Coarse Sediment Transport Measurement in Rivers and on Coasts Using Advanced Particle Tracing
Technologies. EPSRC Review Report for GR/L94987/01.
D.A. Sear1, M.B.Collins2, P. A. Carling1, M.W.E.Lee1, R.J.Oakey3
1. Dept. of Geography, University of Southampton, Highfield, Southampton, SO171BJ, UK.
2. School of Ocean & Earth Sciences, University of Southampton, Southampton Oceanography Centre,
European Way, Southampton, SO14JZH, UK.
3. Dept. of Geography, University of Lancaster, Bailrigg, Lancaster, LN142LX, UK.
Background and Context for the Research
The evolution of morphology and physical habitat in river and coastal environments is linked intrinsically to the
processes of sediment transport. In many cases, these environments are dominated by coarse sediments (here,
defined as particles having a diameter of 8 mm and greater), (Leopold, 1992; Rosen & Brenninkineyar, 1989).
Understanding coarse sediment transport processes has direct economic and ecological benefits, via
improvements in the prediction of changes in physical habitat and the location and magnitude of erosion and
deposition; the latter accounting for substantial costs in terms of land loss and disruption of infrastructure (Sear et
al, 2000; Cooper, 1996).
The study of sediment transport processes has become increasingly sophisticated (McEwan et al, 2001;
Habersack, 2001), with progress in both littoral and fluvial research converging on the prediction of sediment
transport rates; this has taken place through improvements in the physically-based understanding of entrainment,
transport and deposition. Sediment tracing provides a non-invasive, cost-effective approach to the determination
of sediment transport rates (Haschenburger & Church, 1998; Habersack, 2001). A major review of published
literature has been undertaken as part of this project (Sear et al., 2000). The principle behind tracer studies is to
introduce material that is easily distinguishable from the natural sediment, but behaves similarly, into an
environment, and to monitor it’s behaviour. In practice, several types of tracer study exist (Madsen, 1989): the
present study focuses upon the Spatial Integration Method (SIM). This approach is the most widely and simply
deployed method, in both littoral and fluvial tracing studies (Lee et al, 2000; Sear et al, 2000). Three quantities
need to be measured during a SIM tracer study: (1) the distance travelled by the tracer’s centre of mass or volume
(centroid), during a given time period (allowing the velocity of longshore or downstream movement (vb) to be
determined); (2) the thickness of the active layer (ds) (deVries, 2001) and (3) the width of the active layer (ws)
(Ashiq, 1999). The relationship is represented mathematically as:
Qb = vb. ds. ws
(1)
where Qb is the bulk transport rate (downstream or longshore). Typically, estimates of sediment transport rate
have been empirically related to certain hydrodynamic variables: in fluvial environments, maximum stream
power; and on beaches, the longshore component of wave power.
To achieve a more detailed, physically-based understanding of the relationship between hydrodynamics and
particle movement requires improvement in the resolution of particle location in time and space. This can be
achieved through the development of particle tracking technology (Habersack, 2001; Ergenzinger et al, 1989). To
date, tracking technology has been limited to relatively few (< 10) particles, with limited spatial resolution (+/- 12 m) almost exclusively in fluvial environments. However, a limited experiment in acoustic particle tracking has
been undertaken in an estuarine environment (Dorey et al, 1972).
On the basis of the above, the primary objectives of the present research project were :
1) to develop tracer theory in support of the collection of field-based reliable estimates of bulk sediment
transport rates using the SIM method (Objectives 2, 4 & 5); and
2) to develop a multiparticle tracking technology, for deployment in littoral and fluvial environments
(Objectives 1, 3, 6).
Key Advances and Supporting Methodology
Objectives 2,4 & 5: Compilation of robust sediment transport datasets (including hydrodynamic and
sedimentological data) from littoral and fluvial field environments for use in model development/verification.
Development of tracers that better represent the indigenous grainsize population (D16-D84).
A major element of the derivation of robust sediment transport datasets was the consideration of detailed
methodological issues. These included: (1) the number of tracers used in determination of transport (2) selection
of appropriate measures of vb, ds and ws used in equation 1; (3) assessment of the potential influence of size,
shape and injection position on transport rate ; and (4) correct representation of particle size and shape.
Figure 1a: Shoreham-by-Sea
1b: Highland Water
1c: River Lune
Three field sites were used to achieve Objectives 2, 4 and 5. A single field site at Shoreham-by-Sea (Fig. 1a)
provided littoral field data, whilst two, the Highland Water (Fig.1b) and the River Lune (Fig. 1c), represent
contrasting coarse sediment fluvial environments (Table 1).
Field Site
NGR
Basin Area
(km2)
Highland Water
264083
11.5
River Lune
658052
26.0
NGR
Beach Length
(km)
Shoreham-by-Sea 523104
1.5
Table 1: Summary field site data
Qbankfull
(m3s-1)
2.80
22.5
Spring Tidal
Range (m)
5.7
Slope
Width (m)
0.0073
0.0048
Beach
slope
0.116
3.5
10.6
Beach Width
(m)
49.0
D16
(mm)
9.9
9.9
D16
(mm)
3.9
D50
(mm)
32.8
57.2
D50
(mm)
11.7
D84
(mm)
46.3
137.2
D84
(mm)
32.0
Compilation of field datasets, during this project, may be summarised as follows:
• Review of 23 littoral and 51 fluvial tracer studies, incorporating (wherever possible) hydrodynamic, together
with sediment transport measurements (Sear et al, 2000, Project Website).
