Flood risk analysis study at the German Bight Coast

From Coastal Wiki
Revision as of 13:29, 26 September 2008 by Cliviahaese (talk | contribs) (→‎Introduction)
Jump to: navigation, search

Introduction

The 2007 report of the Intergovernmental Panel on Climate Change (IPCC 2007[1]) made evident that an ongoing global climate change will cause increased storminess and sea level rise in coastal zones. There is little doubt that the North Sea will also be affected by an accelerating rise of the sea-level, an increase in extreme weather events and a greater tidal range. In order to be prepared for future conditions, prevention measures have to be improved and methodologies to assess and manage upcoming risks have to be further developed. As there are still deficits in assessing the full range of flood impacts, new approaches have been developed for hazard analysis, vulnerability assessment, and flood risk management in the framework of the EU-project FLOODsite (Integrated Flood Risk Analysis and Management Methodologies). In order to apply some of these new methodologies, a pilot site application was conducted for the community of St. Peter-Ording at the German North Sea Coast combining failure probabilities of the coastal defence system with micro-scale socio-economic vulnerability analysis.

Study area

Along the German Bight Coast vast low-lying areas are threatened by recurring storm floods and are thus at risk of being flooded. Storm surges with a water level much higher than mean high tide (Mhwl) are a major factor inducing flood risks in coastal areas in Germany. The water level in the North Sea depends primarily on tides and the direction, intensity and duration of winds. The German North Sea Coast is at risk by winds mainly coming from the west and northwest.

Figure 1: Map and coastal defence structure of the pilot site St. Peter-Ording.

St. Peter-Ording is one of the largest communities at the German Bight Coast with 7,278 inhabitants, whereof 4,022 are permanent residents; the others have a temporary home or summer residence. The local economy relies heavily on tourism with over 100,000 guests each year. Furthermore, the municipality has an important regional and national function as health centre with various hospitals and other health facilities. The community is located very exposed on the west coast of Eiderstedt peninsula (cp. Figure 1). The coastal landscape of this investigation area is dominated by dunes, which at least in the north of the town are high enough to serve as a natural coastal protection structure. The size of the study area is approximately 6000 ha of which about 4000 ha are considered to be flood-prone due to the respective elevation distribution. Hence, a flooding of the municipality could spread far into the hinterland of Schleswig-Holstein. Strong storm surges, which may occur several times a year, pose a serious threat to the community. Severe storm surges have occurred in 1962 and 1976, where the former caused heavy damages in the German North Sea Region while during the latter the highest storm surge water levels ever were recorded with a water level up to 4.8 m above mean sea level. In reaction to the 1962 flood, the protective measures were increased along the North Sea coast and the dikes were significantly heightened in St. Peter-Ording, resulting in only minor damages in 1976. Three other storm surges in 1962, 1981, and 1999 did pass the 4.0 m mark and a general increase of storm surge frequency and severity over the last decades can be recorded.

The community is protected against storm floods by a complex coastal defence system (cp. Figure 1). It is divided into a foreland, dune structures (>2.5 km length, between ~10 and 18 m high), a major dike line and a second dike line. The major dike line is 12.5 km long and about 8.0 m high, although not constant over its length. Furthermore, there is a 2 km long so called overtopping dike. This type of dike is designed to withstand wave overtopping and wave overflow. It is therefore considerably lower than standard dikes and is protected by a very solid asphalt cover layer. Risk management, including coastal defence, has to be steadily adjusted to be prepared for future climate conditions.

Concept of flood risk analysis

Figure 2: Risk analysis framework.

