Long term recovery rates obtained using RFID technology at a mixed beach

Abstract

Recovery rates were obtained by radio frequency identification (RFID) technology in pebbles and cobbles at San Felipe beach, Gran Canaria. The aim of this work was to define which factors affected the recovery of tagged gravels. Several tests were performed to determine the detection depth threshold, and 16 field experiments were carried out over seventeen months after tracer deployment on the beach. Recovery rates are highly variable with time, ranging from 72.2% in the first recovery session to 25.8% in the last one. Nevertheless, a nearly stable situation was found for the final eight months. Apart from the effect of time, there were several factors that affected the recovery rate. Some of these were related to the particle, such as the position of the tag within the particle, as well as its weight, size and shape. Two environmental factors were considered. First, the elevation of the tracer on the beach showed that the recovery rate was higher with particles located above the storm berm. Second, wave height, which showed no relation with recovery rates even though during the experiment significant storms and periods of calm took place.

Full text

ARTICLE IN PRESS
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
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
ARTICLE IN PRESS
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
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
1
INTRODUCTION
Sediment transport is one of the fundamental aspects
necessary in order to understand the processes that take
place on beaches, and it is therefore indispensable to
properly manage these coastal areas. Sandy beaches have
been studied for many years and there is much available
information on the mechanisms of sediment transport.
However, there are few studies on coarse-grained beaches
due to the logistical difculties involved, such as the rough
dynamic conditions and the fragility of the instrumentation.
One of the most important characteristics of coarse-grained
beaches are their high efciency at protecting shore from
extreme wave conditions, due to their high degree of
stability (Carter and Orford, 1984), so that in many places
these beaches are considered as a management tool against
coastal erosion (Allan and Komar, 2004; Dickson et al.,
2011). Interest in these kinds of beaches has increased
thanks to technological advances that allow for the study
of transport of individual particles over long periods of
time (Allan et al., 2006; Curtiss et al., 2009; Bertoni et al.,
2010, 2012; Dickson et al., 2011).
There are three predominant types of tracers (Allan et
al., 2006). The rst group are visual tracers that consist
of painted coarse-grained particles or exotic lithologies for
Geologica Acta, Vol.13, Nº 2, June 2015, xx-xxxx
DOI: 10.1344/GeologicaActa2015.13.2.nº manuscript
Long term recovery rates obtained using RFID technology at a
mixed beach
M. CASAMAYOR1* I. ALONSO1 J. CABRERA2 S. RODRÍGUEZ1 M.J. SÁNCHEZ-GARCÍA1
1Instituto de Oceanografía y Cambio Global
Universidad de Las Palmas de Gran Canaria, Campus Universitario Tafira, 35017 Las Palmas de Gran Canaria.
E-mail: marionacasamayor@gmail.com
2Instituto Universitario de Sistemas Inteligentes y Aplicaciones Numéricas en Ingeniería
Universidad de Las Palmas de Gran Canaria, Campus Universitario de Tafira, 35017 Las Palmas de Gran Canaria.
Recovery rates were obtained by radio frequency identication (RFID) technology in pebbles and cobbles at San
Felipe beach, Gran Canaria. The aim of this work was to dene which factors affected the recovery of tagged
gravels. Several tests were performed to determine the detection depth threshold, and 16 eld experiments were
carried out over seventeen months after tracer deployment on the beach. Recovery rates are highly variable with
time, ranging from 72.2% in the rst recovery session to 25.8% in the last one. Nevertheless, a nearly stable
situation was found for the nal eight months. Apart from the effect of time, there were several factors that
affected the recovery rate. Some of these were related to the particle, such as the position of the tag within the
particle, as well as its weight, size and shape. Two environmental factors were considered. First, the elevation of
the tracer on the beach showed that the recovery rate was higher with particles located above the storm berm.
Second, wave height, which showed no relation with recovery rates even though during the experiment signicant
storms and periods of calm took place.
Tracer recovery. Pebbles. Detection. Sediment transport. Gran Canaria.
KEYWORDS
ABSTRACT

ARTICLE IN PRESS
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
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
ARTICLE IN PRESS
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
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
M. Casamayor et al.
Geologica Acta, 13(2), xxxxxxx (2015)
DOI: 10.1344/GeologicaActa2015.13.2.nº manuscript
Short-title
2
the quick identication of the particles. Such tracers have
different problems, such as those derived from their burial
and the loss of paint with the pass of time. These problems
lead to low recovery rates (between 5 and 50%) (Joliffe,
1964; Sear et al., 2002; Dornbusch et al., 2002; Allan et
al., 2006; Ciavola and Castiglione, 2009). The second
group are called passive tracers, and may include different
techniques such as radioactive, aluminium, magnetic or
electronic tracking technologies. These techniques are a
signicant improvement in tracking because they allow the
user to locate particles even when they are buried. In this
case, recovery rates show great variability, ranging from
40 to 85%, depending on the hydrodynamic conditions at
the study area (Voulgaris et al., 1999; Sear et al., 2002;
Osborne, 2005). Finally, the third group are active tracers
which are composed of an electronic tracking system
powered by a battery that is encapsulated within a resin
to produce a tracer with the size and shape characteristic
of beach particles (Bray et al., 1996; Voulgaris et al.,
1999). The great advantage of these tracers is that they can
be detected up to a depth of 1m, rising recovery rates to
values greater than 70%. However, their drawbacks are
their costs and articial nature, which can inuence the
tracer response to local hydrodynamic conditions (Allan
et al., 2006).
The tracers used in this study are passive tracers,
the radio frequency identication (RFID) tracers. This
technique presents a high ratio of efciency to cost, since it
is possible to obtain high recovery rates with low operating
costs (Allan et al., 2006; Bertoni et al., 2010).
This paper presents the results of applying RFID
technology to the mixed beach of San Felipe, Gran Canaria,
to study the recovery rates of the pebbles and cobbles over
one and a half year. The different factors that can affect the
recovery rate are considered. Factors related to the tracer
are weight, size, shape and the axis through which the tag
is inserted; and those related to the environment are wave
conditions and elevation of the tracer on the beach prole.
REGIONAL SETTING
The study area, San Felipe beach, is located on the north
coast of the island of Gran Canaria (Spain), at the western
14°0'0"W16°0'0"W18°0'0"W
30°0'0"N28°0'0"N
A B
15°20'0"W15°30'0"W15°40'0"W15°50'0"W
28°10'0"N28°0'0"N27°50'0"N
0 20 4010 Meters
N
C
FIGURE 1. A) Map of Canary Islands. B) The star close to the island of Gran Canaria indicates the wave buoy location used in this work, and the square
shows the location of the study area. C) Aerial photograph of San Felipe beach. The line is the beach profile shown in Figure 3.