Ω
Recovery
No. Mobile
Qs
SITE/EVENT/ Experiment/ Qmax
(Wm-1)
(%)
Tracers
(kgs-1)
No. TRACERS Technology (m3s-1)
HW/1/ – 302
SSLN/Al,F
1.19
71.2
100
46
0.0018
HW/2 /– 302
SSLN/Al,F
0.94
63.6
100
22
0.0035
HW/3 /– 302
SSLN/Al,F
2.36
201.8
99
302
0.0078
HW/4 /– 302
SSLN/Al,F
0.80
52.5
100
26
0.0135
HW/5 /– 302
SSLN/Al,F
3.68
319.5
91
261
0.0592
RL/1/ – 336
SSL/Al,Mg
39.32
1853.5
91
208
0.0975
RL/2/ – 426
SSL/Al,Mg
Tracers Lost Tracers Lost Tracers Lost Tracers Lost Tracers Lost
RL/3/ – 426
SSL/Al,Mg
14.95
704.1
100
35
0.0123
RL/4/ – 231
SSL/Al,Mg
19.57
922.5
100
48
0.0090
RL/5/ – 231
SSL/Al,Mg
35.71
1683.3
89
110
0.2088
RL/6/ – 231
SSL/Al,Mg
55.21
2602.5
77
160
0.2852
HW = Highland Water, Hampshire; RL = River Lune, Cumbria., SSL = Size, Shape & Location influence on Transport
rate, N = Tracer Number Experiment, Al = Cast Aluminium Tracer, F = Foil Wrapped Tracer, Mg = Magnetic Insert
Tracer.Note: Calibrated CFD Data is also available for both field sites.
Table 2: Summary data for the fluvial field experiments undertaken during the project
•
•
Collection of new field data, covering 11 sediment transport events of contrasting magnitude and duration,
from two different fluvial environments: sediment transport rates, hydrodynamics, topography and
sedimentology were measured (Table 2).
Collection of new field data, covering 14 sediment transport events, using a combination of electronic and Alfoil particles: sediment transport rates, hydrodynamics, topography and sedimentology (Table 3).
Pl
Wave Angle
Qs
SITE/EVENT/ Experiment /
Recovery
Il
(%)
(Wm-1)
(Degrees)
(m3hr-1)
No. TRACERS Technology
(Ns-1)
SBS/1 - 45
SSLN/F
100
-2.31
-194.85
13 SSE
-0.77
SBS/2 - 45
SSLN/F
89
28.49
332.54
10 SSW
9.49
SBS/3 - 85
SSL/TP
96
0.57
-167.45
13 SSE
0.19
SBS/4 - 85
SSL/TP
99
0.54
-77.19
11 SSE
0.18
SBS/5 - 85
SSL/TP
91
0.57
-20.83
4 SSE
0.19
SBS/6 - 82
SSL/TP
82
4.14
0.00
0
1.38
SBS/7 – 60
SSLN/F
92
2.19 -4221.49
10 SSE
0.73
SBS/8 – 60
SSLN/F
93
3.45
-78.64
7 SSE
1.15
SBS/9 – 60
SSLN/F
90
3.72
52.20
5.5 SSW
1.24
SBS/10 – 90
SSL/TP
100
0.12
-23.02
6 SSE
0.04
SBS/11 – 90
SSL/TP
100
-0.15
-36.36
11 SSE
-0.05
SBS/12 - 90
SSL/TP
100
0.15
-292.44
9 SSE
0.05
SBS/13 – 74
SSL/TP
92
67.87
575.29
7 SSW
22.61
SBS/14 – 71
SSL/TP
94
78.20
82.99
1 SSW
26.05
SBS = Shoreham-by-Sea, E.Sussex. SSL = Size, Shape & Location influence on transport rate, N = Tracer
Number Experiment, TP = Transmitting Pebble, LP = Logging Pebble, F = Foil Wrapped Tracer. Negative
values of Qs and Il indicate transport towards West. Negative values of Pl indicate wave approach from SE.
Table 3: Summary data for the littoral field experiments undertaken during the project
•
•
Acquisition of data showing intra-tidal particle movement, for 5 (newly-developed) logging pebbles over a
single tidal cycle: simultaneous hydrodynamic measurements were made (See Objectives 1,3 & 6).
Tracer recovery rates in all deployments were greater than 70% and for most, exceeded 90%.
Five different tracer technologies were utilised in this project (Table 4). Three of the five technologies were able
to represent the 16th percentile of the indigenous material at the study sites. A 56% reduction in the size of
existing transmitting pebbles was achieved, extending the grainsize range for this technology. Aluminium foilwrapped tracers developed during this project, provide a tracer that enables 3-D detection at low cost, with high
grainsize/shape representativeness.
Tracer Technology
Size Range Detection Depth
Cost / Tracer
Littoral (L) /Fluvial(F)
(m)
(£)
Logging Pebble (L)
>50mm
0.7m *
500+**
Transmitting Pebble (L)
>14mm
0.7m *
25**
Magnetic Pebble (F)
>8 mm
0.4m
0.75
Cast Aluminium Pebble (F)
>4 mm
0.4m
4.0**
Foil-Wrapped Pebble (F/L)
>4mm
0.3m
<0.3
* Average figure - detection possible up to 1.2m. ** Dependent on numbers.
Table 4: Tracer technologies deployed during the project.
Manufacture
Complexity Rating
High
High – Moderate
Moderate – Low
Moderate
Low
Use
Complexity Rating
High
Low
Low
Low
Low
Theoretical advances focussed initially on manufacturing tracers that represented the indigenous material in terms
of size and shape. The accurate measurement of bulk grainsize was undertaken, for each study site, using the ISO
lower precision curves published by Church et al, (1987). Sizes that represented the 16th, 50th and 84th percentiles
were identified and manufactured where tracing technology allowed. A new, methodology for the determination
of representative tracer shape was developed by the Lancaster studentship (Oakey et al., in review). This
methodology can be used to determine the shape of tracers in any percentile size class; as such it is reproducible,
statistically representative of the river bed material and has minimal bias.
