple-forms or
megaripples, and (3) trails formed by the temporary
grounding of ice blocks during the spring flood.
Topographic depressions are present downstream of
several wave-built ridges, between the megaripples,
and between some parallel ridges of ice trails.
Within these depressions, settling basins form during
spring flooding and local patches of sand, silt, and
clay are deposited on the surface of the gravels.
The gravels are composed predominantly of subangular
to subrounded pebbles of chert and minor amounts of
quartzite and silicified conglomerate. The downstream
terminus of the gravel sheet is marked by an abrupt
change from gravel to sand.
Subarea 4. Llnguoid Gravel Ridge
On the surface of the longitudinal bar, the linguoid
shaped gravel ridge is separated from the gravel sheet
of subarea 3 by approximately 1000 ft (0.3 km).
Topographically, the linguoid gravel ridge has a
bar chan shape with the horns pointing upstream. The
highest elevation, 8 ft (2.4 m), occurs on the down
stream end. For convenience, the lower boundary is
taken as sea level. The gravel of the ridge does
extend into the adjacent portion of the main channel. 13
Subarea 5. Unvegetated sand, silt, and clay
In general, the unvegetated sand, silt, and clay refer
red to as nongravels on Figure 6 extends from sea level
to approximately 8 ft (2.4 m). However, on the unpro
tected stoss side of the longitudinal bar, subarea 5
reaches an elevation of 12 ft (3.6 ra) and consists
largely of sand and silt with a minor amount of clay.
Except for the discontinuity caused by the presence
of the linguoid gravel ridge, this subarea almost
completely encircles the longitudinal bar.
Subarea 6. Grasses
An area of grasses forms a discontinuous band encircling
the longitudinal bar. On the west side of the bar,
subarea 6 occurs between elevations of 5 and 6 ft (1.5
and 1.8 m). On the east side, it occurs between
elevations of 7 and 9 ft (2.1 and 2.7 m). The grasses
of subarea 6 grade into the adjoining upslope subarea
of willows.
Subarea 7. Willows
Willows cover the highest elevations on the longitudinal
bar. At the upstream end of the bar, unprotected from
the prevailing current during flooding, the willows begin
at an elevation of 11 ft (3.4 m). At the downstream
end, they begin slightly below 8 ft (2.4 m). The willow
subarea consists of a central core of taller willows
which generally exceeds 4 ft (1.2 m) in height. Above 14
13 ft (4 m), these taller willows form a complete
thicket (Fig. 6). Outside of the central thicket,
the willows become less numerous and are gradually
replaced by grasses. Numerous sediment ridges formed
by ice shove occur in this subarea. A prominent ice
shove ridge is located just north of sample site 781
(Fig. 6). On the downcurrent side of this ice shove
ridge, there is a prominent topographic depression.
The two large depressions located east of sample
site 795 are downcurrent of several sm aller.ice shove
ridges. The topographic depression east of sample
site 778 was formed by scour during spring flooding.
S id e B ar
Subarea 8. Unvegetated sand, silt, and clay
Subarea 8 is referred to on Figure 6 as unvegetated
nongravels and consists of predominantly sand and silt
with a minor amount of clay. This subarea extends
from sea level to approximately 4 ft (1.2 m). The
dominant topographic feature is a series of wave cut
terraces formed during fluctuations in stage.
Subarea 9. Willows
The willows of the side bar occur above 4 ft (1.2 m)
on a series of terraces which adjoin the base of the
adjacent river bluff. PROCEDURES
Samples Collected
Field studies were conducted from early May to early August of
1971. The field work Included: determination of amount of snow cover, recording of morphologic changes centered around breakup, determina tion of amount of sediment thaw, surveying of the bars, and collection of sediment samples.
One hundred and thirty-eight sediment samples were collected including: (1) 103 surficial sediment samples from the longitudinal bar, (2) 5 pairs of samples, each including a sample of ice rafted sediments from the longitudinal bar during breakup and a sample of surficial bar sediments at the same location, (3) 1 sample of sus pended sediment deposited upon willow branches on the longitudinal bar during the spring flood of 1971, (4) 1 sample of sediment deposited on a screen on the surface of the longitudinal bar by the spring flood of 1971, (5) 9 surficial sediment samples from the side bar, (6) 13 grab samples of bottom sediments of the Colville River, and (7) 1 sample of wind-blown sediment incorporated into a snowdrift at Putu approximately 3.1 mi (5 km) downstream from the bar.
The location of each of the 103 surficial sediment samples is plotted on Figure 6. For purpose of analysis, the surficial samples are classified according to (1) environmental subareas, (2) elevation of the sample site, and (3) distance downstream. Sea level is the datum used for elevation and distances are established from the upstream end of the study area which corresponds with the first river bottom
15 16 sediment sample site, 748 (Pig. 6).
Method Employed
The grain size statistical parameters of mean diameter, standard deviation, skewness, and kurtosis employed herein are derived from the methods developed by Folk (1968, p. 45-48). The verbal classification used for the grain size statistical parameters is according to that of
Folk (1968). These measures were devised by Folk and Ward (1957) in order to identify the characteristics of bimodal sediments in the
Brazos River, a braided river. Jaquet and Vernet (1976) found that essentially the same interpretation can be obtained from the Folk sta tistics of mean diameter, standard deviation, and skewness as from moment measures. For kurtosis, however, the Folk statistic is a ratio of the sorting of the tails of the distribution over the sorting of the center of the distribution; whereas Jacquet and Vernet (1976, p. 309), reporting from the work of Jones, define moment kurtosis as a measure of the length of the tails relative to a normal distribution of the same variance.
The Folk grain size statistical parameters were obtained via computer analysis.
Further details of the methods and procedures employed are pre sented in Appendices II, III, IV, V, and VI. MORPHOLOGIC CHANGES
Before Spring Flooding
Thickness of Snow Cover
Field studies began on May 6th. At this time only the tallest willows on the longitudinal bar projected above the snow (Fig. 7) which covered the longitudinal bar, side bar, and the river ice.
The first half of May was a time of occasional light snows and predominantly northeast winds. The light snows had little effect on the net snow depths. The winds, however, did alter the forms developed on the snow cover (Fig. 8) and occasionally exposed isolated areas of limited extent on the longitudinal bar (Fig. 9). For the movement of the frozen bar sediments, the wind was almost totally in e f f e c t i v e .
On May 13th and 14th, the thickness of the snow cover was determined at 5 foot intervals along 5 traverses (Figs. 10-14) across the longitudinal bar and the adjacent ice cover of the side channel.
The figures show the influence of topography, direction of exposure, and vegetation on the thickness of snow cover. Snow depths on the longitudinal bar are summarized in Table 1.
Northeast winds, which prevailed during the winter months, removed snow from the tundra surface east of the study area. Some of this snow accumulated as thick drifts on the lee of the bluff adjacent to the side channel. These d rifts, which persisted throughout the month of May, extended along the right bank and completely covered the
17 18
Figure 7. In early May the complete snow cover of the longitudinal bar. Only the taller willows (i.e., greater than 4 feet) project approxi mately two feet above the snow.
Figure 8. Effects of deflation on a snowdrift, approxi mately one foot high, by the prevailing northeast w in d s . 19
Figure 9. Deflation of snow from a northeast facing slope of the longitudinal bar in mid-May. The taller willows (i.e., greater than 4 feet) comprise the vegetation in the background. A five foot long rod can be seen in the foreground.
Figure 10. Snow cover in mid-May, 1971, profile A-A * of F ig u re 6 . 20
n o f c i r
S N O W G O V t R MO-MAY >971
Figure 11. Snow cover in mid-May, 1971, profile B-B' of F ig u re 6.
Figure 12. Snow cover in mid-May, 1971, profile C-C' of Figure 6. Figure 13. Snow cover in mid-May, 9171, profile D-D' of F ig u re 6 .
SNOW COVEN MID-MAY 1971
Figure 14. Snow cover in mid-May, 1971, profile E-E1 of F ig u re 6 . 22
TABLE 1 . SNOW COVER OF LONGITUDINAL BAR1
MAX MIN AV PROFILE2 DESCRIPTION in cm in cm i n cm A-A1 18 4 5 .7 0 0 7 .3 1 8 .5 across unvegetated gravel sheet B-B' 32 8 1 .3 0 0 9.9 25.1 across unvegetated gravel sheet C-C' 78 1 9 8 .1 1 2 .5 2 3 .1 5 8 .7 through taller willows D-D' 40 1 0 1 .6 0 0 1 4 .3 3 6 .3 through taller willows E-E' 23 5 8 .4 0 0 1 1 .0 2 7 .9 through smaller willows
^Does not include: (1) snow cover above river ice (2) drifts at base of bluffs
^See Figure 6 for profile locations side bar. Therefore, until the spring flood, the side bar remained virtually unchanged.
From the 5 traverses, the average depth of snow cover on the bar was found to be 14.2 in (36 cm) with a range of 0 in (0 cm) to 78 in
(198 cm). Traverses A-A' and B-B', which are across the unvegetated gravel sheet (subarea 3) at the south end of the bar, had the minimum average thicknesses, 7.3 in (18.5 cm) and 9.9 in (25.1 cm), respec tively. The maximum thicknesses on the gravel sheet are 18 in (45.7 cm) and 32 in (81.3 cm) respectively. The greatest snow accumulation was found along traverses C-C' and D-D1 in the area of the taller willows, those more than 4 feet in height, which occur at the higher elevations on the bar. The maximum accumulations for C-C' and D-D' were 78 in
(198.1 cm) and 40 in (101.6 cm), respectively, with an average of 23.1 in (57.7 cm) and 14.3 in (36.3 cm), respectively. Traverse E-E' through the small willows, less than 4 feet in height, had a maximum thickness of 23 in (58.4 cm) and an average thickness of 11 in (27.9 cm).
The snow depth data indicates that within the study area the 23
thickest snow accumulations occur on the lee of the bluff and within
the large willows at the higher elevations of the longitudinal bar.
These general characteristics of snow accumulation, as recorded in mid-
May of 1971, probably do not differ appreciably from those that existed
throughout the winter months.
Ablation of Snow Cover
By May 13th, portions of the longitudinal bar had been exposed
and remained clear of snow. The exposed areas of the bar included the west facing slope of the linguoid gravel ridge, the unprotected north
east facing slopes, and the wave built ridges along strand lines
formed during varying river-stages of previous years. The exposed
strand lines occurred primarily on the gently sloping southern and
northern extremities of the bar. Except for the west facing slope of
the linguoid gravel ridge, which was exposed primarily due to the
effectiveness of insolation on the thin snow cover, the exposed areas were formed by deflation of the snow cover by northeast winds and by
subsequent enlargement by melting.
With the removal of the protective snow cover from portions of northeast facing slopes, the exposed surficial sediments which had thawed locally were partly removed by the wind. Some of the deflated
sediments were deposited downwind upon the snow cover of the main channel. In snow covered areas, where the albedo was decreased by the deposition of a thin veneer of wind blown silt, surface temperatures
increased more rapidly than on clean snow and isolated pools of m elt- water formed on the river ice. Had the layer of wind blown sediment been thick (i.e., several cm) the effect on the buried snow would have 24
been to insulate the snow and retard melting (Walker and McCloy, 1969).
Between elevations of 0 ft and 8 ft (On and 2.4 m) extensive areas of the northeast facing slope of the longitudinal bar were exposed by May 20th. At this time the thick accumulations of snow in the willows, located above the 8 foot contour, were actively melting. As the meltwater flowed downslope and through the exposed areas of north east facing slopes,.it formed small stream lets which eroded surface sediments on the bar. As the stream lets flowed downslope onto lower- gradient slopes, small delta-shaped lobes of sediment were deposited
(Fig. 15). These lobes occurred at the bases of terracets cut by wave action during previous years, at the edges of small pools of water impounded behind wave-built ridges along strand lines formed in previous years, on the surface of snow cover remaining on the gentle lower slopes of the bar, and on the adjacent river ice. Although this erosion occurred on other parts of the bar, it was most extensive on northeast facing slopes. The likelihood of the delta-shaped lobes surviving the spring flood is remote.
Thickness of River Ice
On May 20th a hole was bored through the ice which was found to be 74 in (1.88 m) thick. At this depth, water was in contact with the ice and had a temperature of 0.0°C. It is to be expected that in sections of channel less than 74 in (1.88 m) in depth, the river ice would have been frozen to the channel bottom sediments.
Throughout the Colville River delta, the thickness of river ice in late winter is found to be approximately 6 ft (1.8 m) (Arnborg, et al., 1967). 25
Figure 15. Sediment exposed in mid-May on a northeast facing slope of the longitudinal bar. Meltwater has built a small delta lobe at the base of a wave cut terrace which formed during the previous year. A rock hammer is present in the center foreground.
Spring Flooding
During spring flooding, changes in stage were not obtained for the study area. However, at Section 1 (Fig. 2), 0.6 mi (1 km) up stream, a complete record was obtained and compiled by Walker (personal communication). Throughout' the following discussion, the stages, as given for Section 1, are considered to be representative of water levels in the study area.
By the latter part of May, meltwater had begun to accumulate on top of the river ice. The first measurable discharge of the Colville
River occurred on May 28th. At this time the stage rose from 3.1 in
(8 cm) to 44.5 in (1.1 m) in a period of 20 hours.
The water level continued to rise and on June 1st at 1700 hrs 26 the stage had reached approximately 11.5 ft (3.5 m). At this time the side bar was entirely submerged. On the longitudinal bar only the higher vegetated areas of grass and willows remained emergent (Fig. 16).
On the river channel the ice frozen to the channel bottom sediments was submerged. However, river ice not frozen to the bottom broke loose from the bottom-fast ice and was buoyed upward by the rising flood waters. This mass of ice floated as a more or less solid sheet which had only slightly disassembled along cracks originating either at or after the time of freezeup (Figs. 17 & 18). The floating ice sheet in most cases still carried its cover of snow (Fig. 19).
Throughout the delta of the Colville River, just prior to break up, bands of floating ice marked the deeper parts of channels (Arnborg, et al., 1966, p. 204). Within the study area, the plan view of the floating ice bands was that of a Y pointing downstream. The Y-shaped area included sinuous bands in both the side channel and the main channel. Downstream from the longitudinal bar, the two bands coalesced.
Almost completely surrounding the bands of floating ice were the flood waters which were nearly devoid of surficial ice.
Until breakup, the flood waters of the Colville River were forced to flow around and tinder the bands of floating ice, which re mained almost stationary. Occasionally a block of floating ice was observed being moved downstream by the current. If the block were located at some position within the center of the Y-shaped bands of ice, it would ultimately come to rest at the junction of the two bands. In this manner the bands of ice became temporary barriers to the down stream movement of ice blocks. Behind this temporary dam, ice blocks 27
Figure 16. Upstream aerial view of the almost totally sub merged, snow covered longitudinal bar taken June 2nd approximately 2 hrs before breakup. Only the willows remain above water at a river stage of 13 ft (4.0 m). The across bar width is approximately 300 ft (91 m).
/ /
Figure 17. Upstream aerial view of prebreakup spring flooding. Floating ice bands approximately 300 ft (91 m) wide mark the deepest parts of the main channel and the side channel. Figure 18. Floating ice blocks impounded at the point of junction of the floating ice bands of Figure 17. Beyond the blocks is the snow covered and partially submerged longitudinal bar. The arcuate area of open water (left foreground) is the position of the submerged side bar.
Figure 19. View of the band of floating ice in the side channel with the slightly disassembled thick snowdrifts resting on the river ice in the lee of the bluffs. 29
continued to collect until the time of breakup (Figs. 17, 18, & 19).
