Fine sediment transport and storage in a gravel bed river, a
pilot study in the Geul River, the Netherlands
Graduation Thesis Master Hydrology
Faculty of Geosciences.
Department of Physical Geography
Utrecht University
Supervisor:
Dr. M. v.d. Perk
Date: 4 July 2013
Jedidja van der Sluis-Stoutjesdijk
3221717
I
Table of contents
List of figures ............................................................................................................................. III
List of tables .............................................................................................................................. III
Summary ................................................................................................................................... IV
1
Introduction to the problem .............................................................................................. 1
1.1
Problem definition ....................................................................................................... 1
2
Study Area .......................................................................................................................... 3
3
Methods ............................................................................................................................. 5
4
3.1
Field measurements .................................................................................................... 5
3.2
Sediment traps............................................................................................................. 5
Results .............................................................................................................................. 17
4.1
Bioturbation .................................................................. Error! Bookmark not defined.
4.2
Fine sediment mass with a gravimetric method ....................................................... 18
4.3
Storage measurements.............................................................................................. 17
5
Model of sediment exchange ........................................................................................... 22
6
Travel time model ............................................................................................................ 26
6.1
Suspended load transport ......................................................................................... 12
6.2
Bed load transport ..................................................................................................... 12
6.3
Results travel time model ............................................. Error! Bookmark not defined.
7
Discussion ............................................................................ Error! Bookmark not defined.
8
Conclusions and recommendations ................................................................................. 29
9
References ........................................................................................................................ 31
10 Appendices ....................................................................................................................... 33
Appendix A ........................................................................................................................... 33
Appendix B ........................................................................................................................... 34
Appendix C............................................................................................................................ 35
Appendix D ........................................................................................................................... 36
Appendix E ............................................................................................................................ 37
Appendix E ............................................................................................................................ 39
Appendix F ............................................................................................................................ 37
Appendix H ........................................................................................................................... 40
Appendix I............................................................................................................................. 40
II
List of figures
Figure 1 Location of the Geul River catchment ....................................................................................... 3
Figure 2: Locations where the samples were taken ................................................................................ 5
Figure 3 a) a trap before it was buried in the river bed. b) a trap after removal from the river. .......... 6
Figure 4 location of the sand used in the traps ....................................................................................... 7
Figure 5 containers with a thin platic film at the bottom containing the sediment. .............................. 8
Figure 6 Filtering installation set up. ....................................................................................................... 9
Figure 7 schematic diagram of the different reservoirs in the sediment travel time model ................ 11
Figure 8 Definition sketch of the bed-load layer (Chanson, 1999)........................................................ 13
Figure 9 Results of the storage measurements..................................................................................... 18
Figure 10 The measured sediment mass versus the duration that the trap was buried. .................... 18
Figure 11 Sediment mass flux measured at each location, during the first fieldwork period. ............ 19
Figure 12 Sediment mass flux measured at each location, during the second fieldwork period. ........ 20
Figure 13 Comparison between the metal concentrations of the sediment from a control trap and
normal traps. ......................................................................................................................................... 22
Figure 14 comparison between the weight of the sediment from a control trap and normal traps. .. 22
Figure 15 ................................................................................................................................................ 21
Figure 16 Weight plotted against time for different discharge values.. ............................................... 23
Figure 17 example of the simulation of one sample modeled using complete discharge time series . 24
Figure 18 The relation of the different fluxes with time (left) and mass (right) are shown for a
discharge value of 1 m3/s. ..................................................................................................................... 24
Figure 19 modeled versus measured data (model using average discharge) ....................................... 25
Figure 20 Modeled mass minus measured mass versus the amount of time the trap was buried. ..... 25
Figure 21 Gamma CDF travel time suspended load per 20 km river reach .......................................... 27
Figure 22 Gamma CDF travel time bed load per 20 km river reach ...................................................... 27
List of tables
Table 1 locations where the traps were placed and amount of traps placed per location ....... 6
Table 2 Discharge measurements (Roer and Overmaas) ......................................................... 19
Table 3 t-test for the mesured weight at the four locations ................................................... 20
Table 4 Input parameters, R2 ans sum os squares after calibration ........................................ 23
Table 5 The input values for the travel time model bed load calculations .............................. 14
Table 6 The computed output values for the travel time model bed load calculations ......... 15
Table 7 amount of sediment, flux, average residence time and velocity per reservoir .......... 26
III
Summary
This pilot study investigates the storage of fine sediments in the river bed of the Geul River,
the Netherlands. The sediment infiltration into the gravel bed is measured at four locations
in the Geul River using two different methods: a gravimetric method and a metal
concentration-based method. Both methods concerned the placement of sediment traps in
the gravel bed, consisting of cylindrical mesh cages with a diameter of 15 cm and a height of
10 cm. In the first method, the cage was filled with clean gravel larger than 12.5 mm (the
size of the mesh openings) collected from the local river bed (with a mean gravel size of 19
mm). After two to sixty-five days, the sediment traps were removed. In the second method,
the sediment traps were filled with clean gravel and 700 grams of fine sand with low metal
concentrations. During the sampling period, this fine sand was contaminated by deposition
of metal-contaminated fine sediment from the Geul River. After four to eight days, the
sediment traps were removed. For both methods the trap is placed in a bag. The bag was
pulled to the bottom of the trap when the trap was placed and was pulled up when the traps
were removed to retain the fine sediment. The fine sediment was washed from the
sediment traps and subsequently dried and weighed. For the second method, the zinc
concentrations of the fine sand and the fine sediment collected from the sediment traps
were measured using a handheld XRF analyser. The sediment flux was then calculated from
the differences between the zinc concentrations in the sediment samples and the fine sand.
The amount of gravel-stored fine sediment was measured with a resuspension cylinder. It
was found that the mean and variation of the fine sediment deposition rates increased with
stream discharge during the sampling period. Changes in the trapped fine sediment weight
were related to changes in discharge. The average fine sediment flux was determined by
fitting a mass balance to the measurements. Based on this flux-discharge relation and
resuspension measurements, average sediment residence times were calculated. A travel
time model for suspended and bed load sediment based on the probability of sediment
having a certain residence time was made. It was found that it takes bed load on average 13
days to travel through a river reach of 20 km and suspended sediment 126 days.
IV
1 Introduction to the problem
1.1 Problem definition
Fine sediment storage in the gravel bed of a river has an effect on many processes, for
example: clogging of pore space in the stream bed; increasing or decreasing the exchange of
water and dissolved constituents between the stream bed and the overlying channel and the
storage of contaminants sorbed to fine sediment. These processes affect the oxygenation of
fish eggs, macroinvertebrate survivorship, nutrient cycling, and pollutant retention (Gartner
et al., 2012).
