Urban River Restoration, a Scenario for Copenhagen

Abstract

Urban flooding is an increasing hazard and expected to aggravate with climate change. The opening of a piped urban river in Copenhagen is explored to mitigate urban flooding as a result of a rising groundwater table. In the Copenhagen case study, a hydrological model has been used to calculate the effect of reopening a currently piped river to its natural environment as a nature-based solution (NBS). From the results of modelling scenarios, we have set up a simple damage function to calculate the avoided costs for the river basin by comparing difference in groundwater levels with and without NBS.

Stakeholders were engaged from the first stages of the case study. The integration of stakeholder’s knowledge in the co-design and implementation process of NBS to support complex decision-making processes was carried out by (1) participatory modelling activities to elicit and structure stakeholder’s risk perception, (2) mapping the interaction among decision-makers and stakeholders, and by (3) deriving Fuzzy Cognitive Maps (FCM) from Group Model Building.

The results show that by reopening the river, an economic benefit is obtained because the river now functions as a drainage channel which prevents flooding by groundwater of subterranean structures, notably housing cellars, and receives stormwater from cloudburst rainfall events. The FCM simulation showed that NBS implementation requires effective cooperation among different decision-makers to define potential interventions and to reduce the level of conflicts and to facilitate collaborative decision-making.

FormalPara Highlights

• Avoided damage to subsurface housing from varying shallow groundwater levels is simulated using a hydrologic model and simplified damage functions.

• Hydrological simulations of restoring an urban river as NBS scenario to mitigate damage resulting from groundwater flooding show promising results with 35% avoided damage

• Social Network Analysis shows that the implementation of the considered restored urban river NBS needs a thorough stakeholder dialogue to obtain high NBS effectiveness by detecting collaborative and institutional barriers and include trade-offs between co-benefits

17.1 Introduction and Characterization

Urban flooding is an increasing hazard and is expected to aggravate with climate change. To mitigate the risk for the urban surroundings, this chapter explores a nature-based solution (NBS) focusing on a potential restoration of an urban river (RUR). The restoration of a currently piped river back to its natural environment, is not only explored to mitigate urban flooding itself, but to create synergies with urban planning including climate change adaptation plans as well. In doing so, we look at quantitative and qualitative co-benefits.

The City of Copenhagen initiated its climate adaptive measures in earnest after hosting the COP 15 climate summit in December 2009. On 25 August 2011, the City Council adopted the Copenhagen Climate Adaptation Plan (City of Copenhagen 2011). This plan sets the framework for the implementation of climate adaptive measures in the City Administration area. Before disastrous flooding events in 2011 caused by a cloudburst, climate adaptation mainly focused on flooding related to storm surge, but then accelerated measures to mitigate urban flooding as well. This was translated in the adoption in 2012 of the Cloudburst Management Plan (CMP) (City of Copenhagen 2012) as an offshoot of the Copenhagen Climate Adaptation Plan from 2011. The CMP outlines the methods, priorities, and measures recommended for the area of climate adaptation including extreme rainfall. With this CMP, decisive steps forward have been taken to protect Copenhagen against high-intensity rain like the ones witnessed in August 2010 and especially in July 2011. Initiatives might include reopening streams, constructing new canals or establishing lakes and more green spaces, and using roads with high curbstones to lead the pluvial stormwater into confined spaces, notably surface water channels following a topographic gradient towards open waters.

Different nature-based solutions (NBS) are already implemented as part of the City’s cloudburst management plan, which will contribute to the RUR and are therefore included in this chapter. Such NBS relevant for the RUR NBS scenario include green areas, retention areas, decoupled rainwater from roofs and blue surface water routing from street level flooding to the RUR recipient and are implemented in the hydrological model described later in this chapter.

The complex decision-making processes, for making such NBS design and implementation, require the collaboration of a wide set of decision-makers and stakeholders. Lack of cooperation could result in barriers to the NBS implementation and/or affect their effectiveness. To detect and analyze these collaborative barriers Social Network Analysis was combined with Group Model Building, in order to identify barriers hampering the implementation of the selected NBS.

