Вы находитесь на странице: 1из 10

World Environmental and Water Resources Congress 2014: Water without Borders ASCE 2014

Downloaded from ascelibrary.org by University of Saskatchewan on 09/21/16. Copyright ASCE. For personal use only; all rights reserved.

MAST-1D: A Size-Specific Sediment Transport and Tracer Model with OffChannel Storage
J. W. Lauer1, C. Li2, E. Viparelli2, and H. Pigay3
1

Department of Civil and Environmental Engineering, Seattle University, 901 12th


Avenue, Seattle, WA 98122-1090; PH (206) 296-5523; Fax (206) 296-2173; email:
lauerj@seattleu.edu
2
Department of Civil and Environmental Engineering, University of Illinois, 205
North Mathews Ave., Urbana, IL 61801-2352
3
Department of Civil and Environmental Engineering, University of South Carolina,
300 Main St., Columbia, SC 29208
4
Ecole Normale Suprieure de Lyon, 15 Parvis Ren Descartes, Bureau R. 241 - Bat.
Recherche BP 7000, 69342 Lyon cedex 07, France
ABSTRACT
Sediment sourced from river banks, either through widening or progressive
meander bend migration, can represent an important component of a rivers bed
material load. However, sediment exchange with a floodplain is not necessarily even,
and both the size distribution and overall rates of transfer into and out of a channel
can influence the overall grain size structure and morphodynamic change rates on the
bed. The Morphodynamic and Sediment Tracers in 1-D (MAST-1D) program is a
numerical model for size-specific sediment transport and morphodynamic evolution
of a coupled river channel and floodplain system over decadal and longer timescales.
The model differs from other size-specific 1-D sediment transport models in that it
allows for uneven, size-specific exchange with an alluvial floodplain and allows
floodplain and channel elevation, hydraulic capacity, and sediment grain size
distributions to evolve over time. It can also be used to track sediment tracer
concentration in any size class. Objects defined in the publically-available source
code (Python or Visual Basic for Applications) potentially allow users to modify the
sediment storage framework without the need to re-derive governing equations and
provide access to model classes for interested developers. Model runs representing
the Ain River, France illustrate the ability of the model to simulate average bed
elevation change associated with imbalances in either gravel, sand, or mud transport.
Results show that the magnitude of bed change is sensitive to the size distributions of
the material supplied laterally to or removed laterally from a channel.

World Environmental and Water Resources Congress 2014

1362

Downloaded from ascelibrary.org by University of Saskatchewan on 09/21/16. Copyright ASCE. For personal use only; all rights reserved.

World Environmental and Water Resources Congress 2014: Water without Borders ASCE 2014

INTRODUCTION
Numerous computational packages are available for performing size-specific
sediment transport computations (Cui, 2007; Papanicolau et al., 2008; Ferguson et al.,
2009; Veerhar et al., 2010; HEC, 2010; Coulthard et al., 2013, etc). Computations are
usually driven by a 1-D or 2-D hydrodynamic model and require the specification of
initial conditions and boundary conditions and the selection of an appropriate
transport equation. While computational improvements are occurring regularly, it is
still difficult to run models over decadal and longer time-scales, particularly 2-D
models that include a representation of floodplain sediment storage. At these
timescales, the floodplain can store significant volumes of sediment, thereby slowing
the rate of bed elevation change relative to that simulated by purely 1-D channel-only
models (Wright and Parker, 2005; Lauer, 2011). Further, even though many existing
models are developed by public agencies, source code is often not available, and even
if it is available, it can be difficult to modify the model to better represent processes
such as a) transfer of sediment between that in transport and the subsurface during
aggradation (Hoey and Ferguson, 1994; Cui, 2007), b) bank erosion and bar
deposition processes (Parker et al., 2011), c) evolution of channel width (Konrad,
2012; Schottler et al., 2014), d) bedrock and/or erodible limits for the channel bed,
which also includes mechanisms for partial transport of alluvial sediment over a
bedrock channel (Nittrouer et al, 2011), e) capability of handling sediment tracers, f)
the impact of vegetation growth, and a host of other issues.
Floodplains represent a storage repository for sediment. While floodplain
sediment is usually finer than the sediment present in appreciable quantities on the
channel bed (i.e. the bed material), floodplains can still represent the major source of
bed material for an alluvial river. The size-specific nature of sediment transport can
mean that disruption of natural bank erosion and deposition processes along actively
migrating rivers can affect in-channel sediment storage (Reid et al., 2007). Because
bank sediment is an important factor in alluvial river bed evolution, it is important for
a modeling framework to account for its presence. In principle, even moderate
amounts of degradation can lead to an increase in bank supply as the banks become
higher (Simon and Klimetz, 2008; Lauer, 2012). Similarly, even small changes in the
amount of bed sediment transferred to the banks of an alluvial river can have impacts
on the in-channel sediment budget (Viparelli et al., 2013). For these reasons, it is
potentially important to allow for bank change in rivers whose channel boundaries
change position laterally over time.
The present paper presents a modeling framework for channel bed evolution
that incorporates simple (but size-specific) exchange between in-channel and
floodplain sediment. In many ways, it is like existing 1-D, size specific transport
models such as the TUGS model of Cui (2007), but with erodible banks. The model,
referred to here as Morphodynamics and Tracers in 1-D (MAST-1D), is presented in

