The effects of climate change are expected to increase the frequency and magnitude of floods, droughts and heat waves. An emerging method termed adaptation tipping point – opportunity (ATP-O) assesses a system's climate-incurred tipping points and uses opportunities arising from urban developments to introduce adaptation strategies while reducing investment costs. The objective of this research was to apply the ATP-O method to the city of Dordrecht in the Netherlands. The results show that the alternative adaptation strategy proposed (an overland drainage system) would be effective in coping with the effects of climate change where the current management strategy (disconnection of impervious surfaces from sewer systems) fails to do so. The ATP-O also proved helpful in identifying opportunities to adapt at lower costs. This research stimulated discussions between stakeholders on performance objectives, policy development, investment strategies, and flood risk management practices. The sensitivity analysis performed to support such discussion revealed that small variations in acceptability thresholds, associated with policy objectives, can have significant impact on ATP occurrence and timing.

INTRODUCTION

Scenarios developed by the Intergovernmental Panel on Climate Change (IPCC) project increasing and intensifying extreme weather events worldwide (Intergovernmental Panel on Climate Change 2007). More frequent and more intense extreme weather events are likely to put more pressure on urban areas, including the different social and ecological subsystems. The effects of changes in pressures are expected to increase the frequency and magnitude of floods, droughts, and heat stress (European Environment Agency 2012). Social and ecological systems could therefore be facing increased climate-related risks and this reveals the need for policy-makers, water managers and urban planners to plan ahead and anticipate extreme weather events. As such, climate change becomes relevant to a range of stakeholders, who need to understand at which point current management strategies will fail to meet their objectives. The questions then become: when and where will current management strategies fail to meet policy objectives, what options are available for adaptation, and what is the right moment to implement these options? While focusing on the management of urban flood risk, this paper demonstrates how it becomes possible to answer such questions by using a novel method for climate change impact and adaptation assessment, called adaptation tipping point – opportunity (ATP-O). Early experiments with the ATP-O performed at the neighborhood-scale showed promising results (Gersonius et al. 2012). This research has applied the ATP-O at the city scale, using Dordrecht in the Netherlands as a case study, to assess the city's potential to increase its degree of resilience to climate change and to optimize the investments in flood risk management and urban development. Resilience is defined here as ‘the capacity of a system to absorb disturbance and reorganize while undergoing change so as to still retain essentially the same function, structure and feedbacks’ (Folke et al. 2010).

The objectives of the research were to determine: (1) to what extent the existing engineering system is capable of coping with climate change; (2) whether, and for how long, the current climate change adaptation strategy will be effective in meeting policy objectives; and (3) when, and where, alternative adaptation strategies are needed to increase the degree of resilience to climate change.

THEORETICAL FRAMEWORK

Traditionally, the assessment and management of climate-related risks consisted of adding a safety factor to the loads (rainfall intensities) acting on an engineering system (e.g. DEFRA 2006). This safety factor against climate change is generally specific to an emission scenario, climate change model, location and time period. However, despite scientific advances, there remain large uncertainties about the direction, rate and magnitude of climate change. Climate change uncertainties introduced in impact models cascade through each step of the assessment and result in large ranges of possible impacts (Schneider 1983). Such ranges commonly become too large for practical application in risk management. Climate change uncertainties thus limit the usefulness of more traditional climate risk assessment and management methods.

In the context of adaptation to climate change, this paper presents and tests a novel approach, the ATP-O. Gersonius et al. (2012) further developed the ATP-O from the ATP as introduced by Kwadijk et al. (2010). In this context, an ATP is defined as ‘the magnitude of change due to changing external drivers such that the current management strategy will no longer be able to meet the objectives’ (Kwadijk et al. 2010). The novelty of ATP-O is that the method does not rely on precise forecasts of climate change, as argued in the remainder of this chapter.

The ATP-O is described as a combined effect-based and bottom-up approach (according to the classification proposed by Carter et al. (2007) and modified by Jones & Preston (2011)) that consists of assessing a system's response to changing external drivers such as climate change, and of developing adaptation strategies that are effective in maintaining or enhancing resilience. The effect-based approach starts by specifying a required performance level, used to define acceptability thresholds to manage the impacts, and then assesses the likelihood of attaining or exceeding this outcome as a result of changes in pressures. Bottom-up approaches begin at the local scale, by determining whether it is feasible to increase the existing system's ability to deal with climate change and its natural variability (Jones & Boer 2005). The bottom-up approach is based on the recognition that adaptation is better conceived as a socio-economic process rather than as a set of stand-alone adjustments. It takes a more practical view on adaptation by combining climate change with socio-economic drivers, such as infrastructure renewal, in order to acquire investment efficiency benefits (Jones & Preston 2011), which has been referred to by Huq & Reid (2004) as ‘adaptation mainstreaming’.