Error analysis from this research and others (Haschenburger & Church, 1998) demonstrate that over 50% of the
total error in SIM estimates of sediment transport rate in fluvial environments, are derived from the estimation of
travel distance vb and ds. Thus a statistical approach was identified for the estimation of the tracer numbers
required to obtain measurements of travel distance, (L) with a given level of accuracy (Lee et al, in review).
Transport rate (kgs-1)
Analysis of the tracer data shows that for the conditions studied, typically 700 tracers are required in fluvial
environments and < 500 in littoral environments in order to achieve estimates of L which are accurate to within
10%. There was no correlation with event magnitude, but as tracers become more dispersed by successive events,
progressively higher number of tracers
will be required. This methodology can
1
Average Values Method
be used to check the minimum
95% Method
accuracy of estimates after a tracer
Max Depth Method
Mean Max. Burial Depth Method
study. In all the experiments in this
Area Method
project, accuracy of measured L was to
Trapped Data
Predicted Trap Data
within +/- 20% attaining and </= 10%
0.1
for three experiments.
The third area of tracer theory explored
the detailed methodology for deriving
the virtual velocity (vb), burial depth
(ds) and active width (ws) for use in the
0.01
SIM. Different methods were used,
and tested against measured transport
yields recorded from traps installed in
the Highland Water fluvial site (traps at
0.001
the Lune field site were not a viable
option). Figure 2, demonstrates the
variability in calculated transport rates.
Unfortunately, for the larger events, the
traps became over-filled. However, a
0.0001
log-linear relationship (r2 = 0.846,
10
100
1000
p<0.001) based on sediment transport
Stream Power (Wm-1)
data from 46 flood events for a reach
350m downstream of the study site,
Figure 2: Variation in predicted transport rates according to method of
enabled a conservative estimate of the
SIM used. Data compared to observed (trapped) sediment. Highland
transport rates to be made (Figure 2).
Water Field site.
Three points emerge from this analysis:
First, there is up to 3 orders of
magnitude variability in transport rate depending on the means of estimating vb , ds or ws for use in the SIM.
Secondly, assuming the trapped data is the “real” value, then the Average Values Method (AVM) which is based
on the average values of ws, ds and vb provides the
most consistent, and accurate predictions of
“observed” transport rates. The 95% Method
(95M), modified from Bray, (1996), also
consistently performs better than other methods.
The Maximum Depth Method (MDM), which is
based on maximum ds and averages of all other
values, over-predicts in all cases. In subsequent
analysis of data, the AVM was applied. Thirdly, the
values predicted by the AVM and 95M are close to
the measured values, demonstrating that the SIM
method, when applied carefully, is a valid method
for estimating sediment transport rates across a
range of stream powers in fluvial environments.
Figure 3: Examples of tracer forms used in fluvial field
The fourth area of tracer theory explicitly tackled in
experiments.
this research was investigation of the effects that
tracer size, shape and deployment location (crossstream or cross-shore) can have on the estimation of transport rates. In each fluvial environment, tracer forms
were constructed that represented the 16th, 50th and 84th percentile diameters wherever possible. In addition 3
shapes were scaled so that each form in each percentile had the same volume and mass (Figure 3). In the fluvial
environments, the resulting 9 forms were deployed in three injection beds located at 25%, 50% and 75% of the
channel width, each contained the same number of tracers and distribution of forms. Tracer form (size and shape)
and injection bed were tested in a 2-Way ANOVA for significant effect on the two dominant determinants of
transport rate vb and ds (Table 5).
Hypothesis
HW4
HW2
HW1
HW3
RL3
HW5
RL4
RL5
RL1
RL6
Reject
(<0.001)
Reject
(<0.001)
Reject
(0.018)
Accept
(0.096)
Reject
(0.004)
Reject
(0.015)
Reject
(<0.001)
Accept
(0.087)
Reject
(<0.001)
Reject
(<0.001)
Reject
(<0.001)
Reject
(<0.001)
Cannot
Test
Cannot
Test
Cannot
Test
Cannot
Test
Cannot
Test
Cannot
Test
Cannot
Test
Cannot
Test
Transport Distance
No difference between
Injection sites
No difference between
forms
Reject
(0.008)
Accept
(0.879)
Reject
(0.002)
Cannot
Test
Reject
(<0.001)
Cannot
Test
No difference between
Injection sites
No difference between
forms
Reject
(0.031)
Accept
(0.032)
Cannot
Test
Cannot
Test
Cannot
Test
Cannot
Test
Reject
(<0.001)
Reject
(<0.001)
Accept
(0.880)
Accept
(0.519)
Tracer Burial Depth
Reject
(0.032)
Reject
(0.807)
Cannot
Test
Cannot
Test
Table 5: Results of 2-Way Analysis of Variance (ANOVA) test for the influence of tracer form (size & shape) and injection
site on transport distance and burial depth. HW1-5 Highland Water Event 1-5, RL1-6 River Lune Event 1-6. Events are
ordered in increasing stream power Left - Right.
The main results may be summarised as follows:
1) Tracer injection position is a significant control on both travel distance and burial depth in the rivers
studied, and is independent of event magnitude.
2) Tracer form (size and shape) has an event specific control on travel distance in both rivers studied;
irrespective of event magnitude. Tracer form also influences burial depth in the Highland Water field site, but
this is again, event specific. The behaviour of forms was not significantly different between injection beds.
These results are directly comparable with previous tracer theory studies undertaken on shingle beaches (Lee et
al, 2000, Bray, 1996), and confirm that tracer injection site, rather than tracer form, is the most consistent
influence on estimated transport rate. However, it is clear that the use of the SIM for estimating sediment
transport rates, must replicate not only spatial controls but also particle size and shape.
The design of the littoral “size, shape and injection position experiments” matched that of the fluvial studies as
closely as possible. Due to the large burial depths experienced during previous experiments (eg Lee et al., 2000),
the optimum tracer technology for use at the chosen site was identified as the electronic (transmitting) pebble.
