THE UNBC BEDLOAD MOVEMENT DETECTOR:
CALIBRATION, INITIAL RESULTS AND ANALYSIS

by
Jon Tunnicliffe

THESIS SUBMITTED IN PARTIAL FULFILMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
in
NATURAL RESOURCES AND ENVIRONMENTAL STUDIES

© Jon Tunnicliffe, 2000
THE UNIVERSITY OF NORTHERN BRITISH COLUMBIA
July 2000
All rights reserved. This work may not be
reproduced in whole or in part, by photocopy or
other means, without the permission of the authorUNIVERSITY OF NORTHERN

BRITISH COLUMBIA
LIBRARY
Prince George, BC

111

Table of Contents
Table of Contents ............................................................................................................. iii
List of Tables .................................................................................................................... vi
List of Figures .................................................................................................................. vii
Acknowledgements .......................................................................................................... xi

Acknowledgements .......................................................................................................... xi
Abstract ............................................................................................................................ xii
Introduction ....................................................................................................................... 1

CHAPTER 1 BEDLOAD TRANSPORT AND ITS MEASUREMENT ................... 6
1.1 Bedload Transport ..... ... .. ........... ................ .. ......................... ............... ..... ... ...... ..... ... 6
1.1.1 The Entrainment Threshold .... ........ ........ .... ............... ........... .......... ... ...... .... ...... 7
1.1.2. The Role ofTurbulence ...... ...... ............................................. ... ....... .... ........ ..... 9
1.1.3 The Effects of Relative Protrusion, Clusters and Bed Annouring .... ............... 10
1.1.4. Multi -phase models of transport ....... .... ..... ........ ....... ....... ........... ......... .... ....... 12
1.2 Measuring Bedload Movement .. ........ ........... .............. ....... ........ ... ..... ..................... 13
1.2.1 Direct Measurement ....... ........ ...... ........ .... .............. .......... ......... ... ........ ........ .... 13
1.2.2 Indirect Measurement .. ......... ........ ..... ... ..... ... ........ ..... .. ........ ... ....... ........... ....... 19
1.3 Stuart-Takla Bedload Studies .................. ........................ ........... ... ....... .................. 21
1.4 New Bedload Sampling Technology .............. ........ ........ ....... .................. ......... ... ... 22
CHAPTER 2 THE UNBC BEDLOAD MOVEMENT DETECTOR ....................... 24
2.1 The Sensor ..... ......... ....... ....... .... ...... ...................... ................. ....... ............. ....... ...... 24
2.2 The Housing Unit. .... ....... .... ...... .... .................................... ....................... .... .......... . 26
2.3 Data Acquisition .... ... .... .... .... ........ .... .... ........ .... ....... ....... ...... ... .... ....... ................... . 28
2.4 Power Supply ..... ........ ........... .................. .... .................... ........... ............ ....... .. .. ...... 29

lV

2.5 The Study Site ............ ...... ...... .............................................. ....... .......... ..... ............. 30
2.5 .1 Sediment Supply .............................................................................................. 31
2.5.2 Lithologies ... ... ....... ........ ........... .. ..... ...... ............... ............ ....... ... .... ... ......... ..... 34
CHAPTER 3 TIME SERIES AND FREQUENCY ANALYSIS ............................ 36
3.1 Identification and Characterization of Signals and Frequencies ............................ 37
3 .1.1. The Influences of Clast Velocity On a Signal ........... ... ...... ... ......... ...... .. ... ..... 42
3 .1.2 The Influences of Clast Size On a Signal ........................................................ 44
3.2 Random Passage of Sediment ............ .......... ........................................ ....... .... .... .... 47
CHAPTER 4 ANALYSIS OF THE FLOOD DATA RECORD ............................. 50
4.1 Nival Flood, 1998 ....... ............ .... ........ ........ ..... .... ........ ....... ......... ...... .......... .... ...... 52
4.2 Nival Flood, 1999 .... .. .. ...... ............. .... .......... ....... .... .... ..... ....... .......... ................... . 55
4.3 Sediment Transport Patterns .............................................. ........... ....... ....... .......... .. 62
4.3.1 Marginal Transport ........ ............ .......... ............. ......... ....... ............................... 62
4.3.2 Clustered Transport ................................................... ....... ........ .. ........ .............. 63
4.3 .3 Continuous Transport ................................................................. ..... ....... .... .... .. 64
4.3 .4 Sand Streets ........... ..... ...... ....... ..................... ..... ... ...... ............................. ........ . 65
CHAPTER 5 PATTERNS OF BEDLOAD TRANSPORT ................................... 68
5.1 Discussion of Sediment Transport in O'Ne-ell Creek ............................................ 68
5.2 Temporal Variation in Transport ............................................................................ 68
5.3 Pulsations in the 1999 transport record ................................................................... 72
5.4 Lateral Instability - The Shifting Locus of Transport .... .. ..... ... ... ..... .... ............... .... 73
5.5 Transport Volumes- A Transport Estimate from the 1999 Nival Event ..... ....... ... 75
5.5.1 Magnitude ofError ........ ............ ..... ... ......................... ... ................ ....... ........... 81
5.5.2 Stream Efficiency.... ................... .......... ................. ................................ ........... 82
5.6 Comparison with Existing Data .............................................................................. 84
CHAPTER 6 FUTURE DIRECTIONS AND CONCLUSION ............................... 86
6.1 Flume Studies ....... ........... .... .................... ...... ... ........................ ...... ................. ...... .. 86
6.2 Device Installation .................. ................... ....... ........ ... .. .. ........... .......................... .. 86

v

603

Automation and Signal Processing

o o o ooo o o o o o o o o o o o o o ooo o o o o o ooo o o o o o o 00 00 0 0 0 00 0 0 0 0 0 00 0 Oo 0 0 0 0 0 00 00 00 0 0 00 000

87

Conclusions ...................................................................................................................... 89

References ........................................................................................................................ 91
Appendix A- Glossary ................................................................................................. 100
Appendix B - Gravel sampling data ........................................................................... 104

Appendix C - Laboratory calibration data ................................................................ 117

Definitions for terms in bold type throughout the text can be found in Appendix A Glossary, on page 1000

Vl

List of Tables

Table 1- Representative lithologies found in the bed material at O'Ne-ell Creek, N=234 .
............................ .......................... ..................... .... ........ ......... ....... ...... ....... ....... ........ 34
Table 2- An estimate of the total number of grains per kilogram of bed material from
O'Ne-ell Creek, based on 4 bulk samples taken at the study site ............................. 80
Table 3 - Daily particle counts, with associated bulk sediment transport estimates. The
record from June 161h is incomplete .......................................................................... 81
Table 4- Summary ofhydraulic parameters at O'Ne-ell1950 ......................................... 83
Table 5 - Calculations of transport rates through three reaches in the Stuart-Takla
Watersheds. Cut/Fill Volumes are taken from morphological estimates made by
Poirier (2000). Morphological Step Lengths and Tracer Step Lengths provide two
estimates of transport distances within the reach. Cut/Fill volumes are accordingly
divided by the number of step lengths in the reach, providing two estimates of
transport volumes ............................ .............. ......... ............. ....... ........ ........ ............. .. 85
Table 6 - Freeze Core sample data .................................................................................. 105
Table 7 - Bulk Sample (15cm deep) Sieve Data - cont on next page ............................. 106
Table 8 - Bulk sample (30cm deep) Sieve Data ............................................................. 108
Table 9- Sample of Magnetic Sampling Data (All clasts< 50mm) ................................ 109
Table 10 - Characteristics of rock samples used in pendulum experiments ................... 117

Vll

List of Figures

Figure 1 -The Stuart-Takla Experimental Watersheds ............. ........... .......... .... ................ 2
Figure 2- O'Ne-ell Creek hydrograph- 1994-1997. Measurements were taken at an
MOF gauging station 300m downstream of the study site. The flows shown here
include Tsitsutl Creek, (approx. 250m downstream of the BMD site) which adds
some 40o/o to the discharge ......................................................................................... 3
Figure 3 - Gluskie, Forfar, O'Ne-ell and VanDecar drainages along Middle River ......... . 4
Figure 4 - Vertical variation in transported material. After Klingeman, 1981. .................. 6
Figure 5 -Force moment balance diagram for the entrainment of a single-sediment grain .
............. ........................ ....... ........................... .... ......... .... ....... ..... ................................. 7
Figure 6 - The VUV pressure difference sampler - top and side views (Hubbell, 1967) 14
Figure 7 - The Helley-Smith portable sampler (Emmett, 1980) ...................................... 15
Figure 8- The initial bedload detector design by Ergenzinger and Custer (1983) .......... 18
Figure 9 - Exploded view of the sensor (above), sensor dimensions and the assembled
product. ..................................................................................................................... 25
Figure 10- Voltage response as ferrous items pass at various points on the sensor........ 26
Figure 11 - Schematic diagram of the BMD installed in the creek bed ........................... 26
Figure 12- Photo of the study site in early August. The instrument shed is located on the
east side of the creek. The BMD is set within the streambed downstream of a group
of mature sockeye salmon. A log and wire-rope cable provided access to the stream
cross-section at all stages .......................................................................................... 27
Figure 13 -The O'Ne-ell Creek watershed ..................................... ................ .................. 30
Figure 14 - Grain size distribution from four bulk sieve samples. Three samples were
excavated to a depth of 15 em and one (represented by a dashed line) went to 25 em .
...................................................... .... .................................... ............ ......................... 32
Figure 15 - Grain size distribution from freeze-core samples. Dashed lines indicate
subsurface samples (below 15 em) and solid lines indicate surface material. .......... 32

Vlll

Figure 16- Texture of the stream bed at the study site, based on a 1998 topographic
survey...... ..... .... .... ....... ...... ........ ............................ ....... ..... ......... ....... ..... ..... .... .......... 33
Figure 17 - Time series x( t) of a sensor's response to a passing clast.. ............................ 3 7
Figure 18 - The signal trace x(t) (black) with its corresponding Hilbert transform h(t)
(dashed) and signal envelope A(t), shown in gray............... ...... ............................... 38
Figure 19 - A sine pulse presented in time and frequency space. The relative brightness
(z-scale magnitude) indicates the intensity of the frequencies. Note that the peak
frequency response of the signal is located at (31.6 ...;- 2 =) 15.8 Hz, and occurs at
·································· 40
15/100 s. ··················································································
Figure 20 - STFT spectrogram of a single sensor trace during a flood event. .......... ....... 41
Figure 21 -The relationship of velocity and signal frequency for clasts of different sizes .
.... ............... .... ........ ...... ................ .... ........ .................................................................. 42
Figure 22- Stage/Discharge rating graph for O'Ne-ell Creek 1950, June 1999 .............. 43
Figure 23- Velocity profile at O'Ne-ell1950, May-June, 1999. The exponential lines
indicate the Universal Velocity Distribution Law for channel geometries based on
low stage (May 1i h - 0.23m) and high stage (June 141h - 0.68m). Four velocity
measurements are indicated for each measurement period- two sets taken at 0.6 and
0.2 depth in the area of maximum flow .................................................... ................ 44
Figure 24 - Signal amplitude diminishes and period increases with lower clast velocities
................................ ......... ...... ........ ......................... ........ ........... .............................. .. 45
Figure 25- A comparison of signal envelopes for different sized clasts .............. ........... . 45
Figure 26 - A comparison of clast size versus the "apparent" clast size calculated from
analysis of the voltage signal from a sensor. .......................... ........ .... ...................... 46
Figure 27 - The probability function facing a particle as it approaches the device. The
dashed line indicates the possible chord length of its transect across a sensor. The
relative sensitivity of the sensor is outlined by the experiments introduced in Section
2.1 ...................................................................................................... ...... .................. 48
Figure 28 - The BMD Record consists of summing the voltage activity in absolute
(positive) units for each channel in 30 second intervals. The resultant sums, showing
the magnitude of activity across the channel width, are shown in the chart on the
right. .......................................................................................................................... 51

lX

Figure 29 - The sums for each 30-seconds interval are then plotted over the hours and
days of the flood event. Here a 2.5 hour section of data from 20:00 to 22:30 on June
11 is shown .......... ........... ...... ........ ............ .... ..... ........ .. ... ......... ......... ..... ..... .... .......... 51
Figure 30 - Stage readings during the 1998 nival flood. The gauge was swamped during
the peak flows, and thus data from the night of May 25 have been reconstructed
from readings taken just upstream of the station. Flow subsided quickly over the
next three days . .... ........... ..................................... .......... ......... .... ....... .... ..... ...... ........ 52
Figure 31- Data from the peak ofthe 1998 flood event, May 25 and 26 ......................... 54
Figure 32 - Stage, bedload activity and suspended sediment readings from the 1999 flood.
·········
························································
·························
················
·························56
Figure 33- BMD Site Water Velocity Cross Section: contours based on maximum stage
measurements in 1999 show the variation in velocity across the channel width. The
vertical dimension is exaggerated threefold ........... ........ ....... ....... ......... ...... ............. 57
Figure 34 (following pages)- The BMD record from June 10 to June 17. The channel
width is measured from the west side of the channel to the east. File lengths vary
from one half hour to four hours. Gaps between the files are from maintenance and
technical problems. Most of the record on June 16 was missed after a disk drive
failure ... ....... ... ....... ......... ............ ....... ...... .................................................................. 58
Figure 35 -Marginal transport .. ............................. .... .............. ......................................... 62
Figure 36 - Clustered transport ................................... ....................... .... .................. ......... 64
Figure 37- Transport at the height ofthe flood event. .............................................. ....... 65
Figure 38 - Steady discharge of sand and gravel along the bed .......... ......................... ..... 66
Figure 39 - The onset of a gravel street, 1.2 m wide. The steady passage of material
lasted for over 10 minutes. This diagram shows the first 90 seconds of the transport
event. ...................... .................. .... ........ ..................................................................... 67
Figure 40 - Figure from Ehrenberger's bedload measurements, taken with a basket
sampler on the River Danube at Vienna National Bridge, 24 June 1931 suggest a
pulsating movement ofbedload. Sampling time=100-300s. Mean transport rate=
0.236 kg/s/m. (Ehrenberger, 1931- Fig. 6). Taken from Gomez, 1991. .................. 69
Figure 41 - Results from magnetic bedload detector at Squaw Creek, Montana (Bunte
1996) ... ................... ......... ........ ............................. ............................................. ........ 70

X

Figure 42 - Total transport rate versus time. Measurements were taken in a gravel-bed
flume, showing transport fluctuations on the order of tens of minutes (Kuhnle and
Southard, 1988) .. .... ............... .. ............................... .. .............. ......... ..... ..... ... ......... .... 70
Figure 43 - Figure showing the pulsing nature of bedload transport. The upper portion of
the graph displays the width of the stream over the course of 4 hours on June 15.
The lower half of the graph is the voltage from a sensor located 3.4 meters from the
right bank of the creek ..... ...... ........... ........ ... ..... ..... ........ ... ........... ........ ...... .... ........... 73
Figure 44 - The wandering locus of transport - the locus of transport is defined here as
the zone of maximum bedload activity throughout each data file from the 1999
Nival event. The locus of transport moved substantially throughout the event,
starting within roughly 1 m of the right bank (see Figure 34, June 11th) and drifting
to within 2m of the right bank (June 15-16) .. ........ ........ .......... ............. ........ ........ ... 74
Figure 45 - The total activity across each sensor is summed for every BMD file. File
lengths are usually 3-4 hours. Summed values were divided by the length of the file
to ensure compatible numbers. The cross section profile (see Figure 33) adds the
context of channel geometry . ........... .................. .................. ...................... ........... .... 74
Figure 46- Distribution of grain geometries based on a sampling of material >50 mm The
average sphericity of particles is 0.7 .................................. ......... ...... .. ..... ................. 77
Figure 47- Size distribution of the theoretical bedload material used in the calculations.
The proportions are based on the 4 bulk samples ofbed material taken at the site.
The coarsest material is 45 mm (5.5 1/J), and the assumed theshold of detection is 4
mm (2.0 1/J) . ............................................................................................ ................... 78
Figure 48 - Rates of particle transport (number of events per second) over the course of
the flood .. ....... ..... ....... ... ........ ...................................... ........ ... ...... ..... .... ..... ............... 79
Figure 49 - The effect of two important parameters in the volumetric transport estimate.
As the assumed threshold of detection by a BMD sensor increases from 0.5 mm to 8
mm, the estimate of transported material grows by several orders of magnitude. The
assumptions regarding the largest particle in motion (y-axis) affect the volumetric
estimate as well - as the assumed size moves from 8 mm to 64 mm, the estimate
increases by a factor of 3 ........ .................... ............. ........... ............. ......... .... .... ...... .. 82
Figure 50 - Locations of gravel sampling sites at the BMD study site .. ........... .............. 104

XI

Acknowledgements
I would like to first of all acknowledge Dr. Allen Gottesfeld for his expert
guidance and contagious enthusiasm for the study of Northern BCs rich legacy of
geomorphology. I would also like to thank Ron Poirier and Tim Dressel for helpful
discussions and technical assistance.
For their helpful comments and encouragement with this study, I would like to
thank my committee members, Moustafa Mohamed and Peter Jackson.
Thanks are due to Rob Huggins (Geometries of Palo Alto, CA) and Dr. Tadeusz
Ulrych (UBC) for advice, readings and encouragement in my crash course in digital
signal processing. I am grateful to Michael Church for the opportunity to present and
discuss the BMD project with graduate colleagues at UBC. For their support in the field,
I would like to thank the staff at the DFO camp, including Erland Macisaac, Herb
Herunter, and John Heinonen. Pierre Beaudry provided assistance with the monitoring
program at O'Ne-ell Creek. Thank you to Dev Khurana for help with the gravel sampling
and site reconnaissance. I am also indebted to the people of the Tl'azten Nation, from
whom I have had the good fortune to learn about the Middle River area, particularly
Wally Joseph and Paul Williams.
Most importantly, thanks are due to my partner, Kristiann Allen, for seeing me
through two field seasons and thesis writing with boundless encouragement, inspiration
and magic.

This research was supported by a grant from the Ecosystems Research Program of
Forest Renewal British Columbia, under research award # OP96043-RE.

Xll

Abstract

In September 1997 the UNBC Bedload Movement Detector was installed in
O'Ne-ell Creek, a mountain gravel bed river in the northern headwaters of the Fraser
River. The device is designed to measure coarse sediment flux by passive induction as
bedload material (> 1 mm), containing natural ferrous minerals, passes over an array of
electromagnetic sensors. The device consists of a series of 82 sensors, housed in an
aluminum beam and placed across the stream, inserted such that its surface is flush with
the gravel bed. The device can be raised or lowered to compensate for bed aggradation
and scour. A data acquisition system collects signals from the device at 100 Hz,
providing a continuous, high-resolution record of any bedload transport event. The device
is sensitive enough to record the movement of most volcanic, metamorphic, granitic, and
ultramafic clasts larger than a few millimeters.
The installation and operation of the device is described, as well as the calibration
and development of signal processing routines for estimating transport rates. Nival event
transport records are presented and discussed. Some of the phenomena apparent in the
records include: a pulsating pattern of activity, discrete 'sweep transport' events, lateral
movement of the transport zone, and a sudden onset ofbed movement with a tapered
cessation. An initial estimate of transport volumes compares favourably with existing
measurements taken at the study site. It is anticipated that more sophisticated calibration
work will continue in the future.

Introduction
The regime of a gravel-bed river system is the product of the complex interaction
between stream morphology and water and sediment discharge. The transport of sediment
along a channel is a critical linkage in this balance, and a decidedly difficult phenomenon
to measure. Although some relations have been established between the fluid forces
exerted by the flow and sediment transport, the relation is not uniform, and is subject to
variations in time and space, even under conditions of steady flow.
Material that is carried along the channel originates from sources both within and
outside the fluvial system. Allocthonous material is carried from hill slopes and colluvial
sources, as well glacial deposits. Autochthonous sediments originate from within the
channel, particularly from storage in banks and bars. The latter generally constitutes the
majority of the transported load, however the mixture varies, reflecting the degree to
which the river is coupled with external sources and the availability of the in-channel
sediments.
The sediment transport characteristics of a stream are important determinants of
channel equilibrium and stability. Since the coarser material in a channel provides both
the structural framework of the channel as well as the medium that accommodates
hydraulic forces, transport of this material is an important process in channel
morphology.
A new device has been developed to quantify and characterize the movement of
coarse sediment along a reach of an undisturbed, salmon-spawning creek. Using an array
of highly responsive magnetic sensors embedded in the stream, passing clasts are
detected by the ferrous minerals they contain. Voltage readings from the sensors are
recorded and stored in a digital format. By processing this information a coherent picture
of transport dynamics may be obtained. In particular, details of the conditions necessary

2
for entrainment, and the spatial and temporal variations ofbedload transport become
remarkably evident. The device is hereafter referred to as a Bedload Movement Detector,
or BMD. It was installed in September 1997 at O'Ne-ell Creek, in the Stuart Takla
Experimental Watershed area (Figure 1).
The present study was undertaken as a component of The Stuart Takla FishForestry Interaction Project. Experiments have been undertaken and baseline data
collected since 1990 on topics such as salmonid populations, riparian vegetation, thermal
regime and hydrology. The area is located 150 kilometers north west of Fort St. James, in
the drainages of the Stuart-Takla watersheds (Figure 1), in Northern BC.

1!25"

,,
/ '~

wl •

,

Stuart-Takla Exp"erimental
Watersheds
British Columbia, Canada ,,
20

30km

I

,

I

-~.: ~~ '
\. '

~

Zl

Tl'emhleur
J.ske

~

.

. . Fort St. James
Figure 1- The Stuart-Takla Experimental Watersheds

3
Sediment movement in the Stuart-Takla streams is extremely episodic. The
majority of fine and coarse sediment moves during the spring snow-melt floods (Beaudry
2000; Gottesfeld, 1998). The discharge hydrographs from 1994-1997 (see Figure 2)
illustrate the dominant role of nival floods in the annual flows . The flooding usually
occurs in late May and early June. Brief summer and fall floods may be observed in the
record, but the volume of coarse material transported during these events is relatively
small.
Substantial amounts of coarse sediment are also moved by salmon redd
excavation in the late summer. One of the initial goals of this project was to compare the
quantity and timing of bedload movement measured in a nival flood event with that
transported by spawning salmon. However, due to very disappointing returns of the Early
Stuart Sockeye run in 1998 and 1999, there was insufficient spawner density to initiate
full motion of the streambed.

-li)

15.00

"'E

-1994
1995
1996
1997

... 10.00
Q)

C)

C'CI

.c
CJ
Ill

c

5.00
0.00
01-Apr

01-May

01-Jun

01-Jul

01-Aug

01-Sep

01-0ct

Figure 2- O'Ne-ell Creek hydrograph -1994-1997. Measurements were taken at an MOF gauging
station 300m downstream of the study site. The flows shown here include Tsitsutl Creek,
(approx. 250m downstream of the BMD site) which adds some 40% to the discharge.

4

Stuart-Takla Experimental
Watersheds
lOOm contours
- - - Watershed Boundary
·•••••••• Logging Access Road
0

1

2

3

4

5 km

Figure 3- Gluskie, Forfar, O'Ne-ell and VanDecar drainages along Middle River.

The scope of the present project is fourfold: firstly, the BMD is a novel instrument
for the quantification of bedload movement. Therefore, I will:
•

Discuss existing bedload sampling techniques (Section 1.2).

•

Evaluate the relative advantages and disadvantages of the BMD as a means of
achieving volumetric estimates of bedload transport (Section 1.4).

•

Present a description of the BMD components, their installation and operation
(Sections 2.1-2.4).

5

Secondly, calibration of the sensors used in the BMD was undertaken in the
laboratory to determine their sensitivity and response. From the experimental results I
will:
•

Show how signals from the transport record vary with changes in particle velocity
(Section 3.1.1).

•

Describe the response of the sensors to rocks of varying sizes and lithologies
(Section 3 .1.2).

•

Demonstrate the time series and frequency analysis of the signal traces from the
BMD sensor (Section 3.1).

Thirdly, bedload movement was sampled with the BMD during the nival events
of 1998 and 1999. From these data I will:
•

Present the record of bedload movement for both flood events (Section 4.1 and
4.2).

•

Compare and contrast examples of the record during periods of marginal
transport, rising-stage and full-stage (Section 4.3).

•

Discuss several 'modes' of transport, from marginal step-and-rest events to
discrete ' sweep' and 'cluster' transport, to the movement of gravel and sand
sheets along the channel (Sections 4.3).

•

Estimate the volume of material transported, and compare this to existing
measurements taken within the Stuart-Takla Watersheds (Sections 5.5).

Fourthly, the goal ofthis endeavour is to apply the device towards the
investigation of a poorly understood phenomenon: the spatial and temporal variation in
bedload transport rates. Some of the characteristics of transport activity during the nival
event records are highlighted.
•

Wave-like patterns in sediment transport rate are shown to exist on several timescales (Sections 5.3).

•

A "wandering" motion of the channel's locus of transport is substantiated
(Section 5.4).

6

Chapter 1 Bedload Transport and Its Measurement

1. 1 Bedload Transport

Bedload is defined as that portion of the sediment load that rolls or slides along
the bed, remaining in nearly continuous contact with the channel bottom at all times. In
the relatively high gradient streams of the Stuart-Takla Experimental Watersheds,
bedload seldom includes sediment that is finer than 1 mm in diameter, since finer sand
particles are easily suspended in fully turbulent flows. Bedload thus comprises coarse
sand, gravel, cobbles and boulders. Fine and medium sands represent a transitional grade,
and may be carried as suspended load or bedload, depending on the channel geometry
and hydraulic conditions (Beschta, 1987).

Relative Concentration
Figure 4- Vertical variation in transported material. After Klingeman, 1981.

At lower flows, a step and rest pattern appears to be the principal mode of bedload
transport (Andrews, 1983; Reid and Frostick, 1994). Observations indicate that this may

7
involve sporadic movement, with bursts of activity by neighbouring particles followed by
a period of relative quiescence (Custer eta/., 1987; Reid and Frostick, 1994). Bedload
transport does not always involve the full range of sediment sizes, however the patterns
and mechanisms of entrainment are the subject of some debate. The effects of turbulence,
relative protrusion, clustering and imbrication complicate the process a great deal. Any
universal function for bedload transport remains a probabilistic, rather than a predictive,
equation.
The following sections provide a brief summary of the theories on bedload
entrainment and transport dynamics.

1.1.1 The Entrainment Threshold
The entrainment of a cohesionless grain, off of the streambed and carried along
with the flow, may be considered the unbalancing of the frictional and gravitational
forces that keep the grain on the streambed. Drag forces acting in a downstream direction,
and lift forces, moving upward, pivot the grain about a point and move it out of position.
Once in motion, the grain will roll or bounce along the streambed until it settles into a
new resting position. Figure 5 illustrates the forces acting on a grain.

Stream Flow

FL
.-- --~~~----

FD
--- ·Theoretical

Bed

Figure 5 - Force moment balance diagram for the entrainment of a single-sediment grain.

Fa is the gravitational force. The drag force, F 0 , is proportional to the crosssectional area of the grain facing perpendicular to flow direction. Lift force, F L, is
proportional to the cross-sectional area of the grain parallel to the bed plane. The
distribution of forces may depend on the pivoting angle, (fJ (also referred to as the angle
of repose). The steeper the angle of repose, the more the particle will tend to resist
entrainment. For particles sitting at a low angle of repose, or for increasingly spherical
particles, the force resisting particle motion is relatively small.

e is the orientation angle,

and FF is the frictional force acting to hold the grain in place.
The transporting ability of a stream has conventionally been described in terms of
unit stream power (Bagnold, 1966; Leopold and Emmett, 1976) or shear stress exerted
on the bed (Shields, 1936; Einstein, 1950). Unit stream power, co, is defined as the
product of the unit weight of water, velocity, depth and water-surface slope (y VDS).
Shear stress at the bed, 't, is defined as the product of the unit weight ofwater, depth and
slope (y DS = co N).
There have been many equations proposed to describe the shear stress conditions
necessary for initial movement of a grain on the streambed, and many bedload transport
equations assume that transport rates vary with some function of excess shear stress
beyond this threshold.
Shields (1936) proposed that the critical shear stress for particle motion 't* ci was
solely a function of the particle Reynolds number, R, which is described as
(1.1)

where u* is the shear velocity, di is the diameter of the ith particle fraction and vis the
kinematic viscosity.

