Upload
others
View
10
Download
0
Embed Size (px)
Citation preview
www.gov.uk/defra
National Risk Assessment: Coastal Flooding
Impact Analysis.
Methodology Report
January 2017
Joint Flood and Coastal Erosion Risk Management
Research and Development Programme
National Risk Assessment: Coastal Flooding
Impact Analysis. Methodology Report
Defra Ref: FD2676 / FD2697
HSL Ref: MSU/2016/29
Produced: January 2017
Funded by the joint Flood and Coastal Erosion Risk Management Research and Development Programme (FCERM R&D). The joint FCERM R&D programme comprises Defra, Environment Agency, Natural Resources Wales and Welsh Government. The programme conducts, manages and promotes flood and coastal erosion risk management research and development.
This is a report of research carried out by the Health and Safety Laboratory and HR
Wallingford, on behalf of the Department for Environment, Food and Rural Affairs
Research contractor: Health and Safety Laboratory and HR Wallingford Ltd.
Authors:
Health and Safety Laboratory: Tim Aldridge, Oliver Gunawan, Kirsty Forder, Peter
Rastall, Helen Balmforth (Technical Review), Charles Oakley (Editorial Review)
HR Wallingford: Ben Gouldby, Dominic Hames, David Wyncoll, Mike Panzeri and Gordon
Glasgow
University of Southampton: Ivan Haigh
Publishing organisation
Department for Environment, Food and Rural Affairs Flood Risk Management Division, Nobel House, 17 Smith Square London SW1P 3JR
© Crown copyright (Defra); 2017
Copyright in the typographical arrangement and design rests with the Crown. This
publication (excluding the logo) may be reproduced free of charge in any format or
medium provided that it is reproduced accurately and not used in a misleading context.
The material must be acknowledged as Crown copyright with the title and source of the
publication specified. The views expressed in this document are not necessarily those of
Defra. Its officers, servants or agents accept no liability whatsoever for any loss or
damage arising from the interpretation or use of the information, or reliance on views
contained herein.
Contents
Executive Summary .............................................................................................................. i
1. Introduction ...................................................................................................................... 1
Background ...................................................................................................................... 1
Aim and objectives ........................................................................................................... 2
2. Scenario Generation ........................................................................................................ 4
Method steps .................................................................................................................... 4
Input Data ......................................................................................................................... 4
Implementation/application ............................................................................................... 7
Example outputs .............................................................................................................. 9
3. Hazard Modelling ........................................................................................................... 11
Method steps .................................................................................................................. 11
Input Data ....................................................................................................................... 11
Implementation/application ............................................................................................. 13
Example outputs ............................................................................................................ 16
Flood hazard rating ........................................................................................................ 18
4. Impact Assessment........................................................................................................ 19
Method steps .................................................................................................................. 19
Input Data ....................................................................................................................... 19
Development of Impact Assessment Metrics ................................................................. 25
Implementation/Application ............................................................................................ 35
5. Discussion ..................................................................................................................... 37
5.1. Flood scenario generation and hazard modelling methods ..................................... 37
5.2. Impact Assessment methods .................................................................................. 38
6. Conclusions and Recommendations.............................................................................. 40
6.1. Conclusions ............................................................................................................. 40
6.2. Recommendations .................................................................................................. 40
7. References .................................................................................................................... 42
Appendix I: Local Resilience Forums and constituent Local Authorities ............................ 46
Appendix II. 2011 Census calculations for population vulnerability .................................... 51
Appendix III. List of impact datasets and sources .............................................................. 53
i
Executive Summary
Background
The risk of a large-scale coastal flood impacting the East coast of England is included in the UK’s National Risk Register, and considered as one of the highest priority risks (Cabinet Office, 2013). Preparing for such scenarios requires an understanding of the risk, based on up-to-date impact assessments that apply the latest understanding in scientific methods and techniques. This report details the methodology applied for assessment of the coastal flooding risks for England and Wales. The assessment was completed by HSL, the Health and Safety Executive’s Laboratory, HR Wallingford and the University of Southampton. The results of the analysis are included in a separate report.
