D 2.5.2 – Identification of GIS applications in cities (applied in commercially available GIS software) The case of Eindhoven

D 2.5.2 b – Identification of GIS applications in cities (applied in commercially available GIS software) The case of Eindhoven

© 2010 PREPARED The European Commission is funding the Collaborative project ‘PREPARED Enabling Change’ (PREPARED, project number 244232) within the context of the Seventh Framework Programme 'Environment'.All rights reserved. No part of this book may be reproduced, stored in a database or retrieval system, or published, in any form or in any way, electronically, mechanically, by print, photoprint, microfilm or any other means without prior written permission from the publisher

Content 1  ‘Flood risk’ mapping GIS applications. Case Studies Eindhoven, The Netherlands (KWR) ...... 5 

1.1  Pluvial flood risk assessment ........................................................................................................ 5 

1.1.1  Introduction ............................................................................................................................. 5 

1.1.2  Background.............................................................................................................................. 5 

1.1.3  Applied GIS Methods ............................................................................................................ 6 

1.1.4  Input data ................................................................................................................................. 7 

1.1.5  Case study results ................................................................................................................... 7 

1.2  Linking hard or paved surfaces to sewer networks .................................................................. 9 

1.2.1  Introduction ............................................................................................................................. 9 

1.2.2  Background.............................................................................................................................. 9 

1.2.3  Applied GIS Methods ............................................................................................................ 9 

1.2.4  Input data ............................................................................................................................... 10 

1.2.5  Implementation ..................................................................................................................... 11 

1.3  Smart Interpolation of groundwater levels ............................................................................... 11 

1.3.1  Introduction ........................................................................................................................... 11 

1.3.2  Background............................................................................................................................ 11 

1.3.3  Applied GIS Methods .......................................................................................................... 12 

1.3.4  Input data ............................................................................................................................... 13 

1.3.5  Evaluation and future research .......................................................................................... 14 

1.4  References ...................................................................................................................................... 14 

1.5  Sitography ...................................................................................................................................... 14 

1 ‘Flood risk’ mapping GIS applications. Case Studies Eindhoven, The Netherlands (KWR)

In this chapter three commercially available GIS applications will be discussed. These applications assist the municipality of Eindhoven in the assessment of flooding risks caused by heavy rainfall or high groundwater levels. Extremely heavy rainfall may cause flooding of streets, pavements and buildings (basement and ground level) in Eindhoven and many other cities. To make a better risk assessment of areas susceptible to flooding a GIS application has been developed; the Flooding Landscape Map (FLM). Based on various FLM scenarios measures can be undertaken to prevent flooding in prioritized areas. Municipalities need a reliable tool to plan for extension and maintenance of the sewage network. For a realistic modeling the unpaved and paved surfaces have to be linked correctly to the nodes in the network. GIS procedures have been developed to maintain the reliability of the model. Areas which are vulnerable to flooding by occasionally high groundwater may be mapped using GIS and a groundwater model. By combining both systems smart interpolation of groundwater levels can be carried out.

1.1 Pluvial flood risk assessment

1.1.1 Introduction Extremely heavy rainfall may cause flooding of streets, pavements and buildings (basement and ground level) in Eindhoven and many other cities. To make a better risk assessment of areas susceptible to flooding a GIS application has been developed; the Flooding Landscape Map (FLM). This application is based on an accurate Digital Elevation Model (DEM, 1x1m) of the city combined with detailed topographic maps. Local rainfall data is used to model extreme events. The FLM reveals areas, streets and buildings with high flooding risks and calculates flood levels and water flow direction. Based on various FLM scenarios measures can be undertaken to prevent flooding in prioritized areas.

1.1.2 Background Climate change will have a considerable impact on urban areas. One of the effects of the global warming for Western Europe is the increase of storm water peak intensities during rainfall and another is likely to be an increase in the frequency of these showers. These extreme events with duration of up to several hours cause the urban storm water drainage system to be overloaded, either due to the intensity of the rain or due to runoff from urban green space which normally infiltrates into the ground. When the sewer discharge capacity is exceeded, storm water will flood the streets or even buildings and homes. At the same time, it creates a possible health risk. Other reasons for more frequent problems due to extreme rainfall in Dutch cities are the increase of paved surfaces and the recent changes to the street profile, for example shopping areas without any kerbs and without storage capacity at street level.

This GIS tool focuses on the problems due to extreme rainfall which accumulates and flows at street level. It excludes other causes of flooding, such as high surface water levels or collapsing sea and river levees.