The ATP-O method uses policy objectives and societal acceptance levels to determine whether the existing system will reach an ATP. ATPs are only dependent on the magnitude of change due to changing external drivers, and not on time. This makes them much less dependent on scenarios in the assessment of adaptation (compared to a cause-based approach that begins by considering the scenarios). Once determined, such points can be positioned in time using scenarios. Combining the defined tipping points with climate change scenarios will provide information about the system robustness to climate change and the potential need for alternative adaptive management strategies. The ATP-O method offers insight (through the analysis of tipping points) into how much change the existing system can cope with, and insight into the effectiveness of a management strategy under changing external drivers. The ATP-O is thus useful in answering some of the basic questions decision makers have: what is the first problem we will have to face as a consequence of climate change, and when will that happen (Kwadijk et al. 2010)? The novelty of the ATP-O over the traditional ATP method lies in identifying and taking advantage of opportunities to introduce adaptation measures. Possible adaptation measures include ‘gray infrastructure’, ‘green infrastructure’ or non-structural approaches (European Commission 2009). A key challenge to adaptation is to implement such measures in existing built environments. Yet there are significant opportunities arising from infrastructure renewal to introduce adaptation measures incrementally over time and to keep adaptation costs low, by timing these measures with broader public and private sector investments. Therefore, the ATP-O method, in contrast with the ATP, recognizes future change not only as a threat, but also as an opportunity to improve the system and to cut down investment costs.

A key variable that defines the characteristics of flood risk management is the system performance, i.e. its capability in terms of reducing flood risk. For flood risk management to be considered resilient, the system performance should be maintained under the different climate change scenarios. The critical threshold occurs where the system performance is outside the acceptable risk level, as defined by law and/or decided by stakeholders. The magnitude of climate change beyond which the characteristics (system performance) change, becomes a fixed point of reference against which the degree of resilience of current flood risk management can be quantified – hereafter referred to as an ATP. When an ATP is reached for the current management practice, then a transformational change may be required. This type of change involves a revision in management practice through the implementation of a different kind of adaptation strategy. It involves ‘breaking up’ structural resilience in order to maintain or enhance functional resilience under future change. In this respect, the objective of this research was, furthermore, to develop and assess an alternative adaptation strategy to better deal with changing risks.

METHOD

The ATP-O is a step-wise method based on three types of activity: (1) setting policies and objectives; (2) ATP analyses; and (3) opportunity analysis (adaptation mainstreaming) (Figure 1). These activities require input from actors of different disciplines to develop the knowledge to support decision-making (Van Herk et al. 2011). The remainder of this section explains the activities and steps in further detail. See also Figure 1 for an illustration of the activities and steps.
Figure 1

Flow-chart of the ATP-O method (adapted from Gersonius et al. 2012).

Figure 1

Flow-chart of the ATP-O method (adapted from Gersonius et al. 2012).

Setting policies and objectives

  • (1) The first step is to specify the functions of the system and climate change effects of interest for the resilience assessment. The objectives for the performance of the system are defined and often translated into acceptable standards. In addition, the current management strategy to achieve the objectives should be identified.

  • (2) Next, the particular threshold values for the acceptable standards are quantified. These threshold values can be defined according to regulation, national law, or determined by the stakeholders, and can change over time.

ATP analysis

  • (3) The ATPs are identified by increasing the design loads on the system (e.g. the rainfall intensity) and by assessing the specific boundary conditions (i.e. the magnitude of climate change) under which the performance objectives are no longer met. This step is similar to performing a sensitivity analysis of the system's performance under possible future design loads. The results of the assessment are then represented in a chart that indicates the occurrence of ATPs.

  • (4) Climate change scenarios are used to transform the specific boundary conditions (i.e. the magnitude of climate change) under which an ATP will occur into an estimate of when it is likely to occur. This can be done by overlaying these scenarios with possible future design loads in the above-mentioned chart. This output provides an estimate of the earliest and latest times when the system performance will no longer be acceptable.

  • (5) To avoid reaching an ATP, a change in adaptation strategy is needed so as to maintain or enhance resilience to climate change. To account for long lead times in implementing structural measures, the adaptation measures should be identified well before the critical ATP occurs. Implementing this strategy will alter the nature and timing of the critical ATPs.

  • (6) A number of alternative adaptation strategies will result from repeating steps 3 and 4. Some adaptation measures will be structural and some non-structural. Engagement with all stakeholders is required in this step to select an adaptation strategy that is realistic and acceptable. However, where the implementation of alternative adaptation strategies is too costly or not acceptable to society or for the environment, the acceptability threshold may be allowed to decline. The use of a sensitivity analysis will be helpful in finding an appropriate balance between adaptation to and acceptance of increased risks.

Opportunity analysis (adaptation mainstreaming)

  • 5A. An understanding of urban infrastructure maintenance and renewal cycles permits identifying opportunities for the mainstreaming of climate change adaptation. The expected physical lifetimes of public infrastructure, established from expert knowledge and/or literature can be used to achieve this step (Langston et al. 2008). For this study, the physical lifetimes of roads, buildings and sewers were estimated from expert knowledge.

  • 5B. Determine the time windows when adaptation opportunities will occur. This can be done by estimating when the existing structures will reach their end-of-life, as deducted from the expected physical lifetimes (as obtained from step 5A) and the construction periods (e.g. Veerbeek et al. 2010). Alternatively, where urban regeneration is concerned, the time windows of opportunities can be identified from existing plans for already planned investment projects. For the analysis of adaptation opportunities it is crucial that all key stakeholders share their plans for already planned investment projects. For this study, adaptation opportunities were identified from the timing of infrastructure end-of-life, which was determined from a database containing the year of construction for all infrastructure.