Objectives
The aims of the work were to prepare improved evidence for coastal flooding hazard and impact affecting the English and Welsh coastline to support the entry for coastal flooding (reference H19) in the UK’s National Risk Assessment (NRA). The assessment modelled five extreme but plausible scenarios based on historical floods and related statistical and hydraulic modelling. Impact models were required to assess impacts on population, property, infrastructure, transport and agriculture. This assessment considered potential hazards and impacts within a maximum timeframe of the next 5 years. Changes in risk as a result of potential changes in climate, such as changes in sea levels, are therefore considered negligible within this timeframe, and not considered.
Main Findings
The aims of the assessment were achieved by identifying and setting up hazard and impact methodologies to support the following tasks:
Task 1: Generation of flood scenarios. Statistical modelling based on historical data was used to develop a wide range of possible flood scenarios. The results of this were analysed and, through consensus across project partners, five scenarios selected. The scenarios provide full coverage of the English and Welsh coastline and provide relevant, realistic worst-case scenarios for provision of NRA evidence. The flood scenarios are not predictions of what will occur - they are hypothetical scenarios that could arise. In reality, future coastal floods could be more or less severe and could influence different geographical regions in different ways.
Task 2: Hazard Modelling. HR Wallingford conducted hydraulic modelling of the five reasonable worst case scenarios to simulate the process of flood inundation. These simulations incorporated a series of flood defence breaches.
ii
The location and number of breaches was based on evidence of historical floods and discussed and agreed with project partners. The hydraulic modelling provided data on flood depths and velocities across the floodplain for the five scenarios.
Task 3: Impact Assessment. This task involved developing an updated, fully national multi-criteria receptor database for England and Wales. Data was sourced and mapped from government departments and infrastructure management organisations. The receptor database includes information required to derive physical and economic impact metrics.
Key metrics were selected to provide a detailed picture of flood impact across receptors. These include physical measurements (counts, percentages, lengths and areas impacted) and economic costs (damage to property and agricultural land). These metrics were produced as impacts in their own right, but also as inputs into more sophisticated metrics that describe wider economic impacts.
Task 4: Interpretation and presentation of results. The outputs of the flood impact assessment were formatted into a spreadsheet-based output template for easy access to the assessment results. The template enables users to access results at a range of spatial scales (from overall scenario results down to Local Authority boundaries) and at varying levels of detail.
Recommendations
The methods described in this report fulfilled the specified aims of the work. In collating the methods, the project team highlighted model sensitivities and issues from which recommendations can be made for future development of the impact assessment:
The NRA requires the estimation of the likelihood of the episode
occurring. It is desirable to adopt a risk based approach to likelihood
specification, whereby likelihood of impact or consequence is the metric
of relevance. This does, however, require the hydraulic simulation and
impact evaluation of many more scenarios, not feasible on this project.
Future work should however, consider the viability of implementing a fully
risk-based approach to scenario likelihood estimation.
The modelling of flood impacts on people remains a challenge and this
research has highlighted this further. The FRTP method uses relevant
information to provide useful indicators that help in the understanding of
the risk but there are still significant uncertainties about the sensitivity of
the fatality and injury estimates, and finding the best way of
communicating these results. Future research could evaluate the
potential to amend FRTP assumptions, equations and parameters as
iii
well as introducing measures of uncertainty into the analysis.
Comparison of FRTP with other methods that provide smaller scale
population impact estimates may help to understand FRTP limitations
and calibrate the model.
Challenges present in the modelling of transport and infrastructure
impacts have highlighted the possibility for more sophisticated network
resilience analysis. This might include the analysis of diversionary or
evacuation routes and related impacts to commuting or emergency
response. There may also be scope to improve analysis of the impact on
utility services including subsequent impacts on supply to residential and
commercial properties, including further economic or social impacts.
Collection of more detailed property data including features such as age
or construction materials could allow for a more sophisticated analysis of
building damage, allowing a deeper use of the Multi-Coloured Manual
methodologies and potentially more information on likely repair/rebuild
times, which have an impact on evacuation and shelter costs.
The response to Environment Agency flood warnings is integrated into
the FRTP methodology, but further acknowledgement of flood response
may be useful for improving the counts of impacted people or impacted
sites. This may require more detailed knowledge of local flood risk plans
or individual infrastructure site flood plans.
The hydraulic modelling has a temporal aspect but this is not included in
the scenario results or applied to the impact assessment. Temporal
analysis of impacts has the potential to provide added value and another
aspect to response prioritisation but challenges are present in applying
this effectively and in communicating the results.
