In the summer of 2021 severe flooding occurred in Belgium, Germany and The Netherlands. High damages have been estimated in the Geul catchment, leading up to 250 million euro. Flood risk management in The Netherlands is mostly focussed on reducing the probability of flooding and in lesser extent on damage mitigation. Especially in areas such as the Geul catchment, reducing the probability of flooding through flood defences is difficult to achieve. Also taking climate change and urban development into consideration, it is challenging to meet flood safety standards in the (near) future, as the frequency and intensity of flooding is expected to increase. Under these circumstances a shift is observed in scientific and policy discourses towards a flood risk reduction strategy that includes resilience, in which the capacity to resist, mitigate and recover has a larger role. As a consequence, there is an increasing interest to include private actors in the strategy by promoting private flood mitigation measures. However, it is difficult to quantify the effects of these private measures and little research in The Netherlands has been performed. This thesis investigates which improvements can be made to the flood risk reduction strategy in the Geul catchment, thereby increasing resilience through private flood mitigation measures. Thereto a method to include and asses the effect of the measures in a risk based approach is derived.

An international literature study about flood resilience and society provided two applicable private mitigation measures and their damage reductive effect. Dry-proofing types of measures are defined as measures that prevent water from entering an object and can reduce damages between 60%-100% up to an inundation depth of 1 meter. Wet-proofing measures reduce damages and recovery time, while allowing water inside the object, by flood adapted use or structural measures. The maximum damage reduction is 53% up to an inundation depth of 2 meter. Investigation of the Protection Motivation Theory provided important factors that improve the uptake of these measures, by identifying connections between awareness, preparedness and self-reliance. The two main findings are that the communication strategy should focus on the relative risk reduction effect of private measures and should provide guidance on implementation to motivate private actors to initiate action. Five interviews were used as a reference to these findings. To describe the relative effect of measures, they are modelled in a publicly available risk assessment method (SSM-2017), based on hazard, exposure and vulnerability. The effect of the measures is calculated for two scenarios, that describe at which locations the measures are implemented. In the first scenario, top risk locations are identified and measures are applied. This is representative for high awareness and object specific guidance. The second scenario randomly distributes measures over the flood prone area and represents the current approach without specific guidance. For the modelling of measures and scenarios, an automatic procedure using a Python script is developed, which is applicable in The Netherlands in combination with SSM-2017.

The results showthat both dry-proofing andwet-proofing houses and shops are applicable in the Geul catchment. At many locations measures have a positive benefit-cost ratio, indicating it is worthwhile to implement them compared to a situation without measures. More importantly, it is found that high awareness of flood risk or object specific guidance increases the effectiveness and cost-efficiency significantly. For residential objects, an optimum was found to implement measures at the top 40% locations most at risk. The achieved risk reduction was 47% and the benefit-cost ratio 1.3-2.2 for dryproofing. At this coverage the benefit-cost ratio ofmeasures is a factor 2 higher compared to a situation without specific guidance. Location specific investigation revealed that the inundation depth is the main driver for determining the most effective measure. Wet-proofing is cost-efficient in a smaller number of areas at risk for inundation depths higher than 1 meter. For most areas dry-proofing is more cost-efficient as the expected inundation depths in the catchment are mostly below 1 meter.

The research demonstrates the potential of privatemeasures with a significant reduction of flood risk. Given the limitations to reduce the probability of flooding in the Geul catchment, the measures are highly interesting. The method itself was able to determine the type of measure thatwas most effective with confidence. However, based on uncertainties in the model, the benefit-cost ratios have to be interpreted carefully. An extension of this research with a higher resolution of input data can address this uncertainty. Current approaches and communication to motivate private measures were found to be ineffective, as they mainly focus on the type of measures and costs instead of the effect and implementation. It is suggested to improve this by explicitly including the before mentioned factors and bymaking use of social network on the spatial scale of a neighbourhood.

