STORM Report Series No. 7
Simulation Tool for River Management
STORM - Rhine
Main Report
Design of the Roleplay
March 2002
STORM Report Series No. 7 Simulation Tool for River Management
STORM - Rhine
Main report
Design of the Roleplay
March 2002
The STORM project is sponsored by:
IRMA-SPONGE: Roleplay for transboundary river management (Project number: 3/NL/1/164/ 99 15 183 01)
Delft Cluster: Ontwikkeling rollenspelen integraal waterbeheer (Project number: 06.01.06)
The STORM projects are carried out in partnership by:
International Institute for Infrastructure, Hydraulics and Environment P.O.Box 3015; 2601 DA Delft; The Netherlands Contact: J.C.Heun, +31-15-2151835, [email protected] Contact: T.D.Schotanus, +31-15-2151853, [email protected]
Resource Analysis Zuiderstraat 110; 2611 SJ Delft; The Netherlands Contact: M.M.de Groen, +31-15-2191526, [email protected]
WL | Delft Hydraulics P.O.Box 177; 2600 MH Delft; The Netherlands Contact: M. Werner, +31-15-2858807; [email protected] STORM Report Series:
1. STORM – BETUWE, Final Report, Design of the Roleplay Schmidt, A.L., October 1998
2. STORM – BETUWE, User Manual, Design of the Roleplay Schmidt, A.L., October 1998
3. STORM – DELTA, Definitiestudie: demonstratiemodel voor Rivierinrichting Schmidt, A.L., December 1998
4. STORM – DELTA, Main Report, Design of the Model De Graaff, B.J., December 1999
5. STORM – DELTA Main Report, User Manual De Graaff, B.J., December 1999
6. STORM – RHINE, Main Report, Executive Summary Heun, J.C., De Groen, M.M., Werner, M., March 2002
7. STORM – RHINE, Main Report, Design of the Roleplay Heun, J.C., De Groen, M.M., Werner, M., Schotanus, T.D.; March 2002
7A. STORM – RHINE, Main Report, Annex A, Documentation on Models on Water Werner, M., March 2002
8. STORM – RHINE, Main Report, User Manual Schotanus, T.D., Versteeg, A., July 2002
9. STORM – RHINE Supporting Document 1: Dutch-German Water Management of the River Rhine, MSc. thesis Bijman, S., January 2002
10. STORM – RHINE Supporting Document 2: Deciding about Safety; Flood Protection & Decision-Making Processes in Germany at the local level; M.Sc.thesis Leene, T.A., August 2002
IRMA project number: 3/NL/1/164/ 99 15 183 01
The IRMA-SPONGE Umbrella Program
In recent years, several developments have contributed not only to an increased public interest in flood risk management issues, but also to a greater awareness of the need for improved knowledge supporting flood risk management. Important factors are: • Recent flooding events and the subsequently developed national action plans. • Socio-economic developments such as the increasing urbanisation of flood-prone areas. • Increased awareness of ecological and socio-economic effects of measures along rivers. • Increased likelihood of future changes in flood risks due to land use and climate changes.
The study leading to this report aimed to fill one of the identified knowledge gaps with respect to flood risk management, and was therefore incorporated in the IRMA-SPONGE Umbrella Program. This program is financed partly by the European INTERREG Rhine-Meuse Activities (IRMA), and managed by the Netherlands Centre for River Studies (NCR). It is the largest and most comprehensive effort of its kind in Europe, bringing together more than 30 European scientific and management organisations in 13 scientific projects researching a wide range of flood risk management issues along the Rivers Rhine and Meuse.
The main aim of IRMA-SPONGE is defined as: “The development of methodologies and tools to assess the impact of flood risk reduction measures and scenarios. This to support the spatial planning process in establishing alternative strategies for an optimal realisation of the hydraulic, economical and ecological functions of the Rhine and Meuse River Basins." A further important objective is to promote transboundary co-operation in flood risk management. Specific fields of interest are: • Flood risk assessment. • Efficiency of flood risk reduction measures. • Sustainable flood risk management. • Public participation in flood management issues.
More detailed information on the IRMA-SPONGE Umbrella Program can be found on our website: www.irma-sponge.org.
We would like to thank the authors of this report for their contribution to the program, and sincerely hope that the information presented here will help the reader to contribute to further developments in sustainable flood risk management.
Ad van Os and Aljosja Hooijer (NCR Secretary and IRMA-SPONGE project manager)
Report Title: STORM-Rhine Main Report – Design of the Roleplay
Contact person : J.C. Heun Date : March 2002 J.C. Heun (IHE) M.M. de Groen (Resource Analysis) Author(s) : M. Werner (Delft Hydraulics) T.D. Schotanus (IHE)
Development of roleplays integrated water DC Project name : management DC Project number : 06.01.06 DC Working Group Theme : Integrated water management DC Theme leaders : E. van Beek, H.H.G. Savenije Report Number : STORM report series no. 7 Number of pages : 64 Number of tables : 32 Number of figures : 21
Keverling Buismanweg 4 Postbus 69 2600 AB Delft
015-2693793 015-2693799
[email protected] www.delftcluster.nl
Delft Cluster verricht lange-termijn fundamenteel strategisch onderzoek op het gebied van duurzame inrichting van deltagebieden. STORM-Rhine, Main Report - Design of the Roleplay Page i
Table of Contents
1 THE STORM PROJECTS 1 1.1 Objectives of the Projects 1 1.2 Objectives of STORM 1 1.3 The STORM Products 1 1.4 Making Use of STORM 2 1.5 Roleplay Setting 2 1.6 STORM Projects Implementation 3
2 THE REPORT 5
3 SCOPE OF STORM-RHINE 7 3.1 The Concept 7 3.2 Scale of the Simulation 7 3.3 River and Floodplain Functions 8 3.4 River and Floodplain management 9 3.5 The Models 9 3.6 The Roles 10 3.7 The Interface 10
4 RIVER MANAGEMENT 11 4.1 Introduction 11 4.2 Interventions in the Main Channel 11 4.3 Interventions in the Floodplain 12 4.4 Summary 13
5 MODELS ON HYDROLOGY 15 5.1 Introduction 15 5.2 Hydrology of the Rhine basin 15 5.3 Extent of the Rhine Basin considered 16 5.4 Modelling Approach 18 5.5 Hydrological Computations 19 5.6 Hydraulic Computations 20 5.7 Morphological Computations 22 5.8 Measures 22 5.9 Hydrological Routing Computations 25 5.10 Hydraulic Computations Upstream 27 5.11 Scenarios 28
6 MODEL ON SHIPPING 31 6.1 Introduction 31 6.2 Objective 31 6.3 Assumptions 32 6.4 Input 33 6.5 Model 34 6.5.1 Relation between water depth and ship load 34 6.5.2 Frequency distribution of normative water depths 36 6.5.3 Frequency distribution of ship loads 38 6.6 Output 42 6.6.1 Introduction 42 6.6.2 Economic effects 42 6.6.3 Environmental effects 44 6.7 Discussion and synthesis 45 6.8 Information on Role 46
7 PRESENT and TARGET LAND COVER 47 7.1 Introduction 47 7.2 General 47 7.3 The Netherlands 47
STORM-Rhine, Main Report - Design of the Roleplay Page ii
7.4 North Rhine Westphalia 48 7.5 Rhineland Palatinate and hessen 49 7.6 Germany Target Distribution 50 7.7 Recommendations 51 7.8 Test of matrix approach 52
8 FLOOD DAMAGE MODULE 55
9 INSTITUTIONS 57 9.1 Water Management in Germany 57 9.2 Water Management in The Netherlands 58 9.3 Institutional Complexity of Inter-departmental Interaction 59 9.4 Comparison between German and Dutch Water Management 59 9.5 Application in the Roleplay 60
REFERENCES 63
ANNEX A MODELS on WATER
LIST of TABLES
Table 3.1 River and floodplain functions 8 Table 3.2 River and floodplain management measures 9 Table 3.3 Modules of the STORM-Rhine system model 9 Table 5.1 Cumulative chainage and catchment area of the Rhine ... 16 Table 5.2 Physical maximum and highest measured discharge 17 Table 5.3 Tributaries explicitly modelled 18 Table 5.4 Overview of branches considered in the model 18 Table 5.5 Geometrical properties of schematised cross sections 21 Table 5.6 Measures acting on geometry and hydraulic roughness 22 Table 5.7 Projected retention areas between Maxau and Lobith 23 Table 5.8 Projected retention downstream of Lobith 24 Table 5.9 Retention basins considered in the STORM-Rhine 24 Table 5.10 Nodes and branches of the routing model 25 Table 5.11 Comparison of observed and routed maximum discharges 27 Table 5.12 Overview of branches and approximate locations 27 Table 5.13 Estimates of climate change for the Rhine basin 29 Table 5.14 Discharges at different tributaries for selected scenarios 30 Table 5.15 Peak discharges at different locations for selected scenarios 30 Table 6.1 Representative sizes of ship classes 33 Table 6.2 Input information for changes in fleet 33 Table 6.3 Ship masses 35 Table 6.4 Number of journeys from ships from different ship classes ...... 43 Table 7.1 Translation of GIS map of NRW to ecotope distribution in STORM 48 Table 7.2 Translation of Gewässerstrukturgütekartiering of Rhineland Paptinate 50 Table 7.3 Matrix to transform current ecotope distribution to target ecotope 51 Table 8.1 Unit flood damages 55 Table 8.2 Areas and land use classes of damage 55 Table 9.1 Water resources management framework in Germany 57 Table 9.2 Water resources management framework in The Netherlands 58 Table 9.3 Water quantity management 59 Table 9.4 River, floodplains and facilities management 60 Table 9.5 Roles in STORM-Rhine 60
LIST of FIGURES
Figure 3.1 Elements of a simulation game 7 Figure 3.2 Linkages between players 7
STORM-Rhine, Main Report - Design of the Roleplay Page iii
Figure 3.3 Reach covered by STORM-Rhine 7 Figure 3.4 Relation between modules and system models 9 Figure 3.5 Compilation of elements from STORM-Rhine interfaces 10 Figure 4.1 Structural measures in the river and floodplain 11 Figure 5.1 The Rhine basin 16 Figure 5.2 Relation between the computational modules 19 Figure 5.3 Example of effect on discharge peaks 20 Figure 5.4 Illustration of link between hydrologic and hydraulic computations 21 Figure 5.5 Schematic cross section used in hydraulic computations 22 Figure 5.6 Structure of the hydrological routing model 25 Figure 5.7 Comparison between Muskingum routing model end SOBEK 26 Figure 6.1 Model results (Eq. 5.2) for different ship classes and empirical relation for 4 barges convoy sets, derived by AVV (2000) 35 Figure 6.2 Derivation of relation between water depth and cumulative probability of this water depth, from the Qh-relation in Lobith and the cumulative probability of certain discharges in Lobith 36 Figure 6.3 Probability distribution of the existing normative water depth (Eq. 3.2) and of the normative water depth that is 0.5 m larger 37 Figure 6.4 The probability that different ship classes can transport maximum loads, plotted as a function of these maximum loads 40 Figure 6.5 Probability density functions of non-maximum loads for different ship classes and a simplified model that uses an estimate of the load which is exceeded with 99% of the time and an estimate of probability that the ship can be fully loaded 40 Figure 6.6 Comparison of analytical and discrete way to determine average load 41 Figure 6.7 Average number of tons transported, under the assumptions that (1) always the maximum possible load is transported and (2) the demand of transport is equally distributed over the year and (3) the loads cannot be buffered 41 Figure 9.1 Denomination of river stretches 62
STORM-Rhine, Main Report – Design of the Roleplay Page 1
1 THE STORM PROJECTS
1.2 Objectives of the projects
The objective of the STORM (Simulation Tool for River Management).projects is to develop simulation games for river management. STORM-Rhine specifically focuses on trans-boundary river and floodplain management, with as case study the River Rhine.
1.3 Objectives of STORM
Considerations Large river basins are complex management systems because of − the highly interrelated hydraulic, morphological and biological processes − the interactions between human activities and the physical system − the management measures may mutually reinforce or counteract each other − the multitude of policy makers and stakeholders with potentially conflicting interests. Simulation games can educate on the issues, provide a uniform frame of reference, and be a catalyst in cross-sectoral and multi-disciplinary dialogues. Simulation games can also explore how stakeholders may react in case of new scenarios developing.
The River Rhine The River Rhine is an interesting case in point, where within less than a decade a complete new approach to river and floodplain management has been introduced with concepts as Room for the River, Flood Retention Areas, Resurrection of former River Channels and Living with the Floods, while at the same time the importance of Nature is widely accepted. Although there is a fair level of consensus amongst experts and politicians, there is still a great deal of understanding and awareness raising to be explored, while the implementation of policy measures will highlight the real conflicts still to come. Besides that there are the new scenarios with higher design floods, and hence flood risks, caused by climate change and changing land use.
Objectives of STORM The simulation game STORM intends to provide a platform for improving insight in river and floodplain management. It sets out: − provide insight in river and floodplain management, showing the links between natural processes, spatial planning, engineering interventions, river functions and stakeholder interests − formulate and analyse alternative management strategies − facilitate debate between different levels of stakeholders − explore the international dimension of river management − highlight upstream – downstream effects on river basin scale − explore river basin scale scenarios affecting river management − create awareness for the Room for the River concepts − raise understanding of river regulation measures; decision-making mechanisms; floodplain use affecting river functions; river regulation affects flood plain use; the functioning of retention areas
1.4 The STORM Products
In the course of the projects three simulation tools were developed, STORM-Betuwe, STORM- Delta and STORM-Rhine, each with its own purpose and characteristics.
The concept The players represent the institutions and stakeholder interests. They have mandates to take measures to pursue their interest in specific river and floodplain functions. The measures may be river engineering oriented, but also land use, economic or administrative interventions are possible. The measures are analysed by a system model and translated into effects, which describe the functioning of the river and floodplain with respect to the interests of the players. The players are linked by a regulatory framework of licensing and budgeting. Scenarios determine under which conditions the game is played for example an increased demand for navigation, a specific design flood or a predicted climate change.
STORM-Rhine, Main Report – Design of the Roleplay Page 2
STORM – BETUWE The STORM-Betuwe simulation game describes a typical stretch of some 60 km length of the Dutch River Rhine. The stretch is divided in eight sections in which the players can define the measures, both for the left and right side of the river. The roles represent four municipalities, each responsible for part of the river stretch and each with its own focus on economic development, nature and agriculture. In addition, the provincial and national government are represented. Because of the limited stretch of the river simulated, upstream-downstream affects do not appear. STORM-Betuwe facilitates most of the simulation game objectives described above. it was especially developed to create awareness for the Integral Reconnaissance Rhine Branches and Room for the River Projects in The Netherlands. Most of the river management measures as given in this report do apply, except that retention areas and green rivers cannot be created.
STORM – DELTA STORM-Delta describes the Dutch Rhine Delta of the rivers IJsel, Lek and Waal from the border with Germany to the boundary of the salt intrusion. All river management measures and criteria listed in the report are available, including the “Green River”, flood retention areas (calamity polders) and redistribution of flows over the three river branches. However, the institutional framework has not been modelled. Consequently, there is no role-playing facility: one user may enter all measures and has access to all output. STORM- Delta is an interactive tool to provide insight in river behaviour, to create awareness for the Room for the River concepts and to test policies to enhance river (floodplain) functions for diverse stakeholder interests.
STORM – RHINE STORM – Rhine is a full-fledged simulation game, which describes the river from the downstream boundary of the salt intrusion in The Netherlands to Maxau in Germany, which is the site of the first weir in the river. It covers a total length of about 650 km, which is divided in 17 sections in which all the typical modern river and floodplain management measures can be taken. A total of eight roles represent the national, regional and local government and specific interest groups. STORM-Rhine incorporates the functionality of both STORM-Betuwe and STORM-Delta. The scope of STORM-Rhine is briefly described in Chapter 3.
1.5 Making use of STORM
Experience shows that simulation games can be used in very different settings, focusing on one or more of the objectives above, depending upon the participants and the time available. STORM is considered particularly suitable for the following settings: − education (particularly to gain insight in complicated multi-disciplinary issues) − training middle and higher level staff of local and regional government, water boards and members of interest groups from across the basin, who deal with particular stretches or functions of the river, but who need (1) to be better aware of the integrated whole, (2) to understand the interests and considerations of others and (3) to experience the mutual benefits of co-operation − the same target group as above + policy makers and experts to explore the consequences of for example (1) different visions on development and policy targets and (2) different natural conditions, such as due to climate change and catchment characteristics − at miscellaneous meetings and conferences for providing a uniform frame of reference or simply “breaking the ice”.
A special challenge would be to use STORM in the following, be it quite different, purposes: − to discover and explore new ways of river management − as a tool to support interactive policy formulation and participatory decision-making in actual plans.
1.6 Roleplay Setting
Each role is preferably played by a small group of participants together. The need for interaction between the various roles will prove to be essential in order to produce an integrated river management approach resulting in performance criteria acceptable to most if not all stakeholders and decision makers
STORM-Rhine, Main Report – Design of the Roleplay Page 3
The simulation game can be ‘played’ with a computer network, whereby each role can directly enter the desired measures it is mandated to enact. However, the respective roles responsible for issuing permits or binding advice, need to have done so on their own computers, before the measure will be executed. Players may communicate via the network with e-mails. The model (on a central database server) provides feedback on the state of the river in the form of tables, graphs and maps showing the performance criteria and state of the river. Based on an individual or communal evaluation, players may decide to take additional measures.
It is also possible to engage in the roleplay with one stand-alone computer which runs the model. The role representatives are then provided with hardcopy input sheets, while a ‘game- leader’ looks after processing and feedback with printed output.