ARTICLE IN PRESS
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
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
ARTICLE IN PRESS
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
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
Geologica Acta, 13(2), xxxxxxx (2015)
DOI: 10.1344/GeologicaActa2015.13.2.nº manuscript
M. Casamayor et al. Short-title
3
limit of San Felipe village (Fig. 1). It is approximately
200m long and has variable width depending on the season,
ranging from 20m wide in winter to 40m in summer,
measured from the upper part of the beach prole to 0m
orthometric height.
San Felipe is a mixed beach consisting of ne black
sand formed by basalt erosion (Balcells and Barrera, 1990),
and coarse-grained particles that are mainly phonolitic and
basaltic pebbles and cobbles. Coarse-grained materials are
distributed over the backshore and upper foreshore, while
sand is found on the lower foreshore and nearshore (Fig.
2). San Felipe beach is subjected to large morphological
changes, which are mostly related to wave climate
variations. In winter time, storm berm is formed, which
increases the foreshore slope. During this season only
cobbles and pebbles are visible along the beach, even at
low tide. Around May, with the beginning of the trade
winds, the submerged sand bar moves onshore, and by the
end of the summer the cobbles and pebbles located on the
foreshore become completely covered by sand, with a clear
reduction in beach slope (Fig. 3).
In a general geological context, the northern coast of
Gran Canaria is characterized by phonolitic Miocene lava
ows that have been eroded by the sea, and constitute a
coastal platform raised 6-15m above mean local sea level
(Balcells and Barrera, 1990). Most of this coastal platform
is covered by crops and villages. The beach is backed by
colluvial deposits and debris. The western edge is bounded
by basaltic lava ow, which enters into the sea, while
the eastern sector of the beach leads to the ravine of San
Felipe, 8,600m in length and maximum 817m in height
(Menéndez et al., 2008).
Gran Canaria has a semidiurnal tidal regime with 2.8m
tidal range at spring tides and 0.8m at neap tides. These are
all astronomic tidal values obtained from the WXTide32
program. Data from wave recorder buoy number 2442
from Puertos del Estado have been used to characterize
the wave climate. This buoy is located northwest of Gran
Canaria, several km offshore and at 780m water depth (see
location in Fig. 1).
Wave data covering the study period (from March
2013 to July 2014) are shown in Figure 4. Signicant
wave height (Hs), peak period (Tp) and wave direction
show a clear seasonal pattern, with calm conditions from
April to October and storm conditions from November
to March. The rst period is under the inuence of the
trade winds. Dominant waves come from the N-NNE,
average Hs=1.52m and average Tp=9.43s. Regarding the
winter season, dominant waves come from N, Hs=1.97m
and Tp=12.08s. There are frequent storms that generate
waves with Hs > 4m. During the study period, 13 stormy
events took place, all of them with Hs higher than 3.0m on
average. The duration of these storms was highly variable,
ranging from 7 to 71h.
METHODS
Particle sampling and preparation
To carry out the experiment 198 particles were collected
over homogeneously distributed region along and across
the beach. Sampling of pebbles and cobbles took place
on November 1, 2012, at low spring tide conditions.
Although sample selection was random, the particles had
to meet certain requirements due to the limitations of the
experiment: the major axis of the particle had to be larger
FIGURE 2. Photograph of San Felipe beach at low tide during summer
conditions. Note i) that pebbles and cobbles are very well rounded
and quite well sorted, and ii) the sand over the lower foreshore.
0 10 20 30 40
Cross-shore distance (m)
-2
0
2
4
6
Elevation (m)
Summer
Winter
FIGURE 3. Beach profile at the central section of San Felipe beach.
Profile location is shown in Figure 1C. Elevation corresponds to
orthometric heights. Note the net accumulation of sediments in the
summer profile, which is mostly related to the onshore migration of
the sand bar.

ARTICLE IN PRESS
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
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
ARTICLE IN PRESS
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
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
M. Casamayor et al.
Geologica Acta, 13(2), xxxxxxx (2015)
DOI: 10.1344/GeologicaActa2015.13.2.nº manuscript
Short-title
4
than 42mm, and cobbles weighing more than 3kg were
discarded. Considering these experimental limitations,
axial lengths of 802 additional particles were measured in
the beach to check if tracer dimensions were representative
of beach particles. These 802 additional particles were
randomly selected along and across the beach.
Mar-13 May-13 Jul-13 Sep-13 Nov-13 Jan-14 Mar-14 May-14 Jul-14
0
1
2
3
4
5
Significant wave height (m)
0
5
10
15
20
25
Peak wave period (s)
0
1
2
3
Tide (m)
0
90
180
270
360
Wave direction (º)
FIGURE 4. Hourly data recorded at the Gran Canaria buoy from March 2013 to August 2014. Tidal values were calculated with the WXtide32 program.
Vertical lines show the date of the different field experiments.

