Environmental Risk Imposed by Diverted Flood Waters on Water and Soils in Emergency Retention Basins

Emergency retention basins (ERB) are diked enclosures alongside rivers into which water from the main river channel is diverted during extreme floods. If the basins are operated during extreme flooding, two negative environmental impacts may occur: 1) contamination of the soils due to their transport by suspended sediments to the basin and 2) depletion of dissolved oxygen in the basin water. A computer-based methodology is presented which was used to assess the environmental risk exhibited by the operation of an ERB system proposed for the Elbe River in Germany. The August 2002 extreme flood event was used as a test case. For such a flood, the results showed that there is a 77% risk that dissolved oxygen levels fall below 2 mg/L in the water and a 48% chance of exceeding the inspection value of 500 mg zinc/kg in the soil.


Introduction
The European Union (EU) has passed a Floods Directive to provide a legislative foundation for flood protection and mitigation in Europe.Since most large river basins in Europe extend over several countries, the directive will help to orient and broaden flood management efforts beyond the municipal and state authorities to the national and international levels and establish an improved and more cost-effective flood protection scheme (see also [1]).Interestingly, the directive includes many references to the environment and the EU Water Framework Directive, e.g.: Article 4 (Preliminary flood risk assessment): …[must consider] potential adverse consequences for human health, the environment  Article 9 (Flood hazard maps and flood risk maps): … shall be coordinated with and may be integrated into the reviews provided for 2000/60/EC [WFD]  Article 7 (Flood risk management plans): … shall take into account costs and benefits, flood extent and conveyance routes, environmental objectives of 2000/60/EC [WFD], soil and water management Hence, it will be mandatory to consider what impact new flood defence measures will have on the environment.Also, these measures are not to conflict with the goals of the EU WFD of achieving a good ecological status of water bodies.
After the extreme flood event along the Elbe River in August 2002, efforts were made to revamp flood management schemes in the river basin.Hence, the construction of an emergency retention basin (ERB) system has been proposed for the middle reaches of the Elbe River to reduce flood risk in the area [2].
Emergency retention basins are diked enclosures alongside rivers used to retain flood water by diverting and storing a portion of the river discharge during floods.The diversion from the main river channel to and from the ERBs is controlled in order to attain maximum capping of the peak discharge volume.The floodwater diversion reduces water levels downstream in the river and inundated areas and alleviates stress and damages in regions prone to high flood risk.
Studies on environmental risk assessment on retention basins by flood events are sparse in the literature.Therefore, the aim of the paper was to provide a possible computer modelling methodology on how such assessments can be tackled in fulfilling the proposed EU directives.Further, a preliminary assessment of the environment risk of heavy metal soil contamination and dissolved oxygen depletion in ERBs should be given.To solve these goals, excerpts have been drawn from [3][4][5][6].
In detail, the following steps were in the focus of the research:  develop and test a quasi-2D (two-dimension) model in which the equations of motion and continuity are calculated in 1D (one-dimension) but the discretisation scheme allows for a 2D representation of flow and substance distribution; this minimises computing and pre-processing expenditure without compensating spatial differentiation of flood behaviour,  test how effective the proposed ERB system is able to cap the peak discharge of the hydrograph from the August 2002 flood,  investigate the environmental risk of heavy-metal contamination of ERB soils after flood water retention, and  assess the environmental risk of dissolved oxygen depletion in the water retained in the basins.

Study Area
Two ERBs were investigated in this studie, P1 and P3 (see Figure 1).