1.7 STORM Projects Implementation
The research project was carried out from 1997-2001 by the following partners: − IHE; International Institute for Infrastructural, Hydraulic and Environmental Engineering; Delft, The Netherlands − RA: Resource Analysis; Delft, The Netherlands − WL | DHL; WL | Delft Hydraulics Laboratory; Delft, The Netherlands
Different phases of the research project were financed by − European Union under the IRMA-Sponge Programme, − Land Water Impuls (LWI) (ICES I), − Delft Cluster (ICES II), − RIZA − the partner institutes. IHE, RA and WL | DHL
STORM-Rhine, Main Report – Design of the Roleplay Page 4
STORM-Rhine, Main Report – Design of the Roleplay Page 5
2 THE REPORT
2.1 This report is technical in nature and its purpose is to describe the design of the main elements of the simulation tool for STORM-Rhine.
Chapter 3 briefly describes the scope of STORM-Rhine.
Chapters 4 to 9 and Annex A describe in detail the system models, which are used in the roleplay: • the model on hydrology • the model on hydraulics and morphology • the model on land use in the floodplain • the model on shipping • the model on damages in case of floods • the roles representing the institutions.
The report is complemented by a STORM-Rhine Roleplay Manual, both for the players and for the facilitators.
STORM-Rhine, Main Report – Design of the Roleplay Page 6
STORM-Rhine, Main Report – Design of the Roleplay Page 7
3 SCOPE OF STORM-RHINE
3.1 The Concept
The main elements of the simulation game are depicted in Figures 3.1 and 3.2 below.
The players represent the institutions Institutions, River River Rhine River Rhine and stakeholder interests. They have Stakeholders Management Model Floodplain mandates to take measures to pursue Interests Strategy Functions their interest in specific river and floodplain functions. The measures Measure Effects may be river engineering oriented, but Player Player System also land use, economic or Player Model administrative interventions are Measure possible. They may reinforce each Effects other or may conflict with each other The measures are analysed by a Evaluation system model and translated into effects, which describe the functioning Figure 3.1 – Elements of a Simulation Game of the river and floodplain with respect to the interests of the players. The game may be played in different Financial rounds representing distinct time periods. Institutional The players are linked by a regulatory framework of licensing and budgeting. Scenarios Scenarios determine under which Budget Budget conditions the game is played for example an increased demand for Objective Actions Actions Objective navigation, a specific design flood or a System model predicted climate change. Criteria Criteria
3.2 Scale of the Simulation Figure 3.2 – Linkages between Players The simulation model of STORM-Rhine is restricted to the area where floodplains are of importance, as depicted in Figure 3.3. The upstream boundary is at Maxau, just upstream of Mannheim and just downstream of the regulated Upper Rhine. The downstream boundary is the limit of the salt intrusion in the three Dutch river branches: Waal, Lek and IJsel. The model is restricted to the main river corridor, with the tributaries incorporated as scenario variables.
The reason for this choice is that it is this reach of the Rhine where the main issues of flood protection, nature, agriculture and navigation are dominant. Directly competing interests exist, especially with respect to land use in the floodplains and between shipping and nature development. Different, interdependent management interventions are available to influence the river functions. These issues and interventions occur at a scale and level, which are recognisable by the intended players of the simulation (see also Box 1, next page) Figure 3.3 – Reach covered by STORM-Rhine
All basin management issues outside this floodplain area are incorporated in the simulation game as scenarios influencing boundary conditions: both for the Rhine proper as for the major tributaries.
STORM-Rhine, Main Report – Design of the Roleplay Page 8
Design Conditions for Simulation Games The design of the simulation game should fulfil a number of conditions, in order to make it interesting to play, and achieve the objectives described above: • Realistic and recognisable: players should be able to recognise their situation and their own decision level. The complexity of river management is that interactions affect different stakes at very different scales. A rather visionary policy game at river basin scale is often not interesting to decision makers at a regional or local scale playing a policy game at a river basin scale if they do not recognise their own decision level. On the other hand, a regional scale policy game raises the awareness of decision makers at a river basin scale of their role in regional decision-making, provided that their role is represented in the simulation. A policy game at a more regional scale with representation of the river basin scale will create a platform for cross-level discussions. • Credible: the system model should report effects, which in their order of magnitude are subscribed to by the “experts”. • Conflict and dependability: competing interests and river functions, which are served by alternative river management strategies have to exist. Box 1
3.3 River and Floodplain Functions
The river and floodplain functions identify the interests of the stakeholders. Criteria indicate the performance of the function. Setting the functions and criteria determines the type of system models, which have to be made in order to be able to give feedback to the players of the roleplay. The functions and criteria represented in STORM-Rhine are listed in Table 3.1.
Table 3.1 River and floodplain functions and criteria represented in STORM-Rhine Function Performance criteria Indicator Flood protection − maximum water level increase m Discharge of Water − unsafe river stretch km − estimated flood damage €/y − cost of all related measures € − cost of flood damages €/y Shipping − change in average turn over per ship/year €/y − additional benefits industrial sector €/y − change in employment in shipping sector persons − change in employment road transport sector persons − energy use % % − CO2 emission, acid rain (NOX + SO2 ) and smog − investments in new ships € € − costs of depth increase Nature − surface area nature ha − ecological chances - − fulfilling of stepping stone function % − presence of indicator species # − costs of acquiring and developing nature area € − cost of nature management €/y Agriculture − (acquired and total) area of agriculture ha − yield ton/yr − cost of acquisition and development € Landscape and − degree of openness % culture − agricultural character landscape % − managed natural character % − spontaneous natural character % Construction − amount of mined clay, sand, gravel m3 3 Materials − amount of mined polluted soil m − cost of cleaning polluted soil € Recreation − land related recreation, water related recreation ha Building − building area in floodplain %
STORM-Rhine, Main Report – Design of the Roleplay Page 9
3.4 River and Floodplain Management
The typical river and floodplain management interventions are described in Chapter 4. Table 3.2 lists the measures presented. The mandates of the different roles to enact management are described in Chapter 9. Not all management interventions can take place in every stretch of the river. The restrictions are also listed in Chapter 9.
Table 3.2 River and floodplain management measures represented in STORM-Rhine River Training Vegetation Management − Construct stone or natural river bank − Spontaneous succession − Change length or height of groynes − Extensive natural grazing − Change height of summer dike − Meadow land management − Change height of winter dike − Wetland management − Change distance of winter dike from river − Forest development − Change depth of river, dredging − Agricultural pasture management − Construct / remove side channel − Agricultural crop management Physical Planning Administrative − Construct retention areas − issue floodplain activity permit − Lower high terrain, excavate sand, clay − issue local land use permit − Remove earthen bridge ramps − issue excavation permit − Build in floodplain (residential, industrial) − change composition shipping fleet − Construct recreational harbour, camp sites − Construct “green river “(outside flood-plain)
3.5 The Models
The models together forming the heart of the simulation tool are shown below in Figure 3.4 and listed in Table 3.3. The models are described in more detail in subsequent chapters.
SCENARIO: Interventions: river training, vegetation management, physical planning
hydrology
shipping Geometry River network ecotope predictor excavation
nature
landscape morphology hydraulics roughnes Land-use costs
mining
agriculture shipping floods agriculture landscape recreation nature
recreation
cost Performance Criteria
Figure 3.4 – Relation between modules of the system models
Table 3.3 Modules of the STORM-Rhine system model Type Water-related Land-use related Other Modules Hydrological routing module Ecotype predictor Excavation module Hydraulic module Nature module Cost module Morphological module Agriculture module Shipping module Recreation module Hydraulic roughness module Landscape module
STORM-Rhine, Main Report – Design of the Roleplay Page 10
3.6 The Roles
The institutions are represented in eight roles as follows:
1 The national government Upstream (Germany) 2 The regional government for the Left River Bank Upstream(Germany) 3 The regional government for the Right River Bank Upstream (Germany) 4 The local government Urban Municipalities Upstream (Germany) 5 The local government Rural Districts Upstream (Germany) 6 The national government Downstream (The Netherlands 7 The local government Rural & Urban (The Netherlands) 8 An international environmental lobby group
The mandates of the different roles are described in Chapter 9.
3.7 The Interface
The interface of the STORM-Rhine model facilitates the selection of interventions and scenarios and gives output in the form of tables and maps showing the performance criteria. The interface also provides the facilities for communication between players so as to account for licensing of interventions. An compilation of the elements of the interface is given in Figure 3.5.
Figure 3.5 - Compilation of elements from STORM-Rhine interfaces
STORM-Rhine, Main Report – Design of the Roleplay Page 11
4 RIVER MANAGEMENT
4.1 Introduction
River Management of the River Rhine is a complex process in which the diverse stakeholders must be considered. These stakeholders include clear economic users of the river such as navigation, drinking water, recreation, but also important issues such as environment and safety. The current situation of the river is by no means a given constant but changes constantly due to for example climatic change, changes in perception of society as well as autonomous changes such as morphological aggregation and degradation (Silva et al, 2001). Perhaps the most apparent change threatening the river is that of the projected climatic change, where it is foreseen that the regime will move away from the snow-melt fed regime to a rain-fed regime, with higher peaks and lower troughs in the discharges as a result (Middelkoop, 1999). One instrument that can be employed to counter these influences and maintain usability of the river is to intervene, either in the geometry, land use or discharge regime.
Any intervention in the river system will affect the river response in terms of water level, discharges, morphology and sediment transport characteristics, and can directly and indirectly change those characteristics, since water movement, sediment transport and river morphology is strongly interrelated. For example, the construction of groynes to ensure sufficient depth for navigation initially leads to higher water levels. This would intuitively lead to the need for dredging. The higher velocities derived from the restriction of the channel, by progressively eroding the riverbed, leads to the ultimate effect of a decrease of the water levels. As a consequence channel dredging may no longer be required. The construction of groynes changes also the channel roughness so that the prediction of water levels should take this into account too. Other examples of river engineering works are channel modifications, the construction of embankments, new or modified infrastructure (flood storage, weirs, barriers, sluices, bridges etc.) and, recently, river restoration and flood plain reforestation. By increasing channel roughness and reducing the cross-section for the water flow, those interventions may have drastic consequences for the water levels during floods.
From these examples it is clear that the consequences of each intervention are very complex, and where the use of the river for one stakeholder may improve, others stakeholders may suffer. In this chapter some of the options available for intervening in the river are discussed. The interventions, or measures, are divided into three main categories, where only options for interventions in the river system itself are considered, as interventions in the catchment to change runoff to the river are beyond the scope of this project.
Figure 4.1 shows the typical river and floodplain management interventions.
Figure 4.1 – Structural measures in the river and flood-plain
4.2 Interventions in the main channel
Interventions in the main channel include measures such as dredging of the main channel, deposition of sediments, construction/amendment of groynes. The main channel is the main conveyer of flow in the river, and flow conditions in the main channel are of particular significance to navigation. The impact of measures in the main channel on flood levels is
STORM-Rhine, Main Report – Design of the Roleplay Page 12
typically felt for a limited reach upstream rather than downstream of the measure, due to the limited change in storage.
Dredging Dredging in the main channel is typically employed to keep the river navigable throughout its entire length, but due to the natural degradation of the river channel in most of the reaches, this measure is generally not well received (Silva et al, 2001). It is clear that the impact on dredging on high flows is also positive, as main channel conveyance is increased. Although dredging is required in some reaches to maintain navigability, application of this approach across long reaches may have undesirable side effects on groundwater table levels along the river, salt intrusion in the delta and long term morphology.
Groynes Construction of groynes along the main channel has been widely employed along the lower parts of the Rhine, most notably the Lower Rhine and the branches of the Rhine Delta. Groynes serve to concentrate the flow in the main channel, and through this help maintain the depth in the main channel and prevent the development of sandbanks that could hinder navigation. Groynes are of great importance to ensuring navigation, particularly during low flows. Construction of groynes may again be considered if the projected climatic change does indeed cause lower flows in the river during the summer. During higher flows groynes have a negative effect, as the conveyance of the main channel is reduced. From the point of view of flood protection construction of groynes is therefore not desirable, and even removing or shortening existing groynes is considered. An alternative to removal of existing groynes, and the negative effects this will have on shipping and stability of the main channel is to lower existing groynes. This option does need to be considered carefully in terms of the impacts on morphology (Silva et al, 2001).
Main channel banks Although the influence on the flow in the main channel may be limited, main channel banks may be either fixed as stone banks or left in a natural state. Stone riverbanks are found along various reaches of the river, particularly in the Upper and Middle Rhine. Obviously natural riverbanks are environmentally appealing, but these do cause the roughness to increase and thus levels during high flow will increase.
4.3 Interventions in the floodplain
A wide variety of interventions may be taken in the floodplain and the number of users and usage’s of the floodplain are much more diverse than the main channel. Clearly the floodplain has limited influence on water levels at low-flow conditions, but for high flows it is of significant importance. Floodplains are used in a variety of ways, with environmental and agricultural being dominant, but also for recreational and even industrial and urban purposes. Broadly speaking, interventions in the floodplain can be categorised into interventions that involve excavation of the floodplains, floodplain landscape planning or the removal of hydraulic bottlenecks. Depending on to what degree the storage in the floodplain is affected, the measures described will predominantly influence flood levels upstream of the location of the intervention. If significant extra storage is created, the levels downstream may also be influenced through increased attenuation of the flood wave.
Floodplain excavation Flood plain excavation is a measure by which the gradual development of heightening by sedimentation on flood plains may be counteracted. This heightening is generally caused by the construction of dikes, ‘normalisation’ of the river and by the construction of submersible embankments.
To influence the hydraulic characteristics of flow significantly, floodplain excavation interventions must be quite rigorous and the consequences on current uses of the floodplain are far reaching. Floodplain excavation may, however, be combined with nature development or even with clay mining, as was recently applied within the framework of the most recent dike reinforcements in the Netherlands. One drawback of floodplain excavation is that some of the soils in the floodplain may be significantly polluted, making the processing of excavated soils difficult and expensive. Once excavation is completed, the excavated floodplain can obviously
STORM-Rhine, Main Report – Design of the Roleplay Page 13
not be left as bare soil, and some form of land-use management should be considered in conjunction with the excavation. Development of nature areas is one example. Construction or maintenance of side-channels, either those connected to the main channel or isolated side channel (e.g. isolated oxbow lakes), can be seen as a special form of floodplain excavation. This intervention is well suited to increasing conveyance of the river during high flows, while promoting the ecological diversity of the floodplain.
Floodplain land-use Land use in the floodplain is an important issue. Encroachment of urban areas on floodplains has not only significantly reduced the extent of floodplains, but urban and industrial development in the floodplain has led to significant damages from flooding. For a large part, particularly in the Rhine branches in the Netherlands, the floodplains are used as agricultural land, often for grazing. Changing perceptions with respect to the environment is leading more and more to a drive for more natural floodplains, with the development of forested floodplains and wetlands very much in public interest. Development of floodplains in this way goes well with interventions such as floodplain excavation, but the combined effect on for example water levels during flood must be studied carefully when these types of interventions are employed.
Removal of bottleneck and submersible (summer) embankments The removal of a bottleneck is a very local intervention. Typical hydraulic bottlenecks are bridge abutments, ferry ramps or flood-free areas in the floodplain. Other examples are sharp bends in the river or very narrow floodplains. Removal of these may be very effective in lowering upstream flood levels, while being economically attractive when compared to measures such as floodplain excavation. Depending on the type of bottleneck, the impacts on other floodplain users may, moreover, be limited. Submersible embankments have been constructed along large reaches of the main channel of the Rhine, and keep the floodplains free from flooding for all except high floods. Removal of these will have significant effects on the frequency of floodplain inundation, as well as possible effects on low flow conditions in the main channel. Increasing the frequency of floodplain inundation may allow for a more natural development of floodplain vegetation, but for other users the effects may less desirable.
Retention basins A very effective method for reducing flood levels downstream is to retard water in the upstream part of the river. Retarding water in the catchments before it reaches the main river may be effective but is not considered here. Large volumes of water may, however, be retarded in off- line retention basins. When used most effectively these can be employed to as it where “shave off” the peak of the flood, thus lowering the flood peak downstream. Retention basins can be constructed either in the floodplain through the construction of a ring dyke, or out of the floodplain (Lammersen et al, 1999). The first category is usually somewhat limited in size and thus capacity to reduce the flood peak. Larger retention areas may be created beyond the floodplain, but this does lay a significant claim on land, with the use of land adjacent to rivers already being under much pressure. Retention areas are controlled through inlet structures, either as a fixed crest weir or an inlet structure that is operated. The water stored in the retention area is allowed to flow back into the river once the peak has passed. Control of the retention area, or the level at which the retention area starts operation is an important factor in determining the effectiveness of the area. If the area is put to use to early then it may be full before the peak arrives, and the effect on the peak may then be negligible. If it is too late then there will be no effect on the peak. Another problem is the location of the basin with respect to the dominant area of flood genesis. Clearly a retention basin in the Upper Rhine will have no effect on a flood generated in the catchment areas of the major tributaries flowing into the Middle Rhine.
4.4 Summary
From this brief overview of available options for intervening in the river it is clear that careful consideration the impact of measures on the broad scale of users of the river must be given. Although the objective of each intervention may be clear, effects are often multi-facetted. The do-nothing options may also be seen as a form of intervention, particularly in terms of doing nothing with respect to the evolution of floodplain vegetation. This aspect referred to, as cyclic floodplain rejuvenation is one topic currently being researched. Interventions of each category
STORM-Rhine, Main Report – Design of the Roleplay Page 14
are also usually not taken in isolation but rather as in the form of extensive plans with interaction of all users of the river as well as local, regional, national and international authorities.
STORM-Rhine, Main Report – Design of the Roleplay Page 15
5 MODELS on HYDROLOGY
5.1 Introduction
The roleplay for transboundary river management considers the Rhine River basin both in Germany and in the Netherlands. Central to the game is the assessment of possible river related measures, where these measures are targeted at one or more of the uses of the river system. Examples of these uses are navigation, safety etc.