Sediment flux between the river flow and the river bed has been correlated to different
hydrodynamic and morphologic factors. Petticrew et al. (2007) found positive relations
between both local flow velocity and Froude number and trapped sediment mass. Gooseff
et al. (2006) used groundwater flow modelling and particle tracking and found that slope
breaks in the longitudinal profile of streams cause zones of upwelling and downwelling in
the river bed. Pettricrew et al. (2007) measured the suspended sediment concentration and
the storage between the gravel during controlled release events. This research shows that
discharge peaks mobilize and redistribute fine sediment that is stored in the channel. Krein
at al. (2003) concluded that fine sediment exchange between the water column and the
gravel bed predominantly takes place during and immediately after storm events, due to the
break up of the amoured layer of the gravel bed during such events.
1.2 Aim and research questions
For this study the Dutch part of the Geul River, situated in the southernmost part (SouthLimburg) of the Netherlands, is chosen as study area. The aim of this study is to gain
quantitative insight in the transient storage of sediment in the river bed and, the exchange
rate of sediment between the flow and the river bed in relation to stream hydrodynamics.
For this, the following research questions were formulated:
•
What hydrodynamic factors control the sediment flux between the bed and the river
flow?
•
How does the fine sediment exchange with the river bed change over time;
•
What are the main factors that control disposition and mobilization of fine sediment;
•
What is the sediment travel time through a river section?
To achieve this aim, the morphology and grain size composition of the bed as well as the
flow conditions at different locations in the Geul River were studied. Sediment traps were
placed in the river bed, stored fine sediment was measured by resuspension and an
empirical model (based on a mass balance) and a stochastic model was developed.
1
2
2 Study Area
This research focuses on the Dutch part of the Geul River, situated in the southernmost part
(South-Limburg) of the Netherlands (figure 1). The Geul River originates in eastern Belgium
near the German border and flows into the Maas River a few kilometers
North of Maastricht. The Geul is a meandering gravel bed river. The river has a length of 56
km, of which 36 km are in the Netherlands and the catchment area is about 380 km 2. The
river channel is 3 to 7 m wide (Leenaers H. , 1989). The Geul’s valley gradient decreases from
0.02 m/m near the source, to 0.005 m/m at the Dutch–Belgian border, to a final value of
0.0015 m/m near the point where the Geul disembogues into the Meuse. The Geul
catchment has a size of 380 km2 and the annual precipitation varies from 750 mm at the
confluence with the Meuse to about 1000 mm near the headwaters. As the Geul is a rain-fed
river, the discharge can change quickly. The bed load sediment grain size is dominated by
coarse sand and gravel is mobilized during peak flood events. The present-day fluvial
processes consist of lateral channel erosion and point-bar sedimentation (van Balen, Kasse,
& De Moor, 2008)
Figure 1 Location of the Geul River catchment (a) and a more detailed map of the Geul River catchment (b) (De Moor,
2007). Locations of former mines are indicated with a dot in the more detailed map (Plombières, La Calamine and
Schmalgraf) after Swennen et al., (1994).
The altitude of the catchment varies from 50 m above sea level near the confluence with the
Meuse River, to 400 m above sea level in the source area. The discharge depends largely on
the amount of rainfall. At the Belgian-Dutch border, the discharge is 1.6 m3/s on average and
26 m3/s at maximum. Near its confluence with the Meuse River the average discharge is 3.4
m3/s (data from Waterboard Roer and Overmaas), while occasional peak discharges of more
than 40 m3/s cause local floods. Small scale (local) floods and bankful discharges occur
3
almost every year (mainly during the winter or after heavy thunderstorms) (Van Damme,
2010) (de Moor, 2006).
Mining activities have played a significant role in the Geul catchment. The main sites of
mining and ore treatment are located in the Belgian part of the catchment. The key mining
centers were La Calamine, Plombières and Schmalgraf (figure 1). The exploitation of zinc and
lead had its origin around the 13th century but it was during the 19th and 20th centuries when
the mining area started working on a commercial scale (Leenaers H. , 1989). Industrial
operations started at La Calamine ore body in 1806 consisting mainly of Zn oxides. Industrial
mining at Plombières and Schmalgraf began in 1844 and 1868 respectively causing pollution
with Pb and Zn sulphides (Swennen, 1994). The last mine closed in 1938 but until 1950s the
remediation of the metal ores continued (Leenaers H. , 1989). Due to the inefficient
techniques used and the dumping of tailings in large piles, pollutants were released directly
into the river and accumulated on the river sediments (Leenaers H. , 1989).
4
3 Methods
3.1 Field measurements
Field measurements and samples were taken
during two field campaigns. The first period lasted
from June 15th 2012 until July 18th 2012. The
second period lasted from September 17th until
September 26th 2012.
3.2 Sediment traps
To determine the fine sediment flux between the
river bed and the water, sediment traps were
placed in the river bed at four different locations
(figure 2 and table 1). Every 7-65 days (first field
period) or every 2-14 days (second field period a
certain amount of traps (table 2) was replaced at
Figure 2: Locations where the samples were taken
each location. The traps consist of a round mesh
cage that has a diameter of 15 cm and has a
height of 10 cm. Around each cage a bag was placed. To place the traps a hole was dug in
the river bed, from which the gravel was washed to remove all fine sediment and sieved to
remove the smaller grain sizes. For the placement of the traps per location and field period,
see appendix A until G. The traps were filled with clean gravel that has a grain size of at least
12.5 mm, which is the size of the mesh. Assuming a porosity of well sorted gravel of 0.35 and
a fine sediment bulk density of 1600 kg/m3 a maximum of 1.13 kg fine sediment can be
collected in a trap, which is equal to 64 kg/m2.
The bag was pulled to the bottom of the trap so water can flow through the trap once it was
buried. A wire attached to the handle of the bag was placed along the side of the cage to
ensure that the bag could be pulled up when the traps were removed. The trap was placed
in the pit and then the pit was further filled with clean gravel. After the placement of the
traps, the water depths above the traps were measured. When the traps were removed, the
bag was pulled up to retain the captured sediment. The removed traps were placed in a
bucket with water, in which the gravel was washed. After waiting at least half an hour for
the sediment to settle, the water was removed through decantation. The remaining
sediment sample was dried and weighed in the lab. This method will be further referred to
as the gravimetric method.