For coastal defence planning and risk management the knowledge and the spatial distribution of risk are compulsory. Hence, the aim of this investigation is developing new methodologies to better estimate flood risk on a micro-scale level where risk is defined as (Gouldby and Samuels 2005[2]):

risk = probability x consequence


This definition includes the probability of flooding of the flood prone area and all kinds of consequences of flooding depending on the vulnerability of the flood prone area. However, the resilience of the coastal system or any management activities is not included. A micro-scale vulnerability analysis together with a full probabilistic approach in determining the flooding of the hinterland is the most efficient way to quantify the magnitude of the flood risk and hence form a sound basis for any risk management activities performed in the area. Hence, the key elements of the analysis are a probabilistic hazard analysis, the determination of flooding scenarios based on the hazard analysis and a micro-scale vulnerability analysis. The vulnerability analysis follows an integrated approach and comprises economic, social, and ecological vulnerability criteria. It is divided into a damage potential analysis for St. Peter-Ording carried out with a standardised methodology and damage estimation for different flooding scenarios (cp. Figure 2).

Finally, the methodology includes a GIS-based approach merging the various levels of the economic, social, and ecological vulnerability with scenario-based probabilities of flooding on a micro-scale level. This was then planned to be used to map different zones of flood risks in the area.

Hazard analysis

This section describes the approach to derive the overall probability of failure for all flood defences in the area. This comprises the development of an algorithm according to which the defence line can be split into different sections, that can be treated independently, and the calculation of the failure probability for each section of the flood defence line.

Figure 3 Typical simplified fault tree for a dike section at “German Bight Coast”.

The methodology applied here is following the source-pathway-receptor model as described in Kundzewicz and Samuels (1997)[3]. Risk sources at the German Bight are resulting from storm surges in the North Sea associated with high water levels and storm waves at the flood defences. Typically, storm surges last not longer than 12 to 24 hours but may increase the water level considerably (up to 3.5 m in the North Sea). The interaction of normal tides (tidal range of 1-2 m is typical in the southern North Sea region), storm surges, and waves is crucial for the determination of the water level at the coast. In addition, the foreshore topography plays a major role when determining the waves at the flood defence structure. In the case of the German Bight, limited water depths over a high foreland will cause the waves to break and will therefore limit the maximum wave heights which reach the flood defence structures. However, the probabilistic hazard analysis only considers single probability distributions for each of the governing variables such as water level, wave height and wave period. No joint or conditional probability density functions were considered. As for risk pathways in the German Bight Coast pilot site, flood defences comprise more than 12 km of dikes (grass and asphalt dike) and a dune area of about 2.5 km length. However, the probabilistic risk assessment (PRA) has focussed on the dikes as the key flood defence structure since the dune belt is extraordinary high and wide and is regarded significantly safer than the dike protection. Laser scan data have been used to determine the exact height of the flood defence line in more detail and to define different ‘homogeneous’ sections of the flood defences. Criteria for distinction of homogeneous dike sections were the type of flood defence, its height (being very different and ranging from 6.22 mNN to 8.43 mNN for the sea dikes), its orientation, the key sea state parameters like water level and wave height and wave period, respectively, and geotechnical parameters. Thirteen sections have been identified using these criteria. Each of these sections is assumed to be identical over its entire length and hence will result in the same probability of failure (cp. Figure 3).

The result of this analysis is an annual probability of flooding of the hinterland for each dike section that has been selected. These flooding probabilities were typically found to range from a probability of 10-4 to 10-6, which means a return period of flooding in the range of 10,000 or 1,000,000 years. These results were found reasonably low and comparable to earlier studies of similar flood defences, although those have been based on different fault trees and failure modes. The overall flooding probability using a fault tree approach for all sections results in Pf = 4 × 10-3.