17.2 Case Study Area and Methodology

The Copenhagen demonstration case involves a test study of the benefits of implementation of the RUR based on a currently piped river or stream running through central parts of the city (See Figs. 17.1 and 17.2). The area along the piped river is 2–4 km long and the catchment covers an area of 80.000–120.000 m² depending on the solution chosen. The primary focus regarding analysis of the benefits related to restoring the river are mitigation of damages avoided from a rising groundwater table. Specifically, the shallow groundwater table in the Copenhagen area are located only a few meters below the groundwater surface. Therefore, it is tested, using an integrated hydrological model, how future climatic changes will affect the shallow groundwater table in a part of the Copenhagen area, both by doing nothing and by reopening the river to make it function as a drainage channel.

Fig. 17.1

Focus area, area where NBS are implemented (green circle). In the NBS area the streams are presently in pipes in the subsurface. In the NBS scenario, the streams are restored to a blue/green water infrastructure (blue line). On the right, the case study is depicted as part of the Greater Copenhagen catchment

Fig. 17.2

Bispeengbuen motorway ramp, suggested to be partially removed to create place for residential and green areas, road tunnels and climate adaptation solutions

The hydrologically based physical methodology used in this case study builds on simulating shallow groundwater with a hydrological model, which shows that the groundwater table under the current conditions and future conditions (see Table 17.1) may rise above surface and become surface inundation and adds to the stormwater flooding from e.g. cloudburst events.

Table 17.1 The scenarios conducted in this study

Furthermore, economic analyses have been carried out to assess the avoided costs by means of the Restored Urban River (RUR) NBS scenario. Avoided costs are considered as the difference between the benefits of the RUR and the costs caused by groundwater flooding without RUR as a result of both cloudburst events by spilling over of excess water from street level and pushed up water from sewage systems to subterranean building infrastructure (e.g. cellars) and a rising groundwater level.

A simple damage function has subsequently been set up to calculate the avoided costs for the river basin.

17.2.1 Hydrological Model

The hydrological effects of the restored urban river in terms of its ability to divert not only rainwater but also groundwater in contact with buildings and critical infrastructure are simulated with a hydrological model, MIKE SHE/HYDRO. MIKE SHE/HYDRO is a physically based fully integrated groundwater surface water model covering the primary aquifer for the Greater Copenhagen area and large parts of Northern Zealand. The set-up of the physical environment and geology hydrological model is described in detail in Pengal et al. (2017), van der Keur et al. (2018) and LeCoent et al. (2019).

Groundwater levels are calculated for a number of deeper layers but only the most upper groundwater table is relevant for the estimation of damage level to urban buildings.

The spatial extension of the model area is in the order of magnitude 5 × 5 km for the entire urban area which is shared by the municipalities of Copenhagen and Frederiksberg (shown in Fig. 17.1) and includes the Bispeengen motorway ramp which in the considered scenario is replaced by road tunnels, residential and green areas and space for climate adaptation, Fig. 17.2.

17.2.2 Assessment of Management Alternatives Using Model Scenarios

To assess the effectiveness of NBS in reducing groundwater levels, we compared the NBS with the business-as-usual (BAU) alternative under different climate scenarios. We defined BAU as the present situation with a piped urban river and no cloudburst management plan projects implemented. The NBS alternative is the one of the Restored Urban River scenario (RUR): the piped urban river is restored and interaction with groundwater enabled. A flow rate of 300 l/sec is targeted to sustain a stable and ecological flow. The restored urban river is fed through the three water sources shown in Fig. 17.3 (hydraulic contact, groundwater pumping and decoupled water).

Fig. 17.3

Water sources to the restored urban river NBS: (1) from groundwater through hydraulic contact with the river; (2) groundwater pumping, shallow and deep; (3) decoupled water from residential areas (e.g. roofs) and other paved areas (e.g. roads)

The performance of both alternatives is compared under two climate scenarios: the reference climate, 1981–2010 and the future climate 2071–2100 using the RCP8.5 projection (Collins et al. 2013). The scenarios are also shown in Table 17.1.