5
Table 1 locations where the traps were placed and amount of traps placed per location
Site town
nr
1 cottessen
Location [RD
coordinates]
193464
307709 m
2 partij
192356
312792 m
3 schweiberg 192675
310211 m
4 schin op
189334
geul
318008 m
Location [lat lon]
50°45'28.88"N,
5°55'55.79"E
50°48'13.62"N,
5°55'1.86"E
50°46'50.27"N,
5°55'16.04"E
50°51'3.58"N,
5°52'29.41"E
Nr of traps 1st field
period
4
Nr of traps 2cnd
field period
4
4
6
4
-
4
8
Figure 3 a) a trap before it was buried in the river bed. b) a trap after removal from the river.
In June and July, the traps were placed in the river for four weeks, during this period traps
were taken out and replaced at intervals of one and two weeks. Two traps were left and
retrieved during the second field period, after 65 days. In total 16 traps were buried in the
gravel bed, at four different locations spread over the transect. The locations were chosen in
a way that different features of the river were represented in the study. The traps were
placed far enough apart and in such a manner that there was no disturbance to the sample
when the next trap was placed.
In September, 18 traps were placed at three of the previously sampled locations. During this
period the intervals at which the traps were replaced were shorter. Most were removed and
replaced after two days. During this period there were also a few traps placed that were
filled with 700 grams of fine sand that has another source than the sediment in the Geul
River. This eolian sand has been retrieved from the Soesterduinen (coordinates 52°9 N 5°17
E) and has very low metal concentrations. The Soesterduinen is an active sand drift area at
the northern fringe of the Utrecht ice pushed ridge (figure 4).
6
Figure 4 location of the sand used in the traps (©2013 Aerodata International Surveys, DigitalGlobe, map data ©2013
Google)
The heavy metal concentration in this sand is very low. The sand was placed between the
gravel to simulate the natural situation in the river bed, where fine sediment fills the spaces
in between the gravel. These samples were retrieved in the same way as the other traps. The
collected samples of the fine sediment were fully dried in an oven for approximately 2 days
(larger samples up to a week) at 70 °C. The samples were weighed and the measured weight
was plotted against the following flow hydraulic parameters: Froude number, water depth,
shear stress, Chezy number and discharge. The discharge data have been retrieved from the
waterboard “Roer en Overmaas”. Grab samples were taken of the gravel at all four locations,
which were sieved and measured to determine the grain size distribution of the gravel, using
methods described by Kondolf (1997).
The trapped sediment samples were afterwards manually homogenized by the use of a
mortar in order to obtain more representative average values of the metal concentrations.
The heavy metal concentration in the sediment was measured with a Thermo Fisher
Scientific Niton® XL3t-600 handheld XRF. The sediment was poured into containers with a
thin plastics film at the bottom (figure 5). Al samples were measured in duplicate or three
times if a large difference occurred between the measurements.
7
Figure 5 containers with a thin platic film at the bottom containing the sediment.
By comparing the heavy metal concentrations of the sediment from the Soesterduinen and
the Geul sediment the sediment flux has been be calculated. For this the following equation
was used:
𝑀1 = ((𝐶3 − 𝐶2 )/(𝐶1 − 𝐶2 )) ∗ 𝑀3
Where M1 [kg/m2] is the added mass through sedimentation [kg/m2/day]; M3 [kg/m2] is the
mass of the mixed sample; C1 [ppm] is the concentration of the added mass; C2 [ppm] is the
concentration of the clean sandy sediment; C3 [ppm] is the concentration of the mixed
sample. This method will be further referred to as the metal concentration based method.
The sediment already stored within the gravel bed could be remobilized due to the
disturbance by the digging. To assure that the trapped sediment was originating from the
river flow, and not from inter gravel flow, a control trap was placed with the bag already
pulled up, thus excluding inter gravel flow. The heavy metal concentration profile and
trapped sediment weight found in this control trap was compared with the other traps.
3.3 Gravel-stored sediment
For the collection of the gravel-stored fine sediment, a re-suspension cylinder was used (a
plastic bottomless trashcan). First the water depth within the re-suspension cylinder was
measured. The gravel bed inside the sampler was stirred with a steel shovel up to a depth of
approximately five cm. Five seconds after the stirring, two bottles with a volume of one litre
were filled with the water containing resuspended sediment. To obtain the amount of stored
fine sediment, the water samples were filtered in the lab. Samples were filtered using
0.45 µm 50 mm white gridded filters manufactured by Millipore with a pump set-up and a
waste flask (figure 6). The filters were dried at room temperature (25°C) for two days and
weighed. By extracting the weight of the filter without sediment (previously measured), the
amount of sediment stored in between the gravel was calculated.
8
Figure 6 Filtering installation set up.
3.4 Morphological mapping
For further insight in the local morphology, maps were made of the four locations in the
Geul River. Here the conditions of the river banks were the most important. The maps
include bank stability, bank height and vegetation. On the map areas of bank erosion and
their activity are included. Is the bank is fully overgrown and not very steep, it is marked as
inactive. When the bank is bare, very steep and shows signs of collapse, it is marked active.
In the map also large pools, gravel bars, obstruction in the channel for example by wooden
debris, channel bifurcations and areas where a large amount of fine sediment is found at the
surface of the bed are included.
3.5 Model of sediment exchange
The model to quantify the sediment flux in the Geul river is based on the results of the
gravimetric method. The underlying assumption of the model is that the measured flux is
initially only the influx and that when the trap is filled the influx will be equal to the outflux.
The process of sediment exchange between the river bottom and the water flow has been
modeled in two ways. The models are based a mass balance where the change of trapped
mass over time depends on an influx and an outflux which both are related to discharge.
When the trap is almost empty (t is small), the outflux (Fout) is assumed to be negligible.
When the trap is full (t is large), the outflux is assumed to be equal to the influx. To
incorporate this in the model the outflux has been modelled as a linear relation of the
trapped mass
9
The influx I depends on the discharge, by multiplying it with the parameter β. To incorporate
the relation between the trapped mass and the outflux, the outflux depends on the
discharge and the trapped mass times a parameter α:
dM
= β ∗ (I − αM)
dt
dM
= Iβ − βαM
dt
dM
dt
=
Change of Mass
over time
Iβ
Influx
(Fin)
-
βαM
Outflux
(Fout)
This nonlinear differential equation has the following analytical solution:
M(t) =
I
I
+ (M0 − ) exp(−βαt)
α
α
Where β is equal to aQb. β represents the increased exchange rate due to discharge for both
the input and the output of sediments.
Here the following boundary condition is assumed: when the maximum trapped mass (M max)
is reached, there is no net change of mass over time and therefore influx is equal to the
outflux:
dM
= Iβ − βαM
dt
0 = Iβ − βαMmax
Iβ = βαMmax
I
=α
Mmax
This resulted in one parameter less to calibrate and assured that the model will reach the
Mmax exactly.
