Modelling floodplain sedimentation using particle tracking

Both climate change and river rehabilitation projects induce changes in floodplain sedimentation. Notably along the lower River Rhine, the sediment deposition patterns and rates are subject to change. To assess the magnitude of these changes, we developed the MoCSED model, a floodplain sedimentation model within a geographical information system for the lower Rhine River. We based MoCSED on the ‘method of characteristics’ (MoC), a particle tracking method that minimizes numerical dispersion. We implemented the MoCSED model in the PCRaster dynamic modelling language. The model input comprises initial suspended sediment concentrations, water levels, flow velocities, and longitudinal and transverse dispersivities. We used a combination of the Krone and Chen concepts to calculate the subsequent sedimentation (SED routine). We compared the model results with sediment trap data for the Bemmel floodplain along the Dutch Waal River during the 2003 inundation. This comparison showed that MoCSED was able to simulate the pattern of sediment deposition. In addition, the model proved to be an improvement in comparison with a conventional raster‐based floodplain sedimentation model for the lower River Rhine. In future, MoCSED may serve well to study the impact of a changing discharge regime due to climate change and floodplain rehabilitation plans on deposition of sediments. Copyright © 2006 John Wiley & Sons, Ltd.     

Reconstruction of floodplain sedimentation rates: a combination of methods to optimize estimates

Reconstruction of overbank sedimentation rates over the past decades gives insight into floodplain dynamics, and thereby provides a basis for efficient and sustainable floodplain management. We compared the results of four independent reconstruction methods – optically stimulated luminescence (OSL) dating, caesium‐137 (¹³⁷Cs) dating, heavy metal analysis, and flood bed interpretation – applied at three embanked floodplain sites along lower Rhine River distributaries in the Netherlands. All methods indicate significant sedimentation rates on the floodplains, varying between 2–7 mm/a in the distal zones and 3–9 mm/a in the proximal zones. On a rapidly developing sand bar along a natural levee sedimentation rates of 9 to 25 mm/a were found. Except for some minor inconsistencies in ¹³⁷Cs dating results, all methods show decreasing sedimentation rates with increasing distance from the river channel. Intercomparison of the results of the different dating methods revealed the potential errors associated with each method, particularly where disagreement among the results were found. Uncertainties may arise due to (1) grain‐size dependent downward migration of ¹³⁷Cs, (2) smoothing of the vertical heavy metal and ¹³⁷Cs profiles, (3) delayed sediment‐associated input of ¹³⁷Cs in addition to direct atmospheric fall‐out, (4) overestimation of the burial age in OSL dating due to incomplete resetting of the OSL signal, or (5) non‐linear relationships between sediment deposition and flood magnitude in the count‐from‐the‐top correlation between sediment lamination and past observed flood records. Still, taking the uncertainties associated with each method into account, the results are generally in good agreement. Using the results we indicate the optimal spatial range of application of each method, depending on sediment texture and sedimentation rate. The optimal spatial and temporal ranges differ for each method, but show significant overlap. A combination of the methods will thus provide maximum information for accurate estimation of sedimentation rates on a decadal time scale. Copyright © 2010 John Wiley & Sons, Ltd.     

A preliminary investigation of the integration of modelled floodplain hydraulics with estimates of overbank floodplain sedimentation derived from Pb-210 and Cs-137 measurements

This paper investigates the possibility of linking finite element modelled two-dimensional hydraulics with floodplain sedimentation rates derived from radionuclide dating techniques. The TELEMAC-2D system is used to simulate the full dynamic velocity vector and water depth distributions for a 1 in 12 year flood event for the River Culm, Devon. These are compared to patterns of medium-term (30– 100 years) sedimentation rates and a degree of correlation is found to exist. This information is used to comment on the complexity of flow and sediment deposition interactions in floodplain environments and to provide a research design for the development of a more fully integrated floodplain morphodynamic model. Copyright © 1999 John Wiley & Sons, Ltd.     