  • 5C. Modify already planned investment projects to incorporate potential adaptation measures. These measures should then be included in the definition of the alternative adaptation strategies at step 5 of the ATP method. Taking account of the possibilities for adaptation mainstreaming in step 6 of the ATP method will lead to a better understanding and quantification of the adaptive potential of the system. In this study, redevelopment plans in two districts were to be revised so as to include terrain modifications.

  • 5D. To realize adaptation mainstreaming, the adaptation implementation process should be tied as closely as possible to urban renewal activities. Adaptation mainstreaming can be realized by comparing the time windows at which renewal is planned for with the moments at which ATPs will occur; this is illustrated in Figure 2. If renewal occurs before the ATP, then adaptation measures can be incorporated in the planned investment projects. In this study, adaptation mainstreaming was possible due to the timing of tipping points that would occur prior to 2050, and adaptation opportunities from urban renewal that would occur city wide until 2050 (refer to Figure 2(a)). As argued by Van de Ven et al. (2011), the costs of carrying out adaptation measures synergistically with already planned investment projects will, in the majority of cases, be of the order of 50–80% lower than the costs of implementing these as stand-alone adjustments. Whether or not adaptation mainstreaming is likely to be cost efficient will, however, also depend on the length of the time period between the time windows of the opportunities and the critical ATPs. With a longer time period, the potential cost savings from adaptation mainstreaming will be off-set by the cost savings from postponing the implementation of adaptation measures until later (refer to Figure 2(b)), that is, until the occurrence of the critical ATP (minus the lead-time). This is because later investments will be discounted more heavily than earlier investments. Hence, the longer the time period, the less attractive adaptation mainstreaming will be. If renewal occurs after the ATP then alternative stand-alone adaptation measures must be defined, and the tipping point analysis has to be repeated to assess their effectiveness in providing resilience for climate change (refer to Figure 2(c)). The required cost saving for cost-efficient adaptation mainstreaming (as a function of the differential time period between the opportunities and the critical ATPs) are given by Gersonius (2012).

Figure 2

Mainstreaming opportunity: (a) urban renewal timed with ATP; (b) urban renewal postponed until occurrence of ATP; (c) ATP occurs after urban renewal.

Figure 2

Mainstreaming opportunity: (a) urban renewal timed with ATP; (b) urban renewal postponed until occurrence of ATP; (c) ATP occurs after urban renewal.

CASE STUDY: URBAN FLOODING IN DORDRECHT

The ATP-O method was applied to the management of flood risk in the city of Dordrecht. In supporting this case study, a Learning and Action Alliance (LAA) was formed within the framework of the Interreg 4B North Sea Region project MARE. This concerns a social learning framework that fosters collaborative planning at the local level. Members of the LAA Dordrecht included the municipality of Dordrecht, the water board Hollandse Delta, knowledge institutes (including Utrecht University and UNESCO-IHE) and other stakeholders. The LAA supported the adoption of an integrated approach to flood risk management and spatial planning, and served a role in (i) system analysis, (ii) collaborative design (including the selection of thresholds) and (iii) governance. In this sense, it provided a vehicle for learning and working together to apply the ATP-O (Van Herk et al. 2011).

In Dordrecht, three types of sewers exist: the combined sewer type, which contains storm water and wastewater and is most widely used; the separate sewer type, where storm water is transported and treated separately from wastewater; and the improved separate sewer type, which includes a provisional system to ensure treatment of contaminated stormwater in case of faulty connections between stormwater pipes and wastewater pipes. Figure 3 shows the sewer network that was used for the tipping point analysis. The sewer network used contains the principal elements of Dordrecht's sewer system: the manholes, pipes, overflows, and pumps. Integral to the system, are some 11,306 manholes and 34 combined sewer overflow outlets which discharge to the open water system.
Figure 3

Sewer system in Dordrecht, The Netherlands: (a) districts and manholes and (b) sewer types.

Figure 3

Sewer system in Dordrecht, The Netherlands: (a) districts and manholes and (b) sewer types.

Setting policies and objectives

Climate change, increasing population densities and changes in land use are some of the variables that can have great impacts on urban flood risk. For this study however, the interest was to assess the effects of climate change, expressed as increasing rainfall intensities, on the drainage system, while keeping other variables fixed. Changes in land use (greater pervious surfaces) were considered to some extent, as a means to decrease runoff reaching the drainage system. The current policy objective set by the municipality and the water board is to have no street flood incidents resulting from an exceeded sewer system capacity, from now until the year 2050. This policy objective was translated into an acceptability threshold, that is, a value beyond which the expectations of performance are likely to be compromised. The indicator selected to determine whether the policy objective would be met was the percentage of overtopped manholes for the design rainfall event. The maximum percentage of overtopped manholes was set to 1% to reflect an appropriate and realistic threshold (i.e. balancing livability, gradual adaptation and acceptable financial investments). Note that the ATP-O will later be used to analyze the capacity of the overland drainage system in coping with climate change. The design load and performance indicator (percentage of flooded buildings) used to analyze this overland drainage system are different than those used for the sewer system. This is because the failure consequences differ in a fundamental way, namely: street flood incidents (for sewer system failure) vs. flooding of buildings (for overland drainage system failure). As a result of these differences, the ATP-O for the overland drainage system will lead to other estimates of where and when an ATP is likely to occur.