Miniaturisation of the circuitry used within this technology (during the course of this project) allowed 35.5% of
the indigenous grainsize distribution to be represented. In common with the fluvial experiments, 9 different tracer
forms were manufactured. Tracer injection was carried out in two different ways during the course of the study.
On 3 occasions, injection was at 3 different cross-shore sites; and on 4 occasions it was at a single site (positioned
at 50% of the active beach width). The single injection site was sometimes used as a consequence of fewer tracers
being available for use in the experiment than had been anticipated (see the Project Plan Review section): it
allowed the number of particles available for assessment of the influence of form (size and shape) to be
maximised, at the expense of information on the influence of injection position. The fact that the littoral
experiments could only be undertaken once the fluvial work was completed (due to logistics), has meant that
analysis of the littoral data is less advanced than that for the fluvial sites. Findings to date can be summarised as
follows:
1) When longshore transport rate is high, large clasts tend to move more quickly than small clasts, while the
reverse is true when transport rates are small (see Fig. 4a). These findings are in agreement with those of
previous studies (eg Jolliffe, 1964; Miller, 1997). Explanations of such results have, in the past, focussed
upon the thresholds and relative exposures of clasts of different sizes (eg Richardson, 1902).
2) Spherical clasts move alongshore more rapidly than discs or rods when overall transport rates are moderate or
high: when overall rates are low, spheres and discs move at a similar speed (greater than that of rods). The
relative speeds of rods and discs show no consistent behaviour (see Fig. 4b). This supports the findings of
previous studies (Lee, submitted).
3) Material on the lower beach tends to move alongshore more slowly than that on the upper beach, regardless
of whether overall transport rates are high or low (see Fig. 4c). These findings are logical from the point of
view that, when the tide is low, wave energy is dissipated on the sand platform immediately to seaward of the
shingle portion of the beach: however, they contradict the results of some previous investigations (eg Lee,
submitted); although not all (eg Bray, 1996 and Kidson and Carr, 1961).
Longshore Centroid Displacement (m
Longshore Centroid Displacement (m
Longshore Centroid Displacement (m)
(4) No consistent trends are apparent with respect to the influence of injection position or clast shape on burial
depth (Fig. 4f & e). In contrast, medium sized particles tend to be buried less deeply than either large or small
clasts (Fig. 4d). Preferential burial of small particles may be due to vibration of grains within the bed
resulting from
wave induced
100
100
100
pressure
gradients
(Madsen, 1974),
10
10
while more
10
rapid settling of
1
1
large material
may account for
Small
Sphere
Upper
Medium
Rod
Mid
its observed
Large
Disc
Lower
0.1
0.1
1
recovery depths
-400 -200
0
200 400 600 800
-400 -200
0
200
400
600
800
-6000
-4000
-2000
0
2000
(Miller, 1997).
Pl (Wm )
Pl (Wm )
Pl (Wm )
-1
-1
a
-1
b
c
These findings
demonstrate the
importance of
accurate
representation of
particle size,
10
10
10
shape and the
active beach
Sphere
Small
width within
Upper
Medium
Rod
Mid
littoral tracer
Large
Disc
Lower
1
1
1
studies if
-400 -200
0
200
400
600
800
-400 -200
0
200
400
600
800
-6000
-4000
-2000
0
2000
Pl (Wm )
Pl (Wm )
Pl (Wm )
reliable transport
rates are to be
d
e
f
derived. They
Figure 4: Variations in longshore centroid displacement and burial depth with longshore
are also of great
component of wave power, illustrating the influence of tracer size (a & d), Shape (b &e) and
value in applied
cross-shore position (c & f).
coastal
engineering, (eg
in the choice of the optimum size and shape of material with which to carry out a beach replenishment, and in
choosing the optimum cross-shore position for its placement).
Longshore Centroid Displacement (m)
100
Longshore Centroid Displacement (m
-1
-1
Using the hydrodynamic and sediment transport data collected
during the littoral field deployment an attempt was made to
calibrate the CERC longshore transport model (Shore Protection
Manual, 1984). Plotting, measured, immersed weight longshore
transport rate (Il) against the longshore component of wave
energy flux (Pl) indicated that the underlying assumption of the
CERC model (that the two variables are linearly related) does
not hold (see Fig. 5). Non-linear relationships between the two
variables have been identified before (eg Workman, 1997 and
Bray, 1996). However, the data presented in Figure 5 are
particularly unusual, in that they indicate little sediment
transport on all occasions when wave approach was from the SE
(i.e. -ve values of Pl). It is possible for tidal currents to result in
asymmetry of transport rates along a beach but tides alone seem
unlikely to account for the magnitude of asymmetry observed.
Further analysis of the hydrodynamic data is required and will
be undertaken as part of the project dissemination. It is
anticipated that the littoral data set will be of great value for the
-1
90
80
70
60
-1 -1
Longshore Centroid Displacement (m)
100
Il (kgm s )
100
50
40
30
20
10
0
-10
-5000
-4000
-3000
-2000
-1000
-1
Pl (Wm )
0
1000
Figure 5: Attempted calibration of the
CERC equation using Shoreham-by-Sea
field data. Linearity assumption is shown
to be invalid.
calibration and development of longshore transport models; only the study of Lee (submitted) is thought to have
been of equivalent quality.
In addition to the hydrodynamic measurements made using pressure sensors and current meters during the littoral
deployment, the use of the SHF Wave Radar to measure angle of wave approach to the shore was also successful.
Further, analysis of the data collected has indicated a new method by which average wave height might be
derived using the instrument. The approach is currently being applied to the dataset collected at Shoreham.
Objectives 1, 3 and 6: Production of a multi-particle (c.250 clasts) tracking system for coarse sediment
transport monitoring, capable of acquiring positional information in 2-D / 3-D and acquisition of intra-event
data on coarse particle motion in littoral and fluvial environments.