9
The Shields equation describes the dimensionless function 8 (Critical
dimensionless shear stress) as the ratio of (erosive) fluid shear stress on the bed to
submerged particle weight;

(1.2)

..
where u. 0 is the threshold shear velocity, 1"0 is threshold shear stress (proportional to the
square of the velocity), g is gravitational acceleration, Dis grain diameter (usually taken
as D 50), Ps is particle density and p is fluid (water) density. Shields calculated the typical
value of this ratio to be 0.06 for spherical grains of equal size in a laboratory flume.
Subsequent field evidence and flume studies have reported values ranging from 0.01 for
an exposed grain on a cohesionless bed (Fenton and Abbott, 1977) to more than 0.25 for
some natural streambeds (Church, 1978; Reid eta!. 1985).
The relationship between shear stress and particle entrainment is complicated by
several factors in natural streambeds, particularly in pool-riffle and braided channels.
The most significant of these are:
•

Turbulence

•

Imbrication, hiding and interlocking of particles

•

The angularity of natural grains

•

Width-wise variations in shear stress across the channel due to flow patterns

1.1.2. The Role of Turbulence
Sediment entrainment is a function not only of the shear stress acting on the bed,
but also the intensity of turbulence above it. The role of turbulence in the entrainment
process is not well understood, but numerous flume studies have shown that 'burstsweep' cycles play a critical role in the initiation of transport (see Grass, 1971; Jackson,
1976). Bagnold (1966) reasoned that the upward thrust of turbulent flow must necessarily

10
be greater, on average, than the downward component; otherwise every suspended
particle would drift back to the river bottom under the influence of gravity in turbulent
flows. This has been shown to be consistent with experimental observations of turbulent
boundary layer flows. Parcels of low-velocity fluid burst upwards within the outer layer
to be replaced by quantities of high velocity fluid that sweep downwards towards the bed,
disrupting the laminar sub-layer. Rapid fluid movements away from the bed are termed
'ejections' and accelerated flow events acting towards the bed are termed 'sweeps'.
Thome et al. (1989) and Williams et al. (1989) have shown that close to the threshold of
bedload movement in marine tidal environments, most of the transport occurs during
sweep events. Grass (1971) and Nino and Garcia (1996) have shown this in flume
experiments as well. We can assume that, in streams, sweep events also move particles
into suspension.

1.1.3 The Effects of Relative Protrusion, Clusters and Bed Armouring
Gravel bed rivers possess a number of sedimentological features that affect the
threshold of entrainment. Depending on the bed material distribution and geometry,
different packing arrangements may create hiding effects, clustering and imbrication.
Particles may become constrained from motion as they are interlocked or 'hidden' from
the shear forces acting on the bed. Therefore, gravel streambeds have susceptible
particles that will move at a predictable threshold, while other particles may be entrained
only as the force over the bed increases to a point where much larger surrounding
material begins to move. The importance of the relative size of particles, as opposed to
their absolute size has been explored extensively in the last 20 years.
In alluvial streams, the portions of the channel surface experiencing the highest
flows are typically described as being armoured, i.e. the surface layer consists of clasts
that are 2 or 3 times coarser than the average size of underlying material (Parker et al. ,
1982; Andrews and Parker, 1987; Sutherland, 1987; Willetts et al. 1987). The armour
layer develops after low and intermediate magnitude floods, when finer material is

11
selectively transported, leaving a residue of well-sorted, coarse immobile particles on the
streambed. The armour layer is usually no more than one grain thick and overlies the subarmour layer, which may be of variable thickness.
In its most pronounced form, the controlling effect of an armour layer is that
particles are interlocked and thus restrained from motion. At extreme flows a critical
point may be reached when the integrity of armour layer is diminished, the armour layer
breaks up, and all sizes of grains are available for transport at the same time (Parker and
Klingeman, 1982; Parker et al., 1982; Andrews, 1983; Andrews and Parker, 1987).
Such conditions of equal mobility may occur in the Stuart-Takla creeks to some
extent, however the effect is much less than we might observe in larger and hence deeper
streams. There are at least some sections of the stream with coarser bed material that do
become armoured after particularly high flows. These reaches did not change very much
during the relatively low-magnitude nival events observed in the present study.
Entrainment of bedload material in the Stuart-Takla streams is dominantly sizeselective. In many reaches, loose packing created by spawning salmon's annual

bioturbation of much of the stream bed encourages selective entrainment. Downstream
fining may be observed, which requires that there be greater mobility of the finer fraction
(either as greater transport distances or more frequent transport). In O'Ne-ell Creek, the
average diameter of cobbles decreases from 14 em to 7 em in the 1.5 km below the
bedload detector site. This accompanies a decrease in gradient from 0.013 to 0.005.
The effect ofroughness elements such as logjams and gravel bars add further complexity
to the entrainment model, as well.

12

1.1.4. Multi-phase models of transport

Bedload transport varies in character depending on the depth of flow and the
nature of the bed material. Some authors have divided bedload transport in alluvial
streams into two phases (Emmett, 1976; Jackson and Beschta, 1982). The first phase is
the mobilization of smaller, loose material that is freely available for transport. Sand and
fines are selectively carried from bars, channel margins, and pools. Gravel-sized
particles, especially smaller sized ones, rotate out of the pockets in which they lie, and
become entrained in the flow. Andrews (1983) points out that a significant portion of
material may be transported during this first, marginal phase.
With much higher flows, a second phase of transport begins: the armour layer is
disrupted. As the whole bed becomes mobile, the size distribution of the bedload
coarsens. The velocities in pool sections exceed those at the riffles, and material is carried
from one riffle to the next. Significant changes in bed morphology may occur during this
period. It requires exceptional flow conditions to reach this stage of transport and large
volumes may be moved.
Subsequent studies and field observations have shown that there are numerous
degrees and variations of this two phase concept (Ashworth and Ferguson, 1989)
depending on the size distribution of the bed material. At O'Ne-ell Creek, for instance,
studies have indicated that a substantial portion of the bed is disturbed even in lesser
flood events (Gottesfeld, 1998; Poirier, 2000). At flood stage, readings from the BMD
indicate mostly finer material in motion, with the sporadic passage of cobbles. No doubt
at sufficiently high flows (such as the 1990 flood) the stream bed is completely disrupted
and equal mobility sensu stricto prevails.
Although the two-phase model is not quite adequate to describe the transport
dynamics at O'Ne-ell Creek, it is a useful concept for understanding shifts in transport
intensity throughout a flood event. It is anticipated that with the UNBC Bedload

13
Movement Detector, we may begin to better understand the factors that control sediment
discharge in the creek.

1.2 Measuring Bedload Movement
Bedload movement is generally monitored using one of two techniques: direct
(net samplers, pit traps, some form of sampling device placed on or above the bed) and
indirect (tracer techniques, repeated channel surveys).
1.2.1 Direct Measurement

Traps and pit samplers of various designs have been employed to sample the
amount of sediment passing through a section of channel bed in a determined time. These
measurements are generally recorded as weight per unit time per unit stream width, i.e.
kg/m-s.
Direct sampling in high gradient mountain streams is often extremely difficult and
hazardous, given the tremendous stream power required to transport cobble and boulder
sized particles. Nevertheless, direct sampling has proven to be the most efficient and
representative methodology.

Box, Basket and Net Samplers
Box and basket samplers have been used since the early 1900s, and the design
evolved toward greater efficiency throughout the 40s and 50s, culminating in the
development of the now standard Helley-Smith sampler (Hubbell, 1964; Helley and
Smith, 1971; Gomez, 1991 ). Several of the most highly developed and tested basket
samplers are those designed by Mtihlhofer (Ehrenberger, 1932), Ehrenberger (1932),
Nesper (1937) and the Swiss Federal Authority (1939). Essentially, these devices have

14
either a solid bottom, or a bottom of fine chain links that conforms to the bed. A mesh
bag collects the coarser material, and water and fines escape at the rear of the device.
Some versions of the basket sampler have weighted frames to keep it lower in the flow,
or a vane mechanism to direct the aperture upstream.
Pressure-differential samplers, such as the Arnhem (Dutch), Karolyi and VUV,
operate by causing the bedload-laden flow passing through the device to slow sufficiently
to deposit material. Water then issues out the rear of the sampling box.

Figure 6- The VUV pressure difference sampler- top and side views (Hubbell, 1967)

Since the mid 1970s, the standard portable sampler for bedload has been the
Helley-Smith pressure-difference sampler. This device has a square 7.62 em entrance
nozzle (a 15.24 em model is also available) and a 46 em long sample bag constructed of
0.2 mm mesh polyester that holds about 10 kg of sediment. A hydraulic efficiency rating
of 1.54 over a range of flows was determined by the calibration work of Emmett (1980).
The trapping efficiency is considered to be 90 to 100 percent efficient for particle sizes
from 0.5 to 32 mm at current velocities up to 1.5 m/s (Hubbell, 1987). Others have found
that since the sampler must sit on the streambed, the geometry and placement of the

15
intake nozzle tends to alter the flow pattern, and thus affect the bedload discharge
estimate. Furthermore, it is particularly difficult to maintain contact with the stream bed
at flood stage.
It is a often a considerable challenge to find the proper sections of channel to

sample in order to account for the random or short-term, often "quasi-cyclic" temporal
and spatial variations in bed load transport rates. Since the rates of transport can change
dramatically even at constant-flow conditions, it is difficult to determine what constitutes
a representative sample. Gomez and Troutman (1997) suggest that a representative
sample of bedload activity should involve four or five traverses of a river, and the
collection of 20-40 samples at a rate of five or six samples per hour.

Figure 7- The Helley-Smith portable sampler (Emmett, 1980)

Other researchers have rigged mesh nets over a solid frame that spans part of the
stream, and these are emptied periodically throughout an event (Bunte, 1996). Clearly,
net samplers give the best representation of pebbles in motion. They do not, however,
gather any of the finer material. They also require a substantial effort, and yet the samples
may not be large enough to be statistically representative and/or frequent enough to
characterize the short period variation.

16

Pit and Bucket Traps
Hollingshead (1971) was able to monitor sediment transport along the Elbow River
in southwestern Alberta by making an excavation 50m long, 20m wide and 1m deep. As
the pit was refilled over the course of a nival event, successive surveys were taken to
determine the total sediment deposition. Smaller traps have been developed, consisting of
a bucket or tube sunk vertically into the bed (Church et al., 1991; Powell and Ashworth,
1995), usually with an inner sleeve that can be emptied. Sampling by this method can
yield a good estimate of the total mass of sediment moved during the flood event. It
provides an estimate of the gross yield, though it provides little information on the timing
and rate of transport. One successful modification of this technique was the placement of
a pressure pillow within the trap well to weigh sediment as it accumulated (Reid et al.,
1980; Kuhnle, 1991; Harris and Richards, 1995). The main disadvantage of pit and
bucket traps is that they may fill up quickly and, unless closely monitored, have little use
in sampling at high rates of sediment discharge. It is also a challenging operation to
remove, empty and replace the buckets in flood conditions.
The largest of pit samplers belongs in a category unto itself. That is the USGS
bedload trap, installed in the East Fork River, Wyoming. A conveyor belt is set within a
large concrete slot, placed perpendicular to the river flow, such that large quantities of
passing sediment could be collected in flood conditions. In this fashion, a continuous
record ofbedload transport volume was maintained (see Leopold and Emmett, 1976).
This device set a standard for bedload measurement by which other devices were
compared and calibrated.

Acoustic and Seismic Equipment
Acoustic devices have been tested in Switzerland to estimate bedload transport in a
mountain torrent by recording the sound generated by gravel collisions (Banziger and
Burch, 1990).

17
Acoustics systems that measure sediment generated noise (SGN) in an ocean
environment have been developed by several researchers in the UK. Results from these
experiments have shown that the SGN spectrum is composed ofband-limited noise, with
a frequency range inversely proportional to particle size, and an acoustic intensity
proportional to the mass of the mobile material. Williams eta!. (1989; see also Thome,
1989) collected acoustic measurements in a broad (4 km) ocean channel at West Solent.
Gravel movement was detected by suspending a hydrophone 0.25m above the bed and
recording the sounds of interparticle collisions. In conjunction with several current meters
and video equipment, some relation between SGN and gravel transport was resolved.
Rouse (1994) reported further calibration ofhydrophones recording intergranular
collisions as particles moved over a fixed flume bed. This could then be related to the
mass flux rate passing through the flume.
In the Italian Alps, Govi eta/. (1993) used a series of seismometers along an
alluvial channel to detect the motion of sediment within the reach. Microseismic data
recordings provided a continuous record of the transport events.

Vortex Tube Traps
Vortex tube traps were originally designed for ejecting sand and silt from irrigation
canals and ditches in agriculture (Hayward, 1974). Their use for measuring bedload
transport rates was pioneered by Klingeman and Milhous (1970) at Oak Creek. By
emptying the collection trap at regular intervals during a storm, they were able to
measure transport rates and particle distribution. This data set remains one of the best
continuous records of transport available anywhere. Similar devices were designed by
Hayward (1974) in New Zealand, and Tacconi and Billi (1987) in Italy.

The device can recover bedload material continuously, with up to 90 percent
efficiency. The construction and implementation of the concrete installation, however, is
a substantial undertaking.

18

In Situ Magnetic Detection Devices
These devices record the magnetic signals of clasts with implanted magnets or, in
the case of more sensitive instruments, the faint signals from remanent magnetism in
naturally-occurring minerals. Ergenzinger and Conrady (1982) installed a device
consisting of four large coils, 250 mm in diameter, attached to an aluminum frame 0.8 m
above the water surface at Buonamico, Italy. It registered the passage of 100 cobbles with
implanted magnets. The Birkbeck Bedload Sampler, installed at Turkey Brook, Enfield
Chase, UK (Reid eta/., 1984) employed a metal detector circuit with a single 2.3 m long
sensor installed within the streambed. This device was sensitive enough to detect
movement of 100 artificial clasts tagged with ferrite rods.
Ergenzinger and Custer (1983) reported a more sensitive instrument consisting of
two 1.25m coils buried horizontally in the streambed at Squaw Creek, Montana. This
device could detect naturally magnetic pebbles and cobbles > 32 mm at a rate of 2 per
second. Since only a portion of the stream was instrumented, bedload volumes were
estimated for the total stream cross section.

SHIEll ALUMINUM
COVER PLATE

CONCRETE RECEI'TICAL

CHAMBER FILLED WITH
ROOFING TAR

Figure 8 -The initial bedload detector design by Ergenzinger and Custer (1983)

19

In 1986 a more sophisticated device was installed in Squaw Creek (Bunte 1996,
Custer eta!. 1986). The device consists of 5 sets of paired multiple choke-coil units
mounted in a beam set lengthwise across the stream. Each module is 1.55 meters long. It
thus provides some spatial resolution for different sections of the channel. The placement
of two arrays of microcoils 15 em apart permits the determination of the travel time of
passing clasts (Custer eta/. 1987; Spieker and Ergenzinger, 1987). The sensitivity of this
device permits detection of an estimated 40% of the coarse material (>32 mm) at Squaw
Creek (Bunte, 1996).
All of these devices have used strip chart recorders, which limit their utility to
several signals/second and a few sensor circuits. Impulses must be evaluated and counted
visually and are usually summed over intervals ranging from 10 minutes to one hour. The
practical limit of resolution of the strip recorder system in use at Squaw Creek may be
200/hr (Bunte 1996). Consequently, the passage of many clasts may be missed due to
limitations of the data logging system. Although the device is well suited to recording the
relative intensity of bedload activity, it remains an onerous task to quantify volumes.

1.2.2 Indirect Measurement

Sediment transfer volumes may be inferred after a transport event by several
means, including tagged sediment tracers, morphological estimates, sonar surveys, and
scour chains.

Tagged Sediment
Tracer studies are one of the most popular tools for tracing the paths and timing of
gravel movement during a flood event. This method generally involves painting a number
of rocks that represent the most common sizes and shapes within the bed material,
numbering them, and then recovering and surveying their positions after a bedload

20
movement event (Crickmore, 1967; Rathbun and Nordin, 1971; Laronne and Carson,
1976). Distance of travel and patterns of deposition may then be inferred. Some of the
strategies for locating the rocks afterward include painting the stones with fluorescent
waterproof paint, inserting a magnet in the rocks for later retrieval with a magnetometer
(Butler, 1977; Hassan, 1990; Gintz et al., 1996; Gottesfeld, 1998), or exposing the rocks
to radioactive treatment (Sobocinsky et al, 1990)
Chacho et al. (1989) presented a methodology whereby radio transmitters were
implanted into rocks in an extraordinarily active glacier-fed river, and then tracked for
kilometers. The monitored particles typically traveled over 1500 min 8 days time.
Similarly, Schmidt and Ergenzinger, (1992) used their Pebble Transmitter System
(PETSY) to track to movement of grains along the Lainbach, a step-pool mountain river
in southern Germany.

Morphological Estimates
Comparing volumes of aggradation and degradation in a channel by repetitive
cross-section or topographic surveys provides some insight into the rates of sediment
exchange within a river system. Griffiths (1979) was one of the first to use this method,
taking cross sections every 1.6 km along a 50 km reach of the braided Waimakariri River
in New Zealand. The density of surveyed sections employed depends on the purpose and
scale of the investigation and practical considerations. For instance, Poirier (2000)
developed a methodology for fine-scale topographic measurements and volume
estimation, taking repetitive surveys at an average density of 4 points per square meter
along several 50 m channel reaches. Lane et al. (1994) recommend at least 3.5 points per
m 2 to detect the subtler morphological changes within a stream reach.
The transport estimates provided by repetitive surveys are necessarily a minimum,
since several cycles of scour and deposition may have occurred throughout the course of
a flood. The final volumetric measurements are therefore the net transport rate. Loci of
erosion and deposition may be matched, providing some indication of transport distance

21
and volumes. When this is not feasible, transport rates may be estimated using reachaverage erosion or deposition and a mean transfer distance (Ashmore and Church, 1999).

' Surveys
Sonar
A variant of the morphological method worth mentioning is the use of sonar to
measure instantaneous elevation changes in a portion of streambed. This is a particularly
useful method for monitoring the passage oflarger-scale bedforms. This method was
successfully used by Dinehart (1989, 1992) at the North Fork Toutle River, Washington.
The channel had a bankfull width of 60m with peak flows up to 330 m 3/s. Coarse gravel
dunes, up to 40 em high, were observed migrating under a fixed sonar observation point
during a flood. A multitransducer sonar altimeter was suspended on a boom, and held 1
meter above the streambed near the center of the channel thalweg.

Scour Chains
Scour chains have been employed in a variety of studies to determine the total
scour and fill occurring during a flow event (Colby, 1964; Leopold eta/., 1966; Carling,
1987 and Hassan, 1990). A length of chain, fastened to an anchor, is inserted through a
steel rod and driven into the streambed. The rod sleeve is then removed and a length of
chain lies along the bed. After a scour and fill event, the chain is located beneath the fresh
deposition with a metal detector. The depth of scour can be inferred from the distance
below the initial datum that the chain length is found (see Laronne et al. 1994).

1.3 Stuart-Takla Bedload Studies

Several studies (Gottesfeld, 1998; Poirier, 2000) have been undertaken in the last
8 years to establish the rates and magnitude ofbedload movement in two streams within
the Stuart-Takla study area. Forfar and O'Ne-ell Creeks (see Figure 3) were chosen as
suitable sites to compare the effects of flooding and salmon spawning activity on bedload

22
movement. Both creeks see an average returning population of approximately 12,000
salmon each year during the Early Stuart sockeye run.
Studies with over 1000 tracer stones were undertaken from 1992 to 1998
(Gottesfeld, 1998). Some 13,000 recoveries of tagged clasts were made after flood events
or sockeye redd excavation. The results from this study posited that the amounts of
bedload moved by spawning salmon bioturbation are similar to those of floods, although
sockeye bioturbation of the streambed did not move the rocks as far as flood events.
A series of intensive total station surveys of stream reaches clearly showed the
annual pattern of stream bedform change (Poirier, 2000). Computer-generated digital
elevation models illustrated the morphological effects that each event exerted on the
streambed, i.e. linearized features after a flood, and more prominent, hummocky features
after the salmon had overturned the gravel substrate.
From these studies some understanding of the magnitude ofbedload movement,
the distance of bedload transport and how it affects channel morphology was gained, but
there was still a lack of information on the timing of the gravel movement. To this end,
we developed a device that records the timing and magnetic intensity of individual
particles passing the detector. In addition this device may offer some indication of how
much gravel is passing a cross-sectional line on the bed.

1.4 New Bedload Sampling Technology
In light of the many inherent drawbacks of existing bedload sampling

methodologies, we have sought to develop a system that will overcome some of these
difficulties. The UNBC Bedload Movement Detector has been developed as an
improvement on in-situ magnetic bedload detection technology, using a series ofhighly
sensitive magnetic sensors, and an industrial data-acquisition system. By placing a row of
these sensors across a streambed, flush with the gravel-bed surface, we will be able to

23
monitor the passage of sediment by recording the voltage signals from the sensors. This
method provides several advantages over the methods previously discussed.

•

It will overcome the numerous concerns about a sampling regimen: how often to
sample, how long to sample, what parts of the creek to sample. The record is
virtually continuous, across the channel.

•

There will be little interference with the water flows or the sediment stream. The
device will have negligible effect on the passing material, and it may be raised or
lowered to accommodate the current bed elevation.

•

Monitoring will provide an instantaneous particle count, and some information
about particle size. The digital data record from flood events provides detailed
information in the spatial, temporal and frequency domains, providing significant
opportunities for analysis. Sediment passage can be observed on the scale of
1/1OOs of a second, to several weeks.

•

The device will have the potential of remote operation. With a triggering
mechanism installed, and sufficient power resources, the device can operate
automatically.

•

No nets, baskets, pits or large installations. The BMD is automated for the most
part, and requires no retrieval of samples. The device is not particularly portable it is intended as a semi-permanent apparatus. However, there is much less
alteration of the channel and construction involved than an installation such as a
vortex tube trap or conveyor belt.

The following chapter will give a detailed description of the device, its installation
and operation.

Chapter 2 The UNBC Bedload Movement Detector

The Bedload Movement Detector was developed principally by Drs. Moustafa
Mohamed and Allen Gottesfeld at The University of Northern British Columbia. Dr.
Mohamed engineered the coil sensors; Dr. Gottesfeld designed the housing unit, wiring
and mounts for the device. Several students (graduate and undergraduate) and technicians
assisted with the device construction and installation at various stages, notably Patrick
Hudson, Tim Dressel, Ronald Poirier, and Jeffrey Gottesfeld. The device was installed at
the study site in September 1997.

2.1 The Sensor
The sensor (Figure 9) consists of a copper-wound coil mounted within a torusshaped magnet. A soft iron casing confines the magnetic field of the sensor so that it is
not influenced by magnetic bodies beyond the diameter of its sensing face. The elements
are set and waterproofed within an epoxy resin. Positive and negative terminals from the
device extend from the rear of the sensor casing.
As a rock, or any ferrous body, passes over the face of the sensor, the magnetic
field of the sensor is distorted. This induces an electromotive force with a potential of 1 x
10-6 to 1 X 10-l VOltS. Jn a laboratory environment with proper VOltmeter equipment,

signals from most rocks can be recognized. The rugged and portable data acquisition
systems that are required in the field environment generate sufficient noise to mask weak
signals such that voltage responses from clasts on the lower end of this scale (i.e. up to
approx. 2 X 1o-4 V) are therefore indiscernible from static.

25

TOROIDAL MAGNET
PUTTY CO IL SEAT

STEEL CASING

CASING :8CM

COIL :

I. 3 CM .... -.. ~~

MAGNET :7CM

SENSOR WATERPROOFED WITH
EPOXY RESIN

Figure 9- Exploded view of the sensor (above), sensor dimensions and the assembled product.

Calibration experiments have determined that the response of the sensor drops off
markedly as the clast moves toward the sensor's edge. Figure 10 shows the corresponding
decline in signal strength as ferrous items are passed over points that are increasingly
distant from the center of the sensor face, at a fixed velocity. Some of the implications of
this will be discussed in Section 3.2.

26

~
Q)

en
c

&. 0.0400 ------~.-----------en

&!

...

Q)

Cl

ns
~ 0.0200 ----

::

' ::------'~---------------

2

3

4

5

6

Position on Sensor Face

Figure 10- Voltage response as ferrous items pass at various points on the sensor.

2.2 The Housing Unit

Cables from sensors

Sensors set in conduit
view)

Concrete footing
(buried in bank)

Figure 11 - Schematic diagram of the BMD installed in the creek bed

27
The sensors are set into a hollow aluminum beam (15 em x 15 em x 8.2 m) that is
constructed of 'l4 inch (0.64 em) aluminum stock (Figure 11 and 12). There are adjusting
mounts on each end of the beam, so that it can be raised or lowered to accommodate
channel scour or fill. The device is set such that the sensing surface is placed flush with
the streambed. Stream flow and sediment then pass over the device with minimal
disturbance of near-bed fluid dynamics and sediment transport.

Figure 12 - Photo of the study site in early August. The instrument shed is located on the east side of
the creek. The BMD is set within the streambed downstream of a group of mature sockeye
salmon. A log and wire-rope cable provided access to the stream cross-section at all stages.

The device was custom built to the dimensions of the creek at the study site. This
prototype spans a stream width of 8.4 meters. A similar design can be employed on most
streams, to a maximum practical installation size of 30 m in width. The prototype
contains 82 sensors, each 8 em in diameter, spaced 10 em from center to center. Wires
from each sensor run the length of the beam and out to an instrument enclosure adjacent

28
to the stream. The conductors are housed in five strands of telecommunications cable,
each one containing 50, 22-gauge insulated copper wires. At least 80 redundant wires
were left in the housing unit as supplementary connections in the event of any wiring
repairs or complications.

2.3 Data Acquisition
The instrument shed was constructed on the west bank of O'Ne-ell Creek adjacent
to the BMD. It houses the data acquisition system and has modest facilities for sleeping
and working. The cables from the BMD are routed to a protective enclosure where the
sensor wires are distributed to their respective channels in the analog-to-digital (A-D)
boxes.
In order to collect data from the array of sensors at a relatively high rate (100Hz
x 82 sensors) an industrial data acquisition system was acquired from Omega
Engineering, Inc. of Stamford, Connecticut. The InstruNet™ (# iNet-100HC) data
acquisition system consists of a 32-bit DSP PCMCIA card and a series of A-D boxes,
each with 16 analog inputs(+ 8 digital). The PCI bus controller contains a 32-bit
microprocessor with 256 Kb of RAM that manages the external network of the device.
All real time tasks are off-loaded to this processor so that the host computer is not
burdened with real-time issues. A 12V power adapter is also included. An optical
isolation unit provides some noise reduction from high frequency disturbances such as
the computer and J azTM drive operating nearby.
Each A-D box contains a signal conditioning amplifier for each channel, therefore
voltage input from the sensor terminal is translated to a digital value which may then be
recorded to a Jaz™ drive. Each channel has a programmable analog low-pass filter, as
well as triggering capabilities. Signals are passed through a low-pass Butterworth filter,
removing signals with a frequency greater than 40kHz. The filtered data are then
recorded to disk for more sophisticated signal processing. A Jaz™ disk system was used

29
for this application, although a more reliable system is being sought. A high capacity disk
drive system is required that can scan at least 20 Kbps, or 240 Mb per hour, without
interruption. New media are inserted every two or three hours, and the written files are
subsequently archived on recordable CDs for storage.
A portable data logger (StarLogger™) operates independently of the data
acquisition system, recording creek stage and suspended sediment data every minute.
Suspended sediment was measured using an optical backscatter (OBS) probe, measuring
turbidity in Jackson Turbidity Units (JTU). These data are downloaded and processed on
a weekly basis. During the Spring 1999 event, stage data were written to CD along with
the BMD files.