Assessment of the effects of measures is accomplished by evaluating these in a hydrological/hydraulic module, where the impacts of measures on different flow conditions are determined. These changed flow conditions are then applied in determining impacts for different users such as the navigation authorities, flood protection authorities etc.
In this chapter the approach taken in modelling the hydrological and hydraulic processes is described. This brief description focuses on the approach, without going into mathematical detail of the concepts applied. These details are given in the applicable sections of Annex A.
A general description of the hydrology of the Rhine River basin is given, with particular focus on that part of the basin downstream of Maxau, being the upper limit of the basin considered. The approach taken in modelling the hydrology, hydraulics and morphology of the basin is introduced, as well as how measures are incorporated. In the last section, the structure of the models as applied in this project are discussed, with particular focus on the determining of boundary conditions for different scenarios.
5.2 Hydrology of the Rhine Basin
The River Rhine has its source in the Alps, and stretches some 1300 km to its mouth in the North Sea. The catchment area is about 185,000 km2. The hydrological regime of the basin is a combination of rainfall and snowmelt. Engel (1997) identifies six major stretches by considering the geo-morphological characteristics. The Alpine Rhine and the High Rhine in the Alps are characterised by high mountains and extensive glaciers. Downstream of Rheinfelden/Basel, the Upper Rhine starts. Here the Rhine flows through a lowland plain at a much smaller gradient. The natural river once was of a braided nature, but in recent centuries elaborate river training works have been undertaken to improve navigability. These river training works are concentrated mainly in the upstream half of the Upper Rhine (above Maxau, Figure 5.1). Below Mainz the river cuts through the Rhenish Slate Mountains, meandering through a relatively narrow gorge. This section is known as the Middle Rhine. Downstream of Bonn, the river once again changes in nature to become the Lower Rhine, a typical lowland river with wide meanders and a relatively wide and flat floodplain. Just after the Dutch/German border the Rhine Delta starts as the river bifurcates into three main branches, Waal, IJsel and the Dutch Rhine (Nederrijn-Lek), before flowing into the North Sea. At the downstream end of these three branches, the tidal influence of the North Sea becomes noticeable. In a transition zone, flood hazard is influenced both by upstream hydrology and downstream tidal conditions. Beyond this transition zone flood hazard is dominated by the tidal influences.
The hydrology of the High and Alpine Rhine is governed by snow-melt, with maximum discharges occurring in the summer months. These may be amplified by intensive, local summer rainfall (Middelkoop, 1999). The major tributaries of the Rhine are the Aare, Neckar, Main and Mosel. While the first is distinctive of the High and Alpine Rhine, the other three are typical rainfall dominated rivers, flowing into the lower half of the Upper Rhine and the Middle Rhine. These rain-fed tributaries contribute significantly to the genesis of the large floods in the Rhine, and most major floods have been caused by extreme discharges in these three tributaries, as well as other predominantly rain-fed tributaries in the Upper, Middle and Lower Rhine (Engel, 1997). These major floods occur predominantly in late autumn or early winter during extensive frontal storms, usually preceded by long spells of wet weather.
The morphology of the Rhine downstream of the extensively regulated section of the Upper Rhine can broadly be divided into two main sections. In the Upper Rhine, the lower Rhine and the Rhine Delta, the river is alluvial, and sedimentation and scour can be considered to act freely, given the hydraulic conditions. In the Middle Rhine, the hard rock of the Rhenish Slate
STORM-Rhine, Main Report – Design of the Roleplay Page 16
Mountains poses clear constraints on the scour of the river bed, and a lower limit must be taken into account. However, in these reaches sedimentation may occur if hydraulic conditions are favourable, and these layers may subsequently erode again.
Figure 5.1 The Rhine basin (from Silva and Dijkman, 2000)
5.3 Extent of Rhine Basin considered
The models for the Rhine basin considered here do not cover the entire Rhine basin. The upstream boundary of the area considered is the gauging station at Maxau (Figure 5.1). Moreover, only the main river is considered, with tributaries dealt with as hydrological inputs. Table 5.1 gives an overview of the river chainage of various stations on the Rhine between Maxau and Lobith, including the locations of major tributaries.
Table 5.1 Cumulative chainage and catchment area of the Rhine between Maxau and Lobtih (adapted from Diermanse et al, 2000 Station Tributary Section Chainage Trib-area Area (km) (km2) (km2) Maxau Upper 362.30 50,343 Speyer Upper 400.60 53,235 Mannheim Upper 424.90 54,136 Neckar 428.20 14,000 68,486 Worms Upper 443.40 68,936 Oppenheim-Nierstein Upper 480.60 70,462 Main 496.60 27,200 98,488 Mainz Upper 498.30 98,488 Bingen Upper/Middle 528.40 99,277 Nahe 530.00 4,100 103,407 Kaub Middle 546.20 103,729 Boppard Middle 570.45 103,981 Lahn 588.00 5,900 109,994 Koblenz Middle 591.50 110,131 Mosel 592.00 28,100 138,231
STORM-Rhine, Main Report – Design of the Roleplay Page 17
Station Tributary Section Chainage Trib-area Area (km) (km2) (km2) Nette&Wied 610.00 1,100 139,586 Andernach Middle 613.80 139,795 Ahr 629.00 850 140,837 Bad Honnef Middle/Lower 641.00 - Bonn Lower 654.70 141,162 Sieg 660.00 2,900 144,217 Koln Lower 688.00 144,612 Wupper 702.00 800 145,618 Erft 738.00 1,800 147,948 Dusseldorf Lower 744.20 148,040 Ruhr 779.00 4,500 153,143 Ruhrort Lower 780.80 153,176 Wesel Lower 814.00 154,528 Lippe 815.00 4,900 159,428 Rees Lower 837.40 159,683 Emmerich Lower 851.90 159,784 Lobith Lower/Dutch Rhine 862.22 160,800 Pannerden Dutch Rhine/Waal 867.20 - Tiel Waal 915.00 - Gorinchem Waal 955.00 - Arnhem Dutch Rhine/ IJssel 878.70 - Amerongen Dutch Rhine 918.00 - Nieuwegein Dutch Rhine 950.00 - Deventer Ijssel 935.00 - Ijselmeer Ijssel 993.00 -
For the sake of simplicity, not all the tributaries given in Table 5.1 are individually modelled. Obviously three of the four main contributors in the genesis of floods (Engel, 1997) that are downstream of Maxau are explicitly considered. Table 5.2 shows estimates of the maximum discharges measured for the various tributaries, and those with estimated maxima in the order of 1000 m3/s or higher are also considered explicitly here. Smaller tributaries are added to the closest larger one such that the volumetric balance is maintained while simplifying the geographic distributions somewhat.
Downstream of the Dutch/German border the river bifurcates into three main branches. These three branches are considered as they were in the original Role play model for the Delta of the Rhine. The downstream boundary of these branches is located such that it is upstream of any tidal influence, with flood hazard therefore being dominated by hydrological conditions upstream. Using the locations of the stations and tributary inflows given in Table 5.1, this led to the definition of 17 main section in the model of the Rhine. Division of the sections discerned by Engel (1999) was driven by the locations of tributary inflows, administrative boundaries (both national and international) as well as other important towns/locations.
Table 5.2 Physical maximum and highest measured discharge (period 1880 - 1995) at various gauging stations on the Rhine and its tributaries (Silva and Dijkman, 2000) River Estimated Highest Gauging station Date Physical Max. measured (m3/s) discharge (m3/s) Oberrhein (Worms, 6300 5600 Worms 550117 incl. Neckar) Main 3000 1980 Frankfurt 950130 Nahe 1200 1150 Grolstein 931221 Lahn 950 840 Kalkofen 460210 Mosel 4800 4170 Cochem 931221 Nette und Wied 400 202 Nettegut, Friedrichshal 840207 Ahr 300 194 Reimerzhoven 840530 Sieg 1600 1053 Menden 840207 Wupper 300 181 Opladen 570923 Erft 800 550 Bliesheim 840531 Ruhr 2300 907 Hattingen 940101 Emscher 430 200 (Estimated) -- Lippe 1200 370 Schermberg 950131 total 22860 16902
STORM-Rhine, Main Report – Design of the Roleplay Page 18
Table 5.3 Tributaries explicitly modelled Tributary Added Tributaries Neckar Main Nahe Lahn Mosel Nette & Wied Sieg Ahr Ruhr Wupper, Emscher and Erft Lippe
Table 5.4 Overview of branches considered in the model Branch Name Start End Start Km End Km Length (approx.) (approx,) (km) U1 Upper Rhine 1 Maxau Mannheim 360 425 65 U2 Upper Rhine 2 Mannheim Mainz 425 497 72 U3 Upper Rhine 3 Mainz Bingen 497 530 33 M1 Middle Rhine 1 Bingen Koblenz 530 592 62 M2 Middle Rhine 2 Koblenz Bad Honnef 592 642 50 M3 Middle Rhine 3 Bad Honnef Köln 642 688 46 L1 Lower Rhine 1 Köln Düsseldorf 688 744 56 L2 Lower Rhine 2 Düsseldorf Wesel 744 813 69 L3 Lower Rhine 3 Wesel Lobith 813 862 49 R1 Dutch Rhine 1 Lobith Pannerden 862 867 5 R2 Dutch Rhine 2 Pannerden Arnhem 867 878 11 R3 Dutch Rhine 3 Arnhem Amerongen 878 918 40 R4 Dutch Rhine 4 Amerongen Nieuwegein 918 950 32 W1 Waal 1 Pannerden Tiel 867 915 48 W2 Waal 2 Tiel Gorinchem 915 955 40 Y1 Ijssel 1 Arnhem Deventer 878 935 57 Y2 Ijssel 2 Deventer IJsselmeer 935 993 58
5.4 Modelling Approach
In the context of the role-play, an attempt is made to keep approach taken in modelling the hydrology, hydraulics and morphology of the river as simple as possible. This section is designed as a brief introduction to the approach, while more detail on the computational methods and the underlying assumptions is given in Annex A. The modelling approach is one of three steps, 1. Hydrological computations for determining discharges in the river system, 2. Hydraulic computations for determining water levels, and, 3. Morphological computations for determining variations in river bed levels.
The relation between these three steps is shown schematically in Figure 5.2. There is a feedback loop between the hydraulic and morphological computations, due to changes in geometry determined in the latter, while the former determines velocities used by the latter. Predicting morphological changes in a first step and subsequently determining the final hydraulic conditions in a second step, using the predicted morphological changes account for this feedback loop. Further iteration is not required under the assumption that differences in the predicted morphological change due to the revised hydraulic conditions of the second step are negligible.
Although full hydrodynamic models are available for integrated calculation of the three processes considered, it was decided that using these would unnecessarily complicate the modelling approach, and lay too much emphasis on these models within the gaming environment. Obviously results derived from the more complex approach could potentially be more accurate, but given the scale at which the different roles interact with the system, it was considered that the level of detail would perhaps be too high. Indeed, in a gaming environment such as that provided here, emphasis should be on general tendencies, rather than minute detail.
STORM-Rhine, Main Report – Design of the Roleplay Page 19
Hydrological Scenarios
Hydrological computations
Discharges
Hydraulic Water computations Levels
River Geometry Velocities
Morphological computations
Figure 5.2 Relation between the computational modules
5.5 Hydrological Computations
In the original role-play model on which the development of this extended model is based, hydrological conditions were relatively simple. The upstream boundary of the model was at Lobith on the Dutch/German border. From a hydrological perspective this single source of flow is easy to deal with. For the Rhine in Germany the hydrology is, however, geographically distributed. Particularly for major flood events, the tributaries of the Upper, Middle and Lower Rhine have a significant influence, and the impacts of floods will be geographically distributed. This entails also that effectiveness of measures designed to alleviate floods will depend very much on where these are taken in comparison to where the flood is generated. As floods may be generated in the geographically distributed tributaries, the occurrence of extreme floods in the main river can be significantly influenced by how well the peaks of these tributaries and in the main river coincide.
To allow for the distributed nature of the tributaries, as well as attenuation of flood peaks, a hydrological routing scheme is used. The well known Muskingum routing method is applied (Fread, 1993). Using this approach the flood wave from the uppermost catchment is routed to the confluence with the first tributary, there the flood wave from this tributary is added, and the combined wave is routed to the next confluence. This is repeated until the downstream end of the model is reached.
The method of routing flood waves through the river network and adding discharges at confluences is applicable as long as the river network is dentritic. Once the river bifurcates, hydraulic conditions at the bifurcation will determine how the flood waves will divide. This would introduce a feedback loop (Figure 5.2) between the hydrological and hydraulic computations. In the Rhine below the bifurcation there are no major tributaries and attenuation effects are limited. To keep the interaction between the hydrological and hydraulic computations simple, the lower boundary of the hydrological computations is thus set between the Lower Rhine and the Rhine Delta on the Dutch/German border at the Lobith gauging station.
An important contribution to the generation of extreme flood events in a dentritic river network as is considered here, may be due to flood peaks from different tributaries coinciding. If peaks coincide, then the resulting floodwave downstream may have a significantly higher peak than if the two contributing waves had not coincided. To allow for "worst case" scenarios to be
STORM-Rhine, Main Report – Design of the Roleplay Page 20
explored the hydrological routine allows for matching peaks of flood waves at the nodes where these are summed.
In Figure 5.3 an example is shown on the effect on hydrographs of peaks coinciding. The first inflow is routed, and the routed hydrograph clearly shows attenuation and a shift of the peak. At this point a tributary inflow is added. A second summed hydrograph may be obtained by synchronising the peak of the routed hydrograph with the peak of the tributary hydrograph.
7000
Q inflow,1 Q 6000 routed Q inflow,2 Q routed,lagged Q +Q routed inflow,2 5000 Q +Q routed,lagged inflow,2
4000
3000
2000
1000
0 50 100 150 200 250 300 350
Figure 5.3 Example of effect on discharge peaks due to peaks of contributing hydrographs coinciding
Generation of flood waves, through the runoff of rainfall or snow-melt is not considered explicitly in the hydrological computations. These deal solely with the routing of flood waves through the river network. As only the main river is considered, flood waves from the tributaries are imposed at the confluence in the main river.
5.6 Hydraulic computations
In the original role-play model, on which the development of this extended model is based, the hydraulic and morphological computations were made based on a steady state approach. It is assumed that the time derivatives of flow conditions are very small, as well as for the flow being sub-critical (low Froude numbers). This allows for significant simplification of the flow equations (Vreugdenhill, 1973). The approach computes the head difference over a reach based on the geometrical properties of that reach. As the method is steady state, no storage or attenuation of discharges is taken into account.
Starting from the downstream end of each branch, the head difference over the whole branch is determined as the sum of the head differences over the sub reaches. Bifurcation of the Rhine into the Waal, Dutch Rhine and IJssel does complicate the solution procedure, as the exact division of discharges over the branches is not known a-priori. This is dealt through the introduction of computational meshes across which the head differences are known (Vreugdenhill, 1973). See Annex A for a full description.
This approach for calculating water levels in the river network is also adopted for the extended model. Discharges to be applied to each reach are sampled from the hydrological model, where the maximum level in the flood wave determined in the hydrological module is taken. These
STORM-Rhine, Main Report – Design of the Roleplay Page 21
discharges are kept constant between each of the sections of the hydrological model. This will give some overestimation at the downstream end of the hydraulic reach due to attenuation, but given the limited length of the reaches in the hydrological model when compared to attenuation effects the overestimation is considered not to be too large. Figure 5.4 shows the principle of sampling the peak discharge of the hydrograph in the hydrological model and imposing these discharges on the hydraulic model. Discharges at intermediate nodes in the hydraulic model are allocated the same discharge as the closest upstream hydrological node. The lengths of the sub-sections in the hydraulic model are typically in the order of 10 km.
Retention Tributary Tributary Basin
Hydrological computation reach/node
Hydraulic computation sub-section/node Sample Qpeak
Figure 5.4 Schematic illustration of link between hydrological and hydraulic computations.
The geometry of the cross section used in the hydraulic computations is a schematised representation when compared to the actual geometry. Figure 5.5 shows this schematic cross section, with the symbols explained in Table 5.5.
Table 5.5 Geometrical properties of schematised cross sections. Description Symbol Levels Bottom level main channel: zm Left floodplain level (horizontal section): zlfp1 Left floodplain level (end of inclined section): zlfp2 Right floodplain level (horizontal section): zrfp1 Right floodplain level (end of inclined section): zrfp2 Left winter-dyke level: zwdl Right winter-dyke level: zwdr Left summer-dyke level: zlsd Right summer-dyke: zrsd Left groin level: zgl Right groin level: zgr Widths Width main channel: bm Width left floodplain (horizontal section): blfp1 Width left floodplain (inclined section): blfp2 Width right floodplain (horizontal section): brfp1 Width right floodplain (inclined section): brfp2 Length left groin: Lgl Length right groin: Lgr
STORM-Rhine, Main Report – Design of the Roleplay Page 22
blfp2 blfp1 bm brfp1 brfp2 Lgl Lgr
zwdl zwdr
zlfp2 zlfp1 zlsd zrsd zrfp zwdr zgl zm zgr
Figure 5.5 Schematic cross section used in hydraulic computations
5.7 Morphological computations
Morphological computations are kept very simple. Sedimentation and erosion are determined using the average velocity in a cross section. This average velocity is compared to a threshold velocity. If the threshold is exceeded the bed level will decrease through erosion while if it is not exceeded the bed level rises through sedimentation. To take the hard rock base level of the Middle Rhine into account, a minimum scour level is assigned. Sedimentation on this hard-base level can occur, and these layers can again erode until this fixed level is again reached.
5.8 Measures
Measures that can be taken to change the flow properties of the river are broadly divided into four categories; retention measures, geometrical measures, measures that have impact on the hydraulic roughness and finally side channels and green rivers.