The Restored Urban River (RUR) scenario receives water from several sources, shown in Fig. 17.3:

• 1. groundwater through urban river interaction (hydraulic contact) with (shallow) groundwater resources

• 2. pumped groundwater from areas within the urban river catchment that experience rising groundwater, including from areas where water is infiltrated locally (Sustainable Drainage Systems, SuDS)

• 3. decoupled water from roofs and paved areas which otherwise would have been routed to the combined drainage-sewer system (combined sewer system)

17.2.3 Assessment of Avoided Damages

The scope of the economic analysis for the NAIAD Copenhagen NBS is to assess the avoided costs by means of the RUR NBS scenario under two climate scenarios with reference periods 1981–2010 and a future period 2071–2100. In Table 17.1 the scenarios are listed.

The damage levels are calculated as 30-year maximum groundwater level for the periods 1981–2010 (reference period) and 2071–2100 (future period). The generated 30-year maximum values are spatially distributed in a 100 × 100 m raster dataset relative to the surface. Then, the dataset is analysed for simulated values above the defined threshold of 2 m below surface, e.g. a number of raster grid cells with a 30-year maximum level above 2 m below surface. This dataset is then applied in GIS analysis to find areas above the defined groundwater thresholds for building damage with (RUR) and without (BAU) implementation of the NBS and with and without the hydrological effect of climate change (as prognosed by an 8.5 RCP scenario (Collins et al. 2013)).

Avoided costs are considered as costs caused by groundwater flooding as a result of both cloudburst events by spilling over of excess water from street level and pushed up water from sewerage systems to below ground building infrastructure (e.g. cellars) and rising groundwater level. Other costs as life cycle costs and opportunity costs as well as an assessment of co-benefits of the RUR – NBS scenario have not been included in the analysis.

The RUR scenario is compared to the BAU scenario with respect to shallow groundwater level and avoided damage is then calculated as the difference in affected area in which the shallow groundwater level causes damage (see Fig. 17.4). The damage caused (total damage) is estimated to € 68/m² (COWI 2014) and implemented in the adopted damage function. The damage function is expressed as:

$$ \left[ BAU- RUR\right]\ \left( Area\ affected\ \left[{m}^2\right]\right)\times unit\ cost\ \left[\in /{m}^2\right] $$

where Area affected is the cellar affected by flooding and unit cost is a constant (estimated unit damage, [€/m²]).

Fig. 17.4

Avoided damage is calculated from the difference in area in which dwellings and businesses are affected by groundwater levels with and without NBS multiplied by a unit damage cost per m²

In this case study the total damage is set to € 68/m², which is a value derived from an analysis by COWI (2014) for the Danish assurance umbrella organisation “Forsikring og Pension” (Danish Insurance and Pension). The unit damage key numbers were estimated from reimbursed compensation from insurance companies for the period 2006 to 2012, including the catastrophic cloudburst event in July 2011.

Damages to cellars and other subsurface infrastructure from elevated groundwater levels caused by cloudburst events, prolonged rainfall, or a combination of both, can be distinguished in two main damage types, which in combination can show cascading effects, reinforcing each other:

(A) water damage due to (i) fast increased street level inundation and spilling over to lower areas, like cellars, via doorsteps, staircases to cellars or other low thresholds, as well as (ii) water pushed up from floor drains and toilets (in case no backflow blocker has been installed). Both types are the result of fast ‘catastrophic’ rain events. (B) slowly rising groundwater as a result of prolonged rainfall, changed (decreased) water abstraction patterns (water savings), and repaired sewage and drainage channels that also functioned as shallow groundwater drainage, all of which exacerbate the effects under (A) as drainage capacity to remove surface water is decreased.

The hydrologic model simulates the effect of (B) whereas mainly (A), and to an unknown extent (B) are previously only considered in the damage functions. This means that the calculated ‘avoided damage’ is considered to be a minimum avoided cost.