2.4 Power Supply
The power requirements for the operation of the BMD are substantial. With two
PC computers, two Jaz™ drives, data acquisition system, CD writer, data logger, and
lights, roughly 13.3 kilowatt hours of energy are required daily while the BMD is
running. To meet the needs of this system, three solar panels were acquired, as well as a
5500-watt generator. Three 12V telecommunications batteries were installed beneath the
cabin to store power from these sources, and discharge them to DC appliances directly,
and to AC appliances through inverters. Typically, the generator had to be run for several
hours each day to maintain the battery array potential at 12.5 volts. While the solar panels
provide up to 200 watts of power during sunny hours, many of the spring days were
overcast and sunlight crested the tree line for only a few hours on clear days.
As we shall see in the presentation of results in Section 4, the generator tended to
affect the signal acquisition by introducing low frequency (~ 10-40 Hz) noise to the
system. However, the functioning of the data acquisition system deteriorated noticeably
as the voltage potential dropped below 12 volts, so at some points it was necessary to use
the generator despite the interference it created in the BMD record.

30
2.5 The Study Site

The bedload movement detector was installed on O'Ne-ell Creek (also known as
Kynoch Creek), approximately 150 kilometers north ofFort St. James in central British
Columbia. It is a third-order tributary of Middle River; Middle River drains Takla
Lake, flowing southward to Stuart Lake, between the Omenica Mountains to the west and
the Hogem Range to the east. The O'Ne-ell Creek watershed (Figure 13) encompasses
about 68 km 2 . The bedload movement detector site is located roughly 200m upstream of
the Tsitsutl Creek confluence, where the drainage area is about 38.5 km 2•

Figure 13- The O'Ne-ell Creek watershed.

The watershed above the detector site is undisturbed by human activities. Large
amounts of woody debris, abundant riparian growth, and ample gravels in the bed make
this river an important spawning locale for sockeye salmon. Frequent inputs of large
woody debris create a forced pool-riffle morphology (Montgomery et al. 1995)
throughout much of the stream. The study reach has a gradient of 1.3%, and though there
is a large logjam 65 m upstream of the detector site, there is little woody debris adjacent
to the bedload movement detector. The device is on a relatively straight reach of about

31
100 meters. Nival discharge through the reach ranges up to about 16 m 3/s. The average
annual precipitation in the O'Ne-ell watershed is about 60 em. Rainfall peaks in June and
early July. Snowfall from November to April accounts for about 50% of the precipitation.
Summer season temperatures frequently exceed 20°; winter temperatures are frequently
below -20° (MacDonald eta/. 1992).
The upper reaches of the watershed are quite steep with alpine ridges up to
2000m. Outcrops ofbedrock along the stream contribute granitic, ultramafic and
metasedimentary rocks to the stream bedload. As the stream approaches its alluvial fan at
about 900m, the gradient becomes increasingly subdued.

2.5.1 Sediment Supply
Numerous processes upstream of the study site determine the particle size
distribution of the streambed and the material in transport. Contributions of tills and
glaciofluvial materials from gullies and scarps are common throughout the course of
O'Ne-ell Creek (Ryder, 1995). Colluvial inputs from bedrock outcrops a few kilometers
upstream of the study site are common. Logjams and woody debris have a great influence
on the patterns of gravel distribution as well (Rice, 1994). Bank failures occur in places
throughout the system - the failure of a section of stream bank above the BMD was
particularly noticeable in the record from the 1999 flood event.
Streambed sediments were sampled by collecting four bulk sieve samples and 12
cores of the streambed with a freeze-core sampler (see Thoms, 1992). The bulk samples
were of the surface 15 em of 1m2 plots, each weighing roughly 270 kg. The large
samples are necessary to provide accurate estimates of the coarsest fraction of the
streambed sediment (Church eta/., 1987). The freeze core samples are 30 em deep, and
weighed roughly 17 kg apiece. They were split into lower and upper 15 em segments.
3

The volume of each freeze core segment is approximately 0.01 m . The surface 1-grain
layer was not segregated in either sampling methodology.

32

100
90
80
c:

70

IU

.r.

...

60

u:::

50

...0

40

1Q)

c:

c
Q)
Q)

11.

30
20
10
0
-4.0

-2.0

0.0

2.0

4.0

6.0

8.0

Grain Size (psi)

Figure 14- Grain size distribution from four bulk sieve samples. Three samples were excavated to a
depth of 15 em and one (represented by a dashed line) went to 25 em.
100
90
80
c:

70

IU

.r.

...

60

u:::
"E
Q)

50

1Q)

c:

...0 40
Q)

11.

30
20
10
0
-4 .0

-2.0

0.0

2.0

4.0

6.0

8.0

Grain Size (psi)

Figure 15- Grain size distribution from freeze-core samples. Dashed lines indicate subsurface
samples (below 15 em) and solid lines indicate surface material.

33
Bed material ranges from coarse sand to small boulders with diameters up to 30
em. Sediment finer than 1 mm makes up 1% to 10% of the streambed. Large pebbles and
cobbles are the most abundant size material. The D 50 of these sediments ranges from 3.5
to 6 psi ( t/1) (11 .3 to 64 mm). The average D 50 is about 42 mm. The D 90 ranges from 5.5
to 7.5t/J (45 .3 to 181 mm). The average value ofthe bulk samples is about 128 mm.
The pattern of textural variation on the streambed is complex. The bed material is
coarsest in the deeper portions of the channel that experience the highest flow velocities
such as along the thalweg and in pools. Areas removed from high velocities because of
shallow depths or obstructions can accumulate finer material. The pattern of average
grain sizes for the relatively linear portion of O'Ne-ell Creek above the BMD is shown in
Figure 16.

FjQe

~

(5 · '30 mm)

Figure 16- Texture of the stream bed at the study site, based on a 1998 topographic survey.

34
2.5.2 Lithologies

The lithologic or mineral character of the bed material varies a great deal, but is
dominated by granitic and metamorphic rock (Table 1). In part this diversity originates
from past regional glacial transport and deposition. Ferromagnetic minerals such as the
common iron oxide accessory minerals magnetite, ilmenite and hematite, and the iron
sulfide, pyrrhotite, are found in many of the clasts in O'Ne-ell Creek. Paramagnetic and
diamagnetic minerals such as pyroxene, biotite, chlorite and and quartz are found in
abundance, but seldom exert sufficient magnetic influence to trigger the sensors. For the
purposes of this study, we have assumed that the distribution of magnetic minerals
remains constant over all size ranges represented within the creek. It is, however, an
important task to quantify the overall percentage of rocks within the study area that
contain sufficient ferromagnetic material to elicit a response from a BMD sensor.
We have recorded the signals of 180 rocks in the laboratory ranging in mass from
0.5 to 1200g. This work and similar field experiments suggests that approximately 30%
of the rocks in O'Ne-ell Creek yield sufficient ferrous material to obtain a voltage

Percent

Lithology

Granodiorite (+ diorites)

26.0%

Cache Creek group metamorphics

25.6%

Metasedimentary

14.1%

Ultra Mafic

9.0%

Dacite

8.1%

Andesite

5.5%

Arenaceous (conglomerates, greywacke)

5.3%

Rhyolite

3.4%

Quartz

3.0%

Table 1 -Representative lithologies found in the bed material at O'Ne-ell Creek, N=234.

35
response greater than 1. 7 x 10-4 V (noise threshold of the data acquisition system) from a
sensor. Clasts that yield strong signals include a variety of lithologies: volcanic rocks,
granodioritic to gabbroic plutonic rocks, serpentinized ultramafic rocks, metasediments
and metavolcanics.
Because of the high sensitivity of the BMD sensors, we expect that this device
will operate effectively in most gravel bed streams, including most of the glaciated
regions on North America and Europe, which have a component of glacially transported
granitic and high-grade metamorphic rocks. Custer (1991) points out that naturally
magnetic clasts can be found in any of the andesitic volcanic terranes of Western North
America (Coastal and Cascade ranges), the Aleutian and Alaska Range in Alaska, the
Andes of South America, and the mountains ofthe Philippines, New Zealand, and Japan.
Basaltic terranes also yield magnetic material (e.g. Snake River Plain, Columbia Plateau,
Iceland, Deccan Plateau of India), and other metamorphic and intrusive terrains may also
contain particles that are sufficiently magnetic to be detectable.

Chapter 3 Time Series and Frequency Analysis
This section will examine some of the strategies that have been employed to
characterize and quantify continuous time series signals from the BMD records. Since all
the voltage information is stored in digital format, there are numerous strategies for
manipulating, processing and visualizing the data. Signals can be analyzed through time,
across the stream width, or in the frequency domain. They may be filtered with respect to
amplitude and frequency or convolved with other signals.
Although these procedures have met with a certain degree of success, there
remain some significant problems. At the simplest level these are due to the chaotic
nature of sediment movement in turbulent flow, especially at high sediment transport
rates. It is difficult to set the rules for particle counting when there seems to be so many
exceptions and permutations. Problems that are more complex have arisen from the
introduction of noise across the whole bandwidth of the system - due in part to the data
acquisition system itself, and also from the operation of the power generator.
The bulk of the signal processing work was accomplished using Lab View™
software, produced by National Instruments of Austin, Texas. The software uses a
graphical programming format, which greatly shortens and simplifies the time one might
otherwise spend generating thousands of lines of code for these applications. The tradeoff
is that many of the signal-processing routines are somewhat of a 'black box', and thus the
mathematical processing routines may only be customized to a certain degree.
The voltage data from the sensors are stored in an interlaced binary format. All
channels are written continuously to one file. This way the hard disk writing head stays in
one place on the platter when writing multiple channels, as opposed to moving back and
forth between separate files for each channel. A matrix operation must then sort the
written voltage values (v) and channels (c) so that VJCJ, v1c2, v1c3, v1c4 ... becomes VJCJ,

37
v2 c 1, v3c 1, v4 c 1 ... and so on. The values are stored as 32-bit floating-point data, in high
endian format.

3.1 Identification and Characterization of Signals and Frequencies
As a clast passes over the device, a signal similar to Figure 17 is recorded. The
signal amplitude, or intensity, is primarily a function of the magnetic properties of the
clast. The period of the signal is a function ofboth the length of the passing axis of the
particle, and the velocity with which it passes the sensing surface of the device. The

i

1.50E-02 . . - - - - - - - - - - - - - - - - - - ,
1.00E-02 - ---------~

~ 5.00E-03

::! O.OOE+OO

~E -5.00E-03

~-------.'

-l--------,-1-- --',k------------1
----

I

- --- - ----

~

-----~--

g -1 .00E-02 ------ ~

---- --- -~~ ......

. ..--

---------

-- _ _ _ _ _ _ _ _-1

-1 .50E-02 . ! . - - - - - - - - - - - - - - - - . . . . l

0

5

10

15

Time (1/100 s)

Figure 17- Time series x(t) of a sensor's response to a passing clast.

positive and negative excursions ofthe signal are related to the change in sign(+/-) for
magnetic flux with respect to time. Although the sensor has a symmetrical magnetic field,
the particles passing over the sensor often lack homogeneity in the distribution of
magnetic elements, or else pass in an irregular fashion, resulting in an unbalanced signal;
i.e., large positive component with little or no negative side, or vice versa. As we shall
see, this presents a problem in terms of detecting and counting events in a continuous
signal record x(t).
The identification of singular pulses in a time series is fairly straightforward; there
are threshold exceedance and peak detection algorithms that are able to detect exceptional
peaks within a record. Peak detection algorithms typically function by fitting a quadratic
polynomial to sequential groups of data points. For each peak, the quadratic fit is tested
against a threshold level, in this case the maximum voltage noise from the data

38
acquisition system (~ 2 x 104 V). Peaks with amplitudes lower than the threshold are
ignored. Peaks are detected only after approximately 4 data points have been processed
beyond the peak locations.
However, if the positive component ofthe signal is absent, a peak detection
algorithm will likely fail (and similarly for the negative component with a valley
detection algorithm). Although a routine might locate a peak followed later in time by a
valley, it is difficult to assert that they come from the same impulse. A useful
mathematical transform for representing a discrete sine pulse as a positive, singular wave
function is the Hilbert Signal Envelope. To envelope a signal x(t), the Hilbert Transform
of the signal is calculated h(t). This consists of the original signal with a 90° phase shift.
Sines are therefore transformed to cosines and vice versa. These signals are then
convolved to produce the envelope function A(t). 1
A(t) = ~ x(t) ® x(t) + h(t) ® h(t)

Figure 18 shows the envelope function A(t) ofthe sine wave from the first
example (Figure 17).

~
Q)

~

0

1.00E-02
5.00E-03

c.

:!l O.OOE+OO

0:::