On the longitudinal bar, two topographic depressions extend
diagonally across the south end of the willow area (Fig. 6). When the
river stage exceeds 11 ft (3.4 m), as it did on June 1st, the most
southerly of the depressions becomes a through channel across the width
of the willows. When the river stage exceeds 13 ft (4.0 m), the second
depression becomes a through channel. As the flood waters rise, not
only is the river current directed through the main channel and side
channel, but also a limited amount of flow occurs across the longi
tudinal bar through the two rather restricted and narrow channels in
the willows. The current in these channels flows from the main channel
into the side channel. These cross-bar flows appear to develop yearly.
This is apparent from small channels cut by stream lets into the north
east facing slope of the longitudinal bar. On aerial photographs
these small channels are found to remain in approximately the same posi
tion from one year to the next.
Within the study area, breakup occurred on June 2nd and 3rd.
At Futu breakup began at 1900 hrs on June 2nd. In 1971 breakup was
observed occurring first at the head of the delta and then progressing downstream (Walker, 1973b). Therefore, it is concluded that breakup in the study area commenced a short time prior to that of Futu.
During breakup, almost the entire surface width of the river channel became a mass of floating ice blocks. These blocks were derived from fragments of the floating ice bands which became disassem bled and from bottom-fast ice which broke lose from the channel bottom.
In some instances, the ice blocks carried sediment which was either 30
deposited on their upper surfaces and/or frozen onto their lower sur
faces. The entire mass of floating ice blocks was being carried down
stream by the current.
The variations in stage which accompany breakup are critical
in determining the ability of the ice blocks to alter topographical
obstructions which they encounter. If the stage is rising, the blocks
are more likely to be floated across obstructions and w ill have little
to no effect on the topography. If the stage is falling, the blocks
are more likely to encounter the bottom and thereby alter obstructions.
When breakup began in 1971, the stage was rising. A few hours
later, the stage reached a maximum and the remainder of breakup corres
ponded to a time of falling stage. This drop in stage facilitated the
grounding of ice blocks on the longitudinal bar.
The maximum height to which the stage rose on the longitudinal
bar was determined by noting the maximum height to which sediment had
been deposited upon the branches of the large willows. The maximum
height was 17.8 ft (5.4 m), which equaled the maximum stage recorded at
Section 1. This stage occurred on June 3rd. At the time of maximum
stage, the longitudinal bar was almost totally submerged.
The effects of submergence of the longitudinal bar on snow
cover can be seen in Figure 20. On June 2nd, just prior to the almost
total submergence of the bar, snow cover was determined at 5 foot
intervals along a traverse across the width of the longitudinal bar.
At this time the maximum snow depth was found to be 50 in (127 cm).
After breakup on June 7th, snow depths were again determined at 5 foot
intervals along the same traverse. At this time a maximum snow cover 31
Jw m 7,1971,
-4
-3
-2
<00 MTtM 400 BOO FCCT 1000
-4
Figure 20. Snow cover and depth of thaw on the longitudinal bar profile F-F' of Figure 6. Water levels are given for June 2, 3, and 7. 32 depth of 36 in (91.4 cm) was recorded. Even after an almost total submergence of the longitudinal bar, flood water was incapable of completely removing the snow cover.
During a series of warm days rapid melting of the snow cover occurred on the longitudinal bar. At this time a maximum temperature of 22.8°C was recorded on June 7th at Putu (McKenzie and Walker, 1974).
By June 9th the longitudinal bar was almost devoid of snow.
With emergence of the longitudinal bar (Figs. 21-24), the effects of spring flooding became apparent. On June 6th, after breakup, a total of 57 ice blocks were found grounded on the bar (Fig. 21). The blocks were prim arily positioned on the bank adjacent to the main channel between elevations of 11 and 13 ft (3.4 and 4.0 m) (Fig. 6).
Of the 57 blocks, only 7 were found grounded on the bank adjacent to the side channel.
When the stage dropped after breakup and the ice blocks melted, depressions a few inches in depth were found at the positions of the former ice blocks. These depressions formed due to sedimentation around the blocks during flooding and were further accentuated by small delta-like sediment lobes deposited by meltwater around the base of the blocks. During melting, sediment present on the blocks was concen trated into cracks which had developed in the ice. When the blocks had fully melted, small ridges of sediment were formed where cracks had e x i s t e d .
Very little visible alteration of surficial sediments of the bar can be directly attributed to the motion of specific ice blocks.
Trails left by ice blocks touching bottom were the most abundant Figure 21. On June 6th, a postbreakup aerial view of the longitudinal bar with the main channel on its lower side and the side channel on its upper side. The small light areas on the bar (lower side) are grounded ice blocks. The larger light areas are snowdrifts. See cross bar flow channel, upper middle right edge.
"V..
Figure 22. Downstream aerial view, June 11th, across the longitudinal bar at a stage of 7 ft (2.1 m). See the partially emergent gravel sheet, subarea 3 (lower left). See the partially emergent linguoid shaped gravel ridge, subarea 4 (upper l e f t ) . Figure 23. Looking upstream, June 11th, across the longi tudinal bar. Faintly outlined below the emer gent portion of the bar is the linguoid shaped gravel ridge, subarea 4 (lower right).
Figure 24. The downstream tip of the longitudinal bar, June 11th, with the emergent portion of the linguoid shaped gravel ridge, subarea 4 (lower right). 35
feature. These trails, developed in both fine and coarse grained materials (Figs. 25 & 26), consisted of two or more parallel ridges having a height of a few inches at most and extending for several feet.
The tra ils were usually somewhat discontinuous and ended abruptly.
Therefore most tra ils must have been formed by blocks which temporarily touched bottom.
The trails formed in sands and silts are not likely to be preserved from one year to the next; whereas, the trails formed within the gravels are more likely to be preserved. A section of a trail
located on the gravel sheet (subarea 3) was painted prior to flooding.
After flooding some of the gravels had been moved but most were s till i n p la c e .
Of the 57 ice blocks grounded on the longitudinal bar in 1971, only one was found associated with its ice-shoved ridge of sediment
(Figs. 27 & 28). This ridge of ice-shoved sediment differs from the ice trails in that it is formed transverse to the direction of block motion and is composed of a single ridge having a length of only a few
feet and a height which may exceed several feet. Above elevations of
12 ft (3.7 m), the longitudinal bar is almost encircled by ridges of sediment formed by ice-shove. If the conditions of 1971 were typical, and they appear to have been, it seems likely that only a few ice-shove ridges form each year. Their present abundance on the bar can be a ttri buted to their preservation from one year to the next.
On the longitudinal bar at the north end of the willows, there are two large depressions. These two depressions may have formed by ice-shove. The sides of the depressions nearest the main channel were 36
F ig u re 2 5 . Trail formed in the graveIs, subarea 3, of the longitudinal bar by an ice block temporarily grounded during b re a k u p .
Figure 26. Trail left in ripple marked silts on the longitudinal bar by an ice block temporarily grounded during b re a k u p . Figure 27. The only grounded ice block on the longitudinal bar directly associated to ice-shoved sediment. It is in an upright position and rests on top of the ice-shoved sediment.
Figure 28. The ice-shoved sediment ridge after the melting of the ice block shown in Figure 27. 38
the sites where the largest number of ice blocks were grounded during
1971. They were also the sites where the only ice-shove ridge was
formed in 1971.
During the higher stages of spring flooding, flood waters
flowed from the main channel into the side channel. A limited flow also occurred through the two narrow channels defined by topographic
depressions across the width of the south end of the willows on the
longitudinal bar. When the river stage fell below 13 ft (4.0 m), the most northerly of the two small channels became separated from the main
channel of the Colville River by the emergence of an ice shove ridge which extended across the west mouth of the channel. When the stage
fell below 11 ft (3.4 m), the more southerly channel became separated
from the main channel and the side channel. As the stage continued to fall, the two depressions, through which the flood waters had flowed, became the site of small ponds perched above the. level of the river.
These ponds drained easterly through small stream lets flowing down the east side of the longitudinal bar. The position of these stream lets appears to fluctuate little from year to year.
After the maximum stage of 17.8 ft (5.4 m) as recorded on June
3rd, the stage dropped rapidly reaching sea level on June 16th (Fig. 4).
This date is considered as marking the end of spring flooding.
While the stage was dropping from the maximum on June 3rd, the river exhibited periods when the stage became stabilized, periods of wind generated waves, and periods of relative calm. The effect of the times of stabilized stage during emergence of the bars was to allow the formation of a series of strand lines. These strand lines are the most 39
extensive and most numerous morphologic features formed on the bar.
Depending upon the location of the strand line and the conditions at
the time of formation, the strand lines consist of wave cut terracets
and/or wave built ridges (Figs. 29 & 30).
Summer Flow Regime
With the end of spring flooding, the stage had reached approxi mately sea level. Throughout the summer of 1971 the river fluctuated about this level. Minor fluctuations were probably caused by localized
summer thundershowers.
During the summer months, the wind became an effective and important agent for sediment erosion, transport, and deposition. On the surface of the bars, especially the longitudinal bar, wind blown sediments formed sand shadow dunes in clumps of grasses and willows
(Fig. 31). These dunes became the most prominent morphologic features
formed at this time. Also, on the east side of the longitudinal bar below the crest of wave cut terracets, wind blown sands formed linear sand bands parallel to the direction of the prevailing wind (Fig. 32).
These bands were only a few sand grains thick, a few inches in width, and a few feet in length. They were transient features which were obliterated with changes in wind direction.
The summer of 1971 was a time characterized by: (1) only slight fluctuations around sea level in the stage, (2) aeolian erosion, trans port, and deposition of sediments, and (3) minor to almost non-existent changes induced by precipitation from localized thundershowers.
The summer flow regime came to an end on October 2, 1971, when Figure 29. Wave built gravel ridges on the surface of the gravel sheet (subarea 3) of the longitudinal bar.
Figure 30. A series of wave cut terraces as viewed in July on the side bar. The terraces are backed by willows and the river bluff. 41
Figure 31. Sand shadow dunes formed around the shorter willows (I.e., less than 4 feet) on the longitudinal bar.
Figure 32. The northeast facing slope of the longitudinal bar and the adjoining side channel. Strong southwesterly winds (from left to right) have deposited long linear bands of sand at the base of the wave cut terrace. 42 freezeup occurred on the channel of the Colville River (Walker, per sonal communication). At that time, the river was at approximately sea l e v e l . SEDIMENTS
On the longitudinal bar, an approximation of the areas of
fluvial sedimentation and the depositional sequence therein was estab
lished by field studies for the spring flood of 1971. As an indication
of deposition from past floods, the sedimentary sequence below the bar
surface was determined to the base of the thawed layer. The results
of this procedure are given below.
During flooding in 1971, gravel particles were neither eroded nor deposited upon the surface of a screen (Fig. 33) anchored on the gravel sheet (subarea 3) near sample site 744 (Fig. 6). This factor appears to indicate that the gravels were not actively transported during flooding.
In contrast, a 1 in (2.5 cm) layer of sediment was deposited by the spring flood on a screen located near sample site 763 (Fig. 34) on the sands, silts, and clays of subarea 5. This deposit includes a basal layer of sand and a surface drape of silt and peat (Fig. 35).
This sequence appears to be fairly representative of the deposits of one spring flood.
A cross-section of sediment to the base of the thawed layer of the longitudinal bar is presented in Figure 36. The columnar sections were obtained by excavating holes below the surface of the bar during the month of July, measuring the units in these sections, and recording the characteristics of the sediments.
From the cross-section, the following can be concluded:
(1) Near sample sites 749 and 750 at the contact of the
43 Figure 33. Looking northeast across the stoss side o£ the longitudinal bar. Except for three gravel particles, the screen (lower left) is free of sediment from the spring flood. The sharp contact between the gravel sheet and the sands of subarea 5 is located at the abrupt change in the surface texture of the bar.
gravel sheet (subarea 3) with the unvegetated sands,
silts, and clays (subarea 5), the gravels appear to
continue beneath the unvegetated nongravels.
(2) Many of the sections outside of gravel subarea 3
consist of repetitive units of a thicker layer of
sand overlain by a thinner layer of silt (e.g.,
sample sites 775, 778, and 802). This sequence is
similar to that observed for the spring flood of 1971
(3) Excluding gravel subarea 3, a general decrease in
grain size occurs with elevation. Progressing from
sample site 830 upward, the columnar sections are
composed prim arily of sands; with increased elevation Figure 34. The outline of a screen, subarea 5, is emerging as sediments deposited by spring flooding are dried and deflated. See Figure 35 for a close-up of lower right.
Figure 35. Close-up of lower right (Fig. 34) showing approximately one inch of sediment in cross-section. See the cross-laminated sand with an overlying drape of peat and silt which has been left overhanging by deflation of underlying sands. I*» METERS .w 1 £ 1 ^ FEET
CD m a m
792 791 790 793
798 706 794 793 p 797, SKSt
17777)
79 801 800
80i
803
803,
EXPLANATIO
GRAVEL rrrir, 6 SAND
SILT □ > > : CLAY
VERTICAL EXAGGERATI
HORIZONTAL SC/ 100 0 100 200 300 HHH-H Hhr I V. .._l= •-= l.~
a O'
792 791 793 m c-z- ~i—z~ V*»“.
MWW
770 J 7 4
VTTTn
EXPLANATION
GRAVEL H PEAT
rrm FROZEN SAND GROUND SILT □UNKNOWN
ERTICAL EXAGGERATION IOOX
HORIZONTAL SCALE 100 200 300 400 500 FEET
0 100 METERS I- 1 — F =1 SAMPLE NUMBER
^ENVIRONMENTAL SUBAREA
Figure 36. Sediment variation to the base of t thaw la y e r , p r o f i l e M-M' o f F ig u re 46 METERS
1/7//,
724
F ig u re 36 Sediment variation to the base of the thaw la y e r , p r o f i l e M-M1 o f F ig u re 6 .
tr m 47
silts become more abundant (e.g., sample site 805);
and at the highest elevations along the cross-
section, clays are abundant (e.g., sample site 791).
Grain Size S tatistical Parameters
Sediments of the study area can be placed into two groups
according to whether or not they are likely to be moved during flooding.
River bottom samples and the samples of the nongravel areas of the bars
from subareas 1, 2, 5, 6, 7, 8, and 9 are composed of particles of
sizes which were actively transported during the spring flood of 1971.
Those samples constitute Group I. The gravel bar samples from sub-
areas 3 and 4 are composed of particles of sizes which were not
actively transported. These samples constitute Group II. The two
groups are analyzed separately to determine if significant relation
ships exist for mean diameter, standard deviation, skewness, and
kurtosis with elevation and longitudinal position.
Sediments whose grain sizes surpass the competency of the river
during spring flooding can be deposited onto the bars by ice rafting.
Differences between bar surface sediments and ice rafting sediments are
com pared.
The influence of environmental factors on the distribution of the four aforementioned parameters is explored via comparisons between
th e 9 s u b a r e a s .
Change With Elevation and Location
For Group I and Group II, respectively, a stepwise regression procedure (Appendix IV) was used to obtain models of best fit for 48 significant changes in mean diameter, standard deviation, skewness, and kurtosis with elevation and distance downstream. Predicted curves with 95% confidence intervals about the mean are presented below.
Group I —Submerged River Sediments and A ll Nongravel Bar Sediments
Stepwise regression analyses show that significant relation ships exist between the vertical position of the sediments and their mean diameter, standard deviation, skewness, and kurtosis. Their longi tudinal position is accompanied by significant changes in mean diameter and kurtosis (Figures 37-40 and Table 2).
The predicted model for change in mean diameter with change in elevation (Fig. 37) is linear and accounts for 50% of the variance.