Current adaptation strategy, ATPs, and adaptation opportunities

Dordrecht's current adaptation strategy consists of gradually disconnecting 40% of publicly owned buildings and paved areas from the sewer system to reduce the amount of storm water runoff that reaches the sewers. This can be achieved for example through green infrastructure, such as rain gardens and bio-swales (e.g. Ashley et al. 2011). Adaptation mainstreaming is realized by combining the activities required for storm water disconnection with already planned renovation works for the public buildings, streets and parking lots. The current adaptation strategy was developed prior to this research (Luijtelaar et al. 2006), however without explicitly identifying specific adaptation opportunities.

The renewal of publicly owned buildings and infrastructure was regarded here as an opportunity to introduce adaptation measures incrementally and to keep additional adaptation costs low. To realize adaptation mainstreaming, the adaptation process should be tied as closely as possible to the time windows of urban renewal. Here, the time windows of opportunities were identified from databases provided by the municipality, which contained spatial and temporal information on publicly owned buildings and infrastructure, identifying the year each building, street and parking lot would undergo maintenance. From expert opinion on expected physical lifetimes and the information contained in the databases, it was possible to determine that all streets and a majority (78%) of buildings were going to need renewal by 2050. Yet because 100% disconnection was not considered feasible, it was assumed that only 40% of these could be disconnected by the year 2050. This assumption was previously made by Luijtelaar et al. (2006) and repeated here.

The tipping point analysis of the system was performed using a hydraulic model developed with the software package SOBEK (Dhondia & Stelling 2002). The one-dimensional (1-D) model called DORD_BAS developed by Luijtelaar et al. (2006) served to analyze the sewer system (refer to supplementary material for model details, available in the online version of this paper). Model simulations were performed for all urban districts in Dordrecht in order to assess the drainage system's response to climate change. The storm series ‘Bui06’, developed by RIONED, the center of expertise in sewer management and urban drainage in the Netherlands, was chosen as design load on the sewer system to run the simulations. Bui06 represents a 75 minutes distributed rainfall event of 17 mm rainfall depth with a 1-year return period (RIONED Foundation 2006), and is the design event adopted by the municipality of Dordrecht for short-duration, intense rainfall under current climate conditions. Climate change was simulated by increasing the rainfall intensity of the design rainfall event, i.e. the amount of rainfall was increased over the same time duration. The model output allowed determining the percentage of flooded manholes resulting from intense rainfall events.

The output of the 1-D model reveals the freeboard height of each manhole being the distance between the street level and the water level inside the sewer pipe. The freeboard height allows determining which manholes overtop, and hence the percentage of overtopped manholes in the city.

Figure 4 shows the ATP graph for the existing sewer system for the 1% threshold of overtopped manholes (current policy objectives). Squares and dots both represent tipping points for the existing sewer system and the current adaptation strategy, respectively. On the vertical axis the districts are listed in order of resilience as a function of climate change, which is expressed on the horizontal axis as a percentage of rainfall intensity increase. The tipping points were translated to a timescale by interpolating two climate change scenarios: G & W, which were developed by the Royal Netherlands Meteorological Institute (KNMI) (van den Hurk et al. 2007). These scenarios are based on a 1 °C temperature increase (G) and a 2 °C temperature increase (W) by 2050. The scenarios G and W were interpolated for 2050 with the rainfall events used for the ATP analysis, in order to provide a timescale for the occurrence of tipping points. These scenarios are shown by the dashed vertical lines in Figure 4. As such, it was possible to estimate whether a management strategy was likely to be effective until or beyond a certain date. The shaded bar at the top of the graph shows the degree of resilience associated with each district. If the tipping point occurs before the rainfall intensities associated with the G2050 scenario, then the system is unlikely to be resilient until 2050. If the tipping point occurs after the W2050 scenario, the system is very likely to be resilient until 2050.
Figure 4

Tipping point analysis using a 1% threshold.

Figure 4

Tipping point analysis using a 1% threshold.

Figure 4 indicates that for the existing sewer system, 10 districts out of 17 (58%) already experience a tipping point under current climate conditions; i.e. 0% rainfall intensity increase. Hence, 10 districts require an adaptation strategy of the existing sewer system under current climate conditions. The current adaptation strategy of disconnecting 40% of public buildings and paved areas helps shift the ATP for only 12 of the districts (70%). For example, in the district of Dubbeldam, the tipping point for the existing sewer system occurs at a 0% rainfall intensity increase, whereas it shifts to 10% rainfall intensity increase with the current adaptation strategy. For these 12 districts the system will be able to cope with greater rainfall intensities. However, despite implementing the current adaptation strategy, four (24%) districts still would not meet current policy objectives under current climate conditions, and an alternative adaptation strategy is therefore required to cope with the impacts of climate change in these districts. Sterrenburg1 illustrates this well; the tipping point is already reached for the existing sewer system and Dordrecht's current climate change adaptation strategy will not change this.