-----~- ----- '---~~-------

-

~~~: :-- -~---.--- --.--- -~~..:: :: : :: ~

E -5.00E-03 + -- - - -------"..---=---f---#-- - - - - -----j
Q)

0

> -1 .00E-02 - ------~~----------------1 .50E-02

..~...-

0

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _____)

5

10

15

Time (1/100 s)

Figure 18- The signal trace x(t) (black) with its corresponding Hilbert transform h(t) (dashed) and
signal envelope A(t), shown in gray.

1

® denotes convolution

39

This transform resolves the ambiguity of missing positive or negative signal
components, and provides a signal that is easy to locate and quantify in terms of
frequency. Of course, as multiple particles cross a sensor in rapid succession, signals are
confounded and this signal processing strategy is no longer effective.
By taking the derivative of the Hilbert Transform, another sine wave is generated,
with balanced components. By finding the "zero-crossing" at either end of this wave, the
effective extents of the original signal may be determined. That is, we find the true
beginning of the signal as it begins its upward trajectory, and the final point of the
signal's negative excursion. This has been an effective strategy for isolating impulse
signals for frequency transforms. Otherwise, they are affected by larger signals within the
time series that is windowed and analyzed.
National Instruments' Joint Time-Frequency Analysis Toolkit (JTFA) TM provides
another useful tool for the analysis of sine wave impulses. The approach used here is to
divide the signal into several blocks that can be overlapped. Then, the Fourier

transform is applied to every individual block of data to indicate the frequency contents
of each. This is known as the Short Time Fourier Transform, or STFf. By squaring the
STFT, the STFT spectrogram is produced. The spectrogram represents the signal in
Time-Frequency or [tJ] space, as illustrated in Figure 19. This representation is
particularly useful for approximating the time of clast passage, and the associated signal
frequency.
The Gabor expansion represents a discrete signal s[i] as the weighted sum of the
frequency modulated and time shifted function h[ i] :

s[i] = L L cm,n h[i- mLW] ej
N-1

211

(3.1)

iiN

m n =O

where the Gabor coefficients Cm,n are computed by the STFT:
Cm .n

2

=STFJ1mLW,n]= Ls[i]y*[i-mLW]e-i nniiN (3.2)
i=O

40
where N denotes the number of frequency points, or "bins" and f:...M indicates the time
sampling interval.

Figure 19- A sine pulse presented in time and frequency space. The relative brightness (z-scale
magnitude) indicates the intensity of the frequencies. Note that the peak frequency response
of the signal is located at (31.6 + 2 =) 15.8 Hz, and occurs at 15/100 s.

An example of JTF A applied to a BMD signal is shown in Figure 20. Estimates of

the frequency content of each impulse can be measured along the Y -axis of the graph,
and the voltage intensity of each signal is indicated by the size and brightness of each
point. The first signal, for instance, has a frequency of 6 Hz. The final signal, on the right
hand portion of the chart, appears to have two frequency components - a different one for
the positive and negative excursions of the signal.

41

- 0.002

~

s

~

O.OOQ-+---........._._..,..._ ,,......_ _ _ _ _

..

.....

.........

~

.....

..

. . ........ ...... ... ....... ... ...... ..

~ ...... ..

~- .
1.0

2.0

3.0

4.0

5.0'

6.0

Time {seconds)

7.!)

8.0

9.0

Figure 20 - STFT spectrogram of a single sensor trace during a flood event.

This example is taken from a 'clean' section of the data. The majority of the data
shows varying degrees of background noise, which are particularly noticeable under
JTFA analysis.
While Joint Time Frequency Analysis is a powerful tool, the cumbersome
calculations involved with the JTF A Toolkit entail prohibitively long processing time. A
single 4 hour file requires over 36 hours of processor time on a 450 MHz Pentium
machine. In order to expedite processing at present, conventional Fast Fourier
Transform (FFT) methodology has been applied to the signals.

42

3.1.1. The Influences of Clast Velocity On a Signal
The time required for a unitary particle (say of 1 mm diameter) traveling at 1 m/s,
to cross the 0.07 m sensor face is .07s. The associated period of a sensor's output voltage
signal would be 14.1 Hz. The general formula for calculating the period for a given
particle moving across a BMD sensor is
1

(3.3)

f {x(t)} = (Dx + 0.07) x v

wheref{x(t)} is the frequency of the signal, Dx is the length in meters of the particle axis
nearest to its vector of motion (referred to hereafter as the passing axis), and vis the
particle velocity in meters per second. We assume that the passing axis of a clast rolling
along a streambed would be equivalent to the B or C axis.
Figure 21 illustrates the relationship between velocity and associated period for 5
different particle sizes. As the passing axis becomes longer, the size ofthe particle
becomes increasingly important in determining the signal duration. For coarse sand and

iii

Passing Axis

:I:

~ 20

s:::::

---------------------------~~~~-----

Q)

::I
C"

~

LL

----------------- ~~---------::: ........"'--- - - --=-

--=--------j

s:::::

.El

--.001 m
- - .01 m

- - .05m
--.1m
.15m

V)

0 ~~------ -------~--------- ------~
0.00

0.50

1.00

1.50

2.00

Grain Velocity (m/s)

Figure 21 -The relationship of velocity and signal frequency for clasts of different sizes.

43
granule particles the signal frequency is proportional to the velocity divided by the width
of the sensor face. As Dx becomes much larger (i.e. larger clasts), the length of the clast
becomes a greater influence on signal length, and the relation departs from this
proportionality.
To put this in a hydraulics perspective, Figure 22 and Figure 23 present the flow
parameters at the study site. Figure 22 shows the stage/discharge graph, relating depth of
flow with corresponding discharge.
Velocity measurements were taken throughout the 1999 flood event, and the
readings are shown in Figure 23. Two sets of measurements, taken at 0.6 and 0.2 of the
flow depth (measured from the bed), are displayed on the graph, and compared to
velocities predicted from the Universal Velocity Distribution Law (Dingman, 1984).
These velocity sampling elevations were erroneously selected (0.8 and 0.2, or 0.4 alone is
generally preferred), however the measurements are still valid. At high stage the water
velocities did not exceed 2.5 m/s, and the velocities in the bedload transport zone (lower
5 em) could not have exceeded 1.3 m/s (likely< 1.0 m/s). One would expect the fastest
moving particle in the transport zone to be moving at a velocity approaching that of the
water.

.

.-----------------------------------------------------~

-lS . ~----------------------~-----------------------------

3:

0

u::
o

~

y = 0.1963Ln(x) + 0.2137

.

---------~~----------------~~ ~
. ~~~----------------~

.

-------------------------~ --------------------------

C1l

c

10

100
3

Discharge (m /s)

Figure 22- Stage/Discharge rating graph for O'Ne-ell Creek 1950, June 1999.

44

0.5 . . . - - - - - - --

- - - - - -- - - -- +--------,

0.4 +-- - - - - - - - - --

I

- . - - - - -+--

- - -----1

-

UV Dist (.68)

-

- uv Dist (.23)

•

May 17th

• May 19th

• May 22nd
• May 23rd

••

0.5

1
1.5
2
2.5
Mean thalweg water velocity (m/s)

3

May 25th

•

June 14th

• June 5th
• June 9th
• June 11th

- - - - Region of BMD Detection - -

0

•

3.5

Figure 23- Velocity profile at O'Ne-ell1950, May-June, 1999. The exponential lines indicate the
Universal Velocity Distribution Law for channel geometries based on low stage (May 17th0.23m) and high stage (June 14th- 0.68m). Four velocity measurements are indicated for
each measurement period- two sets taken at 0.6 and 0.2 depth in the area of maximum flow.

From this reasoning, the shortest frequency signal we could expect to see in the
BMD record from a grain passing the full width of a sensor would come from a 1 mm
grain of sand traveling at 1.0 rn!s, which would have an associated period of 14.1 Hz.
This is substantially less than the Nyquist frequency ofthe sampling rate (104.2 -:- 2 =
52.1) and less than the suggested threshold of0.2 ofthe sampling frequency, which
would be 20.8 Hz. Thus, a sampling frequency of 104.2 Hz allows plenty ofbandwidth
for digitizing signals from the passage of even the fastest particles.
3.1.2 The Influences of Clast Size On a Signal
Figure 24 shows the resultant signals from three trials of a single mafic rock
passing over the sensor on a pendulum. The signals have been enveloped with the Hilbert

45
transform, as discussed above. As the velocity of the particle diminishes, the signal
intensity is attenuated and the signal period increases. The area under each curve is
roughly similar, but clearly the associated period changes substantially (see Appendix C).

- 1 . 4 m/s

-

GO

ID

C\1
"'

Ol
"'

CD

(")

M
'<t

O.Bmls
0.4m/s

0
I!)

Time (1/100 s)

Figure 24 -Signal amplitude diminishes and period increases with lower clast velocities

In Figure 25, results are shown from the passage 8 particles, all of the same
lithology and passed over a sensor at the same velocity. With the increasing length of the
particles' passing axis, the period of the signal broadens correspondingly.

1.40E-02

Passing
Axis

1.20E-02

~
Cll
Ul

--110 mm
104mm
99mm
- - 95mm
--62mm
55mm
- -47mm
44mm

1.00E-02

1:

0
c. B.OOE-03

Ul
Cll

0::

>
Cll

6.00E-03

C)

ca

0

4.00E-03
2.00E-03
O.OOE+OO
(")

I!)

1'--

Ol

Time (1/100 s)

Figure 25- A comparison of signal envelopes for different sized clasts.

46

The amplitude of the signal is dependent on the strength of the magnetic field of
the clast and its velocity. In Figure 25, the sampled particles are all ultramafic dunites and
consequently have similar magnetic characteristics. The amplitudes of the sensor
responses are roughly proportional to the third power of their lengths. The volumes of all
the rocks may not be proportional and thus we see at least one exception, in the rank of
the 95 mm particle; it has a lower amplitude value than the 62 mm sample.
Figure 26 shows a graph of 71 rocks passed over the sensors at a constant velocity
of 2.25 m/s. By measuring the duration of the sensor's voltage response for each particle,
an "apparent" size of the passing clast is calculated. Despite much experimental error the
slope of the fitted curve (1.014) is very close to 1. Although some stones have complex
magnetic properties with varying magnetic strengths across the clast, there is a direct
relationship between the size of the clast and the duration of the voltage signal induced
by its passage over the sensor

120

E'

•

100

.§.
.c
....

C)

1:
Cll

• •

80

....1

.
"C
Cll

....ca

E

••
•
•• •
•

60

Ill

w

40

y =1.0144 X
R2 =0.6823

20

20

40

60

80

100

120

Actual Length (mm)

Figure 26 - A comparison of clast size versus the "apparent" clast size calculated from analysis of the
voltage signal from a sensor.

47

Thus, the frequency of the signal from a passing clast is a function of both its size
and velocity. Since we can neither assume that the size of the particles is constant, nor
that their velocities are constant, we must develop some statistical algorithms to extract
pertinent information from the data. Given that we know the size distribution of the bed
material and can measure water velocities, the distribution of signal durations in a given
period will likely inform us about the range of velocities and/or sizes of clasts in motion.
Furthermore, when particles in the cobble range are entrained, we should expect that at
least two adjacent sensors would peak at the same instant a cobble traverses the device,
leaving another clue as to the range of sizes in motion.

3.2 Random Passage of Sediment
As stated earlier, the BMD signals are the product of the size and velocity of
material moving over a sensor. While we can make some reasonable assumptions about
both of these variables, the margin of error remains at least an order of magnitude.
There is also the matter of the trajectory of particles passing over the device.
Particles that wobble in their course, or pass between sensors will not provide a
representative signal. Figure 27 illustrates the results from laboratory test of the voltage
response from a BMD sensor (see Figure 10) as various magnetic items are passed over
the sensor at a fixed velocity. Although the sensor lane is 10 em wide (i.e. the sensors are
spaced every 10 em) and the sensor face is 8 em wide, the area of highest response is
only 6 em wide. Thus, the probability of a particle passing over the device and registering
a substantial signal is only about 60%.

48

100.0
Length of
Sensor Face

Q)

Ill

c
0

c.

., •

75.0

Ill
Q)

0::

>-

:t:

,

~~

,

....

' •., '

"\

50.0
I

....:c 25.0
Ill

Most Sensitive
Region

I

Q)

I

Magnet
Nail Tip

'
'•

•

~
0

•Sensor Face

•
'•

0.0
0

2

4

6

8

10

Width of Sensor Lane

Figure 27 - The probability function facing a particle as it approaches the device. The dashed line
indicates the possible chord length of its transect across a sensor. The relative sensitivity of
the sensor is outlined by the experiments introduced in Section 2.1.

Another substantial problem for deciphering signals is the phenomenon of clusters
and sweep events. In these cases, distinguishing individual clasts becomes a difficult task.
Particles are frequently swept up in events that trigger up to half the device in one or two
hundredths of a second. This signal-processing problem is small, however, compared to
sections recording the passage of gravel sheets, when the transport record is all but static
(the sensors are overwhelmed by multiple transport events). One would expect, though,
that the total voltage recorded might bear a positive relationship with the total volume of
sediment passing at this point.
Along with signals that can be assigned to a particle length and velocity, there are
numerous high frequency signals in the record. These are signals greater than 12Hz.
These might arise from three circumstances:
•

Passing grains not traversing the central part of a sensor. This is by far the
most common reason, and presents the greatest problem in terms of
estimating the total transport.

49

•

Particles can move faster than the local mean fluid thrust at the bed.
Instantaneous velocities may be as much as double the mean velocity due
to high-speed sweeps.

•

Noise from the system (considerable in some instances) generates discrete
random signals that vary over the breadth of the BMD record's frequency
spectrum. Extraneous signals from system noise will be the easiest
problem to eliminate, with an improved power source and better data
acquisition equipment.

As shall be discussed in the next section, longer signals arise from larger particles,
but turbulence will also slow particles, and drag resistance increases with particle
diameter. It is for these reasons that size estimates from the BMD sensor signals are
considered gross estimates only.

50

Chapter 4 Analysis of the Flood Data Record
Two nival floods have been recorded to date with the BMD. The floods of 1998
and 1999 show considerable contrast in their timing, length, and sediment-moving
capability. DFO data show that record low flows were reported in the Fraser River during
the spring of 1998, while the flow volumes of 1999 were third highest on record (DFO,
2000). This is certainly consistent with the data presented here.
During the peak of sediment transport activity, the device may register over
200,000 signals per hour. The simplest method of depicting the relative amount of
bedload transport activity across the channel throughout a flood event is to show the total
voltage recorded at each sensor integrated over time. The voltage readings from the BMD
are taken to be a reasonable proxy for the rate of sediment flux. It is emphasized that
although this is not entirely accurate, it is reasonable for illustrative purposes. If it is
assumed that the distribution of magnetic particles varies uniformly over time, then
statistically these summations are a good approximation of the bedload activity.
Much of the initial data analysis has consisted of summing the many hours and
days of data into 30-second voltage sums for each sensor. Sensor 1 is on the west bank
and sensor 82 on the east. Charts ofthe total transport activity over time are then plotted
onto graphs similar to Figure 29. Obvious spatial and temporal trends in transport activity
then become evident, and the active periods of transport can be identified and further
analyzed.

51
·-------------- 30 seconds - Sensor# 1 ~~ -'~~ . .. ~- ~ '~
· .. ·l .
2 ~~..... .. . ~... .... ' ' . ~.. .. ·· · ·l
3 ~~ ' ' . ' ~ ' '- '.. . ..... ~ - ........~' ' · · · ·"2:
4 ~' .--. ' .. .. ~ .. ~' . '- ' '. '
· · · ·4
5 ~~~~~~~...~ .. . . ·4
6 ' ~. ~' ' ~~.. - ~.
' · · · ·L

CIO

0.10

Channel Width lSensar No.1
Figure 28- The BMD Record consists of summing the voltage activity in absolute (positive) units for
each channel in 30 second intervals. The resultant sums, showing the magnitude of activity
across the channel width, are shown in the chart on the right.

2 hrs--~--~
. .J

~~

::>

Channel
Width
ISensor No.J

Figure 29 - The sums for each 30-seconds interval are then plotted over the hours and days of the
flood event. Here a 2.5 hour section of data from 20:00 to 22:30 on June 11 is shown.

52

4. 1 Nival Flood, 1998
The nival flood event in 1998 took place between May 17 and May 27.
Intermittent transport activity was recorded for several days as the flood waters rose.
Scattered activity across the channel consisted primarily of high frequency signals that
are interpreted as clusters of sand and pebbles that were available for easy entrainment
and were swept downstream along with loose organic debris and litter. On May 25, after
a significant rain-on-snow event, a flood wave surged down the channel during the
evening. The stage was on the order of a one in three year flood, but water levels rose
very quickly. The great majority of sediment that moved in 1998 traveled within the
ensuing 24 hours.

May
20

May

0.9

18

May
22

May
24

0.8

I.r:. o.1

...g. 0.6

0

~
0

u::

0.5
0.4
0.3
0.2

0

'?.
0

0
0

c'i
....

0
0

0

0

'?.

....
N

0
0

0

0
0

N
....

0
0

0

0
0

N
....

0
0

0

0
0

c'i

....

0

'?.
0

0
0

c'i

....

0
0

0

0
0

c'i

....

0
0

0

0
0

N
....

0

'?.
0

0
0

c'i

....

0
0

0

0
0

N
....

Figure 30- Stage readings during the 1998 nival flood. The gauge was swamped during the peak
flows, and thus data from the night of May 25 have been reconstructed from readings taken
j ust upstream of the station. Flow subsided quickly over the next three days.

Figure 31 shows the BMD record from the peak of the flood event. Data were
recorded at an acquisition rate of 30 Hz per channel. They are presented in the diagram as
voltage values integrated over 30 seconds.
While shorter, more sporadic occurrences of activity could be seen before and
after this period, by far the bulk of the activity took place in the May 25-26 segment. As
the event began, most of the activity registered in the center-left bank section of the

53
channel. As the event progressed, approximately 80 percent of the channel width was
mobilized, until the activity shifted to the right hand side at about 12:30 AM on May 26.
Three large pulses can be discerned in the graph - one starting at midnight, the next at
about 8:45AM, and another, smaller one at about 4:45PM. In each case the onset of
activity is quite abrupt, and tapers off over the course of 8 hours or so. Gaps along the
length of the transport record probably indicate obstruction of sediment transport,
possibly due to boulders coming to rest upstream of the device.
A one-clast thick carpet of sediment 1.5m wide covered a section of the west part
of the channel, which is seen in the chart as a gray zone in the lower part of the strip
diagram. The low intensity is likely due to the increased distance of moving clasts from
the sensors. Brighter, high intensity patches represent movement of clasts when patches
of the covering layer were removed and/or unusually strong signals. At about 3:30PM
the cover was disrupted and normal signals resumed. At the end of the transport event,
another blanket of sediment on the device in the same area was cleared off, and the
device was raised approximately 12 em to accommodate the slight aggradation in that
section.

54

Figure 31 -Data from the peak of the 1998 flood event, May 25 and 26.

55
4.2 Nival Flood, 1999
The spring of 1999 was somewhat unusual. The accumulated snow pack was
above normal, and the spring was mostly cool, with a few warm spells. The snow melt
flow was spread over more than a month. When the weather did finally warm in early
June, the floodwaters rose over the course of a week. Since the flood was attenuated over
many days, we did not see a dramatic flood peak such as the 1998 event. The passage of
sediment was monitored continuously, but the record shows significant movement only
from June 9 to June 16. A wide range of flow conditions was observed over the course of
the flood, enabling us to identify a number of distinct patterns or phases of sediment
transport.
Figure 32 shows the hydrograph from this period, along with the total bedload
activity and suspended sediment from June 9 to June 18, 1999. The strong diurnal signal
is due to the higher altitude snowmelt events in the afternoon that create a flood wave,
passing by the study site some 6 to 8 hours later. The suspended sediment index rises
with discharge levels most days, although there is a subdued response on the night of the
13th, when the availability of fine sediment appears to decline following a period of
prolonged bedload activity.
From May 23 to June 8, the transporting ability of the creek was relatively low.
The bulk of the transport was sand-sized sediments, likely carried from storage in pools,
channel edges and bar margins. Organic debris, clusters of fine material and a few small
pebbles were also noted. By June 9 the stage rose to a point where larger clasts (up to 2
em) were transported over the device. Throughout the flood the stage remained at about
one halfbank full, never quite reaching the point of picking up structural stones, but
easily moving whatever small, loose material was available.

0:00

12:00

0:00

12:00

0:00

12:00
0:00

12 00
0:00

12:00

0:00

12:00

0:00

12:00

0:00

12:00

0:00

12:00

0:0

I

0

June 11

June 13

0 '

= jtM!J:t!:tl t ...... , ••

i
1

I

I

----~-

'

1111 Jll L.Wj il I

Suspended Sediment (JTU)

I

,;

June 17

I

I

-

-------~------~-------

Bedload Activity (Summed Voltage over 30 s)

----------------- -------~------~~~----~----- ~

June 9

0

I

I
I
I

I
Ill

Figure 32 - Stage, bedload activity and suspended sediment readings from the 1999 flood.

..till!

---------~----- --~----- --~------~

1uu .------------------------- -----------~------------~----------~------------ -----------~------------~----------~------------

I

5

10

OA

0. 5

12:00

.--------------------- ---------~--------~---------- ---------- ---------- --------------------- ---------

~

.

.

Stage Level (m)

57
O'Ne-ell creek has a remarkably low amount of fine sediment transport. Turbidity
values remained low throughout the flood. At the highest turbidity levels, the streambed
was visible (but yellowish), and between the turbidity episodes the streambed was clearly
visible.
Discharge through the study section reached a peak of8.6 m 3/s on June 15.
Instantaneous velocities in the creek were measured daily, and ranged up to 2.6 m/s. The
channel cross section at peak discharge is shown in Figure 33. The depth and velocity
measurements were taken roughly 30 em downstream ofthe BMD, every Yz meter at 0.2
and 0.6 depth. The relatively flat cross section of this spot was present before installation
of the BMD.
Left Bank

Right Bank

J

0.25m

0
I

2m
I

Figure 33- BMD Site Water Velocity Cross Section: contours based on maximum stage
measurements in 1999 show the variation in velocity across the channel width. The vertical
dimension is exaggerated threefold.

The following figure shows the 8 days of bedload activity throughout the 1999
nival event. Gaps in the record are due to periods of maintenance and technical problems.
Some portions of the record have been removed due to excessive noise levels. Most of
the record on Jun 16 was missed after a disk drive failure.
By June 12, particles with an a-axis larger than 6 em were being recovered from
the flow with a small hand-held basket sampler. As the stage rose during the afternoon
and evening of June 12, increasingly intense waves of sediment passed over the device.

58
There is a lapse in the sampling record from about midnight to 1 AM, but we then see the
activity continue past roughly 10 AM, when the amount of material in transit subsides.
There is a strong diurnal signal in the stage record and, quite clearly, the bedload
transport pattern responds to this trend. There is an evident hysteresis effect, wherein the
bedload discharge rate responds proportionately to the rising flow, but then continues at a
high rate even as the stage level diminishes.
Interestingly, even with the lesser flows on June 13 (see hydrograph, Figure 32)
the peak rates of bedload discharge were quite similar to the previous days' record. The
duration of the transport event was much shorter, however.

Figure 34 (following pages) -The BMD record from June 10 to June 17. The channel width is
measured from the west side of the channel to the east. File lengths vary from one half hour
to four hours. Gaps between the files are from maintenance and technical problems. Most of
the record on June 16 was missed after a disk drive failure.

59

June l 0

June ll
Generator
Interference

June 12

0·6
0·6
O·A

b«'

~:~ •"'
"'f'r)

').«- '

~ o«'

June 12

June 13

I

June 14

61

JLJne 15

JuJne 16

JlJne 17
o.1
0 ·6
0 •6

b«'

Stt"> ~ '
~'

'J.«'

~~

'

12:0Q,

62

4.3 Sediment Transport Patterns
The 1999 flood data were recorded at a higher rate of signal acquisition ( 104.2 Hz
per channel) than those in 1998 (30Hz), thus presenting a clearer view of transport
activity, and affording better signal processing resolution. As the data from the 1999
event were analyzed, at least four patterns or "modes" of transport became evident; these
are explored in detail here.
4.3.1 Marginal Transport

From May 23 to June 10, the creek discharge was close to the threshold of
competence. Only small amounts of organic debris, clusters of fine grains and pebbles

were transported. For a few evenings between May 30 and June 5, the stage rose to a
point where particles in the pebble range were transported, but it is mostly sand-sized
particles that are evident in the record (i.e. generally few signal frequencies < 14Hz).
Transport was generally sporadic, and when the discharge and turbulence increased,
particles showed a tendency to pass as clusters.

w

June 10, 23:00
Figure 35 - Marginal transport.

--------.~-

63

Figure 35 shows a sample of the record during this early phase of transport. The
majority of the transport activity was on the east side of the channel where flows were
slightly faster. The narrow band of activity on the west part of the channel is sandy, easily
transported material from a bar immediately upstream of the BMD. This portion of the
record is the easiest to analyze, since the events are discrete, and usually affect only one
sensor.
On June 9, a heavy rain fell for most of the morning. This contributed to the
snowmelt flow, which was probably accelerated by substantial melt in the higher reaches
of the basin. The stage rose to a point where larger clasts up to 2 em were transported
over the device. Two days later, temperatures rose to 25 C, and material began passing in
greater volumes.

4.3.2 Clustered Transport
As increasing numbers of particles were picked up and carried along the bed, fluid
turbulence and particle interaction caused grains to cluster together. There are many
points in the record where it becomes difficult to distinguish among several particles
passing at once.
Figure 36 shows an example of transport during clustered transport when material
was moving faster, with particles ranging up to 2-3 em diameter. Some of the bundles of
sand and pebbles washed over the device in swaths up to 3m wide.

64

w

June 12, 04:00
Figure 36 - Clustered transport.

4.3.3 Continuous Transport
As the stage rose during the afternoon and evening of June 12, increasingly
intense waves of sediment passed over the device. Inter-granular collisions, and the sheer
volume of material in motion, blur the distinction of passing clusters. At times, the
material seemed to pass in steady sheets. The activity continued until about mid-day on
June 13, when the amount of material in transit subsided for a few hours and then
resumed and reached a maximum on the evening of June 13. There was clearly an
abundant supply of material entrained by the high flows, and the bedload discharge
continued even as the stage level dropped during the mid-day period of June 13.
With the ample availability of material for transport, the clustering effect becomes
less pronounced, and we observe the onset of pulses lasting for 10-20 minutes. In the
brief intervals between pulses, the transport rate drops almost to zero.

65

w

June 13, 19:00
Figure 37 - Transport at the height of the flood event.

4.3.4 Sand Streets
The final piece of the record (June 14th to June 1ih) was dominated by the
passage of sediment stringers, originating from a collapsed section of the bank
approximately 65 m upstream. A shift in a logjam diverted flow against the bank and
caused a large failure, sending roughly 4 m 3 of sandy material downstream. The transport
was confined to one or two narrow transport zones, 1-2 meters wide, which migrated
laterally by about one meter. Material was passing as sheets of finer material, as could be
observed from the catwalk at the site. Portions of the device became draped in sediment,
although this blanket of material moved laterally as the hours passed, indicating the
active nature of these "streets." The position of the transport streets is determined by
newly developed minor bars upstream of the device and likely by individual small
boulders. The width of the streets appears to vary from 10-20 em to over a meter. The
recorded signals indicate that the stringers were comprised mostly of sand and gravel
material.

66

w

June 15, 23:00
Figure 38 - Steady discharge of sand and gravel along the bed.

Figure 39 shows an excellent example ofthe onset of a gravel street, after a few
minutes of relative quiescence. Once the material was set in motion, it continued in an
uninterrupted stream for approximately 10 minutes. The width of this street is
approximately 1.2 m.
By June 181h, the waters subsided to lower levels, and the transport ofmaterial
gradually diminished to nothing. As the water level dropped, it became evident that a few

larger clasts immediately downstream of the device had been transported leaving
somewhat of a 'drop' at the site, roughly 15 em deep and 2m wide. A small hydraulic
jump was evident in this area during the last portion of the flood. By placing some
boulders, the scoured portion could be filled; however, this did highlight the necessity of
maintaining the site.

67
19:49:42

June 14, 1999

w
E
Figure 39- The onset of a gravel street, 1.2 m wide. The steady passage of material lasted for over 10
minutes. This diagram shows the first 90 seconds of the transport event.

Otherwise, there seemed to have been little geomorphic work done on the stream
throughout the 1999 nival flood. The flood was unusual in that it was so protracted and
that the peak discharge was relatively low. The 1998 flood had a short period but with