The maximum predicted mean diameter, X ^ » 0.3 0, occurs at -16.1 ft
(-4.9 m). With increasing elevation, the predicted mean diameter decreases. The minimum value, X ^ a 5.0 0, occurs at 13.1 ft (4.0 m).
The predicted model for standard deviation with change in elevation
(Fig. 38) accounts for 26% of the variance. The maximum predicted value of standard deviation, Xg = 2.3 0 (very poorly sorted), occurs I at the lowest elevation, -16.1 ft (-4.9 m). As elevation increases within the river channel, the sediments become better sorted. Between
-8 ft (-2.4 m) and -5 ft (-1.5 m), the mean predicted value is 1 0
(moderately sorted). From -5 ft (-1.5 m) to 9.5 ft (2.9 m) sorting again decreases to an approximate value of 1.9 0 (poorly sorted).
From 9.5 ft (2.9 m) to 13.1 ft (4.0 m) sorting improves slightly, reaching a value of 1.7 0 at the terminus of the curve. The predicted relationship between skewness and elevation (Fig. 39) accounts for 13% 49
MEAN DIAMETER
** , ELEV. IN ET. A a* * A» * A
Figure 37. Group I predicted curve with 95% confidence intervals about the mean for mean diameter versus elevation. In Figures 37-44, all predicted curves for Group I, river samples plus a ll nongravel bar samples, are determined only from the position of dots. All pre dicted curves for Group II, gravel bar deposits, are determined only from the position of triangles. Equations for models of best fit are given in Table 2.
STANDARD DEVIATION
i W
ELEV. IN FT.
Figure 38. Group I predicted curve with 95% confidence intervals about the mean for standard deviation (given in 0) versus elevation. For further explanation see Figure 37. 50
SKEWNESS
0.5
o.o -iis -is -ii -II —oh- a< i -i -i ELEV. IN FT.
-0.5
- 1.0
Figure 39. Group I predicted curve with 95% confidence intervals about the mean for skewness. For further explanation see Figure 37.
KURTOSIS
ELEV. IN FT. Figure 40. Group I predicted curve with 95% confidence intervals about the mean for kurtosis. For further explanation see Figure 37. 51
TABLE 2 . EQUATIONS FOR PREDICTED CURVES
I I RIVER PLUS NONGRAVEL BAR DEPOSITS GRAVEL DEPOSITS « r 2 Model of Best Fit Xr2 Model of Best Fit
M ean D iam eter 50 5?M i - 2 .9 + Q .2 E I
z S tandard - 26 o Deviation Kt( - 1.3+0.8xld,EI+a2xld2EI2-0Axl0"3EI3 t< - > U i _i U i Skewness 13 *S k,“ 0.2+0.3x 1 o 'e I _ 0.1 x 103EI3
K urtosis 9
M ean D iameter 39 - -2A+O^DIS-O.8x l()2DIS2+0«4xld*D S3 58 - 1.9 + 0.4 x 10”' ° DIS3
UJ U S tandard Z D eviation 30 = 2.3 + 0.4xl0'7 DIS2 < m o Skewness 22 XSk( “ 0.4-0.5xlon DIS3
Kurtosis 12 * K G " 2-5 "0'8xlff,DIS+0.2 xld2DIS2 -0.1xl04DIS3
of the variance. The undulating predicted curve for skewness varies
between 0.29 (fine skewed) at -16.1 ft (-4.9 m), 0.05 (near symmetri
cal) at -9 ft (-2.7 m), 0.38 (strongly fine-skewed) at 8.7 ft (2.7 m) and 0.29 (fine skewed) at 13.1 ft (4.0 m). The predicted model for
kurtosis with change in elevation (Fig. 40) is linear. The model, accounting for 9% of the variance, varies from a maximum predicted mean kurtosis value of 1.87 (very leptokurtic) at -16.1 ft (-4.9 m)
to a minimum of 1.28 (leptokurtic) at 13.1 ft (4.0 m).
With distance downstream, the predicted changes for mean 52
diameter and for kurtosis are not simple relationships (Figs. 41 & 42).
The maximum predicted grain size (X^ = -2.5 0) occurs at the origin of
the predicted curve for mean diameter (Fig. 41). The curve, which accounts for 39% of the variance, exhibits a decrease in grain size through the first one-third. In the central one-third the finest predicted values occur (Xj^ = 4.3 0) and in the last one-third grain
size again increases to a predicted value of X^z a 2.6 0. The curve
for kurtosis (Fig. 42) accounts for only 12% of the variance. At the origin of the curve the predicted value of Xj^ =2.5 corresponds to a very leptokurtic distribution. Progressing downstream the central portion of the predicted curve has approximate mean kurtosis value of
1.5. At the most distant point downstream the predicted value, Xj^ =
1.0, corresponds to the kurtosis of a normal distribution.
Group II—Gravel Deposits of the Longitudinal Bar
Stepwise regression analyses indicate that no significant rela tionships exist between elevation and the following: mean diameter, standard deviation, skewness, and kurtosis. With change in distance downstream significant changes do occur in mean diameter, standard deviation, and skewness. However, no significant relationship exists between distance downstream and kurtosis (Figs. 41, 43, 44 and Table 2).
For mean diameter the predicted curve (Fig. 41), which accounts for 58% of the population variance, exhibits a slight decrease in grain size with distance downstream. Mean diameter varies from XMz = -2.0 0 at the upstream end of the curve to X^ = -1.0 0 at the downstream end.
The predicted curve for standard deviation (Fig. 43) lies entirely within the region of poorly sorted sediments. The curve indicates that 53
MEAN DIAMETER
8000 6000 7000 8 0 0 0 9 0 0 0 Ui -I DISTANCE DOWNSTREAM IN FT.
Figure 41. For Group I and Group II, the predicted curves with 95% confidence intervals about the mean for mean diameter versus longitudinal position. For further explanation see Figure 37.
KURTOSIS
1000 2000 3 0 0 0 4 0 0 0 0000 6000 70006 0 0 0 9000
DISTANCE DOWNSTREAM IN FT. Figure 42. Group I predicted curve with 95% confidence intervals about the mean for kurtosis versus longitudinal position. For further explanation see Figure 37. STANDARD DEVIATION IE
#v»Q? t
1000 2000 3005 40(56 5606 DISTANCE DOWNSTREAM IN FT.
Figure 43. Group II predicted curve with 95% confidence intervals about the mean for standard devia tion versus longitudinal position. For further explanation see Figure 37.
SKEWNESS
0 .5
0.0 1000 2000 3000 '4000 5 0 0 0 DISTANCE DOWNSTREAM IN F T X
-0 .5
- 1.0
Figure 44. Group II predicted curve with 95% confidence intervals about the mean for skewness versus longitudinal position. For further explanation see Figure 37. 55 the gravel becomes more poorly sorted downstream (X- has a range L:rom I 2.3 0 to 3.2 0). The predicted curve for skewness (Fig. 44) falls within the range from strongly fine skewed (Xg^ = 0.42) at the up stream end to coarse skewed (Xskj a _0.01) at the downstream end.
Further analyses are necessary in order to interpret the shapes of the other predicted curves.
Environments1 Subareas and Grain Size S tatistical Parameters
Comparisons between environmental subareas are presented in
Appendix V and summarized in Tables 3 and 4. Only the comparisons which are pertinent to the determination of the predicted regression curves w ill be discussed herein.
The results of Group I comparisons (Table 4) between subareas indicate the following:
(1) no significant differences exist between vegetated subareas
6 and 7 on the longitudinal bar and vegetated subarea 9 on
the side bar,
(2) no significant differences exist between unvegetated
subarea 5 on the longitudinal bar and unvegetatcd subarea
8 on the side bar.
( 3) unvegetated subareas 5 and 8 of the longitudinal and side
bar are coarser than vegetated subareas 6, 7, ami y.
(4) Lhe longitudinal bar sediments, subareas 5, 6, and 7, are
coarser than those of the side bar, subareas 8 and y. TABLE 3. COMPARISONS BETWEEN ALL SUBAREAS
*M z 5T x S k t I RIVER 1.2A 1.18 0.07 1.59 3 . - I S . 1 ,2 BAR 2 . 8 6 2 . 0 3 0 . 2 8 1 .2 1 1 sas. 3,A,5,6,7,8,9 • F 2 2 2 . 3 1 * * 5 5 9 .9 5 * * 6 8 . 5 0 * * 1 1 1 .9 5 * *
RIVER 1 .2 A 1 .1 8 0 . 0 7 1 . 5 9 s a s . 1 , 2 BAR NONGRAVELS A . 15 1 .7 3 0 . 3 1 1 .3 A sas. 5,6,7,8,9 F 8 8 . 6 2 * * 12.00** 9.57** 5 .A 2 *
MAIN CHANNEL 1.88 s n . 1 SIDE CHANNEL N.S, N.S. N.S. 1.30 s a . 2 F 1 .A 7 0 . 7 3 0.08 A. 7 A*
R IV ER 1 .2 A 1 .1 8 0 . 0 7 1 .5 9 s a s . 1 , 2 GRAVEL DEPOSITS -0.38 2.76 0.19 0.87 s a s . 3 , A F A6.38** 72.1A** 9.00** 37.26**
GRAVEL SHEET - 1 . 6 9 2 .A 2 0 . 3 9 s a . 3 LINGUOID CRAVEL 0 . 9 3 3 . 1 0 - 0 . 0 1 N .S . s a . A F 2 A . 9 9 * * ' 8 . 8 1 * * 8 . 9 0 * * 0 . 2 0
GRAVEL DEPOSITS - 0 . 3 8 2 . 7 6 0 . 8 7 ■ s a s . 3 , A ALL OTHER DEPOSITS 3 . 3 2 1 .5 7 N.S. 1 .A t sas. 1,2,5,6,7,8,9 F . 1 0 3 . 2 A * * 9 0 . 0 0 * * 0 .7 1 A 9 .2 6 * *
I ...... - sas. = subareas ** p <.01 sa. » suburca * p <,05
TABLE A. COMPARISONS OF AVERAGE MEAN DIAMETERS OF ALL SUBAREAS EXCLUDING GRAVELS
! V T D .1 UVTD/ V T D . S, UVTD. S D . BAR L TD . b a r !LTD.J & SD.4 L T D . BAR L TD . BAR BARS s a .- * 9 sa. 6 sas. 6,7,9 s a . 5 saa.6 5,6,7
LGT. BAR N.S. N . S . ! i s , 7 F '0 . 0 2 F - 0 .7 A
LGT. ft N . S . : VTD. 3D. BARS 1 F - 0 . 7 3 ; i s s . 7.9! N.S. 5 3 . 7 5 LGT. BAR 6 , 7 A ,6 8 j a s . 6 , 7 F - 0 . 0 0 i F “ 1 6 .9 $ * * LGT. ft 6,7,9 A.70 • 3D. BARS 5 , 8 3 . 3 3 s a s . 5 . 8 F - 2 3 . 6 7 * * 9 A .7 A i 3 3D. BAR N.S. s a . 8 8 2 . 9 1 j F - A .A 3 * j F“ 3 .5 A 'id 3D. BAR » 5,6,7 A.37 -t t* 8 , 9 3 . 8 3 : a s . 8 , 9 1 i p - 5 . 8 2 * *VTD. "VEGETATED ** p <.01 “UVTD. » UNVEGETATED * p <.05 JLGT. » LONGITUDINAL 4SD. = SIDE -*sa. ** subarea &s a s . = s u b a i u a s Scatter Plots
Scatter plots (Figs. 45-48) show the interrelationship between selected grain size statistical parameters.
In Figure 45 medium sands (1 to 2 0) are the best sorted (i.e., lowest standard deviation). With decreased grain size, sorting be comes poor and with increased grain size sorting also becomes poor.
However, for grain sizes of -1 0 and coarser, sorting improves. This relationship between mean diameter and standard deviation is a reflec tion of the dominance of a gravel mode, a sand mode, and the mixing of the two. The general shape formed by the points in the scatter diagram is that of a sine curve.
In the plot of mean diameter versus skewness (Fig. 46) the general pattern is that of a sine curve. For both gravel (i.e., -1 0) and fine sand, silt, and clay (i.e., 2 0), skewness is positive indi cating a tail toward the finer end of the distribution. In the inter vening area, -1 0 to 2 0, skewness varies from 0 (that of a normal distribution) to that of approximately -0.5 0 and back to 0. The variation can again be explained by the dominance of a gravel mode and a sand mode and the intermixing of the two.
The plot of mean diameter versus kurtosis (Fig. 47) shows that most kurtosis values fall between 0.6 and 1.5. Kurtosis values for the gravels tend to be less than one. This indicates that the tails of the gravel distributions are better sorted than the center and are a pro duct of their bimodality. For the sand, silt, and clay of the river and bars, the kurtosis values are, in general, greater than one indi cating that the centers of the distributions are better sorted than the t a i l s . 58
Figure 4 5 . Scatter plot for mean diameter versus standard deviation. Dots represent samples from river > 3 and bar sands, silts, and clays (Group I). Triangles represent samples from gravels (Group II). Dashed line represents the trend established by Folk and Ward (1957) for a gravel bar in the Brazos River.
MEAN DIAMETER (Mz) IN 4
Figure 4 6 . Scatter plot for mean diameter versus skewness. For further explanation see Figure 45.
MEAN DIAMETER (Mz ) IN 0
Figure 4 7 . Scatter plot for i.'..>.an diameter versus kurtosis. For further explanation see Figure 45,
2-10123 MEAN DIAMETER (Mz) IN 59
• A Figure 48. Scatter plot for skewness versus kurtosis. The region enclosed by ~ o 3 - * the cross formed by solid CO • • lines represents an area
• • * of normal skewness and h 2 ■** a: kurtosis. For further 3 . • • • • explanation see Figure 45. ▲ t o a & f e - 4 1 - *
* A A A A
I 1 i i -.0 D.5 0.0 0.5 1.0 SKEWNESS (SKX)
The scatter plot of skewness versus kurtosis illustrates a lack of points having a symmetrical distribution with high values of kurtosis
(the upper central part of the plot). This is apparently due to the lack of trimodal sediments (Folk and Ward, 1957, p. 21).
Comparisons of Scatter Plots and Regression Curves
Tables 5 and 6 present the results of comparisons between the predicted regression curves of the grain size statistical parameters and the corresponding relationship exhibited by the grain size para meters in the scatter plots. The tables were constructed in the following manner. From the Group I predicted curve for change in standard deviation with elevation, the predicted values of standard deviation at inflection points were read and tabulated in column 3 of
Table 5. The appropriate elevations for the predicted values of standard deviation were tabulated in column 1. Using the Group I pre dicted curve for mean diameter versus elevation, the predicted mean diameters for the given elevations of column 1 were recorded in column
2. Column 4 was determined by using the mean diameters given in column
2 and approximating the corresponding value of the standard deviation 60
TABLE 5 . GROUP X — REGRESSION CURVES VERSUS SCATTER PLOTS
REGRESSION CURVES SCATTER PREDICTED VALUES PLOTS
1 2 3 4 E le v . Mz v s . 2 T i n f t . * Mz in 0 value of 2j -16.1 to -8 0 .3 5 -1 .6 2 .3 - 1 .0 2 .0 - 0 .6 -8 to -5 1.6 -2.1 1 .0 m in. 0 .6 - 1 .0 -5 to 9.5 2.1 -4.4 1.0-1.9 1.0 - 1.9
S k j Mz v s . S k j value of Skj -16.1 to -9 0 .3 5 -1 .5 0 .3 0 -0 .4 5 - 0 .3 5 —0 .1 9
-9 t o 8 .5 1 .5 -4 .3 0 .0 5 -0 .3 8 -0.19- 0.40 8.5 to 13.1 4.3 -5.0 0 .3 8 -0 .2 9 0 .4 0 - 0 .3 0
TABLE 6 . GROUP I I — REGRESSION CURVES VERSUS SCATTER PLOTS
REGRESSION CURVE SCATTER PREDICTED VALUES PLOTS 1 2 3 4 D is ta n c e Mz v s . 2 t i n f t . I Mz in 0 2 I v a lu e o f 750 to 4350 -2.0 - -1.0 2 .3 - 3 .0 5 2 .3 - 2 .7 5
s k i Mz v s . S k j
value of Skj 750 to 2500 -2.0 - -1.4 0.42 - 0.34 0 .4 5 -0 .3 5 2500 to 4350 -1.4 - -1.0 0.34- -0.02 0 .3 5 -0 .1 5 61
from the scatter diagram. A like procedure was used in constructing
the remainder of Table 6 and also Table 7.