For determining parameter b literature was used. The suspended sediment concentration
(SSC) in the Geul river is related to the discharge (Q) as follows:
𝑆𝑆𝐶 = 100.897 𝑄1.693 (Leenaers, 1989)
Assuming that the influx is linearly related to the suspended sediment concentration, b=
1.693
First the data are sorted in classes depending on the average discharge that was measured
during the period the trap was buried. These classes are 0.5 m3/s wide. The model is run for
the average discharge of each class, and the model outputs are compared with that
particular class.
10
However, two samples can have the same average discharge, but still have a very different
timing of discharge peaks. If a peak would occur early in the period that the trap is buried,
than the fluxes will reach equilibrium much earlier than when this peak would have occurred
later. To eliminate this problem, a second approach is used. Each sample is simulated with
the model individually, using the discharge data of the individual time the trap was in place.
3.6 Travel time model
To quantify the sediment travel time in the Geul River a model based on the results of the
first model, the resuspension measurements and formula’s from literature was developed.
For the calculation of the residence time it is assumed that the sediment stored in the
bottom is fully mixed.
The travel time model is based on the probability of transition of a particle of sediment
between different sediment storage reservoirs (figure 7). These probabilities are derived
from calculated sediment residence times. The model is based on the probability of
exchange of particles from four different reservoirs: the river bed and the active flow for
both bed load and suspended load. The suspended transport and bed load transport were
modeled separately, assuming there is no interaction between those reservoirs.
Sediment <8 µm stored in
river flow as suspended
load.
Sediment >8 µm stored in river flow as
bed load.
Sediment >8 µm stored in river bed.
77.5 % (Van Damme, 2010)
Sediment <8 µm stored in river bed.
22.5 % (Van Damme, 2010)
Figure 7 schematic diagram of the different reservoirs in the sediment travel time model
The flux between the reservoirs is derived from the flux as modeled in chapter 5. The
behavior of a particle in a reach of 20 km is modeled. In this model it is assumed that the fine
sediment in the bed is fully mixed. For the residence time calculations the results of the
resuspension measurements are used. The sediment measured using the resuspension
cylinder is classified as clay (<8 µm, Van Damme (2010)). This grain size is assumed to
dominate the suspended load. In order to determine the residence time of the <8 µm
11
sediment in the bed the amount of sediment stored is divided by the sediment flux (aIQb). In
order to determine which part of the flux is responsible for moving which grain sizes the
river sediment size distribution measured by (Van Damme, 2010) is used. It is assumed that
if a certain percentage of the sediment stored in the bed is smaller than 8 µm that the same
percentage of the flux is responsible for the exchange of sediment smaller than 8 µm
between the bed and the flow. The same assumption applies for sediment larger than 8 µm.
By dividing the amount of sediment in each reservoir by the flux an average residence time is
computed. The average residence time is used to compute an exponential distribution. From
this distribution a residence time is selected using random sampling. This way it can be
calculated how long the particle stays in the river bed and how long it is mobile. By running
this model 500 times a travel time distribution is computed, which has the characteristics of
a gamma distribution.
3.6.1 Suspended load transport
The average residence time of the suspended load stored in the bed is calculated by dividing
the average amount of resuspendable sediment (0.54 kg/m2) by the flux. The amount of
sediment in the water column is calculated using the relation between the discharge and the
suspended sediment concentration (SSC) as defined by Leenears (1989).
𝑆𝑆𝐶 = 100.897 𝑄1.693
Then the amount of suspended sediment per m2 over a water depth of 0.5 m is calculated
and divided by the flux. The velocity of the suspended sediment in the water column is
assumed to be the same as the flow velocity, which is calculated by dividing the discharge by
the water depth and width (see table 5 for used input values).
3.6.2 Bed load transport
For bed-load transport, the basic modes of particle motion are rolling motion, sliding motion
and saltation motion (Figure 8). To calculate how much sediment larger than 8 µm it is
assumes that the average amount of resuspendable sediment (< 8 µm) is 23 % of the stored
sediment mass and the sediment larger than 8 µm is 73 % of the stored mass (Van Damme,
2010) and is there for 1.85 kg/m2. This amount is divided by the flux to calculate the average
residence time of sediment larger than 8 µm in the bed. For the calculation of the bed load
concentration in the river flow several formulations are used as stated in the following of
this paragraph.
12
Figure 8 Definition sketch of the bed-load layer (Chanson, 1999).
The following formulations were used:
The bed load transport rate per unit width
3⁄
2
4𝜏0
𝑞𝑠 = √(𝑠 − 1)𝑔𝑑𝑠3 (
− 0.188)
𝜌(𝑠 − 1)𝑔𝑑𝑠
(Meyer-Peter & Mueller, 1948) where qs is the sediment transport, s is the submerged
specific gravity of the sediment, g is acceleration due to gravity, ds is the average sediment
size, ρ is the density of water and τ0 is the boundary shear stress (Chanson, 1999).
The submerged specific gravity of the sediment
𝜌𝑠
𝑠 = ≅ 2.65
𝜌
(Recking, Liébault, Peteuil, & Jolimet, 2012)
Shear stress
𝜏0 = 𝜌𝑔𝑠𝑖𝑛𝜃
Where ρ is the water density, g is the g is the acceleration due to gravity, d is depth and sinθ
is the bed slope.
Shear velocity
𝑉∗ = √𝑔𝑠𝑖𝑛𝜃
The particle Reynolds number
𝑉∗ 𝑑𝑠
𝑅𝑒∗ =
𝑣
Where v is the kinematic viscosity
13
The average bed-load layer thickness, which is equivalent to the average saltation height
measured normal to the bed
1⁄
3
(𝑠 − 1)𝑔
𝛿𝑠 = 𝑑𝑠 0.3 (𝑑𝑠 (
)
𝑣2
0.7
)
√
𝜏∗
−1
(𝜏∗ )𝑐
(van Rijn, 1984) where τ* is the Shields parameter and (τ*)c is the critical Shields parameter
for initiation of bed load transport.
The bed load concentration
𝑄𝑏
𝐶𝑏 =
𝑄
(Belaud & Paquier, 2000) The bed load concentration is defined as the ratio of bed load
discharge to liquid discharge.
The bed load velocity
𝑞𝑠
𝑣𝑠 =
𝑑𝑠
The bed load discharge
𝑄𝑠 = 𝑞𝑠 ∗ 𝑤
Where w is width
In tables 5 and 6 the input and output values for the bed load calculations are given. The
values for the discharge, depth and width are also used for the suspended sediment
calculations.