ARTICLE IN PRESS
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
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
ARTICLE IN PRESS
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
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
Geologica Acta, 13(2), xxxxxxx (2015)
DOI: 10.1344/GeologicaActa2015.13.2.nº manuscript
M. Casamayor et al. Short-title
5
The three axes (long, intermediate and short) of each
individual particle were measured by means of a vernier
calliper with 0.1mm precision and weighed on a scale with
precision of 0.01g.
Preparation of tracer particles was undertaken in three
phases: drilling, tag introduction and sealing. The rst
phase consisted of drilling a hole of 6mm diameter and
40mm in length in each particle. Following Allan et al.
(2006), Dickson et al. (2011), Miller and Warrick (2012),
the holes were drilled through the short-axis of the particle,
but when it was smaller than 42mm, it was necessary
to drill the hole through the intermediate-axis, or even
through the major axis when the intermediate one did not
reach the required length.
The tag dimensions were 30.2±0.6mm long and
3.8±0.05mm in diameter and were protected by a glass
capsule. To avoid breakage through impact with other
coarse grains, tags were introduced into a plastic hose with
a section of 4x6mm. Once this process was nished, both
ends of the hose were sealed with universal acid silicone
and the tag, properly protected, introduced into the cobble.
Finally, the particles were sealed with epoxy resin of two
components. This product was used because of its high
resistance to impacts and temperature changes.
Tracers were regularly deployed on the beach on March
14, 2013 at low spring tide. The entire beach was surveyed
sixteen times from March 19, 2013 to July 28, 2014 with a
detection system to locate the tagged particles.
Detection system
This study uses RFID technology following Allan et al.
(2006), who investigated a mixed sand and gravel beach
with this method. The detection system is composed of a
base station and one mobile detection device. The different
elements of this system are listed on Table 1. Once a certain
tag is detected, its identication number is stored in a
micro SD card by the mobile device and transmitted to the
base station in real time using an industrial, scientic and
medical (ISM) radio band (869MHz). Tracer positioning
was carried out by means of an electronic total station.
The detection range of the RFID reader increases
with supplied voltage. However, a high voltage generates
heat that must be removed from the box that contains the
electronic components. After experimenting with various
voltages, it was found that 18V gave the best result, so that
the detector temperature was kept within reasonable limits
and the detection distance was close to the maximum.
The antenna is a circular structure of PVC tube 80cm in
diameter containing a 1.88mm copper wire looped around
three times. A synchronization module was used to adjust
the radio frequency circuit to the resonant frequency of the
antenna. When it detects a tag, the mobile detection device
emits an acoustic signal and then its position is recorded by
the total station.
Detection range
Several tests were carried out in order to determine the
maximum detection depth of our antenna. Considering
that the beach under study contains different types of
sediment, such as sand at the lower foreshore and cobbles
and pebbles at the backshore, as well as water if the beach
becomes submerged, some experiments were designed
to assess the detection capability of the antenna in these
three different media. In each medium, a tracer was buried
at different depths, being the maximum detection depth
considered the detection threshold. The relative position of
the tag was considered: the tag could be perpendicular to
the beach surface (which means that it was parallel to the
short-axis of the pebble) or parallel to the surface (which
may happen when the tag is either in the intermediate or in
the long-axis).
TRACER CHARACTERISTICS
Since this work mostly deals with recovery rates of
tracers, it is necessary to know the properties of these
particles.
Weight
Tracer weight ranges between 82 and 2837g with an
average of 450.58g. These values were obtained once
the particles had been drilled, the tag inserted and the
hole sealed. Due to the large range of weights, it was
necessary to make a transformation of the data. We chose
the same transformation used by Krumbein (1936), who
established the phi (Ø) scale for grain size. The proposed
scale for particle weight is named phi-weight (Øw), and
is dened as
SYSTEM SUB-SYSTEM COMPONENTS
Base station
Waspmote 12 channel SirFIII GPS receiver and radio
modem XBee 868 Pro
Laptop With direct USB connections to waspmote
Total station
Mobile device
Battery 24V 12V/7Ah lead-gel batteries
RFID reader RI-CTL-MB2B from Texas Instruments
RFM module RI-RFM-008B from Texas Instruments
Synchronization
module RI-ACC-008B from Texas Instruments
Antenna In-house design
Waspmote 12 channels SirFIII GPS receiver and
radio modem XBee 868 Pro
RS232-3V3 adapter
Table 1. Summary of the different components that form the detection system.


PARTICLE ID
AXIS WEIGHT
SHORT
(mm)
INTERMEDIATE
(mm)
LONG
(mm)
INITIAL
(g)
FINAL
(g)
LOSS
(%)
175 48 64.0 78.5 308.06 306.25 0.59
176 42 65.5 83.5 345.11 343.06 0.59
177 44 65.1 80.6 336.94 335.09 0.55
Average 3
particles 44.67 64.87 80.87 330.04 328.13 0.58
Average 198
particles 44.71 65.28 85.73 452.48 450.58 0.42
Table 2. Morphometric properties and weight of selected particles used to determine weight loss
resulting from the preparation of the tracers.


ENVIRONMENT MEDIA AXIS MAXIMUM DETECTION
DEPTH (cm)
Backshore Pebbles and
cobbles
Short 46
Intermediate 25
Long 15
Lower foreshore
Short 35
Sand Intermediate 21
Long 17
Submerged
beach
Short 22
Water Intermediate 16
Long 13.5
Table 3. Detection thresholds in three different media.

TABLE 1. Summary of the different components that form the detection
system.