For the river deposits plus the sand, silt, and clay bar deposits (Group I), the predicted standard deviation with elevation above the bed is found to follow a sine curve. The undulations in the
curve agree closely with the undulations noted in the scatter diagram
of mean diameter versus standard deviation. The predicted skewness, however, only agrees with the scatter diagram for grain sizes of approximately 4.3 0 and finer.
For the gravel deposits (Group II), it is found that the pre dicted value of standard deviation with distance downstream agrees closely with the scatter diagram. For variation in skewness with distance downstream the predicted value is found to agree for mean diameters of -1.5 0 and coarser.
Discussion
Group JE
Transport of Group I sediments occurred as particles in con tinuous contact with the bed, in intermittent suspension, and/or in continuous suspension. From the grain size parameters of mean diameter, standard deviation, skewness, and kurtosis, it is not possible to dis tinguish the modes of transport. The predicted curves, however should yield the net effect of the transport mechanism.
In the predicted curves for mean diameter versus elevation for
Group I (Fig. 37), the decrease in grain size is a produce of decreased competency along the stream bed with elevation and a decrease in 62
suspended particle size with elevation. Suspended sediments of the
Colville River in 1962 were found to decrease in grain size with eleva
tion, however, not with distance downstream (Arnborg, et a l., 1967).
Information on bed load transport with elevation is not available.
The effects of elevation upon mean diameter are reflected in the variation of mean grain size with distance downstream (Fig. 41).
The parabolic shape of the predicted curve is a reflection of the change in elevation of the sediment sample collecting sites within the study area. The highest part of the curve is the finest predicted mean diameter and corresponds with the highest elevations on the longi tudinal bar.
In the columnar sections illustrating changes in sediments below the surface of the longitudinal bar (Fig. 36), it can be seen that the finest sediments are found at the highest elevations and that the coarsest are, in general, at the lowest elevations.
An analysis of specific differences between mean diameters of various environmental subareas reveals the influence of elevation on mean grain size. Comparisons between the longitudinal bar and the side bar reveal no significant difference in mean diameter of unvegetated sand, silt, and clay subareas. No significant differences are found between vegetated subareas. On the longitudinal bar and the side bar, however, a significant difference is found in the mean diameter of unvegetated sand, silt, and clay subareas and vegetated subareas. In both cases the coarsest sediment occurs in unvegetated subareas which correspond with lower elevations. When all sand, silt, and clay sedi ments of the longitudinal bar are compared with the side bar, it is 63 found that longitudinal bar sediments are somewhat coarser than side bar sediments. This finding is reflected in the parabolic curve of predicted mean diameter for distance downstream (Fig. 41). The upstream end of the curve, which is influenced by the samples from the longitudinal bar, predicts a larger mean diameter than the downstream end, which is influenced by the side bar.
Elevation above the bed appears to be the controlling factor in determining the site of deposition of sediments of specific grain sizes during the spring flood of 1971. A very close approximation of the predicted standard deviation with change in elevation (Fig. 38) can be derived from the scatter plot of mean diameter versus standard deviation (Fig. 45). The predicted curve for skewness versus elevation
(Fig. 40) can be partly derived from the scatter plot of mean diameter versus skewness (Fig. 46). In this case, however, only about half of the predicted skewness curve is found to agree with the values from the scatter plot. Because the predicted curve for kurtosis only accounts for 9 percent of the variance, it is difficult to evaluate the importance of the curve. Further difficulties in the analysis result from the complicated scatter diagram for mean diameter versus kurtosis
( F ig . 4 7 ) .
Group I I
For Group II, the predicted mean diameter of the gravels de creases with distance downstream (Fig. 44). No significant change, however, occurs with elevation. This may result from either the limited range of the gravels, 0-9 ft (0-2.7 m), or from the mode of transport. If the gravels were transported from an upstream source 64 during a brief surge in current velocity, it would be expected that rapid deceleration of current would occur in the area of the longitu dinal bar due to channel widening. This would result in a segregation of particles according to size but not necessarily with elevation. The predicted curve for standard deviation (Fig. 43) can be established from the scatter plot of mean diameter versus standard deviation (Fig.
45). For skewness it is possible to establish part of the predicted curve (Fig. 44) from the scatter plot of mean diameter versus skewness.
Effects of Ice Rafted Sediments
On June 6th, four days after breakup, a total of 57 blocks of river ice were observed on the longitudinal bar. These blocks grounded by the spring flood occurred at elevations of 11 to 13 ft (3.4 to
4 .0 m ).
A sufficient quantity of sediment for the collection of a sample was found on only 5 of the 57 blocks. At the site of each of these 5 blocks, paired samples were collected. A sample pair consisted of sedi ment from the ice block itself and sediment from the bar surface.
The cumulative probability curves shown in Figure 49 demonstrate the grouping of curves according to whether the sediment was rafted or nonrafted. The curves for the rafted sediments have a greater spread than those for the nonrafted sediments.
In order to determine what kind of differences existed in the
four grain size statistical parameters of the rafted and nonrafted
samples, a t test was employed. Results indicated that the mean dia meter of rafted samples (Mz^S^l 0) was significantly higher (p <0.0061)
than for nonrafted samples (Mz=2.87 0). For standard deviation, 65 skewness, and kurtosis, no significant differences were found. In order to determine if the amount of variability differed within each group, an F test for differences in variance was employed. This test indicated that mean diameter, standard deviation, and kurtosis
(respective F values of 37.89, 182.78 and 19.75 c 4, 4 d.f.) had higher variability within the rafted samples than within the nonrafted samples.
For skewness no significant difference was found (F=5.58 c 4, 4 d .f.).
From the cumulative curves and the statistical tests, it can be concluded that sediments transported by ice rafting onto the bar differ most from the bar surface sediments in their mean diameter.
Even though more variability is found in the ice rafted sediments, statistically no significant mean difference exists between the two groups in standard deviation, skewness, and kurtosis.
Sediment Thaw
From July 5th through 7th, depths to the permafrost table were determined at 50 fbot intervals along the profiles of the longitudinal bar (Fig. 6). At this time thaw depths on the bar ranged from a mini mum of 1.58 ft (0.5 m) to 4.25 ft (1.3 m). The data are complete for all areas with the exception of the gravels of subareas 3 and 4.
Because of the difficulties of probing in gravels, data were collected only at selected locations. Profiles with depth of thaw are presented in Appendix VII.
At the 74 sediment sample locations with known depths to the
frozen table, thaw was compared with the four grain size statistical parameters and elevation of the sample site. Table 7 gives the correla tion coefficients, probabilities, and number of samples for each 66
99.8 99.5 9 9 9 8
9 5
9 0
- 8 0 $ 7 0 S 6 0 5 0 lU > 4 0 5 3 0
0.5 — RAFTED SEDIMENTS NONRAFTED SEDIMENTS 0.1
0.01 1 J___ L i » i i J L L -4 .0 -2 .0 0.0 2.0 4.0 6 .0 8.0 10.0 GRAIN SIZE IN $
Figure 49. Cumulative probability curves of ice rafted sediments and adjacent nonrafted bar sediments. 67
TABLE 7 . ACTIVE LAYER CORRELATIONS
DEPTH OF THAW r = -0 .7 4 Mean p = <0.0001 D ia . N - 74 r » 0 .3 6 S ta n d a rd p =<0.0019 D e v ia tio n N a 74
i r - -0 .3 4 Skew ness p =<0.0036 N = 74 r =« -0 .3 6
K u rto s is p = < 0 . 0021 N = 74 r = 0 .0 4 E le v a tio n p =< 0.7398 N = 74
comparison.
Mean diameter was found to have a highly significant (p<.0001)
negative correlation (r=-.74) with thaw, indicating that as mean grain
size increases the depth of thaw increases. No relationship was found
in the comparison of thaw with elevation.
Due to the respective low r values of .36, -.34, and -.36 and
the large sample size, the apparent correlations between thaw and
standard deviation, skewness, and kurtosis as seen in Table 7 were
concluded to be unimportant.
Thaw depths on the longitudinal bar were further divided into
thosQ on west facing slopes and those on east facing slopes. An analysis of,variance using a completely randomized block design indi
cates that no significant difference existed in the direction of
exposure (F=.67 c 1, 72 d .f.). 68
On the longitudinal bar the' slopes are usually less than one degree. Therefore, the effectiveness of insolation is probably fairly evenly distributed resulting in similar depths of thaw on slopes of different exposure. This may also account for the lack of correlation between thaw and elevation.
With increased grain size the depth of thaw increases. This is readily apparent from the gravel subareas where the greatest thaw occurs.
For fluvial activity, the ultimate thaw depth is of no impor tance. The amount of thaw occurring prior to and during spring flooding is, however, significant. PROCESSES INTERPRETED FROM GRAIN SIZE DATA
Fluvial and aeolian processes are the dominant agents active within the study area. The duration, timing, and effectiveness of these
processes vary seasonally. Using grain size data derived from these deposits, the modes of sediment transport are considered (Appendix VI).
Modes of fluvial transport are wash load and bed load. Wash
load consists of particles transported in continuous suspension, whereas bed load consists of particles transported in contact with the bed and of particles placed into interm ittent suspension. Where possible hydrodynamic theory is used to determine the boundary between
these modes of transport.
Fluvial Processes Interpreted from C-M Diagram
An approximation of the fluvial processes active within environ mental subareas is derived from a plot (Fig. 50) of the coarsest one percentile versus mean diameter. For the Colville River data the diagram yields an S-shaped plot similar to that reported by Passega
(1964, p. 831) for the M ississippi River. From the Colville River plot, the fluvial processes operating within the study reach are interpreted as following:
1. . The gravels of subareas 3 and 4 represent transport as bed
load. The one sample lying outside of the zone formed by
samples from subareas 3 and 4 is due to the difficulty of
establishing the subarea boundaries which are often
transitional.
2. The river sediments of subareas 1 and 2 represent transport
69 70
-2 -S -4
-2
; ■** M i »
A9
Figure 50. C-M plot of sediment samples grouped according to environ- menta1 subareas.
as bed load and in continuous suspension.
3. The unvegetated sands, silts, and clays of the longi
tudinal bar, subarea 5, represent transport as bed load
and in continuous suspension.
4. Subareas 6 through 8 represent transport as continuous
suspension.
Between subareas some differences indicated by the plot are as follows: 71
1. The river samples (subareas 1 and 2) are similar to the
samples from subarea 5. However, the mean diameter of
subarea 5 is less than that of subareas 1 and 2.
2. The mean diameter of the unvegetated sands, silts, and
clays of the longitudinal bar (subarea 5) and of the side
bar (subarea 8) are sim ilar. However, the coarsest one
percentile of subarea 8 is finer than subarea 5. The
deposits of subarea 8 are prim arily derived from continu
ous suspension; whereas, for subarea 5, they are derived
from both continuous suspension and bed load.
Truncation Points and Hydrodynamic Principles
The significance of various truncation points between log
normal subpopulations (Appendix VI) can be determined by using hydro-
dynamic principles in conjunction with grain size distributions and
field data.
On the main channel of the Colville River approximately 0.6 mi
(1 km) south of the study area, Arnborg, Walker, and Peippo (1966, p.
197) report the hydrologic characteristics of the 1962 spring flood.
On June 14, 1962, when the maximum discharge of 212,370 ft^/s (6010 q m /s) occurred, it was accompanied by a maximum recorded current velocity of 2.3 m/s, an average current velocity of 1.5 m/s, and a maxi mum stage of 11.9 ft (3.6 m) above sea level. For the middle of June,
1962, the average water temperature was estimated from the graph of
Arnborg, Walker, and Peippo (1966, p. 205) to be 8°C. In the following discussion, these values are used in the calculations of the necessary 72 shear stress, shear velocity, and current velocity for sediment trans p o r t .
For each sample the truncation points were determined by f it ting straight lines to log-probability plots.
Figure 51 shows the reoccurrence of sim ilar points within sub- areas 3 and 4, the gravel subareas of the longitudinal bar. It can be seen that the truncation points fall into distinct groups with very little overlap. These observations are used to determine the average value for each grouping.
According to subarea (i.e ., subareas 1-9), Table 8 presents the average values for the truncation points between subpopulations.
This table also gives the number of observations used in determining those averages. From the table it appears that subareas which are environmentally similar have similar truncation points. This is illu s trated by subareas 7 and 9, the willow subareas, and also by subareas 1
u ir i n m in n m m i i i n u m in i in
0
O - “ ‘ - a — \r~ Figure 51. Truncation points occur / 0 - ring in the 35 samples z taken from subareas 3 w and 4. The truncation o points from one sample a . 0 - z can be established by a o vertical line connecting o< •A/ a tick mark at the top j . A • y J \ f . and bottom of the Hte -\ g ra p h . 3 *
S LJ-LX1 mill 111 t 111 73
TABLE 8 . TRUNCATION POINTS PER SUBAREA
S u b area 1 2 3 4 5 6 7 8 9 No. o f 9 4 30 5 40 10 28 7 2 Sam ples # T .P . # T .P . # T .P . # T .P . # T .P . # T .P . # T.P. # T.P. # T.P. 1 - 3 .0 1 -2 .6 30 - 3 .0 4 -3 .1 2 -2 .8 3 -1 .5 1 -1 .6 30 - 1 .4 4 -1 .5 7 -1 .8 7 0 .3 5 1 .0 3 0 .9 30 1 .1 5 1 .8 25 1 .4 6 1 .5 6 1 .0 4 0 .6 14 2 .3 2 2 .3 9 3 .1 3 3 .1 22 3 .1 2 3 .3 27 2 .9 4 3 .1 5 2 .8 10 4 .1 4 4 .2 1 4 .0 6 5 .1 3 5 .2 24 4 .8 5 5 .2 39 5 .0 10 5 .2 26 5 .8 4 5 .9 2 6 .0
# Number of times the truncation point was observed T.P. Average truncation point for subarea through 5, the unvegetated subareas. Subarea 8, the unvegetated sands, silts, and clays of the side bar, appears to be somewhat transitional.
Subarea 6, the grasses of the longitudinal bar, has greater sim ilarity to the unvegetated subareas. The truncation points may reflect varia tions in energy according to environment especially if they are the boundaries between subpopulations corresponding to modes of sediment t r a n s p o r t .
Table 9 presents the averages of groupings of truncation points which are found to reoccur in the 135 samples and summarizes their significance.
Aeolian Transport
Periodically throughout the year, aeolian transport occurs within the study area. A sample which can be related specifically to aeolian transport is lacking from the study area. However, during the spring of 1971 a sample of wind blown sediment was collected from a 74
TABLE 9 . AVERAGE TRUNCATION FOR ALL SAMPLES
N.O. T .P . POSSIBLE SIGNIFICANCE, FLUVIAL TRANSPORT
38 - 3 .0
45 - 1 .5 Interm ittent Suspension, Coarse Boundary, When Gravels were Transported
7 0 .3
84 1.2 Interm ittent Suspension, Coarse Boundary
16 2 .3
72 3.0 Continuous Suspension, Coarse Boundary, Subareas 1-6
15 4.0 Continuous Suspension, Coarse Boundary, Subareas 7,8, & 9
119 5 .0
Total No. of Samples 135 N.O. = Number of Times Truncation Pt. was Observed T.P. = Average Truncation Point snow d rift located near Putu, approximately 3.1 mi (5 km) downstream.