World Environmental and Water Resources Congress 2014

1363

Downloaded from ascelibrary.org by University of Saskatchewan on 09/21/16. Copyright ASCE. For personal use only; all rights reserved.

World Environmental and Water Resources Congress 2014: Water without Borders ASCE 2014

an initial form in Viparelli et al. (2013) that considers a single sediment size for the
channel bed material. The primary contribution of the present paper is the adaptation
of a formulation to multiple size classes. The code (available in VBA or Python) has
been developed using an object-oriented computer programming framework that can
be extended to include more complex representations of exchange processes or inchannel sediment transport. The present paper describes the basic sediment reservoir
structure and presents model runs illustrating the relatively complex response that can
result from including bank sediment storage and resupply.
SEDIMENT STORAGE RESERVOIR SYSTEM
MAST-1D is based on a conceptual river valley cross section that includes the
following sediment storage reservoirs: Active layer representing channel bed,
floodplain, and a number of substrate reservoirs representing sediment stored beneath
channel and floodplain zones. Figure 1a shows the rectangular cross-section used in
the model, and Figure 1b illustrates the sediment reservoirs at each node. Lateral and
vertical exchange mechanisms vary from reservoir to reservoir. Nodes are connected
in series in the manner common to all 1-D transport models, with sediment moved
downstream from active layer reservoir to active layer reservoir at the computed
transport capacity (although there is a mechanism for a partially alluvial channel). A
single conservation equation applies to each reservoir, greatly reducing numerical
complexity of the governing equations relative to many existing 1-D formulations.
Substrate reservoirs are subdivided into an arbitrary number of layers, the uppermost
of which has thickness that varies over time as the channel aggrades or degrades
(Viparelli et al., 2010). If aggradation is sufficiently high, a new layer is formed, and
if degradation is sufficiently great, the layer is merged with the next lower layer.
b.

a.

BF=Bv-Bc

Overbank

Bc
Floodplain, F

Floodplain, F
Active

Hpb

LA

Substrate, S, split into multiple


layers and into channel (C)
and floodplain (F) zones.
Bv

LF

Bar

Sn,F

A
Sn,C

Sn-1,F

Sn-1,C

Sn-2,F

Sn-2,C

Figure 1. a. Valley cross section and geometry and b. associated sediment reservoirs
and lateral exchange mechanisms applied at each node in the river reach.
In the present implementation of the code, sediment can be exchanged
between reservoirs either by channel migration (in which the volume of the active
layer reservoir does not change), by change in channel width (in which the width of

World Environmental and Water Resources Congress 2014

1364

Downloaded from ascelibrary.org by University of Saskatchewan on 09/21/16. Copyright ASCE. For personal use only; all rights reserved.

World Environmental and Water Resources Congress 2014: Water without Borders ASCE 2014

the active layer does change), or by vertical aggradation or degradation of the bed.
For a degrading system, both the lower boundary of the active layer and the lower
boundary of the floodplain reservoir are assumed to pass into the substrate reservoir,
resulting in a net removal of material from the appropriate part of the substrate and
addition of an equivalent amount of material in each size class to the active layer and
floodplain reservoirs, respectively. Because the horizontal boundaries of the substrate
zones track the channel boundary, exchange occurs between channel and floodplain
zones in each substrate reservoir. This simulates a de-correlation of surface and
substrate size distributions over time as the channel migrates away from recent
sedimentary deposits while still allowing for short-term storage of recently deposited
material in a zone of the substrate that is easily accessed by the channel.
COMPUTATIONS
Sediment is transferred into the active layer reservoir from both upstream and
lateral sources. For most nodes, the upstream source is specified as the computed
transport capacity for the upstream node, a function of discharge and sediment size
distributions computed using Wilcock and Crowe (2003), with shear stress
determined using a standard backwater formulation. Each node object has selfcontained methods for performing these computations, so the programming
framework can easily be extended to allow for more complicated representations of
hydraulics or new transport capacity formulations. Nodes are grouped into reach
objects which, among other things, contain methods for performing hydraulic
computations. For the upstream-most node in a reach, sediment input is whatever is
specified at the boundary. It is also possible to specify lateral sources or sinks in any
size class at each node.
Bed elevation is updated according to a simplified Exner equation:

d
1 1
=
(Qs,in,k Qs,out ,k )
dt 1 dA k

(1)

where is the bed elevation, is porosity of the bed, dA is the bed area represented
by a node (dA = dxc*Bc), Qs,in,k and Qs,out,k are the volumetric bed material sediment
fluxes in size class k into or out of the node, respectively. The sediment efflux term
includes transport to the next downstream node and lateral deposition.
Lateral fluxes to and from the active layer are required for computing Qs,in,k,
and Qs,out,k for the floodplain and active layer. For these two sediment reservoirs,
influx and efflux due to lateral boundary movement are not assumed to be balanced
even when summed across all sediment sizes. Instead, as in Lauer (2011) and
Viparelli et al. (2013), it is assumed that lateral boundary movement of the channel
zone through migration and/or narrowing leads to the development of a laterally
accreted bar whose surface is a specified distance, Hpb, above the surface of the active

World Environmental and Water Resources Congress 2014

1365

World Environmental and Water Resources Congress 2014: Water without Borders ASCE 2014

layer. This distance may be larger or smaller than the average thickness of the
floodplain, depending on floodplain history. The volume of material transferred is:

Downloaded from ascelibrary.org by University of Saskatchewan on 09/21/16. Copyright ASCE. For personal use only; all rights reserved.

Vsed = (1 )s&Hdxc dt

(2)

where Vsed is the total volume of sediment particles transferred, s& is the lateral
change rate (e.g. channel migration or widening rate), H is a thickness, and dt is the
computational time step. For material transferred from channel to floodplain, H = La
+ Hpb, and for material transferred out of any reservoir except the active layer, H =
reservoir thickness (i.e., the volume of sediment plus voids in the deposit divided by
its valley area). Additional volumetric transfer from channel to floodplain is allowed
through overbank deposition, where the overbank deposition rate is computed based
on sediment flux through the floodplain zone a trap efficiency factor for the
floodplain, as described in Viparelli et al. (2013).
For material deposited laterally in bars or in the bed under aggrading
conditions, the size distribution of the transferred material is a mixture of sediment in
the load and that in the active layer, following Hoey and Ferguson (1994) and
Viparelli et al. (2013):
Fk ,deposit = (1 Fwash ,deposit )(Fk , Activelaye r + (1 ) Fk , Load )

(3)

where Fk represents the fraction of bed material in size class k in the deposit, active
layer, or load (as indicated by the subscript) and alpha is a constant between 0 and 1
(set independently for bar deposition and bed aggradation), and Fwash,deposit is the
fraction of washload in the deposit. Fwash,deposit is assumed zero for bed aggradation,
but for lateral deposition in point bars, it is computed as follows:
1
(4)
Fwash ,deposit = 1
Qs , washload
1 + k bar
Qs ,bedmaterial
where kbar is a constant, Qs,washload is the volumetric flux of washload (i.e. mud) , and
Qs,bedmaterial is the total volumetric bed material sediment flux. For kbar > 0, this has the
effect of making point bar deposits somewhat muddier as overall wash load increases.
Once influx and efflux of sediment in all size classes are computed, sediment
volume (voids plus sediment) in all reservoirs can be updated as follows:
1
(Vsed ,migration + Vsed ,widthchange + Vsed ,verticalchange + Vsed ,latsource )in,k
Vk =
1
1
(Vsed ,migration + Vsed ,widthchange + Vsed ,verticalchange + Vsed ,latsource )out ,k (5)

1
1
(Qs, feed ,k Qs ,load ,k )dt
+
1

World Environmental and Water Resources Congress 2014

1366

World Environmental and Water Resources Congress 2014: Water without Borders ASCE 2014

Downloaded from ascelibrary.org by University of Saskatchewan on 09/21/16. Copyright ASCE. For personal use only; all rights reserved.