Impact of natural reforestation on floodplain sedimentation in the Dragonja basin,SW Slovenia

Changes in floodplain sediment dynamics have profound effects on riverine habitats and riparian biodiversity. Depopulation due to socio‐economic changes in the Dragonja catchment (91 km²) in southwestern Slovenia resulted in the abandonment of agricultural fields, followed by natural reforestation since 1945. This profoundly changed the water and sediment supply to the streams, as well as floodplain sediment deposition. This paper presents a reconstruction of the development of the Dragonja floodplain due to these land use changes during the last 60 years. The reconstruction is based on dating of floodplain sediments using ¹³⁷Cs profiles, measurement of actual sedimentation rates using artificial grass sedimentation mats, and linking this information to the present‐day hydrological behaviour of the river. The sedimentation mats showed that floodplain sedimentation was restricted to peak flows of considerable magnitude. Due to the reforestation, the return period of such high flows increased from 0·31 year in the period 1960–1985 to 0·81 year between 1986 and 2003, with commensurate changes in sedimentation rates. At the 1·5 m river terrace (formed about 60 years ago), ¹³⁷Cs‐based sedimentation rates (1960–1986) were roughly twice the rates inferred from the artificial grass mats (2001–2003). This finding matches the increase in the return period for larger peak events during the 1986–2003 period, which caused fewer major inundations at this level. Conversely, sedimentation rates determined for the lowest terrace at 0·5 m were similar for both techniques (and periods) because the return periods of the peak events responsible for sediment deposition at this lower level did not change much over the period 1986–2003. Copyright © 2006 John Wiley & Sons, Ltd.     

Spatial variability of floodplain sedimentation at the event scale in the Rhine–Meuse delta,The Netherlands

This article addresses spatial variability of comtemporary floodplain sedimentation at the event scale. Measurements of overbank deposition were carried out using sediment traps on 11 floodplain sections along the rivers Waal and Meuse in The Netherlands during the high-magnitude flood of December 1993. During the flood, sand sheets were locally deposited behind a natural levee. At distances greater than 50 to 100 m from the river channel the deposits consisted mainly of silt- and clay-sized material. Observed patterns of deposition were related to floodplain topography and sediment transporting mechanisms. Though at several sites patterns were observed that suggest transport by turbulent diffusion, convection seems the dominant transporting mechanism, in particular in sections that are bordered by minor embankments. The average deposition of overbank fines ranged between 1·2 and 4·0 kg m⁻² along the river Waal, and between 1·0 and 2·0 kg m⁻² along the river Meuse. The estimated total accumulation of overbank fines (not including sand sheets) on the entire river Waal floodplain was 0·24 Mton, which is 19 per cent of the total suspended sediment load transported through the river Waal during the flood. © 1998 John Wiley & Sons, Ltd.     

Amazonian floodplain water balance based on modelling and analyses of hydrologic and electrical conductivity data

In the Amazon basin, floodplains form a complex mosaic of freshwater systems with differing morphologies, resulting in varied inundation patterns and heterogeneous chemical and ecological characteristics. In this study, we focused on the Janauacá floodplain, a medium‐sized system (786 km², including the local watershed) located along the Solimões River. Based on in situ and satellite observations acquired from November 2006 to November 2011, we computed water fluxes between the mainstream and the floodplain and examined the temporal dynamics of floodplain storage from river flooding, rainfall, runoff, and exchanges with groundwater through bank seepage for the 5 years from 2006 to 2011. The mainstream was the main input of water to the flooded area, accounting on average for 93% of total water inputs by the end of the water year. Direct precipitation and runoff from uplands contributed less than or equal to 5% and 10%, respectively. The seepage contribution was less than 1%. Model uncertainties, evaluated using Monte Carlo analysis of the input data and model parameters, showed that all water fluxes were relatively well constrained except for outflow through seepage, which had a standard deviation across simulations greater than 60%. The water balance computation was verified using electrical conductivity as an assumed non‐reactive tracer. Except during periods of very low water, the simulated and measured conductivities agreed well. Moreover, conductivity data analysis confirmed that the Janauacá system can be considered homogeneous in terms of electrical conductivity for filling percentages equal to or greater than 40% (i.e., when the water level is above 19.5 m, generally from April to August) but presented large heterogeneities during the rest of the hydrological cycle.     