ARTICLE IN PRESS
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
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
ARTICLE IN PRESS
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
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
M. Casamayor et al.
Geologica Acta, 13(2), xxxxxxx (2015)
DOI: 10.1344/GeologicaActa2015.13.2.nº manuscript
Short-title
6
Øw = log2(W) W the particle weight in grams (1)
After applying the above transformation and considering
that the different categories derived from the Øw scale are
delimited by integer units, all our tracers lay in the small
range between 6 and 12 Øw. There are two main categories
(7-8 and 8-9 Øw) of very similar abundances, with the
average value corresponding to 8.82 Øw, equivalent to
450.58g (Fig. 5A).
The weight loss during preparation of the tracer was
calculated measuring the initial and nal weight of three
particles. The nal weight takes into account the drilling,
the introduction of the tag and sealing of the hole. Table 2
shows that the preparation process represents an average
mass loss of 0.58% of the initial weight for these three
particles. Considering that these particles are lighter than
the average, and that the weight loss is similar in each
particle and does not depend on its size, the estimate
average loss of weight of the whole of the particles is only
0.42%.
Size
Tracer particle size is in the range of cobbles and
pebbles. The major axis ranges between 49.2 to 160.0mm,
with an average length of 85.73mm, the intermediate axis
is in the range 42.5-135.0mm, with an average length of
65.28mm, and the minor axis measures between 26.0 and
90.0mm (average length of 44.70mm).
Figure 6 shows the three axis length distribution for
two types of particles: those used as tracers (198 particles)
and 802 additional particles whose axes were measured.
It is shown that both populations follow a normal
distribution with regard to any of the three axes. Note that
the distribution of the three axes in the tracer population is
very similar to that in the population of 802 particles, and
Shape
0
20
40
60
80
100
Number of particles
7
31
53
27
17
50
13
-7.5 -7 -6.5 -6 -5.5 -5
Size (Ø)
0
20
40
60
80
100
Number of particles
2
15
73
85
23
6 7 8 9 10 11 12
Weight (Øw)
0
20
40
60
80
100
Number of particles
9
65 66
43
12
3
C CP CB CE P B E
FIGURE 5. A) Distribution of tracer weights using the Øw scale. B) Distribution of tracer sizes, in Ø. C) Distribution of tracer shapes according to the
Sneed and Folk (1958) classification (C: compact, P:platy, B:bladed, E:elongate).
-8 -7 -6 -5 -4 -3
Size (
0
10
20
30
40
50
60
Percentage
Short-axis
-8 -7 -6 -5 -4 -3
Size (
0
10
20
30
40
50
60
Percentage
Intermediate-axis
-8 -7 -6 -5 -4 -3
Size (
0
10
20
30
40
50
60
Percentage
Long-axis
FIGURE 6. Axis length distributions for both populations: 198 particles used as tracers (lines) and 802 particles whose axes were measured (histograms).

ARTICLE IN PRESS
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
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
ARTICLE IN PRESS
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
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
Geologica Acta, 13(2), xxxxxxx (2015)
DOI: 10.1344/GeologicaActa2015.13.2.nº manuscript
M. Casamayor et al. Short-title
7
therefore it is very reasonable to assume that the selected
pebbles and cobbles used as tracers are representative of
the particles on the beach. In fact, the average grain size
is -5.95Ø and -5.78Ø for the tracer population and for the
additional 802 beach particles respectively. Considering the
intermediate axis as the average particle size, the dominant
size of tracers corresponds to very coarse pebbles (between
-5.5 to -6.0Ø), based on the Blair and McPherson (1999)
classication (Fig. 5B).
Shape
The shape of the particles was analysed using the
approach of Sneed and Folk (1958) based on the ratio
between their three axes. Figure 7 shows the predominant
shapes of 1000 particles plotted on the diagram proposed
by Sneed and Folk (1958). There are slight differences
between additional beach particles and tracers. In most
parts of the beach particles are bladed, and tracers are either
bladed or compact-bladed (Fig. 5C). Moreover, tracers are
not very-platy, very-bladed or very-elongate since these
categories are not suitable for particle preparation.
RESULTS
Detection threshold
Several tests were performed on the beach to determine
the detection threshold of the system. Table 3 shows
that the greatest detection depth occurs when the tag is
buried in gravel decreases in the case of sand, and has the
lowest values when it is immersed in water. These results
conrm that the propagation of electromagnetic waves in
water is worse than in solid media. On the other hand, the
detection range is greater when the tags are in the short-
axis of the pebbles, which is the direction normal to the
electromagnetic eld generated by the antenna, while the
lowest range occurs when the tags are in the long-axis.
Tracer recovery
Sixteen eld experiments covering the period between
March 2013 and July 2014 were carried out to determine
the movement of the tracers on the beach. The number
of tracers recovered (detected by RFID system) shows a
Beach particles
Tracers
Mean tracer value
Mean beach particle value
c / a b / a
(a - b) / (a - c)
C
CP
VP
PBE
CB
VB
CE
VE
FIGURE 7. Tri-plot diagram based on Sneed and Folk (1958), which
represents particle shapes according to their orthogonal axes:
198 tracers, and 802 additional beach particles. a: long-axis, b:
intermediate-axis, c: short-axis. C: compact, P: platy, B: bladed, E:
elongate and V: very. Diagram generated after Graham and Midgley
(2000).
Mar-13 Jun-13 Sep-13 Dec-13 Mar-14 Jun-14 Sep-14 Dec-14 Mar-15
0
40
80
120
160
Number of particles
FIGURE 8. Number of particles recovered in each field session.
Continues and dashed lines are the linear and exponential fits,
respectively. R2= 0.82 in both cases.
SYSTEM SUB-SYSTEM COMPONENTS
Base station
Waspmote 12 channel SirFIII GPS receiver and radio
modem XBee 868 Pro
Laptop With direct USB connections to waspmote
Total station
Mobile device
Battery 24V 12V/7Ah lead-gel batteries
RFID reader RI-CTL-MB2B from Texas Instruments
RFM module RI-RFM-008B from Texas Instruments
Synchronization
module RI-ACC-008B from Texas Instruments
Antenna In-house design
Waspmote 12 channels SirFIII GPS receiver and
radio modem XBee 868 Pro
RS232-3V3 adapter
Table 1. Summary of the different components that form the detection system.


PARTICLE ID
AXIS WEIGHT
SHORT
(mm)
INTERMEDIATE
(mm)
LONG
(mm)
INITIAL
(g)
FINAL
(g)
LOSS
(%)
175 48 64.0 78.5 308.06 306.25 0.59
176 42 65.5 83.5 345.11 343.06 0.59
177 44 65.1 80.6 336.94 335.09 0.55
Average 3
particles 44.67 64.87 80.87 330.04 328.13 0.58
Average 198
particles 44.71 65.28 85.73 452.48 450.58 0.42
Table 2. Morphometric properties and weight of selected particles used to determine weight loss
resulting from the preparation of the tracers.


