Vulnerability and Risk Assessment for the PFRA in Ireland

Transcription

1 Vulnerability and Risk Assessment for the PFRA in Ireland Mark Adamson 1, Anthony Badcock 2 1 Office of Public Works, Ireland 2 Mott MacDonald Group, U.K. Abstract The Preliminary Flood Risk Assessment (PFRA) in Ireland is being undertaken based on three approaches; a review of historic floods, a predictive assessment of flood risk, and consultation with stakeholders and the public. The second of these approaches (the predictive assessment) involves assessing risk in terms of probability / hazard (based on indicative flood maps) and consequences. The consequences that can arise from flooding include those that to human health, the environment, cultural heritage and economic activity, and are very diverse in their nature and consequently the way in which they can be measured. Detailed methodologies are available for assessing risk to some sectors, although monetarised assessments of risk are difficult to achieve with a high degree of confidence for many types of consequence. For the purposes of the PFRA, a simple but yet robust and transparent methodology is required. To undertake the PFRA in Ireland, the wide range of potential consequences has initially been identified based on the OPW Scoping Report on Flood Risk Indicators, Methods and Datasets (2008). These consequences have been linked to indicators of risk, based on geo-spatially identifiable receptors, and assessed in terms of the potential vulnerability and importance of the receptors that could suffer harm in the event of a flood, using a common, numeric, but non-monetary, metric to enable a consistent comparison of vulnerability and risk between sectors and to enable an integrated measure of risk to be established. This integrated measure of risk forms the basis for the identification of Areas of Potentially Significant Flood Risk. This paper will set out how this process has been undertaken, focusing on the common metric methodology adopted and how vulnerability classifications were assigned across the different sectors, including those of the environment and cultural heritage. Introduction The Preliminary Flood Risk Assessment (PFRA), as required under the Floods Directive [2007/60/EC], involves the analysis of potentially significant flood risk, based on an assessment of past and potential future floods and their adverse consequences (according to the needs of each Member State), using available and readily-derivable information. An assessment of risk needs to consider the two elements of the risk function, namely probability of a flood event (or range of events), and the consequences or degree of damage that occurred, or is expected to occur, as a result of the event(s). The PFRA in Ireland is being undertaken using three approaches: Use of historic data on past floods A predictive assessment of flood risk Consultation with relevant stakeholders and the public The first of these has involved a review of information collated on historic flood events, as collated and verified for the development of the national historic flood database, and published through the flood maps website ( There is however very little direct risk data available within the reports on past floods that were collated for the database with, for example, only 5% of reports containing information on the number of properties affected, and less than 1% of reports providing information on the scale of economic damage. This data source is hence of limited value in directly estimating flood risk, although the direct risk information has been used where available to determine past floods that had significant adverse consequences and hence areas where the flood risk might be deemed significant. The remainder of the data has not been discarded however, and has rather been used as a surrogate for flood hazard, with the number of events recorded within an area that could be vulnerable to damage in the event of a flood being used to validate the predictive assessment.