Modeling Tool and Setup
The hydrodynamic module DYNHYD (based on St. Venant equations), the sediment and contaminant transport module TOXI and the water-quality module EUTRO from the WASP5 (Water quality Analysis Simulation Program v.5) modelling package, developed by the U.S. Environmental Protection Agency [7] was used for the floodwater simulations.Figure 2 (left panel) shows the discretisation of the river-ERB system domain in junctions and channels.The junctions ensure volume continuity, hence all junction water volumes represent the total water volume in the system.The channels are the basis for computing the flow in 1D between the junctions.Many channels may branch from a junction allowing the discretisation to take a 2D characterisation.The segment discretisation for the contaminant transport and water-quality model, shown in Figure 2 (right panel), corresponds 1-to-1 the junction discretisation.
In TOXI, processes that describe the transport and fate of heavy-metal contaminants include:  longitudinal dispersion in the water column,  vertical diffusion of dissolved substances between the water column and the bottom sediments,  sedimentation of suspended solids,  re-suspension of solids from the bottom sediments to the water column, and  sorption of dissolved substances to suspended and deposited sediments.EUTRO was used to simulate water quality pertaining to the oxygen balance in a river and ERBs using the cycles of dissolved oxygen (DO) decomposition and nutrient-limited phytoplankton growth.Oxygen is also consumed in the sediments, which is described in the model by the sediment oxygen demand.An important source of oxygen into the water body is re-aeration via the water surface from the atmosphere.
An important source and sink of DO are phytoplankton photosynthesis and respiration, respectively.Phytoplankton growth is also light limited and its loss rate is governed by respiration, death, settling and zooplankton grazing.
The DYNHYD, EUTRO and TOXI modules were structured in a simulation platform HLA (High Level Architecture) to allow Monte-Carlo analyses (MOCA) to be carried out [8].The MOCA allowed uncertainty bounds to be computed in order to assess the degree of risk in exceeding environmental thresholds.For the environmental risk appraisal the following were investigated: 1) the amount of zinc, with high contamination potential along the Elbe [9], sorbed and deposited from the diverted flood waters and 2) the minimum concentration of DO attained in the basin waters.

Results
For a comprehensive description of the development, calibration and validation of the models with corresponding sensitivity analyses of the parameters and boundary conditions, the reader is referred to [6] for the hydrodynamic modelling, [4] for contaminant transport modelling and [5] for the eutrophication modelling.Only the key results to support the environmental assessments will be presented here.

Soil Contamination
In order to make an environmental risk assessment of contamination in the ERBs, the deposition of pollutants onto the ERB surfaces must be placed within the legal context of the German sewage sludge ordinance.An example is given here using zinc as the heavy-metal contaminant.The ordinance provides threshold values of zinc content allowed in the soils and sediment ("sludge") deposited onto the land surfaces within the basins (summarised in Table 2).
In accordance to the German Sewage Sludge Ordinance, the background zinc concentrations in the soils, where the ERBs are to be situated, must be determined.The geogenic value of 150 mg zinc/kg sediment [10] for the Elbe catchment is used here for illustration.This value does not exceed the maximum allowable zinc content in the soils of 200 mg/kg.Hence, sediment may be deposited from the river water diverted into the ERBs during extreme flooding.The content of zinc within the deposited sediment may not exceed 2500 mg/kg, if the land surfaces are to continue to be used for agricultural purposes.Table 3 summarises the results from the MOCA showing that the probability of reaching this value is minute (< 1%).A different approach based on standards provided by the German Soil Protection and Contaminated Sites Ordinance is given in [5].
An important question that arises is how many floodings of the ERB can occur before the concentrations in the soil increase from the geogenic level of 150 mg/kg to the threshold value of 200 mg/kg.Figure 4 (right panel) shows a spatial distribution of zinc bound to deposited sediment, which corresponds to an 8% exceedence probability extracted from the MOCA.The zinc content in the soil increases by 5 mg/kg.Hence, approximately ten ERB floodings may cause zinc accumulation to reach the threshold value of 200 mg/kg.
Although the environmental risk for zinc contamination is low, it should be noted that an amendment to the sewage sludge ordinance is currently being discussed in which the threshold values for allowed zinc content in soils and deposited sediment ("sludge") are lower (see Table 2).Careful measurements of the background zinc levels in the soils of ERB sites should be carried out to determine how close these values are to the maximum allowable content in soils.
The environmental risk analyses presented here shows that there is minimal risk of plant contamination by zinc deposited onto the ERBs during floods.However, the risk may be substantially increased if the ERBs are located further downstream.Figure 6 shows the zinc content in deposited sediment along the portion of the Elbe River in Germany.Imbedded in the graph are the results  obtained from the MOCA, the range of which complies with the range of all measurements taken upstream from the Mulde River inflow.For the river stretch downstream from the Mulde and Saale tributaries, the range and the maximum of zinc content in deposited sediments is twice as high as for the upstream reach.