The project has developed and deployed a
novel multi-particle tracking system, suitable
for use in littoral and fluvial environments.
The results provide data on the first ever
tracking of coarse gravel/cobbles from within
swash, surf and breaker zones, over a tidal
cycle.
Unlike previous methodologies
(Ergenzinger et al, 1989; Habersack, 2001;
Dorey et al, 1972), the technology is based
upon the pebble receiving and storing
transmitted signals from a grid of wires
buried within, or suspended above, the beach
/ river bed (Fig. 6). This arrangement has the
advantage of minimising the requisite onboard power supply within the pebble (size
reduction), whilst allowing considerable
power to be supplied to the transmitting
coils. The methodology, in theory, permit the
deployment of an infinite number of
particles.
The system, as developed, is based upon the
generation of magnetic fields, by loops of
wire buried within the beach or river bed
(Fig. 6). If the power supplied to a wire loop
is known, then it is possible to calculate the
electro-magnetic field at any position and use
this to estimate the location of an object that
records field strength.
The tracking system uses four independent,
rectangular, wire transmitting loops (Fig. 6).
Figure 6 : Transmitting grid for logging pebble tests as deployed
These transmit, in sequence, with a pause
at Shoreham-by-Sea field trials.
between sequences. The total sequencing
takes 2.88s. The logging pebble detects the
transmitted field, using 3 orthogonally-mounted receiving coils of 0.03m diameter (Fig. 7). Field strength, time,
battery output and tilt-switch output (a measure of whether the pebble is moving or not) are recorded to nonvolatile EPROM memory circuits and downloaded via a resin-covered serial port on the pebble (Fig. 7). The
logging pebble was configured to record every 6 s; this provides a memory life of 8 hours, and a battery lifetime
of 180 hours. The pebble could be re-programmed, to provide different logging sequences; these could extend
operational lifetime to over 1 month. This approach is particularly suitable for use in fluvial deployments, where
flooding may occur on an irregular basis.
Once downloaded, the data are filtered (usable data are only recorded when the pebble is stationary for a whole
loop sequence); and processed via a custom-built MATHCAD programme, which converts stored data into
electromagnetic field strength. The pebbles position is then calculated via a series of sub-routines. The data are
exported subsequently as a single file, containing
time (s), and calculated x, y position (m). A full
description of the software processing protocols
and hardware are available upon request from the
P.I.’s.
Field trials were undertaken initially on land. The
prototype circuit was moved to fixed locations,
within and outside of the transmitting loop, and the
signal strength was recorded. These trials
demonstrated the feasibility of the system,
recording positional errors of +/-0.08m, at a
maximum distance of 8m from the transmitting
loop. Subsequent land-based tests, using 4 loops
and a 3-coil detecting pebble, gave rms errors of
0.1m and 0.07m, in x and y directions, respectively.
Figure 7: Logging pebble components, illustrating from
R-L: circuitry and receiving coils; encapsulated circuitry;
and data download jig.
Cross-shore distance (m)
Beach trials were undertaken on 4 separate
occasions using, initially, a single loop/coil set-up. Initial tests undertaken at Hordle Beach (Southern UK) were
affected significantly by the presence of buried metal, or cabling. Subsequent tests using a circuit fixed, in place
within a single transmitting loop, demonstrated that at peak immersion by sea water (1.0m depth), the field
magnitude was increased by 1%, resulting in an 0.04m change in the estimated position. The results of both
theoretical and field tests were considered satisfactory in terms of positional accuracy and proof of concept;
subsequently 20 logging pebble circuits were constructed, of which 14 were deployed in field trials; the remaining
seven became non-functional during encapsulation. The final pebbles used two different coil configurations, in
order to achieve spherical and
discoid-shaped particles.
The
-12
circuit and receiving coils were
Ewan Incoming
coated in waterproof resin, and
Ewan Outgoing
-10
wrapped in protective film. A
Olly Incoming
mixture of barytes (BaSO4)
Olly Outgoing
-8
powder
and
a
waterproof
Pete Incoming
modelling material was used to
Pete Outgoing
-6
Quin Incoming
encapsulate and mould the pebble
Quin Outgoing
shapes, before coating with
-4
Tony Incoming
fibreglass and resin. The pebbles
Tony Outgoing
had a density of 2600-2730 kgm-3;
-2
this is similar to that of the
indigenous material at the field
0
sites (2650 kgm-3).
2
4
6
8
-6
-4
-2
0
2
4
6
8
10
12
Longshore distance (m)
Figure 8: Transport paths of 5 logging pebbles during a single tide,
illustrating differential movement between incoming (flood) and out-going
(ebb ) tides, and position relative to transmitting grid (dashed lines).
14
Two field trials were undertaken,
using
6
and
7
pebbles,
respectively. During the first test
undertaken at Shoreham-by-Sea,
no useful data were logged; this
was related to the signal power
being set too low by the operators.
However, the tests demonstrated:
(a) that the transmitting grid could
be installed between tides and
operate for at least two full tidal
cycles; (b) the pebbles logged the
transmitted signals and that these
data could, after a full tide, be downloaded to a PC; and (c) that the pebbles could be deployed and recovered
successfully. The second deployment, used 7 pebbles. Data were recorded by all of the pebbles, but two only
operated intermittently due to minor circuit faults. The remaining 5 pebbles had full data logs.
The diameter of the pebbles deployed in this experiment was similar to that of radio-tagged pebbles deployed in
fluvial environments (Habersack, 2001), (72mm B-axis). The coarsest 1.5% of the indigenous beach material was
represented. Scope for improvements using the existing circuitry and batteries, could reduce tracer size to 50mm
diameter (top 8% of indigenous material), and with AMIC circuit design, a reduction down to 20mm (top 35%) is
possible. However, costs of the latter would probably be prohibitive. Note the representativeness would be
improved at coarser-grained sites.