Analysis of recharge conceptualization in inverse groundwater modelling

ABSTRACT

This study presents a systematic illustration quantifying how misleading the calibration results of a groundwater simulation model can be when recharge rates are considered as the model parameters to be estimated by inverse modelling. Three approaches to recharge estimation are compared: autocalibration (Model 1), the empirical return coefficient method (Model 2), and distributed hydrological modelling using the Soil and Water Assessment Tool, SWAT (Model 3). The methodology was applied in the Dehloran Plain, western Iran, using the MODFLOW modular flow simulator and the PEST method for autocalibration. The results indicate that, although Model 1 performed the best in simulating water levels at observation wells in the calibration stage, it did not perform satisfactorily in real future scenarios. Model 3, with SWAT-based recharge rates, performed better than the other models in the validation stage. By not evaluating the model performance solely on calibration results, we demonstrate the relative significance of using more accurate recharge estimates when calibrating groundwater simulation models.

EDITOR D. Koutsoyiannis; ASSOCIATE EDITOR M. Besbes     

The use of remotely sensed land cover to derive floodplain friction coefficients for flood inundation modelling

Remotely sensed land cover was used to generate spatially‐distributed friction coefficients for use in a two‐dimensional model of flood inundation. Such models are at the forefront of research into the prediction of river flooding. Standard practice, however, is to use single (static) friction coefficients on both the channel and floodplain, which are varied in a calibration procedure to provide a “best fit” to a known inundation extent. Spatially‐distributed friction provides a physically grounded estimate of friction that does not require fitting to a known inundation extent, but which can be fitted if desired. Remote sensing offers the opportunity to map these friction coefficients relatively straightforwardly and for low cost. Inundation was predicted using the LISFLOOD‐FP model for a reach on the River Nene, UK. Friction coefficients were produced from land cover predicted from Landsat TM imagery using both ML and fuzzy c‐means classifiction. The elevetion data used were from combined contour and differential global positioning system (GPS) elevation data. Predicted inundation using spatially‐distributed and static friction were compared. Spatially‐distributed friction had the greatest effect on the timing of flood inundation, but a small effect on predicted inundation extent. The results indicate that spatially‐distributed friction should be considered where the timing of initial flooding (e.g. for early warning) is important. Copyright © 2007 John Wiley & Sons, Ltd.     

Reconstructing the early 19th‐century Waal River by means of a 2D physics‐based numerical model

Suspended‐sediment concentration data are a missing link in reconstructions of the River Waal in the early 1800s. These reconstructions serve as a basis for assessing the long‐term effects of major interventions carried out between 1850 AD and the early 20th century. We used a 2D physics‐based morphodynamic model accounting for the influence of floodplain vegetation to fill in this gap. Historical discharge hydrographs were derived from a correlation between flow discharge records at Cologne and water level measurements of the Rhine branches in the Netherlands, taking into account the discharge distribution between the branches. Historical floodplain sedimentation rates were estimated using old cartographic information and recent geomorphologic field work. The computed historical sedimentation rates are found to be within the range of measured data, which suggests that fine suspended sediment concentrations in the early 1800s were comparable to contemporary ones. The computations show also how vegetation enhances the formation of natural levees close to the main channel and at the same time decreases the sedimentation rates in farther areas of the floodplain. A sensitivity analysis shows suspended sediment composition to have a strong influence on the resulting quantities and patterns of floodplain deposition. The reconstruction has also provided validation of the modelling tools to reproduce the effects of vegetation on sediment dynamics, enabling their implementation to study other cases. Copyright © 2010 John Wiley & Sons, Ltd.     