The methodology outlined in this report and the associated code has
been developed for repeatability across different types of flooding, where
the extent, onset and composition of the flood waters are likely to differ.
The concepts of the impact assessment component are common to
wider non-flooding contexts and the authors would also encourage
adaptation of the methodology for other applications including other
natural hazards or industrial accidents. Further, the receptor database
created collates information for different types of property, key
infrastructure and service categories and a range of different population
types. Much of the information is not flooding-specific.
The impacts methodology could be simplified and applied to statistical
flood scenario generation to produce a novel impact-based risk
assessment. Applied in this way, the model would allow for the
iv
estimation of risk, based on impact severity and likelihood of occurrence.
This could be a valuable tool for development of evidence for the
National Risk Assessment and for other flood impact applications.
Environmental impacts beyond those to agriculture are not currently
included in the assessment. These may include the impacts of prolonged
salt water inundation or the impacts of the release and diffusion of
pollutants and other dangerous materials by floodwater into the wider
environment.
The current model does not yet consider the social or psychological
impacts of flooding. Awareness is steadily growing of these chronic
impacts, which include stress, anxiety and depression. These impacts
are amongst the most challenging features to measure and quantify. It is
anticipated that further research into these areas could build on current
indices based on community characterisation by socio-economic data,
there is also potential to build collaborations with organisations of social
scientists and psychologists to explore alternative approaches.
1
1. Introduction
Background
The risk of a large-scale coastal flood impacting the East coast of England is included in
the UK’s National Risk Register, and considered as one of the highest priority risks
(Cabinet Office, 2013). Such flooding has precedence in recent history. The impacts of
the large-scale flooding of the night of 31 January 1953 are well known; 307 people lost
their lives in England, and there was widespread damage to property and disruption to
normal activities (Baxter, 2005). Significant damage and loss of life was also experienced
in the Netherlands where a further 1835 people perished (Jonkman and Vrijling, 2008).
This flood was unusual in its magnitude, but it was the result of tidal and meteorological
conditions that are likely to reoccur.
The storm surge of 5-6 December 2013 on the East coast of England and Scotland
resulted in the largest scale flood of that type since 1953. The forecast and flooding
prompted severe flood warnings to be issued by the Environment Agency (EA). Reports
from the incident outline less severe impacts than were forecast. Impacts included flooded
homes, evacuation of residents, and the collapse of property (Met Office, 2014; BBC,
2013).
Organising a response for such a flood requires planning and cooperation at the national
and local level. Since 1953, improvements to understanding and planning for a large scale
flood have been extensive, as have the improvements to flood defences, mitigation and
awareness schemes; this is based on a better understanding of the risk, which has been
informed by research work and impacts analyses. Due to the high risk of coastal flooding
highlighted in the National Risk Register, there is continual demand for up-to-date flood
scenario assessments that apply the latest understanding in scientific methods and
techniques.
This report details the methodology applied for the development of the flood scenarios and
impact of coastal flooding along the East, South and West coastline of England and
Wales. The assessment was completed by HSL, the Health and Safety Executive’s
Laboratory, HR Wallingford and the University of Southampton. Analysis of the historical
coastal flood record allied to statistical modelling has defined the five extreme but plausible
flood hazard scenarios. Impact information was collected for five groups of receptors:
Population, Property, Infrastructure, Transport and Agriculture. Mathematical algorithms
were applied to convert direct impacts into tangible metrics within the five receptor groups,
which were then translated into wider economic costs. Flood impact assessment results
are presented for each of the scenarios in an accompanying results report.
2
Aim and objectives
The latest NRA update required improved evidence for coastal flooding with extended
coverage along the entire English and Welsh coastline. Flood scenario modelling needed
to be improved to increase the relevance of the impact analysis, and to ensure that
subsequent NRA guidance is proportional and grounded in current science. This aim was
achieved through completion of four tasks as demonstrated in Figure 1.1 and outlined in
more detail below.