Geometrical measures and measures influencing hydraulic roughness Two categories of geometrical measures can be given; measures acting on the main channel and measures acting on the floodplain. Measures acting on the geometry of the cross sections are applied to the schematised cross section (Figure 5.5). Measures influencing the vegetation result in a change in the distribution and type of floodplain vegetation, and based on this new distribution amended floodplain roughness coefficients are calculated. These measures may in principle be taken at any location along the entire river system, excepting locations where geometrical properties make application of a given measure irrelevant (e.g. removal of groynes will have no effect for river stretches where no groynes are present). Other restrictions may also apply for specific reaches. In the Middle Rhine, dredging of the main channel is limited to the layer of hard rock. An overview of the measures supported is given in Table 5.6.
Table 5.6 Measures acting on geometry and hydraulic roughness Measure type Measure Acting on River training construction of stone river bank main channel roughness construction of natural river bank main channel roughness increasing length of groins main channel geometry dredging main channel geometry deposition of sediment main channel geometry removal of summer-dykes floodplain geometry raising of summer-dykes floodplain geometry construction of side-channel (see next paragraphs) lowering of floodplain floodplain geometry inland moving of winter-dyke floodplain geometry raising of winter-dyke floodplain geometry construction of Green River (see next paragraphs)
STORM-Rhine, Main Report – Design of the Roleplay Page 23
Measure type Measure Acting on Vegetation spontaneous succession floodplain roughness management extensive natural grazing floodplain roughness meadow land management floodplain roughness swamp management floodplain roughness forest development floodplain roughness agriculture: pasture floodplain roughness agriculture: crops floodplain roughness Physical planning construction of retention area (see next paragraphs) construction of recreational harbour floodplain roughness construction of campsite floodplain roughness lowering of high terrain lowering of high terrain removal of earthen bridge ramps removal of earthen bridge ramps building in floodplain floodplain roughness
Retention measures Retention basins are off-line reservoirs that can be employed to “shave” off the peak of a flood wave. This is achieved by diverting a portion of discharges in excess of a threshold discharge at which the basin starts operation to a low-lying area adjacent to the river (Silva et al, 2000). The volume thus diverted is later returned to the river when conditions are more favourable. Obviously the impact of these retention measures depends very much on where the measure is taken in relation to where the flood is generated. For this reason, the impact of retention basins on the reduction of flood peak is considered within the hydrological computations, rather than in the hydraulic module.
This approach is applicable only for those retention basins in the reaches considered by the hydrological model, i.e. those in the Upper, Middle and Lower Rhine. For Retention basins projected in the Rhine Delta, retention basins are dealt with by considering the effect these have on the peak flow. Based on the threshold discharge and the volume that can be stored, a weir with a fixed level above which the basin comes into operation is applied. Clearly this threshold is set at the equivalent level of the threshold discharge, which is only exceeded for the extreme flood events. Table 5.7 shows the retention basins that have been projected or completed in the Rhine between Maxau and Lobith (Lammersen et al, 1999, Silva and Dijkman, 2000)
Table 5.7 Projected retention areas between Maxau and Lobith (Lammersen et al, 1999, Silva and Dijkman, 2000) Measure Constructio Area volume Approx n (Ha) (million m3) Location Upper Rhine Baden-Württemberg Elisabethenwört 2005-2020 400 11.9 387 Rheinschanzinsel 2005-2020 210 6.2 390 Rheinland-Pfalz Wörth/Jockgrim gestuurd 2005-2020 275 12.0 373 Mechtersheim 2005-2020 225 7.4 390 Flotzgrün 2000-2005 165 5.0 397 Kollerinsel 2000-2005 235 6.1 410 Waldsee/Altrip/Neuhofen 2005-2020 265 8.1 421 Bodenheim/Laubenheim 2005-2020 200 6 491 Ingelheim 2005-2020 130 3.8 518
Hessen between confuences of unknown ? (Est. 80) 472 Neckar and Main (estimated location at Kühkopf) Middle Rhine - - - - Lower Rhine Nordrhein-Westfalen Köln-Langel 2000-2005 500 6 673 Worringer Bruch 2000-2005 600 29 707 Ilverricher Bruch 2000-2005 400 15 760 Bylerward 2000-2005 1.100 25 850
STORM-Rhine, Main Report – Design of the Roleplay Page 24
In the Rhine Delta, downstream of Lobith, further retention basins have been projected. The two main retention basins are shown in Table 5.8. Numerous other small basins have been projected on each of the three river branches.
Table 5.8 Projected retention areas between Maxau and Lobith (Silva and Dijkman, 2001) Measure Construction Area Volume Location (Ha) (million m3) (approx.) Rhine Delta Rijnstrangen unknown 3100 150 865 Ooijpolder/Duffelt unknown 1400 65 Waal (875)
In the role-play model for the Rhine Delta, the retention basin projected at “Rijnstrangen” (Table 5.8) is considered as a green river, with the outlet of the basin on the IJssel river. Rather than take the distributed retention basins actually considered, the approach taken in the role-play is to define retention basins along each of the main sections. In the Rhine Delta, approximate volumes are estimated at 100m3/s on each of the three branches, where for the upper three reaches, an “average” retention basin is estimated for each of the four federal states. This approach is followed as it allows for a clear selection by the different role-players, without making the number of options unnecessarily large. This leads to the retention basins in Table 5.9 being considered. Threshold discharges for the retention basin are derived from Lammersen et al (1999) for the basins in the Upper Rhine. For the lower Rhine, the 100 year discharge for Lobith is applied in the same way as in the original STORM model for the Rhine Delta (de Graaf, 1999)
Table 5.9 Retention basins considered in the STORM model of the Rhine Retention basin Section Area Volume Location QStart (Ha) Million m3 (km) m3/s Rheinland Pfalz Upper Rhine 1495 48.4 373 4500 Baden-Württenberg Upper Rhine 610 18.1 395 4500 Hessen Upper Rhine 2700 80.0 468 5200 Nordrhein-Westfalen Lower Rhine 2600 75.0 846 13500 Waal Waal 100 13500 Dutch Rhine Dutch Rhine 100 13500 IJsel IJsel 100 13500
The discharge flowing into the retention basin is determined during the computations. For the retention basins considered in the hydrological routing model (i.e. the first four in Table 5.9), the discharge is calculated through a balance of flows into the retention basin and in the river downstream of the basin. On the basis of a volume balance and the ratio of the stage discharge relationship of the river at the location of the retention basin and the stage discharge over the inlet weir of the basin, an estimate of the discharge into the basin can be established. The procedure considers the water level in the basin, and once this equals the level in the river, the flow into the basin is stopped. No return flow from the basin to the river is currently considered. For further details on this algorithm as well as the derivation of stage discharge curves reference is made to Annex A.
Side channels and green rivers Side channels are channels excavated in the floodplain, parallel to the main channel. Rather than change the geometry of the river, side channels are incorporated in the hydraulic computation by estimating the discharge through the channel and subtracting this from the discharge in the river. Side channels may be applied along any reach of the river, given adequate space in the floodplain.
Green rivers are similar to side channels, but are excavated outside the floodplain. Green rivers need not start and end within the same river branch as is the case with side channels, but may start on one branch and end in another. Excavation of green rivers in the Middle Rhine is obviously not feasible due to the Rhenish Slate Mountains nor in most of the Upper and Lower Rhine as the area outside the floodplains is heavily urbanised. Currently only one realistic option for the creation of a green river is considered, the Rijnstrangen area, starting approximately at Lobith and flowing into the IJssel downstream of the bifurcation of the IJssel and the Dutch Rhine. Details of the location and method of calculation for side channels and green rivers can be found in Annex A.
STORM-Rhine, Main Report – Design of the Roleplay Page 25
5.9 Hydrological computations
The schematisation of the hydrological routing model covers the sections of the Upper, Middle and Lower Rhine, with the downstream boundary at Lobith. Further extension of the approach is difficult to apply due to the bifurcation downstream of Lobith. At this point the steady state approach is again adopted, as attenuation effects are now much smaller, and the problem of distributed inflows found upstream of Lobith does not occur.
For the structure of the routing model, calculation nodes defined at locations where tributary inflows are to be accounted for, locations where retention basins are defined, as well as at a number of intermediate stations. These intermediate stations are either points of interest or are chosen such that the length of routing branches is not excessive. The nodes and branches of the routing model are given in Table 5.10 and a schematic diagram of the structure is given in Figure 5.6.
> Maxau
> Retention Rheinland-Pfalz Retention Baden- > Wurtenberg
Neckar >
Retention > Hessen
Main >
Nahe >
Lahn > > Mosel
Sieg >
Routing branch with nodes > > Input hydrograph Ruhr
> > Lippe Retention Basin
Retention > Nordrhein-Westfalen
Lobith >
Figure 5.6 Structure of the hydrological routing model (not to scale)
Table 5.10 Nodes and branches of the routing model Location Start Node End Node Extra condition at start node Branch 1 Maxau Retention Rheinland Pfalz Upstream boundary Branch 2 Retention Rheinland Retention Baden Wurttenberg Retention basin Pfalz Branch 3 Retention Baden Mannheim Retention basin Wurttenberg Branch 4 Mannheim Retention Hessen Inflow of Neckar Branch 5 Retention Hessen Mainz Retention basin
STORM-Rhine, Main Report – Design of the Roleplay Page 26
Location Start Node End Node Extra condition at start node Branch 6 Mainz Bingen Inflow of Main Branch 7 Bingen Koblenz Inflow of Nahe Branch 8 Koblenz Bad Honnef Inflow of Mosel and Lahn Branch 9 Bad Honnef Bonn Branch 10 Bonn Koeln Inflow of Sieg Branch 11 Koeln Duesseldorf Branch 12 Duesseldorf Ruhrort Branch 13 Ruhrort Wesel Inflow of Ruhr Branch 14 Wesel Retention Nordrhein-Westfalen Inflow of Lippe Branch 15 Retention Nordrhein- Lobith Retention basin Westfalen
The two parameters of the Muskingum routing equation (Fread, 1992 and also Annex A) for each branch of the routing model were calibrated using the SOBEK model of the Rhine to generate hydrographs at each of the nodes. This allowed for an inflow and outflow hydrograph to be established for each branch. A gradient based optimisation method was applied to find optimal parameter values. Discharge data for the 1993 flood event were applied in the calibration, and the parameter values were verified using discharge data for the 1995 flood event.
Comparison of the hydrographs which were calculated with the hydrological routing module and SOBEK are given in Figure 5.7 and Table 5.11. The comparison shows a very good fit between the two hydrographs, particularly as this is data for the verification event, the calibration event giving even better results. Errors between the two models are indeed smaller than between the calibrated SOBEK model and observed events. Interesting to note is that using the option to synchronise the peaks of the contributing hydrographs results in a discharge at Lobith of some 5% higher than without the synchronisation. This indicates that these floods may indeed have been more severe had an even less favourable synchronisation of peaks of the contributing tributaries occurred.
(a) Bad Honnef (b) Koeln (c) Lobith 12000 12000
12000
11000 11000 11000
10000 10000 10000 s s 3/ 3/ m m 9000 9000 ge ge r r 9000 a a h h c c si si D 8000 D 8000 8000
7000 7000 7000 Musk Musk Musk Sobek Sobek Sobek 6000 6000 6000 200 300 400 500 200 300 400 500 200 300 400 500 timestep (h) timestep (h) timestep (h)
Figure 5.7 Comparison between Muskingum routing model and SOBEK for the 1995 event (verification)
STORM-Rhine, Main Report – Design of the Roleplay Page 27
Table 5.11 Comparison of observed, Muskingum routed and Sobek routed sampled maximum discharges from the 1995 event at locations along the Rhine. Location Observed Sobek Muskingum Difference Muskingum Difference (Synchronised) (Synchronised) Maxau 3770 3770 3770 0.00% 3770 0.00% Retention Rheinland Pfalz 3756 3756 -0.02% 3756 -0.02% Retention Baden Wurttenberg 3742 3735 -0.20% 3735 -0.20% Mannheim 4483 4457 -0.59% 4836 7.87% Retention Hessen 4456 4437 -0.43% 4720 5.93% Mainz 5920 6103 6093 -0.16% 6675 9.38% Bingen 7127 7125 -0.02% 7695 7.98% Koblenz 10983 10999 0.14% 11688 6.42% Bad Honnef 10966 10985 0.17% 11628 6.03% Bonn 11113 11142 0.26% 11845 6.58% Koeln 10800 11080 11117 0.33% 11771 6.23% Duesseldorf 11014 11026 0.10% 11600 5.32% Ruhrort 11931 11929 -0.02% 12754 6.90% Wesel 11962 11973 0.09% 12816 7.14% Retention Nordrhein-Westfalen 11953 11963 0.09% 12649 5.82% Lobith 11885 11947 11959 0.10% 12591 5.39%
5.10 Hydraulic computations
Once an estimate of the discharges throughout the river network has been established using the hydrological module, the hydraulic module is applied to determine water levels. The schematisation for the routing model covers again the Rhine downstream of Maxau, including the Upper, Middle and Lower Rhine, as well as the branches of the Rhine Delta. The 17 branches forming the aggregation level of the role-play was found to be too rough for the hydraulic computations, with typical branch lengths in the order of 60-70km. As a consequence each branch is divided into a number of sub-sections. Table 5.12 shows the 17 branches included in the model, as well as an indication of the length and the number of sub-branches.
Table 5.12 Overview of branches and approximate locations No Branch Branch Start End River Length Number Start End (location) (location) Km (km) Sub- ID ID branch 1 U1 Upper Rhine 1 Maxau Mannheim 360-425 65 6 1 7 2 U2 Upper Rhine 2 Mannheim Mainz 425-497 72 7 7 14 3 U3 Upper Rhine 3 Mainz Bingen 497-530 33 3 14 17 4 M1 Middle Rhine 1 Bingen Koblenz 530-592 62 6 17 23 5 M2 Middle Rhine 2 Koblenz Bad Honnef 592-642 50 5 23 28 6 M3 Middle Rhine 3 Bad Honnef Köln 642-688 46 5 28 33 7 L1 Lower Rhine 1 Köln Düsseldorf 688-744 56 6 33 39 8 L2 Lower Rhine 2 Düsseldorf Wesel 744-813 69 8 39 47 9 L3 Lower Rhine 3 Wesel Lobith 813-862 49 6 47 53 10 R1 Dutch Rhine 1 Lobith Pannerden 862-867 5 1 53 54 11 R2 Dutch Rhine 2 Pannerden Arnhem 867-878 11 1 54 63 12 R3 Dutch Rhine 3 Arnhem Amerongen 878-918 40 3 63 66 13 R4 Dutch Rhine 4 Amerongen Nieuwegein 918-950 32 8 66 74 14 W1 Waal 1 Pannerden Tiel 867-915 48 4 54 58 15 W2 Waal 2 Tiel Gorinchem 915-955 40 4 58 62 16 Y1 IJssel 1 Arnhem Deventer 878-935 57 6 63 80 17 Y2 IJssel 2 Deventer IJsselmeer 935-993 58 6 80 86
For each of the sub-branches a representative cross section is calculated using the geometric data of the 1D Hydrodynamic model SOBEK model of the Rhine. For each of the calculation points the discharge derived from the hydrological routing module is applied (see Figure 5.2). Starting from the given levels at the downstream boundaries at Nieuwegein, Tiel and IJsselmeer, it then proceeds to calculate water levels at each point upstream until the upstream boundary is reached. During this procedure the division of discharge across the bifurcation may be adjusted such that level continuity as well as volume continuity are maintained across the bifurcation. A full explanation of the procedure is given in Annex A.
STORM-Rhine, Main Report – Design of the Roleplay Page 28
Besides the geometrical properties of cross sections as derived from the SOBEK model as mentioned, the hydraulic module applies roughness coefficients for each of the sub-branches. Roughness values for the floodplain are derived by the roughness module from the given floodplain land use in each sub-section, while for the main channel these are allocated depending on the type of river bank. Again details of the derivation, as well as coefficients used are given in Annex A. Allocating coefficients in this way differs in principle from the method applied when modelling the river in for example SOBEK. Although roughness coefficients may also be related to land use, the values are typically refined through calibration until the modelled water levels match those observed in reality as closely as possible. Such a calibration exercise is not applied here, as roughness values are derived directly through the interpretation of land use distribution. The advantage of this approach is that impacts due to land use change are more meaningful. This is because the influence on water levels from geometry and roughness are separated as much as possible. In the case of calibration being applied to estimate value of roughness coefficients, these coefficients may well be used to cover up uncertainties in the geometry and deficiencies of the 1D approach. One disadvantage is, however, that water levels calculated are likely to deviate significantly from measured level. This is not as big a disadvantage as it may seem, though, as the aim of the model is to investigate changes in water levels rather than absolute water levels.
5.11 Scenarios
The domain of the hydrological and hydraulic models is such that water levels and discharges are dominated by the discharge series applied as boundary conditions at the upstream end of the model and for the tributary inputs. Downstream boundaries have been chosen such that tidal influences from the North Sea can be considered negligible. To investigate the impacts of different discharges, three distinct discharge conditions are considered;
Critical flow. This is the lowest discharge condition, and is particularly relevant to navigation due to restricted depths. In the Netherlands the shipping authorities use a criterion of a discharge that is exceeded 97,5% of the time to assess the navigability of the river. This discharge is evaluated at the Lobith gauging station, and is currently taken to be in the order of 1000 m3/s. Upstream of Lobith the same criterion is applied here. As low discharges occur predominantly in the summer it may be assumed that the discharge is generated primarily by melt from alpine glaciers, augmented with base flow from the various tributaries. Low discharges such as these are fairly persistent when compared to flood pulses, and the limited storage in the river at low water levels, resulting in negligible storage effects. The combination of this persistency and lack of storage allows for the consideration of the discharge as a steady state event.