Furthermore, focusing on a single district, Figure 5 shows the percentage of overtopped manholes for the Vissershoek district for the existing sewer system and for the current adaptation strategy. On the horizontal axis the climate change impacts are expressed as rainfall intensity increase per 5% increments, and corresponding accumulated rainfall quantities in millimeters per 75 minutes. The horizontal dotted line represents the 1% threshold. The system's tipping points occur where the curves for the existing sewer system and for the current adaptation strategy intersect the threshold line. For the existing sewer system, the tipping point with a 1% threshold will occur at 10% rainfall intensity increase, while for the current climate adaptation strategy the tipping point occurs at 45% rainfall intensity increase. The shift in tipping point from left to right indicates exceedance of the 1% threshold occurs at much higher rainfall intensities after implementation of the current adaptation strategy than with the existing sewer system. As shown in Figure 4, other districts exhibit similar patterns of threshold exceedance as Vissershoek.
Figure 5

Percentage of overtopped manholes in Visserhoek district for the existing sewer system and the current adaptation strategy.

Figure 5

Percentage of overtopped manholes in Visserhoek district for the existing sewer system and the current adaptation strategy.

A sensitivity analysis was performed to understand the effects of changing acceptability thresholds on ATP occurrence and timing. The tipping point analysis was thus repeated using acceptability thresholds set to 0% and 5% of overtopped manholes. The number of districts that exceeded these thresholds at current climate conditions for both the current system and current adaptation strategy are compared in Figure 6. Refer to the supplementary material (available online) for the charts showing the tipping point analysis with 0 and 5% thresholds. When setting the acceptability threshold to 0%, a majority of districts (15 out of 17) fail to meet the performance objectives. Setting such strict performance objectives would likely translate into large financial investments to develop adaptation measures. On the other hand, setting performance objectives to 5% of overtopped manholes show that only four districts would fail to meet performance objectives, and would therefore required fewer adaptation measures; however this would allow for a greater flood extent.
Figure 6

Sensitivity analysis using varying acceptability thresholds.

Figure 6

Sensitivity analysis using varying acceptability thresholds.

The results of the sensitivity analysis demonstrate how sensitive the tipping point analysis is to small threshold variations and stress the importance of carefully selecting thresholds that are deemed acceptable during policy formulation. These results allowed initiating a dialog between the stakeholders on future objectives and system performance levels. The 1% threshold, which was deemed more realistic with the infrastructure currently in place in Dordrecht, is used in the remainder of the study.

Alternative adaptation strategy, adaptation opportunities, and ATPs

Figure 7 shows a map of districts that exhibit low to moderate resilience under the current adaptation strategy and the 1% threshold, and thus are identified as requiring an alternative adaptation strategy. A bottom-up process to develop an alternative adaptation strategy was set with the LAA Dordrecht. The LAA used the ATP-O approach to deal with climate change and to integrate flood risk management in urban development planning at the local level. As a result of this collaborative process, the alternative adaptation strategy proposed and assessed is that of using the overland drainage system to manage excess runoff that the existing drainage system cannot accommodate for. An overland drainage system consists of using or modifying street surfaces and other areas to temporarily store runoff or to direct runoff to an area where it can be stored or infiltrated without inconvenience, health or security threats to the public (Balmforth et al. 2006).
Figure 7

Map of districts requiring an alternative adaptation strategy.

Figure 7

Map of districts requiring an alternative adaptation strategy.

The alternative adaptation strategy involves a transformational change that will potentially take place over longer time periods, as it involves changing social tolerance levels (Folke et al. 2010). The alternative strategy requires individuals to accept giving water more room at the surface, rather than attempting to convey it underground. Although managing excess runoff on street surfaces is generally accepted in the Netherlands for extreme events, there used to be no objectives and acceptability thresholds for this. This situation is currently changing, and this paper contributes to this advancement. The alternative strategy is proposed here as complementary to the current strategy to help the municipality of Dordrecht to cope with climate change.

As implied by the ATP-O method, the alternative strategy (use of an overland drainage system) is proposed here to effectively cope with the effects of climate change and enhance resilience where the current management strategy (disconnection of impervious surfaces from sewer system) fails to do so. In this case, the ‘Opportunity’ identification phase of ATP-O took place during discussions with the municipality regarding upcoming redevelopment plans for two districts: Sterrenburg1 and Sterrenburg2. Those redevelopment plans were regarded as the time windows of opportunity to perform adaptation mainstreaming with the proposed terrain modifications for the overland drainage system.