Task 1: Generation of flood scenarios. Statistical modelling based on historical data
was used to develop a wide range of possible flood scenarios. The results of this were
analysed and, through consensus across project partners, five scenarios selected. The
scenarios provide full coverage of the English and Welsh coastline and provide relevant,
realistic worst-case scenarios for provision of NRA evidence. The flood scenarios are not
predictions of what will occur - they are hypothetical scenarios that could arise. In reality,
future coastal floods could be more or less severe and could influence different
geographical regions in different ways (Chapter 2).
Task 2: Hazard Modelling. HR Wallingford conducted hydraulic modelling of the five
reasonable worst case scenarios to simulate the process of flood inundation. These
simulations incorporated a series of flood defence breaches. The location and number of
breaches was based on evidence of historical floods and discussed and agreed with
project partners (Chapter 3)
Task 3: Impact Assessment. This task involved developing an updated, national multi-
criteria receptor database for England and Wales. Data was sourced and mapped from
government departments and infrastructure management organisations. The receptor
database includes information required to derive physical and economic impact metrics.
Key metrics were selected to provide a detailed picture of flood impact across receptors.
These include physical measurements (counts, percentages, lengths and areas impacted)
and economic costs (damage to property and agricultural land). These metrics were
produced as impacts in their own right, but also as inputs into more sophisticated metrics
that describe wider economic impacts (Chapter 4).
Task 4: Interpretation and presentation of results. The outputs of the flood impact
assessment were formatted into a spreadsheet-based output template for easy access to
the assessment results. The template enables users to access results at a range of spatial
scales (from overall scenario results down to Local Authority boundaries) and at varying
levels of detail.
3
Figure 1.1. Description of tasks required for flood impact assessment including key
processes. The Tasks within the red box are covered in this report.
Throughout the project, emphasis was placed on using well-established flood risk
methodologies and information sources. This includes alignment with National Flood Risk
Assessment (NaFRA) methods (EA, 2009a), evaluation of surface water flooding (EA,
2014), impact assessment methods recommended by the Multi-Coloured Manual
(Penning-Rowsell et al. 2013) and The Flood Risk to People methodology (HR
Wallingford, 2006).
Discussions with Defra and EA highlighted significant benefits in aligning H19 coastal flood
risk assessment to EA H21 project, which focuses on widespread inland and coastal
flooding. This alignment applies to the general approach and the Impact Assessment task
undertaken by HSL which uses the same methods and datasets. The EA H21 work is
being documented in a separate report.
Task 1:
Scenario generation
Task 2:
Hazard Modelling
Task 3:
Impact Assessment
Task 4:
Interpretation and
presentation of results
Analysis of historical events
Fitting a multivariate statistical model and simulation of a large sample of events
Scenario selection
Transformation of offshore conditions to nearshore
Assessing and identifying potential breach locations
Transformation of nearshore conditions into wave overtopping/overflow rates
Simulation of water propagation across floodplain
Collation and formatting of receptor datasets
Depth damage curves and agricultural damage
parameters from Multi-Coloured Manual
Flood Risk to People calculations
Intersection of flood hazard and impact datasets
Calculation of flood impacts
Aggregation of results to Local Authority level
Description of multi-scalar results template
Presentation of headline statistics and maps
High level narratives
Chapter 2
Chapter 3
Chapter 4
Separate results report
4
2. Scenario Generation
Method steps
The methodology for the extreme flood scenario generation comprised three main
components:
Analysis of historic coastal floods, including storm tracks and surge levels.
Fitting a multivariate extreme value statistical model and simulation of a large
sample of extreme sea levels from the fitted model.
Selection of the scenarios used for the extreme flood generation.
The relevant data sources and the implementation of these steps are described below.
Input Data
Analysis of historical coastal floods
There is a long history of coastal flooding in the UK and this information can be used to
assess the validity of the scenarios that are used. The intention of the statistical model is
to simulate floods that have the same spatial characteristics as floods that have been
observed in the past but are more extreme. With coastal flooding being driven by large
surges in many areas, this assessment utilised a database of historical coastal floods
collated by The University of Southampton (Haigh et al. 2015). This database has
identified a total of 96 events and comprises information on:
Storm tracks that caused coastal flooding
Surface pressure fields associated with each occurrence of coastal flooding
Surge levels associated with each coastal flood
Qualitative information on flood impacts
This information is illustrated in Figures 2.1, 2.2 and 2.3. This information acts to provide a
source of verification of the statistical modelling work that has been undertaken. This is
described further below.