Dominant flow. This discharge condition is primarily applied to evaluate morphological changes in the river. The discharge is selected such that the river is considered to carry the maximum sediment load, and was estimated in the STORM Delta model at 4500m3/s at Lobith. Although this is well above the average annual flow at Lobith, it is not high enough to cause significant out of bank flow, thus limiting attenuation effects due to storage. Again flows of this magnitude can be assumed to be generated through a combination of snow-melt and base flows from tributaries, and thus persistent when compared to flood pulses. This allows for also treating the dominant flow condition as a steady state event.
Design flow. This discharge condition is the highest considered. In the original STORM model this condition was selected as the design discharge for the upstream end of the model at Lobith, hence its name. With the extension of the model upstream, and the spatially and temporally distributed nature of the contributing flows, the validity of a single design discharge for the entire basin is less apparent. The name is maintained, but different scenarios will be investigated such that the impact of variation in magnitude and source of contributing flood hydrographs can be explored. Obviously for the high flow conditions flow will not be in bank. This necessitates taking attenuation and lag effects into account. For each of the scenarios described below hydrographs as a function of time are given for the tributaries and upstream boundary.
STORM-Rhine, Main Report – Design of the Roleplay Page 29
The first three scenarios considered are developed to offer insight into both the impact of historical flood events and the synchronisation of hydrographs within these events. The last two scenarios consider the much more extreme event resulting in the 1/1250 year design discharge at Lobith and a variation of this event due to the projected effects of climatic change.
Scenario 1: 1993 Flood Event Scenario 2: 1995 Flood Event
The first two scenarios consider two recent major flood events, the first in December 1993 and the second in January 1995. Both scenarios led to extensive flooding and damage. The genesis of the events was relatively similar (Engel, 1997), with the Mosel as one of the major contributors.
Scenario 3: 1995 Flood with synchronisation of hydrographs
In this scenario the impact of synchronisation of hydrographs from the contributing tributaries is investigated. The inflow hydrographs are the same as for the second scenario, but at the confluences the hydrograph from the tributary is shifted such that its peak coincided exactly with the maximum of the hydrograph in the main river.
Scenario 4: 1/1250 year discharge at Lobith In the Netherlands a design discharge equivalent to a return period of on average once every 1250 years is taken into account. This design discharge has recently been established at 16000m3/s (Silva and Dijkman, 2001). A design hydrograph resulting in this peak has also been estimated. Obviously no historical records are available for the contributing hydrographs. An estimate for the average volume fraction each tributary contributed to the 1993 and 1995 flood events was established, and this fraction was then applied to the design hydrograph at Lobith to determine the contributing hydrographs from each of the tributaries. Small corrections were applied to account for attenuation effects. In the scenario the hydrographs at tributaries are synchronised to minimise attenuation due to differences in timing.
Scenario 5: 1/1250 year discharge at Lobith including climatic change
A relevant topic of research in recent years is effect the projected climatic changes will have on the discharge regime of the Rhine. Investigating the effects of climate change is a complex scientific challenge. Middelkoop (1999) discusses the impacts of climate change on the peak flows in the Rhine river basin using the Climate Change scenarios run with the UK Hadley centre's high resolution model (UKHI). For the Rhine basin estimates of change in temperature for 2050 range between +0.5oC and +2 oC with a central estimate of +1 oC are quoted, while for 2100 the range is between +1 oC and +4 oC with a central estimate of +2 oC. Obviously if the importance of rainfall and snow melt are taken into account, then it is important that impacts on climate change are not only assessed for the basin as a whole, but also for sub-basins where rainfall fed floods or snow-melt fed floods may dominate. Table 5.13 gives examples of climatic change impacts for different areas of the Rhine, as well as the difference in summer and winter.
Table 5.13 Estimates of climatic change for the Rhine basin using UKHI model. Projected changes given for upper estimate for 2050 and central estimate for 2100 (values identical). Adapted from Middelkoop (1999). Alpine Area Central Germany Lower Area Year Winter Summer Year Winter Summer Year Winter Summer Temperature Change (oC) 2.2 2.3 2.0 2.1 2.4 1.9 1.9 2.2 1.7 Precipitation Change (%) 1.8 8.6 -5.1 5.4 12.6 -1.9 11.3 16.9 5.6
Although these figures will change depending on the emission scenario selected, the climate model applied and the method of downscaling (Middelkoop, 1999, RIZA, 1999, Asselman, 1997), generally the results do indicate a shift of the hydrological regime of the Rhine, with snow-melt regimes diminishing and rain-fed regimes increasing. This typically is projected to higher winter discharges and lower summer discharges (Middelkoop, 1999). Again numerous figures of the impact on the discharge for the lower Rhine are given, Middelkoop (1999) quotes figures of between +12% for the 1000 year recurrence interval for the central 2050 estimate using UKHI to +25% for the upper 2100 estimate. Silva and Dijkman (2001) quote 12,5% for
STORM-Rhine, Main Report – Design of the Roleplay Page 30
2050 and 20% for 2100, while RIZA (1999) quote increases of high flows of between 5% and 20%, with low flows dropping by 10% to 20%.
Giving exact figures is obviously not possible when climate change impacts are concerned. Rather than apply a complicated derivation process, the impacts for this scenario are estimated here to be 10% increase for high (design) flows and 10% decrease for low (critical) flows. For the high flows each of the contributing tributaries is equally amended, while for the low flows it is considered that the change will affect snow-melt primarily, and only the upper boundary discharge is reduced. With respect to the discharge for dominant flow, no clear indications of the shift in regime were found in literature, and as such it was decided to keep the values the same for this event as in all other scenarios.
Table 5.14 gives the discharges at the different inflows for the selected scenarios. Bear in mind that for some tributaries these figures are the sum of a number of smaller inflows (e.g. the discharge for the Ruhr include the Wupper, Emscher and Erft), and thus do not necessarily reflect the observed peak derived from the measured series. The discharges derived through application of the routing model for different locations on the Rhine are given in Table 5.15
Table 5.14 Discharges at different tributaries for the selected scenarios (for design discharges these are the peaks of the hydrograph) Location Design Critical Dominant Critical Scenario 1993 1995 Lobith1600 Climate Change All All Climate 0 Scenarios Scenarios Change m3/s m3/s m3/s m3/s m3/s m3/s m3/s Maxau 2930 3770 4618 5079 800 3000 700 Neckar 2140 1130 1920 2112 50 250 50 Main 1225 1988 1920 2112 50 750 50 Nahe 185 1058 800 880 0 0 0 Lahn 576 528 640 704 0 0 0 Mosel 4169 3536 4800 5280 100 500 100 Sieg 87 224 160 176 0 0 0 Ruhr 1265 1220 1600 1760 0 0 0 Lippe 101 138 160 176 0 0 0
Table 5.15 Peak discharges at different locations along the Rhine for the selected scenarios at design flow Location 1993 Event 1995 Event 1995 Lobith Lobith Event+ 16000 16000 Climate Change Maxau 2930 3770 3770 4618 5079 Retention Rheinland Pfalz 2921 3756 3756 4616 5078 Retention Baden Wurttenberg 2905 3735 3735 4609 5070 Mannheim 4826 4457 4836 6512 7163 Retention Hessen 4797 4437 4720 6482 7130 Mainz 5915 6093 6675 8389 9228 Bingen 6075 7125 7695 9179 10097 Koblenz 10733 10999 11688 14592 16052 Bad Honnef 10715 10985 11628 14572 16029 Bonn 10752 11142 11845 14729 16202 Koeln 10717 11117 11771 14692 16161 Duesseldorf 10581 11026 11600 14568 16025 Ruhrort 11233 11929 12754 16115 17726 Wesel 11211 11973 12816 16220 17842 Retention Nordrhein-Westfalen 11049 11963 12649 16062 17668 Lobith 10988 11959 12591 16000 17600
STORM-Rhine, Main Report – Design of the Roleplay Page 31
6 MODEL on SHIPPING
6.1 Introduction
Measures to change the geometry of the river Rhine have been and are still being influenced by the shipping functions of the river Rhine. A role play on the management of the river Rhine should therefore give the users sufficient feedback on the consequences of a certain geometry for the shipping function of the river.
A change in the intensity of shipping affects the shipping sector itself, the transport sector in general, as well as the industrial activities along the river Rhine, because transport by ship is cheaper per tonkm. Transport by ship is more environmentally friendly than transport by road (energy, CO2, NOx, smog).
Ships reduce their load whenever the minimum depth on their journey is not sufficient for a full load. The minimum depth on a river stretch is called the normative depth (in Dutch MGD: maatgevende diepte). The frequency distribution of normative depths plays an important role. Therefore climate changes and river engineering measures affect the shipping sector.
Instead of investing in changes in the geometry, shipping can be increased/decreased by changes in the fleet. In the past years a European regulation has only allowed new ships to be built if this was compensated by an equal decrease in tons capacity. Therefore the number of ships has decreased and ships in general have become larger. Larger ships are more sensitive to low normative depths.
The shipping module presented here: • gives the player a mandate to change the fleet and • translates combinations of changes in the fleet and river engineering measures to changes in the shipping sector, the industry and the environment.
In this version of STORM – Rhine the role of the shipping function is managed by the National Government of The Netherlands and the Federal Government of Germany. They both get the same output. Each manages part of the fleet and is responsible for the river engineering measures in its own country.
The shippers are assumed to adjust their load to the normative depth of either the stretch Rotterdam-Duisburg, the stretch Duisburg-Mannheim or the stretch Rotterdam-Mannheim. Frequency distributions of normative depths are related to the Qh-relation and the frequency distribution of discharges in Lobith. The normative depth on the stretch Duisburg-Mannheim is taken 0.60 m less than the one of Rotterdam-Duisburg.
A full argumentation of the above choices is given in this report. The shipping module does not account for the influence of the width of the river on shipping manoeuvre. The module neither accounts for the reduction in shipping during extremely high discharges. The river engineering measures that are possible in the role play are not so drastic that manoeuvre and shipping during high discharges are much affected. At very low water levels, shippers get a subsidy on their travel. This neither has been taken into account in the model as it is an insurance rather than an external funding.
The model that is presented is a modified and elaborated version of a benefit analysis for a decision support system for dredging of the River Waal (WaalBOS, AVV, 2000, Resource Analysis, 2000). While the model for the river Waal only takes one normative water depth for the whole summer season, in this model the full probability distribution of discharges is taken into account.
6.2 Objective
The shipping module should give an estimate of the consequences for the shipping sector, the industry and the environment of changes in the normative depth and changes in the fleet
STORM-Rhine, Main Report – Design of the Roleplay Page 32
6.3 Assumptions
The following assumptions are used:
• A distinction is made between 5 ship types Europe ship, Tank ship, Dortmunder, 2- barge convoy set, 4-barge convoy set, 6-barge convoy set. These ship types are each assumed to represent the ship class they are in and for which data on transport are available (AVV, 2000, www.inlandshipping.com, see Table 6.1). Because the model does not account for manoeuvre, no distinction was made between barges in front or next to each other. At any depth, (4 barge convoy set) (CEMTVIB) can transport twice the amount of 2 barge convoy set. The benefit analysis in Resource Analysis, 2000, has only been carried out for the four barge convoy set. Between Rotterdam and Duisburg also 6 barge convoy sets are used. In reality the ship types are each increasing in size, apart from the convoy sets. Thus, new Europe ships are larger than older ones (De Vries, 2000).
• The normative depth in three stretches (Rotterdam-Duisburg, Duisburg-Mannheim, Rotterdam-Mannheim) determines the shipping in those stretches. In the model all normative depths have been related to the discharge in Lobith (km 860). It has been assumed that at the start of the play the normative depth on the stretch Duisburg-Mannheim is 0.60 m lower than on the stretch Rotterdam-Duisburg. This is in line with the differences in normative depth at the Agreed Low Water Level 1992 (Overeengekomen Lage Rivierstand, CCR, 2000). In this agreement the normative depth in the stretch downstream of Mainz (km 508-557) is restricting shipping with a depth of 1.90 m, while downstream of Cologne (km 688) the normative depth is 2.50 m.
• The ships are always loaded at the maximum amount that both the normative depth of the river and the capacity of the ship allow. It is proven in the next chapter that this is a realistic assumption; the relation between the minimum depth in the river stretch Rotterdam Nijmegen and the actual loads, as derived from actual data (AVV, 2000), is very similar to the relation assuming maximum loads.
• Ships need 0.3 m freeboard between their bottom and the river bottom (AVV, 2000).
• The normative water depth is based on the Q-h relation and the frequency distribution of Lobith. Arguments follow in section 6.3.
• Transport by water is limited by the capacity not by the demand, unless the maximum amount of cargo available at the sea harbours is reached. Thus, transport by water is assumed to have preference over transport by road, because it is cheaper per tonkm. There is a maximum amount of mixed cargo, wet bulk and dry bulk to be transported and these different types each have their own ship types, see Table 6.2. In reality the flexibility and speed of transport by road make it a very competitive way of transport. Here the demand for transport of goods is assumed constant. In reality, the demand for transport of bulk goods by ship changes. For example, the coal mining in the catchment has decreased considerably and has had influence on the shipping sector (CCR, 2000). The shipping sector expects much from developments in harbours, which make a combination of water and road transport more feasible (De Vries, 2000). The large increase in container transport by ship shows that developments are in this direction. The ship types in the STORM model only transport one type of cargo. However, although not yet often done, container transport is already possible with convoy sets (De Vries, 2000).
STORM-Rhine, Main Report – Design of the Roleplay Page 33
Table 6.1 Representative sizes of ship classes Width Length Max Depth full load number Type of Type Class B L d M of trucks goods full full (m) (m) (m) (ton) (25 t / truck) Dortmunder <= CEMT III dry bulk 8.2 67 2.5 900 36 Europaship CEMT IV mixed cargo 9.5 85 3.0 1500 60 Tankship CEMT Va wet bulk 11.4 100 3.5 3000 120 2 barges long CEMT VIa dry bulk 11.4 193 4.0 5100 204 2-barges long 2 barges wide CEMT Via dry bulk 22.8 117 4.0 5100 204 2-barges wide 4 barges CEMT VIb dry bulk 22.8 193 4.0 10200 408 4-barges* 6 barges CEMT VI? dry bulk 22.8 270 4.0 15300 612 6-barges* Source:www.inlandshipping.com, except for CEMT Vib
* derived from AVV (2000): average load when depth not normative = 9,800 ton, max. load = 10,600 ton. From the two figures the average is taken. N.B. Length of ship that pushes the barges is estimated 40 m.
6.4 Input
The national/federal government has the mandate: • to change the fleet of its country; • to implement river engineering measures to remove bottlenecks.
Each ship type has a default number of ships Ni and the player adds a change in this number ∆Ni. Each ship type has a default number of journeys Ji which changes proportionally to the fleet size.
The input for the implementation of river engineering measures is not described in this report, as it is standard for all kinds of functions. Table 3-1 shows which information is on the input screen of the fleet. The player immediately gets feedback on the costs of the investment and the capacity increase.
Apart from the input by the player, the module uses the output of the hydraulic module for the critical discharge Q95 (discharge with 95% probability of exceedance) and for the dominant discharge Q5, which both come from the routing module. Moreover, the hydraulic module uses the changes in normative depths at Q95 over the stretches Rotterdam-Duisburg and Duisburg- Mannheim. This change in normative depth is here annotated as dchange and is a change in comparison to the situation at the start of the play. These changes can be either due to a changing discharge regime or due to the river engineering measures implemented by the player.
Table 6.2 Input information for changes in fleet. (the player can adjust the number of ships by adding input of the column 'change') To be transported Costs Start Change incl. transport by road (106 €) (nr.) (nr.) mixed cargo 39 Europe ship 2.5 150 30 wet bulk 117 Tankship 7 1000 10 dry bulk 104 Dortmunder 1.5 2200 1 Two barge convoy set 5 1000 0 Four barge convoy set 8 600 2 Six barge convoy set 11 200 0 260 Total capacity 19.5 *106 ton Change in capacity 0.5 % Costs 163 * 106 €
STORM-Rhine, Main Report – Design of the Roleplay Page 34
6.5 Model
6.5.1 Relation between water depth and ship load
The physical law of floating is simple: 1 m3 of water has a mass of 1 ton. If a ship is loaded with an additional ton, the ship sinks until an additional 1 m3 of the ship is below the water level. Therefore, with use of the representative sizes of the ship the maximum load M with any water depth h can be computed.
Ship characteristics that are readily available are the width B, the length L, the full load Mfull and the depth at the full load dfull. Moreover, transport river ships have parallel vertical side walls. The ship bow (front side) and escutcheon (back side) are not exactly flat, but because of the large lengths of the ship, the top view of the ship can be assumed a rectangle of width B and length L.
This makes that the ship mass without any load can be derived:
Mship = B * L * dfull * ρ − Mfull Eq. 6 -1
where
Mship ship mass when ship is empty (ton) B ship width (m) L ship length (m)
dfull depth when fully loaded (m) ρ density of water = 1 (ton/m3)
Mfull load when fully loaded (ton)
The dry load of the ships is only derived once, see values in Table. The ship that pushes the barges convoys does not sink with higher loads; only the barges sink. Therefore, only the barge length and mass is taken into account.
For any water depth, the maximum load is computed as follows:
h − d M = M + M * Max Min over ,1,0 − M ()ship full ship Eq. 5-2 d full
where M maximum load of ship mass in case of minimum water depth h (ton)
dover required distance between bottom ship and bottom level river = 0.3 (m) h normative depth in stretch (m)
Figure 6-1 shows the model results for the different ship classes. Also the polynomial function that AVV fitted to the empirical data on the actual transport of 4 barge convoy sets is plotted (based on some 5100 journeys, Resource Analysis, 2000, based on reports AVV). The linear relation of Eq. 6-2 is far more transparent than a polynomial with parameters that have no physical meaning. The empirical data through which the polynomial is fitted are only for depths up to 4.5 m, because for larger depths the water levels do not restrict the load.