2 The second approach has involved the development of national indicative flood maps for a range of types of flooding that could give rise to significant flood risk (e.g., fluvial, coastal, pluvial and groundwater flooding), and the assessment of the potential damages such floods could cause. This approach is described in more detail herein. The third of the approaches is used to ensure that the views and knowledge of stakeholders and the public are captured to enhance the evidence-base of the overall process, and in particular to validate (or otherwise) the findings of the first two approaches, and to highlight any areas of potentially significant risk that might not have been captured by these other approaches. This paper sets out how the risk, and particularly the consequences, has been assessed in Ireland in the implementation of the predictive assessment for the PFRA. Overview of the Predictive Flood Risk Assessment The predictive flood risk assessment for the PFRA comprises three main elements. These cover the definition of 'flood hazard areas and of 'flood receptors'. The third stage involves combining these two data sets to evaluate 'flood risk'. 'Flood hazard areas' are the predicted extents of flood inundation for extreme events of a range of probabilities. Flood hazard areas have been defined nationally for a range of flood types 'Flood receptors' can be defined as people, places, objects or activities that would suffer harm or damage in the event of a flood. The receptors reflect the impact of flooding on human health, the environment, cultural heritage and economic activity. 'Flood risk' is a combination of the probability of an area or receptor being flooded with the impact, or consequences, if that area or receptor were to flood. Both the flood hazard areas and the flood receptors have been compiled and reviewed in a Geographical Information System (GIS). This approach has enabled layers of spatial data covering different flood mechanisms and different types of receptors to be used to define flood risk. Development of Indicative Flood Mapping The first stage or element of the process involves the development of indicative flood mapping to define the projected probability of flooding, or flood hazard. This has been done for a range of different sources and for at least three flood event probabilities, typically the flood events with a 10%, 1% and 0.1% annual exceedance probability (AEP) of occurrence. Fluvial Flooding A national set of indicative fluvial flood maps were developed based on estimated flood flows at nodes at 500m centres along the national river network (derived from a parallel programme of hydrological research), and the application of the normal-depth method using cross-sections derived at each node through a national DTM. The levels derived from the normal-depth method were extrapolated out along the cross-sections until they intersected the DTM (the valley edge) and then interpolated between sections to create a continuous map of flood extents. Validation of this method against recorded flood extents and against maps or predicted flood extents developed through the use of detailed hydraulic modelling have shown acceptable degrees of accuracy for the purposes of the PFRA. Coastal Flooding Under a parallel project, the OPW has been developing strategic coastal flood mapping based on 2- dimensional off-shore modelling and the horizontal inland extrapolation of the projected extreme sea levels over the national DTM to derive indicative flood maps for the relevant flood events.

3 Pluvial Flooding The OPW has commissioned a national pluvial screening project, whereby indicative pluvial flood extents and depths have been identified using a range of rainfall durations and intensities that are run through a rapid flood spreading model (using the national DTM as terrain data). Groundwater Flooding The OPW has commissioned a scoping study of groundwater flooding in Ireland, which has concluded that the principal source of groundwater flooding is turloughs; a karstified limestone feature common in the west and north west of Ireland where groundwater within the karst system emerges above ground following prolonged periods of wet weather. The potentially flood-prone areas from the turloughs have been determined based on historic evidence and some indicative projections of peak groundwater elevations (and the spatial projection of resultant flooding) derived from analysis of recorded data. Other Sources The potential for flooding from a range of other sources, such as raised canals, reservoirs, etc., are also being assessed using suitable methods as relevant to each source. Assessment Of Potential For Flood Damage Flood Risk Criteria In order to determine areas of flood risk across the country, data on flood risk receptors is required that can then be assessed in combination with the flood hazard extents described above. The receptors have been classified under the following four broad categories aligned to the groups of consequences referenced in the Floods Directive: Social: Social receptors include the impact of flooding on human health and community infrastructure. The effects on human health could range from death and personal injury to stress and anxiety due to the loss of personal possessions and, in some cases, enforced relocation to alternative properties whilst repairs to homes are completed. Community infrastructure comprises social services including education, emergency services, government administration buildings and emergency and long-term health service organisations. Economic: This covers infrastructure with economic importance such as utility supplies and transport services, plus agricultural areas. In addition, the flooding of commercial properties will impact on the function of the business and would also cause damage to 'matters of value' located inside the property. Environment: For the PFRA, receptor data for the natural environment has been obtained for areas of designated habitats and species across Ireland. Cultural Heritage: These receptors cover sites of importance from the man-made environment including architecturally significant sites, national monuments, museums and galleries. The receptors, and receptor date, used for the purposes of the PFRA were based on the research previously commissioned by the OPW into potential indicators of flood risk (OPW, 2008), taking into account a range of factors including availability, format, spatial representation, national coverage and data content (classification, description, level of detail, etc.). Assessment of Receptor Vulnerability There is a great diversity of flood receptors that have been considered under the PFRA, with impacts or damage associated with each receptor manifested in a very different ways. There are methods available for the detailed and / or quantified assessment of certain aspects of the damages associated with some of these receptors, such as the monetarised assessment of economic damages to properties. Methods are also emerging or under development for some aspects of risk, such as intangible damages for human health. There are not however such methods available for all categories of risk, and these methods can be data intensive to apply (e.g., require detailed local information), which might be possible or considered appropriate for a broad scale screening exercise such as the PFRA. These methods also measure risk in different ways, i.e., use different indicators or scales.