5
Figure 2.1 Summary of return period1 of historical surges. Haigh et al (2015):
http://www.surgewatch.org/
Figure 2.2 “Footprint” of the 5th/6th of December 2013 East Coast event (return period
(years), surge levels (m)). Haigh et al (2015): http://www.surgewatch.org/
1 Return period is the reciprocal of the annual exceedance probability. Hence a 1 in 100 year return period event has
an annual exceedance probability of 1%.
6
Figure 2.3 Example storm track and pressure field for the January 2014 event. Haigh et al
(2015): http://www.surgewatch.org/
Waves, wind and sea level data
The relevant variables used in this analysis include information on waves, winds, surges
and tides. These data have initially been analysed offshore. The data sets containing
these variables have been obtained from the “A Class” National Tide and Sea Level
Facility network of tide gauges managed by EA and an historic hindcast analysis of waves
and winds derived by the Met Office using the WaveWatch III (WW III) model. The
locations of the various data sets used in the analysis are shown in Figure 2.4. The
locations of SWAN wave transformation models used in the analysis are also shown. The
SWAN models were not set up specifically for this project. Rather they were re-used from
existing studies. The SWAN model for Wales was set up by Deltares on behalf of Natural
Resources Wales (NRW). The SWAN models covering England were set up by HR
Wallingford on behalf of EA, under their State of the Nation flood risk project. The choice
of offshore locations differed between England and Wales. For England, the locations
used for EA’s existing State of the Nation project were re-used for this study, since these
already coincided with the boundary of the SWAN wave models. For Wales however, it
was necessary to define a new set of offshore locations to provide appropriate boundary
conditions for the SWAN model. Discussions were held with NRW representatives to
define and agree as appropriate a set of offshore locations at which to undertake the
statistical modelling.
7
Figure 2.4 Locations of data sets used within the statistical model.
Implementation/application
Statistical model
The objective of the statistical modelling was to extrapolate observations of the variables
that influence coastal flooding to obtain floods that are more extreme than past
observations. The floods should however, contain the same spatial characteristics and
dependencies that are observed within the data. Extrapolation of the data requires the use
of specialist extreme value statistical techniques. As it is a requirement to simulate floods
that cover large parts of the country it is a requirement for the statistical model to capture
dependencies, or spatial correlation between the different variables at different locations.
8
The statistical model applied has been developed by Lancaster University (Heffernan and
Tawn, 2004), and is well established for use in flood risk analysis (EA, 2011a; Lamb et al.
2011; Wyncoll and Gouldby, 2013) and specifically for coastal applications (Gouldby et al.
2014). The statistical model was fitted to the data sets and used to generate a large
sample of synthetic extreme floods. Example outputs from the statistical model are shown
below.
Selecting the scenarios
Whilst the statistical model generates many thousands of extreme scenarios, for the
purposes of this study, only five scenarios were to be simulated. Analysis of storm tracks
from the historical database of coastal floods showed clusters of floods that affect different
areas of the coastline (Figure 2.5).
Figure 2.5 Analysis of storm track clusters Haigh et al. (2015): http://www.surgewatch.org/
Based on the analysis of the storm track clusters, scenarios were chosen that were
centred on different locations and affected different geographic regions:
East Coast (North): Humber
East Coast (South): East Anglia and Thames
South Coast: Solent
Severn Estuary
Liverpool Bay
Although they were centred on a particular location, resultant floods can also be extreme
at different places along the coastline because the statistical model captures the spatial
dependence. The scenarios are defined in terms of offshore wave height, period and
direction, wind speed and direction and surge/sea level at an “A” Class gauge. These
9
variables are specified at all locations around the coastline (Figure 2.4). The observed
storm track information has been used to help inform the narratives that have been
developed to accompany each scenario.
Example outputs
An example of the synthetic scenarios that have been generated from the fitted statistical
model is shown in Figure 2.6, for one location off the coast of East Anglia. Also shown are
the observed conditions (green points) that occurred on 5-6 December 2013 and the
conditions selected for simulation as part of the NRA (orange point). Whilst this shows the
degree of dependence between the variables at a particular site, it is also important to
capture the degree of dependence between the variables at different spatial locations.