The avoided damage is calculated by analysing the area in which shallow groundwater level exceeds a certain threshold value where damage to cellars is probable and for which we have estimates depending on the type of built environment for that particular location. The threshold value of 2.0 m below surface has been determined from averaging known cellar depths (data provided by HOFOR, 2008) and expert judgement on expected shallow groundwater level in the RUR catchment. Groundwater may rise to this level where it interferes with the drainage and sewer system, primarily combined, and where it may cause water being pushed up to cellar or street level. In that case it adds to the stormwater flooding from e.g. cloudburst events. The applied threshold of 2.0 m below surface above which unit damage per square meter does not change. This assumption is in reality an underestimation, due to that the damage value probably will change with level higher above the threshold value, therefore the avoided damage calculated in this way must be regarded with caution and represents a minimum value (best case). Hence, the rule is damage = 0 if depth <2 m and damage = constant if depth <2 m.

Cellar depth is calculated as an overall average depth for all 100 m by 100 m model grids for which an occupation percentage of 7.5% is applied. For example, for a 100 m by 100 m grid amounting to 10,000 m² there is a building which is at risk for being flooded as a result of rising groundwater level, or spill over water from the street level. This building is assumed to have a cellar surface of 0.075*10,000 m² = 750 m², which is the value used for calculating damage for cellar flooding.

To estimate the cost of damages the calculation looks like this:

$$ Damage=\left( number\ of\ pixel\right)\times 750\ {m}^2\times \in 68/{m}^2 $$

for both the RUR and BAU scenarios.

The avoided damage is the difference between the two scenarios:

$$ Avoided\ damage= BAU\_ damage- RUR\_ damage. $$

Other avoided costs with respect to the implementation of the restored urban river RUR scenario are costs associated with climate adaptation alternatives and grey solutions. Alternatives to the opening of the urban river and its tributaries include cloudburst tunnels and retention basins, the costs of which have been estimated (Ernst and Young and Moe 2019). Before the release of the cloudburst management plan there also was the option of upgrading the traditional combined sewer system in Copenhagen which appeared to be a very costly solution (Ernst and Young and Moe 2019) as compared to alternative climate adaptation solutions. Therefore, avoided costs associated with upgrading and operating the traditional sewer solutions are not considered in this analysis for the urban river RUR NBS scenario.

17.3 Integrating Stakeholder’s Knowledge in the NBS Designing Process

To successfully implement NBS designs for controlling the flooding from groundwater and cloudburst, the collaboration of a wide set of decision-makers and stakeholders is required. Lack of cooperation could result in barriers to the NBS implementation and/or affect their effectiveness (see Chaps. 7 and 19 for a broader discussion).

In the Copenhagen case study, stakeholders were engaged from the first stages of the project. The integration of stakeholder’s knowledge in the co-design process of NBS had multiple elements of relevance. Firstly, potential barriers hampering NBS implementation were identified. Secondly, participatory modelling activities were carried out to elicit and structure stakeholder’s risk perception. Finally, NBS effectiveness was assessed from a shared stakeholder perspective. To this aim, Fuzzy Cognitive Maps (FCM) (Kosko 1986) and Social Network Analysis were carried out starting from a Group Model Building (GMB) phase.

17.3.1 Fuzzy Cognitive Maps as a Tool for Trade-Offs Identification

Group Model Building (GMB) was used to conceptualize and understand, from a shared stakeholder perspective, different NBS strategies as well as the system where they would be applied. During a GMB Workshop key stakeholders participated in the identification of key factors and issues relevant for the modelling of Nature Based Solutions (NBS). Special attention was paid to the identification of NBS co-benefits and the main relationships between them. The main aim of the activity was to elicit and structure stakeholder’s risk perception and to assess the trade-offs and synergies of the co-benefits associated to NBS implementation. For this, a conceptual qualitative model was developed in a 2 hours workshop with key stakeholders of the Copenhagen case study (Stakeholders include representatives from the City of Copenhagen (partner, but others invited to capture multiple views), the surrounding Capital administrative region of which the City of Copenhagen is part. Others represent utility companies, neighborhoods within the city area, insurance, enterprises working with innovative climate adaptation solutions, business models and a cooperation between municipalities, utility services and the Capital Region of Denmark to support efficient and sustainable climate adaptation solutions in the region. In total ca. 25 people participated). The resulting qualitative model represented key factors and variables of the system as well as the main relationships between them (See Fig. 17.5).