Flood scenarios and inundation simulation

The results of calculating the flooding probability were used to derive flooding/breaching scenarios, which up to now have been based on experience and expert knowledge. The section with the highest probability of failure for breaching of the dike was taken as the section where a breach location was assumed. The detailed location of the breach was defined after visual inspection of the relevant section and consultation with the local authorities. Additional analysis of other sections of the flood defence line has shown that the lowest part of the dikes is overtopped for relatively low storm surge water levels. Hence, a first flooding scenario was assumed, which includes a water level of 5.30 mNN (design water level for this area), a breach location in the south of the area as described above and initiated by wave overtopping, and wave overtopping at a low asphalt dike near the village of Ording. This scenario was also used as the standard flooding scenario for estimating the consequences. The probability of this flooding scenario is Pflooding = 9.6 × 10-8, which is much lower than the results obtained by the hazard analysis. Preliminary versions of breach models developed under FLOODsite were used to calculate the expected breach dimensions (final breach width and depth). These parameters were used as input boundary conditions for the flood inundation model. The numerical non-linear shallow water (NLSW) model SOBEK (see Delft Hydraulics Software) was used to perform the flood inundation simulation. SOBEK models the details of the flooding process and hence provide inundation depths, velocities, and duration of flooding for any location of interest in the flood prone area. Ditches and channels were simulated using the 1D flow module of SOBEK and were found to be relevant for distributing the flood wave into the area. Boundary conditions were the time series of the storm surge water level on the one hand and the mean overtopping rates over the lowest part of the defence line on the other hand.

Vulnerability analysis

To assess the flood risk and to define specific risk zones, the estimation of the expected damages and its spatial distribution is crucial in addition to the hazard analysis. The total flood damage of a specific flood event depends on the vulnerability of the socio-economic and the ecological system. Hence, a detailed vulnerability analysis was conducted for St. Peter-Ording focussing on two major deficits of former vulnerability assessment studies (see text box).

Shortcomings of vulnerability assessment
1.) The scale of vulnerability assessments is substantial as it is directly linked to the application of the results in practice. At the German Bight Coast, vulnerability assessment studies have been conducted on macro- (IPCC CZMS 1992 and Sterr, 2008[4]), meso- (Hamann & Klug, 1998[5]) and micro-scale (Reese & Markau, 2002[6]). In comparison, these studies have shown that the expenditure and accordingly the precision of the macro- and meso-scale vulnerability assessment decreases, since methods are generally based on aggregated data. Micro-scale methods can achieve a high level of precision, as it is possible to identify the actual existing conditions in the areas at risk on an object orientated level. This approach is however costly in terms of time and money. Hence, a simplification of micro-scale approaches towards a quick economic feasible instrument is necessary.
2.) So far, most methodologies for the assessment of vulnerability were designed according to economic criteria, which can be described in monetary terms, whereas intangible values, social characteristics and ecological values have been widely neglected. However, flood risk is determined by more than economic losses, rather it comprehends all kinds of consequences of flooding. In order to understand the interrelations of socio-economic and ecological dynamics and the impacts of floods on intangible values, the integration of physical, social, and economic processes at the coast is crucial.

Multi-criteria vulnerability assessment for St. Peter-Ording

According to a multi-criteria risk assessment the spatial distribution of the three main dimensions of flood risk - economic, social, and ecological risk - were investigated at the pilot site in order to identify specific risk zones. At first, the selection of the vulnerability criteria is important as the choice of indicators influences the results. As the focus of this vulnerability analysis is on simplifying the assessment methodology, the choice of the vulnerability criteria was done according to five selection criteria. They should be complete, covering all dimensions of vulnerability, available at a reasonable cost-benefit ratio, comparable in different periods and places, measurable, i.e. statistically sound, and minimal in order to be easily applied.

To assess the overall vulnerability the following vulnerability criteria were assessed in St. Peter-Ording:

  • Economic vulnerability criteria (Buildings, private inventory, stock value, gross value added)
  • Social vulnerability criteria (population at risk (& risk to life), vulnerable people, social hotspots)
  • Ecological vulnerability criteria (number of sensitive coastal biotopes (dunes, forest, wetlands, grassland))

Economic vulnerability

Economic vulnerability relies primarily on the description and quantification of the economic risk potentials. The economic vulnerability in this investigation is assessed by a damage potential analysis, estimating the sum of existing monetary values, which could suffer damages in case of a flood event on an object level, and a damage estimation with relative depth damage curves to calculate the damage of the values depending on inundation depth. On the basis of different flooding scenarios and the calculated damage potential, an economic vulnerability could be subsequently assessed by means of an estimation of the possible damages based upon depth-damage-functions.