For this study, the use of an overland drainage system was associated with an extreme rainfall event, ‘Storm50’, rather than the design rainfall event. Storm50 is a 2 hour distributed rainfall event with 43 mm rainfall depth, having a 50-year return period. During such events, a large percentage of manholes overtop. Overtopped manholes were assumed to be socially acceptable, whereas flooded buildings were assumed to be socially unacceptable. This (i.e. the alternative strategy) also implies a change in objective. The objective selected for the alternative strategy was to prevent flooding of buildings for extreme rainfall events (with a 1 in 50 year frequency). The performance indicator was the percentage of flooded buildings rather than the percentage of overtopped manholes. A doorstep height of 5 cm was assumed, and buildings were considered flooded when the runoff height exceeded 5 cm. The threshold was set to 1% of flooded buildings, which amounts to approximately 100 buildings. The use of an overland drainage system was tested for the existing terrain profile in Dordrecht which served as point of reference, as well as for an adapted terrain (alternative adaptation strategy). A digital elevation map of the area, the Dutch AHN2 map with elevation accuracy of ca. 5 cm reshuffled to 2 × 2 m resolution, was used to simulate runoff over the terrain. For this study, a 1-D/two-dimensional (2-D) model called DORD_ODS was developed with the SOBEK software to simulate interactions between the sewer system and the major overland drainage system (runoff triggered by overtopped manholes and flowing over the cells of the AHN2 terrain). Refer to supplementary material (available online) for model details. Elevations at specific areas were then edited in ArcGIS 10.0 to simulate modifications brought to the terrain with the alternative adaptation strategy. The terrain modifications that were simulated consisted of lowering parking lots and public park grounds, increasing sidewalk curb heights, constructing speed bumps according to Dutch road standards, modifying the road gradient, installing urban rain gardens and digging swales on the side of the roads. Moreover, property dry-proofing, where no water is able to enter the building and property wet-proofing, where water is able to enter the building, with reduced flood damage effects through improved building design, were included in the simulation for specific buildings. The location and type of terrain modifications were developed based on the flow patterns of the model runs with the current terrain, and on the existence of public parks and parking lots identifiable with Google Earth (for example, a speed bump in addition to increasing the sidewalk curb height to an acceptable height was simulated to temporarily store runoff locally). As such, the conveyance paths were chosen so as to avoid inundating private property and critical public infrastructure. However, it should be noted that several model runs were performed in order to select the most appropriate adaptation measure. For example, if the terrain was modified to include an additional speed bump and an increased curb height, and the simulation revealed that the runoff could not properly be stored, than the adaptation measure was revised to possibly modifying the road gradient to direct the runoff to a nearby parking lot. The terrain was therefore edited again in ArcGIS and the model simulation was run again, until the runoff was effectively managed, that is, without inundating critical infrastructure.

The alternative adaptation strategy was tested with this model for districts Sterrenburg1 and Sterrenburg2, as upcoming redevelopment plans in these districts provide adaptation opportunities. Design loads on the overland drainage system were simulated using a 1 in 50 year synthetic storm event with 43 mm of rainfall depth distributed over a 2-hour period. This is the acceptable standard for the overland drainage system, as proposed by RIONED Foundation (2006). The 1-D/2-D model outputs allowed determining the percentage of flooded buildings resulting from the design loads.

Figure 8 shows a map of the flood extent in Sterrenburg2 for the design load under current climate conditions. When comparing Figure 8(a) and 8(b) it is apparent that fewer building blocks are flooded with modifications brought to the terrain (alternative adaptation strategy); for the current terrain 15 building blocks (out of 1,683) are flooded, as opposed to only 7 with the modified terrain.
Figure 8

Map of flooding in Sterrenburg2 for (a) current terrain and (b) modified terrain under the alternative adaptation strategy.

Figure 8

Map of flooding in Sterrenburg2 for (a) current terrain and (b) modified terrain under the alternative adaptation strategy.

Climate change was simulated by increasing the rainfall intensity of the synthetic rainfall event. Figure 9 presents a tipping point analysis for the current terrain and the alternative adaptation strategy. The results show that with the current terrain, the rainfall intensity can increase by 20% and 10% for Sterrenburg1 and 2, respectively, before a tipping point occurs; that is, before more than 1% of buildings flood. The modified terrain under the alternative adaptation strategy would substantially increase each district's degree of resilience to climate change by 10% and 20%, respectively. Sterrenburg1 and 2 could therefore cope with greater rainfall intensities with the alternative adaptation strategy (55.9 mm) rather than with the current terrain (51.6 mm and 47.3 mm, respectively) or the current management strategy (17 mm and 19.55 mm, respectively), as seen from Figures 4 and 9.
Figure 9

Tipping point analysis for current terrain and modified terrain under the alternative adaptation strategy.

Figure 9

Tipping point analysis for current terrain and modified terrain under the alternative adaptation strategy.

DISCUSSION

The use of the ATP-O in this case study has provided valuable information on the options to enhance resilience of flood risk management to climate change. The tipping point analysis provided insight into the effectiveness of the current climate change adaptation strategy in meeting flood risk management objectives over the next decades for each of the districts in Dordrecht. From there, an alternative adaptation strategy was developed for districts likely to experience a tipping point before 2050, and tested for resilience to climate change. Finally, opportunities for adaptation arising from planned renewal activities were identified, so as to encourage investment optimization.

The tipping point analysis shows that Dordrecht's existing sewer system already fails to meet suggested policy objectives in 10 out of 17 (58%) districts and is therefore not effective in dealing with current and future climate conditions. The analysis of Dordrecht's current adaptation strategy, i.e. disconnection, revealed its effectiveness in postponing ATPs in 70% of the districts. Therefore, the implementation of this strategy would be effective in enhancing resilience to climate change in the majority of districts. These results were expected and in accordance with other examples of impervious surface disconnection from sewer systems found in the literature. Digman et al. (2012) for example, discuss how the city of Portland (USA) has successfully dealt with sewer flooding backing up into basements and on roads by performing down-pipe disconnections of multiple commercial and residential properties since 1993.