where Vsed represents the sediment volume transferred into or out of a given
reservoir due to migration, width change, vertical change of the boundary, or from a
lateral source, as indicated by the subscript, Qs,feed,k represents sediment input from
upstream, and Qs,load,k represents sediment transferred from the reservoir downstream
(Qs,load,k and Qs,feed,k are zero except when performing computations for the active
layer). Then, to update volume V and fraction F for any size class k in any reservoir:

Vk = Vk + Vk and Fk = Vk

(6, 7)

INITIAL CONDITIONS AND SPIN-UP REQUIREMENTS


An advantage of a 1-D numerical framework is that the assumptions made
regarding channel-floodplain exchange result in a well-defined steady state. For a
floodplain undergoing exchange with the channel through lateral migration, sediment
is removed from floodplain zone to channel zone when the channel boundary moves.
At steady state, an equivalent amount of sediment in each size class should be
transferred from channel to floodplain (Lauer and Parker, 2008) through a
combination of point bar and overbank deposition. It is not clear whether such steady
states are actually reached, but because it is simple to solve for them analytically with
knowledge of the size distributions of sediment in the channel and floodplain
reservoirs, this allows the model to be initialized from a condition of perfect steady
state, thereby ensuring that any changes simulated in a run are a result of a change in
boundary conditions rather than simply initial model adjustment. This greatly
simplifies the interpretation of results and minimizes the need for spin-up runs.
In the model, lateral erosion consists of a mixture of sediment deposited on
the floodplain through overbank processes and sediment deposited through lateral
accretion. This is shown in Figure 1, where Hpb represents bar sediment, La represents
active layer sediment, and the rest of the floodplain deposit is assumed to consist of
overbank sediment. In the model, the size distributions for these layers at steady state
can be found using equations 3 and 4 and the overbank deposition function (Viparelli
et al., 2013). An approximation of the long-term overbank deposition rate can be
estimated by dividing overbank thickness, as measured in the field, by an estimate of
the average age of the floodplain (i.e. floodplain area divided by the areal reworking
rate), which allows floodplain trap efficiency (Viparelli et al, 2013) and floodplain
roughness to be calibrated. This results in an analytical solution for the size
distribution of floodplain sediment that would be required for lateral exchange to be
perfectly balanced on a size-specific volumetric basis, providing a convenient initial
condition that allows the effect of changes in boundary conditions to be explored
unambiguously. While it is not clear whether floodplains ever achieve such a
condition, this is the approach taken in the model runs presented below.

World Environmental and Water Resources Congress 2014

1367

Downloaded from ascelibrary.org by University of Saskatchewan on 09/21/16. Copyright ASCE. For personal use only; all rights reserved.

World Environmental and Water Resources Congress 2014: Water without Borders ASCE 2014

EXAMPLE APPLICATION
The model has been used to represent the lower 40 km of the Ain River, a
meandering gravel-bed tributary of the Rhone River draining the Jura region of
eastern France. Sediment supply to this reach of the Ain has been cut off due to the
construction of several large dams. The sediment deficit probably became most
severe after the 1960 completion of Allement Dam, which is located at the upstream
boundary of the study area. Under existing conditions, the dams cut off essentially all
bedload, resulting in a condition where bank erosion (and,in recent years, gravel
augmentation) represents the primary sediment source. Significant incision and
coarsening of the bed has been documented in the 12 km immediately downstream of
Allement Dam, mostly upstream of the area where channel migration is most
extensive (Rollet, 2007).
Three runs are considered here, with input data derived from the description
of the system provided by Rollet (2007). Key input parameters common to each run
are provided in Table 1. The only difference between runs relates to the spatial
distribution of lateral exchange.
Table 1. Selected Input Parameters
Variable

Symbol

Value

Units

Floodplain Width

Bf

1000

Channel Width
Initial Slope
Sinuosity 1
Node spacing

Bc
Slope
ChSin
dx c

60
0.00121
1.2
833

m
m

Initial Floodplain Thickness

Lf

Active Layer Thickness

La

0.2

Point Bar Thickness


Initial Median Sediment Size in Active Layer2

Hpb

2.7

D50

kbar

nc

17.6
0.2
0.01
0.1
0.036

mm
sm-1/3

nf

0.1
36.5

sm-1/3

Porosity
Bar washload parameter
Sorting parameter
Manning's n for channel3
Manning's n for floodplain
timestep
1.
2.
3.

dt

days

Affects hydraulic computations by increasing node spacing down channel relative to spacing down floodplain.
Median of 8 bins used to characterize size distribution in active layer.
Varies over run since total roughness includes grain roughness and additions for form drag.