Fig. 17.5

Showing the final qualitative group model integrating stakeholders and expert’s knowledge. In green are highlighted NBS co-benefits

The qualitative model was used firstly, to identify and visualize the main causes-effects processes occurring in the system. Secondly, to increase the communication and shared vision among stakeholders in order to facilitate the consensus agreement process. Finally, the group model was used to develop a semi-quantitative Fuzzy Cognitive Map (FCM) (Kosko 1986), that was used to simulate the effect of different NBS strategies in the production of co-benefits. In Table 17.2 the simulated scenarios in the Fuzzy Cognitive Model are shown.

Table 17.2. Showing scenarios simulated in the Fuzzy Cognitive Map model

17.3.2 Fuzzy Cognitive Map Development

FCM is a methodology that allows the representation of social knowledge in order to model the decision-making process in a socio-political system (Kosko 1986; Dickerson and Kosko 1994). It has proven to be an easy-to-use method to engage stakeholders and to acquire knowledge from them. It is also useful to aggregate knowledge coming from different sources and people with different degrees of expertise (Park and Kim 1995). FCM consists of nodes representing key concepts of the simulated system and fuzzy links connecting various nodes. The links represent positive or negative causal relationships among concepts and the fuzzy part permits the description of degrees of causality (Sokar et al. 2011). The strength of a link between two concepts indicates the intensity of the relationship between them, that is to say, how strong is the influence of one concept on the other. The weights range in a normalized interval of {−1, 0, 1}. The value 1 and − 1 represents a positive and negative causality respectively and the strongest relationship. The closer the values approach to 0, the weaker the relationships are. The relationships between variables can be represented in an adjacency matrix. In the FCM, this matrix allows the overall effects of an action on the elements in the map to be inferred qualitatively. The activation of a concept occurs when the value of a concept changes from 0 to 1. See Chap. 5 for more details.

The qualitative model developed during the group model building (GMB) exercise was post-processed and analysed and validated by a group of experts who were involved in the GMB and with experience in the system and in the FCM development. The role of the experts was assigning a weight to the causal relationship (weak, medium, strong). The strength of a relationship was represented by changing the thickness of the links between concepts composing the FCM (See Fig. 17.5). The experts also indicated potential ‘delays’ on the system, this means, causal processes that require time to occur. In the graphical representation of the FCM this was indicated with a delay mark (\\). Important feedback loops were also identified in order to facilitate the visualization of the causes-effect of the processes occurring in the system. Following the work developed by Giordano and colleagues in 2020 the FCM was simulated.

Identifying barriers hampering the NBS implementation using a collaborative social Network analysis approach.

The second phase of the stakeholders’ engagement process in the CPH case study concerned the mapping and the complex network of interactions taking place among different decision-makers. This phase aimed at detecting potential barriers to cooperation that could hamper the actual and effective NBS implementation. Starting from the results of the Group Model Building exercise, Social Network Analysis was adopted (See Chap. 5 for more details).

To this aim, a participatory mapping exercise was organized with a group of selected stakeholders. At the beginning, participants were required to identify the key issues that need to be addressed for enabling the NBS implementation. Therefore, participants agreed upon the following list of issues: (i) water quality (discharge); (ii) traffic management and parking areas; (iii) Cost affordability and finance; (iv) quality of the urban environment; (v) control of the water flow; (vi) private property right issues; (vii) regulatory and legal issues. The discussion was organized around these issues. The main scope of the participatory mapping exercise was to model the way decision-makers interact with each other, sharing and exchanging information, and cooperating for carrying out specific tasks for addressing the above-mentioned issues.