Nevertheless, despite disconnecting 40% of publicly owned buildings and paved areas from the sewer system, four districts (24%) would still fail to meet policy objectives. This reveals the need to develop alternative adaptation strategies in these districts. An overland drainage system proved effective in dealing with climate change, where the current adaptation strategy failed to do so. The alternative adaptation strategy proposed should be regarded as complementary to the current strategy, as both are effective in providing resilience to climate change, but to different degrees and at different locations. The results of the tipping point analyses performed with the alternative adaptation strategy were expected and in accordance with examples found in the literature. For example the ‘Urban flooding retrofit’ project of Devonshire Park in Yorkshire, which partially consisted of controlling the path of excess runoff to Devonshire Park in order to deal with annual flooding of properties, has been successful in protecting properties from being flooded (Digman et al. 2012). In this respect, the alternative adaptation strategy offers increased flexibility in the way flood risk is managed, as it widens the range of adaptation options available and helps achieve policy objectives.

The sensitivity analysis demonstrated the importance of setting appropriate acceptability thresholds. Selecting different threshold values can produce very different results when performing the tipping point analysis of the system in question. This is illustrated by the large difference obtained for the existing sewer system and current adaptation strategy for the 0, 1 and 5% thresholds. Setting stringent performance criteria will ensure having little flood damage but investment costs will be higher than with moderate performance criteria. The trade off between flood damage and investment costs should be carefully evaluated by the stakeholders concerned. Based on the result of this case study such policy debates are currently being addressed by the collaborative network in Dordrecht. Moreover, the municipality of Dordrecht is currently updating the sewer model, as reports of sewer flood incidents indicate that the sewer system is likely performing better than the model indicates.

Applying the ATP-O and developing site-specific adaptation strategies can be time consuming, especially if each district requires a different strategy. However, the analysis only needs to be performed once, even if new climate change information becomes available. The time invested in performing the analysis is therefore likely to pay off in social, environmental and economic benefits, as long as alternative adaptation strategies are implemented before the system reaches a tipping point and during planned renewal activities where possible. Within the scope of this research, it has not yet been possible to modify the actual renewal projects to incorporate the adaptation options identified, and to thereby assess the impacts of adaptation mainstreaming on investment costs.

CONCLUSION

This research shows that the application of the ATP-O increases the understanding of the system dynamics and its response to a range of external pressures. This method helps to identify the threats of climate change as well as opportunities to adapt. The results of the tipping point analysis show how much longer the current management strategy will be effective in meeting policy objectives and societal preferences. The sensitivity analysis further revealed the importance of setting appropriate and realistic policy objectives that allow balancing livability, gradual adaptation and acceptable financial investments among other stakeholder priorities. In Dordrecht, setting policy objectives and acceptability thresholds to no street floods resulting from overtopping manholes until 2050 is unrealistic considering the existing infrastructure and immediate investments required to achieve this. An acceptability threshold of 1% of overtopping manholes is more realistic as it provides room to balance gradual adaptation with financial investments.

Moreover, the results of such studies allow presenting managers and policy-makers with improved information on climate-related risks and on the consequences of delaying adaptation activities. The ATP-O also helps identify key moments during which adaptation can be realized and timed with already planned urban renewal activities, so as to allow for a gradual adaptation process. The comparative assessment of different management strategies provides detailed information that can support stakeholders in making decisions for climate risk management.

Despite the advancements realized through this research, performing additional case studies and conducting further research would provide greater empirical evidence of the applicability of the ATP-O to climate change impact and adaptation assessments. In addition, a cost-benefit analysis of the adaptation strategies identified should be performed in order to weigh the environmental, financial, social and technical costs as well as benefits of each measure. Moreover, an in-depth analysis of the effects on investment costs of alternatives is still required. Finally, additional work is required in order to organize the collaborative process effectively.

ACKNOWLEDGEMENTS

This work has been carried out as a part of the EU Interreg IVB project MARE (Managing Adaptive Responses to changing flood risk) and the Dutch Knowledge for Climate project CPC (Climate Proof Cities).