STORM-Rhine, Main Report – Design of the Roleplay Page 35
Actual loads can be smaller than maximum loads for the following reasons (Resource Analysis, 2000): • bulky load; volume is normative factor, • the water depth in another river stretch or in the harbour is limiting, • the shipment is smaller than the capacity of the ship.
Apart from the reasons above, the differences may be due to differences in ship sizes within the different classes. Judging from visits to the Rhine, most ships seem maximum loaded, apart from container ships.
Table 6 - 3 Ship masses, see Eq. 6.1 Ship mass Class Mship (ton) <= CEMT III 474 CEMT IV 923 CEMT Va 990 CEMT VIa 2-barges 1395 CEMT VIb 4-barges 2791 CEMT VIc 6-barges 4186
12000 <= CEMT III_load 10000 CEMT IV_load ) n
o 8000 t
( CEMT Va_load ip
h 6000 s
f 2 barges o d
a 4000
lo 4 barges
2000 polynomial AVV 0 1.5 2 2.5 3 3.5 4 4.5 5 minimum depth in branch (m)
Figure 6.1 Model results (Eq. 6.2) for different ship classes and empirical relation for 4 barges convoy sets, derived by AVV (2000)
STORM-Rhine, Main Report – Design of the Roleplay Page 36
ps. Lobith
2000 2000 1800 1800 1600 1600 m) m)
1400 c 1400 ( (c . . P P . . 1200 1200 A A . . N l N l 1000 1000 e e v v
e 800 800 le l r r e t te 600 600 a a w 400 w 400 200 200 0 0 0 5000 10000 15000 20000 02468101214 Q (m3/s) h (m)
1.20E+00 ps. Lobith 1.20E+00 ) 1.00E+00 - 1.00E+00 ( ) - e ( c e n c 8.00E-01 a 8.00E-01 d an e e ed c 6.00E-01 x
e 6.00E-01 ps. Lobith exce of f y t o i . 4.00E-01 l b 4.00E-01 bi o r P 2.00E-01 oba
pr 2.00E-01
0.00E+00 0.00E+00 0 5000 10000 15000 20000 0 2 4 6 8 10 12 14 Q (m3/s) h (m) Figure 6.2 Derivation of relation between water depth and cumulative probability of this water depth, from the Qh-relation in Lobith and the cumulative probability of certain discharges in Lobith
6.5.2 Frequency distribution of normative water depths
The Qh-relation and the cumulative probability distribution of the discharges can be combined to a cumulative probability distribution for certain water levels. This is corresponding with the tactics of the shippers: They determine their load using the gauge levels along their journey, they assume that the bottleneck in their journey is related to the gauge level 1:1.
The gauges Maxau (km 326), Kaub (km 546), Ruhrort (km 780 in Duisburg) are important. However, the STORM shipping module only relates to the cumulative probability function of Lobith, f or the following reasons: (1) only using Lobith keeps it simple, (2) Only for Lobith gauge data were readily available, (3) in times of low water, the spatial correlation is very high, thus upstream water levels can also be assumed low.
In the STORM hydraulic module scenarios for the dominant discharge (Q5 is 5 % probability of exceedance) and the critical discharge (Q95) are used. For the shipping module Q99 is necessary. In line with the histogram of Lobith, it can be assumed that linear interpolation in frequencies is possible between the critical discharge Q95 and the discharge Q40 and that
Q99 = 0.75 *Q95 Eq. 6.3
Q40 = 0.50 *Q5 Eq. 6.4
It has been assumed that the normative depth on the stretch Duisburg-Mannheim is always 0.60 m lower than on the stretch Rotterdam-Duisburg. This corresponds to the differences in normative depth at the Agreed Low Water Level 1992 (Overeengekomen Lage Rivierstand, CCR, 2000), in which case the stretch downstream of Mainz (km 508-557) is restricting shipping with a depth of 1.90 m, while downstream of Cologne (km 688) the normative depth is 2.50 m. Figure 5-1 shows how the cumulative probability distribution of the water depth is derived. The cumulative probability distribution of the discharges is readily available, based on daily data
STORM-Rhine, Main Report – Design of the Roleplay Page 37
from the period 1901-1995. The normative depth of 4.5 m has a cumulative probability of 60% (stretch Rotterdam-Duisburg, AVV, 2000, subsection 1.4). The water level at Lobith with the same cumulative probability is 9.70 m. Therefore, the relationship between normative depth between Rotterdam and Duisburg and water level in Lobith is assumed:
h = ()h − 9.70 + 4.50 + d Rotterdam−Duisburg lobith change Rotterdam−Du Eq. 6.5
where
hRotterdam-Duisburg is the normative depth between Rotterdam and Duisburg. (m)
hlobith is the water level in Lobith. (m)
dchange Rotterdam-Duisburg is the increase in normative depth due to river engineering measures in the stretch Rotterdam-Duisburg. (m)
Because of the assumption that at the start of the play the normative depth on the stretch Duisburg-Mannheim is 0.60 m lower than on the stretch Rotterdam-Duisburg for all discharges.
hDuisburg−Mannheim = (hlobith −9.70)+ 4.50+ dchangeDuisburg−Mannheim −0.60 Eq. 6.6
hDuisburg-Mannheim is the normative depth between Duisburg and Mannheim (m)
dchange Duisburg-Mannheim is the increase in normative depth due to river engineering measures in the stretch Duisburg-Mannheim. (m)
Figure 6.3 illustrates the form of the probability density function, which directly follows from the cumulative probability distribution. (In statistical terminology: Fig. 6.3 gives the histogram, from which the probability density function can be derived.) For low water depths the frequency of occurrence is almost linearly related to the water depth. As only for low water depths the normative depth is really determining the load, this will serve in the derivation of the frequency distribution of ship loads.
ps. Lobith
6.E-01
after increase of 5.E-01 normative depth with 0.5
) m -
ce ( 4.E-01 en r r u
c 3.E-01 c o
f o
. 2.E-01 eq r f 1.E-01
0.E+00 02468101214 h (m)
Figure 6.3 Probability distribution of the existing normative water depth (Eq. 6.2) and of the normative water depth that is 0.5 m larger
STORM-Rhine, Main Report – Design of the Roleplay Page 38
6.5.3 Frequency distribution of ship loads
The cumulative probability distribution of ship load follows from the cumulative frequency distribution of the normative water depth and the relation between water depth and ship load. Ships with smaller maximum depths have a larger probability that the normative water depth is not restricting their load. With the scenarios for scale increase in the shipping sector (De Vries, 2000) the sensitivity for low water occurrences becomes higher.
Figure 6.4 shows for the different ship classes the probability that the normative depth is not restricting the load, for the present situation in Lobith. This is the probability that the normative depth exceeds the maximum depth of the ship plus a free board of 0.3 m (dover).
Figure 6.5 shows the probability density functions of loads that are not the maximum ship loads. Together with the probability of having a full load, the average load for a certain morphology can be computed:
Lim(M→Mfull ) M = M * f (M) dM + M * p(M ) Eq. 6.7 ∫ full full 0 where
p(Mfull) probability of occurrence of full load:
p(Mfull ) = 1− F(h = dfull + dover ) Eq. 6.8
Thus, p(Mfull) equals the probability of exceedance of a normative depth that is the sum of the maximum depth of the ship (dfull) and the freeboard necessary between the bottom of the ship and the bottom level of the river (dover = 0.3 m).
In the Lobith Qh-relation the Qx is looked up that corresponds with hlobith = dfull + dover - dchange Rotterdam-Duisburg –4.50 + 9.70 for the stretch Rotterdam-Duisburg (see Eq. 6.7) and hlobith = dfull + dover - dchange Duisburg -Mannheim–4.50 + 9.70 + 0.60 for the stretch Duisburg-Mannheim (See Eq. 6.8). Because of the linearity in the cumulative frequency distribution between 40% and 95% probability of exceedance, the probability of exceedance x is calculated through:
Qx −Q95 95 − 40 p(Mfull ) = x = 0.95 − * Eq. 6.9 Q40 −Q95 100
To estimate the integral in Eq. 6.8, it is possible to use a discrete version of the probability density functions in Figure 6.4 and Figure 6.5 However, to evaluate quickly the impact of a change in normative water depth and to make the model easier applicable in other locations, it is preferred that the model can be computed analytically. For this purpose, a linear estimate of the probability density function is used.
f (M) = Max(M − Mmin ,0) * fmax Eq. 6.10
Mmin Ship load with a probability of exceedance of 99% ton
h − d M = M + M * Max Min 99 over ,1,0 − M Eq. 6.11 min ()ship full ship d full
where h99 is the normative depth corresponding to Q99 at Lobith, which follows from Eq. 6.5 for Rotterdam-Duisburg and 5-6 for Duisburg-Mannheim through substitution of hlobith with h99lobith , which is the water level at Lobith with a probability of exceedance of 99%. To find the h99Lobith, the Q99 as obtained with Eq. 5-3 is used. Subsequently the Qh-relationship at Lobith is linearly extrapolated, using the two pairs of Q and h with the highest probability of exceedance. This is
STORM-Rhine, Main Report – Design of the Roleplay Page 39
necessary, because for climate change scenarios the Q99 is lower than the present Q99 and no water levels are available for such low discharges.
fmax Probability of occurrence of load that is close to maximum load -
Because the area under the total probability function is necessarily 1, fmax amounts:
1− p(Mfull ) fmax = * 2 Eq. 6.12 Mfull − Mmin
A comparison of the dots with the drawn lines in Figure 6.5, illustrates the accuracy of the estimates.
As a consequence the analytical estimate of Eq. amounts:
Lim(M→Mfull ) fmax M av = M * M − M dM + M * p(M ) ∫ ()min full full 0 Mfull − Mmin M f 1 1 full Eq. 6.13 = max M 3 − M 2 * M + M * p(M ) M − M 3 2 min full full full min Mmin
1− p()Mfull 1 3 1 2 1 3 M av = * 2 * M − M * M + M 2 full full min min ()Mfull − Mmin 3 2 6 Eq. 6.14
+Mfull * p(Mfull )
Eq. 6-5 and 6-6 can give a negative size for Mmin, the load with a probability of exceedance of 99%. This negative load is used in the integration, but is corrected by an additional term that computes the cumulative probability of a negative load and assumes that this load is zero: − M min M = M av + if Mmin < 0,p(Mfull ) * 0.5 * ,0 Eq. 6.15 Mfull − Mmin
Figure 6.6 shows that the estimate of the average load in the analytical way reasonably agrees to the estimate in the discrete way. The analytical estimates are slightly higher than the discrete estimates. It is concluded that the analytical estimates can be used.1
In the modelling approach suggested here, the ships are travelling with a constant frequency and are always maximum loaded. Thus, the frequency of travelling is not larger in times of low water. It is assumed it is not feasible that ships lay idle. In the study on river transport in The Waal (Resource Analysis, 2000), two alternative cases were taken: (1) buffering is possible (frequencies will be constant), (2) only direct transport is possible (more ships will sail in times of low water). This second possibility is thus not taken into account.
1 In the Waal study (Resource Analysis, 2000), the normative water depth was assumed constant over 3 months. This makes the average load transported very sensitive for dredging. Due to an increase in normative depth by 0.2 m (3.6 > 3.8 m) the average load transported by a 4 barges convoy is supposed to increase with 6% over the three months or 1.4 % over the whole year (change from 3/4*9800 + 1/4 * 8187 = 9397 ton to 3/4*9800 + 1/4 *8724) = 9531 ton. In the case described in this report, the average load is anyway less than the one suggested by the Waal study, as the normative depth is assumed lower. In the model of this report, for a four barges convoy set, an increase of normative depth of 0.2 m lead to an increase of average load of 3% (8703 becomes 8978 ton).
STORM-Rhine, Main Report – Design of the Roleplay Page 40
1.00
s 0.90 i <= CEMT III d
a 0.80 CEMT IV lo ) . 0.70 x (- a d 0.60 CEMT Va te m t r a o 0.50 h p 2 barges 4 barges t s
y 0.40 n ilit
tra 0.30 b a
b 0.20 o r
p 0.10 0.00 0 2000 4000 6000 8000 10000 12000 load (ton)
Figure 6.4 The probability that different ship classes can transport maximum loads, plotted as a function of these maximum loads
8.00E-04 <= CEMT III 7.00E-04 CEMT IV
6.00E-04 CEMT Va
) 2 barges -
( 5.00E-04 y
t 4 barges i l
bi 4.00E-04 modelled 0 probability
oba 3.00E-04 modelled max. probability pr
2.00E-04
1.00E-04
0.00E+00 0 2000 4000 6000 8000 10000 12000 load (ton)
Figure 6.5 Probability density functions of non-maximum loads for different ship classes and a simplified model that uses an estimate of the load which is exceeded with 99% of the time (♦) and an estimate of probability that the ship can be fully loaded (•)
STORM-Rhine, Main Report – Design of the Roleplay Page 41
ly 10000 ical
alyt 8000 an
y b
d 6000 e t a on) t m i ( t
s 4000 e ad o l
2000 e g a r 0 ave 0 2000 4000 6000 8000 10000 average load estimated by discrete solution (ton)
Figure 6.6 Comparison of analytical and discrete way to determine average load
12000 <= CEMT III CEMT IV 10000 CEMT Va on) t 2 barges d ( e t 8000 4 barges por s n a
r 6000 on t t
e 4000 g a r e v
a 2000
0 0 0.1 0.2 0.3 0.4 0.5 0.6 increase in normative depth through dredging (m)
Figure 6.7 Average nr of tons transported, under the assumptions that (1) always the maximum possible load is transported and (2) the demand of transport is equally distributed over the year and (3) the loads cannot be buffered
STORM-Rhine, Main Report – Design of the Roleplay Page 42
6.6 Output
6.6.1 Introduction
The output relates to economic effects and evironmental effects.
6.6.2 Economic effects
The model computes the following indicators: • change in average turn over per ship/year, €/y • additional benefits industrial sector total, 106 €/y • change in employment shipping sector (1 ship = 2 persons) persons • change in employment road transport sector (1 truck = 25 ton = 1 person) persons
For the above indicators a distinction is made between 'due to fleet changes' and 'due to fleet and geometry'-changes. Apart from these performance indicators total sums of financial costs are given: • investment in new ships 106 • costs of depth increase 106 €
The benefits of an increase in average ship load translate directly to the shipping sector, because with the same number of ships they can make a larger turn-over. The player gets the average turnover per ship: m n ∑∑()M * ()Jij + ∆Jij * D j * Pw stretches j shiptypes i
T = n Eq. 6.16 ∑()Ni + ∆Ni shiptypes i where
T average turnover per ship per year (€/ship/y)
M average load per ship (ton)
Jij number of journeys of particular ship type i in particular stretch j (-)
Ni + ∆Ni Jij + ∆Jij = * Jij Eq. 6.17 Ni D distance of stretch. Either Rotterdam-Duisburg, Duisburg-Mannheim or Rotterdam- Mannheim (km)
Pw price per tonkm for transport via water (0.006) (€/tonkm)
Ni number of ships of certain type (-)
In the model, for the average ship turn over this version of STORM does not make a distinction between German and Dutch ships. The number of journeys are given in Table 6.3. The numbers in STORM Rhine are less than in reality to get the total number of tonkm correct and to have sufficient sensitivity for measures.
For 4 barge convoy sets the average price is about 0.006 €/tonkm. (For smaller ships than convoy sets, it is about 0.011 €/tonkm). In STORM the price of 0.006 €/tonkm is used. This implies that for the 4 barge convoy set the benefit for the shipping sector for travels between Rotterdam and Duisburg is about 4800*225*0.006 *∆M Euro/year. Each ton change in the average weight is thus giving a change in turnover of € 6480. A change in the normative depth of 0.2 m, yields a change of 175 ton in the average load (175 = 8978-8703). This yields a change in benefits of 1,134,000 €/y.2 .
2 Although assumptions are quite different, finally the result is not so different from the 1,016,400 Euro/y by Resource Analsyis, 2000. In that case the benefit per journey saved of 44,000 NLG. However, this benefit is based on saving a journey with a full ship, while the journeys saved by definition had less load than the full load, because the transported nr of tons stays the same.
STORM-Rhine, Main Report – Design of the Roleplay Page 43
Because transport by truck is more expensive per tonkm than transport by ship, the benefits for the shipping sector are also considerable.
m n
B = ∑∑(M * J * D * (Pr − Pw )) Eq. 6.18 stretches j shiptypes i where B benefits of industrial sector from shipping (€/y)
Pr price per tonkm for transport by road (0.045) (€/tonkm)
The other parameters are as above. The total cargo transported is limited per type of cargo; it cannot become higher than the amount offered at the Rotterdam harbour. A distinction is made per type of cargo: mixed cargo, wet bulk, dry bulk.
The price of road transport is about 0.045 €/tonkm. This makes the impact for the industry 7 times as sensitive. For simplicity and to make the model sensitive, it is assumed that the shipping sector replaces the road transport sector over the same stretches, while in reality the goods often have to be reloaded at the harbour to a truck.
The changes in employment in the shipping and in the road sector directly depend on the change in fleet size: every additional ship gives employment of 2 extra persons.
n
∆Eship = 2 ∑ ∆Ni Eq. 6.19 shiptypei
where ∆Eship is the number of additional jobs in the shipping sector due to the change in fleet size.
However, each 25 ton increase in fleet size means the loss of employment for 1 person in the road sector.
1 n
∆Eroad = − ∑ ∆Ni * Mfull,i Eq. 6.20 25 shiptypei
where ∆Eroad is the number of additional jobs in the road sector due to the change in fleet size.
In this assumption just the drivers/shippers are counted, not the supporting staff. It is implicitly assumed that the change in that this will be the same per tonkm for both sectors and that the national government, who plays the role in STORM, will not care too much about a shift between sectors.