4 To derive an indicative estimate of risk across a broad range of categories, it is considered useful to use a consistent metric or measure, such that risk under one category can be directly compared against another, or indeed that accumulations of risk in area covering different categories can be combined, and compared against other such area. One such consistent measure is based on monetarisation of risks, which is compatible with assessments of economic risk, but is difficult to achieve without gross assumption and with a sufficient degree of accuracy and confidence for some other categories of risk (e.g., environmental risk). It was concluded in Ireland that a monetary or fully quantitative assessment using a consistent metric of flood impact across each and all of the receptor types, including risk to society, human health, cultural heritage and the environment, would not have been possible, or indeed necessary, at the PFRA stage with sufficient reliability based on available or readily-derivable information. The not insignificant work involved would further not be warranted given the residual uncertainties in the flood hazard extents. It was hence decided that an alternative semi-quantitative method, that was relatively simple yet transparent and robust, of evaluating the impact of flooding across all the receptor types would be developed. The first part of this method, with the aim of distinguishing the impact of flooding between the different receptor types, is summarised in the following stages: Stage (i) The Principles of Defining the Impact of Flooding for Each Receptor The flood impact for each receptor, or sub-group of a receptor, can be represented as a function of the receptor importance and potential damage from flooding. The importance of a receptor is based on its function, scale of influence and ease of replacement. For example, an international airport providing national and international flights will have a greater importance than a local, private airfield. Potential damage from flooding could, for example, be defined as the degree of permanent damage to a receptor or the impairment of the ability of that receptor to function as a service to the local, or wider, population. Higher importance and higher potential damage equate to higher levels of vulnerability (see the description of the following stage) to flooding for the receptor. Stage (ii) The Basis for Assessing Receptor Vulnerability The combined function of level of importance and the potential damage from flooding has been used to define the flood vulnerability of each receptor. The vulnerability scale used in this method comprises five categories ranging from Critical to Low as listed in Table 1. Table 1: List of Vulnerability Classifications Critical Vulnerability Extreme Vulnerability High Vulnerability Moderate Vulnerability Low Vulnerability Stage (iii) Defining Vulnerabilities According to the Types of Flood Impact Flood vulnerability has been systematically assessed using a set of tables to relate receptor importance and potential damage. The specific layout of these tables depends on the type of receptor and therefore the impact of flooding involved, with tables developed for impact on risk to life, damage to matters of value, loss of services and impact on emergency response, and with each receptor type or sub-type being assessed against the tables to define a vulnerability classification. As an example, the table for the damage to matters of value is set out in Table 2 below. An area of vulnerability assessment that has been a particular challenge has been the area of potential environmental and cultural risk. Exploration has revealed very limited information on the vulnerability of receptors within these sectors to flood damage, and potential for damage, or adverse consequences, arising from flooding can vary enormously between the hundreds of different sub-types of environmental and cultural receptor. To address this information gap an assessment has been undertaken as part of the work for the PFRA in Ireland to derive a vulnerability classification for each sub-type of cultural and environmental receptor as they occur in Ireland.