In Figure 52 the log-probability distribution is interpreted as two log-normal subpopulations truncated at 3.4 0. The coarser sub population comprises 98 percent of the sample. The results given agree favorably with the findings of Visher (1969, p. 1104) for coastal dune sands. He found that the truncation point between the saltating and suspended populations occurs at 3 to 4 0 with the saltating population composing 97 to 99 percent of the population.
Fluvial Transport of Sands, S ilts, and Clays
Using the average current velocity of 1.5 m/s, the shear velo city attained at maximum discharge is found via Keulagan's equation 75
9 9 8 9 9 9 9 9
Figure 52. Log-normal sub- populations and modes of sedi ment transport SO for aeolian sed im en t collected from a snowdrift at P u tu . AEOLIAN SEOIMENT
OS
4 0 ■2.0 00 2 0 4 0 6 0 SO 10.0 GRAIN SIZE IN $
(equation 6, Appendix VI) to be 6.67 cm/s for a kQ of 0.4 for clear flow,
and 5.27 cm/s for a kc of 0.2 for sediment laden flow (Table 10). The average shear velocity is 3.87 cm/s.
The calculated total bottom shear velocity attained is found
to be 5.75 cm/s (equation 8, Appendix VI). Ackers and White (1975, p.
625) report that the total shear velocity should average 7 percent more
than the shear velocity with respect to a grain. It appears that the
calculated shear velocity for clear flow and for sediment laden flow
are respectively too large and too small. The average shear velocity
of 5.27 cm/s, however, is 8.4 percent less than the total shear velo
city and is probably closer to being correct.
For bed load transport the truncation point between the subpopu
lation being moved in contact with the bed and the subpopulation being
transported in interm ittent suspension can now be predicted (equation 76
TABLE 10. CALCULATED SHEAR VELOCITIES AT SECTION 1 , 1962
SHEAR VELOCITY WHEN U =1.5m /s % DIFFERENCE BETWEEN k = 0 .4 k = 0 .2 AV. U.. . •“ o o t o t a l Uav.& Uto tal
6.67 3.87 5.27 5.75 8 .3 5
Calculated from Keulagan's equation
9, Appendix VI). The shear velocity is 5.27 cm/s. From Rouse's curve
( B l a t t , e t a l . , 1972, p. 54) it is found that the quartz particle with a settling velocity of 5.27 cm/s in water at 8°C has a diameter of
1.2 0 (.44 mm). From Table 9 it can be seen that a truncation point having an average value of 1.2 0 occurs in 84 samples.
This finding is further substantiated by the grain size distri butions. Curve A in Figure 53 is the log-probability plot of a com posite sample of suspended sediment deposited upon willow branches
C U R V E V 8 v »m r **4 SMIj m M From Willow* 99.5 C U R V E 'B " 3 « r f ld it O f L v n f IM In a l B ar 9 9 0 » p eilt« 4 ln 1971
9 0
Figure 53. Log-normal sub populations occur ring in suspended sediment from w illo w s (Curve A) O 10 and in surficial sediment of the longitudinal bar (C urve B ). 0.5
001 o.o ao io.o GRAIN SIZE IN 4 77
(Fig. 54) of subarea 7 during the spring flood of 1971. The sediment samples, which were collected from the willows at elevations of 2 in
(5.1 cm) to 18 in (45.7 cm) above the longitudinal bar surface, are composed of grains transported by both interm ittent suspension and con tinuous suspension. From Figure 53, curve A, it can be seen that no grains coarser than 1 0 occur in the sample.
The competency of the Colville River within the study reach can be determined when the shear velocity is 5.27 cm/s (Shields' curve and equation 3, Appendix VI). The maximum size of the particle for which the stream can initiate motion is found to be 0.9 0 (.54 mm).
An indication that the Colville River was capable of moving only particles of 0.9 0 and less may be seen on curve B, Figure 53.
Curve B represents the log-probability plot of sediment deposited in subarea 5 during the spring flood of 1971. The sample, collected from
1 in (2.5 cm) of sediment deposited upon a screen on the stoss side of the longitudinal bar (Figs. 34 & 35), consisted of a ripple marked layer of sand with a surface drape of silt and peat. The particles composing the sample are all less than 0.5 0 in diameter (i.e ., samples were sieved at 0.5 0 intervals).
Additional information on the mode of sediment transport for the subpopulations can be derived from curve A. The first truncation point occurs at 4 0. Arnborg, Walker, and Peippo (1967, p. 139) provide information on suspended sediment samples from the spring flood of 1962.
The samples collected approximately 0.6 mi (1 km) upstream (Section 1) from the study area are from within the river and represent sediment transported in continuous suspension. The reported sediment samples 78
Figure 54. Suspended sediments deposited on the branches of the taller willows (i.e., greater than 4 feet) of the longitudinal bar by the spring flooding. are all composed of particles less than 4 0 in diameter. It can,
therefore, be concluded that 4 0 represents the truncation point be
tween interm ittent suspension and continuous suspension.
Table 8 indicates that a truncation point occurs at approxi mately 4 0 in subareas 7, 8, and 9. However, in subareas 1 through 6 a truncation point occurs at approximately 3 0. This truncation point may represent either the boundary between saltation and suspension during aeolian transport or it may indicate greater energy within sub- areas 1 through 6 during fluvial transport. The higher energy would allow larger particles to be placed into suspension. Arnborg, Walker, and Peippo (1967, p. 140) found that in the Colville River the greater the discharge the larger the grain size of the suspended sediment. 79
Their measurements, however, were of sediment held in continuous suspen sion. Visher (1969, p. 1104) gives the generalization that the trunca tion point for continuous suspension in fluvial transport should occur between 2.75 and 3.5 0. The cross-section of the longitudinal bar illustrated in Figure 36 reveals that within subareas 5 and 6 the general tendency is for the surficial deposits to consist of a sand layer with silt above. This would appear to indicate that the dominant depositional process is fluvial. The truncation point at 3 0 probably does represent the boundary between interm ittent suspension and con tinuous suspension during fluvial transport.
The boundary of interm ittent suspension and continuous suspen sion appears to vary according to elevation above the bed and exposure to the prevailing current during spring flooding. The environmental subareas of 1 through 6 correspond with the lower elevations on the longitudinal bar and have a truncation point of 3 0. Environmental subarea 7 corresponds to the highest elevations on the longitudinal bar and has a truncation point of 4 0. On the side bar, environmental subareas 8 and 9, which are somewhat protected during spring flooding by the longitudinal bar, have a truncation point of 4 0.
The truncation points listed are gathered from samples obtained by penetrating the surface to a depth of 4 in (10.2 cm). These samples are not of a single sedimentary unit but rather from multiple units.
Some of these samples may contain sediments deposited over 10 or more years. The reoccurrence of truncation points of approximately the same value probably indicates that the hydrodynamic conditions of the
Colville River do not fluctuate appreciably from year to year. 80
Fluvial Transport of Gravels
Table 11 presents the shear stresses and shear velocities neces sary to initiate movement of the coarsest particles occurring in subareas 3 and 4 . The shear stress and shear velocities are given for
- 3 .0 0, the average of the first truncation point in the grain size distributions of subareas 3 and 4 ; - 4 .3 0, the average value of the coarsest one percentile; and - 4 .5 0, the maximum grain size occurring in subareas 3 and 4 . The calculated shear velocities range from 8 .8 cm/s for - 3 .0 0 to 1 4 .8 cm/s for -4 .5 0. As can be seen from Table 10 the calculated shear velocities attained under the conditions of 1962 were insufficient to initiate motion of the gravels. The necessary shear velocities for movement, however, are not exceptional. Shear velocities in excess of 14 cm/s are reported in the Mississippi River at St. Louis when the discharge exceeds 2 2 ,6 4 0 m^/s and for the Otowi reach of the Rio Grande when the discharge exceeds 85 nrfys (Middleton,
1 9 7 6 ).
Keulagan's equation (equation 6, Appendix VI) is used to calcu late the stream velocities necessary to attain the required shear velocities for the initiation of particle motion for -3.0 0, -4 .3 0, and - 4 .5 0. For each particle size the current velocities are given for a kQ of 0.2 and 0.4. The two current velocities are viewed as the maximum and minimum necessary for particle motion.
The gravels of subareas 3 and 4 are either in place or have been transported to their present position from upstream. The lack of gravels outcropping along the bluffs of the study area appears to indi cate an upstream source. However, gravels do occur at the entrance to 81
TABLE 11. SHEAR STRESS AND SHEAR VELOCITY FOR GRAVEL MOVEMENT
SHEAR SHEAR STRESS VELOCITY a**o u* mm 0 d y n es cm /s
MAXIMUM GRAIN SIZE 22.6 -4 .5 220 14 .8
COARSEST ONE PERCENTILE 19.7 - 4 .3 191 13 .8
FIRST TRUNCATION 8.0 - 3 .0 78 8.8 POINT
Calculated from Shields' diagram the West Channel (Fig. 3). If the gravels have been derived from an upstream source, it is necessary to move the gravels through Section 1, approximately 0.6 mi (1 km) upstream from the study area. Table 12 gives the current velocities necessary to move the gravels through
Section 1 in 1962 and 1971. Current velocities are also given for the downstream movement of the gravels from their present location at sub- area 3. At Section 1, a higher velocity was required for 1971 than was necessary in 1962 due to the higher stage attained in 1971 (Fig. 4).
For each of the vertical columns of Table 12, the current velo cities given for the two stages at Section 1 and at subarea 3 are so close that probably no valid difference exists between them. Keulagan's equation is an empirical relationship and the average current velo cities for a column are probably as close to being correct as any single value in the column.
The average current velocity of 1.5 m/s reported for 1962
(Arnborg, et a l., 1966, p. 205) at 0.6 mi (1 km) upstream from the study 82
TABLE 1 2 . AVERAGE CURRENT VELOCITIES FOR GRAVEL MOVEMENT
U cm /s
-3 0 - 4 .3 0 -4 .5 0 MEAN U*= 8.8 U*=13.8 U*=14.8 DEPTH m ko=0.4 ko=0.2 ko=0.4 ko=0.2 ko=0.4 ko=0.2
SECTION 1 1962 4 .2 1 198 341 311 536 333 574
NEAR SECTION 1 1971 7 .1 1 209 364 329 572 353 613
ACROSS SUBAREA 3 5 .5 0 204 353 320 554 343 594
AV. 204 353 320 554 343 594
Calc, from Keulagan's equation using a water temperature of 8 °C .
reach is insufficient to move the gravels of subareas 3 and 4. The
maximum reported current velocity of 2.3 m/s, however, is sufficient
to move particles of -3.0 0 under conditions of clear flow but not
sufficient to move larger particles. These results indicate that the
current velocities need not be increased greatly to move gravels of
- 3 .0 0 .
In Table 13 the shear velocities necessary to initiate parti
cle motion for grain sizes of -3.0 0, -4.3 0, and -4.5 0 are presented.