Figure 1-1: Extreme rainfall in the municipality of Eindhoven (august 2011) caused a large number of streets

to flood. It is becoming commonly accepted that these kind of problems need to be solved by providing more space for water at ground level. The European Flood Risk Directive [EU, 2007] promotes flood risk management plans with non-structural measures aiming at resilience of urban infrastructures and preparedness of the social system. There is a growing consensus that sewer capacity is limited and that there is a need to consider all aspects of water drainage during extreme rainfall events. The drainage system is designed at a certain drainage standard. In the Netherlands the general standard is that all rain is drained without water on the streets for an extreme storm event with a return period of 2 years (20 mm in one hour). This is achieved with technical facilities such as sewers. Next to the functioning of sewers during standard rainfall events, also the functioning during far more extreme rainfall events should be evaluated and measures to reduce flooding should be evaluated. These measures will mainly consist of above ground measures and require spatial and urban planning. The measures to reduce problems due to flooding should also contribute to a well functioning urban living environment: e.g. it is more likely that a low laying area for storing water will be accepted and be a success if it is also set up as a nice park or a square. The main problem is how to convince other professionals, concerned with urban planning and maintenance, of the importance of space for water. To support this discussion, Tauw bv, a Dutch engineering company, has developed a GIS-based method for mapping urban storm water. This urban storm water flooding map shows in detail the water flow paths during extreme rainfall events, as well as highlighting the depressions in the urban landscape where water pools and illustrating the overall extent of the problem. The model and its resulting maps make it possible to identify potential and observed locations of flooding, and implement and test the effects of mitigation measures at street level.

1.1.3 Applied GIS Methods The first step in the modeling process is removal of any “noise” from a Digital Elevation Model. This is needed because of the airborne laser (LIDAR) mapping method which is used to obtain the

elevation data. The laser data bounces back from the surface, but also from treetops or parked cars. This is undesirable, because it gives an impression that the water cannot flow into another direction while in reality it can. To remove those obstacles a semi-automatic method has been set up. All points within a certain frame which are higher than the average surface level plus a threshold are simply removed. Afterwards a check for the situation around dikes, elevated roads and bridges needs to be made. But this is necessary anyway. These steps may be skipped when the newer elevation data in the Netherlands (AHN2) is available, since trees and cars are supposed to be already filtered out. A careful assessment of the elevation map will still be needed in order to decide on how to deal with bridges and roads crossing at different levels. After the surface map has been improved, a land use map layer (accounting for impervious and paved surfaces) is used to define from which surfaces water will flow off. At this point it is possible to assume that only water from impervious or paved areas will be available for run off (assuming that the water from the other areas will infiltrate). It is however also possible to assume that from green areas a certain percentage of the rainfall will be available for runoff. Once the model (ArcGIS with Spatial Analyst and ModelBuilder) has been run, the generated output needs to be presented in such a way that flow direction and water accumulation are readily identifiable. Measures to reduce flooding can be implemented in the model and the effects of these measures can be simulated in a second phase of modeling.

1.1.4 Input data The urban storm water model is a GIS-based model, which doesn’t require a lot of data. However, the required data needs to be as accurate as possible. Accurate DTM of the Netherlands is available with average 1 point per m2 (AHN1) and partly already 10 points per m2 (AHN2) [AHN, 2009], providing a solid base for this GIS-model. To solve the inconvenient situation of flooding due to an overloaded sewer system, modifications at ground level can be implemented into the DTM to redirect water to less vulnerable areas, where it can be drained or temporarily stored. Before measures to efficiently and effectively guide the water away from inconvenient locations can be made, the ground level of the urban area needs to be mapped properly. Other input for the model includes: - Land use: Impervious and paved surface map, surface water and green areas. - External input to support the review of the model output and complete the urban storm water map includes registration of any known data about flooding in the urban area. This data can be obtained from registered complaints made by residents and from specific knowledge of the involved professionals.

1.1.5 Case study results The model presented is a useful tool in finding solutions to prevent flooding, even far upstream from the actual location of the risk areas. Measures that can control and redirect the water flow can be investigated and used to support a case for tackling the actual problems at its source in order to prevent vulnerable areas downstream from flooding. The model setup is deliberately simple in order not to raise high expectations of accuracy. It is more important to “get the feeling” and visualization of the situation and to initiate discussions, than to correctly compute the overland water flows in a complicated 3-dimensional sewer model. The accuracy of more fundamental approaches is limited to the accuracy of the input data.