5 Table 2: Flood Vulnerability for Damage to Matters of Value PROBABLE IMPACT DEGREE OF RECEPTOR IMPORTANCE / DESIGNATION International National Regional Local Total Loss Critical Critical Extreme High Severe Degradation Critical Extreme High Moderate Moderate Damage Extreme High Moderate Low Minor Damage High Moderate Low Low No / Negligible Damage Low Low Low Low An example of the application of this method is given below with respect to environmental risk, although similar approaches were taken with respect to other types and sub-types of receptor under each category of risk. Application of Method for Environmental Risk In relation to the application of this method with respect to environmental risk, a stepped process has been followed, as set out below. The list of all protected species and habitats under the Habitats [92/43/EEC] and Birds [79/409/EEC] Directives were available from national datasets, along with the Natura 2000 sites where they are present. Each of the species and habitats has been assessed for its probable impact (or inherent vulnerability) from the effects of flooding. In some instances, species or habitats are water-based, would benefit from flooding or would be resilient to the effects of flooding, flood and were hence determined to suffer no or negligible damage in the event of a flood. Such species and habitats include the Harbour Porpoise (Phocoena phocoena, a marine species found in coastal waters), the Petalwort (Petalophyllum ralfsii, that is present in damp, calcareous dunes and often subject to coastal flooding) and Alluvial Forests with Alnus Glutinosa and Fraxinus Excelsior (Alno-Padion, Alnion incanae, Salicion albae, found on floodplains and adapted to occasional flooding). For other species or habitats however it was deemed that the degree of probable adverse impact of flooding would be more severe, such as for the Corncrake (Crex crex, a groundnesting bird that can nest in floodplain areas, for which flooding could cause severe damage during the nesting season) or Fixed Coastal Dunes with Herbaceous Vegetation (grey dunes, for which severe coastal flooding could cause moderate damage). Having assessed the probable impact, the vulnerability classification for the species or habitat was determined based on the degree of importance. All listed species and habitats are deemed to have national importance, while priority species and habitats, or those that are of particular rarity in Ireland, were designated as being of international importance. Hence those listed species and habitats of no or negligible damage were assigned a Low vulnerability classification, while those of Severe or Moderate probable impact that were listed (but not priority) would have been assigned Extreme or High vulnerability classifications respectively. There often several listed species or habitats within a protected Natura 2000 site. On the basis of the precautionary principle, the vulnerability classification of each site was taken as the highest of those for the species or habitats within the site. The area of any given site that was prone to flooding did however vary significantly, from less than 0.1% up to 100%, and this needs to be taken into account in assigning the final vulnerability classification for the site, which is the spatial unit assessed as to whether it would constitute an area of potentially significant flood risk. Summary As set out above, it was not considered that the implementation of a monetarised or fully quantitative approach using a consistent metric of risk for all categories of consequences could not be reliably implemented with the information that is available or readily-derivable. An alternative approach was hence developed that is transparent and robust, and can be readiliy-impletmented using the information available. It is clearly recognised that this approach is subjective, although consultation with experts is underway to ensure that assessments of potential vulnerability are reasonable, and broader consultation has been, and will be, undertaken to ensure general agreement with the relative degrees of vulnerability between sectors.

6 Predictive Flood Risk Assessment The Flood Risk Index A Flood Risk Index scoring system was developed to determine the degree of risk combining the vulnerabilities of the different receptors with the probability of the flood concerned, where higher receptor vulnerability classifications and lower probabilities of flooding (i.e., more frequent flooding) translate to higher levels of flood risk. The basis for the Flood Risk Index system is shown in Table 3. This approach means that every receptor located inside a flood extent can be assigned a Flood Risk Index score, the magnitude of which depends on the vulnerability class and the probability of flooding of that receptor. For each vulnerability class, the flood risk index score varies according to the probability of the flood event concerned. The multiplying factor of ten between the scores for the different flood events is a reflection in the difference in percentage probability of the flood extents, between the 10%, 1% and 0.1% AEP events. For example, a receptor classified with Extreme Vulnerability to flooding would receive a score of 2500 if it is located inside the 10% AEP flood extent, but only a score of 25 if is located within the lower probability 0.1% AEP flood extent (i.e. with the former flood event 100 times more likely to occur in any given year than the latter event, hence the score that is 100 times greater). Table 3: Vulnerability Class Flood Risk Index Scores Derived from Vulnerability Classes and the Probability of Flooding Vulnerability Class Factor Probability of Flood Event (Annual Exceedance Probability) 10% - High 1% - Medium 0.1% - Low Critical Vulnerability Extreme Vulnerability High Vulnerability Moderate Vulnerability Low Vulnerability Areal Assessments of Risk The Flood Risk Index method assesses the risk for any individual risk receptor. While a single risk receptor might be deemed to constitute, or be at, significant flood risk in its own right, it is equally recognised that a group of receptors may exist within a defined area which individually might not constitute significant risk, but which if considered together could constitute significant risk. There is therefore a need to assess accumulations of risk on an areal basis. The following stages were developed to identify spatially combined flood risk within defined areas: Stage (i) Stage (ii) Stage (iii) A 500m by 500m grid system was developed across the country in GIS to provide a systematic method for evaluating the aggregate flood risk at a settlement scale. For a given flood type, the receptors located inside the flood extents were identified. Any grid square containing one or more flooded receptor was identified as a risk square. From the defined vulnerability classes for each of the receptors and on the basis of the probability of flooding, each flooded receptor was assigned a Flood Risk Index score in accordance with Table 3. The different scores for the flooded receptors for each individual grid square were summed to determine a total Flood Risk Index score for each grid square. A GIS routine was developed to combine any adjacent (horizontally, vertically or diagonally) risk squares into 'amalgamated areas'. The Flood Risk Index scores for each of the amalgamated sets of grid squares were determined from a sum of the scores of the constituent grid squares. It was also possible that individual risk squares may be left from this process where no nearby receptors in the surrounding grid squares are inside the flood hazard areas. These individual squares would be assessed on the basis of the Flood Risk Index score for that specific square.