Figure 2.7 only shows wave conditions but demonstrates the dependence between wave
conditions at different locations. As is expected, wave data from nearby points shows a
high degree of dependence (top left panel). Spatially disparate data points on opposite
coastlines (top right panel) show less dependence.
Figure 2.6 Example output from the statistical for a site off East Anglia.
10
Figure 2.7 Spatial dependence between wave conditions around the coastline
(axis labels Hs(m)).
11
3. Hazard Modelling
Method steps
The scenario generation stage produces estimates of extreme offshore waves, winds and
sea levels at each location around England and Wales identified in Figure 2.4, for each of
the five different scenarios. The objective of the hazard modelling stage is to translate this
extreme boundary condition information into estimates of flood depths and velocities over
the floodplain area. The key stages in this analysis are:
Transformation of the offshore conditions to the nearshore, taking account of
processes such as wave refraction and shoaling.
Assessing the potential for breaches to occur and identifying potential breach
locations.
Transformation of the nearshore conditions into wave overtopping or overflow rates
(i.e. rates of water flowing over or through the defences into the floodplain), to form
the boundary conditions to the inundation modelling.
Simulating the propagation of water across the floodplain using a flood inundation
model.
Input Data
Wave transformation modelling
Wave transformation models covering the coastline of England and Wales were already
available from previous studies. The model domains are shown in Figure 2.4.
Breach location identification
The data used to inform the breach location analysis comprised:
Information on location, type and condition of coastal flood defences stored within
EA Asset Information Management System (AIMS) database and NRW’s Asset
Management eXpert (AMX) system.
EA Continuous Defence Line (CDL) data set.
Topography from EA LiDAR data set.
Information on breach occurrences during the 2013/2014 Winter floods (Figure 3.1).
12
Figure 3.1 Defence breach locations, Winter 2013/2014 (EA).
Defence data and Wave overtopping rates
Information on location, type and geometry of coastal flood defences stored within EA
AIMS database and NRW’s AMX system was used as input to the wave overtopping and
overflow calculations. There were known deficiencies in relation to the crest level of
defences in some specific locations, particularly within the EA’s AIMS database. In these
areas, crest level information was adjusted, in discussion with EA representatives, to be
more appropriate based on extreme sea levels information and knowledge of historical
flooding and general standards of protection. NRW’s AMX data was supplemented with
additional information from NRW’s ongoing data collection programme to ensure the best
available information was utilised.
13
Information on sea conditions output from the wave transformation modelling was used to
provide the boundary conditions for the overtopping model. HR Wallingford’s BAYONET
model (Kingston et al. 2008), was used to undertake the wave overtopping calculations.
Flood Inundation Modelling
The input data relating to flood inundation modelling comprised EA’s and NRW’s 2m
resolution LiDAR data set. The hydraulic boundary conditions were provided by the
aforementioned wave overtopping and flood defence overflow calculations.
Implementation/application
Wave Transformation modelling
For flood inundation simulation it was necessary to transform the offshore wave conditions
to the nearshore, taking account of processes like refraction, wave growth and breaking.
Existing SWAN 2D wave models from other projects (eg. Environment Agency’s State of
the Nation project; HR Wallingford 2015) were used for simulating this process. The
boundaries of the wave models used are shown in Figure 3.4. The offshore information
from each scenario was transformed to the nearshore to a series of points on
approximately the -5m Ordnance Datum Newlyn (ODN) contour. This information was
then transformed through the surfzone using the SWAN 1D model. This then formed the
input to the wave overtopping model (Figure 3.2).
14
Figure 3.2 Conceptual diagram of the nearshore SWAN 1D modelling, linking with the
BAYONET wave overtopping model.
Breach location Identification
Breaches of flood defences can and do occur during extreme conditions, particular when
defences are tested beyond their design standard. During the East Coast surge event of
5th December 2013, a series of breaches occurred. Further breaches occurred during
subsequent storms that followed over the winter period. It is important to consider
breaches during the development of extreme but plausible flood scenarios.
The approach used to specify the number and location of breaches within the scenarios
was judgement-based supported by available evidence and simplified modelling. It is
important to note that the breach locations specified within the modelling are not
predictions of where breaches will occur but rather possible scenarios.
The first stage in the breach identification analysis was to undertake a screening of
topographic information to identify low lying coastal areas that are potentially susceptible
to large scale coastal flooding. These are shown in Figure 3.4 which simply shows land
levels below 10m ODN.