The table also gives the grain size of the particle whose settling velocity is numerically equal to the shear velocity. From the rela
tionship given for particle suspension (equation 9, Appendix VI), it
can be concluded that when the shear velocity is 8.8 cm/s, particles having a diameter of -3.0 0 w ill begin to move and particles having a 83
TABLE 1 3 . GRAIN SIZES WHEN U* = uu
- DIA. OF PARTICLE 1 FOR WHICH U -fo s u l DIA. OF PARTICLE U* WILL cm /s WHOSE INITIATE MOTION SETTLING VEL. =ui
mm 0 mm 0
minimum U* (k Q= 0 .2 ) 3.87 0.35 1.5 r-H g a v e ra g e U* 0 .5 4 0 .9 5 .2 7 0 .4 3 1.2 O 04 H VO total bottom shear vel. 5.75 0.48 1.0 Wes CO maximum U* (k Q= 0 .4 ) 6 .6 7 0 .5 4 0 .9
first truncation point CO 8 .0 - 3 .0 8.8 0.68 0.6 2 ^Diameter determined at 8 °C from Rouse's curve as presented in Blatt, et al., 1972, p. 54). diameter of -0.6 0 will be placed into interm ittent suspension. For particles of -4.3 0 and -4.5 0 the respective grain sizes which w ill be placed into interm ittent suspension are 0.2 0 and -0.1 0. By referring to T ab le 8 it can be seen that in subarea 5 a truncation point aver aging 0.3 0 occurs in seven samples. This is the only indication of a truncation point having a value close to that predicted from equation 9 of Appendix VI. In subareas 1 through 5 a truncation point, observed in 45 samples, occurs at -1.5 0. This truncation point represents a grain size greater than the competency of the Colville River as determined for 1962 and 1971. Visher (1969, p. 1104) reports that, in general, for fluvial transport a truncation point occurs at -1.5 0 to -1.0 0 marking 84 the boundary between bed load and transport by interm ittent suspension. When shear velocities within the study reach of the Colville River attain a magnitude great enough to place particles of - 1.5 0 in t o interm ittent suspension, the competency of the river is sufficient to move even the coarsest particles occurring in the gravel regions, sub- areas 3 and 4, of the longitudinal bar. ORIGIN AND DEVELOPMENT OF THE BARS The physiography, hydrodynamics, and sediments of the Colville River hold the key to the origin and development of the side and longitudinal bars. The hydrodynamics of the Colville River and its sediments are important indicators of present and past conditions. During the spring floods of 1967 and 1971, the Colville River was incapable of moving the gravels of the longitudinal bar. This could indicate either that the gravels originated from erosion of the bluffs at the side of the longitudinal bar with removal of the fines by fluvial action or that the gravels have been transported from an up stream source to their present location during a time of high stream velocities. The former, however, appears somewhat unlikely due to three f a c t o r s : ( 1) no outcroppings of gravels were observed in the bluff adjacent to the site of the longitudinal bar, ( 2) the streamlined shape of the linguoid shaped gravel ridge probably indicates that it was formed during movement of the gravels, and (3) the presence in the grain size distributions of a truncation point at -1.5 0 probably represents the coarse boundary of bed load transport at time of gravel movement. The Colville River flows from a relatively constricted channel at Section 1 (Fig. 2) into a much wider channel in the study area. This channel widening occurs abruptly just upstream from the longitudinal bar. As the river flows from the constricted channel of Section 1 into the wider channel downstream, a decrease in stream velocity is expected. During spring flooding in the arctic, river channels may become restricted by ice jams. The effect of an ice jam is initially felt 85 86 upstream where an increase in river stage above that of normal may occur. In the Tana River of Norway ice jams are known to increase river stage by several inches (Collinson, 1971, p. 559). In the Yukon River it has been found that in every five to fifteen years an ice jam occurs during spring flooding which may increase the stage locally by 40 ft (12 m) or more (Eardly, 1938, p. 351). During the spring flood of 1966 on the Colville River, an ice jam formed which caused a locally higher than normal stage and a localized increase in erosion near the ice jam (Walker and McCloy, 1969, p. 72). At the site of blockage by an ice jam increased velocities around the restriction may cause erosion and transport of gravels the size of those found on the longitudinal bar. At the time of ice jam breakup current velocities w ill cause transport of the gravels. Thus if an ice jam forms in the vicinity of Section 1, gravels w ill even tually be transported downstream. With the channel widened at the present site of the longitudinal bar, current velocities would decrease rapidly resulting in deposition of the gravels. The gravels form the nucleus for bar development. This initial state of the Colville River longitudinal bar is sim ilar to that described for braid bar development by Leopold and Wolman (1957). When the gravels were initially deposited, the right river bluff in plan view must have had a gently curving convex face toward the river channel. After deposition of the gravels the Colville River was forced to flow around the obstruction. In this manner the side channel was established. Stream flow through the side channel now eroded the adjoining bluff and altered its shape from that of convex to concave ( F ig . 3 ) . 87 Progressing downstream, the topographic expression of the gravels after their deposition may have been that of a gravel sheet rising toward the north succeeded by a topographic low which culmi nated in the linguoid shaped gravel ridge at the north end of the deposit. This gravel deposit formed the foundation upon which the longitudinal bar was built. During each succeeding spring*flood sedi ments were deposited in the topographic lows on the lee of the gravel sheet (subarea 3) and the linguoid shaped gravel ridge (subarea 4). At the south end of the longitudinal bar the gravel sheet (subarea 3) acted as a rampart that protected the downstream area which increased in elevation after each spring flood. Grass and willows became estab lished at higher elevation and aided in trapping suspended sediments which further increased the elevation. Also aiding in increasing the elevation was the occasional formation of ice-shove ridges during breakup. In this manner the longitudinal bar grew to an elevation which exceeded the maximum stage occurring during the observed breakup o f 1971. When the longitudinal bar gained sufficient elevation, its presence aided in protecting the portion of the right bank downstream adjacent to the present side bar. In this protected area the side bar fo rm ed . SUMMARY AND CONCLUSIONS Fluvial and aeolian processes are responsible for the develop ment and modification of the morphology of the bars. Aeolian activity is likely to occur at any time of the year. The dominant process, how ever, is fluvial with its activity confined to the relatively short period of spring flooding. Modes of sediment transport are determined by using the grain size distributions from fluvial deposits. Presented in the following summary are conclusions regarding morphologic changes centered around spring flooding, factors influ encing active layer development, significance of grain size statistical parameters, effects of ice rafted sediments, and fluvial processes. Morphologic Changes (1) The effectiveness of fluvial and aeolian processes in the modification of bar sediments is dependent upon the distribution, thickness, and duration of snow cover. As long as a snow cover is present, both fluvial and aeolian processes have little ability to modify bar surfaces. (2) In the spring, meltwater derived from the snow cover becomes an active agent for the downslope movement of exposed, thawed sediments. These sediments, which are easily removed during spring flooding, become part of the sediment load of the Colville River. In this manner the longitudinal bar contributes sediment to the river. (3) Active layer development prior to and during spring flooding facilitates the modification of bar surface 88 89 sediments. The ultimate thickness achieved by the thawed layer at the end of the summer has little to no effect on the morphology of the bar. (4) During spring flooding, the longitudinal bar is almost entirely submerged. At this time the bar becomes primarily an area of deposition. The depositional sequence observed is a unit composed of two layers: a lower thicker layer of coarser sediment followed upward by a thinner layer of finer sediment. (5) Ice shove of sediment during spring flooding is responsible for significant increases in the elevation of the longitudinal bar. It appears that very few ice shove ridges of sediment form during a spring flood. Once they have originated, their likelihood of remain ing as prominant topographic features of the bar is g r e a t . (6) As the spring flood subsides, fluvial morphologic change is confined primarily to the strand lines corresponding to fluctuations in stage. Aeolian processes now become important agents modifying portions of the longitudinal bar surface. The dominant physiographic features on the bars, however, are those formed by fluvial activity and, in particular, those formed during the spring flood. Depth of Thaw (1) The depth of thaw of the bar sediments does not correlate with direction of slope exposure. The slopes on the 90 longitudinal bar are, in general, less than one degree. The difference of insolation on such gentle slopes is probably insignificant. (2) The depth of thaw of the bar sediments does correlate with mean diameter. It does not, however, correlate with e l e v a ti o n . Significance of Grain Size S tatistical Parameters (1) The predicted mean diameter curve for the river plus nongravel bar deposits (Group I) reflects the conditions of the spring flood of 1971. The predicted mean dia meter curve for the gravel deposits of the gravel sheet and the linguoid gravel ridge (Group II) reflects conditions occurring at some time in the past. (2) For the river plus nongravel bar deposits (Group I), mean diameter decreases with elevation above the bed. The relationship is linear. Also, the differences and sim ilarities in the grain size statistical parameters between environmental subareas are a reflection of the elevation of the subarea and the predicted mean diameter which w ill occur at that elevation. Effects of Ice Rafted Sediments (1) On the surface of the bars, abrupt changes in the sedi ment are possible due to deposition of ice rafted sediments. If the results obtained in 1971 are typical, the variations are most likely to occur at or between 91 elevations of 11 and 13 ft (3.4 and 4.0 m). The statis tical parameters of standard deviation, skewness, and kurtosis do not differ significantly between the rafted and nonrafted sediments. The mean diameter of the rafted sediments is, however, found to be significantly coarser than for the nonrafted sediments. The effect of ice rafting is therefore reflected in the occurrence of areas of greater than normal mean grain size. Fluvial Processes (1) The dominant fluvial process active in each environmental subarea is revealed by the C-M plot. However, a more complete determination of fluvial processes is obtainable from a study of subpopulations in log-probability plots of grain size distributions. Further, studies of sub populations grouped according to environmental subareas reveal differences between subareas which are not obtain able from comparisons between grain size statistical parameters of environmental subareas. (2) From a hydrodynamic consideration and a study of the points of truncation between subpopulations of grain size distributions, it is apparent that the portion of the stream load being transported as bed load and wash load can be determined specifically. For 1971, the specific bound ary between transport as bed load in contact with the bed and as interm ittent suspension occurred at 1.2 0. The m ax i mum grain size which the river was capable of transporting 92 was 0.9 0. Particles less than 3 0 are held in continu ous suspension in areas of high energy. In areas of less energy, particles less than 4 0 are held in continuous suspension. Particles between 1.2 0 and either 3 0 or 4 0 (depending upon the environment) are transported in interm ittent suspension. (3) The phi boundaries between the portions of the sediment load being transported as bed load, interm ittent suspension,and continuous suspension are found to have approximately the same value whether the sediment sample was taken from sediment deposited during the spring flood of 1971 or whether the sediment sample was taken from laminae representing deposition over many years. From this, it is concluded that the flow regime from year to year has not fluctuated appreciably from that of the present. (4) In areas containing grain sizes which surpass the present compentency of the river, subpopulations are encountered which (a) have corresponding boundaries with subpopulations of grain size distributions of finer sedi ments and (b) have subpopulations occurring in the coarser parts of the distribution (i.e., the part lacking in the finer areas). It is believed that these truncation points in coarser sediments are a legacy from the time of trans port of the coarser materials. (5) The general characteristics of the sediment samples 93 reflect the Influence of sedimentation during the spring flood. With emergence of the bars, modification by other agents (e.g., aeolian) occurs in the morphology of the bars. However, when the characteristics of all the samples are averaged, the differences introduced by these other agents disappear. Thus, the major deter minant of the characteristics of the bar sediments appears to be fluvial activity during spring flooding, a process which embraces as little as 4 percent of an entire year and as little as 13 percent of the period of active discharge. (6) The field data and the hydrodynamic analysis indicate that the Colville River was incapable of moving the gravels of the longitudinal bar either from their present location in subareas 3 and 4 or from the location of Section 1, a probable source. The truncation point of -1.5 0 found in the gravel deposits may be a legacy from the time of gravel transports and may represent the boundary between transport as bed load and as interm ittent suspension. (7) From the hydrodynamic analysis, it is apparent that the maximum recorded stream velocity at Section 1 in 1962 was sufficient to move the gravels of -3.0 0. However, the velocity was incapable of moving the coarser gravels which have a maximum size of -4.5 0. From these findings, it is concluded that only a slight increase in stream 94 velocity is necessary in order to move the finer gravels (i.e., -3.0 0). Much greater increases in velocity are necessary to move the coarser gravels. It is concluded that the presence of the gravels on the longitu dinal bar is primarily responsible for maintaining the stability of the bar. Further, it is proposed that these gravels were transported to their present location on the bar from an upstream source during a major ice jam breakup which offered a time of high stream velocity. Thus, it is concluded that the deposition of these gravels created the nucleus for the development of the bar. REFERENCES Ackers, P., and White, W. R., 1973, Sediment transport; New approach and analysis; Jour, of the Hydraulics Division, Am. Soc. of Civil Engineers, v. 99, HY 11, p. 2041-2060. 1975, Sediment transport; New approach and analysis, Discussion: Jour, of the Hydraulics Division, Am. Soc. of Civil Engineers, v 101, HY 5, p. 621-625. Arnborg, L ., Walker, H. J ., and Peippo, J., 1966, Water discharge in the Colville River, 1962; Geografiska Annaler, v. 48 A, p. 195-210. 1967, Suspended load in the Colville River, Alaska, 1962; Geografiska Annaler, v. 49 A, p. 131-144. Barr, A. J., and Goodnight, J. H., 1972, Statistical analysis system; Dept, of S tatistics, North Carolina State University. 260 p. B latt, H., Middleton, G., and Murray, R., 1972, Origin of sedimentary rocks; Englewood C liffs, N .J., Prentice-Hall Inc., 634 p. Bluck, B. J., 1974, Structure and directional properties of some valley sandur deposits in southern Iceland: Sedimentology, v. 21, p. 533-554. Brice, J. C., 1964, Channel patterns and terraces of the Loup River in Nebraska: U.S. Geol. Survey Prof. Paper 422-D, 41 p. Briggs, L., and Middleton, G., 1965, Hydromechanical principles of sediment structure formation, in Middleton, G., ed., Primary sedimentary structures and their hydrodynamic interpretation: Soc. of Econ. Paleontologist and M ineralogists, Spec. Pub. 12, p. 5-16. Chow, V., 1959, Open-channel hydraulics: New York, McGraw-Hill Book Co., I n c . , 680 p . Collinson, J. D., 1970, Bed forms of the Tana River, Norway: Geografiska Annaler, v. 52A, p. 31-56. 1971, Some effects of ice on a river bed: Jour. Sed. Petrology, v. 41, p. 557-564. Eardley, A. J ., 1938, Yukon channel shifting: Geol. Soc. America B ull., v. 49, p. 343-358. Einstein, H., 1950, The bed-load function for sediment transportation in open channel flows: U.S. Dept, of Agriculture, Soil Conservation Service, Tech. Bull. 1026, 71 p. 95 96 Folk, R. L., 1968, Petrology of sedimentary rocks: Austin, Texas, Hemphill's, 170 p. Folk, R. L., and Ward, W. C., 1957, Brazos River bar: A study in the significance of grain size parameters: Jour. Sed. Petrology, v. 27, p. 3-26. Galehouse, J. S., 1971, Sedimentation analysis, in Carver, R. E., ed., Procedures in sedimentary petrology: New York, W iley-Interscience, p . 6 9 -9 4 . Gray, D., and Wigman, J., 1973, Peak flow - Rainfall events, in Gray, D., ed., Handbook on the principles of hydrology: Port-Washington, N. Y., Water Information Center, Inc., p. 8.1-8.102. Ingram, R. L., 1971, Sieve analysis, in Carver, R. E., ed., Procedures in sedimentary petrology: New York, W iley-Interscience, p . 4 9 -6 7 . Jaquet, J. M., and Vernet, J. P., 1976, Moment and graphic size parameters in the sediments of Lake Geneva (Switzerland): Jour, of Sed. Petrology, v. 46, p. 305-312. Jordan, P. R., 1965, Fluvial sediments of the M ississippi River at St. Louis, Missouri: U.S. Geol. Survey Water-Supply Paper 1802, 89 p. Krigstrom, A., 1962, Geomorphological studies of sandur plains and their braided rivers in Iceland: Geografiska Annaler, v. 44, 328-346. Leliavsky, S., 1966, An introduction to fluvial hydraulics: New York, N.Y., Dover Publications, Inc., 257 p. Leopold, L. and Wolman, M. G., 1957, River channel patterns: braided, meandering, and straight; U.S. Geological Survey Professional Paper 282-B, p. 39-73. Leopold, L., Wolman, M., and M iller, J., 1964, Fluvial processes in geomorphology: San Francisco, C alif., W. H. Freeman and Co., 522 p. Li, J. C. R., 1964, Statistical inference I: Ann Arbor, Michigan, Edwards Brothers, Inc., 658 p. McKenzie, L. S., and Walker, H. J ., 1974, Morphology of an Arctic river bar: Coastal Studies Institute, Louisiana State Univ., Technical Report No. 172, 32 p. McKinney, T., and Friedman, G., 1970, Continental shelf sediments of Long Island, New York: Jour, of Sed. Petrology, v. 40, p. 213-248. 