Table 2 The input values for the travel time model bed load calculations
Vwater
Q
depth
slope
dbedload
width
g
viscosity
s
0.25
1
0.5
0.004
0.00075
8
9.81
1.01E-06
2.65
m/s
m3/s
M
M
M
m/s2
m2/s
14
Table 3 The computed output values for the travel time model bed load calculations
tau 0
V*
tau *
Re*
qs
Qs
Cs
Vs
19.58
0.14
1.62
104.32
0.0013
0.01
0.27
0.13
Pa
m/s
m2/s
m3/s
kg/m2
m/s
15
16
4 Results and discussion
4.1 Field observations
In the traps large amounts of amphipods were found. Amphipods are typically less than 10
millimeters long. In one case, the bottom of the bucket used for decanting was so full of
Amphipods that the whole bottom was covered with them. The Amphipods were set free
before decanting. Occasionally there were also fish (< 10 cm) found in the traps, hiding in
between the gravel. This shows that there is bioturbation taking place in the river bed of the
Geul River.
At the location at Cottessen (location 1) the river banks had an approximate height of 2 m
with average to high slopes. The banks were mainly vegetated with high grass and also trees.
The bank erosion was mainly inactive.
At the location at Partij (location 2) the river banks had an approximate height of 1 - >3 m
with average to high slopes. The banks were vegetated with high grass and trees. The bank
erosion directly upstream of the sampling location was mainly active.
At the location at Schweiberg (location 3) the river banks had an approximate height of 2 >3 m with average to high slopes. The banks were vegetated with high grass and trees. The
majority of the banks are indicated as inactive for erosion.
At the location at Schin op Geul (location 4) the river banks had an approximate height
varying from <1 m up till >3 m with low to high slopes. The banks were mainly vegetated
with trees. The bank erosion was mainly inactive.
4.2 Storage measurements
As explained in chapter 3, a resuspension technique was used to measure the gravel-stored
fine sediment. With this method the amount of fine sediment stored in the upper
centimeters of the river bed is measured. At the four locations measurements were made on
three morphological units. Measurements were made in bends of the channel, on straight
parts of the channel and on pointbars. Results of the storage measurements are given in
figure 9. The highest amounts of gravel-stored fine sediment were measured on the
pointbars. There is much of variation in the measurements, also within the same
morphological unit.
storage
storage (kg/m2)
2.0
1.5
1.0
0.5
0.0
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16
measurement
17
Figure 9 Results of the storage measurements. Red are measurements made in a bend of the river, measurements made
in a straight part of the river are blue and those made on a point bar are green.
4.3 Fine sediment mass with a gravimetric method
It was found that water discharge explained the largest part of the variation in the measured
amounts of trapped sediment. There is much scatter in the data, especially for those
measurements with a duration of around a week (figure 10). After circa ten weeks the traps
have collected virtually the maximum amount of fine sediment (64 kg/m2). In some traps this
weight of collected sediment is reached even faster, as at location 3 a weight of 54.7 kg/m2 is
reached after 15 days.
70
M measured [kg/m2]
60
50
40
Location 1 Cottessen
30
Location 2 Partij
Location 3 Schweibergen
20
Location 4 Schin op Geul
10
0
0
20
40
60
80
Duration [day]
Figure 10 The measured sediment mass versus the duration that the trap was buried.
18
The measured masses are plotted with the measured discharge (figures 11 and 12). The
changes in the sediment mass flux coincide with changes in discharge.
To compare the results with the discharge several examples are described below. In these
examples, the mass (not the mass flux) is taken into account, while only looking at samples
with the the same burial duration. This is to exclude the effect of the duration of the trap
being buried on the flux, because, as the duration approaches the moment the influx and
outflux are in equilibrium, the net flux decreases. For the first part of the second field period
(until September 22) the discharge conditions were very stable and calm, and the measured
mass of the samples are low and stable, namely 0.83 ± 0.33 kg/m2 (only considering traps
which have been buried for 2 days). Once small peaks in discharge started to occur, the
average measured mass increases, but also the variation in the data increases. The average
trapped sediment mass during the last part of the second field campaign is 1.68 ± 2.83 kg/m2
(only considering traps which have been buried for 2 days). The higher average sediment
mass during the first field campaign (June-July) can also be related to the higher discharge
during this period. The average discharge was higher (table 2) than during the second period
and the peaks were also higher (figures 11 and 12).
Table 4 Discharge measurements (Roer and Overmaas)
Discharge (m3/s)
min.
max.
Average
SD
Field period 1
0.532
3.098
0.919
0.462
Field period 2
0.468
1.330
0.653
0.133
Flux [kg/m2/day]
8
6
4
2
0
6/5/2012 6/12/2012 6/19/2012 6/26/2012 7/3/2012 7/10/2012 7/17/2012
flux location 1
Flux location 2
Flux location 3
Flux location 4
Discharge Cottessen [m3/s]
4
3.6
3.2
2.8
2.4
2
1.6
1.2
0.8
0.4
Discharge [m3/s]
First fieldcampaign
Figure 11 Sediment mass flux measured at each location, during the first fieldwork period, plotted with the discharge
measured at Cottessen.
19
Flux [kg/m2/day]
4
3.6
3.2
2.8
2.4
2
1.6
1.2
0.8
0.4
6
4
2
0
9/16/2012
9/18/2012 9/20/2012
Flux location 1
9/22/2012 9/24/2012
Flux location 2
Flux location 4
Discharge [m3/s]
Second fieldcampaign
8
9/26/2012
Discharge Cottessen [m3/s]
Figure 12 Sediment mass flux measured at each location, during the second fieldwork period, plotted with the discharge
measured at Cottessen.
The weight of the trapped mass is compared for the different locations using a paired twosample t-tests. The data are paired when they are measured over the same period. In table 3
the P values for the different pairs and the distance between the different locations is
shown. Location 3 shows the most different results.
Table 5 t-test for the mesured weight at the four locations
Location
2 vs 4
1 vs 2
1 vs 4
2 vs 3
3 vs 4
1 vs 3
P value
distance [m]
0.68
0.63
0.59
0.17
0.13
0.07
9775
8250
18025
3578
13353
4672
The results of the gravimetric method show a lot of scatter. It is, however, evident that the
flux increases when the discharge increases. An explanation for these higher sediment fluxes
when the discharges increase is that the sediment transport capacity will increase as well.
When more sediment is being transported, there is more sediment available for deposition
and infiltration. A higher suspended load during higher discharge is confirmed by
measurements of the Roer en Overmaas Regional Water Authority (appendix I). Pettricrew
et al. (2007) measured the suspended sediment concentration and the storage between the
gravel during controlled release events. This research shows that discharge peaks mobilize
and redistribute fine sediment that is stored in the channel. Krein at al. (2003) concluded
that fine sediment exchange between the water column and the gravel bed predominantly
takes place during and immediately after storm events, due to the break up of the amoured
layer of the gravel bed during such events. During normal discharge conditions there seems
to be a dynamic equilibrium between the sediment in the river flow and the sediment stored
in the river bed. This is the flux that is calculated with the sediment exchange model.