ENVIRONMENT MEDIA AXIS MAXIMUM DETECTION
DEPTH (cm)
Backshore Pebbles and
cobbles
Short 46
Intermediate 25
Long 15
Lower foreshore
Short 35
Sand Intermediate 21
Long 17
Submerged
beach
Short 22
Water Intermediate 16
Long 13.5
Table 3. Detection thresholds in three different media.

TABLE 2. Morphometric properties and weight of selected particles
used to determine weight loss resulting from the preparation of the
tracers.
SYSTEM SUB-SYSTEM COMPONENTS
Base station
Waspmote 12 channel SirFIII GPS receiver and radio
modem XBee 868 Pro
Laptop With direct USB connections to waspmote
Total station
Mobile device
Battery 24V 12V/7Ah lead-gel batteries
RFID reader RI-CTL-MB2B from Texas Instruments
RFM module RI-RFM-008B from Texas Instruments
Synchronization
module RI-ACC-008B from Texas Instruments
Antenna In-house design
Waspmote 12 channels SirFIII GPS receiver and
radio modem XBee 868 Pro
RS232-3V3 adapter
Table 1. Summary of the different components that form the detection system.


PARTICLE ID
AXIS WEIGHT
SHORT
(mm)
INTERMEDIATE
(mm)
LONG
(mm)
INITIAL
(g)
FINAL
(g)
LOSS
(%)
175 48 64.0 78.5 308.06 306.25 0.59
176 42 65.5 83.5 345.11 343.06 0.59
177 44 65.1 80.6 336.94 335.09 0.55
Average 3
particles 44.67 64.87 80.87 330.04 328.13 0.58
Average 198
particles 44.71 65.28 85.73 452.48 450.58 0.42
Table 2. Morphometric properties and weight of selected particles used to determine weight loss
resulting from the preparation of the tracers.


ENVIRONMENT MEDIA AXIS MAXIMUM DETECTION
DEPTH (cm)
Backshore Pebbles and
cobbles
Short 46
Intermediate 25
Long 15
Lower foreshore
Short 35
Sand Intermediate 21
Long 17
Submerged
beach
Short 22
Water Intermediate 16
Long 13.5
Table 3. Detection thresholds in three different media.

TABLE 3. Detection thresholds in three different media.

ARTICLE IN PRESS
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
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
ARTICLE IN PRESS
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
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
M. Casamayor et al.
Geologica Acta, 13(2), xxxxxxx (2015)
DOI: 10.1344/GeologicaActa2015.13.2.nº manuscript
Short-title
8
clear decrease with time, so that the longer the elapsed
time between the deployment and recovery, the lower
the amount of tagged particles recovered (Fig. 8). In
fact, during the first field session carried out five days
after deployment (March 19th), only 143 tracers from
198 were found, and three weeks later the number
had decreased to 114. From these numbers it could be
guessed that in three months all tracers should have
been removed, but what we really found was that
nearly 18 months later there were still 51 tracers left
on the beach. In fact, the cloud of dots (Fig. 8) may be
fitted by a linear trend, which indicates that the total
disappearance of tracers would happen by March 2015
(two years after deployment), or by an exponential fit
in which case two and a half years after deployment we
still should have 10% of the initial number of tracers.
Factors related to the tracer
The tag may be inserted into the particle through the
long, intermediate or the short axis. The inuence of this
factor on the recovery rate is shown in Figure 9, from
which it is clear that the number of recovered particles
decreases with time for the three axes. Nevertheless,
the linear ts (Fig. 9) clearly indicate that the particles
in which the tag was inserted through the short-axis
lasted longer compared to those particles with the tag
in the intermediate or long axis. It should be noted that
the initial number of tracers is not the same for each
category. Tags in the long axis occur in only 27 particles,
while there are 97 with the tag in the intermediate axis
and 74 in the short axis. Relative to these numbers, the
percentage of recovered particles with the tag in the
short-axis is 37.8%, 36.7% in the intermediate-axis and
35% in the long-axis. These values conrm that the best
position in terms of recovery is when the tag is inserted
through the short-axis.
Regarding the weight of the particles, recovery
rates of all categories of Øw (light, medium and heavy
clasts) decrease with time. Nevertheless, particles with
Øw between 7 and 8 (128-256g) have a higher negative
slope, which indicates that this category will disappear
sooner than the others (Fig. 10).
Recovery rates versus time and size of the pebbles
and cobbles show similar patterns to those described
for weight. The dominant categories of the initial 198
particles (Fig. 5B) are those with higher recovery
rates (Fig. 11). However, the category corresponding
to tracers in the range between -5.5 and -6Ø (45.3-
64mm) has the most negative slope, and therefore this
is the least favourable size range for long-term RFID
experiments.
Regarding the different shapes of the tagged
particles, no clear pattern is seen when they are plotted
against time, despite the great diversity of shapes
(Figs. 5C and 7). All shapes present a negative slope,
which corresponds to a decrease in the recovery rate.
The only difference is that platy and elongate particles
show a more stable trend. Therefore, these are the best
shapes for this kind of experiment (Fig. 12).
Mar-13 Jun-13 Sep-13 Dec-13 Mar-14 Jun-14
0
10
20
30
40
50
Number of particles
6 - 7
7 - 8
8 - 9
9 - 10
10 - 11
11 - 12
FIGURE 10. Time evolution of the number of recovered particles for the
different weight categories (Øw).
Mar-13 Jun-13 Sep-13 Dec-13 Mar-14 Jun-14
0
20
40
60
Number of particles
-7,5 - -7
-7 - -6,5
-6,5 - -6
-6 - -5,5
-5,5 - -5
FIGURE 11. Recovered particles for each field session as a function of
the length of the intermediate-axis.
Short-axis
Intermediate-axis
Long-axis
Mar-13 Jun-13 Sep-13 Dec-13 Mar-14 Jun-14 Sep-14 Dec-14 Mar-15 Jun-15
0
20
40
60
80
Number of particles
FIGURE 9. Recovered tracers during the time span of the experiment
depending on the axis through which the tag was inserted.