Participants were requested to introduce in the map the actors that need to be involved in order to address the above-mentioned issues, the tasks that need to be carried out and the information needed/used by each actor in order to carry out the requested tasks. Please, refer to Chap. 5 of this book for a more detailed description of the methodology for the participatory mapping. The following Tables 17.3, 17.4 and 17.5 show the results of this first phase of the discussion.

Table 17.3 Selected tasks within the Social Network Analysis for the Copenhagen study case group work

Table 17.4 Selected actors within the Social Network Analysis for the Copenhagen case study group work

Table 17.5 Selected information within the Social Network Analysis for the Copenhagen study case group work

Then, participants were requested to define the links between actors - that is, what are the actors that cooperate in order to address the issues? - between actors and tasks - i.e. who is supposed to do what? - between knowledge and tasks - i.e. what pieces of knowledge are needed to perform a certain task? - and between different pieces of knowledge - i.e. What piece of knowledge is needed in order to access another needed knowledge? Fig. 17.6 shows the Actors X Actors map of interactions among the different actors cited in Table 17.4.

Fig. 17.6

Maps of interactions between the different actors and the information

17.4 Results and Discussion

17.4.1 Hydrological Modelling and Damage Functions

Figure 17.7 shows the simulated change between the future groundwater levels with and without implementation of RUR. As expected, and intended, the shallow groundwater level rise is reduced in case a RUR-NBS is implemented (the blue pixels have lowest groundwater level).

Fig. 17.7

Difference in groundwater max levels with and without implementation of NBS (RUR) for the period 2070–2100 simulated the RCP 8.5 emission scenario (mean ensemble member)

Within the catchment area of RUR NBS (see Fig. 17.7) for the future climate, the area with GW damage is reduced from 1,620,000 m² (162 cells × 10,000 m²) to 1,060,000 m² (106 cells × 10,000 m²), which equals a reduction of 35%.

This means that the avoided damage under the conditions of future climate can be calculated from the reduced number of cells for which the cellar area is 750 m² per cell. Therefore, the avoided damage is (162–106) × 750 m² × € 68/m² is equal to € 2.856.000.

Likewise, for the reference period, the area is reduced from 1,460,000 to 930,000 m² which results in a somewhat lower avoided cost of € 2.715.700 (Table 17.6).

Table 17.6 Avoided cost of the two overall scenarios (see Table 17.1)

Besides the direct benefits of lowering the risk of groundwater flooding by restoring the urban river, there are also many other additional benefits or co-benefits associated with the RUR scenario. The conversion of the area from an area with a heavy traffic to a greener and quieter and more ambient place, will expect to increase the liveability of residents in terms of better health because of less air pollution, outdoor facilities and the opportunity for more exercise. The temperature will be reduced during the summertime, due to the cooling effect of the greenery and reduction of concrete and asphalt (urban heat island effect). Other co-benefits are increase in real estate prices which will provide a higher tax revenue for the municipality; businesses can flourish and create more jobs and crime will go down, due to the openness of the altered area.

17.4.2 Participatory Modelling

The results of the Social Network Analysis combined with the Group Model Building analysis allowed us to identify the EU directive (EUDIR) as a key barrier in the network of interaction. The model shows that this information is poorly shared among the different actors, although it plays a key role in the Knowledge X Task network. This means that several tasks cannot be properly implemented due to the limited access to the EU directives. This barrier is confirmed by the low level of centrality of the EU in the Agent X Agent network. Similarly, the cost/benefits analysis (COSBEN) is poorly shared in the Agent x Knowledge map. This could represent a limit, given the high level of importance of this information. Finally, although the City Council (CCOUNS) has frequent connections with the other actors, its capability to access key pieces of information is rather low. This potentially represents a barrier in the NBS implementation.

17.4.3 Uncertainties and Usefulness of the Approach in Assessing NAS Effectiveness

The methodology to assess avoided damage is conducted under several sources of uncertainties.

The hydrologic model has been calibrated against discharge of the piped river to an open water recipient, but since the urban river NBS is not implemented no calibration for the anticipated urban river has been possible. However, the applied model is physically based, i.e. physical processes in the integrated groundwater and surface water environment are state-of-the-art represented. Nevertheless, assumptions have been made including the exact position and depth of the restored river, groundwater-surface water interactions, groundwater well extractions and how cloudburst management climate adaptation measures are connected to the urban river, all of them add to assessment uncertainty.