In all cases, the model is started from a condition of equilibrium at the time
sediment supply at the upper end of the reach is dropped to zero due to dam closure.
The model is then run for 50 years under sediment-depleted conditions. The first run
represents a system that does not undergo lateral migration or widening over the run.
This run is similar to what can be performed with most size-specific 1-D
morphodynamic models (e.g. TUGS, HEC-RAS, etc.). The second run represents a

World Environmental and Water Resources Congress 2014

1368

Downloaded from ascelibrary.org by University of Saskatchewan on 09/21/16. Copyright ASCE. For personal use only; all rights reserved.

World Environmental and Water Resources Congress 2014: Water without Borders ASCE 2014

system that migrates at a constant rate across its floodplain, again with no widening.
The third represents a system that migrates across its floodplain at a spatially variable
rate, with constant migration specified for each of five subreaches based on observed
channel change over the period from 1945 to 2000. The model is initialized using a
size distribution for the active layer that is based on pebble counts taken in the lower
reach, where dam-related impacts are presumably minimal (initial active layer D50 =
17.6 mm). For runs 2 and 3, the initial size distribution for the floodplain is computed
based on the assumption of perfect equilibrium under pre-dam conditions. Using this
approach, computation of the size distributions for bar subsurface and floodplain
require specification of a flow duration distribution, channel geometry, friction
parameters (e.g. mannings n for channel and floodplain), sorting coefficients and
kbar, and an initial average thickness of overbank sediment (computed here based on
observed bank geometry). The resulting initial floodplain D50 = 9.2 mm.
RESULTS AND DISCUSSION
Figure 3 illustrates bed elevation and median active layer sediment size (D50)
for each run 50 simulated years after sediment starvation begins.

Figure 3. Results after 50 years of sediment starvation. a. Specified migration rate


input for each run. b. Bed change relative to initial bed position. c. D50 in active layer.
In run 1, the bed incises most dramatically near the upper boundary, but
coarsening of the sediment surface occurs throughout the system. In run 2, the bed
incises more rapidly than in run 1 near the upper boundary but at rates similar to run 1
further downstream, with coarsening limited to the upper portion of the system. In run

World Environmental and Water Resources Congress 2014

1369

Downloaded from ascelibrary.org by University of Saskatchewan on 09/21/16. Copyright ASCE. For personal use only; all rights reserved.

World Environmental and Water Resources Congress 2014: Water without Borders ASCE 2014

3, channel incision is again greatest near the upper boundary, but incision is very low
in reaches that do not migrate rapidly and high near the upper end of reaches that do
migrate rapidly. Downstream of the active reach, coarsening is minimal in run 3.
The difference between the results of runs 2 and 3 illustrates the importance of
the size distribution of the laterally-exchanged sediment. In run 3, the floodplain
supplies much more fine sediment in actively migrating areas than in laterally
inactive reaches. The additional fine sediment mobilizes the coarse fraction, as is well
known regarding the Wilcock and Crowe (2003) transport equation (Cui, 2007).
Perhaps the most important potential contribution of simple models like
MAST-1D is their ability to quickly evaluate the potential importance of lateral
exchange in controlling bed elevation and size structure. In many cases, more
complex 1-D or 2-D hydrodynamic models could drive transport, but it is not clear
what level of effort is warranted to fully characterize exchange. MAST-1D, which
can perform runs like those presented here in few minutes, can be used for a rapid
initial sensitivity analysis, thereby informing the choice of potentially more
complicated numerical modeling approaches. The rapid run times also allow MAST1D to simulate much longer time periods than is feasible using many other models.
CONCLUSIONS
The MAST-1D model represents an actively migrating alluvial river as a
series of nodes, each of which contains a set of sediment storage reservoirs that
experience size-specific sediment exchange as a function of in-channel hydraulics and
specified lateral movement of the channel. Model runs on the Ain River, France help
characterize the importance of eroding streambanks in mitigating the effect of
sediment starvation downstream from a series of dams. The simulated response
depends on spatial variability in migration rates, indicating that bank sediment storage
and resupply have the potential to influence in-channel sediment transport dynamics
on this and other alluvial systems.
ACKNOWLEDGEMENTS
This work was supported by Seattle University and a research grant provided by
cole Normale Suprieure, Lyon, France.