significantly higher flows . Larger floods (> 10 year return interval) will have a larger
transporting capacity with greater, and more stochastic, sedimentary inputs. We might
then expect to see roughly similar patterns of transport, with coarser material, higher rates
of flux and greater variability in the transport record. At least four distinct phases of
transport are discernible from the 1999 flood, although it can be assumed that with
different bed material composition and changing hydraulic conditions, other phases
would likely develop.

68

Chapter 5 Patterns of Bedload Transport

5.1 Discussion of Sediment Transport in O'Ne-e/1 Creek
A broad comparison of the 1998 and 1999 nival events shows that they both
exhibit clear pulsating or unsteady transport patterns. Transport may falter or cease
during flood stage, subject to channel conditions or sediment availability. The transport
zone clearly widens as more material is set into motion, and then tapers as sediment
transport declines. We observe that the transport zone may move considerably throughout
an event. These patterns of activity are discussed in sections 5.2, 5.3 and 5.4, below.
Although a number of problems have arisen applying the signal processing
algorithms (described in Chapter 3) to the analysis of the flood record, we have attempted
to estimate the total transport volume based on the number of particles counted as they
pass over the device. The assumptions and calculations involved with this estimate are
found in Section 5.5. The estimates obtained are then compared with existing transport
measurements taken at 3 study reaches within the Stuart-Takla watersheds.

5.2 Temporal Variation in Transport
In O'Ne-ell Creek we observe pulses on several time scales. Firstly, there are
daily pulses observed, as sediment moves in response to the diurnal pattern of snow melt.
As the melt water discharge rises in the late afternoon and peaks around midnight, rates
of sediment movement generally increase proportionally. Secondly, pulses are clearly
visible on a time-scale of tens of minutes. As the volume of sediment in transit reaches a
certain threshold of congestion, the transport becomes unsteady and moves in pulses with

69
intervals of 10 to 30 minutes. Finally, at the instantaneous time scale, sediment is
observed moving as clustered material.
The various types of short term, quasi-cyclic pulses discussed here are distinct
from larger multi-event cycles of degradation and aggradation known as 'slugs' ,
'translational waves' or 'megaform pulses' which exist on a timescale of months or years
(Nicholas eta!, 1995). These may occur as a result oflarge, discrete sediment inputs,
such as landslides and debris torrents, or from industrial activity such as mining. They are
also associated with the development of gravel bars and other morphological features
such as logjams.
With respect to variation in transport rates at the tens of minutes scale, this pattern
has been reported in the literature for almost 70 years. Work done in the early 1930s with
basket samplers was the first to provide clear indication that bedload transport occurs in
waves rather than as a steady discharge. Ehrenberger (1931) documented the oscillatory
behaviour of bedload discharge in the Danube and Inn rivers (Figure 40), taking
repetitive samples from a limited number of fixed stations. Similarly, Mulhofer's (1933)
and Nesper's (1937) data showed patterns that suggest a variable periodicity in sediment
discharge rates even at constant water flow .
o.o.----.-----.-------,
'E

:!:

.;

O.f

Figure 40 - Figure from Ehrenberger's bedload measurements, taken with a basket sampler on the
River Danube at Vienna National Bridge, 24 June 1931 suggest a pulsating movement of
bedload. Sampling time=100-300s. Mean transport rate= 0.236 kg/s/m. (Ehrenberger, 1931 Fig. 6). Taken from Gomez, 1991.

70
Results from more recently developed, direct sampling apparatus such as slot
traps with pressure pillows (Kuhnle, 1991; Harris and Richards, 1995), vortex tube traps
(Klingeman and Milhous, 1970; Taconni and Billi, 1989), and in situ magnetic sensor
devices (Reid et al., 1984; Ergenzinger and Custer, 1983; Bunte, 1996) have all clearly
illustrated the passage of distinct pulses of sediment on a number of different time scales.
Measurements taken by Bunte (1996) at Squaw Creek, Montana, are particularly clear,
showing the presence ofbedload pulses often followed by smaller 'secondary' and
'tertiary' pulses during a nival flood event (Figure 41).

Figure 41- Results from magnetic bedload detector at Squaw Creek, Montana (Bunte 1996).

Flume studies have amplified certain details of this phenomena, beginning with
the work ofSchoklitsch (1934) and Einstein (1937). Work by Hamamori (1962) sought to
model the frequency of bedload pulses based on the movement of coherent bedforms
such as ripples, dunes or bars. More recently, Iseya and Ikeda (1987), Kuhnle and
Southard (1988), Gomez et al. (1989), Hoey and Sutherland (1991), Seminara et al.
(1996) and others have all sought to model bedload waves in heterogeneous sediments

~ -~--- --~--~- --~--~- --~--~- --~--~- ~
o.o

4.0

a.o

12.0

1e.o

20.10

24.0 2a.o
'T'M: ~

32.0

3e.o

4o.o

44.0

418.0

no

!Us.o

so.o

Figure 42 -Total transport rate versus time. Measurements were taken in a gravel-bed flume,
showing transport fluctuations on the order of tens of minutes (Kuhnle and Southard, 1988).

71
based on contributing factors such as irregularities in streambed surface and variations in
sediment supply. These functions operate at a number of time scales and within a number
of different river types.
Part of the variable nature of sediment discharge is attributed to changes in the
rate of sediment supply during the flood event. Supply may vary in response to the scour
and fill cycles that take place upstream. Pulses may be caused by the periodic evacuation
of storage structures such as step-pools and logjams. Exogenous inputs, such as collapse
of the stream bank section noted above, are another typical source of sediment pulses.
Lateral and vertical grain sorting effects will also affect sediment supply. Gomez
(1983) points out that the development of an armour coat may eventually lead to the
exhaustion of mobile material. Kuhnle and Southard (1988) note that the episodic
dislodging of larger structural clasts from their stable position will cause variability in
transport rates as well, since all the imbricated clasts around them immediately become
available for transport. Large woody debris, bed armouring and bar development
undoubtedly play a role as well.
Braided rivers present a special case of pulsing sediment transport. Kang (1982)
and Southard et al. (1984) have both suggested that the basic processes ofbraiding
produce bedload pulses as a matter of course. Scour occurring in anabranch confluences
may also produce bedload pulses (Ashmore, 1988).
Perturbations or roughness elements along the streambed lead to instability of
flow and sediment transport. As sediment discharge increases, the transport rate becomes
out of phase with the existing bedforms. As sediment accumulates or disperses along the
channel, the bed surface is altered, in turn changing the flow patterns and shear stresses
exerted by water. As material is transferred from one reach to the next in greater and
greater volumes, traffic jams begin to occur.

72

Numerous field studies have related this pattern to the migration of distinct bedforms
such as dunes or sheets of sediment (Klingeman and Milhous (1970); Tacconi and Billi
(1987), and Gomez et al. (1983, 1989, 1997), although such coherent forms are not often
observed in coarse gravel systems such as O'Ne-ell Creek. Whiting et al. (1988)
observed the transport of fine gravels, under conditions where the bed could be clearly
seen. They noted that the bed material was organized into waves with distinct coarse
fronts, traveling as a "bedload sheet" or moving carpet. The sheets are coarsest at the
leading edge, the length of which is much greater than the height, and the height of which
is less than three coarse-grain diameters. This spatial and temporal organization of
moving sheets was further analyzed by Seminara et al. (1996) in flumes, with a view to
understanding the sorting function they perform in heterogeneous sediments as they
migrate along the channel. These sheets usually develop in streams with sand and fine
gravels.

5.3 Pulsations in the 1999 transport record
Figure 43 shows a typical data record, this one recorded on June 15, which
exemplifies the pulsing nature ofthe bedload transport in O'Ne-ell Creek. The peak of
the hydrograph occurs at about 23 :00 (off the graph). The lower portion of the graph is
the output from a single sensor, located 3.4m from right ban1c This roughly corresponds
to the locus of transport for this period, and gives a good indication of the general pattern
of movement. Pulses seem to occur every 5-20 minutes
The upper portion of the graph shows the spatial variation of the transport
intensity. The stream width is measured from the east (right) bank of the study site. The
active transport zone is mostly confined to a 1 m lane, although activity can be seen along
the whole length of the device.

73
.

------

-------

-------.------.-------

-------.-----~~---

2.
~

g

"0
Q)

E

E
:J

(f)

18:00

19:00

20:00

21:00

Figure 43 - Figure showing the pulsing nature of bedload transport. The upper portion of the graph
displays the width of the stream over the course of 4 hours on June 15. The lower half of the
graph is the voltage from a sensor located 3.4 meters from the right bank of the creek.

5.4 Latera/Instability- The Shifting Locus of Transport
The variability of transport conditions shown within the record includes not only
temporal variations in intensity, but also in the lateral migration of the locus of
transport. Distinct from the thalweg, which is defined as the thread of deepest water
(Leopold, 1994), the locus of transport is a term used to denote the portion of the channel
that accommodates the most material in transit (Church, 2000 personal communication).
Leopold (1994) points out that in many mountain streams, 80 percent of bedload moves
in 40 percent of the channel width. It is therefore important in modeling rates of transport
with empirical equations to understand where the majority of bedload is moving.
The following figure was calculated by summing up activity in each BMD file to
determine which sensor lane experienced the most activity. As one may observe, the
principal area of activity shifted considerably throughout the course of the flood. At first,

74

0
z 70 -- ~ ---~-----------------------------------------------------~

...

~

1:

60

-------~~~----~~---------------------------------------~

--

~

---------. --~~~~~-~~ ------------------------~~~--------

£"C 40

-----------------------~ -~------~----------------~------------

~

E

-...
~

20

-------------------------~~~~~

~~--~~--------~

---------------------------------------------------~------------

C/)10 +---------------------------------------------------------------l

0

-----------~--------~----------~----------~---------- -~

6/8/99 12:00

6/1 0/99 12:00

6/12/99 12:00

6/14/99 12:00

6/16/99 12:00

6/18/99 12:00

Figure 44 - The wandering locus of transport - the locus of transport is defined here as the zone of
maximum bedload activity throughout each data file from the 1999 Nival event. The locus of
transport moved substantially throughout the event, starting within roughly 1 m of the right
bank (see Figure 34, June 11th) and drifting to within 2m of the right bank (June 15-16).

E Bank
Figure 45 - The total activity across each sensor is summed for every BMD file. File lengths are
usually 3-4 hours. Summed values were divided by the length of the file to ensure compatible
numbers. The cross section profile (see Figure 33) adds the context of channel geometry.

75

most of the material was carried along the deeper, eastern portion of the creek, where
flows were swifter. As the flood progresses, we see the activity shifting towards the
center, and eventually over to the western side of the creek.
Figure 45 shows this activity trend in three dimensions. Summations from each
file are plotted along the length of the 8 day event, showing the relative amount of
activity at each sensor. The active zone changes from a relatively narrow band at the
beginning to a broader field of motion mid-event, which then constricts again to a small
lane of transport towards the end of the flood.

5.5 Transport Volumes- A Transport Estimate from the 1999 Nival Event
Ideally, the BMD would be used as a device to determine the instantaneous rate of
sediment transport across the stream width. With the proper calibration and improved
design, it is hoped that we might be able to collect information on the size of passing
particles, providing vital information about sedimentological processes occurring during
a flood event. This is difficult with the existing record because much of the record from
the 1999 event is contaminated with background interference from the electrical systems,
thus complicating much of the signal processing work. The high variability of magnetic
intensity of particles further complicates the modeling.
However, an order-of-magnitude estimate for the bulk transport over the course of
the event can be made by making a number of simplifying assumptions and using the
particle size distribution from the bed material bulk samples to estimate the total number
of grains in a single kilogram of bedload. The simplifying assumptions are:
•

The bedload material has the same particle size distribution as the bed material.
As flood stage rises in a gravel-bed river, researchers have noted that the number
of particles transported in each grain-size class may vary by nearly two orders of

76
magnitude (Parker, Klingeman and Maclean, 1982; Bunte, 1992). Particle size
distributions vary with changing hydraulic conditions, sediment availability and
clustering or hiding effects. However, since we are looking at the smaller end of
the scale (ie. particles smaller than 45 mm), this material is less subject to
selective transport, and we may assume that the bed material distribution is a
useful approximation.
•

Variation in magnetic intensity is uniform for various size clasts, and is similar in
character to the bed material. Although lithologies tend to disintegrate and
become sorted at different rates, we assume here that each lithologic group, with
its distinct magnetic properties, is uniformly distributed throughout all size ranges
ofbedload material at O'Ne-ell Creek. We may further assume that since most of
the transported material is endogenous to the creek channel, it has a magnetic
character that is statistically similar to the samples taken from the stream bed.

•

The volume ofparticles is similar to that ofspheres of the same caliber. The
larger clasts are subrounded to sub angular and generally have 70% of the volume
of equivalent spheres (n=442, St.Dev.=1 0.3%, see Figure 46). However, since the
sieving process selects particles by their b- and c- axes, it is likely that their
volume averages that of a sphere. With smaller sized materials, grains will
increasingly approach sphericity.

Figure 46- Distribution of grain geometries based on a sampling of material >50 mm The average
sphericity of particles is 0.7.

•

The density ofparticles is uniform (2. 65). This density value is for the most
common minerals such as feldspar and quartz, but there are a number of
lithologies in the stream that are denser such as andesite, and ultramafic rocks that
likely range closer to 3.0.

77

•

Only particles larger than 4 mm are able to induce a sufficiently large voltage
response to exceed the threshold of detection. Since the volumetric estimate is
based on the number of particles passing, this is certainly the most sensitive
variable in the equation. While 2 mm particles have been detected in the
laboratory, this case is generally rare; a threshold of 4 mm appears to be a better
approximation.

Figure 47- Size distribution of the theoretical bedload material used in the calculations. The
proportions are based on the 4 bulk samples of bed material taken at the site. The coarsest
material is 45 mm (5.5 t/J), and the assumed theshold of detection is 4 mm (2.0 t/J).

Based on random bedload sampling with a 10-inch (25 em) basket sampler
throughout the course ofthe event, the largest clast captured had an a-axis of6.61j1(100
mm). By far the majority of the material did not exceed the medium gravel size range
(>12 mm).
During periods of marginal transport, we observe particle counts at rates of up to
35 per second. At the peak of transport activity, rates reach up to 100 or 150 per second
(Figure 48). Theoretically, the detection algorithm can detect up to 600 events/ sec.
However, there may be times where the sensors are overwhelmed by multiple events. On
the other hand, there are a few points in the records where random noise interferes with
particle counting, and this may add up to 25% more particles than were actually passing.

78

'0

§
0

Q)

V'J

1

Cli

0.60

c. 1

t:n

~

ru

·;;; 1

:0

(f)

co

(t)

.50

!J..

gj

~

r

C'O

~

3"

'E
ru

a..

....0
Cii

.0

E
:::;

z

I

17:00:00
06/12

I

17:00:00
06/13

I

17:00:00
06/14

I

17:00:00
06/15

I

17:00:00
06/16

I

17:00:00
06/17

I I

20:00:00
06/18

Figure 48- Rates of particle transport (number of events per second) over the course of the flood.

Assuming that the device is only able to distinguish individual particles larger
than 4 mm in size, there are approximately 564 7 particles per kilogram of the bed
material (Table 2). Ofthose particles, we assume that only 30% are sufficiently magnetic
to initiate a voltage signal over the threshold of detection. Furthermore, since particles
must generally pass over the more sensitive 6 em central diameter of the sensor to
register (the sensors are spaced at 10 em intervals), there is a further exclusion of 40% of
the sample. Therefore, a correction factor of (0.30) x (0.60) = 0.18 is applied to this
value. The total number of detectable particles per kilogram is thus 1017.
Other complicating factors include cases where two signals arise on adjacent
sensors from a larger clast passing between them, and sections of mostly static, where
sand grains are passing in large volumes.

- l
79

% of Sample Proportional Density

Sieve

Single Grain

Single Grain

#of Grains

Size (mm)

Weight

Weight

kg/m3

Volume (m3)

Weight (kg)

in 1kg Sample

45.3

17.78

0.1778

2650

3.02E-05

8.01E-02

2.22

32.0

15.99

0.1599

2650

1.07E-05

2.82E-02

5.66

22.6

11.22

0.1122

2650

3.76E-06

9.97E-03

11.25

16.0

8.75

0.0875

2650

1.33E-06

3.53E-03

24.79

11.3

9.52

0.0952

2650

4.71E-07

1.25E-03

76.32

8.0

14.70

0.1470

2650

5.32E-08

1.41 E-04

1042.71

4.0

7.35

0.0735

2650

6.18-E-09

1.64E-05

4484.20

2.0

8.95

0.0895

2650

1.49E-09

3.94E-06

-Not

0.83

3.05

0.0305

2650

1.55E-10

4.11 E-07

Included-

0.5

2.69

0.0269

2650

6.54E-11

1.73E-07
5647.15

100.00%
x 18% detected =

1016.49

Table 2 -An estimate of the total number of grains per kilogram of bed material from O'Ne-ell
Creek, based on 4 bulk samples taken at the study site.

Over the course of 6 days (June 10 to June 15), some 14 x 106 particle events
were registered. Typical rates for a day were 1.7 x 105 on June 10 (low) and 4.2 x 106 on
June 13 (high). Making the gross assumption that the particle distributions for these days
were similar, and that 1016.5 events marked the passage of 1 kg of material, the volumes
break down as follows:

I

80
Date

# of Particles

Tons of Material

Unit Trans. Rate

June 10

171859

0.169

0.023 kg/m-s

June 11

1203123

1.184

0.164

June 12

1504929

1.480

0.206

June 13

4202050

4.134

0.574

June 14

2058716

2.025

0.281

June 15

3230275

3.178

0.441

June 16

1271613

1.251

------------

June 17

771466

0.759

0.057

Total

14414031

14.18

Table 3 - Daily particle counts, with associated bulk sediment transport estimates. The record from
June 161h is incomplete.

This would imply that at the height of transport activity, bedload was moving at a
rate of2.87 kg/s, or 0.574 kg/m-s, given an active transport zone width of 5.0m. In terms
of the volume of material passing during the flood, assuming an aggregate density of 1.8,
14.18 tonnes equals 7.88 m 3 of material passing over the device

5.5.1 Magnitude of Error
Although larger particles are not likely in motion during the early part of the
event, we are assuming a static distribution of material throughout. With marginal

amounts of material moving, the inclusion oflarger clasts (i.e. 45 mm) does affect the
volumetric estimate somewhat, even through they comprise such a small fraction of the
particle count. By changing the largest-sized clast assumed to be in motion to 22.6 mm,
for example, the volumetric estimate for the given period declines to 66% of our original
estimate.
The calculations are rather more sensitive to the decision of the minimum size
clast detectable. If a smaller size limit is chosen, the calculated volume decreases greatly.
If we had chosen a detection threshold of 0.833 mm instead of 4 mm for instance, the

81
transport estimate for the nival event would have been 781 kg- just over 5% of the
present weight estimate of 14 180 kg. With a threshold of 8 mm, however, the estimate
would be been 68 857 kg, almost 5 times the present value. The assumption of minimum
threshold of detection is thus of considerable importance in these calculations.

65

e

.§.

c:

.!:! 45
0

::::!!

c:

.II:
1.1

0

0:: 25

iii
Cll

...Cl
CQ

...J

5

100

1000

10000

100000

1999 Nival Event Transport Estimate (kg)

Figure 49 - The effect of two important parameters in the volumetric transport estimate. As the
assumed threshold of detection by a BMD sensor increases from 0.5 mm to 8 mm, the
estimate of transported material grows by several orders of magnitude. The assumptions
regarding the largest particle in motion (y-axis) affect the volumetric estimate as well - as the
assumed size moves from 8 mm to 64 mm, the estimate increases by a factor of 3.

Ideally, data from basket sampling should be paired with the BMD calculations so
that some estimate of the largest grain in motion can be made for as many periods of the
record as possible. We could then adjust the estimate to reflect the different conditions
encountered throughout the flood event.

5.5.2 Stream Efficiency
Table 4 shows some of the hydraulic parameters at the BMD site. The watersurface slope was not adequately measured, and is thus assumed to be similar to the bed
slope of0.013 . Unit stream power is defined as the product of the unit weight of water,
velocity, depth and slope (w = yVDS). The mean shear stress at the bed is defined as the

82
product of the unit weight of water depth, and slope (yDS =roN). The stream efficiency
is the unit bedload-transport rate (estimated in Table 3) divided by unit stream power
(Emmett, 1976).
It is interesting to note that Leopold and Emmett (1975) have developed a

relationship between bedload transport rates and stream power, based on the hypotheses
ofBagnold (1973). They suggest that as stream power increases, and sediment is set in
motion, the stream efficiency will approach a constant. The efficiency rating generally
found in coarse gravel rivers (32-64 mm) is 2-3 percent, barring any effects such as
armouring, that further diminish the transport capacity of the stream. These values are
close to the range of the present transport estimates, 0.2% to 4.8%.

Mean Flow

Unit Stream

Mean Shear

Stream

Velocity

Power

Stress

Efficiency

(m/s)

(kg/m-s)

2

(g/cm )

%

10.55

1.5

10.14

0.676

0.2316

0.55

11.73

1.6

11.44

0.715

1.4370

June 12

0.62

14.49

2.0

16.12

0.806

1.2756

June 13

0.57

12.52

1.6

11.856

0.741

4.8427

June 14

0.63

14.88

2.0

16.38

0.819

1.7173

I June 15

0.64

15.27

2.1

17.472

0.832

2.5261

June 16

-

-

-

-

-

-

June 17

0.44

7.41

1.4

8.01

0.572

0.7152

Stage

Q

(m)

(m /s)

June 10

0.52

June 11

Date

3

Table 4- Summary of hydraulic parameters at O'Ne-ell1950.

83
5. 6 Comparison with Existing Data

The best available data on rates of bedload transport in O'Ne-ell Creek have been
collected by Gottesfeld (1998) and Poirier (2000) (See Section 1.3). Based on their
measurements taken over the course of the 1996 and 1997 floods at O'Ne-ell Creek, we
are able to determine approximate volumes of material carried through a given reach. The
reaches chosen for comparison are O'Ne-ell1550 and O'Ne-ell9502 , roughly 400m and
1000m downstream ofthe BMD site (O'Ne-ell1950), respectively. Tsitsutl Creek joins
O'Ne-ell at approximately 1750, adding substantial flow to that ofO'Ne-ell. Combined
with a subdued gradient (>.005), somewhat finer material and different years of
measurement, we would expect the transport estimates from these reaches to vary
appreciably. Data from Forfar Creek (see Figure 3) is also shown for comparison. The
flood discharge ~

m 3/s) and slope (0.015) ofForfar 1545 are close to that ofthe O'Ne-

ell BMD site.
By measuring the typical travel distances of tracer sediment particles, Gottesfeld
was able to determine an average step length of clasts within the study reaches. The
smallest particles that were used as tracers were 50 mm (i.e. mostly larger than the gravel
and sand material that was measured by the BMD in 1999). The travel distances
measured are referred to later as 'Tracer Step Lengths'.
From topographic surveys of the stream bed, Poirier was able to determine loci of
deposition and erosion, corresponding closely to pool-riffle units within each study reach.
The lengths between these points are referred to in the calculations as 'Morphological
Step Lengths'. These agree substantially with the Tracer Step Lengths measured by
Gottesfeld. The volumes of cut and fill surveyed by Poirier represent the net amount of
material moved, i.e. the lowest possible estimate of transport.

2

Reach labels refer to the chainage distance measured from the mouth of the creek.

84
To convert these reach transport estimates to at-a-point transport estimates, the
total amount of material transported within the reach is divided by the number of steplengths. The 'tracer' and 'morphological' values offer slightly different estimated
transport volumes, but they are largely in agreement with the BMD estimates. If
anything, the 7.88 m 3 moved during the 1999 flood event appears slightly low. This may
be attributed to the lower magnitude flood that passed through the study reach in 1999.

O'Ne-e111550
Nival Flood '96
O'Ne-ell1550
Nival Flood '97
O'Ne-e11925
Nival Flood '96
O'Ne-ell 925
~

Flood '97

Forfar 1545
Nival Flood '96
Vorfar 1545
Nival Flood '97

Cut I Fill

Reach

#Pool/Riffle Units

# Tracer Steps

Transport Vol

Vol. (m 3)

Length

(Morph. Step Length)

(T Step Length)

Estimates (m 3)

8.82

30.87
68m
32.07

2.5

12.35

(19.42 m)

(26.78 m)

9.16
12.83

15.82
37m
11.21

2
(18.5 m)

7.91

-5.61
3.00

7.51
47m
11.60

3.5

2.5

4.8

1.56

(23 .5 m)

(9.65 m)

4.64
2.42

Table 5- Calculations of net transport rates through three reaches in the Stuart-Takla Watersheds.
Cut/Fill Volumes are taken from morphological estimates made by Poirier (2000).
Morphological Step Lengths and Tracer Step Lengths provide two estimates of transport
distances within the reach. Cut/Fill volumes are accordingly divided by the number of step
lengths in the reach, providing two estimates of transport volumes.

85

Chapter 6 Future Directions and Conclusion
There are two major components to be undertaken for the completion of the BMD
project. The first is the proper calibration of this device, likely in a flume with video
capabilities. We are presently only able to guess at the effectiveness of this device further work must be done toward generating accurate volumetric estimates. The second
is the device installation, which must be improved in a number of respects. This section
will examine some of the opportunities and barriers that face the project at this point.

6. 1 Flume Studies
The next stage for calibrating the BMD is to place it on the bed of a laboratory
flume, and pass known quantities of material over the sensors and relate this to the
volume of passing sediment. Church (2000, personal communication) has suggested that
the proper methodology to accomplish this is to cast sediment particles out of concrete
with varying amounts of magnetite mixed into them. By controlling the magnetic content
of the particles in this way, and measuring velocities by means of a video camera system,
all the variables should be accounted for.
Another important calibration experiment will be to place a conventional sampler
such as the Helley-Smith device downstream of the BMD during a transport event, and
compare the BMD record with the volumes of sediment retrieved from the flow.

6.2 Device Installation

Several improvements for the device have been suggested over the course of our
investigative work (1998-2000). Since the velocity of individual particles is far from
certain throughout the course of a flood event, a second row of sensors would be
extremely helpful in interpreting particle trajectory. This could be placed directly

86
downstream of the present row, with a 10 em distance from center to center. Data from
the two rows of sensors could be convolved to estimate a mean particle transport
velocity. The second row of sensors would double the number of channels in the data
acquisition system (and thus the data storage requirements), so some thought must be
given as to how to implement this.

It is presumed that ifthis type of bedload movement detector were to be installed
in a stream with clasts of uniform magnetic permeability, then the summed voltage
signals could be interpreted as transport mass per unit time.
The device has been described as somewhat of a 'sieve', since particles passing
between sensors may register much less of a signal than those passing directly over a
sensor (see section 2.1 ). What is required is a 'discrete' field such that a passing particle
would induce the same signal no matter which way it passed over the device.
As a compensatory measure, the second row of sensors could be placed in a
staggered position, such that a particle passing between sensors on one row would pass
directly over the next sensor downstream. At whichever point that it passes along the
device, the sum of the upstream and downstream signals may theoretically be constant.
Further additions to the study apparatus should include some device for
measuring the instantaneous shear stress close to the boundary layer, so that the subtle
turbulent fluctuations may be recorded in conjunction with flow velocities. A number of
strategies involving Pitot tubes or pressure transducers have been suggested, although the
implementation of these devices in the active transport zone presents some problems.

6.3 Automation and Signal Processing

We have not yet been able to capture a summer or fall flood, of which there are
usually few that will initiate bedload transport. It is hoped that some means of automating

87
the device will be developed in the future. We have experimented with a "knock
detector," a piezoelectric device that emits a voltage signal when triggered by vibration.
This will provide an acceptable cue to start recording, however, we have not been able to
arrange a failsafe computer system that will initialize itself and begin recording.
Ultimately, the BMD must be able to process the incoming data 'on the fly', and
provide a summarized record of a flood event, rather than spooling hundreds of gigabytes
of data that need to be processed. If the magnetic character of the stream bed sediments is
known, it may be possible to set the data acquisition software to the task of 'counting'