Another source of uncertainty is the missing information on cellar depths for the municipality of Frederiksberg for which similarity with data from Copenhagen has been assumed, i.e. the average depth for the Copenhagen part of 1.78 m has been assumed for the Frederiksberg part, which is probable. The model setup and overall approach however represents a credible tool which can be updated when new information becomes available.

To calculate the avoided damage, data on insured damage reimbursements for the period 2006 to 2012 were crucial. Reported damage from citizens to insurance companies has been analyzed by a consultant (COWI 2014) for which available damage data from the insurance companies to the built environment has been aggregated as ‘water damage’. As a result of this, damage to cellars from both surface water overflow and from the subsurface sewerage system could not be distinguished from each other and the damage function was aggregated at the scale of residential buildings (€ 68/m²). This value does not consider whether flooded areas and buildings include businesses for which the damage costs may be higher. Also, the unit damage cost of € 68/m² does not vary with depth over the threshold value (2.0 m below surface), which represents an uncertainty and means the estimated avoided damage is a minimum value.

In addition, groundwater flooding damage as a result of sewer overflow probably also causes damage to the structure of buildings, not taken into consideration in the simple damage function used, which is why the estimated value of damage, a bit more than € 2.7 million, alone for this reason among others, probably is much higher.

When calculating the cost of avoided damages, a value of 7.5% for the affected area (e.g. cellars) is used. This value is a premise based on the best qualified assessment and is also subject to uncertainty. The affected area is likely higher, which means the calculated cost of avoided cost is higher than shown.

All this adds substantially to the overall uncertainty, but again the conceptual approach can be improved with new and improved data. Therefore, improved and more data, especially for the development of accurate damage functions are needed to reduce the overall uncertainty.

17.5 Conclusion

The RUR NBS scenario clearly shows that restoring (opening) a piped urban river to its natural position influences the groundwater level. The simulations show that the restored river will act as a drainage channel for both cloudbursts induced surface water flooding and a rising groundwater level and thereby avoid groundwater flooding and subsequently damage to subterranean structures.

Calculations show that by restoring the urban river to its natural position, 35% of the catchment area for the RUR NBS scenario will remain below the applied threshold of 2 m, where no damage will occur, compared to the BAU scenario.

Economic analyses show that the avoided costs by restoring the urban river will be at least € 2.7 million for the catchment area of the RUR scenario, and probably higher considering the overall uncertainty.

Besides the direct benefits of lowering the risk of groundwater flooding by restoring the urban river, there are also many other additional benefits or co-benefits associated with the RUR scenario. These co-benefits have not been converted into direct economic values in this study, but they will certainly mean a lot in the overall assessment of RUR NBS.

The restoration of the urban river (RUR) combined with an adjacent urban green area is perceived by the stakeholder as an effective solution capable to deliver a wide range of co-benefits. The results of the simulation show that the integration of both measures is likely to be more effective in the reduction of groundwater and surface flooding. The effectiveness of NBS increases considerably when good strong institutional collaboration and good green management follows. The FCM simulation has also revealed potential trade-offs between co-benefits. For example, an increase of the recreational and aesthetic value of the area where the NBS are applied could lead to a decrease of the social justice and cohesion co-benefit. This is because green spaces tend to increase property value. The results of the study have emphasized the need of adopting an integrated view of the system to assess and compare different NBS and adaptation strategies.

However, NBS implementation requires effective cooperation among different decision-makers. Therefore, efforts are needed to overcome the existing collaborative barriers, in order to (i) enhance the flow of information among the different decision-makers, (ii) reduce the level of conflicts and (iii) to facilitate collaborative decision-making. The implemented SNA helped to detect those barriers. A dialogue is needed among the different decision-makers for defining potential interventions – i.e. networking interventions – whose main scope is to enhance the collaboration over NBS design and implementation.