REFERENCES

REFERENCES
Ashley
R. M.
Nowell
R.
Gersonius
B.
Walker
L.
2011
Surface water management and urban green infrastructure: A review of potential benefits and UK and international practices. FR/R0014, Evans T.D. editing, Foundation for Water Research
.
Balmforth
D.
Digman
C.
Kellagher
R.
Butler
D.
2006
Designing for Exceedance in Urban Drainage: Good Practice
,
C635
.
CIRIA
,
London
.
Carter
T. R.
Jones
R.
Lu
X.
Bhadwal
S.
Conde
C.
Mearns
L.
O'Neill
B.
Rounsevell
M.
Zurek
M.
2007
New assessment methods and the characterization of future conditions. Climate change 2007: Impacts, adaptation and vulnerability
. In:
Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change
(
Parry
M. L.
Canziani
O. F.
Palutikof
J. P.
van der Linden
P. J.
Hanson
C. E.
, eds).
Cambridge University Press
,
Cambridge
,
UK
, pp.
133
171
.
DEFRA
2006
Flood and Coastal Defense Appraisal Guidance FCDPAG3 Economic Appraisal
.
Supplementary note to operating authorities – climate change impacts
.
DEFRA
,
London
.
Dhondia
J.
Stelling
G.
2002
Application of one-dimensional-two-dimensional integrated hydraulic model for flood simulation and damage assessment
. In
Bibliography | 123 Proceedings of the 5th Hydroinformatics Conference
,
Cardiff
,
United Kingdom
, pp.
265
276
.
Digman
C.
Ashley
R.
Balmforth
D.
Balmforth
D.
Stovin
V.
Glerum
J.
2012
Retrofitting to Manage Surface Water
,
C713
.
CIRIA
,
London
.
European Commission
2009
White Paper on Adapting to Climate Change: Towards a European Framework for Action.
European Commission
,
Brussels
.
European Environment Agency
2012
Climate Change, Impacts and Vulnerability in Europe 2012: An Indicator-based Report
.
EEA Report No 12/2012
.
EEA
,
Brussels
.
Folke
C.
Carpenter
S. R.
Walker
B.
Scheffer
M.
Chapin
T.
Rockström
J.
2010
Resilience thinking: Integrating resilience, adaptability and transformability
.
Ecology and Society
15
(
4
),
20
. .
Gersonius
B.
2012
The Resilience Approach to Climate Adaptation: Applied for Flood Risk
.
PhD Thesis, TUDelft, CRC Press/Balkema
,
Leiden, NL
.
Gersonius
B.
Nasruddin
F.
Ashley
R.
Jeuken
A.
Pathirana
A.
Zevenbergen
C.
2012
Developing the evidence base for mainstreaming adaptation of stormwater systems to climate change
.
Water Research
46
(
20
),
6824
6835
.
Huq
S.
Reid
H.
2004
Mainstreaming adaptation in development
.
IDS bulletin
35
(
3
),
15
21
.
Intergovernmental Panel on Climate Change
2007
Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, 2007
.
IPCC, Cambridge University Press
,
UK
.
Jones
R.
Boer
R.
2005
Assessing Current Climate Risks. Adaptation Policy Framework for Climate Change: Developing Strategies, Policies and Measures
.
Cambridge
:
Cambridge University Press
,
UK
.
Jones
R. N.
Preston
B. L.
2011
Adaptation and risk management
.
Wiley Interdisciplinary Reviews: Climate Change
2
(
2
),
296
308
.
Kwadijk
J. C. J.
Haasnoot
M.
Mulder
J. P. M.
Hoogvliet
M.
Jeuken
A.
Van Der Krogt
R. A. A.
Van Oostrom
N. G. C.
Schelfhout
H. A.
Van Velzen
E. H.
Van Waveren
H.
2010
Using adaptation tipping points to prepare for climate change and sea level rise: a case study in the Netherlands
.
Wiley Interdisciplinary Reviews: Climate Change
1
,
729
740
.
Langston
C.
Wong
F.
Hui
E.
Shen
L. Y.
2008
Strategic assessment of building adaptive reuse opportunities in Hong Kong
.
Building and Environment
43
(
10
),
1709
1718
.
Luijtelaar
H.
van Luijendijk
J.
Lee
A. v. d.
2006
Optimalisatie Watersysteem Dordrecht [Optimisation Water System in Dordrecht]. TAUW BV sponsored by Waterschap Hollandse Delta and Gementee Dordrecht
.
RIONED Foundation
2006
Stedelijke Wateropgave, Vergelijking normen voor water op straat en inundatie [in Dutch]. Online article.
(accessed 7 September 2012)
.
Schneider
S.
1983
CO2, climate and society: A brief overview
. In:
Social Science Research and Climate Change: An Interdisciplinary Appraisal
(
Reidel
D.
, ed.).
D. Reidel Publishing Company
,
Boston
, pp.
9
15
.
van den Hurk
B.
Tank
A. K.
Lenderink
G.
van Ulden
A.
van Oldenborgh
G. J.
Katsman
C.
van den Brink
H.
Keller
F.
Bessembinder
J.
Burgers
G.
Komen
G.
Hazeleger
W.
Drijfhout
S.
2007
New climate change scenarios for the Netherlands
.
Water Science and Technology
56
(
4
),
27
33
.
Van de Ven
F.
Van Nieuwkerk
E.
Stone
K.
Zevenbergen
C.
Veerbeek
W.
Rijke
J.
Van Herk
S.
2011
Building the Netherlands Climate Proof. Urban Areas
,
1201082-000-VEB-0003
,
Deltares/UNESCO-IHE
,
Delft/Utrecht
.
Veerbeek
W.
Ashley
R.
Zevenbergen
C.
Gersonius
B.
Rijke
J.
2010
Building adaptive capacity for flood proofing in urban areas through synergistic interventions
. In:
First International Conference on Sustainable Urbanization (ICSU 2010)
,
Hong Kong
,
China
,
15–17 December
.

Supplementary data