events and estimating their frequency. As better signal processing and spectral analyses
tools evolve, it will become easier to provide a statistical summary that need not be larger
than a gigabyte in total, even for several weeks of monitoring. The only remaining
logistical problem then is the power supply.
The power supply is certainly the weak link in the installation. In sections, the
data is so badly degraded from sources such as a running generator (or possibly a VGA
monitor) that the data are unsalvageable. The generator needs to be shielded from the
data acquisition's power supply, perhaps by connecting it to the last battery in the array,
rather than the same charge/discharge terminals as it is presently configured. Despite this
measure, it may still be quite difficult to isolate the data acquisition system from the
voltage irregularities of a solar system. As far as automating power supply, some means
of triggering the generator when power runs low must be found.

Conclusions
The UNBC BMD is a novel instrument for measuring the passage of bedload
material in a coarse gravel bed creek. With newly patented sensor technology devised at
the University ofNorthem British Columbia, and a high-speed data acquisition system,
the instantaneous observation of particle motion along an alluvial channel is now
possible. Initial data sets recorded from nival events in 1998 and 1999 show transport
dynamics that are consistent with a century of field observations, with a spatial and
temporal resolution never achieved before.
In light of existing technologies that have been used for measuring bedload, the
BMD presents several advantages. The device provides a continuous record, and samples
the full width of the stream. It is able to detect 30 percent of the material in transit, with
minimal interference of the sediment passage. It is expected that the sensors are capable
of detecting a higher percentage of the material - there is always the potential of filtering
the data in order to reduce the quantities detected, and thus simplify processing.
The BMD is a semi-permanent installation, requiring considerable effort to install
the footings, frame, cables, and wiring. The power requirements are substantial. An array
of solar panels and a backup generator are required to run the computers and data logging
equipment. The device still requires extensive calibration in order to better relate the
signals recorded to the actual quantities of sediment in transport. It is hoped that with a
better data acquisition system and faster computer systems, the device will be a viable
instrument for sediment transport studies.
The calibration work that has been performed shows that there exists a strong
relationship between the size of passing clasts and the period of the signal that they
generate. The confounding factor is velocity; as particles move more quickly, the period
is proportionally shortened. If an assumption is made about the likely transport velocity

89
of particles, there are a number of signal processing tools for calculating the frequency of
the recorded signals. A rough volumetric estimate may then be made.
The results from the data recorded over the course of two nival flood events offer
a wealth of information on bedload transport. Bedload sampling has never been achieved
at this resolution, and thus we are uniquely capable of verifying hypotheses regarding the
rates and distribution of sediment moving along the bed. The flood records clearly show
how the locus of transport varies from one side of the creek to the other, and that the
response of transport rates to water discharge can be quite variable. The semi-quantitative
data collected on pulsations and bedload streets raise new questions about the transport
processes in gravel bed rivers such as O'Ne-ell Creek.
It is anticipated that as the device is refined, gathers more and more data, and

becomes applied to different fluvial environments, a more comprehensive picture of
sediment transport will emerge.

References

Andrews, E.D. 1983. 'Entrainment of gravel from naturally sorted riverbed material. '
Geol. Soc. Am. Bull. 94. 1225-1231.
Andrews, E.D. and G. Parker. 1987. 'Formation of a coarse surface layer as the response
to gravel mobility.' In : C.R. Thorne, J.C. Bathurst and R.D. Hey (Eds). Sediment
Transport in Gravel Bed Rivers. Wiley, Chichester. 43-79.
Ashmore, P.E. 1988. 'Channel morphology and bedload pulses in braided, gravel-bed
streams', Geogr. Annlr. 73A(l), 37-52
Ashmore, P.E. and M.A. Church. 1999. 'Sediment transport and river morphology: a
paradigm for study.' In : Gravel Bed Rivers in the Environment. P. Klingeman, P.
Komar, W. Bradley (Eds). John Wiley & Sons, Chichester. 115-148.
Ashworth, P .J. and R.I. F erguson.1989. 'Size-selective entrainment of bed load in gravel
bed streams.' Water Resour. Res. 25 (4). 627-634.
Bagnold, R.A. 1966. 'An approach to the sediment transport problem from general
physics.' US Geol. Surv. Pap. 422-1.
Banziger, R. and H. Burch, 1990. Acoustic sensors (hydrophones) as indicators for bed
load transport in a mountain torrent. In: Hydrology in Mountainous Regions
(Proc. Lausanne Symp., August 1990) 207-214. IAHS Publ. no. 193.
Beaudry, P. 2000. 'Fine sediments in the Stuart-Takla Watersheds.' Stuart-TaklaFishForestry Interaction Study: 10 Year Report. In press
Beschta, R.L. 1987. 'Conceptual models of sediment transport in streams' . In: C.R.
Thorne, J.C. Bathurst and R.D. Hey (Eds). Sediment Transport in Gravel Bed
Rivers. Wiley, Chichester. 378-408.
Billi, P. and P. Tacconi. 1987. 'Bedload transport processes monitored at Virginio Creek
Measuring Station, Italy' . In : V. Gardiner (Ed). International Geomorphology.
Wiley, Chichester. 549-559.
Bridge, J.S. and S.J. Bennett. 1992. 'A model for the entrainment and transport of
sediment grains of mixed sizes, shapes and densities.' Water Resour. Res. 28 (2).
337-363.

91
Bunte, K. 1992. 'Particle number grain-size composition ofbedload in a mountain
stream.' In : Billi, P., Hey, R.D., Thome, C.R., and Tacconi, P., (Eds), Dynamics
of Gravel Bed Rivers. John Wiley & Sons, Chichester, 55-68.
Bunte, K. 1996. 'Analyses ofthe Temporal Variation of Coarse Bedload Transport and
Its Grain Size Distribution', General Technical Report RM-GTR-288, USDA
Forest Service, Fort Collins, 123 pp.
Butler, P.R. 1977. 'Movement of cobbles in a gravel-bed stream during a flood season.'
Geol. Soc. Am. Bull. 88. 1072-1074.
Carling, P.A., 1987. 'Bed stability in gravel streams with reference to stream regulation
and ecology.' In: K Richards (Ed). River Channels, Environment and Process.
Blackwell, Oxford, 272-294.
Carling, P.A., Williams, J.J., Kelsey, A., Glaister, M.S. and Orr, H.G. 1998. 'Coarse
bedload transport in a mountain river', Earth Surf Proc. Landf 23, 141-157.
Chacho, E.F., R.L. Burrows and W.W. Emmett. 1988. 'Detection of coarse sediment
movement using radio transmitters.' International Association for Hydraulic
Research XXIII Congress August 21-25, 1989, Ottawa, Canada. 1989. p B367B373 .
Church, M.A. 1978. 'Palaeohydrological reconstructions from a Holocene valley fill.' In:
Fluvial Sedimentology. A.D. Miall (Ed.) Mem. Can. Soc. Pet. Geol. 5. 743-772.
Church, M.A., D.G. McLean and J.F. Wolcott. 1987. 'River bed gravels: sampling and
analysis. ' In: C.R. Thome, J.C. Bathurst and R.D. Hey (Eds). Sediment Transport
in Gravel Bed Rivers. Wiley, Chichester. 43-79.
Church, M.A., Wolcott, J.F., and Fletcher, W.K. 1991. 'A test of equal mobility in fluvial
sediment transport: behavior of the sand fraction.' Water Resour. Res. 27. 29412951.
Church, M.A. and M.A. Hassan. 1998. 'Stabilizing self-organized structures in gravekbed stream channels: Field and experimental observations.' Wat. Resour. Res.
34(11). 3169-3179.
Colby, B.R., 1964. Scour and fill in sand bed rivers. US Geol. Surv. Pap. 462-D, 32 pp.
Crickmore, M.K. 1967. 'Measurement of sand transport in rivers with special reference to
tracer methods.' Sedimentology. 8. 175-228.
Custer, S.G. 1991. A review of natural gravel transport detection experiments at Squaw
Creek, Montana, 1981-1991. Unpublished manuscript. Dept. of Earth Sciences,
Montana State University, Bozeman, MT.

92

Custer, S.G., Bunte, K., Spieker, R., and Ergenzinger, P. 1986. 'Timing and location of
coarse bedload transport: Squaw Creek, Montana.' EOS, Trans. Am. Geophys.
Union. 67(44). 943.
Custer, S.G., Ergenzinger, P., Bugosh, N., and Anderson, B.C. 1987. 'Electromagnetic
detection of pebble transport in streams: a method for measurement of sediment
transport waves.' In Ethridge, F. and Flores, R. (Eds.) Recent Development in
Fluvial Sedimentology. Soc. Paleont. Mineralog. Sp. Pub. 39: 21-26.
Dingman, S.L. 1984. Fluvial Hydrology. W.H. Freeman & Co. N.Y.
Dietrich, W.E., J.W. Kirchner, H. Ikeda and F. Iseya. 1989. 'Sediment supply and the
development of the coarse surface layer in gravel bedded rivers.' Nature. 340.
215-217.
Dinehart, R.L. 1989. 'Dune migration in a steep, coarse-bedded stream.' Water Resour.
Res. 25(5). 911-923.
Dinehart, R.L. 1992. ' Evolution of coarse bed forms: field measurements at flood stage.'
Water Resour. Res. 28(10). 2667-2689.
Ehrenberger, R. 1931. 'Direkte geschiebemessungen an der Donau bei wein und deren
bisherige Ergebnisse [Direct bedload measurement on the Danube at Vienna and
their results to data]' . Vienna, Die Wasserwirtschaft, Issue 34, p.1-9. Translation
no.39-20, U.S. Waterways Expt. Sta., Viscksburg, Miss.
Einstein, H.A., 1937. 'Der Geschiebetrieb als Wahrscheinlichkeitsproblem [Bedload
transport as a probability problem]'. Mitt. Versuchsanst. Wasserbau Eidg. Tech.
Hocsch., Verlag Rascher, Zurich. [English Translation in: H.W. Shen (ed), 1972.
Sedimentation. Water Resour. Publ., Fort Collins, C1-C105].
Einstein, H.A. 1950. 'The bedload function for sediment transportation in open channel
flows' . US. Dept. ofAgriculture Tech. Bull. 1026. 71pp.
Emmett, W.W. 1976. 'Bedload transport in two large, gravel-bed rivers, Idaho and
Washington.' Proc. 3rd Fed. Int. Sed. Conf 4-101-4-114.
Emmett, W.W. 1980. 'A field calibration of the sediment-trapping characteristics of the
Helley-Smith bedload sampler.' US. Geol. Surv. Prof Pap. 1139. 28pp.
Ergenzinger, P., and J. Comady, 1982. 'A new tracer technique for measuring bedload in
natural channels.' Catena. 9. 76-93 .

93
Ergenzinger, P., and S. Custer, 1982. 'First experiences measuring coarse bedload
material transport with a magnetic device.' In: Mechanics of Sediment Transport.
Proc. Euromech. 156.223-227.
Ergenzinger, P and S.G. Custer. 1983. 'Determination ofbedload transport using
naturally magnetic tracers: first experiences at Squaw Creek, Gallatin County,
Montana.' Water Resour. Res. 19(1). 187-193.
Ergenzinger, P., C. de Jong, and G. Christaller. 1994. 'Interrelationships between bedload
transfer and river-bed adjustment in mountain rivers: an example from Squaw
Creek, Montana.' In: Kirby, M.J., (ed). Process Models and Theoretical
Geomorphology. John Wiley & Sons. Chichester. 141-158.
Fenton, J. and J.E. Abbott. 1977. 'Initial movement of grain on a stream bed: The effect
of relative protrusion.' Proc. R. Soc. London, Ser. A. 352. 523-537.
Ferguson, R.I., and Ashworth, P.J., 1992. 'Spatial patterns ofbedload transport and
channel change in braided and near-braided rivers.' in Billi, P., Hey, R.D.,
Thome, C.R., and Tacconi, P., (Eds), Dynamics of Gravel Bed Rivers. John
Wiley & Sons. Chichester. 477-495 .
Gintz, D., M.A. Hassan and K-H Schmidt. 1996. 'Frequency and magnitude ofbedload
transport in a mountain river.' Earth Surf Proc. Landf 21. 433-445.
Goff, J.R. and P.E. Ashmore. 1994. 'Gravel transport and morphological change in
braided Sunwapta River, Alberta, Canada.' Earth Surf Proc. Landf 19. 195-212.
Gomez, B. 1983. 'Temporal variations in bedload transport rates: the effect of
progressive bed armouring.' Earth Surf Proc. Landf 8. 41-54.
Gomez, B., R.L. Naff, and D. Hubbell. 1989. 'Temporal variations in bedload transport
rates associated with the migration of bedforms'. Earth Surf Proc. Landf 14.
135-156.
Gomez, B. 1991. 'Bedload transport' . Earth-Science Reviews. 31. 89-132.
Gomez, B. and Troutman, B.M., 1997. 'Evaluation of process errors in bed load sampling
using a dune model'. Water Resour. Res. 33(10). 2387-2398.
Gottesfeld, A.S., 1998. 'Bedload transport during sockeye redd excavation and by floods
in the Stuart-Takla experimental watersheds, British Columbia'. In Brewin, M.K.
and Monita, D.M.A. (tech. coords). Forest-fish conference: land management
practices affecting aquatic ecosystems. Proc. Forest-Fish Conf, May 1-4, 1996,
Calgary, Alberta. Nat. Resour. Can. For. Serv., North. For. Cent. Edmonton,
Alberta. Inf. Rep. NOR-X-356. 249-258.

94
Govi, M., F. Maraga and F. Moia. 1993. 'Seismic detectors for continuous bed load
monitoring in a gravel stream'. Hydr. Sci. J. 38 (2), 123-132.
Grass, 1971 . 'Structural features of turbulent flow over smooth and rough boundaries.' J
Fluid Mech. 50. 233-255.
Griffiths, G.A. 1979. 'Recent sedimentation history of the Waimakariri River, New
Zealand'. J Hydro!. NZ. 18. 6-28.
Hamamori, A. 1962. 'A theoretical investigation on the fluctuations ofbedload
transport.' Delft Hydraulics Laboratory Rept., R4.
Harris, T. and K.S. Richards. 1995. 'Design and calibration of a recording bedload trap.'
Earth Surf Proc. Landf 20. 711-720.
Hassan, M.A., 1990. 'Scour, fill and burial depth of coarse material in gravel bed
streams'. Earth Surf Processes Landf, 15. 341-162.
Hayward, J.A. and A.J. Sutherland. 1974. 'The Torlesse Stream vortex-tube sediment
trap.' J Hydro!. 13 (1). 41-53.
Helley, E.J. and W. Smith. 1971. 'Development and calibration of a pressure-difference
bedload sampler.' US. Geol. Surv. open-file report, 18 pp.
Hoey, T. and A.J. Sutherland. 1991. 'Channel morphology and bedload pulses in braided
rivers: a laboratory study.' Earth Surf Processes Landf 16. 447-462.
Hoey, T. 1992. 'Temporal variations in bedload transport rates and sediment storage in
gravel-bed rivers.' Prog. in Phys. Geog. 16 (3). 319-338.
Hollingshead, A.B. 1971. 'Sediment transport measurements in a gravel river'. Journal of
the Hydraulics Division, ASCE. Nov 1971, 97 (HYll), 1817-1834.
Hubbell, D.W. 1964. 'Apparatus and techniques for measuring bedload.' US. Geol. Surv.
Water-Supply Pap. 1748.
Hubbell, D.W. 1987. 'Bedload sampling and analysis.' In : C.R. Thome, J.C. Bathurst and
R.D. Hey (eds). Sediment Transport in Gravel Bed Rivers. John Wiley and Sons.
Chichester. 90-120.
Iseya, F. and H. Ikeda, 1987. 'Pulsations in bedload transport rates induced by a
longitudinal sediment sorting: a flume study using sand and gravel mixtures.'
Geogr. Annlr. 69A. 15-27.
Jackson, R.G. 1976. 'Sedimentological and fluid-dynamic implications of the turbulent
bursting phenomenon in geophysical flows.' J Fluid Mech. 77. 531-560.

95

Jackson, W.L. and R.L. Beschta. 1982. 'A model of two-phase bed load transport in an
Oregon Coast Range stream.' Earth Surf Proc. Landf 7. 517-527.
Kang, S. 1982 'Sediment transport in a small glacial stream: Hilda Creek, Alberta'. MSc
thesis. 265 pp. Univ. oflll. at Chicago.
Klingeman, P.C. 1981. 'Short-term considerations on river gravel supply.' In: SalmonSpawning Gravel: A Renewable Resource in the Pacific Northwest? Water
Research Center, Washington State University. Pullman, Washington. 123-139.
Klingeman, P.C. and R.T. Milhous. 1970. 'Oak Creek vortex bed load sampler.' Jih
Annual Pacific North-West Meeting of the Am. Geophys. Union, University of
Puget Sound. Tacoma, Washington. 17 pp.
Knighton, D. 1998. Fluvial Forms and Processes: A New Perspective. John Wiley and
Sons Inc. New York. 383 pp.
Kondolf, G.M. and W.V.G. Matthews. 1986. 'Transport of tracer gravels on a coastal
California river.' J Hydro!. 85. 265-280.
Kuhnle, R.A. and J.B. Southard. 1988. 'Bedload transport fluctuations in a gravel bed
laboratory channel.' Water Resour. Res. 24(2). 247-260.
Kuhnle, R.A., 1991 . 'Bedload transport on two small streams', Proc. of the 51h Fed.
Interagency Sed. Conf 1. 4-139-4-146.
Kuhnle, R.A., 1992.' Bedload transport during rising and falling stages on two small
streams'. Earth Surf Proc. Landf 17. 191-197.
Lane, S.N., K.S. Richards and H. Chandler. 1994. 'Development in the monitoring and
terrain modeling small-scale bed topography'. Earth Surf Proc. Landf 19. 349368
Lane, S.N., K.S. Richards and H. Chandler. 1995. 'Morphological estimation ofthe timeintegrated bed load transport rate.' Water Resour. Res. 31(3). 761-72.
Laronne, J.B. and M.A. Carson. 1976. 'Interrelationship between bed morphology and
bed material transport in a small, gravel-bed channel'. Sedimentology. 23. 67-85.
Laronne, J.B., D.N. Outhet, P.A. Carling and T.J. McCabe (1994). Technical note- scour
chain deployment in gravel bed rivers. Catena. 22(4). 299-306.
Leopold, L.B. 1994. A View of the River. Harvard University Press. Cambridge,
Massachsetts.

96
Leopold, L.B., W.W. Emmett, and R.M. Myrick, 1966. Fluvial Processes in
Geomorphology. Freeman. San Francisco, CA.
Leopold, L.B. and W.W. Emmett. 1976. 'Bedload measurements, East Fork River,
Wyoming.' Proc. Nat. Acad. Sci. 73. 1000-1004.
Macdonald, J.S., Scrivener, J.C., and Smith, G. 1992. 'The Stuart-Takla
Fisheries/Forestry Interaction project: study description and design'. Can. Tech.
Rep. Fish. Aquat. Sci. No. 1899.
Montgomery, D.R., J.M. Buffington, RD. Smith, K.M. Schmidt and G. Pess. 1995. 'Pool
spacing in forest channels.' Water Resour. Res. 31(4). 1097-1105.
MUhlhofer, L. 1933. 'Untersuchungen uber der Schwebstoffund Geschiebefurung des
Inn nachst Kirchbichl, Tirol' [Investigations into suspended load and bedload of
the River Inn, near Kirchbichl, Tirol]. Wasserwirtschaft, Heft 1-6.
Nesper, F. 1937.'Ergebnisse der Messungen uber die Geschiebe- und Schlammfuhrung
des Rheins an der Brugger Rheinbrucke [Results of bedload and silt movement
observations on the Rhine at the Brugg Bridge]. Schweiz. Bauz., Heft 100.
Nicholas, A.P., P.J. Ashworth, M.J. Kirkby, M.G. Macklin and T. Murray. 1995.
'Sediment slugs: large-scale fluctuations in fluvial sediment transport rates and
storage volumes.' Prog. Phys. Geog. 10 (4). 500-519.Parker, G. and P.C.
Klingeman. 1982. 'On why gravel bed streams are paved.' Water Resour. Res. 18. 14091423.
Nino, Y. and M.H. Garcia. 1996. 'Experiments on particle-turbulence interactions in the
near-wall region of an open channel flow: implications for sediment transport.' J
Fluid Mech. 326. 285-319.
Parker, G., P.C. Klingeman and D.L. McLean. 1982. 'Bedload and size distribution in
paved gravel bed streams.' J Hydraul. Eng. 108. 544-571.
Poirier, Ronald W. 2000. The effects of sockeye spawning on streambed morphology.
M.Sc. thesis. University of Northern British Columbia. Prince George, B.C.,
Canada (in preparation).
Powell, D.M. and Ashworth, P.J. 1995. 'Spatial pattern of flow competence and bedload
transport in a divided gravel bed river.' Water Resour. Res. 31(3). 741-752.
Qian, S. and D. Chen. 1996. Joint Time-Frequency Analysis. Prentice Hall, New York.
Rathbun, R.E. and C.F. Nordin. 1971. 'Tracer studies of sediment transport processes.'
Proc. ASCE, J Hydrau/. Div. 97 HY9. 1305-16.

97

Reid, I., J.T. Layman, and L.E. Frostick. 1980. 'The continuous measurement ofbedload
discharge.' J Hydraul. Res. 18(3). 243-249.
Reid, I., A.C. Brayshaw and L.E. Frostick.1984. 'An electromagnetic device for
automatic detection of bedload motion and its field applications.' Sedimentology.
31. 269-276.
Reid, I., L.E. Frostick, and J.T. Layman. 1985. 'The incidence and nature ofbedload
transport during flood flows in coarse grained alluvial channels.' Earth Surf
Proc. and Landf 10. 33-44.
Reid, I. and Fro stick, L.E. , 1994. 'Fundamental properties of fluids and their relation to
sediment transport processes. ' In: Pye, K. (Ed.), Sediment Transport and
Depositional Processes. Blackwell Scientific Publications. Boston. 25-60.
Rice, S.P. 1994. 'Towards a model of changes in bed material texture at the drainage
basin scale.' In : Kirkby, M.J. (Ed.). Process Models and Theoretical
Geomorphology. John Wiley and Sons. Chichester. 159-172.
Rouse, H.L., 1994. 'Measurement ofbedload gravel transport: the calibration of a selfgenerated noise system.' Earth Surf Proc. Landf 19. 789-800.
Ryder, J.M. 1995. 'Stuart-Takla watersheds: terrain and sediment sources. ' Res. Br., BC
Min. For. Victoria, BC. Work. Pap. 03/ 1995.
Sayre, W .W. and D.W. Hubbell, 1965. 'Transport and dispersion oflabeled bed material,
North Loup River, Nebraska. ' US. Geol. Surv. Prof Pap. 433-C.
Schmidt, K-H. and P. Ergenzinger. 1992. 'Bedload entrainment, travel lengths, step
lengths, rest periods-studied with passive (iron, magnetic) and active (radio) tracer
techniques.' Earth Surf Proc. Landf 17. 147-165.
Schoklitsch, A., 1934. Der geschiebetrieb und die geschiebefracht [Bedload and bedload
movement] . Wasserwirtschaft, Heft 4.
Seminara, G., M. Columbini and G. Parker. 1996. 'Nearly pure sorting waves and
formation of bedload sheets.' J Fluid Mech. 312. 253-278.
Shields, A. 1936. Anwendung der Aehnlichkeitsmechanik der Turbulenzforschung auf
die Geschiebebewegung, Mitt. Preuss. Versuchsanst. Wasserbau Schifjbau . 36. 524. (English translation by W.P. Ott and J.C. van Uchelen, 43 pp., U.S. Dept. of
Agric. Soil Conser. Serv. Coop. Lab., Calif. Inst. ofTechnol., Pasadena, 1950).

98
Sobocinsky, R.W., T.E. Cerling, S.J. Morrison and I.L. Larsen. 1990. 'Sediment transport
in a small stream based on 137 Cs inventories of the bed load fraction.' Water
Resour. Res. 26(6). 1177-1187.
Spieker, R. and Ergenzinger, P. 1987. 'New developments in measuring bedload by the
magnetic tracer technique.' In : Erosion, Transport and Deposition Processes.
Proceedings ofthe Jerusalem Workshop, March-April1987. IAHS Publ. No. 189.
169-178.
Sutherland, A.J. 1987. 'Static armour layers by selective erosion.' In: C.R. Thome, J.C.
Bathurst and RD. Hey (Eds). Sediment Transport in Gravel Bed Rivers. John
Wiley and Sons. Chichester. 243-260.
Southard, J.B., N.D. Smith and R.A. Kuhnle. 1984. 'Chutes and lobes: newly identified
elements of braiding in shallow gravelly streams.' In: Koster, E.H. and Steel, R.J.
(Eds). Sedimentology of Gravels and Conglomerates. Can. Soc. Petr. Geol. Mem.
10. 51-59.
Tacconi, P. and P. Billi. 1987. 'Bedload transport measurement by the vortex trap on
Virginio Creek, Italy.' In: C.R. Thome, J.C. Bathurst and RD. Hey (Eds).
Sediment Transport in Gravel Bed Rivers. John Wiley and Sons. Chichester. 583616.
Thoms, M. C. 1992. 'A comparison of grab- and freeze-sampling techniques in the
collection of gravel-bed river sediment.' Sediment. Geol. 78. 191-200.
Thome, P.D., J.J. Williams, A.D. Heathershaw. 1989. 'In situ acoustic measurements of
marine gravel threshold and transport.' Sedimentology. 36. 61-7 4.
Whiting, P.J., W.E. Dietrich, L.B. Leopold, T.G. Drake and R.L. Shreve. 1988. 'Bedload
sheets in heterogeneous sediment.' Geology. 16. 105-108.
Wilcock, P.R. and J.B. Southard. 1989. 'Experimental study of incipient motion in
mixed-size sediment.' Water Resour. Res. 24. 1137-1151.
Willetts, B.B., J.K. Maizels and J. Florence. 1987. 'The simulation of stream bed
armouring and its consequences.' Proc. Inst. Civ. Eng., Part 1,82. 799-814.
Williams, J.J, P.D. Thome, A.D. Heathershaw. 1989. 'Comparisons between acoustic
measurements and predictions of the bedload transport of marine gravels.'
Sedimentology. 36. 973-979.
Yilmaz, Ozdogan. 1987. Seismic Data Processing. S. Doherty (Ed). Society of
Exploration Geophysicists. Tulsa.

99

Appendix A- Glossary
a-, b-and c- axes: 3 measurements describing the dimensions of a clast. The a-axis of a
clast is the longest axis of the grain, and the longest axis orthogonal to that is the
b-axis. The two dimensional area described by the a-b plane of a clast is its
maximum projection. The c-axis is normal to the a-b plane.
Aggradation: The process of building up a surface by deposition.
Allochthonous : Derived from outside a system, for instance, colluvial material from a
cliff face contributing to the bedload of a stream.
Anabranching River: Multiple channels separated by stable islands which are large
relative to the size of the channels, and which divide the flow at discharges up to
and including bankfull. Consequently the flow patterns in adjacent channel
segments are essentially independent of one another, unlike those in braided
channels (Knighton, 1998).
Armour layer: Erosion-resistant layer of relatively large particles on the surface of a
streambed. Such layers typically result from removal of finer particles by erosion.
Autochthonous: Derived from within a system, for instance, sediment transported from
one gravel bar to the next.
Bankfull: Most commonly defined as the point at which discharge in the creek has filled
the channel to capacity, occurring on the average once every 1.5 years.
Bioturbation: Disturbance of a natural system by an organism; in this case, the
disruption of the streambed gravels by spawning salmon.
BMD: The Bedload Movement Detector, developed at UNBC. The device consists of an
array of magnetic sensors housed in an aluminum beam. The beam is set into the
stream bed such that its upper surface is flush with the gravel bed, and signals
from the sensors are recorded during a bedload transport event. An estimate of
bedload transport may then be obtained.
Braided Stream: Stream that forms an interlacing network of branching and
recombining channels separated by bars and contained within a dominant pair of
floodplain banks.
Competence: A threshold of stream discharge at which particles may be entrained and
carried with the flow.

100
DFO: The Department of Fisheries and Oceans, recently renamed Fisheries and Oceans
Canada (FOC).
Equal Mobility: A hypothesis stating that coarse sediment particles will resist
entrainment due to their being hidden or imbricated. Once flows reach a critical
threshold, however, the resistance ofthe larger, structural stones is overcome, and
all size ranges of material become available for transport at roughly the same
time.
Fast Fourier Transform (FFT): A fast method for calculating the Discrete Fourier
Transform. It is used when the number of samples in a time series is a power of2.
Fluvial Entrainment: Suspension and transport of solid materials by stream flow.
Fourier Transform: Usually referred to as the Discrete Fourier Transform or DFT. A
mathematical process whereby a time series, or signal, is uniquely and completely
described as a sum of a number of sinusoids, each with a unique peak amplitude,
frequency, and a phase lag (relative alignment). The resultant series provides a
representation of frequencies present within the original signal.
Freeze Core Sampling: A method of sampling streambed gravels. The device used in the
present study consisted of a 1m long, 5 em diameter, hollow aluminum standpipe
with a bowl-like flared open end. The pipe has a pointed tip that is driven into the
streambed some 30 em. Acetone is then poured into the large flared end, filling
the standpipe, and dry ice is placed in the bowl. As the acetone cools, the
interstitial water and sediment next to the pipe are quickly frozen (1 0-20 min) and
then removed as a core.
Hilbert Transform: The Hilbert Transform (.Yt) returns a complex helical sequence,
sometimes called the analytic signal, from a real data sequence. The analytic
signal has a real part, which is the original data, and an imaginary part, which
contains the Hilbert Transform. The imaginary part is a version of the original real
sequence with a 90° phase shift. Sines are therefore transformed to cosines, and
vice versa. The Hilbert transformed series has the same amplitude and frequency
content as the original real data and includes phase information that depends on
the phase of the original data (Mitra, 1990).
Hydraulic Efficiency: The ratio of the mean velocity of water discharge through the
sampler to the mean velocity of the water discharge which would have occurred
through the area occupied by the opening in the sampler nozzle had the sampler
not been there (Hubbell, 1964).
Imbrication: The interlocking, or "shingling" of particles, as they are naturally deposed
in similar orientations. Freshly deposited material becomes more tightly packed as
marginal flows shake and settle them into place.

101
JTFA: Joint Time-Frequency Analysis. A method of analysis that simultaneously
provides both time and frequency information. It shows how the frequency
spectrum of a signal varies with time.
Large Woody Debris (L WD): Any large piece of woody material that intrudes into a
stream channel, whose smallest diameter is greater than 10 em, and whose length
is greater than 2m. LWD is important for providing fish habitat and in influencing
channel morphology.
Locus of Transport: The portion of a stream cross-section that exhibits the highest rates
ofbedload transport.
Low Pass Filter: A filter that passes frequencies below a certain cutoff frequency. It
passes low frequencies but attenuates high frequencies.
MOF: The British Columbia Ministry of Forests.
Nival Flood: The spring flood event, generated by snow-melt, in mountain catchments.
Nyquist Frequency: Half of the sampling frequency. The Nyquist theorem states that to
recover an analog signal from its samples, the sampling frequency should be at
least twice the highest frequency in the signal. Therefore, in order to capture
signals up to 50Hz, the BMD sampled the voltage data at a rate of 100 Hz.
Phi ( ¢) Scale: A logarithmic scale used to describe the size of sediment grains. The
relation between grain size (D) and phi ( ¢) units is as follows:

¢ = -log2D(mm) = -3.3219 log 10D(mm)
Pool-Riffle Morphology: Pool-riffle channels have an undulating bed that defines a
sequence of bars, pools and riffles. Pools are topographic depressions within the
channel, spaced every five to seven channel widths. Riffles are shallow sections
of the stream with rapid currents and a surface broken by gravel, rubble, or
boulders.
Psi ( l/1) Scale: The inverse (negative) representation of the logarithmic Phi ( ¢) Scale.
Redd: Nest made in gravel, consisting of a depression hydraulically dug by a fish for egg
deposition (and then filled) and associated gravel mounds.
Shear Stress: The product of the unit weight of water (1000 kg!m\ depth (m) and slope
(m/m), and is a parameter useful in predicting the maximum size of bedload
particles that may be entrained.

102
Short Time Fourier Transform {STFT}: The term for taking a Fourier transform of
shorter time intervals of samples of a signal, rather than on the entire set of
samples. Also known as the windowed Fourier transform.
Stage Level: The depth of stream flow at a monitoring station. The stage level is usually
calibrated to previous measurements of discharge.
Stream Efficiency: Expressed as the unit bedload transport rate divided by unit stream
power.
Stream Order: A number from 1 to 6 or higher, ranked from headwaters to river
terminus, that designates the relative position of a stream or stream segment in a
drainage basin. First-order streams have no discrete tributaries; the junction of
two first-order streams produces a second-order stream; the junction of two
second-order streams produces a third-order stream; etc.
Stream Power: The product of the unit weight of water (1000 kglm\ velocity (rnls),
flow depth (m) and slope (rnlm). This is a useful parameter in presenting the
transport rate ofbedload sediment.
Stream Reach : A relatively homogeneous section of a stream having a sequence of
repeating structural characteristics (or processes) and fish habitat types.
Thalweg: A line connecting the deepest parts of a stream channel.
Third Order Tributary: (see Stream Order)

103

Appendix B - Gravel sampling data

19+40

cabin

bedload
detector

19+60

19+70
Freeze Core
Sample Locations,
September 20-21 , 1998
One'ell Project Site

D bulk sample site

5 ~ freeze core site
10m grid

t

19+80

Flow
Direction

19+90

Figure 50 - Locations of gravel sampling sites at the BMD study site.

98/09/20 -8325.4
98/09/20 -7426 .8
98/09/20 -5040 .1