Social vulnerability

To determine the social risk, the social vulnerability criteria people at risk (& risk to life), vulnerable people, and social hotspots were assessed. Community statistics were used to obtain these data on a micro-scale level. The spatial distribution of people at risk was taken from statistics, including seasonal differences due to tourism. People with a special vulnerability include invalid persons, people older than 70 years and children younger than 8 years as they are assumed to be more vulnerable than others in case of a flood disaster. Social hotspots in St. Peter Ording are schools, kindergartens, a nursing home, a youth recreation home and clinics. These places are assumed to be more affected in case of a flooding than others.

Ecological vulnerability

Ecological vulnerability describes the susceptibility of ecological values, protected areas, or biotopes towards adverse impacts. Coastal biotopes like dunes or wetlands have not only a high ecological value but also a buffer and protection function. At the same time, they are highly sensible to disturbances or changes by e.g. inundation, saltwater intrusion, or wave impact which may lead to a loss of their function.

In this study, ecological vulnerability criteria were assessed rather simply by mapping the size and extension of coastal biotopes at the pilot site. Land use categories and biotopes were mapped, classified in beach area (e.g. moor, salt marshes, dunes, cliffs, outlets, tidal flats), natural habitats (e.g. reed belts, bog forest), semi-natural habitats (e.g. coniferous forest, mixed forest), grassland and acres, recreational areas, traffic areas and settlement areas. However, in this simple approach only the size of these biotopes was considered but not their susceptibility.
Figure 4: Spatial distribution of economic (l.), social (m.), ecological (r.) vulnerability in St. Peter Ording.

From the selected classes, grassland covers the largest areas followed by forest, dunes, and wetlands. However, the size alone does not describe the degree of ecological vulnerability, as smaller areas like the dune belt are ecosystems that are more sensible. Hence, a weighting of the different coastal biotopes had to be done in a next step.

Results of the vulnerability analysis

The flooding scenarios together with the related probabilities provided the basis for combining the hazard and the vulnerability analyses. Different scenarios were calculated taking into consideration the various locations where breaching or severe overtopping may occur. These scenarios were linked to the damage potential in the risk zone to estimate the specific damage. From the inundation depth and the duration, the expected damage could be calculated (cp. Figure 4).

Risk estimation and mapping

Figure 5: Risk categories according to the MAUT approach.

In order to define risk zones it is necessary to quantify the flood risk as exactly as possible by weighting the different vulnerability criteria. This was done by a multi-criteria risk assessment using a comparable risk rating system including economic, social, and ecological damage categories. A GIS based map output allows a ranking of risk zones. Therefore, the area map of St. Peter-Ording is divided into a raster with a size of 300 x 300 m and all rating is based on a raster cell. The Multi-Attribute Utility Theory (MAUT) approach (Meyer et al., 2007[7]) was used as an evaluation scheme with which all criteria can be aggregated to a single scalar factor. Two weighting approaches have been tested, a simple ranking and a swing weight approach. The latter was more convincing; it is based on a decision of the importance of each criterion in relation to the others. The considered vulnerability criteria are all criteria from the vulnerability analysis above (buildings, private inventory, stock value, gross value added, people at risk, vulnerable people, hotspots, dunes, forest, wetland, grassland). Each criterion is given a value from 0 to 10, depending on its value in a raster cell. Then, using the swing weight approach, a weighting and normalization within each category is carried out. Thus each category again can have a value from 0 to 10. Depending on the stakeholders’ interest, the different categories can now be weighted differently. In Figure 5, each category is weighted as one third as an example.


Finally, the approach results in comparable risk maps for different scenarios of flooding in St. Peter-Ording. The integration of economic, social, and ecological vulnerability criteria showed that there has been a shift in risk zones, compared to the mere economic risk assessment as for example areas with ecologically vulnerable dunes became a risk zone although no economic assets are in that region.