Experimental and numerical modelling of sedimentation in a rectangular shallow basin

Numerical simulation of flows in shallow reservoirs has to be checked for its consistency in predicting real flow conditions and sedimentation patterns. Typical flow patterns may exhibit flow separation at the inlet, accompanied by several recirculation and stagnation areas all over the reservoir surface. The aim of the present research project is to study the influence of the geometry of a reservoir on sediment transport and deposition numerically and experimentally, focusing on a prototype reservoir depth between 5 and 15 m as well as suspended sediment transport.
A series of numerical simulations is presented and compared with scaled laboratory experiments, with the objective of testing the sensitivity to different flow and sediment parameters and different turbulence closure schemes. Different scenarios are analyzed and a detailed comparison of preliminary laboratory tests and some selected simulations are presented.
The laboratory experiments show that suspended sediment transport and deposition are determined by the initial flow pattern and by the upstream and downstream boundary conditions. In the experiments, deposition in the rectangular basin systematically developed along the left bank, although inflow and outflow were positioned symmetrically along the centre of the basin. Three major horizontal eddies developed influencing the sediment deposition pattern. Although asymmetric flow patterns are privileged, a symmetric pattern can appear from time to time. This particular behaviour could also be reproduced by a two-dimensional depth-averaged flow and sediment transport model （CCHE2D）. The paper presents numerical simulations using different turbulence closure schemes （k-ε and eddy viscosity models）. In spite of the symmetric setup, these generally produced an asymmetric flow pattern that can easily switch sides depending on the assumptions made for the initial and boundary conditions. When using the laboratory experiment as a reference, the most reliable numerical results have been obtai     

Methods to calculate sedimentation rates of floodplain soils in the middle region of the Elbe River

This study presents different methods to quantify the historic and recent sedimentation of floodplain soils along the Elbe River. These methods include the comparison of surface elevations, the quantification of sedimentation with the aid of anthropogenic and geogenic tracers, sediment trap studies, and the calculation of load balances. Selected results from sites at the lower section of the middle Elbe River are presented and verified. The results show that several methods are suitable. In future work it should be possible, depending on the available soil and sediment data, to calculate sedimentation in Elbe River floodplain and the loss of retention volume for larger areas.     

Heavy metal transport in large river systems: heavy metal emissions and loads in the Rhine and Elbe river basins

Pollutant transport and management in the Rhine and Elbe basins is still of international concern, since certain target levels set by the international committees for protection of both rivers have not been reached. The analysis of the chain of emissions of point and diffuse sources to river loads will provide policy makers with a tool for effective management of river basins. The analysis of large river basins such as the Elbe and Rhine requires information on the spatial and temporal characteristics of both emissions and physical information of the entire river basin. In this paper, an analysis has been made of heavy metal emissions from various point and diffuse sources in the Rhine and Elbe drainage areas. Different point and diffuse pathways are considered in the model, such as inputs from industry, wastewater treatment plants, urban areas, erosion, groundwater, atmospheric deposition, tile drainage, and runoff. In most cases the measured heavy metal loads at monitoring stations are lower than the sum of the heavy metal emissions. This behaviour in large river systems can largely be explained by retention processes (e.g. sedimentation) and is dependent on the specific runoff of a catchment. Independent of the method used to estimate emissions, the source apportionment analysis of observed loads was used to determine the share of point and diffuse sources in the heavy metal load at a monitoring station by establishing a discharge dependency. The results from both the emission analysis and the source apportionment analysis of observed loads were compared and gave similar results. Between 51% (for Hg) and 74% (for Pb) of the total transport in the Elbe basin is supplied by inputs from diffuse sources. In the Rhine basin diffuse source inputs dominate the total transport and deliver more than 70% of the total transport. The diffuse hydrological pathways with the highest share are erosion and urban areas. Copyright © 2002 John Wiley & Sons, Ltd.     

Application of inverse geochemical modelling for predicting surface water chemistry in Ekbatan watershed,Hamedan, western Iran

ABSTRACT

A study of surface water chemistry evolution was conducted by multivariate statistical analysis and inverse geochemical modelling using the PHREEQC computer program. Using hierarchical cluster analysis the 14 sampling sites were classified into three groups (recharge, transition and discharge areas). Water chemistry changed along a flow path so that waters with Ca–HCO₃ and Mg–Cl composition changed to Mg–Cl–HCO₃ waters. The order of abundance of the major cations was Mg > Ca > Na > K. Their average concentrations were 21, 19, 3.6 and 2.5 mg L^-1, respectively. Inverse geochemical modelling along flow paths indicated that the dissolution of sylvite and kaolinite, and precipitation of feldspars and andalusite, happened with Na entering the solution and Ca, Mg and K leaving the solution.