20
4.4 Fine sediment mass with a metal concentration based method
Flux [kg/m2/day] measured with concentration
based method
Sediment mass fluxes were calculated using four elements that have high concentrations in
the Geul river sediment and low concentrations in the clean sand. These elements are zinc,
titanium, iron and lead. The fluxes calculated with the metal concentration based method
for each element can be found in figure 13 plotted against their control samples which are
retrieved using the gravimetric method. Using a T-test it can be concluded that the fluxes
calculated using zinc show the most similarity with those measured with the gravimetric
method.
2.5
2
1.5
Zn
Ti
1
Fe
Pb
0.5
0
0
0.5
1
1.5
2
2.5
Flux [kg/m2/day] measured with gravimetric method
Figure 13 Results of the gravimetric method plotted against the results of the concentration based method for different
metals.
Sediment flux based on metal concentrations are studied in further detail in the MSc thesis
by Van der Werf (2013). Graphs and tables of the calculations and results can be found in the
mentioned study.
The metal concentration profile and trapped sediment weight, found in the trap with the
bag already pulled up at location 2, was compared with the other traps (figures 14 and 15).
From figure 14 it can be concluded that the metal concentrations in the control bag show no
relevant difference with the concentration profile of the other traps at the same location.
This suggests that the trapped sediment originates by vertical sedimentation rather than
lateral supply by hyporheic flow in the gravel bed.
21
concentration [ppm]
25000
20000
15000
10000
5000
Fe
Pb
Zn
average values location 2.5,
Partij
Bag pulled up
0
S
Weight sediment per trap [g]
Figure 14 Comparison between the metal concentrations of the sediment from a control trap and normal traps.
70
60
50
40
30
20
10
0
0
1
2
3
4
5
6
7
8
9
Duration [days]
2.2.2 17-9-'12
2.2.2 19-9-'12
2.2.2 21-9-'12
2.2.4 17-9-'12
2.2.4 19-9-'12
2.2.4 21-9-'12
2.2.5 17-9-'12
2.2.6 17-9-'12
2.2.5 bag pulled up 21-9-'12
Figure 15 Comparison between the weight of the sediment from a control trap and normal traps.
The metal concentration based method, used in this study to calculate fine sediment fluxes,
much more resembles the natural situation. In this method the trap does not act as a sink
until it is filled with sediment, but it directly measures the equilibrium flux. The fluxes
calculated with this method result in values in the same range as those measured with the
gravimetric method. As only a small amount of traps were used for this method, more data
is needed to determine how accurate it is.
22
4.5 Model of sediment exchange
The calibration of the model to the data resulted in the following parameters and sum of
squares and R2:
Table 6 Input parameters, R2 ans sum os squares after calibration
model using average discharge
model using complete discharge
time series
I*a
b
alpha
R2
SS
3.47E-05 1.693
3.01E-08 0.6531 6225.79
3.09E-05 1.693
3.01E-08 0.6352 6238.62
The result of the model using average discharge is depicted in figure 16.
60
Weight [kg/m2]
50
40
30
20
10
0
0
1000000
2000000
3000000
4000000
5000000
6000000
7000000
8000000
Time [s]
Q=0.55-0.6
Q=0.6-0.65
Q=0.65-0.7
Q=0.7-0.75
Q=0.75-0.8
Q=0.8-0.85
Q=0.85-0.9
Q=0.9-0.95
Q=0.95-1
Q=1-1.05
Q=1.3-1.35
Q=1.8-1.85
0,575
0,625
0,675
0,725
0,775
0,825
0,875
0,925
0,975
1,025
1,825
1,325
Figure 16 Weight plotted against time for different discharge values. The points are the measurements, the lines are the
model results from the model using average discharge.
23
Figure 17 depicts an example of the simulation of one sample modeled individually.
1.1.1 15/6/'12-30/6/'12
2.5
20
2
15
1.5
10
1
5
0.5
0
10/6/2012 0:00
20/6/2012 0:00
modeled weight [kg/m2]
0
10/7/2012 0:00
30/6/2012 0:00
Measured weight [kg/m2]
Discharge Cottessen[m3/s]
Figure 17 example of the simulation of one sample modeled using complete discharge time series
In this model the inlfux is constant if the discharge is constant. The outflux increases
exponential with time and linear with mass (figure 18). If the discharge increases, the
equilibrium level is reached faster.
3.00E-05
3.00E-05
2.50E-05
2.50E-05
Flux [kg/m2/s]
Flux [kg/m2/s]
Discharge [m3/s]
Weight [kg/m2]
25
2.00E-05
1.50E-05
1.00E-05
5.00E-06
2.00E-05
1.50E-05
1.00E-05
5.00E-06
0.00E+00
0.00E+00
0
10000000
20000000
0
Time [s]
influx
outflux
20
40
60
80
Time [s]
som
influx
outflux
som
Figure 18 The relation of the different fluxes with time (left) and mass (right) are shown for a discharge value of 1 m3/s.
From the sum of squares (table 4) it can be noted that the second approach where each
sample is simulated by the model using complete discharge time series, is fitting the date
worse than the approach using average discharge values. A cause for this could be the
hysteresis. Many catchments, particularly small ones, exhibit clockwise or positive
hysteresis. The sediment concentrations increase more rapidly than the discharge, peaks
before the discharge does and shows much lower concentrations at the same discharge on
the falling limb (Burt & Allison, 2010).
24
In figure 19 the modelled weight versus the measured weight is plotted. The data from
location 3 are on average underestimated by the model.
70
M modeled [kg/m2]
60
50
40
Location 1 Cottessen
30
Location 2 Partij
Location 3 Schweibergen
20
Location 4 Schin op Geul
10
0
0
20
40
60
80
M measured [kg/m2]
Figure 19 modeled versus measured data (model using average discharge)
In figure 20 the modeled mass minus measured mass versus the amount of time the trap
was buried. The variation first increases and than decreases again.
M modeled - M measured [kg/m2]
40
30
20
10
Location 1 Cottessen
Location 2 Partij
0
-10
0
20
40
60
80
Location 3 Schweibergen
Location 4 Schin op Geul
-20
-30
-40
Duration [days]
Figure 20 Modeled mass minus measured mass versus the amount of time the trap was buried.
Now that the parameters are known, the equilibrium flux can be calculated as follows: aIQb.