ARTICLE IN PRESS
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
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
ARTICLE IN PRESS
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
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
Geologica Acta, 13(2), xxxxxxx (2015)
DOI: 10.1344/GeologicaActa2015.13.2.nº manuscript
M. Casamayor et al. Short-title
9
Factors related to the environment
One of the environmental variables to be considered
when analyzing the recovery rates of tracers is the wave
climate. However, to cope with this effect it is necessary to
know what was the time elapsed during which waves were
responsible for tracer movement. Four different cases have
been considered: the average of hourly wave data during 1,
5, 10 and 20 days prior to the eld experiment.
Figure 13 shows the number of recovered particles vs
signicant wave height (Hs) for the four mentioned cases.
No clear relationship is found for any of them, which
indicates that under high waves and in mild conditions the
recovery rates were quite similar. Obviously, waves are the
main agent of particles movement, but it seems to have no
direct effect on tracer recovery.
The second factor to be considered is the elevation of
the tracers on the beach. For this purpose we divided the
beach into four regions. Tracers found below 0m were
below mean low level , those between 0-2m were along the
lower and central part of the foreshore, particles between
2-4m were in the upper foreshore and the summer berm,
and those above 4m were above the crest of the winter
berm. Despite the general decreasing trend in the number
of retrieved particles, we found a positive trend for tracers
found above 4m (Fig. 14).
DISCUSSION
The entire process of tracer preparation is associated
with a certain mass loss of the sample due to the drilling
for the tag introduction. It has been calculated that this loss
of weight only represents 0.42% on average of the mass of
the whole particles (Table 2). This result is similar to that
of Miller et al. (2011), who state that mass loss after RFID
placement in most cases was less than 1%. Therefore, it
can be considered that this process does not affect the basic
characteristics of the particles.
Results of the detection threshold in depth (Table 3)
match those presented by Bertoni et al. (2010) who obtained
a detection range of 0.40m and 0.35m under water, as well
as those from Curtiss et al. (2009) who obtained a detection
range of 0.40m. However, other authors such as Allan et
al. (2006) and Dickson et al. (2011) obtained ranges from
0.8 to 1m. The detection depth in water is lower than that
in other media (sand, gravel), since the signal attenuation
in water depends on the frequency of the electromagnetic
eld and the water conductivity (Bertoni et al., 2010).
Therefore, results show that the type of media where the
tracer was immersed produces a decrease in the detection
threshold in depth. However, Chapuis et al. (2014)
concluded that the sediment material does not hinder the
Mar-13 Jun-13 Sep-13 Dec-13 Mar-14 Jun-14
0
10
20
30
40
Number of particles
Compact
Compact-platy
Compact-bladed
Compact-elongate
Platy
Bladed
Elongate
FIGURE 12. Recovered tracers as a function of particle shape.
01234
Hs (m)
0
40
80
120
160
Number of particles
20 days
01234
Hs (m)
0
40
80
120
160
Number of particles
10 days
01234
Hs (m)
0
40
80
120
160
Number of particles
5 days
01234
Hs (m)
0
40
80
120
160
Number of particles
1 day
FIGURE 13. Relationship between recovered particles and average
significant wave height (Hs) for the four considered cases.
Mar-13 Jun-13 Sep-13 Dec-13 Mar-14 Jun-14
0
20
40
60
80
Number of particles
< 0 m
0 - 2 m
2 - 4 m
> 4 m
FIGURE 14. Time evolution of the number of recovered tracers as a
function of the elevation where they were found.