Editor D. Koutsoyiannis; Associate editor not assigned     

Conditionality of inverse solutions to discretised integral equations in geoid modelling from local gravity data

The eigenvalue decomposition technique is used for analysis of conditionality of two alternative solutions for a determination of the geoid from local gravity data. The first solution is based on the standard two-step approach utilising the inverse of the Abel-Poisson integral equation (downward continuation) and consequently the Stokes/Hotine integration (gravity inversion). The second solution is based on a single integral that combines the downward continuation and the gravity inversion in one integral equation. Extreme eigenvalues and corresponding condition numbers of matrix operators are investigated to compare the stability of inverse problems of the above-mentioned computational models. To preserve a dominantly diagonal structure of the matrices for inverse solutions, the horizontal positions of the parameterised solution on the geoid and of data points are identical. The numerical experiments using real data reveal that the direct gravity inversion is numerically more stable than the downward continuation procedure in the two-step approach.     

Assessment of the quantitative and qualitative buffer function of an alluvial wetland: hydrological modelling of a large floodplain (Garonne River,France)

For 2 years, water flow‐patterns in the Garonne floodplain of south‐western France were studied in the field and through hydrodynamic modelling (MARTHE Hydrodynamic Software developed by BRGM). Water flow‐paths and the transport of dissolved elements between river and aquifer have been investigated and modelled. In order to quantify the buffer function of the alluvial floodplain, we focused our work on the effect of a major flood on the water flow‐direction, and on nitrate transport. Thus, we showed that the effect of a large flood in the river was rapidly lost with increasing distance from the river. During the observation period, a hydrologically active strip only 300 m wide on either side of the riverbed played a buffering role in absorbing the flood crest. It was also found that meanders favour the exchange between river and alluvial aquifer, shown by the creation of bypasses between the upstream and downstream parts of meanders. This, in turn, contributes to a dilution of nitrates in the phreatic aquifer, which here has higher nitrate content than the surface water; such dilution may result in an overestimation of the denitrification process in the wooded riverbanks. The coupling of chemical measurements—especially of chlorides and nitrate—with modelling of the dissolved‐element transport allows us to establish the water balance for the riparian wetland, and to separate the effect of dilution and denitrification on nitrate concentration. This indicated the existence of areas in the riparian wetlands where denitrification is particularly strong, leading to reductions in nitrate concentrations of 10 to 30 mg/l NO₃^? during the flood. Copyright © 2003 John Wiley & Sons, Ltd.     

Landscape transformation of an Alpine floodplain influenced by humans: historical analyses from aerial images

Natural floodplains are spatially heterogeneous and dynamic ecosystems but at the same time, a highly endangered landscape feature due to climate change and human impacts such as water storage, flood control and hydropower production. Flow is considered a master variable that shapes channel morphology and the heterogeneity, distribution, and turnover of floodplain habitats. Despite their highly dynamic nature, the relative abundance of different habitat elements (islands, gravel bars) in natural floodplains seems to remain relatively constant over ecological periods and is referred to as the shifting mosaic steady state concept. In this conceptual context, we analysed spatiotemporal changes in relative habitat abundance and channel complexity of an alpine floodplain from its near natural state in 1940 before water abstraction and levee construction until 2007 using historical aerial images. Within the first decades of impairment, the relative abundance of floodplain habitats that depend on flood and flow pulses such as parafluvial channels and islands shifted toward a greater abundance of terrestrial forest and grassland habitats. After 1986, the duration and frequencies of high‐precipitation events (>60 mm 24 h^–1) triggering major, channel‐reworking floods increased substantially and caused a restructuring of the floodplain and decrease in the abundance of more terrestrial habitat types. These results are contrary to expectations of the shifting mosaic steady state concept yet suggest its potential application as an indicator of landscape transformation and human impacts on floodplain ecosystems. Last, the results raise the applied question as to whether an increased frequency of high flow events induced by climate change can contribute to floodplain restoration. Copyright © 2011 John Wiley & Sons, Ltd.     