Table 6.4 Number of journeys from ships from different ship classes, historical and derivations for STORM. Ships nr of journeys (rounded average of figures 1998 simplifications for STORM model & 1999, see table 3 & 3 report AVV, 2000) km 150 225 600 375 ARA*- ARA*- Netherlands R'dam- R'dam- Duisburg- Waal Duitsland Duisburg local Duisburg Mannheim Mannheim CEMT IV 45000 19000 2500 2,000 2,500 8,000 6,000 CEMT Va 28000 16000 2300 12,000 2,300 13,700 2,000 <= CEMT III 96000 19000 3000 15,000 3,000 8,000 2,000 CEMT VIa 2-barges long 3200 2000 350 1,800 700 2,530 300 CEMT VIa 2-barges wide 1800 1200 320 CEMT VIb 4-barges 7000 6000 4800 500 2,400 1,200 0 6 barges 0 200 300 0 * ARA = Antwerpen / Rotterdam / Amsterdam
STORM-Rhine, Main Report – Design of the Roleplay Page 44
6.6.3 Environmental effects
Transport by ship has a few advantages for the environment compared with transport by road. These advantages are used in the lobby by the shipping sector (De Vries, 2000). Therefore STORM gives the following indicators for environmental effects. • energy use %
• CO2 emission %
• acid rain (NOx + SO2) % • smog % The computation of the indicators is based on the change in tonkm transferred between transport by road and transport by water, with and without measures. This is based on the tons that should be transported from Rotterdam harbour. Of a total amount of 260 * 10 6 ton, 15% is mixed cargo, 45% is wet bulk, 40% is dry bulk (De Vries, 2000). Each ship type has a certain type of cargo. Because the number of journeys from Rotterdam and also the average load are known, the tons transported by water from Rotterdam can be easily computed.
n m−1 Mtotal water = ∑∑()M * J Eq. 6.21 ship types i stretches j
where Mtotal water is the total amount transported from Rotterdam by ship (tons)
Thus, the stretch Duisburg-Mannheim is not taken into account.
The ratio of energy use by shipping to energy use of transport by truck is only 17%, for CO2- release only 18%, for acid rain (SO2+NOx) 50% and for smog production 45%.
The relative benefit of shipping for the environment is:
Mtotal water *θ + (Mtotal offered − Mtotal water ) *100 θ total = 100 − Eq. 6.22 Mtotal offered where
θ % of parameter (energy use, CO2 emission, acid rain, smog) of one tonkm if cargo is transported by water instead of transported by road
θtotal % of parameter if particular amount Mtotal water is transported by water compared with if all offered cargo is transported by by road Mtotal offered total amount of cargo offered in Rotterdam (260) (tons)
The rate of change of θtotal caused by the fleet change, the river engineering measures and the original θtotal is the percentage given in the output.
STORM-Rhine, Main Report – Design of the Roleplay Page 45
6.7 Discussion and synthesis
In the model the consequence of a measure that changes the normative depth, translates into a change in the probability that the maximum load can be transported p(Mfull).
The model only depends on: • the probability distribution of the exceedance of a normative water depth (see Eq.4.5); • the scenarios for ship sizes and ship numbers. It is assumed that the demand for river transport is not restricting, but such a restriction can be easily incorporated.
The sensitivities of the average loads are not very high, not even for the deepest ships. However, a small percentage wise change in load already has a large economic impact in absolute sense. Therefore, the modeloutput is very sensitive for the model assumptions. The model is directly proportional to the price per tonkm and to the travel distance assumed. The STORM shipping module makes a distinction between different ship types.
It is not sufficient to derive one relationship for the whole fleet. Different influences play a role: • Due to climate change and upstream land use changes, the cumulative frequency distribution of the discharges change. River training measures may change the Q-h relationship. Through computation with a few representative discharges, the hydraulic model of STORM Rhine derives new probability distributions of the exceedance of a water depth. • There is a trend to move to bigger ships, in particular barges. The number of ships active in the Dutch inland shipping sector with less than 1000 ton capacity, decreased between 1985 and 1998 from 4600 to 2800. The number above 1000 ton increased from 1600 to 2200 between 1985 and 1990 and is quite stable since. (De Vries, 2000). Thus, the sensitivity for low water depths is increasing. • Container transport is increasing steadily (De Vries, 2000). For containerships the volume is often the normative factor instead of the depth. This is not taken into account in the model.
STORM-Rhine, Main Report – Design of the Roleplay Page 46
6.8 Information on role
Introduction
Shippers reduce their load whenever the minimum depth on their journey (the normative depth) is not sufficient for a full load. The changing climate gives a higher probability of low discharges, which means that in the future more often ships cannot be fully loaded. Through river engineering measures the bottlenecks on the journey can be removed, but that costs money.
Instead of investing in changes in the geometry, shipping can be increased/decreased through changes in the fleet. In the past years a European regulation only allowed new ships to be built if this was compensated by an equal decrease in tons capacity. Therefore the number of ships has decreased and ships in general have become larger. Larger ships are more sensitive to low normative depths.
A change in the intensity of shipping affects the shipping sector itself, the transport sector in general, as well as the industrial activities along the river Rhine, because transport by ship is cheaper per tonkm than transport by road. Transport by ship is also more environmentally friendly than transport by road (energy, CO2, NOx, smog).
It is assumed that transport by water is limited by its capacity and will have preference over transport by road. There is a maximum amount of mixed cargo, wet bulk and dry bulk to be transported and these different types each have their own ship types.
You have a mandate to: • change the fleet; • implement river engineering measures to remove bottlenecks.
The indicators that give you a feedback on the changes in the shipping sector are:
economic changes
• change in av. turnover per ship/year, €/y • additional benefits industrial sector total, 106 €/y • change in employment shipping sector (1 ship = 2 persons), persons • change in employment road transport sector (1 truck = 25 ton = 1 person), persons • investment in new ships 106 € • costs of depth increase 106 €
environmental changes • energy use %
• CO2 emission %
• acid rain (NOx + SO2) % • smog %
STORM-Rhine, Main Report – Design of the Roleplay Page 47
7 PRESENT and TARGET LAND COVER
7.1 Introduction
In the role play STORM Rhine, the land cover in the floodplains is an important element. This report describes the distribution of the 'present' and of the 'target' land cover and how they were derived. The 'present' land cover is the land cover at the start of the play. The 'target' land cover is the land cover which is supposed to be the land cover that is best from an ecological perspective, without being unrealistic. The STORM Rhine module uses this 'target' land cover for comparison when evaluating the measures that the players have taken.
An ecotope is a spatial unit of a corresponding ecosystem which is a combination of abiotic site conditions and related plant and animal communities. STORM Rhine's description of land cover is based on an ecotope distribution. The ecotope distribution method that is used in STORM Rhine is based on work by Pedroli and Rademakers (1995), who developed a system that is called RES (Rivier Ecotopen Stelsel)
For The Netherlands the ecotope distribution was readily available, as determined by Pedroli and Rademakers and implemented in STORM DEMO. However, for the German parts of the area the ecotope distribution was not readily available. A translation had to be made for the descriptions of land covers in North Rhine Westphalia, for which GIS maps with biotypen were available, and for Rhineland Palatinate, where the Gewässerstrukturgütekartierung (STRUKA) was available. This report describes this translation.
7.2 General
Each hydraulic branch of the river system has:
• a 100% natural or a 100% stone river bank (e2 = 1 and e3 = 0 or vice versa); • a distribution of the percentages of the floodplain that are occupied by a certain ecotope.
In STORM Delta for each sub-branch of the river the percentages of certain ecotopes are given.
For the purpose of the role play, the RES-system was slightly altered. The ecotopes campsite (e19) and recreational harbour (e20) were added to give more meaning to the recreational function of the river. The ecotope earthen bridge ramp was added to make available the measure 'removal of earthen bridge ramps'. An overview of all ecotopes and their description is given in Appendix A.
7.3 The Netherlands
For the Dutch part of the STORM Rhine the present and target ecotope distribution by Pedroli and Rademakers was used (1995), like was done in STORM Delta. Only slight alterations are made to introduce the ecotopes campsite (e19), recreational harbour (e20) and earthen bridge ramps (e21). The RES 'present' ecotope distribution was based on:
• hydrodynamics (water levels and durations of flooding), • flow dynamics (velocity of flow), and • management intensity (resulting in vegetation structure).
Pedroli and Rademakers in first instance based their target ecotope distribution on geomorphological trajects. However, such an approach resulted in a reference ecotope distribution which had so much softwood forest that it was not considered realistic. Therefore, using some rules of thumb and the conditions at the design discharge, they manually determined the target ecotope distribution.
STORM-Rhine, Main Report – Design of the Roleplay Page 48
7.4 North Rhine Westphalia
This subsection describes the methodology that was used to determine the 'present' ecotope distribution of North Rhine Westphalia. A description of the 'target' ecotope distribution is described in chapter 0, for the whole of Germany.
The present ecotope distribution in North Rhine Westphalia was determined using the preliminary LEDESS-ecotope GIS map (Alterra, 2001). LEDESS is an alternative methodology for determining ecotopes and stands for Landscape Ecological DEcision Support System. LEDESS gives each piece of land two qualifications: the type of fysiotope (water levels, type of soil, geomorphology etc.) and the type of vegetation. Alterra also made a translation table between the LEDESS and the RES-ecotopes. Alterra based its LEDESS ecotope distribution map on a land cover map of the Ministry of Environment of North Rhine Westphalia, which uses again another classification. In Table 7.1 the translation for STORM Rhine is given.
This algorithm does not give a key to all ecotopes. The areal percentages of the following ecotopes were manually estimated, subtracting areas from similar ecotopes: e07 Production forest e08 Herbaceous floodplain e10 Natural levee e19 Camping/park e20 Recreational harbour e21 Earth ramp e22 Industrial harbour
If the area in either the left or the right half of a STORM sub branch has more than 800 m2 that is classified as Annuellenfluren this half of the sub branch is considered to have a natural river bank.
The percentages of the floodplain area occupied by a certain ecotope were determined with the use of ARC-View. The reclassification tables are in
Table 7.1 Translation of GIS map of NRW to ecotope distribution in STORM Rhine NRW GIS map STORM Rhine (RES)
System of NRW LEDESS system Legende Ldss_bg Ldss_fysiotopen r_type A Waldflächen 19 0 e5 hardwood floodplain forest B Gehölze 15 0 e6 softwood floodplain forest C Grosseggenried/ 5 9 e9 back swamp Röhrichtbestand D Trockenrasen 12 0 e11 natural grassland E Grünland 8 0 e12 production grassland F Gewässer 1 16 diep zomerbed e_river N.B. not to be 1 19 nevengeul e13 considered in computation of percentages 1 20 plas open e16 side channel verbinding 1 21 plas geïsïoleerd e16 gravel pit lake 1 22 dynamische strang e14 periodical flowing side-channel 1 23 geïsoleerde strang e15 isolated oxbow lake
H anthropogen 6 13 e17 crop cultivation geprägte Standorte
STORM-Rhine, Main Report – Design of the Roleplay Page 49
L Annuellenfluren 3 0 e2 natural river bank K Hochstaudenfluren 4 0 e4 river dune KB Siedlungen etc. 22 14 e18 building area
7.5 Rhineland Palatinate and Hessen
In this chapter only the methodology to determine the 'present' ecotope distribution of Rhineland Palatinate and Hessen is described. A description of the target ecotope distribution is described in chapter 0, for the whole of Germany.
For Rhineland Palatinate and Hessen, the Gewässerstrukturgütekartierung of the Landesamt für Wasserwirtschaft of Rhineland Palatinate (STRUKA V) formed the basis for the determination of the ecotope distribution. STRUKA V contains for each river side for river branches of in between 0.2 and 5 km long whether the river bank or floodplain satisfies a particular criterion.
For STORM, a river bank was considered natural if less than 50% of the STRUKA branches in the STORM subbranch satisfied the criterion that '5.2 Uferverbau –Kein Verbau' amounted more than 50%. For some of the STORM ecotopes, the STRUKA V did not offer a key to the percentage. In that case the percentage of the area was guessed or manually estimated using the Rhine Atlas. The percentage of the floodplain that was not classified manually was distributed over the remaining ecotopes. STRUKA makes a distinction between > 50% or 10- 50% of the river bank or floodplain satisfying a certain criterion. For STORM, the 10-50% was assumed to count half of the > 50% criterion. The number of STRUKA branches satisfying a criterion was summed and ecotopes were attributed percentages of the area proportional to their score. Of the Bodenständiger Wald, which is 'indigenous forest', 2/3 was assumed hardwood floodplain forest and 1/3 softwood floodplain forest. The Nicht- Bodenständiger Wald was assumed production forest. The algorithm is explained in Table 7.2.
In particular in the area upstream of Mainz, the algorithm yielded that most of the floodplain was building area or recreational area. This is not realistic as in that area most of the floodplain is nature reserve (the Rhine Atlas, 1998). For the role play it is not very interesting to have such a large area of nature reserve, already in the start situation. Therefore most of thearea that the algorithm considered building area was manually changed to production forest. The other ecotopes were attributed area proportional to their scores. On other locations the scores were also very unrealistic and again manually changes were made. These can be looked up in
It is feared that the reason for the unrealistic results is a fault in the way that the author coupled different parts of the STRUKA database. For the author to apply the algorithm to the STORM subbranches required the export of data from the STRUKA tool to a GIS attribute file for each parameter separately. These attribute files were subsequently imported in EXCEL for database- computation. The coupling between the Rhine km's of the STRUKA database and the GIS attribute files of STRUKA were supposed to match (respectively field Gewässer to field Gewässer and field Abschnitt to field Nr.). If this coupling has been done wrong, the ecotope distribution in STORM in Rhineland Palatinate and Hessen is a fictive one. However, the ecotope distribution seemed realistic to serve as an ecotope distribution in the role play. Therefore not further efforts have been made to verify if the coupling had been done wrongly.
STORM-Rhine, Main Report – Design of the Roleplay Page 50
Table 7.2 Translation of Gewässerstrukturgütekartierung of Rhineland Palatinate to ecotope distribution in STORM Rhine STORM Rhine (RES) STRUKA V
River bank e02 Natural river bank 5.2 Uferverbau kein Verbau Floodplain e04 River dune guessed e05 Hardwood floodplain 6.1 Flächennützung bodenständiger Wald forest e06 Softwood floodplain guessed: part of hardwood floodplain forest forest e07 Production forest 6.1 Flächennützung nicht bodenst. Wald e08 Herbaceous floodplain guessed e09 Back swamp manually estimated using Rhine atlas e10 Natural levee guessed e11 Natural grassland 6.1 Flächennützung Extensivgrünland e12 Production grass 6.1 Flächennützung Intensivgrünland e13 Permanent flowing side 1.1 Laufform mit Nebengerinnen channel e14 Periodical flowing side manually estimated using Rhine atlas channel e15 Isolated oxbow lake manually estimated using Rhine atlas e16 Gravel pit lake 6.1 Flächennützung Abgrabung e17 Crop cultivation 6.1 Flächennützung Acker e18 Building area 6.6 Überflutungsfläche Bebauung e19 Camping 6.1 Flächennützung Park, Grünanlage e20 Recreational harbour manually estimated using Rhine atlas e21 Earth ramp manually estimated using Rhine atlas e22 Industrial harbour manually estimated using Rhine atlas
7.6 Germany target distribution
For Germany a 'target' ecotope distribution was not readily available. To assess what a realistic 'target' ecotope distribution could be, the 'present' ecotope distribution was multiplied with a matrix that distributed the percentages of a certain ecotope into percentages of other ecotope(s). Similar to Pedroli and Rademakers, the 'building area', including the industrial harbours, remained untouched.
Table 7.3 shows the matrix.
To determine the suitability of the matrix approach, the method has also been applied to the Dutch region and has been compared with the target distribution by Pedroli and Rademakers (1995). In Appendix A the result of this comparison are shown. As expected, the method is not correct in simulating the target distribution of Pedroli and Rademakers. However, for the individual ecotopes the average simulated is the about the same as the average of the target distribution of Pedroli and Rademakers.
STORM-Rhine, Main Report – Design of the Roleplay Page 51
Table 7.3 Matrix to transform current ecotope distribution to target ecotope distribution
nnel st
e
r ain n fore u o g sid k lake ing side cha oodpl dplain forest ass l w odplai rest
on w ban ank
r o
b ea ati flo r p us f x v ver ne amp ent flowin l eo tional harb cal l ri w ng e c ction g ction fo a a n el pit lake culti v n p cre ck s a a h Natur Stone river b River du Hardwood flo Softwood floo Produ Herba B Natural levee Natural grassland Produ Perman c Periodi Isolated o Gra Cro Building a Campi Re Earth ram Current use Target use Eco- e02 e03 e04 e05 e06 e07 e08 e09 e10 e11 e12 e13 e14 e15 e16 e17 e18 e19 e20 e21 tope e02 1 1 e03 0 e04 1 0.1 e05 1 0.5 0.2 0 0.1 e06 1 0.5 0.3 0 0.2 e07 0 e08 1 0.4 0.2 0 0.2 e09 1 0.5 1 e10 1 e11 1 0.4 0.3 0 0.3 1 e12 0 e13 1 0.5 0.3 0.2 0 0.2 e14 0.5 0.7 e15 0 0.5 e16 0 e17 0 e18 1 e19 0 e20 0 e21 0
7.7 Recommendations
In Germany the Bundesanstalt für Gewässerkunde uses an alternative method to evaluate the ecological performance. This method is called the Gewässerstrukturgütekartierung. For Rhineland Pfalz a classification based on this method was available in time for this report. The method is also available for North Rhine Westphalia since November. At the time of writing The Netherlands is applying this method on the River Rhine in The Netherlands. In comparison with the RES-method, the methodology of Gewässerstrukturgütekartierung has a more transparent relation between ecological performance and the geometry and morphology of the river. This method will probably become more influential in the international discussion on river management. Therefore the use of this methodology should be considered in a next phase of development of STORM Rhine.