In figures 16-20 the weight is modelled for 10 cm depth. It is assumed that the flux resulting
from calibrating the model with the measurements would also have been found if the trap
would have been of different depth. This is, assuming that the time Mmax is reached is
scaled with the depth of the trap. So if for instance if the depth of the trap would have been
5 cm, Mmax would have been reached twice as fast as in the 10 cm deep traps and the
Mmax itself would have been halved. The graph of figures 16-20 would look different, but
25
the parameters a, I and b would remain unchanged. Alpha would be changed, as alpha is
dependent on Mmax. For traps of every size to result in the same Mass versus time graph
both the mass and time should be divided by the depth of the trap.
The naturally occurring porosity has been disturbed in the sediment traps, this is because
the fine gravel (<12mm) is sieved out and because the gravel is repositioned and therefor
ordered differently. This will result in a higher amount of inter gravel stored sediment when
the trap is full, than in the natural situation. Therefore, in natural conditions, after a storm
event, when the armoured layer is broken up, it will take less time for the equilibrium flux to
be reached.
It was found that the control bag showed no deviation compared with the other
measurements, implying that the contribution of inter gravel flow is negligible. However
(Petticrew, Krein, & Walling, 2007) found that there was no difference between
measurements made with lidded and non-lidded sediment traps, which led to their
conclusion that inter gavel flow very much contributed to the sediment flux between the
river bed and the flow.
4.6 Travel time model
By carrying out 500 simulations a probability distribution of possible model outcomes arises.
In table 7 the compartments and their characteristics are listed.
Table 7 amount of sediment, flux, average residence time and velocity per reservoir
reservoir
bed
flow
bed
flow
Sediment
size
> 8 µm
> 8 µm
< 8 µm
< 8 µm
Average amount of
sediment stored [kg/m2]
1.85
0.27
0.54
0.004
Flux
[kg/m2/s]
2.69E-05
2.69E-05
7.81E-06
7.81E-06
Average residence
time [s]
68696.70
10042.19
68696.70
505.22
Velocity
[m/s]
0
0.13
0
0.25
In figure 21 and 22 the cumulative probability of the travel time of suspended load and bed
load are depicted. Based on the travel time model it is found that it takes bed load on
average 13 days to travel through a river reach of 20 km and suspended sediment 126 days.
The distribution of the travel times show a gamma distribution, because the probability
distribution used for the residence times per reservoir are exponential.
The bed load travel time is almost 10 times shorter than the suspended load travel time. The
reason for this is that the ratios between the amounts of sediment in the flow and the
amount of sediment in the bed are very different. For bed load, there are 6 times more
particles stored in the bed than in the flow, whereas for suspended sediment this is 135
times. Therefor the suspended sediment is stored in the bed much longer, as the change of a
particle being picked up by the flow is much smaller than for bed load particles.
26
Travel time suspended load per 20 km river reach
1
0.9
0.8
Probabillity
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
60
80
100
120
140
160
180
200
Travel time suspended load [days]
Modeled data
Gamma CDF (alfa=80.48;beta=1.57)
Figure 21 Gamma CDF travel time suspended load per 20 km river reach
Travel time bed load per 20 km river reach
1
0.9
0.8
Probabillity
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0
10
20
30
Travel time bedload [days]
Gamma CDF (alfa=10.56;beta=1.30)
Modeled data
Figure 22 Gamma CDF travel time bed load per 20 km river reach
27
40
As a first approximation, exchange between suspended and bed load sediment is not taken
into account in the travel time model. However in a natural situation this does happen
(García, 2008). This assumption has most likely resulted in the travel times for particles
smaller than 8 µm and particles larger than 8 µm to be more different than they would have
been if exchange between suspended and bed load sediment was taken into account in the
model. This is because than for example a small particle (<8 µm) could also end up in the >8
µm reservoir, where other residence times are used as input values.
The study of Arribas Arcos (2011) predicted that it takes over 1000 years to remover 70% of
the zinc and lead contamination in the Geul river system. Comparing this with the results of
the travel time model, contaminants sorbed to suspended sediment in the river will flush
through the system at an average speed of 6.3 km/day. This means that the storage of
contaminants in the river itself is a very small part of the total residence time of the
contaminants in the river system.
28
5 Conclusions and recommendations
The analysis of the data compiled for this study gained quantitative insight in the transient
storage of sediment in the river bed and the exchange rate of sediment between the flow
and the river bed in relation to stream hydrodynamics. In this study two methods were used
to determine the infiltration of fine sediment in a gravel bed. The first method is a
gravimetric method, in which the weight of the fine sediment caught by traps that were
buried in the gravel bed was used to determine the fine sediment flux. The second is a metal
concentration based method. Using the results of the gravimetric method, two models were
developed. A sediment exchange model and a sediment travel time model.
A significant positive correlation was found between river discharge and sediment flux.
Furthermore it is found that bed load sediment is stored much shorter in the river bed than
suspended load sediment. The conclusions for each subquestion are elaborated in more
detail below.

What hydrodynamic factors control the sediment flux between the bed and the river
flow?
The mean and variation of the fine sediment deposition rates increased with stream
discharge during the sampling period. It was found that the hydrodynamic factor
mostly correlated to the sediment flux between the bed and the river flow is the
discharge.

How does the fine sediment exchange with the river bed change over time?
Under normal flow conditions, the influx and outflux are in equilibrium. The fine
sediment flux is determined by fitting a mass balance through the results of the
gravimetric method, resulting in a relation between sediment flux and river
discharge.

What are the main factors that control disposition and mobilization of fine sediment?
The river discharge, which is a direct result of rainfall intensity, explained the largest
part of the variation in the measured amounts of trapped sediment. Especially high
flow conditions cause the fine sediment that is stored in the channel to mobilize and
redistribute, as the armoured layer of the gravel bed can then be broken up. Bankfull
conditions occur almost every year.

What is the sediment travel time?
Based on the flux-discharge relation found in this study, river sediment size
distribution and resuspension measurements, the average sediment residence times
were calculated. A travel time model for suspended and bed load sediment, based on
the probability of sediment having a certain residence time, was made. It was found
that it takes bed load on average 13 days to travel through a river reach of 20 km and
suspended sediment 126 days.
29
For future research it will be valuable to measure the ratio of gravel and fine sediment in the
riverbed, so that the difference in porosity between the gravel in the traps and the natural
situation can be quantified. As the traps filled with sand simulate a more natural situation it
is worthwhile to further apply this technique. It might also be useful to pour the clean sand
around the trap, together with the clean gravel. This will prevent a sudden change in
porosity around the trap. When these traps are buried for several week/moths the
equilibrium flux can be studied. However, when looking at the influence of changes in
discharge it is more useful to bury the traps for a shorter amount of time, in the order of
days.