REFERENCES
Coulthard, T., Neal, J., Bates, P., Ramirez, J., Almeida, G., Hancock, G., 2013, Integrating
the LISFLOODFP 2D hydrodynamic model with the CAESAR model: implications
for modelling landscape evolution. Earth Surf. Process. Landf. 38, 1897-1906.
Cui, Y. 2007. The Unified Gravel-Sand (TUGS) model: simulating sediment transport and
gravel/sand grain size distribution in gravel-bedded rivers. Water Res.Res. 43:
WR005330.
Ferguson, R., Church, M., 2009. A critical perspective on 1-D modeling of river processes:
Gravel load and aggradation in lower Fraser River. Water Res. Res. 45: W11424

World Environmental and Water Resources Congress 2014

1370

Downloaded from ascelibrary.org by University of Saskatchewan on 09/21/16. Copyright ASCE. For personal use only; all rights reserved.

World Environmental and Water Resources Congress 2014: Water without Borders ASCE 2014

Hoey, T. B., R. I. Ferguson. 1994. Numerical simulation of downstream fining by selective


transport in gravel bed rivers: Model development and illustration, Water Res. Res. 30,
2251 2260.
Konrad, C.P., 2012. Reoccupation of floodplains by rivers and its relation to the age structure
of floodplain vegetation. J.Geophysical Research Earth Surface 117, G00N13
Lauer, J.W., 2012. The importance of off-channel sediment storage in 1-D morphodynamic
modeling. In: Church, M., P. Biron, and A. Roy, eds., Gravel Bed Rivers: Processes,
Tools, Environments, Wiley, Chichester
Lauer, J.W., Parker, G., 2008. Net local removal of floodplain sediment by river meander
migration. Geomorphology 96: 123-149.
Nittrouer, J.A., Mohrig, D., Allison, M.A., Peyret, A.B. 2011. The lower Mississippi River, a
mixed bedrock-alluvial channel. Sedimentology 58, 1914-1934.
Papanicolaou, A. N., Elhakeem, M., Krallis, G., Prakash, S., Edinger, J., 2008. Sediment
transport modeling reviewcurrent and future developments. J. Hyd.Eng.134: 1-14.
Parker, G., Shimizu, Y., Wilkerson, G.V., Eke, E.C., Abad, J.D., Lauer, J.W., Paola, C.,
Dietrich, W.E., Voller, V.R., 2011, A new framework for modeling the migration of
meandering rivers. Earth Surface Processes and Landforms 36: 70-86.
Reid, S.C., S. N. Lane, J. M. Berney, and J. Holden. 2007. The timing and magnitude of
coarse sediment transport events within an upland, temperate gravel-bed river.
Geomorphology 83:152 182.
Rollet, A.J. 2007. Etude et gestion de la dynamique sedimen-taire sur une section fluviale a
l'aval d'un barrage: le cas de la basse vallee de l'Ain Ph.D. Dissertation. Universite
Lyon.
Schottler, S.P., J. Ulrich, P. Belmont, R. Moore, J.W. Lauer, D. R. Engstrom, J.E.
Almendigner. 2014. Twentieth century agricultural drainage creates more erosive
rivers. Hydrological Processes 28: 1951-1961.
Simon, A., Klimetz, L., 2008. Relative magnitudes and sources of sediment in benchmark
watersheds of the Conservation Effects Assessment Project. Journal of Soil and Water
Conservatio 63: 504-522.
U.S. Army Corps of Engineers, Hydrologic Engineering Center (HEC). 2010. HEC RAS
River Analysis System Hydraulic Reference Manual, Version 4.1.
Verhaar, P.M., Biron, P.M., Ferguson, R.I., Hoey, T.B., 2010. Numerical modelling of
climate change impacts on Saint-Lawrence River tributaries. Earth Surface Processes
and Landforms 35: 1184-1198.
Viparelli, E., Sequeiros, O.E., Cantelli, A., Wilcock, P.R., Parker, G. 2010. River
morphodynamics with creation/consumption of grain size stratigraphy 2: numerical
model. Journal of Hydraulic Research 48: 727-741.
Viparelli, E., J.W. Lauer, P. Belmont, and G. Parker, 2013. A numerical model to develop
long-term sediment budgets using isotopic sediment fingerprints. Computers and
Geosciences 53: 114-122.
Wilcock, P., Crowe, J. 2003. Surface-based Transport Model for Mixed-Size Sediment. J. of
Hyd. Eng. 129: 120128.
Wright, S., Parker, G., 2005. Modeling downstream fining in sand-bed rivers. I: Formulation.
Journal of Hydraulic Research 43: 612-619.

World Environmental and Water Resources Congress 2014

1371

Вам также может понравиться