The second stage was to assume an extreme coastal flood (>1000 year sea level) and
undertake a simplified flood inundation calculation to understand the potential impacts of
breaches to a range of receptors. These receptors included transport infrastructure (rail-
lines and roads) and properties (including hospitals and schools). An example of the
analysis for motorways is shown in Figure 3.5, with the highlighted red areas indicating
motorway stretches that are potentially susceptible using the screening impact analysis.
15
Wave overtopping/overflow rate analysis
Even when the sea level is below the crest level of the flood defences, water can still be
discharged into the floodplain due to wave overtopping process. Overtopping rates were
calculated using the BAYONET wave overtopping model (Kingston et al. 2008), which
relates closely to the standard Clash method, described in the EurOTop manual (Pullen et
al. 2010).
The boundary conditions of the flood inundation model were therefore formed by a time
series water level at the coastal defences. When this exceeded the crest level of the
defences, the standard weir equation was used to calculate flow into the floodplain. When
the water level was below the crest level, a time series of wave overtopping rate was
discharged into the cell just landward of the flood defence. These processes are
illustrated in Figure 3.3.
Flood inundation model
The flood inundation model, Caesg, solves the shallow water equations to simulate the
propagation of water over the floodplain. The model was constructed using the best
available information on flood defences, including defence data from the EA’s Asset
Information Management System (AIMS), NRW’s AMX system and topographical
information from EA and NRW’s composite 2m LiDAR dataset. The model used a 50m
regular grid. The model takes as input time series boundary overtopping/overflow
information over each flood defence. Where relevant, breaches were introduced into the
model assuming the initiation occurred at the peak of the hydrograph inflow. The output of
the inundation model is a time series of depth and velocity for each grid cell, over the
flooded area. This information has been used in subsequent impact analysis.
16
Figure 3.3 Conceptual diagram showing wave overtopping and breaching calculations
Example outputs
Example outputs from various processes are provided below to facilitate understanding of
the modelling activities. Example outputs from the breach location identification process
are shown in Figures 3.4 and 3.5 respectively. They show screening analysis undertaken
to identify locations where breaches have the most potential impact.
17
Figure 3.4 Coastal areas below 10m ODN (in blue).
Figure 3.5 An example of the simplified impact of coastal flooding showing motorway
stretches (in red) that are potentially susceptible to extreme flooding.
18
Flood hazard rating
The inundation model outputs were used to provide depth and velocity data for each 50m
cell across the duration of the flood. A flood hazard rating was calculated using the depth and velocity values at the time of the maximum hazard rating over the full flood scenario, with a depth-related debris coefficient. The hazard rating was then classified into categories corresponding to increasing hazard severity (Table 3.1). Hazard scores below 0.575 were removed in alignment with methods used for the updated Flood Map for Surface Water (uFMfSW, EA, 2013).
Table 3.1 Hazard categories (EA and HR Wallingford, 2008).
Hazard
Rating
Degree of
Flood Hazard
Description
0.575 – 0.75 Low Caution
“Flood zone with shallow flowing water or deep standing
water”
0.75 – 1.25 Moderate Dangerous for some (i.e. children)
“Danger: Flood zone with deep or fast flowing water”
1.25 – 2.00 Significant Dangerous for most people
“Danger: flood zone with deep fast flowing water”
>2.00 Extreme Dangerous for all
“Extreme danger: flood zone with deep fast flowing water”
19
4. Impact Assessment
Method steps
The methodology for the impact assessment comprised three main components:
Collection and formatting of receptor datasets into a single standardised receptor
database
Development of impact assessment metrics
Implementation of the impact assessment model including aggregation of results to
Local Authority Boundaries.
The implementation of these steps is described below.
Input Data
Collection of receptor datasets
Receptors are features or elements that are potentially exposed and vulnerable to the
flood hazard. The receptors included in this impact assessment can be categorised into
five groups: Population, Property, Infrastructure, Transport and Agriculture. The best
available information was sourced from government organisations (including the EA,
Department for Transport (DfT), Health and Safety Executive (HSE)), national data
providers (Ordnance Survey (OS), Office of National Statistics (ONS)), and infrastructure
asset owners. Direction was provided by previous flood risk assessment work (EA, 2009a,
Aldridge et al. 2011, Aldridge et al. 2015), as well infrastructure-specific work on climate
change (ITRC, 2013; HR Wallingford, 2014).