It is expected that in complex or specific cases more detailed simulations will be needed for a good representation of the actual discharge. The GIS-analyses will provide the first insight in which areas are vulnerable to flooding and which solutions are feasible. However, up to now these first analyses appeared to be so important and giving so much direction and possibilities that a more detailed modeling has not been started yet. Figures 3-2 and 3-3 show results for the case study in Eindhoven. A commonly used rain peak in this model is a T=100 years rainfall event. This is 60 mm in one hour, of which it is presumed that 20 mm is drained by the sewer system. The sewer system is hereby considered overloaded and temporarily “out of function”. The remaining 40mm will be used as input for the calculation model, without infiltration into the ground and sewer system, and spread over the entire paved area of interest. This simplification provides a clear and relatively fast insight into how the storm water runs off at street level. Due to the assumptions named above, the amount of water which is stored at street level is exaggerated, as well as the length of the stream tracks; the catchment areas are possibly very large for an urban situation.

Figure 1-2: Output of the FLM model for the city of Eindhoven (red = high risk of flooding)

Figure 1-3: Detailed output of the FLM model for the city of Eindhoven (red = high risk of flooding)

The GIS-based method of modeling and mapping urban storm water flooding is an effective tool in deciding how to prevent urban storm water flooding in a cost effective way. The output of the model is generated in a manner which allows any specialist (of various disciplines) involved in determining measures at ground level, to comprehend the process. The urban storm water flood maps improve the communication between various disciplines, generating an environment for fast, successful and cost effective decision-making in prevention of urban storm water flooding.

1.2 Linking hard or paved surfaces to sewer networks

1.2.1 Introduction Municipalities need a reliable tool to plan for extension and maintenance of the sewage network. For a realistic modeling the unpaved and paved surfaces have to be linked correctly to the nodes in the network. In existing and new building projects some hard surfaces like roofs may be detached from the main sewer network and connected to a separate surface water drainage network. GIS procedures have been developed to maintain a correct coupling of hard surfaces to the sewer network model.

1.2.2 Background In 2006 the sewage network of the city of Eindhoven has been modeled in Sobek, a hydraulic modeling environment (Deltares). The main purpose was to have a reliable tool to plan for extension and maintenance of the sewage network. In 2006 a number of assumptions had to be made because detailed geographical data about the paved and unpaved surfaces were not available. These assumptions may have a great impact on the outcome of the modeling in Sobek. In 2009 more detailed data became available in the GIS databases of Eindhoven. Nelen & Schuurmans consultants were contracted to improve the input data of Sobek and to develop a GIS method for easy updating of these data. With this tool the reliability of the model may be improved, especially because locally detached surfaces (roofs) are taken into account.

1.2.3 Applied GIS Methods A toolbox called “Turtle Urban” has been developed in ArcGIS ModelBuilder by Nelen&Schuurmans. It consists of a number of Python scripts which incorporate generic ArcGIS

tools (e.g. overlays and spatial joins) as well as special modules for data input and output conversion. A number of subsequent tools may be launched in this ModelBuilder environment. Typically each tool has a user friendly dialog and help-function The procedure in “Turtle Urban” is illustrated in Figure 3-4 and the steps are explained below. For each step a separate ModelBuilder model has been developed. Step 1: Data conversion. The sewage network is imported into a geodatabase in GIS. Step 2: Topographic objects are classified in different classes ranging from completely pervious to impervious. The Dutch guidelines for sewage (C2100) are applied in this classification. Step 3: A map with “Sewage water units” is constructed to facilitate the correct links between the network nodes and the topographic objects (step 4). This map is derived from a DSM (Digital Surface Model, elevation data) and a land use map showing paved and unpaved areas. Per sewage water unit a certain percentage of the hard/paved surfaces can be detached from the main sewer system. This percentage is an important parameter in the hydraulic model. Step 4: In the geodatabase the correct link is established between each topographic object and the sewage network. The percentage of detachment (step 3) is stored in the geodatabase. More detailed information about local detachments (roofs, pavements) may be incorporated and manually adapted.

Figure 1-4: The GIS concept in Turtle Urban (Nelen&Schuurmans)

1.2.4 Input data The required input for the model includes: - The Sewage network in a geodatabase (ArcGIS) - Detailed digital topographic map of the urban area, scale 1:1000 (GBKN) - High resolution digital surface map, AHN in the Netherlands [AHN, 2009] and other LiDAR based, preferably with a resolution of at least 1 point per m2. - Land use: Impervious and paved surface map, surface water and green areas.

1.2.5 Implementation The Turtle Urban toolbox has been tested and implemented in the city of Eindhoven. Some minor adaptations were necessary to meet the specifications of the local sewer system. Also a small adaptation was necessary to establish a reliable Object-Network linkage procedure. Figure 3-5 shows an impression of the Turtle Urban toolbox and results.