For the German part the areas occupied by industrial harbours were identified, but in this version of STORM Rhine, industrial harbours have the ecotope nr. 18 'building area'. This information could in a next phase of STORM Rhine be used to make a kind of 'stepping stone'- function for harbours, as one of the constraints to the growth of transport by ship is the availability of harbours (De Vries, 2000).
STORM-Rhine, Main Report – Design of the Roleplay Page 52
7.8 Test of matrix approach to determine target ecotope distribution
e04 River dune e11 Natural grassland 10 50 40 30 5 20 10
0 0 0510 0 1020304050
e05 Hardwood floodplain e13 Permanent flowing forest side channel 30 10
20 5
pe % 10 to
t eco 0 0 0510
esen 0102030
Pr ο e06 Softwood floodplain e14 Periodical flowing side forest channel 20 20 e % or 15 otop c 10 10
target e 5 ted 0 0 0 5 10 15 20 01020
Simula • Target land cover % of total, acc. to Pedroli and Rademakers (1995)
STORM-Rhine, Main Report – Design of the Roleplay Page 53
e08 Herbaceous floodplain e15 Isolated oxbow lake
50 5 40 30
20
e % 10
otop 0 0 0 1020304050 05 nt ec
e
s
e09 Back swamp or Pre • 10 % e p oto 5 rget ec d ta 0 0510
Simulate Target land cover % of total, acc. to Pedroli and Rademakers (1995)
STORM-Rhine, Main Report – Design of the Roleplay Page 54
STORM-Rhine, Main Report – Design of the Roleplay Page 55
8 FLOOD DAMAGE MODULE
The damage through inundation when a main embankment fails has been modelled based on the following assumptions. The potential areas to suffer from flooding are taken to be identical to the areas which according to the IKSR (19..) Rhine Atlas suffered from the 1926 flood. Per river stretch as used in the STORM model the total area potentially inundated were identified for the left and right banks with a division between 5 main damage categories: • Residential • Commercial • Infrastructure • Agriculture • Nature
The cost of damage figures used in STORM are given in Table 1 and considered equal (per category) for the whole catchment. These figures do not take into account incremental damage from progressive depth of inundation and are simplified from DWW 2000, DWW 2001 and NRW 2000.1
Table 8.1 - Unit flood damages Cost of flood damage Type of area Million Euro/km2 Residential 150 Commercial 200 Infra 100 Agriculture 50 Nature 0
The actual figures used for the surface areas and division of categories under threat in case of embankment failure is given per river stretch for the Left and Right banks in Table 8.2. These river stretches are in fact, as with the Hydraulic module, subdivided into sub-stretches for a somewhat more geographically specific calculation of the damage. The figures in Table 2 are sums (for the surface area and averages for the categories)
Table 8.2 Areas and land use classes of damage in case of overtopping winter dike Left Bank, percentages of area (%) Right Bank, percentages of area (%) ) 2 ) 2 (km (km k ture ture ce ce u ntial ntial
er er truc trc e e side side Re Comm Infras Agriculture Natur Section Left bank Re Comm Infras Agriculture Natur Right Ban U1 128 13 13 13 45 17 170 18 18 13 17 35 U2 133 17 17 11 46 9 257 10 10 10 37 33 U3 18 15 15 17 30 23 9 15 15 30 0 40 M1 7 7 7 67 10 10 7 7 7 67 10 10 M2 15 14 14 40 14 18 16 14 14 44 16 12 M3 15 32 32 12 24 0 11 20 27 17 22 14 L1 61 24 24 20 22 10 53 31 31 22 12 5 L2 400 22 22 20 28 9 82 21 21 19 26 14 L3 198 11 11 10 48 20 300 10 10 10 50 20
1 Dienst Weg- en Waterbouwkunde, 2000. Schade na overstroming. Ministerie van Verkeer en Waterstaat, Directoraat- Generaal Rijkswaterstaat. Dienst Weg- en Waterbouwkunde, 2001. Standaardmethode Schade en Slachtoffers als gevolg van overstromingen. Ministerie van Verkeer en Waterstaat, Directoraat-Generaal Rijkswaterstaat. Nordrhein- Westfalen, 2000. Potentielle Hichwasserschaden am Rhein in NRW. Ministerium fur Umwelt, Raumordnung und Landwirtschaft.
STORM-Rhine, Main Report – Design of the Roleplay Page 56
Left Bank, percentages of area (%) Right Bank, percentages of area (%) ) 2 ) 2 Residential Residential Commerce Infrastrcuture Agriculture Nature Section (km Left bank Residential Residential Commerce Infrastructure Agriculture Nature Right Bank (km R11 36 5 5 10 20 60 50 15 15 10 60 0 R21 33 15 15 10 60 0 50 15 15 10 60 0 R3 70 10 10 7 40 33 11 13 13 7 0 67 R4 210 11 11 8 45 25 871 13 13 9 39 26 W1 254 15 15 10 48 13 153 14 14 10 63 0 W2 243 15 15 10 60 0 160 15 15 10 60 0 Y1 223 14 14 12 60 0 345 15 15 10 60 0 Y2 252 5 5 7 50 33 216 6 6 7 48 33
For the purpose of the simulation it is assumed that the adjacent areas inland from the river embankments are separated from each other by dikes perpendicular to the river per sub-stretch. This means that if the embankment fails in one particular sub-stretch it does not automatically mean that the adjacent sub-stretch is affected as well. However, since measures taken to counter the threat of flooding are taken per main river stretch, in practice it may well prove that the embankment along a complete river stretch is likely to fail in one go. This is not necessarily so for adjacent “main” river stretches, subject to measures taken by the actors. An exception to this independence of flood threats is the dike rings which are primarily found in the Dutch part of the catchment. Here, if one part of the dike ring fails, the whole area will be liable to flooding. Table 8.2 shows the subdivision in flood- threatened areas per river stretch. For the total damage, the different areas comprising an area surrounded by a dike ring must be totalled.
STORM-Rhine, Main Report – Design of the Roleplay Page 57
9 INSTITUTIONS
9.1 German Water Management
German river management is embedded in the federal political, legislative and administrative structure. This means that there is a very high level of autonomy in decision-making at the individual State level (German: Länder) as guaranteed in the federal constitution. This means that at the federal level water management is only dealt with in broad framework legislation. At the State level the federal framework is worked out into specific water policies and institutional water management mechanisms. Local authorities (Kommunen) within the States are accorded municipal autonomy in dealing with the local environment and providing for those services necessary for the adequate living conditions within their territories and may choose freely as to which institutional arrangement to apply for water management. This can be implemented individually or in combination with other municipalities. In addition many States have opted for an extra regional government level (Bezirksregierungen or Regierungspräsidien) between the State and the municipalities.
The States have formed a interstatal body LAWA (Länderarbeitsgemeinschaft Wasser) to harmonize and coordinate intergovernmental cooperation. This body also takes a leading role in dealing with policy development by the European Union.
Given the constitutional guarantees for the high level of autonomy at each level, this three or four tier structure is not explicitly hierarchical, but better characterized by its division of responsibilities. The federal level, which establishes framework legislation, represents the states at the European level, but can at all times be held accountable by the individual States for its actions. The States and regional governments take a leading role in the policy development and legislation. The municipalities play a leading role in the execution and enforcement of water management legislation, although in many cases the actual work is done by specialized public authorities, agencies and associations.
At all levels there is extensive horizontal cooperation between the responsible institutions formalized in treaties and contracts. Table 9.1 gives an overview of the institutional levels and cooperation with regard to mandates and objectives for water resources protection and management.
Table 9.1 Water resources management framework in Germany (after Nunes Correia, ) Institution Mandates & Legislation Objectives Federal level Transposition of European Harmonisation of water protection; trans- legislation to Federal framework boundary water supply, sewerage and laws water resource management; administration of federal waterways Co-operation between Marine protection; monitoring Federal & State level programmes Treaties & co-operation Framework for trans-boundary Trans-boundary water supply, sewerage between States inter-municipal associations; and water resources management; Establishing trans-boundary water Harmonising of legislation and management institutions; implementation; Co-ordinating specific management tasks State Institutions River basin management State Parliaments & Transposition of European Laws; Governments State water laws Municipalities Local statutes and bye-laws Water supply, sewerage and water resources management; Flood control; Drainage; Dyke maintenance; Management of water courses not under the control of the Federal or State level or a water association Inter—Municipal Statutes Water supply and sewerage Associations Water Authorities & Statutes Implementation of federal legislation; Agencies Management of large water courses; Collecting effluent charges; Collecting abstraction charges; Monitoring and enforcement; Data and information
STORM-Rhine, Main Report – Design of the Roleplay Page 58
9.2 Water Management in The Netherlands
9.2.1 Description
Water resources management in the Netherlands is primarily a government task. Responsibilities are divided between the various levels of public administration. The administrative structure of water management in the Netherlands consist the following levels: the national, the provincial and the local.
National waters, i.e. the main rivers, lakes, as well as coastal waters are the direct responsibility of the Ministry of Transport, Public Works and Water Management (V&W: Verkeer en Waterstaat), in particular the operational department Rijkswaterstaat. It prepares national policy on flood protection and water management and supervises the implementation of the water policy by provinces and water boards. Legislation in the field of water can often only be enacted in co-operation with other ministries because responsibilities overlap. A close relation exists with the ministries of Housing, Spatial Planning and the Environment (VROM) and of Agriculture, Nature Management and Fisheries (LNV). Besides these ministries tuning is required with the ministries of Economic Affairs (EZ) and of Foreign Affairs (BUZA).
The province is the level where most of the vertical and horizontal co-ordination of government is concentrated. The provinces can make their own policies but will also receive policies and directives from the central government and ensure that they are carried forward to municipalities and water-boards for implementation. In the same way the provinces pass the interests of lower governments to the central government. An important task of the provinces moreover is to co- ordinate the policies of the different sectors of government, such as water management, environment, nature conservation, housing, physical planning and transport. Most provinces have delegated the majority of the surface water quantity and quality management tasks to the water-boards (in accordance with the Water-boards Act.
The task of the provinces is the quantitative and qualitative groundwater management. The daily water management department called the Provinciale waterstaat executes this groundwater management.
The local level consists of the water boards, the municipalities and drinking water companies. Through mergers both water boards and drinking water companies now cover geographic areas generally at a scale somewhat smaller than a province, but often much larger than individual municipalities.
Responsibilities for Dutch water management are shown in Table 9.2.
Table 9.2 Water resources management in The Netherlands (after Bijman, 2002) Authority Competence’s Tasks National level Responsible for efficient executing of the Management of national surface waters entire care of water management Flood control Enacts policy documents on water Quantitative and qualitative aspects of management water Direct responsibility for legislative policy Concern for navigation and harbours concerning water Concern for roads and cross river-, Navigation channel connections Responsible for safety against water Data collection and research (dikes)
STORM-Rhine, Main Report – Design of the Roleplay Page 59
Provincial level Responsible for non state waters Responsible for the quantity and quality Formulates Provincial Policy Document on management of non state waters1 Water Management Provides permits for water extraction or Formulates Provincial Policy Document on recharge and waste disposal on non state Groundwater management waters1 Formulates ordinance for creating water Sets up levies for disposals on non state board and their tasks waters1 Formulates ordinances on water Responsible for groundwater quantity and management quality Supervises water boards and municipalities Provides permits with regard to Executes administrative force1 groundwater Sets up levies with regard to groundwater Water boards Responsible for non state waters Responsible for the quantity and quality Formulates Management Plan Remaining management of non state waters Waters Provides permits for water extraction or Formulates ordinances on water recharge and waste disposal on non state management waters Applies administrative force to execute Sets up levies for disposals on non state legislation, Amvb’s or provincial ordinances waters Sets up water board taxes Responsible for water-control structures Purification of waste water
Municipal level Responsible for sewerage Construction and management of sewers Responsible for small municipal waters Responsible for small municipal waters Formulates ordinances for sewerage Provides permits for the construction of or management operations on hydraulic engineering Formulates Land Use Allocation Plan2 structures
1 These tasks are mostly mandated to the water boards 2 The Land Use Allocation Plan is for spatial planning in the municipality, which can be of importance for groundwater management. It also determines for which operations to hydraulic engineering structures need a permit of the municipality.
9.3 Institutional Complexity of Interdepartmental and Inter-Agency Interaction
The actual involvement of various government agencies at the different levels of government is in both countries much more complex as suggested above. Especially horizontally across various ministries and departments overlaps in responsibilities exist and policies from one agency affect the area of responsibility of other government agencies. In both countries more and more emphasis is given to integrating planning and decision-making. This level of detail is however beyond the immediate scope of the STORM project, but described in detail in the attached Bijman report and in the literature (e.g. Nunes Correia).
9.4 Comparison between German and Dutch Water Management Institutions
A comparison of German and Dutch quantitative aspects of water resources management is presented in Table 9.3 (modified after Bijman).
Table 9.3 Water Quantity Management Germany The Netherlands Federal Level National Level Strategic and operational responsibility for state Strategic and operational responsibility for state waters to guarantee navigation waters Responsible for flood control State Level Provincial Level Strategic responsibility for all first order waters Strategic responsibility for regional surface waters (These tasks can be delegated to the Upper State Strategic and operational responsibility for ground Authority) waters
Regional Level Water-board Strategic and operational responsibility for all waters, Operational responsibility for regional surface waters except state waters in respect to navigation Provides permits for water extraction Provides permits for water extraction and drainage Responsible for water-control structures
STORM-Rhine, Main Report – Design of the Roleplay Page 60
Municipal Level Municipal Level Provides permits for water extraction and drainage for Operational responsibility for small waters smaller uses Water Association Can have responsibilities in water extraction and drainage
Table 9.4 River, Floodplains & Facilities Management Germany The Netherlands Federal Level National Level Construction of river banks and bank pavements Construction of dikes, roads, cross river-channel Provides permits for constructions that can influence connections and harbours navigation along state waters State Level Province
Regional Level Water-board Responsible for construction and maintenance of Responsible for water-control structures winter dikes Construction of water purification installations Appoints contours of flood zones Municipality Municipality Provides permits for construction and use of Construction and maintenance of sewerage treatment constructions in floodplains facilities Maintains floodplains of all order waters Provides permits for construction and use of Responsible for construction and maintenance of constructions summer dikes Water association Can have responsibilities in the construction and maintenance of summer dikes
9.5 Application in the Roleplay
Within STORM Rhine the river and floodplain management process is attributed to eight organizations. Five organizations represent the upstream country, Germany, two the downstream country, the Netherlands and one is an international non-governmental organization. The organizations are listed in Table 9.5, together with their main objectives, criteria and area of interest.
Figures 9.1 and 9.2 show the map of the river and floodplains branches and the areas of responsibility for the different organizations.
Table 9.5 Roles in STORM Rhine Organization Objectives Criteria Area of Interest Federal • Guarantee safety from flooding • Maximum water level in river U 1-3, L&R Government for residential and commercial • Unsafe dike length M 1-3, L&R Upstream interests • Tonnage shipped L 1-3, L&R • Maximize navigation on river • Shipping constraints • Ensure renaturalization of floodplains Regional • Guarantee safety from flooding • Maximum water level in river U1L – L3L Government for residential and commercial • Unsafe dike length Left Bank interests • Area for Recreation Upstream • Develop river-based recreational • Volume excavated facilities • Area under cultivation • Income from excavation • Area returned to nature • Reduce Agriculture in favour of Nature development Regional • Guarantee safety from flooding • Maximum water level in river U1R – L3R Government for residential and commercial • Unsafe dike length Right Bank interests • Tonnage shipped, Shipping Upstream • Maximize navigation on the river constraints • Maximize agricultural output • Area under cultivation • Promote nature development • Area for recreation together with retention and • Area for nature
STORM-Rhine, Main Report – Design of the Roleplay Page 61
Organization Objectives Criteria Area of Interest recreation Urban • Maximize navigation on the river • Tonnage shipped U1R, U3L Municipalities • Optimize commercial and • Shipping constraints M3L, M3R Upstream residential construction in • Area for construction L1L, L1R, L2R floodplains • Unsafe dike length • Protect against flooding Rural • Maintain agricultural activities in • Area under cultivation U1L, U2L, U2R, Districts floodplains • Length & height of summer U3R Upstream • Maintain function of summer embankments M1L, M1R, M2L, dikes • Subsidy from higher M2R • Financial compensation for government L2L, L3L, L3R renaturalization National • Guarantee safety from flooding • Maximum water level in river R1 – R4, L&R Government for residential and commercial • Unsafe dike lengths W1 – W2, L&R Downstream interests (without heightening • Tonnage shipped Y1 – Y2, L&R dikes) • Shipping Constraints • Maximize navigation on river • Area under cultivation • Safeguard / increase • Area returned to nature navigational capacity • Renaturalization of floodplains Municipalities • Guarantee safety from flooding • Maximum water level in river R1 – R4, L&R Downstream for residential and commercial • Unsafe dike lengths W1 – W2, L&R interests • Area under cultivation Y1 – Y2, L&R • Maintain agricultural activities • Area returned to nature • Maintain function of summer • Area for construction dikes • Financial compensation for • Length & height of renaturalization and removal of summer dikes buildings from floodplains • Area of retention areas • Oppose new retention and emergency flooding areas within dike-rings Environmental • Renaturalization of river • Area returned to nature Whole area Interest Groups • Removal of summer • Length of summer dikes embankments removed • No growth of navigation industry • Area removed from • Oppose construction in construction purposes floodplains • Area under cultivation • Promote shift from intensive • Area for recreation agriculture to extensive meadow and range-lands • Promote eco-tourism instead of mass recreation
STORM-Rhine, Main Report – Design of the Roleplay Page 62
Figures 9.1 and 9.2 Denomination of river stretches
DOWNSTREAM
Right Bank
Left Bank UPSTREAM