During the second Shoreham experiment, hydrodynamic conditions changed between the incoming and outgoing
tides (Table 6). This is reflected in the behaviour of the logging pebbles (Fig. 8). During the flood tide, the
movement of all Logging Pebbles (LP’s) was onshore and to the East as would be expected from the angle of
wave approach to the shore. Typical flood tide net transport distances average 2m longshore, but were much
more varied cross-shore, depending on pebble position on the beach. During the ebb tide, net transport is again
east but changes to offshore and the transport distances increase. This is consistent with a change in longshore
component of wave power recorded during this period (Table 6).
Variable
Wave Angle
(degrees)
Hs
(m)
Pl
(Wm-1)
No.
Observations
Mean
(Standard deviation)
Flood Tide
Rest period (s)
62
62.5 (60.5)
7 SSW
0.40
28.5
Step length (m)
70
0.53 (0.46)
Particle Velocity (ms-1)
70
0.042 (0.041)
Ebb Tide
Rest period (s)
50
60.7 (42.1)
4 SSW
1.61
533.6
Step length (m)
66
1.04 (0.78)
Particle Velocity (ms-1)
66
0.079 (0.077)
Hs = significant wave height, Pl = longshore component of wave power.
Table 6: Summary data values for particle transport characteristics and flood/ebb tide hydrodynamics
Min – Max
Values
6 – 231
0.07 – 2.57
0.002 - 0.233
6 – 174
0.10 – 4.21
0.005 – 0.39
Step Length (m) / PT Output (mV)
Net transport distances, measured using standard tracers during the period that the logging pebbles were
deployed, were consistent with those given by the new technology. Those pebbles that were positioned, or moved,
higher up the beach
during the flood
3.00
tide,
experienced
much longer step
2.50
lengths
and
transport distances.
This is consistent
2.00
with
the
conventional tracer
1.50
experiments
undertaken at the
1.00
same time. None of
the
LP’s
were
buried during the
0.50
tidal cycle. Data on
particle
velocity,
0.00
rest periods and
12:00:00
13:12:00
14:24:00
15:36:00
16:48:00
18:00:00
step lengths are
Time (BST)
available at a 6s
Step length
Pressure Transducer output
resolution for all
five particles. These
Figure 9: Particle movement during a tidal cycle, illustrating the variation in step lengths
indicate
that
and relatively short period of motion.
particle motion is
limited to relatively
short periods when the pebbles are in the Swash, Surf or Breaker zone, with no transport outside of these zones
(Figure 9).
In fluvial research, particle step length is typically modelled using exponential or 2-parameter Gamma function
distributions (Einstein, 1942, Ergenzinger et al, 1987, Habersack, 1999). These assumptions form the basis of
stochastic sediment transport models such Einsteins (1942). The logging pebble technology provides, for the first
time, the ability to test this assumption applied to coarse particle movements on shingle beaches.
Project Plan Review
Table 7 details the contributions made to achieving the project objectives. However, the project has been affected
by a number of developments, as outlined below.
1. The nature of the technological development, including periods of limited progress (i.e. Logging Pebble
Technology slippage of c.12 months, Electronic Pebble slippage c. 8 months).
Significant technical challenges arose from the decision to produce a prototype logging pebble capable of
delivering real-time information on its position during tidal events. Correspondingly, the time required to develop
this particular technology, from scratch, was underestimated (by the electronic specialists involved). Repeated
Contribution
Objectives
Achieved
Southampton (Geography & Ocean & Earth Sciences) Phobox Electronics, WS Ocean Systems, O.T.D.
Development and manufacture of a multi-particle, logging pebble technology.
1, 6
YES
Generation & processing of high quality field datasets demonstrating “real-time”multi3, 6
YES
particle movements during a tidal cycle with concurrent hydrodynamic data
(Shoreham-by-Sea).
Development of miniaturised electronic pebble technology to permit > representation
5
YES
of indigenous grainsize/grain shape populations in littoral environments.
Generation & processing of 5 sediment transport datasets from a moderate energy,
2, 4
YES
lowland gravel-bed stream with concurrent hydrodynamic data (Highland Water).
Generation and preliminary processing of 14 high quality sediment transport datasets
2, 4
YES
for low – moderate wave energy conditions, on a mixed beach including concurrent
hydrodynamic data (Shoreham-by-Sea).
Tracer Theory – Assessment of the potential influence of methodological variability
2
YES
within the SIM upon derived transport rates.
Tracer Theory – Development and application of a method for determining tracer
2
YES
numbers required for a given accuracy of transport distance.
Tracer Theory - Assessment of the influence of particle characteristics on the
2, 5
YES
estimation of transport rates.
Tracer Theory – Assessment of the influence of tracer injection location on the
2, 5
YES
estimation of transport rates.
Assessment of the suitability of the littoral data ( following preliminary processing )
2
YES
for calibration of the CERC longshore transport model.
Collation of existing field data on fluvial and littoral tracer experiments and
2, 4
YES
calculation of transport rates. Datasets include hydrodynamic and sedimentological
data.
Lancaster Postgraduate (Geography)
Tracer Theory – Development of methods to represent indigenous grainsize and grain
5
YES
shape populations.
Generation of 6 high quality field data sets of sediment transport rates from High
2, 4
YES
Energy cobble/boulder bedded river including concurrent hydrodynamic datasets
(River Lune).
Tracer Theory – Development and application of a new method for calculating
2
YES
sediment transport rate from tracers.
Development and testing of freeze-coring technique for determining active layer
2
On-going (due
thickness.
Feb 02)
Development and calibration (using ADV) of a Computational Fluid Dynamics model
4
On-going (due
of the reach to explain observations of tracer movement.
Feb. 02)
Testing of Magnetic and Cast-Aluminium tracer technology.