98/09/20 -9289.3
98/09/20 -14699.2
98/09/20 -8600.6
98/09/20 -7077 .0
98/09/20 -5226 .2
98/09/20 -8148.3
98/09/20 -3616 .8
98/09/20 -9241.0
98/09/20 -13245.4
-8534 .7
98/09/21
98/09/21 -12663.1
98/09/20 . -13018.3
98/09/20 -7781 .9
98/09/20 -10234 .2
98/09/20 -10960.6
98/09/20 -3477.1
98/09/20 -15371 .6
98/09/20 -11962.7
98/09/20 -10502.7

Date

7324.3
6449.5
4542 .1
8274.9
13204.7
7618.5
6113 .1
4709.0
7162.0
3119.6
8235.7
11784.0
7568.3
11177.8
11534.8
6776.7
8750.3
9464.0
2990.0
13868.1
10961 .6
9525.4

n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n

Total
Sample
Total
Weight
Split
preshake Sample
Weight (g) (y/n)
(g)

Table 6 - Freeze Core sample data

6.3
8.5
3.0
8.2
13.8
8.0
6.9
3.7
6.5
11 .3
12.7
28.2
9.7
31 .3
37 .7
7.9
43 .3
10.5
0.7
20.4
8.6
20 .7

0.00.075mm
(g)
61.3
72 .8
41 .9
62 .2
56.9
59.0
28 .9
17.0
45.8
133.3
90.2
207.5
65.6
214.0
215.3
64.2
276.5
83 .0
1.4
144.1
59.6
183.7

.0750.3mm
(g)
57.4
240.1
60.2
643 .2
348.4
278 .1
156.6
22 .1
147.6
283.7
694.2
918.4
390.5
969.0
1349.1
310 .3
997.5
662 .3
5.4
881 .3
274 .5
543 .0

0.31.18mm
(g)
130.4
379 .1
91 .7
644.8
541.4
416.1
221 .3
28.8
142.3
272.5
627.1
910 .2
579.3
627.0
978.0
324 .2
788 .0
638 .6
12.2
855.2
358.1
553 .1

1.182.36mm
(g)
254.8
595.4
190.8
543.3
582.4
563.8
246 .7
27 .1
128.0
332.2
546.9
1275.7
636.3
900.6
1364.2
447.8
1090.0
674.3
32.4
1133.3
357 .5
583.1

2.365.6mm
(g)
224.7
469.9
186.5
502.5
429.4
452 .6
287.3
41 .9
160.0
305.5
565.4
910.2
487 .0
828 .3
1058.4
381.5
842 .3
526 .9
46.7
961 .6
355.6
553.4

5.69.5mm
(g)
142.5
297.1
95.2
351 .8
278.0
278 .3
272 .2
45.0
131.4
182.2
377.8
533.3
286.9
484 .5
654.0
220.7
501 .5
377 .8
62 .9
599.5
236.4
399 .8

9.5 12.5mm
(g)
909.5
773.2
446 .7
1734.6
1401 .6
1266.9
1910.2
1034.7
518 .8
915.7
1854.6
2308 .0
1230.4
2245.8
2236.9
906.1
2175.4
1854.9
640.1
2476 .2
1324.8
1895.6

12.525mm
(g)
1054.4
412 .0
712 .5
1313.9
1116.7
918.0
1309.4
663 .5
1153.9
683.0
1400.9
2081 .2
935.0
1552.2
854.4
958.4
1382.2
1751.7
692.7
1397.2
348 .2
810.6

25-50
mm
(g)

Sum of
Seive
fractions
7326.8
6447.2
4542 .5
8274.4
13199.6
7618 .6
6112.9
4709.1
7162.4
3119.4
8234.2
11782.3
7567.2
11177.4
11531 .2
6776.4
8748.5
9463.2
2992 .8
13861 .8
10960.7
9522 .0

> 50mm
(g)
4485.5
3199.1
2714.0
2469.9
8431 .0
3377 .8
1673.4
2825.3
4728.1
0.0
2064.4
2609.6
2946 .5
3324 .7
2783.2
3155.3
651.8
2883 .2
1498.3
5393.0
7637.4
3979.0

l

Coarse

I

.J

Selit four way_s, so multielied by_ 4

27197.1

46072.3

5866 .5
6346.0
6380.0

18592.5

4439.2
4481 .7
4437.0
4708.4
4852 .8
3528.9

26448.0

13286.7

5990.0
4837.7
2459.0

31224.0

14258.71 31224.0

I

F"

L

7404 .8
6853 .9

9591.2

9591.2

3655.2

3655.2

I

30J

3005.61

7034.2
9428.5
9762.5
4850.6

7644.6
7728 .8
8324 .1
4542.4

7580.4
7875.6
8017.5
5468.5

6563.2
7495.0
6035.9

6435.8
5378.3
3891 .2

6468.9r .. 40626.0
6132.3
3659 .0.

L

17964.0

5337.2

6398.4.

Passed #1 Passed #2 Passed#3 Passed#4 Passed#5 Passed#6 1Passed #7 Passed#8 Passed#9 Passed #101

5086 .8
4580 .8
5050 .1
3986.2
4134.3
4358 .9

5438.8
3520 .3
5032 .0
6251 .5
6631 .2
7784 .2
6744.1
4670.2

64.0
45.0
32.0
22 .6
16.0
11 .31
8.0
2.0
0.833
.~
Passed #1 Passed #2 Passed#3 Passed#4 Passed#5 Passed#6 !passed #7 Passed#8 Passed#9 Passed #1 0:

102176.7

Coarse

Sample #1 taken from gravel bar near bank failure, ?Om upstream
Sample #2 taken from gravel bar 50m downstream from kitchen
Sample #3 taken from Wside bar, 1Om upstream of the detector

Table 7- Bulk Sample (15 em deep) Sieve Data- coot on next page

295508.0

TOTAL:

105

L

81219.0

Coarse

34126 .0

28239.9

28942.0

20094.1

15705.3

16260.21 40626.0

17964.0

5337.2
6398.4!

8680.0
9493.2
8675.0

26848.2

10518.6
11667.0
9820.0
10149.7

42155.3

19463.1

9277 .6
6851 .7
3333.8

13808.0

8006 .0
5802 .0

11889.1

7456.0
4433.1

31326.0

13759.0l 31326.0

5705 .0j
6121 .3
1932.7

14497.2

14497.2

5164.4

5164.4

3473.2

.~

Passed #1 Passed #2 Passed#3 Passed#4 Passed#5 Passed#6l assed #7Passed#8 Passed#9 Passed #101

31075.8

263602 .5

244768.9

106

f-

5768.0

11247.0

23047 .3

otal Sample

Omm

1.18 mm
.36mm
.70mm
.60mm
.80mm
.65mm

5768 .0

5315.4
5680 .1
251 .5

10736.8
6914.7
5395.8

Sieve Work- August 4-6, 1998
0.833 mm
1.18 mm
2.36 mm

3039.6

3039.6

357315.4

23047.3
11247.0
5768.0
21955.7
3039.6
16182.4
2175.1
12069.5
40096.8
85646.8
136087.2
4.53,
0.61
3.38
11 .22
23.97
38.09

. ~

6.45
3.15
1.61
6.14

Weight of
Percentage
Fraction {g). By Weight

21955.7

3892 .0
4929.5
5288.3
4254.6
3591 .3

16182.4

7720 .8
8461 .6

All weights measured in grams (g)
2.70 mm
5.60 mm
5.80 mm

Table 8- Bulk sample (30 em deep) Sieve Data

2175.1

820 .8
1354.3

8.65 mm

12069.5

6851.0
5218 .5

9.5mm

40096.8

1954.6
4925.2
6300 .7
6566.0
4913.6
6600 .0
5231 .3
3605.4

12.3 mm
8307 .6
8860.2
8004.8
9346.8
8172 .5
10116.2
9784.8
9090.6
6995.5
6967.8
85646 .8

24.08 mm

136087.2

136087.2

50mm

107

-

phi

-5.95
-5 .94
-6.58
-6.15
-6.18
-6.90
-6.69
-6.48
-6 .30
-6.40
-6.48
-5.93
-6.55
-6.54
-5.87
-6.11
-5.82
-6 .19
-6 .19
-6.14
-5.87
-6.71
-6.11
-6.34
-5.79
-6.11

Rock#

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26

61.66
61.48
95.97
71 .18
72.44
119.67
103.51
89.12
78.70
84.74
89.26
60 .95
93.88
92 .90
58.42
68.96
56.59
72 .93
73.15
70.39
58 .65
105.01
68.88
81 .06
55.40
68.94

Equiv.
Sieve
Size (mm)

111.6
86
154.4
108.4
128.9
157.2
155.3
134.1
122.2
95
153
117.1
118.7
133
78.2
113.2
88.3
106.5
145.4
91 .1
134
155.2
117.21
101.3
85.1
96.2
77 .9
74.4
119.2
78.1
88.5
141 .2
116
106.5
96.7
87.4
99 .8
64.4
104.5
109.9
68.5
78.7
67 .5
81 .2
95.4
77.4
68
125.3
81 .2
96.3
65.1
78
39.2
45
64 .9
63.5
51 .6
93.3
89.3
67.4
55.1
82
77.3
57 .3
81.9
72
46.2
57.6
43
63 .6
40
62 .6
47.5
79.7
53.8
62 .2
43 .6
58 .5

Caliper Measure (mm)
A
B
c

0.56
0.68
0.61
0.78
0.62
0.73
0.76
0.69
0.64
0.93
0.73
0.76
0.82
0.71
0.74
0.72
0.68
0.78
0.49
0.82
0.63
0.69
0.68
0.74
0.70
0.77

Sphericity

Table 9- Sample of Magnetic Sampling Data (All clasts < SOmm)

12
5
13
23
15
14
18
14
9
24
27
17
18
23
27
8
12
7
7
6
13
25
18
7
10
12

(mm)

ness

Round

-0 .533
-0.492
-0 .521
-0 .595
-0 .343
-0 .612
-0.193
-0.225
-0.456
-0.454
-0.448
-0.544
-0.010
-0.526
-0.445
-0.524
-0.497
-0.484
-0.433
-0.514
-0.619
-0.484
-0 .673
-0 .548
-0.499
-0.485
-0.029
-0.473
-0.494
-0.419
-0.032
-0 .560
-0 .073
-0.080
-0.388
-0.419
-0.425
-0.442
0.056
-0 .513
-0.424
-0 .510
-0.489
-0.471
-0.404
-0.503
-0.146
-0.463
-0.664
-0.535
-0.463
-0.475
-0.141
-0.483
-0 .509
-0.474
-0.076
-0 .593
-0.129
-0 .152
-0.424
-0.437
-0.440
-0.498
0.024
-0 .521
-0.437
-0.517
-0.494
-0.479
-0.422
-0 .510
-0.441
-0.471
-0.669
-0.543
-0.478
-0.480

Magnetic Intensity
Min
Max
Avg

0.392
0.009
0.012
0.121
0.267
0.019
0.064
0.073
0.032
0.017
0.008
0.046
0.034
0.005
0.008
0.007
0.003
0.005
0.011
0.004
0.178
0.013
0.004
0.005
0.021
0.005

452 .1
346.4
1703.3
714.1
1022.7
2232.7
1742.3
1510.2
840.7
710.6
1796.0
693.6
1036.6
1559.4
332.8
595.2
257.4
542 .6
909.0
505.5
678 .0
2367.2
641.7
707.7
301.4
580.5

Intensity
Weight (g)
Value

Greeny UM
Andesite
Meta-shale
Grano-diorite
Quartzose
Graywacke
C.C. Quartzite
Rhyolite w/Quartz bands
Grano-diorite
Grano-diorite
Meta-graywacke
Andesite
Rhyolite
Grano-diorite
Grano-diorite
C.C. Quartzose
Mudstone
Volcanic
Graywacke
Meta-green?
Greeny UM
Grano-diorite
Meta-shale
C.C. Green
Rhyolite
Diorite

Note: C.C . =cache creek
metamorphics
Lithology

108

phi

-5.80
-6.83
-5.84
-5.99
-6.35
-6.00
-5.75
-6.14
-5.69
-5.73
-5.80
-5.96
-5.95
-5.71
-6.71
-6.34
-5.66
-6.27
-5.77
-6.16
-5.67
-6.39
-5.87
-6 .67
-5.86
-6.43
-5.93
-6.60
-5.87
-6.78

Rock#

27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56

55.83
113.98
57.47
63.57
81.38
64.11
53.98
70.61
51 .73
53 .14
55.67
62 .34
61.77
52.34
104.64
81.04
50.45
77 .07
54.49
71 .74
50.91
84.06
58 .54
101 .51
58.23
86.14
60.79
97.05
58 .37
110.20

Equiv.
Sieve
Size (mm)
83.5
189.2
108.8
100.3
97.8
90.1
85.8
95.4
62.3
68 .2
85.8
90.9
86.16
78 .7
152.5
109
77 .8
123.1
85.8
108.7
82.1
151.4
83.6
141 .3
103.1
106.2
119.2
130
87 .7
145
68 .7
116.6
70 .3
75.18
93 .5
72 .9
67
83.8
64.4
55 .2
69.7
78 .72
67.7
65
113.8
96.9
55.2
94.5
68.5
84.8
57.6
102.8
72.4
115.9
73
97.9
69.2
112.1
70.7
117.3
38 .9
111 .3
40.8
49 .3
67.1
53.9
36.6
54.3
34.7
51
36.6
39 .7
55.2
35.4
94 .6
61 .2
45.2
54 .3
35.3
55.7
43 .2
59.7
40.15
84.7
38.1
72 .5
51
79 .2
42.6
102.6

Caliper Measure (mm)
A
B
c
0.64
0.83
0.60
0.69
0.79
0.76
0.62
0.72
0.67
0.89
0.61
0.61
0.81
0.63
0.80
0.71
0.78
0.64
0.60
0.70
0.74
0.61
0.65
0.76
0.58
0.80
0.68
0.76
0.67
0.85

Sphericity
8
13
11
8
4
15
14
11
13
8
7
13
5
14
32
10
9
12
5
6
9
22
6
7
13
11
13
22
10
16

ness

Round

-0.428
-1 .243
-0.555
-0.501
-0.519
-0.562
-1 .016
-0 .713
-0 .703
-2 .584
-0 .683
-0 .721
-0.920
-3 .046
-0.488
-0.092
-2.160
-0 .699
-0 .035
-0.538
-0.466
-0.471
-0.458
-0.482
-0.422
-0.525
-0.482
-0.398
-2.970
-0.527
-0.412
-0.067
-0 .519
-0.485
-0 .508
-0 .510
-0.103
-0.686
-0 .624
-0.127
-0 .647
-0.709
-0.555
0.062
-0.474
-0.060
-0 .042
-0.671
0.053
-0 .137
-0.445
-0.449
-0.441
-0.465
-0.402
-0.492
-0.458
-0.381
0.125
-0.498

-0.418
-0.661
-0.535
-0.492
-0.514
-0.541
-0.538
-0.703
-0.660
-0.696
-0.666
-0 .716
-0.710
-0 .881
-0.481
-0.075
-0.643
-0.688
0.009
-0.415
-0.455
-0.460
-0.451
-0.472
-0.412
-0.510
-0.472
-0.387
-0.659
-0 .508

Magnetic Intensity
Min
Max
Avg
0.010
0.582
0.020
0.009
0.005
0.021
0.478
0.010
0.043
1.888
0.017
0.005
0.210
2.165
0.007
0.017
1.517
0.011
0.044
0.123
0.011
0.011
0.007
0.010
0.010
0.015
0.010
0.011
2.311
0.019
299.7
2909.2
474 .0
506.8
711.0
575.7
274 .5
501.7
191 .7
311.2
352.7
323.2
325.7
259 .1
2362 .8
714.9
300.5
865.5
299.7
526.8
252.1
1364.8
261 .3
1633.5
374.6
767.1
569.7
1660.0
330 .3
1620.8

Intensity
Weight (g)
Value

Grano-diorite
Greeny UM (Meta?)
Grano-diorite
Grano-diorite
C.C. Quartzose
Meta-greywacke
Andesite
C.C. Quartzose
Greenstone
Porph Andesite
Andesite
Grano-diorite
Dacite
Porph Andesite
Grano-diorite
Grano-diorite
Andesite
Grano-diorite
C.C. Quartz
C.C . Green
C.C. Green
Grano-diorite
Grano-diorite
Meta-Quartzy ?
C.C. Green
Grano-diorite
Dacite
Grano-diorite
Meta-Quartzy ?
Grano-diorite

Note: C.C. =cache creek
metamorphics
Lithology

109

phi

-5.87
-5.93
-5.61
-5 .97
-5.92
-5.96
-6.37
-6 .21
-6.14
-6.24
-5.53
-5.80
-5.85
-6 .11
-5.90
-6.12
-5.70
-5.92
-6 .27
-6 .04
-5.98
-5.97
-6.26
-6.19
-5.98
-5.81
-5.89
-6.00
-6.18
-6.30

Rock#

57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86

58.48
61 .09
48 .91
62 .73
60 .51
62.44
82 .58
74 .16
70.58
75.82
46.12
55.87
57.57
68.95
59.57
69 .54
52.15
60.65
77.42
65.72
62 .94
62.66
76 .39
72.92
63 .20
56 .06
59 .26
64.02
72.58
78.73

Equiv.
Sieve
Size (mm)
98 .2
82.4
87.7
87.1
88.4
88.4
116.1
109.2
106.5
97.3
105.8
71 .9
101.2
87.7
90.7
131.9
87.5
105
131 .7
99.5
104.2
122.6
96.4
102.9
94 .9
76.5
94.4
102.14
111 .06
108.8
61.5
78.9
59 .9
72.8
69 .3
70.7
100.5
83.4
84 .2
89.8
53.5
64.1
64 .8
79.4
71.6
80.56
66.1
75.5
88.4
73.8
67.7
72.4
89.9
88.4
83.3
56.8
66.5
78.1
85.5
96 .7

55.3
35.2
34.6
50.7
50.2
52.9
59 .5
63 .6
53.6
58.6
37.3
46.2
49.3
56.6
44.4
56.4
32.7
40.7
64.6
56.5
57.8
51 .1
59 .9
53 .1
32.4
55.3
51
45.8
56.8
55.2

Caliper Measure (mm)
A
B
c
0.80
0.58
0.61
0.74
0.75
0.77
0.67
0.77
0.69
0.73
0.63
0.78
0.72
0.77
0.67
0.67
0.57
0.60
0.71
0.76
0.78
0.67
0.75
0.68
0.51
0.89
0.75
0.64
0.70
0.66

Sphericity
12
4
6
8
21
8
18
22
9
11
8
15
6
8
4
13
14
3
17
3
8
18
13
16
5
6
6
4
8
5

ness

Round
-0.502
-0.492
-0.706
-0.758
-1 .025
-0.678
-0.803
-0.725
-0.753
-0.704
-0.713
-0 .713
-0.729
-0.718
-0.693
-0 .907
-0.931
-0.713
-0.725
-1.122
-0.797
-0.809
-0 .858
-1 .844
-0.565
-0.780
-3.030
-0.782
-0 .868
-1.967
-0.472
-0.464
-0 .668
-0.530
-1.005
-0.661
-0.513
-0.634
-0 .551
-0.690
-0.701
-0.701
-0.718
-0.703
-0.636
-0.443
-0.431
-0.699
-0.692
-0.463
-0.783
-0.767
-0.777
-0.111
-0.539
-0 .762
0.101
-0.767
-0.626
-0.041

-0.484
-0.476
-0.679
-0.668
-1 .011
-0.671
-0.652
-0.686
-0.696
-0 .697
-0 .708
-0.710
-0 .723
-0 .710
-0 .676
-0 .720
-0 .726
-0 .706
-0 .703
-0 .825
-0 .790
-0 .784
-0.800
-0.874
-0.554
-0.772
-0.602
-0.776
-0.773
-0.944

Magnetic Intensity
Min
Max
Avg
0.018
0.016
0.027
0.090
0.014
0.007
0.151
0.039
0.057
0.007
0.005
0.003
0.006
0.008
0.017
0.187
0.205
0.007
0.022
0.297
0.007
0.025
0.058
0.970
0.011
0.008
2.428
0.006
0.095
1.023
454.9
296.4
234.9
423 .0
413 .3
392 .1
931.5
755.2
586.2
662.4
265.5
292.9
388 .9
433 .7
310 .5
679.0
243 .7
450 .9
1020.7
441.6
547 .8
620 .3
654.8
722.4
355.3
259.7
342 .7
466 .2
550.5
657 .8

Intensity
Weight (g)
Value

Meta-Sandstone
Mudstone
C.C. Green
C.C. Green
Meta-Sed-Quartzose
Green/Mafic
Porph Andesite
Porph Dacite
C.C. Green
Grano-diorite
Meta-Sandstone
Quartzite
Dacite
C.C . Quartzose
Meta-Sed.
Serpentine
Greenstone/Quartzose
Meta-Shale
C.C. Green
UM Green
Grano-diorite
Grano-diorite
Grano-diorite
UM Green
Meta-Sed.
C.C. Quartzose
Porph Andesite
C.C. Meta-Sed .
Greenstone
Greeny UM

Note: C.C. =cache creek
metamorphics
Lithology

110

Equiv.
Sieve
Size (mm)

50.07
90.06
80.39
65.07
49.97
54.84
76.11
74.18
58.84
82 .23
77 .13
66 .90
63.85
60.81
83.14
56.97
89.63
48.96
90.74
60.41
85.04
57.05
76 .95
51.52
57.76
68 .96
58.90
48.51
56.87
83.57

phi

-5.65
-6.49
-6 .33
-6.02
-5.64
-5.78
-6.25
-6.21
-5.88
-6.36
-6 .27
-6.06
-6.00
-5.93
-6.38
-5.83
-6.49
-5.61
-6.50
-5.92
-6.41
-5.83
-6.27
-5.69
-5.85
-6.11
-5.88
-5.60
-5.83
-6.38

Rock#

87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116

89.4
174.2
123.1
91 .3
68.2
70.2
109.8
169.5
87.3
93
116.1
88 .2
84
110.5
140.2
83.6
130.5
73 .1
121 .1
88 .8
128.2
87.3
105.1
82 .6
102.8
91.1
84.5
95.4
80
117.1
59.5
105.8
89.3
76.4
51.5
67.5
104.8
85.7
78 .6
88 .1
97.6
85.6
81
61 .6
92.1
58.5
95.8
64.3
106.7
71.8
99.6
74.2
80.8
61.6
66.1
80.4
78.3
58.4
63
85.5

38.4
70 .9
70.36
51 .3
48.4
38.2
24.54
60.5
27.3
75.9
48.7
40.3
39.9
60
73 .1
55.4
83
25.7
71.3
46 .3
67.4
31.7
72 .9
38.9
48
55.2
28.4
36
50
81.6

Caliper Measure (mm)
A
B
c

0.65
0.65
0.77
0.72
0.87
0.68
0.38
0.63
0.48
0.89
0.60
0.60
0.62
0.81
0.75
0.86
0.82
0.52
0.74
0.70
0.71
0.54
0.86
0.67
0.70
0.75
0.50
0.62
0.79
0.87

Sphericity

4
19
9
12
7
8
5
10
5
8
12
5
7
4
20
5
14
8
27
11
7
6
6
3
7
6
6
6
11
19

ness

Round

-0.754
-0.805
-0.760
-0.929
-0.817
-0.727
-0.749
-0.732
-0.655
-0.739
-0 .678
-0.878
-0.868
-0.896
-0.783
-0 .908
-0.689
-0.625
-0.646
-0.598
-2.287
-0 .756
-0.656
-0.623
-0.624
-0.647
-0 .654
-0.716
-0.712
-0.669
-0.733
-0.724
-0.730
-0 .530
-0.723
-0.710
-0.729
-0.657
-0.642
-0.730
-0.666
-0 .864
-0.857
-0 .883
-0.698
-0 .512
-0.636
-0.613
-0.618
-0.576
-0.030
-0.733
-0.634
-0.595
-0 .592
-0.616
-0 .644
-0 .626
-0.702
-0.624

-0 .745
-0.778
-0.747
-0.079
-0.779
-0.720
-0.741
-0.694
-0.649
-0.734
-0 .672
-0.870
-0 .862
-0 .889
-0.722
-0 .784
-0 .667
-0 .619
-0 .635
-0 .586
-0.591
-0.745
-0.645
-0.611
-0.599
-0.626
-0.649
-0.696
-0.708
-0 .648

Magnetic Intensity
Min
Max
Avg

0.009
0.027
0.013
0.850
0.038
0.007
0.008
0.038
0.006
0.005
0.006
0.008
0.006
0.007
0.061
0.124
0.022
0.006
0.011
0.012
1.696
0.011
0.011
0.012
0.025
0.021
0.005
0.020
0.004
0.021
278 .6
1793.8
895.5
485.9
235.5
215.3
364.5
1153.5
247 .8
684.7
729 .3
460 .5
338.4
492 .7
1605.2
381 .5
1254.4
153.8
1227.3
378.9
853.6
278.0
735.2
230 .5
442 .5
587.2
227.2
278.8
322.2
1220.7

Intensity
Weight (g)
Value

Porph Dacite
Grano-diorite
Rhyolite
C.C. Green
C.C. Quartzose
Greywacke
Quartzite
C.C. Meta-Sed.
Soapstone
Grano-diorite
Meta-Quartz Sandstone
C.C . Quartzose
Dacite
Meta-Quartz Sandstone
C.C Meta Quartzose
Greeny UM
Grano-diorite
Meta-mudstone
Rhyolite
Quartzite
Greeny UM
Porph Andesite
Greenstone
Meta-Sed.
Grano-diorite
C.C. Green
Grano-diorite
C.C. Green
Meta-conglomerate
Grano-diorite

Note: C.C. =cache creek
metamorphics
Lithology

111

Equiv.
Sieve
Size (mm)

55.84
51 .10
74.37
50.21
52 .57
57.09
67.44
53.63
69 .11
47.47
52 .22
75.04
50.89
56.39
88.10
50 .32
60.68
56.60
67.01
57 .96
58.48
51.80
68.14
76.23
62.47
52 .38
80.33
67.09
70.27
88.42

phi

-5.80
-5.68
-6.22
-5.65
-5.72
-5.84
-6 .08
-5.75
-6.11
-5.57
-5.71
-6.23
-5.67
-5.82
-6.46
-5.65
-5.92
-5.82
-6 .07
-5.86
-5.87
-5.69
-6 .09
-6 .25
-5.96
-5.71
-6 .33
-6 .07
-6 .1 3
-6.47

Rock#

117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146

73.8
69.9
80.8
75.1
78.1
89.3
139.3
80
94.7
66.5
126.8
113.2
82.1
82 .8
156.1
77.4
76.1
80.8
93.6
80.4
84.9
81 .9
100.6
123.6
102.5
86.3
127.9
106.3
108.6
124.8
69
58 .6
84.3
66.2
64.6
67.5
71 .6
67.5
91 .6
63.9
62.4
88.8
55.1
65
104
56
73
67
86
71 .2
68 .6
51.9
70.4
89.7
70 .3
65.6
99 .9
77 .1
86.4
107.3

38.4
42 .3
62.9
25.7
36.8
44.3
63
34 .6
34 .1
20 .6
39 .5
58.1
46 .3
46 .2
68 .6
43.9
45.1
43.8
39.8
40.6
46.2
51 .7
65.8
59.8
53.5
34.4
54.1
55.3
49.1
64.2

Caliper Measure (mm)
A
B
c
0.66
0.76
0.84
0.51
0.65
0.69
0.74
0.61
0.52
0.47
0.59
0.70
0.78
0.74
0.66
0.77
0.72
0.71
0.58
0.66
0.72
0.86
0.85
0.69
0.74
0.60
0.61
0.72
0.64
0.68

Sphericity
8
5
5
3
5
3
5
7
9
8
6
15
1
9
21
6
9
5
11
5
10
7
11
9
4
2
6
9
5
8

ness

Round
-0 .633
-0.649
-0.710
-0.648
-0.646
-0.867
-0.660
-0.627
-0.631
-0.623
-0.644
-0.774
-1.283
-1 .094
-0 .716
-0 .723
-0.734
-0.744
-0.699
-0.828
-0.685
-1.443
-0.698
-0.546
-0.513
-0.516
-0.499
-0.457
-0.864
-0.811

-0.600
-0.635
-0.596
-0.626
-0.636
-0.493
-0.608
-0.609
-0.608
-0.607
-0.598
-0.587
-0 .394
-0.531
-0 .690
-0.691
-0.685
-0.716
-0.671
-0.575
-0.659
-0 .292
-0.675
-0.471
-0.471
-0.464
-0.458
-0.431
-0.072
-0.069

-0 .617
-0.641
-0.658
-0.634
-0.641
-0 .695
-0 .641
-0.619
-0.621
-0.616
-0.618
-0.704
-0.696
-0.750
-0.706
-0.707
-0.714
-0.730
-0.683
-0.713
-0.674
-0.684
-0.689
-0 .518
-0.493
-0.497
-0.478
-0.446
-0.451
-0.635

Magnetic Intensity
Min
Max
Avg
0.016
0.008
0.052
0.014
0.005
0.172
0.019
0.008
0.010
0.007
0.026
0.070
0.587
0.344
0.010
0.016
0.020
0.014
0.016
0.115
0.011
0.759
0.009
0.028
0.020
0.019
0.021
0.011
0.413
0.176

236.9
238 .0
532.1
216.8
233.8
300 .6
845.1
275.0
359 .1
138.4
409.0
641 .3
347.9
395 .3
1593.8
261 .9
310.7
343.8
415.9
248.3
357.7
333.2
621 .0
779.6
480.7
196.0
778 .5
587.9
516 .1
814.7

Intensity
Weight (g)
Value

Grano-diorite
Meta-greywacke
C.C. Volcanic (?)
Meta-Sed.
Dacite
C.C. Green
C.C . Quartzose
Meta-Sed .
Grano-diorite
Andesite
C.C. Green
C.C. Conglomerate
Rhyolite
Greeny UM
Grano-diorite
C.C.Granitic
Greenstone
C.C. Green
Grano-diorite
Grano-diorite
Greywacke
Greeny UM
Grano-diorite
Grano-diorite
Grano-diorite
Meta-Quartz
C.C. Quartzose
Dacite
Greeny UM
C.C. Green

Note: C.C. =cache creek
metamorphics
Lithology

112

phi

-5.81
-5.85
-6 .37
-6.56
-5.81
-5.59
-5.62
-6.05
-6.50
-5.66
-6.44
-6.12
-5.87
-6 .38
-6 .10
-6.40
-6 .08
-6 .35
-5.87
-6.40
-6 .17
-5.81
-6 .33
-6.26
-5.77
-5.87
-6.26
-6.17
-5.88
-5.62

Rock#

147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176

56.09
57.71
82 .54
94.26
55.99
48.33
49 .11
66 .18
90.49
50 .69
86 .90
69 .58
58.54
83.19
68.76
84.61
67.74
81 .83
58.56
84.37
72.10
56.10
80 .24
76.45
54.56
58.66
76 .88
72.22
58 .85
49 .22

Equiv.
Sieve
Size (mm)
76.3
87.7
108.7
144.2
87.7
66.4
76.7
100.9
130.3
71 .2
126.3
98
94 .1
124.7
115.4
118.4
93.9
120.3
100.3
128.1
88.6
79.1
99.5
140.9
103
87.8
104.9
94.7
89.3
73.4
67.9
63 .8
92 .5
111.6
70
51
56.2
83.13
103.7
63.8
102.3
73.7
69.3
95
78 .2
102.1
78.5
95.2
66.7
94
82.8
67
91 .8
93.2
64.2
62.8
90 .8
80.7
72 .2
60.4
41
50.9
71.2
72 .9
37
45.5
40.8
43
75
32 .7
68 .1
65.2
45 .3
69.4
57.8
62.4
54 .9
65.8
49 .1
73.5
59.5
42 .5
66.7
54.8
42 .8
54.2
59 .8
62.6
41.4
34.6

Caliper Measure (mm)
A
B
c
0.69
0.78
0.80
0.69
0.61
0.85
0.73
0.61
0.75
0.62
0.71
0.84
0.68
0.74
0.72
0.69
0.74
0.73
0.71
0.77
0.79
0.70
0.79
0.61
0.65
0.81
0.72
0.80
0.65
0.65

Sphericity
8
8
14
11
5
15
7
7
10
6
8
10
5
13
3
8
3
19
5
7
8
5
13
11
8
8
13
7
6
4

ness

Round
-0.631
-0.884
-0.599
-0.602
-0.586
-0.630
-0.642
-0.625
-0.632
-0.678
-0.661
-0 .641
-0 .676
-0.711
-0.879
-0.729
-0.767
-0.603
-0.629
-0.619
-0.911
-0 .614
-0 .615
-0 .604
-0.621
-0.643
-0 .642
-0.784
-0.585
-0.821
-0 .610
-0.315
-0 .581
-0 .565
-0 .563
-0.616
-0 .621
-0 .606
-0.603
-0.599
-0.569
-0.583
-0.659
-0.670
-0.622
-0.678
-0 .662
-0 .582
-0.562
-0 .585
-0 .039
-0.594
-0.603
-0.592
-0.611
-0.617
-0.629
-0.547
-0.566
-0.606

-0 .618
-0.596
-0.592
-0 .577
-0 .575
-0.622
-0.634
-0 .616
-0 .618
-0.625
-0.614
-0 .618
-0 .670
-0 .680
-0 .733
-0 .695
-0 .699
-0.591
-0.610
-0 .603
-0.587
-0.606
-0 .609
-0.599
-0.616
-0.630
-0.635
-0.654
-0.577
-0.709

Magnetic Intensity
Min
Max
Avg
0.013
0.288
0.007
0.025
0.011
0.008
0.008
0.009
0.014
0.053
0.047
0.023
0.006
0.031
0.146
0.034
0.068
0.012
0.019
0.016
0.324
0.008
0.006
0.005
0.005
0.013
0.007
0.130
0.008
0.112
307 .5
352.7
877.1
1537.4
255.2
230 .5
216.5
442.0
1039.3
212 .3
1011 .6
613 .7
329.3
1151 .3
611.2
775.2
566.2
920.0
342 .9
964.1
485.6
265.5
661.4
909.4
286 .1
334.5
593 .7
327 .0
418.3
214 .6

Intensity
Weight (g)
Value

Grano-diorite
Meta'd UM
Porph Rhyolite
Grano-diorite
Dacite, w/quartz
Grano-diorite
Meta-Quartz Sandstone
C.C. Sed
Greywacke
Rhyolite
Dacite
Diorite
Quartzite
Meta-Greywacke
Diorite w/ Greenstone
Meta-Sandstone
C.C . Quartzose
Grano-diorite
C.C. Green
Meta-Greywacke
Greeny UM
Grano-diorite
Grano-diorite
Andesite
Meta-Quartz Sandstone
Dacite
Greywacke
Grano-diorite
C.C. Mafic
Dacite

Note: C.C. =cache creek
metamorphics
Lithology

113

phi

-5.93
-5.72
-6.01
-5.95
-6.19
-6.26
-6.19
-5.92
-6.14
-5.99
-5.54
-5.93
-5.69
-5.92
-5.81
-6 .29
-6.70
-6.13
-5.61
-5.62
-5.91
-5.76
-5.73
-5.67
-5.69
-5.67
-6.09
-6 .56
-6.16
-6.69

Rock#

177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206

60.85
52.76
64.43
61.78
73 .13
76.56
73.00
60.55
70.32
63.67
46.69
61 .02
51.64
60.60
56.02
78 .01
103.93
70 .03
48 .99
49.18
60.24
54.33
52.98
50.76
51 .52
50.99
68.26
94.57
71 .37
102.91

Equiv.
Sieve
Size (mm)

86
115
80.9
74.9
117.2
115.8
108.2
89.3
97.7
75.8
103.7
120.5
96.4
93.8
83.5
143.18
139.4
123.6
71
66.4
111 .3
81.2
83.1
117.6
71 .2
101 .9
90.6
137.8
127.5
145.7

76 .3
60.4
80 .7
69
75.2
97.4
91.4
75 .5
90.7
73.8
53.5
66.9
62.3
65.8
67.6
84.7
125.5
81 .2
61 .6
50.8
74 .9
74
63.6
51 .5
56.5
58.4
83 .7
118.4
83.5
111 .2
39.8
43.8
42 .3
53.6
71
47.3
48
40.4
40 .8
51 .6
38.7
54.5
38.1
54.9
41 .3
70.7
76.5
56.7
31.7
47.5
40.6
20.7
39.6
50
46
42.3
48.1
62.2
56.7
93 .9

Caliper Measure (mm)
A
B
c

0.63
0.65
0.65
0.82
0.83
0.59
0.62
0.63
0.58
0.78
0.65
0.72
0.63
0.79
0.67
0.75
0.70
0.69
0.62
0.88
0.59
0.42
0.67
0.75
0.81
0.67
0.68
0.62
0.67
0.82

Sphericity

5
8
7
13
10
9
3
8
5
12
6
7
3
14
3
5
21
15
3
10
3
1
6
3
5
3
6
9
8
6

ness

Round

-0.736
-0.731
-0.689
-0.691
-0.688
-0.710
-0.778
-0.750
-0.730
-0.773
-0.744
-0.890
-4.359
-0.850
-0.842
-0.811
-0.819
-0.784
-0.782
-4.994
-3.999
-0.771
-0.771
-0.772
-0.871
-0.932
-0.811
-0.799
-0.777
-0.792
-0.707
-0.706
-0.675
-0.681
-0.666
-0.686
-0.637
-0.600
-0.693
-0.509
-0.692
-0.808
0.785
-0.827
-0.825
-0.762
-0.778
-0.768
-0.773
1.556
1.226
-0.754
-0.760
-0.761
-0.608
-0.612
-0.793
-0.756
-0.732
-0.712
-0.720
-0 .717
-0 .680
-0 .686
-0 .674
-0 .697
-0 .696
-0 .692
-0 .714
-0.718
-0 .720
-0 .854
-0 .670
-0.839
-0.835
-0.788
-0 .801
-0.776
-0.778
-0.575
-0.816
-0.767
-0.766
-0.765
-0.748
-0.702
-0.800
-0.787
-0.753
-0 .758

Magnetic Intensity
Min
Max
Avg

0.016
0.014
0.009
0.005
0.014
0.013
0.082
0.058
0.016
0.055
0.024
0.036
3.689
0.011
0.007
0.023
0.018
0.008
0.004
4.419
3.183
0.004
0.005
0.007
0.123
0.230
0.011
0.012
0.024
0.034
315.1
411 .3
382.7
320.9
851.8
709.5
492.4
353.5
411.0
350.7
299.8
585.3
281.9
516.6
275.4
1235.0
192.6
918 .2
190.2
199.2
428 .5
133.2
293 .1
388.1
250.4
354.1
373.5
1226.3
911.3
1765.1

Intensity
Weight (g)
Value

Rhyolite w/Quartz bands
C.C . Quartzose
Grano-diorite
C.C. Volcanic
Grano-diorite
Grano-diorite
Meta-Sed .
Granitic (?)
Porph Rhyolite
C.C. Quartzose
C.C. Sed
Porph Dacite
Greeny UM
Diorite
C.C. Quartzose
Meta-Greywacke
Grano-diorite
C.C. Sed
C.C. Conglomerate
Greeny UM
Greeny UM
C.C. Volcanic
C.C. Sed
C.C. Quartz/Sed
Diorite
C.C. Quartz/Sed
Porph Dacite
Granitic (light col.}
Greywacke
Grano-diorite

Note: C.C. =cache creek
metamorphics
Lithology

114

phi

-6.42
-5.83
-5.72
-5.89
-6.03
-5.97
-5.49
-5.73
-5.69
-6 .04
-6 .16
-5.69
-5.97
-6.06
-5.93
-5.65
-5.57
-5.78
-5.82
-5.99
-5.63
-5.63
-5.70
-5.73
-5.72
-5 .88
-5.66
-5.59

Rock#

207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234

85.48
57 .05
52 .78
59 .34
65.32
62.64
45.02
53 .02
51.45
65.82
71.47
51.47
62.88
66.67
60.77
50.28
47 .37
54 .97
56 .31
63.51
49.65
49.44
51.83
52.90
52.81
58 .70
50.56
48 .15

Equiv.
Sieve
Size (mm)
121 .2
77.3
79.9
85.2
98.3
101 .1
83.9
92.4
69
99 .9
116.3
79.9
100.9
81.6
94.4
80.1
102
84.6
94.1
85.8
71.2
79.5
73.1
70
72 .9
83 .1
96 .9
80 .8
93 .6
65
61.7
72 .3
83.7
80.9
53 .9
63.1
50.9
84
84.4
52.8
73.6
77.7
68.8
55.3
59.6
69
66.5
74
55.2
57
59
61.7
62 .1
75
61 .3
52 .8
76 .5
47.8
42
42 .6
39.1
36.1
33.9
40.5
52
40.1
55.6
50.1
49.9
53.4
51.5
44.7
30.6
35.8
43.8
50.9
43.4
40.5
43.5
42.3
41.5
35.6
36.8
43

Caliper Measure (mm)
A
B
c
0.80
0.77
0.71
0.67
0.57
0.55
0.64
0.66
0.92
0.58
0.68
0.84
0.70
0.77
0.74
0.77
0.54
0.61
0.68
0.74
0.78
0.72
0.76
0.75
0.73
0.59
0.61
0.76

Sphericity
2
5
5
4
3
2
5
4
5
4
2
9
4
13
3
2
4
3
1
7
5
8
4
2
4
5
4
3

ness

Round
-0.727
-0.680
-0.738
-0.828
-0.769
-0.823
-0.780
-0.781
-1 .282
-0 .751
-0.815
-0 .776
-1.184
-0.753
-0 .742
-0.705
-0 .695
-0.748
-0 .681
-0.690
-0.609
-0.600
-0 .619
-0 .768
-0.766
-0 .772
-0 .851
-0.753
-0.717
-0 .670
-0.482
-0.626
-0 .597
-0 .624
-0.764
-0.746
0.064
-0.717
-0.483
-0.745
0.212
-0.737
-0.697
-0.685
-0.670
-0.559
-0.636
-0.465
-0.583
-0.577
-0.547
-0.751
-0.754
-0.736
-0.660
-0.730

-0 .723
-0.676
-0.669
-0.723
-0.705
-0 .783
-0.772
-0.768
-0.763
-0.733
-0.710
-0.764
-0.742
-0.744
-0.723
-0.697
-0.683
-0.681
-0.652
-0.616
-0.593
-0.590
-0.585
-0.761
-0.750
-0.756
-0.753
-0.743

Magnetic Intensity
Min
Max
Avg
0.004
0.004
0.069
0.105
0.064
0.040
0.008
0.013
0.519
0.018
0.105
0.012
0.442
0.009
0.019
0.008
0.012
0.067
0.029
0.074
0.016
0.010
0.034
0.007
0.016
0.016
0.098
0.010
668.8
311.5
297.4
301 .8
367 .7
388 .5
260.2
305.9
212 .8
396.4
550 .2
320 .3
585.4
426.8
388 .5
243 .8
316.8
270.5
390.5
450.7
265.3
287.1
232 .7
268.6
259.2
233 .0
261.5
239 .3

Intensity
Weight (g)
Value

C.C. Sed
Grano-diorite
Porph Dacite
C.C . Quartzose
Grano-diorite
C.C. Sed
Grano-diorite
C.C. Quartzose
Greeny UM
Andesite (pink,v.fine)
C.C. UM (?)
C.C. Quartzose, Porph
C.C . UM
Grano-diorite
Quartzite
Dacite (greenish)
Quartzite
C.C. Andesite
C.C. Quartzose
C.C. UM
C.C. Quartzose
Meta-Sandstone
Grano-diorite
Porph Dacite
C.C. Quartz/Sed
Dacite
C.C. Quartz/UM
Diorite w/Quartz bands

Note: C.C. =cache creek
metamorphics
Lithology

115

Appendix C - Laboratory calibration data

Tests were done on the BMD sensors using 6 rocks of varying size and lithology.
The rocks were affixed to a pendulum and passed over two sensors at varying speeds.
Three positions were tested: across the middle of sensor 1, at one quarter the width of
sensor 1 (toward sensor 2), and exactly between the two sensors. The position is indicated
by the picture at the top of each set of diagrams. The clasts were passed at three speeds:
1.4 rn!s, 0.8 rn!s and 0.4 rnls. Note that the voltage scale is the same for all diagrams.
The rocks used were as follows:
Sample

Weight

Short Circumf.

Long Circumf.

Rock Type

A

214 g

15.2 em

18.7 em

Metapyroxini te

B

388 g

17.5 em

22.8 em

Andesite

c

22.1 g

7.0cm

8.9cm

Serpentine(+?)

D

1355 g

30.4 em

34.9 em

Granodiorite

E

552 g

19.5 em

30.2 em

Greenstone

F

69.1 g

17.3 em

14.6 em

Serpentine(+?)

Table 10- Characteristics of rock samples used in pendulum experiments.

117

Rock A 1.4 m/s (5s interval)
1.50E-02
1.00E-02 ......................
S.OOE-03

-Sensor 1
----·- Sensor 2

O.OOE+OO
-S.OOE-03
-1 .00E-02

r----

--

_________ ___ _ _______________,
,

,

Rock A .8 m/s (5s interval)
1.50E-02 -, . - - - - - - - - - - - - - - - - - - - ,

S.OOE-03 -/-111- - - -1- - ---!11- - --f!- - - -n- 1

-S .OOE-03

.........________ ...................................u. ................ .......... . ................

-1.00E-02
-1.50E-02 ' - - - - - - - - - - - - - - - - - - - '
..........................................................................- ..........

_________________

,

..................... _ , __ ................,

Rock A .4 m/s (5s interval)
1.50E-02 ·-, ---

-

--------------,

S.OOE-03

-S.OOE-03
-1.00E-02 ·

························-····---

-1.50E-02 ..' - - - - - - - - - - - - - - - - - - - '

__

118

Rock A 1.4 m/s (5s interval)

1.00E-02
5 .00E-03

- --

-- -

- - - -- + - - --11----··---

-5.00E-03
-1 .00E-02
-1.50E-02 - - - - - - - - - - - - --- - ----·- - -·

Rock A .8 m/s (5s interval)
1.50E-02
1 .OOE-02
5.00E-03

-·--·---------··-··- ···-...···--····-------------··- ···--··-··-··-···-··--··-··-··--··-··-

~- ... . ...

~~
~

.....
·······_
- -N

..

- -~'.- ···_···-··_··············....q
······ ~~~

~ :~~

O.OOE+OO
-5.00E-03
-1 .OOE -02

·-------··--····--····-··········-··--··--····--······--··-··--··--····---··········--··--··------------·-·
··--····--

-1 .50E-02

Rock A .4 m/s (5s interval)
1 .50E-02 ·· · - - -....·---·--·-·----·-..--·--·--·----··---.._ ...._.._ ____ _
1.00E-02 ··
5 .00E-03 - ...._.._..____________.,__________..._..___________..._................--................____ _

-Sensor 1
- Sensor2

-1.00E-02 ·
-1 .50E-02 ' - - - - - - - - - - - - - - - - - - '

119

Rock A 1.4 m/s (Ss interval)
1.50E-02 · . . . - - - - - - - - - - - - - - - - - - ,
1.00E-02
S.OOE-03 · ...·--·---·-..................................................................................___ ____ ................
O.OOE+OO ·1-(f).....y---d>-;-,._....
·
·

'---

- '--- ~

•

Sensor 1
.......... Sensor 2

·

-S.OOE-03 -1- - - - - - - - - - - - - - --1
-1.00E-02 - -

- -- ····································-·················-··································-

-1.50E-02

· - - - - - - - - - - - - · - -.................- ...............

__ __
,_

.............

,,,

Rock A .8 m/s (Ss interval)
1.50E-02 , - - - - - - - - - - - - - - - - - - ,
1.00E-02 +-- - - - - - - - - - - - - ---1
S.OOE-03 ··

-

--· ··································

-

---...... ~---

-----

O.OOE+OO

--

-S.OOE-03

- - - - - · - · - - - - - - - - - - - - - --1

: ----

-

Sensor 1

--- Sensor 2

-1 .00E-02
-1.50E-02 -' - - - - - - - - - - - - - - - - - - l

Rock A .4 m/s (5s interval)
1.50E-02
1.00E-02

-------------·----·--····--------!

S.OOE-03
O.OOE+OO
-S.OOE-03 ·f--- - - - - - - - - - - - - - - 1
-1 .00E-02
-1.50E-02 _L,__ _ _ _ _ _ _ _ _ _ _ _ _ ___J

-

Sensor 1

-

Sensor2

120

Rock 8

1.4 m/s (5s interval)

1.50E-02 -r------------------..

-----------------------------------1

1

1.00E-02 ·
S.OOE-03 ··
O.OOE+OO -S.OOE-03 -

A
A
-~
--- ~ ---.....
.--- ~ ----- ~ 1
v
- Sensor 2
r
~

'

-1.00E-02 ··
-1 .50E-02 ··

Rock 8

.8 m/s (5s interval)

1.50E-02
1.00E-02
S.OOE-03
O.OOE+OO

,

l

lA

1l/

,
I

~

1j '

-S.OOE-03
-1.00E-02
-1 .50E-02

Rock 8

.4 m/s (5s interval)

1.50E-02 ·-------------------------·
1 .OOE-02 · ----------- ---------------------·-··--·
S.OOE-03
O.OOE+OO ·· .. •• • ~

'~ '

• .. •

.,.;v-.. - """

-S.OOE-03 ·
-1 .00E-02
-1 .50E-02 ·' - - - - - - - - - - - - - - - - - - - - '

121

Rock 8

1.4 m/s (5s interval)

1.50E-02 . . . - - - - - - - - - - - - - - - - - - ,
1 .OOE-02 · -- ··········-·············--···········
5 .00E-03 ··f
O.OOE+OO

················--···---·-----

·················································································································································································
l

-~~-~------~ -----~~----~~--~i ~ -- Senwr2
r-

-5.00E-03 · --·

•
•

y

Y

y

j

.......... ____ ··--·---···- · - - - - - -!

I

-1.00E-02 ................_. __________________________.............

·j

- ------------

-1.50E-02 - - - - -..·-·--·--

Rock 8

I

.~

.8 m/s (5s interval)

1.50E-02
1.00E-02
5 .00E-03
O.OOE+OO

v

...

v

-5.00E-03
-1 .00E-02
-1 .50E-02

· - - - - - - - - - - - ----------------------Rock 8 .4 m/s (5s interval)
1.50E-02 - - - - - - - - - - - - - - - - - - - - - - .
1.00E-02
5.00E-03 ----- - - - - - - - - - - - - - - - - -4

--Senwr 1
......... Senwr 2

O.OOE+OO
-5.00E-03
-1 .00E-02

____

__ _____________

-1.50E-02 · · ' - - - - - - - - - - - - - - - - - - - - '
.............

,

.......

122

Rock 8
1.50E-02

1.4 m/s (Ss interval)

·r-----------------,

1.00E-02 ·1--- - - - - 5.00E-03
O.OOE+OO