Temporal distribution of heavy metal concentrations in oysters Crassostrea rhizophorae from the central Venezuelan coast

The oyster Crassostrea rhizophorae is a bivalve abundant in Venezuelan estuaries and consumed by local populations. No known values have been reported on trace metals in oysters from the central Venezuelan coast. We report the concentrations of Al, Cd, Co, Cr, Cu, Fe, Mn, Ni, Pb, Sr, V and Zn in the soft parts of C. rhizophorae, which were collected bimonthly between March 2008 and March 2009, at two sampling areas from the Central Venezuelan Coast: Buche estuary and Mochima estuary. Our results show that for each metal there is a similar temporal variation pattern. The concentrations of the heavy metals reported in this work are useful as reliable baselines and can be used for comparison in future environment studies. Concentrations in C. rhizophorae from the Buche estuary can be interpreted to be high on a global scale for Cd, Cu, Ni and Mn, indicating atypically raised bioavailabilities.     

Hydrological runoff modelling by the use of remote sensing data with reference to the 1993–1994 and 1995 floods in the river Rhine catchment

In hydrological modelling of runoff processes, including water balance, various input data and parameters can be acquired or estimated by the use of remote sensing (RS) techniques.The acquisition and use of synoptic RS areal information rather than traditional point information is an important issue in hydrology. Hydrological models allow runoff/water balance in catchments to be calculated and flow routing within flow channels to be done. For runoff and water balance computations land use, soil moisture, detection of snow and ice, digital terrain models (DTM), as well as hydrometeorological information and discharge are important. For flow routing, water level information, geometric–topographic information such as cross-sections for normal and flood conditions, coefficient of roughness and velocity of flow and its cross-sectional distribution are required. In addition, water level information (lower and upper level) is needed for shipping and for design purposes. In the German part of the River Rhine catchment, several focus areas in the December 1993–January 1994 and January 1995 floods were covered with RS data [ERS-1 and airborne SAR, both C-band VV, passive microwave (18·7, 36·5, 89 GHz), TIR, UV, aerial photographs (b/w PAN, b/w NIR)], giving a good opportunity for a comparison of methods. Evaluation is still continuing. The importance of soil saturation for flood generation and, therefore, for flood monitoring, was shown on this occasion. The use of ERS SAR data for soil moisture estimation is currently being investigated by the Federal Institute of Hydrology. Also, the need for emergency schemes for data acquisition and easy, quick and affordable RS data dissemination was demonstrated. The assimilation of RS data with GIS information such as DTMs, including relevant topographic features like dams, which is omitted in currently available raster digital elevation models, is promising. RS altimetry techniques can be a step towards high resolution DTMs for hydrological purposes. Ground truth reference data are still needed. © 1997 John Wiley & Sons, Ltd.     

A Linking Test that investigates the feasibility of inverse modelling: application to a simple rainfall interception model for Mt Gambier,southeast South Australia

Interception loss has an important influence on the water yield of forested areas. Nevertheless, in most studies stemflow is not measured, therefore the question of how to determine the feasibility of optimizing interception and stemflow parameters simultaneously by matching daily simulated throughfall to fortnightly measurements of cumulative throughfall is an important one. By applying a daily empirical interception model, a goodness fit of 2·2 mm/day is obtained between observed and simulated cumulative throughfall. However, by applying the simple but robust Linking Test, it was shown that the parameters are non‐unique and falsely linked, i.e. inter‐relationships between different vegetation parameter sets give similar throughfall but non‐unique net precipitation. The Linking Test investigates the causes of obtaining falsely linked parameters and shows that objective equifinality is not the source of the problem. Objective equifinality occurs when an inappropriate objective function is used. The Linking Test also shows that falsely linked parameters are not caused by measuring throughfall on a non‐daily basis (termed frequency sampling equifinality). By expanding the interception model to the second degree, it was found that the non‐uniqueness is due to the inherent nature of interception and stemflow functions that behave similarly and therefore can easily compensate each other (termed similarity equifinality). It is also shown that a simple daily empirical exponential interception model developed for conifers in the uplands of the United Kingdom is suitable to model interception in Pinus radiata plantations in the Mediterranean climate of southern Australia by using only daily gross precipitation data and two parameters. Copyright © 2009 John Wiley & Sons, Ltd.