5
YES
Table 7: Components of the research project indicating the division of responsibility, objectives tackled and achieved.
attempts were made to bring work back on to schedule, these included project planning meetings and
development of continuously updated time-charts. Likewise, additional specialist input was brought in, in an
attempt to break through the technological impasse. Less progress than originally anticipated has resulted; this has
impacted upon achieving objectives (details are available from the Minutes of the 6 Steering Group meetings,
held over the course of the project). The decision to maintain this aspect of the project arose from the lack of any
other comparable (pebble) technology, for deployment in the littoral zone.
Research Impact and Benefits to Society
This research has provided four key outcomes as summarised below.
1) Provision of the first particle tracking technology, capable of operating within the surf, swash and breaker
(SSB) hydrodynamic zones on shingle beaches. The technology developed permits any number of particles
to be deployed, but is limited to particle diameters of >50mm. The technology has already provided new
information on particle step-lengths and velocities during a tide. The technology has also demonstrated that
for particles larger than the 98th percentile, movement during tidal cycles is limited to relatively short periods
associated with the passage of the SSB zones over the beach, with significant periods during which no
transport occurs.
2) Provision of a miniaturised electronic tracer tagging system that enables detection down to 0.7m burial
depths, with the potential to extend this further. The electronic tags can now be deployed in particles as small
as 14 mm b-axis, representing a 56% improvement in the range of sizes that may be represented.
3) Assessment of the Spatial Integration Method (SIM) for the determination of coarse sediment flux in fluvial
and littoral environments, has shown that experimental design should include accurate size and, to a lesser
extent shape replication of indigenous material. The tracer numbers deployed should, as a first order
approximation, be around 700, and injection procedure should account for cross-section or cross-shore
variability in transport rates. Average Value Methods or 95% Methods typically produce the most accurate
and consistent estimates of bulk transport rate of the results of fluvial trapping experiments are taken to be
correct.
4) Rejection of acoustic tracking technology as it currently exists for the tracking of coarse sediment particles in
shallow gravel-bed rivers. Existing radio-technology offers potential, but is currently limited to poorer than 1
metre positional accuracy.
The research undertaken will contribute to the science base that underpins river and coastal management, in
coarse sediment dominated environments, as outlined below.
1) The provision of an assessment of tracing as a technique for determining sediment transport flux on shingle
beaches and in gravel-bed rivers. Included are: estimation of tracer numbers required; methods for production
of representative size, and shape populations; and, guidance on the most appropriate methods for calculating
sediment transport flux using the SIM.
2) Development of a technology that, for the first time, permits the assessment of particle behaviour in relation
to the hydrodynamics, at a temporal resolution of seconds/minutes in littoral environments. Potential exists
for application of the system to fluvial environments.
3) Provision of high quality field datasets for model calibration obtained from both fluvial and littoral
environments, for use by other researchers.
Explanation of Expenditure
The expenditure did not differ by more than 20% in any heading. Several key divergences occurred, arising from
the cost-neutral extension. Budgetary reviews and staffing issues were presented and agreed at the 6 Project
Steering Group Meetings. To facilitate the cost-neutral extension to the project, the RA post was reduced to 50%,
to provide a longer period of commitment; this was supported (for 5 months) by an additional 50% RA post, to
support the beach deployment. Funds to offset these increases were obtained from the equipment heading.
The technological delays and inability to utilise the acoustic particle tracking system, released some of the
committed equipment costs. These were used to support the Lancaster studentship with water level recorders,
logging software, and the manufacture of replacement tracers (substantial losses were incurred, in both of the
field seasons) together with support for the additional 0.5 RA.
Further Research and Dissemination Activity
The research undertaken during the project has provided a foundation for further work.
1. The new electronic pebble technology and aluminium tracers are part of a NERC grant bid (in review)
involving staff from the Southampton and Aberystwyth Universities.
2. Evaluation of the technology involved in the logging pebble has highlighted areas for improvement in data
logging and processing that will form the basis of new research proposals in 2002 to specifically examine the
behaviour of a range of particle forms in progressively higher energy tidal conditions.
3. The science gain arising from the logging pebble technology is yet to be realised in fluvial environments.
Building on existing research, a proposal to deploy the logging pebble for intra-event tracking of particles in
pool-riffle sequences will be developed in collaboration with researchers involved in EPSRC CEWE project
GR/L90590 Flow & Sediment Dynamics of the Pool-Riffle Unit.
Further work is already being undertaken, while new research proposals are being developed for submission in
the short to medium term. Results have been presented to scientific audiences at the British Geomorphological
Research Group annual conference Sept. 2001, the International Conference on Tracers in Geomorphology in
Sept.1998. Papers were submitted to the Annual MAFF conference on Flooding and Coastal Defence in 1999,
2000. Writing of a technical manual for the logging pebble, together with an end-user guide to the application of
tracing for sediment transport estimation in fluvial and littoral environments are in progress. Dissemination of
these will be via the Project Website and end-user contacts.
The technology developed on this project has been widely disseminated in the popular science and media,
including leader articles in the Times (Sept. 2000) and Guardian (Sept. 2000), as well as TV Broadcasts on BBC
News 24 Science (Aug. 2001) and BBC Newsround (Sept.2001). Further international dissemination has been
achieved via articles for BBC World Service, American Radio Broadcasting (live to 200 radio stations across the
US) and through the Home Office Press Association. Two journal papers have been submitted to Journal of
Sedimentary Research and Marine Geology, Two peer-reviewed book chapters has been published (Sear et al,
2000; Lee et al 2000). It is the intention to publish the research in both scientific and practitioner Engineering and
Geomorphological Journals. Results form the Logging Pebble will be submitted to the journal Nature. A project
web site facilitating dissemination of data, and project information to specialists and the public including full
project datasets will be made available 12 months from September 2001, to allow for publication of results.