7 An example of an amalgamated area is shown in Figure 1, with the at risk squares highlighted in red. Figure 1: Example of an Amalgamated Area Definition of Significant Risk The Flood Risk Index provides a consistent metric for the measurement of risk across all sectors, and for individual risk receptors as well as for whole communities or other defined spatial areas. On the basis of the Flood Risk Index, it is then necessary to define a threshold, or value of the Flood Risk Index, that would constitute Significant risk such that these areas or receptors would be defined as areas of potentially significant flood risk (APSRs), as required under Article 5 of the Directive. This is not a simple question to address. To define a threshold for significant risk that is too high will result in an excessively limited number of APSRs, with other areas not defined as APSRs even though there is a widely held public or political view that flood risk in these areas is a problem that needs to be addressed. Conversely, if the threshold is set too low, then an excessive number of APSRs will be defined, leading to a major increase in the cost of implementing the remaining elements of the Directive (flood mapping and preparation of flood risk management plans), and possibly more importantly an unrealistic expectation would be set that measures would be viable for all of the APSRs (noting that with a low threshold of significance the actual degree of risk might be quite low, potentially limiting the scope for economically viable measures in some areas).

8 A provisional threshold for significance has been set in Ireland that has been determined to be appropriate in relation to historic experience of flooding, perceptions of flood risk and realistic expectations that some form of flood risk management measures will be economically viable for the area. This is however subject to ongoing consultation and review and hence is not stated herein. Summary The PFRA is a high-level screening exercise, based on available and readily-derivable information, to identify those areas where potentially significant flood risk exists, and where more detailed assessment is required. It has been concluded that robust and quantitative methods for determining flood risk across all sectors of possible risk using a consistent metric are not currently available, or could not readily be implemented. The OPW has therefore developed a system whereby flood risk can be assessed across all sectors to provide a single metric of risk, based on: Indicative flood mapping to determine the probability of flooding for each receptor Assessments of vulnerability for each type and sub-type of receptor An integration of probability and vulnerability for each receptor to determine a measure of flood risk as a Flood Risk Index score An assessment of the spatially-cumulative degree of flood risk for groups of receptors The identification of APSRs based on individual or cumulative Flood Risk Index scores While the method developed is based on some subjectivity in terms of the assignment of the flood risk vulnerability classification, and the integration of vulnerability and probability to derive a Flood Risk Index score, it is considered that this process is transparent and appropriate for the purposes of the PFRA. The subjectivity and the process as a whole is being, and will be, tested through consultation with experts, stakeholders and the public to ensure that assumptions made, while subjective, are generally accepted or supported. It should be noted that the methodology and findings of the assessment of flood risk under the PFRA as set out herein is work-in-progress and subject to ongoing consultation, and hence should not be taken as final or definitive. References OPW, 2008: Halcrow Group Ireland Ltd for The Office of Public Works (September 2008), Flood Risk Assessment: Indicators, Methods and Data Sets Scoring Study Acknowledgements The authors would like to thank the staff of the OPW and of Mott MacDonald Group Ltd who have contributed to the development and implementation of the methods set out herein.