ARTICLE IN PRESS
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
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
ARTICLE IN PRESS
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
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
M. Casamayor et al.
Geologica Acta, 13(2), xxxxxxx (2015)
DOI: 10.1344/GeologicaActa2015.13.2.nº manuscript
Short-title
10
signal transmission and thus burial does not signicantly
affect detection other than to increase the vertical distance
between the transponder and the antenna. One of the
most important technical factors that could determine the
detection threshold is the diameter of the antenna. Allan
et al. (2006) indicate that reducing the antenna diameter
by 70% causes a decrease in surface detection distance of
about 50%. Since the different authors use antennas
of different diameter sizes, this could explain the
different detection thresholds. Nevertheless, there are
some other antenna characteristics that are determinant
in detection, such as the copper wire diameter, the
number of loops and the voltage. None of these have
been considered in this paper.
Another important aspect that affects the detection
threshold is the tag orientation inside the pebble. Most
authors normally obtain higher detection distances (both
at surface and at depth) when the tag is perpendicular
to the beach surface, which indicates that this is the
most favourable position, and this is achieved when
the tag is inserted in the short-axis of the particle
(Allan et al., 2006; Dickson et al., 2011; Chapuis et
al., 2014). The tests we have carried out to calibrate
the instrumentation follow the same pattern (Table
3), and the same result was found when the recovery
rate obtained from the sixteen field experiments was
related to the axis through which the tag was inserted
(Fig. 9). Therefore, it is confirmed that particles with
tags inserted through the short-axis are more efficient
as tracers, since they can be detected when buried
in deeper positions and, additionally, they may be
recovered after longer periods.
During the first five days the decrease in detection
rate was very high: 27.8% in 5 days, 5.6% every day.
Between March 19 and April 2013, the decrease in
recovery rate was not so strong: 14.1% in 14 days, 1%
daily. During the last period the recovery rate was 39%
in October 2013 and 26% in July 2014, with a reduction
of 0.05% daily. This pattern seems to indicate that
certain stability in the recovery rate could be reached
if the experiment would continue long enough.
Three main possible factors may explain the
decrease of recovery rates with time: i) tracers may
have been transported out of the survey area; ii) tracers
may become buried at greater depths than the detection
threshold; and iii) tracers may eventually be destroyed
due to abrasion or collision against other particles. Since
detected pebbles are not always the same, the stability
we have found in the last period could indicate that after
a certain period tracers become completely distributed
in the three dimensions (alongshore, cross-shore, and in
depth). Probably, most tracers that were not retrieved
were outside of the survey area or too deeply buried,
but they were still active and may eventually be moved
again to the survey area. Therefore, the arrival of tracers
that were located too far away is compensated by the
exit of other tracers that were in the survey area and, as
a result, it is possible to achieve certain stability in the
number of retrieved particles.
In this study 51 tracers were found seventeen months
after deployment, which represents a 25.8% recovery rate.
This value is very similar to that of Allan et al. (2006),
who obtained 24% and 25% recoveries at the lower and
upper beaches respectively on the high-energy Oregon
coast (USA) after eighteen months. Bertoni et al. (2010)
obtained 77% two months after tracer deployment, while
Curtiss et al. (2009) obtained recovery rates higher than
80%, with a minimum of 73% fourteen months after the
gravel release. Nevertheless, comparison of data is not
easy, since the factors that must be taken into consideration
include not only the time duration of the experiment, but
also the number of storms that occurred and the presence
of protective structures on the beach (Benelli et al., 2012).
Considering these variables, only the results from Allan
et al. (2006) are comparable to ours, since the work by
Bertoni et al. (2010) was carried out on a beach between
lateral groins and bounded seaward by a submerged
breakwater, and the study site of Curtiss et al. (2009) was a
mixed sand gravel beach located near a narrow navigation
channel, which is completely different to the open ocean
beach of San Felipe.
In this study it has been found that there is no signicant
inuence of particle weight and size on the recovery rate
(Figs. 10 and 11), since both large and heavy cobbles, as
well as small and light pebbles, show a decrease in recovery
rates with time. Nevertheless, results obtained in this
paper indicate that particles between 7 and 8 Øw, as well
as those between -5.5 and -6 Ø, have the less favourable
weights and sizes for long-term RFID experiments. This
result agrees with Osborne (2005), who reported a very
signicant decrease in the recovery rate for small particles.
He explained his results arguing that the smaller particles
were more favourable for burial than the larger ones, and
therefore more difcult to be detected.
Tracer detection rates are higher at the upper part of
the beach prole than below the berm (Fig. 14). This
pattern is related to particle transport, since those located
along the lower part are more heavily pulled by waves, so
that they can be moved larger distances, and some of them
may eventually leave the survey area. On the contrary,
it is more difcult for particles above the berm to be
moved by waves, and therefore they remain in the same
location. However, those tracers located close to the berm
crest may fall down to the lower part due to berm retreat.

ARTICLE IN PRESS
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
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
ARTICLE IN PRESS
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
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
Geologica Acta, 13(2), xxxxxxx (2015)
DOI: 10.1344/GeologicaActa2015.13.2.nº manuscript
M. Casamayor et al. Short-title
11
Despite this process, the linear t for particles retrieved
at >4m elevation (Fig. 14) shows a positive slope, which
indicates that the number of particles retrieved above the
berm is gradually increasing. This can only be explained
by onshore transport from the foreshore.
Regarding the effect of waves, even though wave
energy is the most important agent of particle transport,
no clear relationship with recovery has been found. Wave
conditions have been very changeable over time, with
signicant storms during the experiment (see Fig. 4), but
the number of detected particles was quite stable. This is
in disagreement with Benelli et al. (2012), who state that
the frequency of storms is a key factor in recovery rates.
CONCLUSIONS
The RFID system developed for this study is based on
that of Allan et al. (2006) and has proved to be appropriate
because it allowed us to locate tracers accurately and with
acceptable recovery rates. Therefore, it permits to study
the movement of cobbles and pebbles over long periods
of time. The sample preparation process did not affect the
basic characteristics of the particles, since the weight loss
originated only represented 0.42% of the whole tracers.
After deployment of tracers on the beach, sixteen eld
experiments were carried out over seventeen months for
tracer positioning. Recovery rates strongly decreased at
the beginning but tended to stabilize after one year from
initial deployment. After seventeen months a recovery
rate of 26% was achieved, which is very similar to those
reported in the literature for similar periods and open
beaches.
Several factors that could affect the recovery rate
have been considered. Some of these are related to the
particle, and it is conrmed that tracers with the tag
inserted through the short-axis are more efcient in
terms of recovery. Furthermore, the worst particles to
be used in RFID long term experiments are those whose
intermediate-axis are between -5.5 and -6 Ø, as well as
those between 7 and 8 Øw in weight. Regarding particle
shape, platy and elongate shapes seem to be the best for
this kind of experiment.
Finally, two factors regarding the environmental
conditions have been considered. The rst was the
elevation of the beach, and it was clear that the most
favorable position to detect a certain tracer is the upper
part of the beach, above the storm berm. The second
factor was wave energy, and no relationship has been
found between this and recovery rates, which were quite
stable both under storm and calm conditions.
ACKNOWLEDGMENTS
We are grateful to the people that participated in the eld
experiments for their invaluable help. Thanks are due to two
anonymous reviewers whose detailed comments have helped to
improve this manuscript. Wave data were provided by Pilar Gil
from Puertos del Estado.
REFERENCES
Allan, J.C., Komar, P.D., 2004. Environmentally compatible
cobble berm and articial dune for shore protection. Shore
and Beach, 72 (1), 9-18.
Allan, J.C., Hart, R., Tranquili, J.V., 2006. The use of Passive
Integrated Transponder (PIT) tags to trace cobble transport
in a mixed sand-and-gravel beach on the high-energy Oregon
coast, USA. Marine Geology, 232, 63-86. DOI: 10.1016/j.
margeo.2006.07.005
Balcells, R., Barrera, J.L., 1990. Mapa Geológico de España
1:25.000, hoja 1101-III-IV (Arucas). Instituto Tecnológico
GeoMinero de España, Madrid.
Benelli, G., Pozzebon, A., Bertoni, D., Sarti, G., 2012. An
RFID-based toolbox for the study of under- and outside-
water movement of pebbles on coarse-grained beaches.
IEEE Journal of selected topic in applied earth observations
and remote sensing, 5 (5), 1474-1482. DOI: 10.1109/
JSTARS.2012.2196499
Bertoni, D., Sarti, G., Benelli, G., Pozzebon, A., Raguseo, G.,
2010. Radio Frequency Identication (RFID) technology
applied to the denition of underwater and subaerial coarse
sediment movement. Sedimentary Geology, 228, 140-150.
DOI: 10.1016/j.sedgeo.2010.04.007
Bertoni, D., Sarti, G., Benelli, G., Pozzebon, A., Raguseo, G.,
2012. Transport trajectories of “smart” pebbles on an articial
coarse-grained beach at Marina di Pisa (Italy): Implications
for beach morphodynamics. Marine Geology, 291-294, 227-
235. DOI: 10.1016/j.margeo.2011.08.004
Blair, T.C., McPherson J.C., 1999. Grain-size and textural
classication of coarse sedimentary particles. Journal of
Sedimentary Research, 1, 6-19. DOI: 10.2110/jsr.69.6
Bray, M.J., Workman, M., Smith, J., Pope, D., 1996. Field
measurements of shingle transport using “electronic” tracers.
Proceedings 31st Ministry of Agriculture, Fisheries and Food
Conference of River and Coastal Engineers. Keele University,
United Kingdom, 10.4.1-10.4.13.
Carter, R.W.G., Orford, J.D., 1984. Coarse clastic barrier
beaches: a discussion of the distinctive dynamic and
morphosedimentary characteristics. Marine Geology, 60,
377-389. DOI: 10.1016/0025-3227(84)90158-0
Chapuis, M., Bright, C.J., Hufnagel, J., MacVicar, B., 2014.
Detection ranges and uncertainty of passive Radio Frequency
Identication (RFID) transponders for sediment tracking in
gravel rivers and coastal environments. Earth Surface Processes
and Landforms, 39, 2109-2120. DOI: 10.1002/esp.3620