Research findings have been incorporated into Bachelors and Masters courses in both University of Southampton
Geography and Ocean & Earth Science Departments. This is important because many of our Graduates eventually
enter careers in river or coastal management with engineering or environmental organisations.
Articles Published or in Review (these are available on request to EPSRC)
Lee, M.W.E., Sear, D.A., Workman, M., Collins, M.B., Atkinson, P.M. & Oakey, R.J. Number of tracers required for the
measurement of longshore transport distance on a shingle beach, Marine Geology, (in review).
Oakey, R.J., Green, M., Carling, P.A., Sear, D.A., Lee, M.W.E., & Warburton, J. Grain shape analysis – New perspectives
and a method that determines representative tracer shapes, Journal of Sedimentary Research, (in review).
Sear, D.A., Lee, M.W.E., Oakey, R.J., Carling, P.A. & Collins, M.B. Coarse sediment tracing technology for littoral and
fluvial environments review, in Foster, I.D.L. (Ed) Tracers in Geomorphology, J.Wiley & Sons, Chichester, UK, 21-55,
2000.
References
Ashiq, M., Role of active bed width in bedload transport – Coarse bed rivers, in Jaywardena et. al. (eds) River Sedimentation,
Balkema, Rotterdam, 951 – 955, 1999.
Bray, M.J., 1996. Beach Budget Analysis and Shingle Transport Dynamics: West Dorset. Unpublished PhD Thesis, London
School of Economics, University of London, 641 pp.
Church, M., Mclean,D.G. & Wolcott,J.F.River bed gravel: sampling and analysis, in Thorne, C.R., Bathurst, J.C. & Hey,
R.D. (Eds) Sediment Transport in Gravel-bed Rivers, J.Wiley & Sons, Chichester, UK, 291-338, 1987.
Cooper, N., Evaluation of the impacts of shoreline management at contrasting sites in southern England, Unpublished Ph.D
thesis, Dept. of Geography, University of Portsmouth, 279pp,1996.
deVries, P., Bedload layer thickness and disturbance depth in gravel-bed streams, Journal of Hydraulic Engineering, (in
press).
Dorey, A.P., Quinn, S.P. & Dyer, K.R., A Transponding Acoustic Pebble, Ultrasonics, 147-148, July, 1972.
Einstein, H.A., Formulas for the transportation of bedload, Transactions of the American Society of Civil Engineers, 107,
561-597, 1942.
Ergenzinger, P., Schmidt, K-H, & Busskamp, R., The Pebble Transmitter System (PETS): first results of a technique for
studying coarse material erosion, transport and deposition, Zeitschrift furGeomorphologie N.F., 33, 503-508, 1989.
Habersack, H.M, Radio-tracking gravel particles in a large braided river in New Zealand: a field test of the stochastic theory
of bedload transport proposed by Einstein, Hydrological Processes, 15, 377-391, 2001.
Haschenburger, J.K. & Church, M., Bed material transport estimated from the virtual velocity of sediment, Earth Surface
Processes & Landforms, 23, 791-808, 1998.
Jolliffe, I.P., 1964. An Experiment Designed to Compare the Relative Rates of Movement of Different Sizes of Beach
Pebbles. Proceedings of the Geologists’ Association, 75, 67-86.
Kidson, C. and Carr, A.P., 1961. Beach Drift Experiments at Bridgewater Bay, Somerset. Proceedings of the Bristol
Naturalists’ Society, 30, 2, 163-180.
Lee, M.W.E., submitted. Tracer-Derived Measurements of Coarse Clastic Sediment Behaviour, in the Littoral Zone.
Unpublished PhD Thesis, School of Ocean and Earth Science, University of Southampton, 75 pp (plus appendices).
Lee, M.W.E., Bray, M.J., Workman, M., Collins, M.B. & Pope, D., Coastal shingle tracing: A case study using the Electronic
Tracer System (ETS), in in Foster, I.D.L. (Ed) Tracers in Geomorphology, J.Wiley & Sons, Chichester, UK, 413435, 2000.
Leopold, L.B., Sediment size that determines channel morphology, in Billi, P., Hey, R.D., Thorne, C.R. & Tacconi, P. (Eds)
Dynamics of Gravel bed Rivers, J.Wiley & Sons, Chichester, 297-311, 1992.
Madsen, O.S., 1974. Stability of a Sand Bed Under Breaking Waves. Proceedings of the 14th Coastal Engineering
Conference, Copenhagen, American Society of Civil Engineers, 776-794.
Madsen, O.S., Tracer Theory, in Seymour, R. (Ed) Nearshore Sediment Transport, Plenum Press, New York, 103 – 114,
1989.
McEwen, I. & Heald, J., Discrete particle modelling of entrainment from flat uniformly sized sediment beds, Journal of
Hydraulic Engineering, 127, 7, 588-597, 2001.
Miller, K.A., 1997. Sediment Dynamics on Wemyss Beach, South Fife. Unpublished MPhil.Thesis, University of St Andrews,
174 pp.
Richardson, N.M., 1902. An Experiment on the Movements of a Load of Brickbats Deposited on the Chesil Beach.
Proceedings Dorset Natural History and Antiquarian Field Club, 23, 123-133.
Rosen, P.S. & Brenninkmeyer, B.M., Transport of coarse material in low energy beach environments, Proceedings of
Coastal Zone ‘89’, New York, ASCE, 1724-1737,1989.
Schmidt, K-H, & Ergenzinger, P., Bedload entrainment, Travel lengths, Step lengths, Rest periods – studied with passive
(Iron, Magnetic) and Active (Radio) tracer techniques, Earth Surface Processes & Landforms, 17, 147-165, 1992.
Shore Protection Manual, 1984, 4th Edition. Coastal Engineering Research Centre. US Army Corps of Engineers,
Washington D.C., Volume I.