Global economic losses due to flooding have increased over the last 50 years, with damages estimated in the billions of dollar each year. Governmental institutions face great challenges to reduce flood risk under the influence of population increase, urbanization and climate change. Interdisciplinary urban planning becomes more important to mitigate flood risk with a multidisciplinary approach and insights in conceptual designs is important to support interdisciplinary cooperation. An example of a country facing great challenges is Albania, where on average 23 flooding events occur each year and at least 54% of the regions have been affected. In addition, the country faces challenges in the disciplines of urban planning, transport and water quality. The capital city Tirana and surroundings were highly affected by flooding events in 2017. The city is prone to flash floods and, especially in the winter months, river flooding. It is crucial to identify areas susceptible to flooding and to quantify the impact on flood risk reduction of potential measures. Data availability imposes a major constraint to quantify flood risk in urban areas with currently available assessment methods. Therefore a flood risk assessment method needs to be identified that is applicable in areas with poor or scarce data. Once the method is identified it can be applied in the Tirana region to identify flood risk and to quantify the impact of measures. Potential measures were developed in an interdisciplinary workshop with students from the TU Delft and Tirana and focussed on the Tirana region, including the Lana river area. The proposed measures for the Lana river are quantified in the risk assessment in this thesis. The identified method defines flood risk as the product of hazard, exposure and vulnerability expressed in Expected Annual Damage (EAD) in Euro. European depth-damage functions for four land use classes are used, in combination with maximum damage values, to express the exposure and vulnerability component. The hazard is described by inundation maps and the development forms the core of this thesis. With the use of a Digital Elevation Model (DEM) a Height Above Nearest Drainage (HAND) map is developed. A Synthetic Rating Curve (SRC) was constructed based on the application of the Manning equation with reach-averaged geometries derived from the HAND map. With a time series analysis of measured discharges, inundation maps could be constructed linked to a return period. The identified method was applicable in the Tirana region. Major benefits are the low computational requirements, easy applicable GIS operations and low data requirements, which make the method easily transferable to other regions. Constraints were found in the use of the coarse DEM with 25 meter resolution, which did not cover river geometries and local measures with high enough accuracy. This was solved by the manual inclusion of river geometries and verification of the developed inundation maps. The Lana catchment is identified as the area most susceptible to flooding with an EAD of approximately one million Euro. One of the proposed measures, a multifunctional cross-section, proved to be most effective with a reduction of 65%. An urbanization scenario, in which the risk increased with 84%, demonstrated the need of the measures. Finally, the assessment revealed potential adverse effects of upstream measures on other areas and proved to add beneficial insights.

Workshop Group 2

Japan and the Netherlands have very different physical, historical and cultural contexts but they share a vulnerability to extreme flood related events and have, in both their (relatively) recent pasts, had to recover from such events: be they the floods of 1953 in the Netherlands or the tsunami that hit Japan’s east coast in 2011. This paper describes the process and results of two workshops investigating flood reconstruction responses undertaken by students representing five disciplines at TU Delft in the Netherlands. A particular workshop method was employed to promote an interdisciplinary design process and then design responses investigated for the (very real) Japanese case were transferred to a hypothetical disaster scenario for Vlissingen, in the south of the Netherlands. The conclusions reached focused as much on the efficacy of the workshop method as the particular design proposals for both cases as well as on what was learnt via the comparison between Japanese and Dutch, contexts and reconstruction philosophies.

The global economical losses caused by flooding have increased over the last 50 years (IPCC, 2014; UNISDR, 2013). Deltares and HKV are developing new methods to quantify flood risk on a global scale. A Flood Risk Assessment Tool (FIAT) is used to performcalculations. The definition of risk is hazard x exposure x vulnerability. FIAT is able to combine these elements and calculate the Expected Annual Damage (EAD) for each city.

Currently two methods are used on a global and continental scale. For Europe a flood risk analysis is performed based on five land uses categories and five depth-damage functions, this method is referred to as the Multiple land use method. On a global scale an analysis is performed based on one land use and one depthdamage function, this method is referred to as the Single land use method. Nootenboom (2015) investigated the difference between the two methods for Europe. He described a good correlation of 0.97. Deltares and
HKV want to extend the Multiple land use method to a global scale.

A flood risk assessment for Australia is performed with the Multiple land use method. Brisbane, Adelaide, Melbourne, Perth and Sydney are reviewed. Recently derived depth-damage functions by Huizinga (2015) were used. Australian functions are available for residential and commercial land use. For other land use classes European functions are used instead. As a source for flood hazard the PCR-GLOBWB hydrological model is used. Exposure maps for Australia are made with OpenStreetMap (OSM). Metadata from OSM was filtered and processed in such a way five land use classes were formed. A total coverage of 52% was achieved.
The areas where no data could be recovered was filled in with ratios of the retrieved data. A sixth layer, named the no-data layer is formed. Visual analysis shows that the assumptions made in the no-data layer are most accurate in densely populated areas.

When the results of the Multiple land use method are compared to the Single land use method a correlation of 0.99905 is found. For every city there was a decrease in flood risk. The depth-damage function for Australia is investigated. Typical for the Australian function is the steep shape compared to the European function. A sensitivity research shows that each city reacts different to changes in the depth-damage function. Especially for cities with low inundation depths the risk is strongly influenced by the use of Australian functions. Cities with high inundation depths, such as Brisbane, are far less susceptible for changes in the depth-damage function. The use of European functions for certain land uses classes is therefore acceptable in Brisbane, as long as the dominant residential and no-data layer have continent specific functions.