~~~~----~--~~~----~~--~ ~ --·-··· Sensor
~ 2

r

-5 OOE-03 ·1---- - - - - - - - - - - - - ---;
-1 .00E-02 · ························
··
-1.50E-02 ··' - - - - - - - - - - - - - - - - - - '

'--------------------------··--Rock 8

.8 m/s (Ss interval)

1.50E-02
1.00E-02
5.00E-03

---------------i

O.OOE+OO t -........

-

~

-- Sensor 2

-5.00E-03 ·!---- - - - - - - - - - - - - - - - ·
-1 .00E-02
-1 .50E-02

'----------------...-J
Rock B .4 m/s (5s interval)

1.50E-02
1.00E-02 ·!---- - - - - - - - - - - · - - - - - - - - - - - - !
5.00E-03 +
O.OOE+OO

·························
·······························································································
························
············.........

I

-1 ----------------1 [ : ~ ~~

-5.00E-03 +-- - - - - - - - - - - - - - - - - 1
-1.00E-02
-1 .50E-02 -' - - - - - - - - - - - - - - - - _ . . . J

123

Rock C 1.4 m/s (5s interval)
1.50E-02 . . . - - - - - - - - - - - - - - - - ,
1 .OOE-02

----- -

---------------- - - - - - ·- - - - -·

5.00E-03
O.OOE +00

· kfl---.,.P..--"11,!...__-"11..__""""".._,_---,ft I ~ ~~~~

-5.00E-03 -1-'--- - - - - - - - - - - - - - ' - - - ---l

-1 .50E-02 ·· -----···--·-····------------·
·-----··--···-····-··--·-··--·········
··

Rock C .8 m/s (Ss interval)
1.50E-02 · · r - - - -- - - - - - - - - - - - ,
1.00E-02 ··
5.00E-03

-5.00E-03 -1. 00E-02
-1 .50E-02

·····------------···············································-·······-···-····-······-·····-

Rock C .4 m/s (5s interval)
1 .50E-02 · -------·----··-·-·-·····-··--·--····
............~..······----·-·----·1.00E-02
5 .00E-03

-----------------------1
-Sensor 1
-·- Sensor 2

-5.00E-03
-1.00E-02
-1.50E-02 ' - - - - - - - - - - - - - - - - - - '

124

Rock C 1.4 m/s (5s interval)

I·

1.50E-02
1.00E-02
S.OOE-03

II

A _
O.OOE+OO hll---111-S.OOE-03

ll

1

__,,p..
i. _ _...,lp.-_._.,1..__-.JI!t ~ -

f

··-- ...........................

-1 .00E-02 ·f - · -

_

1

-

..........

J

.........

l

···-···-··-,

-

Sensor 2

··-··-..

. . . . . . .-----···-··. ·-··-·---·---------------1
J

-1.50E-02 ·

Rock C .8 mls (5s interval)
1.50E-02 , . . . . . . . - - - - - - - - - - - - - - - - .
1.00E-02

O.OOE+OO

--~--Y

~-~
- - --- ~'---'

V

Y

I,.............- .... -

Y

V

-5. OOE -03 f.....................- .......... _____ ...................... _ .......................................................................... 1
-1 .00E-02
-1.50E-02

Rock C .4 mls (Ss interval)
1.50E-02

............,

' ... . ..... ~ - Sensor 1

-----------------,

1.00E-02 1.............................................................................................................................................................................. ,
S.OOE-03 +-- - - - - - - - · - - - - · - - - 1

-S.OOE-03
-1. 00E-02
- 1.50E-02 · ' - - - - - - - - - - - - - - - - - '

-

Sensor 2

125

,------------------------------------------------Rock C 1.4 m/s (Ss interval}

1.50E-02

--

1 .OOE-02 · -----------------------------------S.OOE-03 - -----

- - ---------....------l ~ -

O.OOE+OO ··1-

1
-- Sensor2

-S.OOE-03 ·1- - - - - - - - - - - - - - - - - - - - - - - - - - - - - --1
-1 .00E-02 · -- ················································

-- --

-1 .SOE-02 ·'--------------------------------'

No discernible data at .8 m/s
1.50E-02 ··, -------------------------------,
1.OOE-02

------·--------------------------------------------

S.OOE-03 ·· ····································· ·
O.OOE+OO

·

................... ............ ........ ....... ............

-------

~:~:~~

--

-S.OOE-03 -1--------------------------------1
-1.00E-02

1··································································································
···················································································1

-1.50E-02 -L--------------------------------l
L-------·-··········---·------···········-·········--······------·--······················-----

No discernible data at .4 m/s
1.50E-02
1.00E-02 -1---------------------------------1
S.OOE-03

!-S.OOE-03 r-------------·-·----------------1
-1 .00E-02
-1 .50E-02 --~-----------------------------

sensor 1l

. . ~-~-~

126

••

Rock D 1.4 m/s (Ss interval)
1.50E-02 . - - - - : - - - - - - - - - - - - - - - - - .
I

1 .OOE-02 + ····-t-···········································
t ·········································
t ·······················································!
S.OOE-03 ·· ··
O.OOE+OO

--~- '------

...

---

' ---

-

--........

-S.OOE-03 ·1-- r- ------ -··- - · - - - - - - - -11-- ···- - -..;

!

Sensor 1
Sensor 2

L...................................................>

-1 .00E-02 ··
-1 .50E-02

---~--

- - - - -·- - - - - · -"'--·---'

Rock D .8 m/s (Ss interval)
1.50E-02 · · · , - - - - - - - - - - - - - - - - - ,
1 .OOE-02 !
S.OOE-03

['

............................

-~- ---- '

O.OOE+OO

.....

-1.00E-02 +
'

[

' '---.... '---~.......- I

j . . . . . . . . . . . . . ·-·

-S.OOE-03

-1.50E-02

·····································
······························································································································
I

\ . . . . . . . . . . . . . . . . . . . ..

····················
·······································
' ····································································
··················
·····'··········

------

-

-

Sensor 1
Sensor2

i

-----~-~-'

Rock D .4 m/s (Ss interval)
1 .SOE-02 · . , - - - - 1 .OOE-02 +

·································································································································································
I

S.OOE-03

-S.OOE-03
-1.00E-02 +

···············································
···································································································································
I

-1.50E-02 .1._- - - - - - - - - - - - - - - - - '

I

127

•

;

l

Rock D 1.4 m/s ( Ss interval)
1.50E-02 . , - - - - - - - - - - - - - - - - - - - ,
1.00E-02 ··
S.OOE-03 + ············+

.

' '---- '''' ''----~---- --- I

-

O.OOE+OO

·······

-S.OOE-03 ·r----11-- - - --n------+- - - - -11---

Sensor
Sensor 2

..........,

-1 .00E-02 - ------------------------- ---------------------·-..--..+ ----....--·---1

Rock D .8 m/s (Ss interval)
1.50E-02 . . - - - - - - - - - - - - - - - - - .
1.00E-02
S.OOE-03
O.OOE+OO
-S.OOE-03
-1.00E-02 ,
-1 .50E-02
....................................................,

........, ,

_____

1.50E-02

!

.....................................................................................................................

_____________

................,

,,,

_____

,,,

.................

________

,

...............

______

,,,

·---------------·-·
Rock D .4 m/s (Ss interval)

·r--------------------,

1 .OOE-02 +

........................................................................................

l

-Sensor 1
............ Sensor 2

-1 .00E-02
-1.50E-02 · ' - - - - - - - - - - - - - - - - - - - '

128

Rock D 1.4 m/s (5s interval)
1.50E-02 · · , . . - - - - - - - - - - - - - - - - - - - . ,
1.00E-02 -1- - - - - - - - - - - - - - --1

O.OOE+OO ··Kfi----··W·-..--<tl-<---rtr - - - - i

~ ---..-..... ~:~:~ ~

-S.OOE-03 ++ - - --+-- - - -fl-- - - -1' - - ---l
-1.00E-02 ·
-1. 50E-02 ..__ _ _ _ _ _ _ _ _ _ _ _ _ ___..

Rock D .8 m/s (5s interval)
1.50E-02 - , . . . - - - - - - - - - - - - - - - - - ,
1.00E-02 +-- - - - - - - - - - - - - ---l
S.OOE-03 ··

-Sensor 1
--·. Sensor 2

-1 .00E-02

-

-1 .50E-02 - ' - - - - - - - - - - - - - - - - - - '

Rock D .4 m/s (5s interval)
1.50E-02 ·· . - - - - - - - - - - - - - - - - - ,
1.00E-02 +--- - - - - - - - - - - - - - ----1
S.OOE-03

- Sensor1
---·- Sensor 2

-1 .00E-02 ··
-1 .50E-02 - ' - - - - - - - - - - - - - - - - - - '

129

RockE 1.4 m/s (Ss interval)
1.50E-02 . - - - - - - - - - - - - - - - - - ,
1 .OOE-02 · ····

· ····- -- -

---

. . . . . . . . . . . !I

S.OOE-03 ··
O.OOE+OO -1-""""''---...._......!4--...-,r-P---"---t+l----l 1' -S.OOE-03 -f -11- - --llf- - --41- - - - -11-----1
-1.00E-02
-1 .50E-02

-------------------------

RockE .8 m/s (5s interval)
1.50E-02
1.00E-02
S.OOE-03

- - - ---

-S.OOE-03
-1.00E-02
-1 .50E-02 · ------- -···...··-·--·-----

-----

RockE .4 m/s (5s interval)
1.50E-02

-S.OOE-03 ·
-1.00E-02 ·

------- ·······················································

-1.50E-02 · · ' - - - - - - - - - - - - -- - ----'

Sensor 1 .

~~~

l

130

RockE 1.4 m/s (Ss interval)
1.50E-02 -. - - - - - - - - - - - - - - - - - - - ,
1 .OOE-02 + ---·-.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .--...........1

•

5.00E-03 ···························

..\

O.OOE+OO
-5.00E-03
-1.00E-02
-1.50E-02

RockE .8 m/s (Ss interval)
1.50E-02 . - - - - - - - - - - - - - - - - - - ,
1.00E-02
5 .00E-03

·········--···-·-····-···-······

O.OOE+OO
-5.00E-03

~-~_

: ~~

_\,"!""""""'"""! [

-1.00E-02
-1 .50E-02
______
___________
,,_,

.................

.._ , , , ........................

______ ___
,_,

,_,,

_____

,

____,

____________ ................-----------------------------------·
,

RockE .4 m/s (Ss interval)
1 .50E-02

------------------..------------·----

1.00E-02

································································
·

5 .00E-03 - f - - - - - - - - - - - - - - - - - 1

-Sensor 1
_ ....... Sensor 2
-5.00E-03 ,

..................................................................................................................................................................................,

-1 .00E-02 ·· ···
-1 .50E-02 - ' - - - - - - - - - - - - - - - - - - - - '

131
I

I

I

----

RockE 1.4 mls (Ss interval)
1.50E-02

I
I

I

1.00E-02

·-·----·

5.00E-03

···-·······

-Sensor 1
--- Sensor2

.il
~

O.OOE+OO ·
-5.00E-03

I

-1.00E-02 - ········
-1 .50E-02
-----·-------

I

RockE .8 mls (5s interval)
1.50E-02

I

1.00E-02
5.00E-03 ··

I

O.OOE+OO
-5.00E-03

I

--··-

-1 .00E-02
-1.50E-02 ·

-···-------··------------------ -------·--------------------------------------

I

Rock E .4 m/s (5s interval)

I

1.50E-02
1.00E-02 --·-··--·- - - - - - -·-··--··· ·------

I

5.00E-03

I

O.OOE+OO ··

I

-5.00E-03 +-1 .00E-02
-1 .50E-02 -

I
I

I

I

-----------

-------------··-------··-- ···········-··-··-·

-Sensor 1
····--Sensor 2

132

I
!
Rock F 1.4 m/s (5s interval)
1.50E-02

I

1.00E-02 ··········-· ..
S.OOE-03

I

I

-Sensor 1
- Sensor2

O.OOE+OO ·
-S.OOE-03

I

I

-1.00E-02 · -- -······
-1 .50E-02

·------

._L.___ _ _ _ _ _ , _ _ _ _ _ _ _ _ _

I
Rock F .8 m/s (5s interval)

I

1.50E-02 ··
1.00E-02

I

S.OOE-03

-

--

• ····························
l ·······...............•....

O.OOE+OO
-S. OOE-03

II

......... -·-·····-.-

-1 .00E-02 ·
-1 .50E-02 ··-··-··

I

-··-·--······

--···------·-·--·--·-·-

· - - - - - - - - - - - -····---·······
···

Rock F .4 m/s (5s interval)

I

1.50E-02 ··----------1.00E-02 ··

I

S.OOE-03 ·
O.OOE+OO ··

I

~

-

Sensor 1
Sensor2

-S.OOE-03 ··
-1 .00E-02 ··

I
I
I

I

-1 .50E-02 ·
···--·----·-·--··----------··-·····-··············--·--·-····
··························-·········

133

Rock F 1.4 m/s (5s interval)
1.50E-02 . . - - - - - - - - - - - - - - - - - - ,
1 .OOE-02 + ·············································-········-···························-·····································································
·
i

,....... .

S.OOE-03
O.OOE+OO

'---

~----'

'---

--

--

-

~~
Sensor 2

-S.OOE-03 -1 .00E-02 ·f

·······--------------·--·-···---------1

-1 .SOE-02 - ~- -

--------- --

-

-

-

--

-

----

-----'

Rock F .8 m/s (5s interval)
1.50E-02
1.00E-02
S.OOE-03

~

O.OOE+OO

~

-S.OOE-03

~

A

v

v

h

v

A

v

-1.00E-02
-1 .50E-02
---------------------------'

-

-----------------------Rock F .4 m/s (5s interval)

1.50E-02 ----------------------------------------------..,
1.00E-02
S.OOE-03

v

-S.OOE-03 +

··································································································································································
·I

v

-1 .00E-02
-1.50E-02 - L - - - - - - - - - - - - - - - - - '

I

-Sensor 1
--- Sensor 2

O.OOE+OO