97 Middleton, G., 1976, Hydraulic interpretation of sand size distribu tions: Jour, of Geology, v. 84, p. 405-426. Ore, H. T., 1964, Some'critera for recognition of braided stream deposits: Contributions to Geology, Univ. of Wyoming, v. 3, p. 1-14. Fassega, R., 1957, Texture as characteristic of clastic deposition: Am. Assoc. Petroleum Geologists B ull., v. 41, p. 1952-1984. 1964, Grain size representation by CM patterns as a geological tool: Jour, of Sedimentary Petrology, v. 34, p. 830-847. Searby, H. W., and Branton, C. I., 1975, Climate conditions in agricultural areas in Alaska, in Weller, G., and Bowling, S. A., eds., Climate of the arctic: Fairbanks, Alaska, Geophysical Institute, Univ. of Alaska, p. 281-292. Shen, H., 1975, Hans A. Einstein's contributions in sedimentation: Jour, of the Hydraulics Division, Am. Soc. of Civil Engineers, v. 101, HY 5, p. 469-488. Short, A. D., 1973, Beach dynamics and nearshore morphology of the Alaskan arctic coast (Ph.D. dissert.): Baton Rouge, Louisiana State Univ., 139 p. Smith, N. D., 1974, Sedimentology and bar formation in the upper Kicking Horse River, a braided outwash stream: Jour, of Geology, v. 82, p. 205-223. Task Committee on Preparation of Sedimentation Manual, Vanoni, V., Chairman, 1966, Sediment transportation mechanics: initiation of motion: Jour, of the Hydraulics Division, Am. Soc. of Civil Engineers, v. 92, HY 2, p. 291-314. Task Force on‘Bed Forms in A lluvial Channels, Kennedy, J. F., Chairman, 1966, Nomenclature for bed forms in alluvial channels: Jour, of the Hydraulics Division, Am. Soc. of Civil Engineers, v. 92, HY 3, p. 51-64. Wahrhaftig, C., 1965, Physiographic divisions of Alaska; U.S. Geological Survey Professional Paper 482, 52 p. Walker, H. J., 1973a, The nature of seawater-freshwater interface during breakup in the Colville River delta, Alaska in Permafrost: The North American contribution to the second international conference: Washington, D. C., National Academy of Science, p. 473-476. 1973b, Spring discharge of an arctic river determined from salinity measurements beneath sea ice; American Geophysical Union, Water Resources Research, v. 9, p. 474-480. 98 Walker, H. J., 1976, Depositional environments in the Colville River delta, in M iller, T. P., ed., Recent and ancient sedimentary environments in Alaska: Alaskan Geological Soc., p. C1-C22. Walker, H. J ., and McCloy, J. M., 1969, Morphologic change in two arctic deltas: Arctic Institute of North America, Research Paper No. 49, 91 p. Weinberg, G. H., and Schumaker, J. A., 1967, S tatistics, An intuitive approach: Belmont, California, Wadsworth Publishing Co., Inc., 338 p . ■ Visher, G. S., 1965, Fluvial processes as interpreted from ancient and recent fluvial deposits, in Middleton, G., ed., Primary sedimentary structures and their hydrodynamic interpretation: Soc. of Econ. Paleontologist and M ineralogists, Spec. Pub. 12, p. 116-132. 1969, Grain size distributions and depositional processes: Jour. Sed. Petrology, v. 39, p. 1074-1106. APPENDICES 100 APPENDIX I DERIVATION OF FLUVIAL BAR TERMINOLOGY A lluvial Bar Terminology One attempt at organizing the terminology of bed forms in alluvial channels was made by the Society of Civil Engineers (Task force on bed forms in alluvial channels, 1966). An examination of this report and a study of the nomenclature presented revealed that adequate terms could not be found to apply to the two bars studied in the Colville River. A more extensive review of the literature revealed t h a t : ( 1) terms would have to be adopted from more than one author and (2) somewhat sim ilar features have been described by various workers without the use of a uniform terminology. Ore (1964, p. 1) defines a longitudinal bar as one which, "has its long axis essentially aligned with stream flow." Bars of other workers which can meet the requirements of this definition are: ( 1) th e spool bar of Kringstrom (1962), (2) the islands considered by Collinson (1970), and (3) the medial bars studied by Bluck (1974). Likewise, when a side bar is designated as one which is attached to the bank (Collinson, 1970), the point bar of Smith (1974) and the lateral bar of Bluck (1974) meet the qualifications of this definition. The bar nomenclature used herein and the associated terms of other workers are, however, not necessarily synonymous. Therefore, the alternate terms discussed below are not employed in this study. For example, the spool bar of Krigstorm (1962) meets the requirements for a longitudinal bar only at certain stages of development. Collinson 101 (1970) in his study of the Tana River uses the term islands for physiog raphy similar to the longitudinal bar of the Colville River. However, it is felt that the word island should be restricted to physiographic features having sufficient topographic height so that they are not subjected to major submergence during flooding. Brice (1964) implies this meaning in presenting his stabilized braid index: B.I. = 2 [sum of lengths of islands and (or) bars in reach] 7 length of reach measured between banks. By definition the stabilized braid index remains constant with varying river stage. The medial bar of Bluck (1974) conveys the impression of being located in mid-channel. Because the longitudinal bar of this study is positioned towards one side of the river channel, the term medial is not appropriate. The side bar of the Colville River could be classified as a point bar as used by Smith (1974, p. 210). Smith's point bar is a gravel deposit, "formed in gently curving channels, commonly separated from the inner convex bank by a smaller channel." At a time when the point bar is not separated from the bank it would be referred to as a side bar. The term point bar is not used in this report in order to avoid indirectly implying ( 1) a meandering stream pattern or ( 2) th e presence of ridge and swale topography. Another possible tern for side bar is Bluck's (1974, p. 541) lateral bar, defined as "units of sediment attached (or about to be) to a bank." The physiographic expression of Bluck's lateral bar is that of bar and swale topography (Bluck, 1974, p. 541). For this reason the term lateral bar is not u s e d . 102 APPENDIX I I FIELD PROCEDURES Studies Prior to Spring Flooding Snow depths and subsequent depths of thaw were determined by using a metal rod marked at 6 in (15.2 cm) intervals. The rod was inserted into either the snow cover or thawed sediments at specified intervals along traverses established with a Brunton compass. Prior to the spring flood, at selected locations on the longi tudinal bar, screening was placed on the bar surface and held in place with one foot iron spikes driven into the bar sediments. The objec tive of the screening was to ( 1) determine areas of sedimentation and/or erosion and ( 2) differentiate sediments deposited by the spring flood of 1971 from sediments deposited during previous floods. Post Breakup Studies Sediment Sample Collecting Beginning in the latter part of June and continuing through July, the surficial sediments of the longitudinal and side bars were sampled. The samples were obtained by inserting a plastic tube of 2.5 in (6.4 cm) in diameter to a depth of 4 in ( 10.2 cm) into the surface of the bar. The tube was then extracted with its sediment content. The sedi ment was removed from the tube, placed into a plastic bag, sealed, labeled,and shipped for analysis to Louisiana State University. The surficial sediment samples were collected at 100 ft (30.5 m) intervals along six traverses established with the use of a 103 Brunton compass and tape on the longitudinal bar (Fig. 6). One traverse of 5200 ft (1585 m) was positioned along the length of the bar. Beginning at 900 ft (274 m) from the south end of the bar, and at each succeeding 900 ft (274 m) interval, a total of five traverses were established across the width of the bar. On the side bar surficial sediments were collected at the following locations: (1) one sample was collected at the southern extremity of the bar and another at the northern extremity, (2) midway along the length of the bar, three samples were collected across the width of the bar, one at the water line, one at midway along the width, and one at the base of the river bluff, and (3) at one-quarter and three-quarters of the length of the bar, two samples were collected at each location along a traverse across the width of the bar, one at the water line, and one at the base of the river bluff. Grab samples of the river bottom sediments were collected from both the main channel and the side channel in mid-July. Bottom sedi ment samples were collected across the width of the channels at each of the following locations: (1) at the upstream and downstream extrem ity of the study area, (2) at the upstream and downstream extremity of the longitudinal bar, (3) at one-third and two-thirds of the length of the longitudinal bar, and (4) at half the distance.between the down stream end of the longitudinal bar and the downstream end of the study a r e a . Sediment Characteristics Below the Surface of the Bars Outside of the gravel areas of the longitudinal bar at each sediment sample collecting site, a hole was excavated to frozen ground. 104 The changes In sediments were visually observed, measured, and recorded. In gravel areas, holes were excavated only at selected loca tions. The total depth of the hole In most cases was determined by the depth at which ground water was encountered. The metal probe was then inserted until penetration was no longer possible. This depth was assumed to be frozen ground. Topographic Survey and Map Construction Throughout the month of July the longitudinal bar and the side bar were surveyed by using a theodolite. Depths in the.river channels were determined by a fathometer. Maps were constructed from the survey and from the aid of low elevation aerial photographs of the study area taken during the 1971 field season. 105 APPENDIX I I I LABORATORY PROCEDURES Analysis of Sediment Samples Sediment samples were air dried for several days and crumbled by hand. After the sample was dried it was weighed. By passing the sediment through a >1 0 screen, gravels were removed from appropriate samples and the weight of the gravel fraction was determined. The gravels were then sieved at 0.5 0 intervals. The fraction, composed of less-than-gravel-size sediments, was split into smaller samples with a sample splitter. The size of the samples depended upon the characteristic sizes of the sediments. For samples of primarily sands, a sample weight of approximately 100 grams was chosen. For finer sediments smaller samples were obtained. The objective was to derive a sample for pipette analysis of approximately 15 g ram s. After the samples had been split, organics (i.e., primarily peats) were partly removed by using a modification of the procedure developed by Jackson, et a l., as presented by Ingram (1971, p. 58). To a known weight of sample in a 400 ml beaker, 30 percent H 2O2 was added until frothing stopped. The sample was heated to 40°C for 10 minutes and allowed to evaporate to a paste. To the beaker, 30 ml of 30 per c e n t H2O2 was added. The beaker was then covered by a watch glass and allowed to digest for several days. The contents were then brought to a slow boil to drive off excess H202* These treated samples were oven- dried at less than 40°C and then were allowed to stand in the lab for 106 several hours in order to achieve room temperature and approach room humidity. The weight of the sample was then determined. The aforementioned method achieves partial removal of organics and aids in dispersing clays. Its main advantage is that it removes organics without leaving residue which might affect the sample weight. The treated samples were wet sieved through a 4 0 screen using deionized water. The sediments which passes through the screen were placed into a 1000 ml cylinder. The sediments retained on the screen were dried, weighed, and sieved at 0.5 0 intervals in a Ro-Tap for 10 minutes (Ingram, 1971, p. 64). The less than 4 0 fraction from the sieve analysis was then added to the 1000 ml cylinder. To the contents of the 1000 ml cylinder, 50 ml of a 10 percent Calgon solution was added in order to disperse the sediments. The water level was brought to exactly 1000 ml by adding deionized water. The contents of the cylinder were then stirred, allowed to stand over night, and checked the next day to see if flocculation had occurred. For the appropriate room temperature, pipette analyses were run using.the time schedule and depths of withdrawal given by Galehouse (1971, p. 80). Beaker weights were determined by using an electrical b a la n c e . 107 APPENDIX IV STATISTICAL METHODS Post anova tests in the form of orthogonal comparisons were used to determine specific subarea (i.e., subareas 1, 2, 3, etc.) differences. M ultipliers for the orthogonal comparisons were weighted (Li, 1964, p. 410-415) to adjust for unequal sample size. For the four grain size statistical parameters, mean diameter, standard deviation, skewness, and kurtosis, a stepwise regression pro cedure (Barr and Goodnight, 1972) was used to determine the model of best fit with respect to distance downstream and elevation above sea level. The model containing a minimum number of independent variables while maximizing the r^ was chosen. Linear, cubic, and quadratic com ponents were included in the procedure. Based on these models' of the existent data, predicted values were obtained. In addition 95% confi dence intervals about the mean were determined. These procedures were performed for Group I (river and all nongravel deposits, subareas 1, 2, 5, 6, 7 , 8 , and 9) and Group II (gravel deposits, subareas 3 & 4) separately due to field data indi cating the gravel was not an active part of the sediment transport system during the breakup observed. A paired t test (Li, 1964, p. 108-111) was used in order to test if a difference existed between the rafted and nonrafted samples in their respective values of mean diameter, standard deviation, skewness, and kurtosis. An F test (Li, 1964, p. 122-124) was used to.determine if the amount of variability was different within each group (rafted and nonrafted). 108 Pearson correlation coefficients (Weinberg and Schumaker, 1967, p. 259-265) were used to ascertain if the depth of the active layer was related to the grain size statistical parameters and the elevation of the sample site. An analysis of variance with a completely randomized block design (Li, 1964, p. 193-197) was used to test if thaw depth differed on- east and west facing slopes. 109 APPENDIX V ORTHOGONAL COMPARISONS River Bar Environments and Bar Gravel Deposits River versus Bar Sediments A comparison of all river samples (subareas 1 and 2) with all bar samples (subareas 3, 4, 5, 6, 7 , 8 , and 9) reveals the following significant (p <. 01) r e s u l t s ; 1. The mean diameter of river sediments (X^ = 1.24 0) is z significantly larger than the mean diameter of bar sediments (%z = 2.86 0). 2. Both river and bar samples are poorly sorted. Bar sedi ments, with a mean standard deviation of 2.03 0, are significantly more poorly sorted than river sediments, with a mean standard deviation of 1 .1 8 0. 3. Bar samples (Xgj^ = 0.28) are significantly finer skewed than river samples (Xgj^ = 0.07) with near symmetrical distribu t i o n s . 4. A significant difference exists between bar samples which are leptokurtic (Xj^ = 1. 21) and river samples which are very leptokurtic (XK = 1.59). G River versus All Nongravel Deposits of the Bars The river sediments (subareas 1 and 2) compared to all non- I gravel bar sediments (subareas 5, 6, 7 , 8 , and 9) yields the following results, all significant at p< . 01: 110 1. The mean diameter of river sediments (X^al.24 0) is sig nificantly larger than the mean diameter of nongravel deposits of the bars (Xjfl = 4.15 0). z 2. River and nongravel deposits of the bar are both poorly sorted. Bar samples (Xg = 1.73 0), however, are significantly more poorly sorted than river samples (Xg = 1.18 0). 3. The nongravels with a mean skewness of 0.31 (strongly fine skewed) are significantly more skewed than river sediments with a mean skewness of 0.07 (nearly symmetrical). 4. Leptokurtic nongravel bar deposits (XK = 1.34) differ G significantly from very leptokurtic river deposits (Xw = 1. 59). G Main Channel versus Side Channel A comparison of the main channel (subarea 1) with the side channel (subarea 2) reveals that the only significant difference (p <.05) occurs between the mean values of kurtosis. Side channel deposits are leptokurtic (X„. = 1.30) whereas main channel deposits *G are very leptokurtic (XK = 1.88). G River versus Gravel Deposits of the Longitudinal Bar The coarsest sediments within the study area occur within river channels (subareas 1 and 2) and on longitudinal bar gravel deposits (subareas 3 and 4). A comparison of these two regions reveals the following significant (p < . 01) r e s u l t s : .111 1. Bar gravel deposits have a significantly larger mean dia meter (XM = -0.38 0) than river deposits (XM = 1.24 0). z z 2. Very poorly sorted gravel deposits (Xg *= 2.75 0) differ I significantly from poorly sorted river deposits (X^ = 1.18 0). I 3. Fine-skewed gravel deposits (Xg^ = 0.19) differ signifi cantly from nearly symmetrical river deposits (Xg^ = 0.07). 4. Platykurtic gravel deposits (Xw- = 0.87) differ signifi- G cantly from very leptokurtic river deposits (X^ = 1.59). G Gravel Sheet Deposit versus Linguoid Gravel Deposit At p <.01 a comparison of the unvegetated gravel sheet (subarea 3) at the south end of the longitudinal bar with the linguoid gravel (subarea 4) on the northwest side of the longitudinal bar indicates the fo llo w in g : 1. The mean diameter (XM <= -1.69 0) of subarea 3 gravels is z significantly coarser than the mean diameter (X jj = 0.93 0) of subarea z 4 g r a v e l s . 2. Subarea 3 gravels (X^ = 2.42 0), which are very poorly sorted, are not as poorly sorted as subarea 4 gravels (X* = 3.10 0). I 3. Subarea 3 gravels, with a mean skewness of 0.39 (strongly fine skewed), differ significantly from subarea 4 gravels with a mean skewness of - 0.01 (near symmetrical). 112 Gravel Deposits versus All Other Deposits The following significant results (p<.01) are indicated for a comparison of gravel deposits (subareas 3 and 4) with all other deposits, (subareas 1, 2 , 5 , 6, 7 , 8 , and 9 ): 1. The mean diameter (Xj^ = -0.38 0) of gravel deposits is significantly coarser than the mean diameter (X^ = 3.32 0) of river and all nongravel deposits. 2. Very poorly sorted (X- = 2.76 0) gravel deposits differ I significantly from poorly sorted (X- = 1.57 0) river plus all non- gravel bar deposits. 