Dissolved Oxygen Depletion
From the MOCA runs, probability distributions of the lowest DO levels attained in the ERBs during the simulation time frame were derived.The cumulative frequency distributions for three locations in the ERB system are shown in Figure 7.The threshold value of the minimum concentration of DO in inland waters used for a risk assessment is 3 mg O 2 /L-the minimum oxygen content with which most fish in the Elbe River can endure for a short term [12].As soon as the oxygen content sinks below approximately 3 mg/L, the fish begin to search for water pockets with better oxygen supply or begin to snap for air at the water surface [13].
For a summer time extreme flood such as the August 2002 event along the Elbe, the probability of the DO values dropping below a threshold value of 3 mg O 2 /L is quite high, 77% for the inlet region of P1, 65% for the far eastern point of the P1 extremity and 54% for P3 (see Figure 7).In general, the DO concentrations are higher more frequently in P3 than in P1 due to the shorter residence times of the water in this ERB and allowing less time for de-oxygenation of the water.The generally higher DO content in P3 is also due to its shallower depth, making it a more favourable aquatic environment for phytoplankton growth.The low DO values are also a result of low re-aeration (wind calm period) and high nitrification.
In contrast to experiences gained from the Havel River during the August 2002 flood, the low oxygenated water emptying into the River Elbe does not appear to have negative implications to the oxygen levels in the river water.Although the water of the Elbe River was under the DO saturation level, DO was still above the threshold value of 3 mg/l at Wittenberg (see Figure 8).
However, if these ERBs were situated further downstream where the DO concentrations of the river water fall below the threshold value, the emptied basin water could reduce the DO concentrations further and further stress the aquatic ecosystem.

Conclusions
The quasi-2D approach was successful in simulating the spatial distribution of flow and substance deposition in ERBs and determining the effectiveness of discharge capping.The environmental risk to heavy-metal contamination is minimal but may increase due to stricter threshold values.Risk may also be higher for ERBs constructed further downstream past the Saale and Mulde river confluences where exposure to contamination is higher.There is a high potential for DO levels in the water retained in the basins to drop to very critical levels (< 3 mg O 2 /L), even after 4 days of flood water retention.From an ecological perspective, the basins should also be emptied as soon as possible after the flood in order to reduce the time of de-oxygenation of the water.

Figure 1 .
Figure 1.Investigated emergency retention basins (ERB) proposed for the Elbe River reach between Torgau and Wittenberg, modified from [2]

Figure 2 .Figure 3 .
Figure 2. Discretisation for the hydrodynamic model with junctions and channels (left panel) and the water-quality model with segments (right panel), modified from [6] and [5]

Figure 3 (Figure 4 (Figure 4 .Figure 5 .Figure 5 (
Figure 3 (left panel) shows the water level hydrographs of the August 2002 flood simulated along the river at Positions A, B and C (see Figure 2 for spatial reference).The dashed line of the hydrograph crests represent water level heights that occurred without the ERB system, the solid line represents the lowering of the water levels that can be obtained when the ERBs are included in the modelling.The capping lies in the range of 35 to 40 cm and progressively recedes as the flood wave moves down-

Figure 6 .
Figure 6.Longitudinal profile of zinc loading on soils along the Elbe River measured in 2003 after the extreme flood in the previous year, modified from [11], p. 242

Figure 7 .
Figure 7. Cumulative frequency distributions of the lowest dissolved oxygen concentrations attained in the ERBs