Figure 1-5: The Turtle Urban toolbox (Dutch version) and visualization of results for a case in Eindhoven. (Nelen&Schuurmans)

1.3 Smart Interpolation of groundwater levels

1.3.1 Introduction Areas which are vulnerable to flooding by occasionally high groundwater may be mapped using GIS and a groundwater model. By combining both systems smart interpolation of groundwater levels can be carried out.

1.3.2 Background In some residential areas in Eindhoven the groundwater may occasionally be at such a high level that basements are prone to flooding. The current policy states that this is acceptable for a maximum of 10 days per year. More insight is needed in where this maximum is exceeded and where the current monitoring system of 150 monitoring wells must be adapted (sample density). Also the effectiveness of measures like artificial drainage systems and perforating clay layers in the subsoil needs to be evaluated. For this purpose a study has been undertaken to map vulnerable areas within the municipality.

1.3.3 Applied GIS Methods In order to obtain a spatial image of the highest occurring groundwater table in a specific year a spatial interpolation on the basis of measurements is used in combination with groundwater modeling. The steps are described below and the concept is illustrated in figure 3-6.

Figure 1-6: The concept of the applied procedure.

Step 1: Groundwater modeling The groundwater model had already been set up by Royal Haskoning for the management area (watershed) of the Dommel Water Board. The city of Eindhoven is located within this watershed. Non stationary model runs were necessary to depict the highest groundwater levels (2,5 % exceedance values). Groundwater model results are imported in ArcGIS as a GRID (raster). Step 2: Selection of groundwater level measurements For all available data points (150) the 2.5% exceedance values of the groundwater levels were determined. The differences between the measured and modeled levels are determined on a point by point basis in GIS using GRID sampling. Step 3: Interpolation of differences The differences between the calculated and measured values are then interpolated using IDW (Inverse Distance Weighted method in ArcGIS) to obtain a difference map (GRID, see Figure 3-7).

Modeled GWL

Measured GWL

Measured GWL

Interpolated GWL

Figure 1-7: Difference between measured value and calculated value (meetpunten = monitoring wells)

Step 4: Estimation of highest groundwater levels This difference map (see Figure X) is then used as a "correction map" in order to obtain the best possible spatial estimation of the "highest" annual groundwater level (see Figure 3-8).

Figure 1-8: Calculated groundwater [meters above NAP] in a wet situation

This method ensures that the differences in the data points between calculated and measured values are by definition zero. The reliability of the estimations depends on the spatial distribution and magnitude of the calculated differences.

1.3.4 Input data The required input for the model includes: - Groundwater levels – time series

meetpuntenmeetpunten

- Detailed digital topographic map of the urban area, scale 1:1000 (GBKN) - High resolution digital surface map, AHN in the Netherlands [AHN, 2009] and other LiDAR based elevation data, preferably with a resolution of at least 1 point per m2. Optional: - Soil and subsurface maps - Locations with artificial drainage systems

1.3.5 Evaluation and future research In this method the interpolated values are compared with the monitoring data and the spatial distribution of differences are used as a correction map. The study clearly indicated vulnerable areas and will be repeated and detailed if necessary. Additional model runs will be added to predict the effects of certain measures, especially artificial drainage systems and piercing the clay layer in the subsurface.

1.4 References Claessen. E.G., Kluck, J. (2009), Verslag workshop wateroverlast met touch table (in Dutch), Tauw Deckers. D. (2009), Turtle Urban Toolbox (project memo), Nelen&Schuurmans, March 2009 European Union (2007). Directive 2007/60/EC of the European Parliament and of the Council, of 23 October 2007, on the assessment and management of flood risks. Kluck J.1,3, Claessen E.G.1, Blok G.M.1, Boogaard F.C.1,2, Modeling and mapping of urban storm water flooding Communication and prioritizing actions through mapping urban flood resilience Geldof, G.D., Kluck, J.(2008), The Three Points Approach, ICUD 11, Edinburgh/Scotland 2008 bv, November 2009 Wal, van der. B.J. (2007), Analysis of groundwater levels in Eindhoven (in Dutch), Royal Haskoning, November 2007

1.5 Sitography SKINT North Sea Skills Integration and New Technologies ‘SKINT’ (http://www.skintwater.eu/),Boogaard, F.C, TU Delft, 2009 AHN Algemeen Hoogtebestand Nederland (http://www.ahn.nl) http://www.nelen-schuurmans.nl http://www.deltaressystems.com/hydro/product/108282/sobek-suite