3. Platykurtic (Xrr = 0.87) gravel deposits differ signifi- G cantly from leptokurtic (Xj^ = 1.41) river plus all nongravel bar d e p o s it s . Nongravel Areas of the Longitudinal and Side Bars Table 3 summarizes the results of orthogonal comparisons be tween the nongravel areas of the two bars. The findings are as follows: 1. A comparison between willows (subarea 7) of the longitudinal bar with willows (subarea 9) of the side bar indicates that no signifi cant differences exist in the grain size statistical parameters. 2. On the longitudinal bar, a comparison between grasses (s u b a re a 6) with willows (subarea 7) indicates that no significant dif ferences exist in the grain size statistical parameters. 3. A comparison of all samples from grasses (subarea 6) w ith all samples from willows (subareas 7 and 9) reveals that no significant 113 differences exist In the grain size statistical parameters. 4. A comparison of vegetated subareas 6 and 7 on th e l o n g i tudinal bar with vegetated subarea 9 of the side bar indicates that no significant differences exist in the grain size statistical parameters. 5. A comparison of unvegetated sand, silt, and clay deposits of the longitudinal bar (subarea 5) with vegetated grasses and willows (s u b a re a s 6 and 7) of the longitudinal bar indicates that the mean diameter (X« = 3.75 0) of the unvegetated sands, silts, and clays is z significantly (p*<. 01) larger than the mean diameter (Xw = 4.68 0) o f z grasses and willows. 6. On the side bar, a comparison of unvegetated deposits (s u b a re a 8 ) with vegetated deposits (subarea 9) reveals the following (p < .0 5 ) : (a) The mean diameter (XM =2.91 0) of unvegetated z deposits is significantly larger than the mean diameter (X^ = 4.74 0) of vegetated deposits. (b) Poorly sorted (Xp = 1.33 0) unvegetated sands, silts, d I and clays differ significantly from very poorly sorted (Xo = 2.16 0) I w illo w s . 7. A comparison of unvegetated sand, silt, and clay deposits (s u b a re a s 5 and 8 ) with vegetated deposits (subareas 6, 7, and 9) reveals that the mean diameter (Xw = 3.33 0) of unvegetated sand, silt, c z and clay deposits is significantly (p«c. 01) larger than the mean diameter (X^ = 4.70 0) of vegetated deposits. 114 8 . A comparison between unvegetated sands, silts, and clays of the longitudinal bar (subarea 5) with unvegetated sands, silts, and clays of the side bar (subarea 8 ) indicates that no significant differ ences exist in the grain size statistical parameters. 9. A comparison of the nongravel deposits (subareas 5, 6, and 7) of the longitudinal bar with deposits of the side bar (subareas 8 and 9) indicates that at p <.05 the longitudinal bar has a finer mean diameter (XM = 4.37 0) than the mean diameter (X j,j = 3.83 0) of the z z s id e b a r . 115 APPENDIX VI METHODS USED FOR DETERMINATION OF FLUVIAL PROCESSES C-M Diagram as an Approximation of Dominant Process As a means of identifying sedimentary processes from grain size data, Passega (1957) developed the C-M diagram which is a logarithmic diagram resulting from a plot of median diameter (M) against the coarsest one percentile (C). The principle upon which the diagram is based is stated by Passega (1964, p. 830) as follows: "Swift sedi mentary agents can be characterized best by parameters which give more information on the coarsest than on the finest fractions of their sediments." An S-shaped zone results with the plotting of a sufficiently large number of samples from a tractive current deposit (defined as one which transports its load by rolling and suspension). Using data from the modern M ississippi River, Passega found that the upper part of the S-shaped zone corresponded with bed load, the lower portion with a uniform suspension, and the intervening area with a graded suspension. Log-Normal Subpopulations of Grain Size Distributions Visher (1969) traces the development of the theory that grain size distributions are composed of log-normal subpopulations which can be related to sedimentary processes. As reported in Visher (1969, p. 1076), Spencer suggests that all clastic sediments consist of three or less log-normal populations which Moss relates to the modes of sediment 116 transport, namely, surface creep, saltation, and suspension. Middleton (1976, p. 407) reports an alternate explanation offered by Shea whereby the segments are attributed to the size distributions inherited from the parent material and the subsequent response of the grains to a t t r i t i o n . On log-probability plots, the subpopulations are represented by straight line segments of a grain size distribution. The segments can be interpreted as representing either ( 1) overlapping normal popula tions or (2) truncated populations. If the segments are considered as overlapping populations, it is necessary to use a graphic dissection technique (e.g., Visher, 1965, p. 126) in order to establish the point of equal overlap of the subpopulations. If the segments are considered as truncated populations, it is only necessary to establish the point of intersection of the segments (e.g., Middleton, 1976, p. 406). In most cases the precision required for the utilization of the graphic dissection technique as outlined by McKinney and Friedman (1970) exceeds the precision of the grain size data. Middleton (1976) demon strates that when the subpopulations are considered to be truncated, the point of inflection between the coarsest segments of an appropriate grain size distribution can be predicted from hydrodynamic theory. Therefore, it is unnecessary to use the laborious graphic dissection technique in order to establish mode of sediment transport. Hydrodynamic Theory Stream Competency The maximum size of sediment which a flow can move is the 117 competency of the flow (Blatt, et a l., 1972, p. 90). Historically stream competency has been approached from either an analysis of stream velocity or from an analysis of bottom shear stress. The former approach is an older method which is now being replaced. However, an academic disagreement exists as to which method is preferable. The Task Committee on Preparation of Sedimentation Manual (1966, p. 301) states, . . . if two flows of different depth have beds of identical sediment and the same bed shear stress, the velocities at any distance y above the bed w ill also be the same in the two flows. However, because the mean velocity occurs at y equal a constant fraction of the depth, the deeper flow will have the larger mean velocity. Thus, it is seen that mean velocity alone cannot express the scouring action of the water at the bed and that to completely specify conditions the depth must also be given. Because of the aforementioned, the Task Committee proposes that shear stress be used as the appropriate quantity to express critical condi tions. However, Ackers and White (1973, p. 2041) return to the older approach and develop a sediment transport function based on average stream velocity. Unfortunately their method was not verified for flow conditions where the Froude numbers are greater than unity. In order to establish the critical conditions for sediment movement within the Colville River study area, it has been found neces sary to use both shear stress and average current velocity. Modes of Sediment Transport The fluvial sediment transport model developed by Middleton (1976) from the research of Einstein and the work of Moss will be used as the framework for the discussion to follow. 118 The sediment load of a stream consisting of particles fine enough to remain in almost continuous suspension during transport con stitutes the wash load. The sediment load being transported in close proximity to the bed constitutes the bed load. The bed load can be con sidered as consisting of two populations which differ in their manner of movement. One population, composed of particles too large to leave the bed, is transported by sliding and rolling. The other population, composed of smaller particles, is interm ittently suspended for short distances and short durations of time. In nature the bed load and wash load probably grade into one another and no sharp boundary exists be tween the two. Under the proper hydraulic conditions (e.g., falling stage) part of the wash load may become part of the bed load. From the work of Einstein and Johnson, Middleton (1976) recog nized that the key to the interpretation of hydrodynamic conditions from fluvial sediments lies not with the wash load but rather with the bed load. Middleton (1976, p. 407) reports that Einstein and Johnson found that the amount of wash load in a stream is determined by the supply and not by the hydraulics. Bed Load, Initiation of Particle Motion At our present state of knowledge, a precise determination of the hydrodynamic conditions necessary for the movement of bed load is not possible. Briggs and Middleton (1965, p. 13) state that no satis factory theory exists; however one approach to the problem has resulted from the work of Shields. This approach to the sediment transport problem yields only the conditions necessary for the initiation of bed load movement on a flat bed. 119 Leliavsky (1966, p. 47-51) presents a simplified version of the Shields’ dimensionless analysis and the Shields’ curve. However, the Shields' curve as modified by Vanoni, w ill be used in this report. It is presented in the following publications: (1) Briggs and Middleton (1965, p. 11), (2) Task Committee on Preparation of Sedimentation Manual (1966, p. 297), and (3) B latt, Middleton, and Murray (1972, p. 91). The Shields' curve as presented in the aforementioned publica tions results from a plot of two dimensionless numbers, the dimension- less critical shear stress (/3) as ordinate against the critical boundary Reynolds number (Rg) as abscissa. The dimensionless critical shear stress and the boundary Reynolds number are given by: 1^3 9 B (1) ( Ye - Y >d 8 ( /O s — /o )d n = U* d . ^ y ( 2 ) w here 6, = the shear stress acting on the bottom d = particle diameter y & a specific weight of the solid Y = specific weight of the fluid g = gravitational constant /% = density of the solid / o = density of the fluid U* = shear velocity V = k i n e t i c v i s c o s i t y d e fin e d a s V® w here f J . - dynamic velocity of the f l u i d 120 When turbulence is fully developed at the bed (Rg 600), the dimensionless critical shear stress (/&) assumes a constant value of 0.06 (Briggs and Middleton, 1965, p. 13). By substituting 0.06 in e q u a tio n 1 and rearranging terms, the critical shear stress necessary to initiate motion (2^), when Re 600, is determined by: i c - 0 .0 6 d g (3) From the known critical shear stress ( 2rc ) fche critical shear velocity is derived by using: U* (4 ) The Shields' diagram is cumbersome to use due to the presence of the unknown shear velocity (U*) in the abscissa and the unknown shear stress ( go) in the ordinate. This problem can be overcome by calculating the quantity: (5 ) and reading the value of either ^or Rg off the graph given by Blatt, Middleton, and Murray (1972, p. 91). Whenever equation 3 is used, equation 5 is employed with Shields' curve to establish that Rg is greater than 600. This procedure is used to establish the validity of the use of equation 3. In order to ascertain if the hydrodynamic conditions in a specific cross-section of a river are sufficient to initiate motion of a particle of given grain size, it is necessary to: either calculate the bottom shear velocity (U*) acting in the given cross-section under 121 the specified conditions and compare the result with the predicted shear velocity given by equation 4, or calculate the average velocity (U) for the cross-section which would be sufficient to attain the necessary bottom shear velocity and compare the result with the average velocity reported from the field. This can be achieved by employing Keulagan's equation (Gray and Wigham, 1973, p. 8.71 - 8.72) which for rough boundaries can be expressed in the general form: U = U*(A 0 + logS:) ( 6) K q k w here A 0 = a function of the channel shape R = hydraulic radius k = channel roughness factor kQ ss von Raman's coefficient for turbulent exchange This equation assumes that uniform flow is occurring within a channel cross-section with a logarithmic velocity profile. The validity of the results obtained from Keulagan's equation can be better understood from a consideration of the derivation and evaluation of each quantity in the relationship. As w ill be seen, it is not possible to be entirely objective in evaluating AQ, k, and kQ. The quantities R and U*, however, are somewhat more narrowly defined and therefore less judgement is required in their evaluation. The value of the hydraulic radius (R) is set by the geometry of the channel cross-section under investigation. For a wide channel, where the width is much greater than the depth, the hydraulic radius 122 is equivalent to the mean depth (Leopold, et a l., 1964, p. 156-7) given by A/w where A is the cross-sectional area and w is the channel width. This can be expressed as follows: R = -A- (7) w The approach used in this paper has one lim itation which should be noted. For the calculation of the average stream velocity (U) in a cross-section, it does not differentiate the critical shear velocity acting on the grain from the total shear velocity acting on the bottom. Using the formula given by Ackers and White (1975, p. 625) the total shear velocity acting on the bottom is given by: " t o t a l - J SD i (8 ) where D = mean depth of flow i = gradient Data of Ackers and White appear to indicate that the shear velocity with respect to the grain is about 7 percent less than the total shear velocity acting on the bottom. In evaluating U* from the Shields’ curve, a further notation should be made. In equation 3 the quantity equals 0.06. The work of Ackers and White (1973, p. 2049) suggest that possibly/3 is less than 0.06. However, in this analysis / 3 is considered to be 0.06. The shape factor A 0 in Keulagan's equation is found to vary little with change in channel cross-section (Chow, 1959, p. 203). How ever, Chow (1959, p. 204-5) found that when A 0 is considered as a function of the Froude number it varies widely. Iwagaki reasoned that at high Froude numbers the resistance to turbulent flow increases due to 123 the increased instability of the free surface (Chow, 1959, p. 204). The quantity AQ is not only a function of channel shape but is also effected by such factors as the instability of the free surface and unequal distribution of the boundary tractive force (Chow, 1959, p. 203). For rough channels a study reported by Chow (1959, p. 204) reveals that AQ has a range from 2.23 to 16.92. Gray and Wigham (1973, p. 8.72) report, "AQ may vary in the order of magnitude of +50 per cent so as to produce an error in U of 125 per cent." In applying Keulagan's equation, Jordan (1965, p. 39) assumed that the mean value of 6.25 for clear flow could be used for the sediment laden water of the M ississippi River. In the present study AQ is considered to be a constant of 6.25 regardless of the concentration of suspended sediment. Einstein (1950, p. 8 ) defines the channel roughness factor k as equal to dgg which is, "the sieve size of which 65 percent of the mixture (by weight) is finer." For k, dg^ can be used when no bed or bank roughness exists except for that due to the stationary grains (Jordan, 1965, p. 43). In this investigation k is assumed to equal dgg. This value is used even though it is recognized that all grain sizes on a bed contribute to the roughness. Shen (1975) reports that no evidence exists which indicates that dgg should or should not be u s e d . From laboratory experiments kQ for clear flow is found to be 0.4 (Blatt, et al., 1972, p. 88 ). Computed kQ from data collected in the M ississippi River at St. Louis, Missouri, vary from 0.4 to less than 0.2 (Jordan, 1975, p. 57). B latt, Middleton, and Murray (1972, p. 88 ) report kQ decreases with increased concentration of sand size 124 suspended sediment. In this investigation computations of velocity are made for kQ equal to 0.4 and 0.2. The respective results are con sidered as indicators of minimum and maximum average current velocities necessary to initiate particle motion. Interm ittent Suspension of Bed Load Particles From the work of Lane and Kalinske, Middleton (1976, p. 410) presents the following criterion for particle suspension: JO - ± 1 (9) U* where u>= particle settling velocity With the use of equation 9 and the graph of Rouse, for particle settling velocity, as presented in Blatt, Middleton, and Murray (1972, p. 54) it should be possible to predict for a given water temperature the maximum particle size which w ill be placed into suspension. This can be accomplished in the following manner. The shear velocity necessary to initiate particle motion is equated to the settling velocity so that «*/U* equals 1. The maximum particle size which w ill be placed into suspension for the appropriate temperature is now determined by referring to Rouse's graph. This particle size should mark the boundary between the bed load being transported by traction and the bed load being transported by interm ittent suspension. METERS 2 - 3 - —12 —6 —8 —2 -2 -4 -6 -10 8 - 100 LONGITUDINAL BAR LONGITUDINALBAR SEDIMENTPROFILES THAWWITH HORIZONTAL SCALE METERS R E T E M 0 0 1 APPENDIX VII APPENDIX FEET E E F 0 0 3 125 126 -14 4 - -12 2 u j t i l HORIZONTAL SCALE SOOFEET 20 IOOMETERS 2 - - 8 —12 127 -8 2- -6 - 4 -2 — 2, HORIZONTAL SCALE SOOFEET 20 - 8 ■3- —12 128 -8 -6 UJ — U J —2 -I- HORIZONTAL SCALE — 6100 500FEET IOOMETERS 2- —8 - 3 - —10 —12 VITA Donald F. Nemeth was born on March 27, 1938, in East Chicago, Indiana. He graduated from Roosevelt High School in East Chicago and attended Regis College in Denver, Colorado. He received a B.S. in Geology from the University of Southern California in 1962. In 1969, he received an M.S. in Geology from the University of Southern California. His thesis was titled , "Sedimentary and Topographic Character of Scree at Three California Localities." During his m aster's studies, Mr. Nemeth also served as Instructor of Geology at California State College at Los Angeles and at Citrus College in Azusa, California. Mr. Nemeth entered the Graduate School of Louisiana State University in June, 1969, and was awarded a teaching and research assistantship by the Department of Geology. In 1971, he accom panied Dr. Harley J. Walker to Alaska on a research expedition. He is the recipient of the Penrose Grant from the Geological Society of America which was awarded to partly finance his dissertation research on the north slope of Alaska. Mr. Nemeth is a member of the Geological Society of America, the American Geological Institute, and the American Association for the Advancement of Science. In September, 1969, Mr. Nemeth married Darlyne Gaynor. Mr. Nemeth is a candidate for the Doctor of Philosophy degree in Geology at the spring commencement. 129 EXAMINATION AND THESIS REPORT Candidate: Donald F. Nemeth Major Field: Geology Title of Thesis: Morphology/ Grain Size Characteristics/ and Fluvial Processes of Two Bars/ Colville River Delta/ Alaska. Approved: ajor Professor and Chairman Dean of the Graduate School EXAMINING COMMITTEE: a J c ( 1 ft. WjJk Date of Examination: April 13, 1977