To improve the travel time model it would be very useful to measure the bed load and
suspended load in the river flow. Also it is recommended to have more input values in the
form of random samples from a probability distribution, as for now the discharge, grain size
distribution, width and depth are average values.
Overall the data set obtained in this research needs to be extended. This will improve the
insight in the effects of the different locations, as at this moment the scatter is so large that
no relations between the morphological features and the measurements where found.
30
6 References
Arribas Arcos, V. (2011). Modeling and prediction of the natural decontamination of the
mining-impacted Geul River floodplain. MSc Thesis Hydrology, Utrecht University.
Belaud, G., & Paquier, A. (2000). Estimation of the total sediment discharge in natural stream
flows using a depth-integrated sampler. Aquatic sciences, Volume: 62, Issue: 1, 39-53.
Burt, T. P., & Allison, R. J. (2010). Sediment cascades : an integrated approach. Chichester,
West sussex: Wiley-Blackwell.
Chanson, H. (1999). The Hydraulics of Open Channel Flow. London: Arnold.
Corazza, M. Z., Abrao, T., Lepri, F. G., Gimenez, S. M., Oliveira, E., & Santos, M. J. (2012).
Monte Carlo method applied to modeling copper transport. Stoch Environ Res Risk
Assess 26, 1063–1079.
de Moor, J. (2006). Human impact on Holocene catchment. Amsterdam: Vrije Universiteit
Faculty of Earth and Life Sciences Department of Palaeoclimatology and
Geomorphology.
De Moor, J. V. (2007). Simulating meander evolution of the Geul. Earth Surface Processes
and Landforms 32, pp. 1077–1093.
García, M. H. (2008). Sedimentation Engineering: Processes, Measurements, Modeling, and
Practice. ASCE Publications.
Gartner, J., Renshaw, C., Dade, W., & Magill, F. (2012). Time and depth scales of fine
sediment delivery into gravel stream beds: Constraints from fallout radionuclides on
fine sediment residence time and delivery. Geomorphology, Volume: 151-152, pp. 3949.
Gartner, J., Renshaw, C., Dade, W., & Magilligan, J. (2012). Time and depth scales of fine
sediment delivery into gravel stream beds: Constraints from fallout radionuclides on
fine sediment residence time and delivery. Geomorphology 151–152, 39–49.
Kondolf, G.M. (1997), Application of the pebble count: notes on purpose, method and
variations, Journal of the American Water Resources Association, vol. 33, no. 1, pp.
79-87
Krein, A., Petticrew, E., Udelhoven, T. (2003). The use of fine sediment fractal dimensions
and colour to determine sediment sources in a small watershed. Catena, 53, pp. 165179
Leenaers, H. (1989). The dispersal of metal mining wastes in the catchment of the river Geul
(Belgium - The Netherlands). Netherlands Geographical Studies, 102, pp 1-200
Leenaers, H., & Schouten, C. (1989). Soil erosion and floodplain soil pollution: Related
problems in the geographical context of a river basin. Sediment and the Environment
(Proceedings of the Baltimore Symposium, May 1989).
Meyer-Peter, E., & Mueller, R. (1948). Formulas for bed-load transport. Proceedings, 2nd
Meeting IAHR, Stockholm, 39–64.
Michael N. Gooseff, J. K. (2006). A modelling study of hyporheic exchange pattern and the
sequence, size, and spacing of stream bedforms in mountain stream networks,
Oregon, USA. Hydrological Processes 20, pp. 2443–2457.
Petticrew, E., Krein, A., & Walling, D. (2007). Evaluating fine sediment mobilization and
storage in a gravel-bed river using controlled reservoir releases. Hydrolgical Processes
21, pp. 198–210.
31
Recking, A., Liébault, F., Peteuil, C., & Jolimet, T. (2012). Testing bedload transport equations
with consideration of time scales. Earth Surface Processes and Landforms 37 (7), 774789 .
Roer and Overmaas, Waterschap. Debiet-, waterstand en neerslaggegevens. Accesed in 2013
at http://wro.lizardsystem.nl/
Swennen, R. V. (1994). Heavy metal contamination in overbank sediment of the Geul river
(East Belgium): Its relation to former Pb-Zn mining activities. Environmental Geology
24, pp. 12-21.
van Balen, R., Kasse, C., and De Moor, J. (2008). Impact of groundwater flow on meandering;
example from the Geul River, The Netherlands. Earth Surface Processes and
Landforms, pp. 2010-2028.
Van Damme, A. (2010). Zinc speciation in overbank sediments contaminated by mining and
smelting activities. PhD Thesis, Katholieke Universiteit Leuven.
van der Werf, M. (2013). Fine sediment transport and contaminant distribution in a gravel
bed river: a pilot study in the Geul river, the Netherlands. MSc Thesis Hydrology,
Utrecht University.
van Rijn, L. (1984). Sediment Transport, Part I: Bed Load Transport. J. Hydraul. Eng., 110(10),
1431–1456.
32
7 Appendices
Appendix A
Schematic map of the placement of the traps at Cottessen, location 1; field period 1 (June July 2 012)
33
Appendix B
Schematic map of the placement of the traps at Cottessen, location 1; field period 2
(September 2012)
34
Appendix C
Schematic map of the placement of the traps at Partij, location 2; field period 1 (June - July
2012)
35
Appendix D
Schematic map of the placement of the traps at Partij, location 2; field period 2
(September 2012)
36
Appendix E
Schematic map of the placement of the traps at Schwijbergen, location 3; field period 1
(June - July 2012)
37
Appendix F
Schematic map of the placement of the traps at Schin op Geul, location 4; field period 1
(June - July 2012)
38
Appendix G
Schematic map of the placement of the traps at Schin op Geul, location 4; field period 2
(September 2012)
39
Appendix H
Size distribution of the gravel bed at the different locations
100.0
90.0
80.0
Location 1
70.0
location 2
%
60.0
location 3
50.0
location4
40.0
location 1 cumm
30.0
location 2 cumm
20.0
location 3 cumm
location 4 cumm
10.0
0.0
0.0
2.0
4.0
6.0
sediment diameter [cm]
Appendix I
Suspended sediment (mg/L)
80
20
18
16
14
12
10
8
6
4
2
0
70
60
50
40
30
20
10
0
14/06/12
04/07/12
24/07/12
13/08/12
02/09/12
Geul Grens
Geul Voor rwz Wijlre
Geul Valkenburg
Geul Bunde
Prediction leenaers (p 80) SSC [mg/L]
Discharge Cottessen
40
22/09/12
Discharge Cotessen[m3/s]
Discharge and suspended load measurements (Roer and Overmaas) and suspended load
predictor (Leenaers H. , 1989).