Conclusions

A detailed flood risk analysis at the German Bight Coast has been performed within the European FLOODsite research project. The study comprises a full probabilistic analysis of the flood defences protecting the hinterland close to the village of St. Peter-Ording on the Eiderstedt peninsula and a micro-scale vulnerability analysis, including economic, ecological, and social aspects of the vulnerability. With respect to the hazard analysis, the results have shown that even though some input parameters were not directly available and had to be estimated the results were believed to be very reliable (Pf in the range of Pf = 10-4 to 10-6). Sensitivity analyses have been performed accounting for the uncertainties of the input parameters. The results of the vulnerability analysis have shown that the integration of economic, social, and ecological vulnerability criteria is feasible as it gives a more complete picture of the overall susceptibility of a coastal site towards flood risk. Compared to studies focusing only at economic vulnerability the risk zones shifted in St. Peter-Ording as they include also risk zones dominated by ecological values. The vulnerability criteria and the simplified assessment method chosen in this investigation are assumed to be easily transferable to other coastal areas. However, the method has to be tested in other coastal sites to validate the transferability.

References

  1. ↑ IPCC (2007): Climate change 2007: The physical science basis - Summary for policymakers. Intergovermental Panel on Climate Change (IPCC), Genf, 21 p.
  2. ↑ Gouldby, B. & Samuels, P. (2005): Language of risk - project definitions. Floodsite project report T32-04-01.
  3. ↑ Kundzewicz, Z.; Samuels, P.G. (1997): Real-time Flood Forecasting and Warning. Conclusions from Workshop and Expert Meeting. Proceedings of Second RIBAMOD Expert Meeting, no. EUR-18853-EN, Published by DG XII, European Commission, Office for Official Publications of the Europ. Communities, Padova, Italy, 277 p.
  4. ↑ Sterr, H. (2008): Assessment of Vulnerability and Adaptation to Sea-Level Rise for the Coastal Zone of Germany. In: Journal of Coastal Research 24 (2): 380-393.
  5. ↑ Hamann, M. & Klug, H. (1998): Wertermittlung für die potentiell sturmflutgefährdeten Gebiete an den Küsten Schleswig-Holsteins. Gutachten im Auftrag des Ministeriums für ländliche Räume, Landwirtschaft, Ernährung und Tourismus des Landes Schleswig-Holstein. Unpublished Final Report.
  6. ↑ Reese, S. & Markau, H. (2002): Risk Handling and Natural Hazards: New Strategies in Coastal Defence – A Case Study from Schleswig-Holstein, Germany. In: Ewing, L. & Wallendorf, L. (eds.): Solutions to Coastal Disasters 2002, San Diego, pp 498-510.
  7. ↑ Meyer. V.; Haase, D.; Scheuer, S. (2007): GIS-based Multicriteria Analysis as Decision Support in Flood Risk Management. Floodsite project report 10-07-06, 55 p. Floodsite

See also

Internal Links

External Links

  • IPCC General website


The main author of this article is Kortenhaus, Andreas
Please note that others may also have edited the contents of this article.
Citation: Kortenhaus, Andreas (2008): Flood risk analysis study at the German Bight Coast. Available from http://www.coastalwiki.org/wiki/Flood_risk_analysis_study_at_the_German_Bight_Coast [accessed on 24-11-2024]


The main author of this article is Kaiser, Gunilla
Please note that others may also have edited the contents of this article.
Citation: Kaiser, Gunilla (2008): Flood risk analysis study at the German Bight Coast. Available from http://www.coastalwiki.org/wiki/Flood_risk_analysis_study_at_the_German_Bight_Coast [accessed on 24-11-2024]
Retrieved from "https://www.coastalwiki.org/w/index.php?title=Flood_risk_analysis_study_at_the_German_Bight_Coast&oldid=23439"