ARTICLE IN PRESS
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
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
ARTICLE IN PRESS
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
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
M. Casamayor et al.
Geologica Acta, 13(2), xxxxxxx (2015)
DOI: 10.1344/GeologicaActa2015.13.2.nº manuscript
Short-title
12
Ciavola, P., Castiglione, E., 2009. Sediment dynamics of mixed
sand and gravel beaches at short time-scales. Journal of
Coastal Research, Special Issue 56, 1751-1755.
Curtiss, G.M., Osborne, P.D., Homer-Devine, A.R., 2009.
Seasonal patterns of coarse sediment transport on a mixed
sand and gravel beach due to vessel wakes, wind waves, and
tidal currents. Marine Geology, 259, 73-85. DOI: 10.1016/j.
margeo.2008.12.009
Dickson, M.E., Kench, P.S., Kantor, M.S, 2011. Longshore
transport of cobbles on a mixed sand and gravel beach,
southern Hawke Bay, New Zealand. Marine Geology, 287,
31-42. DOI: 10.1016/j.margeo.2011.06.009
Dornbusch, U., Williams, R.B.G, Moses, C., Robinson, D.A.,
2002. Life expectancy of shingle beaches: measuring in situ
abrasion. Journal of Coastal Research, Special Issue 36, 249-
255.
Graham, D.J., Midgley, N.G., 2000. Graphical representation of
particle shape using triangular diagrams: an excel spreadsheet
method. Earth Surface Processes and Landforms, 25, 1473-
1477. DOI: 10.1002/1096-9837(200012)25:133.0.CO;2-C
Jolliffe, I.P., 1964. An experiment designed to compare the
relative rates of movement of different sizes of beach pebble.
Proceedings of the Geologists’ Association, 75, 67-86. DOI:
10.1016/S0016-7878(64)80012-2
Krumbein, W.C., 1936. Application of logarithmic moments
to size-frequency distributions of sediments. Journal of
Sedimentary Research, 6 (1), 35-47. DOI: 10.1306/d4268f59-
2b26-11d7-8648000102c1865d
Menéndez, I., Silva, P.G., Martín-Betancor, M., Pérez-Torrado,
F.J., Guillou, H., Scaillet, S., 2008. Fluvial dissection, iostatic
uplift, and geomorphological evolution of volcanic islands
(Gran Canaria, Canary Islands, Spain). Geomoprhology, 102,
189-203. DOI: 10.1016/j.geomorph.2007.06.022
Miller, I.M., Warwick, J.A., 2012. Measuring sediment transport
and bed disturbance with tracers on a mixed beach. Marine
Geology, 299-302, 1-17. DOI: 10.1016/j.margeo.2012.01.002
Miller, I.M., Warwick, J.A., Morgan, C., 2011. Observations of
coarse sediment movements on the mixed beach of the Elwha
Delta, Washington. Marine Geology, 282, 201-214. DOI:
10.1016/j.margeo.2011.02.012
Osborne, P.D., 2005. Transport of gravel and cobble on a mixed-
sediment inner bank shoreline of a large inlet, Grays Harbor,
Washington. Marine Geology, 224, 145-156. DOI: 10.1016/j.
margeo.2005.08.004
Sear, D.A., Lee, M.W.E., Oakey, R.J., Carling, P.A., Collins,
M.B., 2002. Coarse sediment tracing technology in littoral
and uvial environments a review. In: Foster, I. (ed.)
Tracers in the environment. Chichester (UK), Earth Surface
Processes and Landforms, Wiley and Sons, Special Issue,
21-55.
Sneed, D., Folk, R.L., 1958. Pebbles in the Lower Colorado
River, Texas a study in particle morphogenesis. The Journal
of Geology, 66 (2), 114-150. DOI: 10.1086/626490
Voulgaris, G., Workman, M., Collins, M.B., 1999. Measurement
techniques of shingle transport in the nearshore zone.
Journal of Coastal Research, 15 (4), 1030-1039.
Manuscript received November 2013;
revision accepted December 2014;
published Online June 2015.