Population
The spatial data used for population receptors were largely derived from the National
Population Database (NPD). The NPD is a GIS database providing spatially-referenced
estimates of population numbers for different population types and scenarios. The NPD
was originally created for HSE by Staffordshire University in 2004 and has since been
adapted and improved by HSL, who continue to develop and maintain it (Smith et al.
2005; Smith and Fairburn, 2008). A list of populations included in this assessment is
contained in Table 4.1. Several population layers were created outside of the NPD and are
described below.
The census catchment of vulnerability information in Table 4.1 details the size of census
boundary used to calculate the proportion of people more vulnerable to flooding for a given
population theme. The different sized catchments reflect the fact that the population within
20
different types of property (schools, workplaces, caravan parks etc.) are likely to be drawn
from different sized spatial catchments depending on population theme.
Census data is collected into hierarchical administrative boundaries to preserve anonymity
and provide appropriate data for different applications. The smallest spatial unit available
is the Output Area (OA), which represents an average of 100 residents (ONS, 2016).
Lower Super Output Areas (LSOA) represent an aggregation of OAs and contain an
average of 1,500 residents. Middle Super Output Areas (MSOA) represent an aggregation
of LSOAs and contain an average of 7,200 residents. Local Authorities (LA) represent an
aggregation of MSOAs and are the largest spatial units used in this analysis.
Where population analyses are based solely on households, the most local-level census
information (e.g. OA) is an appropriate choice. LSOAs were used to represent wider
spread daytime residential populations. Evidence from the National Travel Survey (DfT,
2014a) suggests that the average commute to an educational establishment in 2013 was 3
miles which is equivalent to a commuter travelling across an average sized Medium Super
Output Area. In the same survey the average commute to work was 8.8km. This roughly
equates to a commuter travelling halfway across an average sized Local Authority (LA).
LAs were also used to represent larger, regional catchments, including stadia populations.
Table 4.1 Impacted populations aggregated from flood impact data into geographical units.
Population Theme
Breakdown Data used Census catchment of vulnerability
Residential Night-time Day time (term time) Day time (non-term time)
NPD NPD NPD
OA LSOA OA
Sensitive Schools Colleges Care Homes Childcare facilities Hospitals Prisons
NPD NPD NPD NPD NPD NPD
MSOA MSOA 100% vulnerability MSOA 100% vulnerability LAD
Working Population
Weekday Workers Saturday Workers Sunday Workers
NPD/NRD NPD/NRD NPD/NRD
LAD LAD LAD
Leisure Caravan/Camping Sites (peak / low season) Other tourist accommodation (peak / low season) Stadiums (capacity)
AddressBase Premium / VOA AddressBase Premium / VOA NPD
National Average National Average LAD
The workforce layer was derived from the NPD and the National Receptors Dataset (NRD)
(described in the Property section), including the NPD’s temporal scenarios, which model
day time, night time and weekend employment levels.
21
The location and population of camping and caravan sites, and other leisure
accommodation were produced specifically for this project. Locations were derived from
OS AddressBase Premium, Valuation Office Agency (VOA) and Camping and Caravan
Club data. Campsite populations were derived from Camping and Caravan Club data and
online campsite directories. Other leisure accommodation populations were derived from
bed space information contained in VOA data.
Property
The NRD property point data formed the basis of the receptor database for this work
(Table 4.2). The NRD was created by the EA and NRW for flood risk assessment (EA,
2011b). The NRD is based on OS datasets, and aims to locate and attribute all properties
in England and Wales that are addressable or have a floor-level footprint over 25m2.
Attributes include residential type and non-residential usage categories, building footprint
size, indicators for the lowest floor of the property, and unique reference identification
codes such as the Unique Property Reference Number (UPRN; Geoplace, 2016). This
provides the basic information required for estimation for property impact analysis. The
NRD property point dataset was filtered to remove points that did not represent buildings
(e.g. advertising hoardings, telephone boxes etc.), and properties recorded as being above
ground floor. Listed buildings are not explicitly included in the NRD, but are considered key
sites in case of a flood. Therefore listed building locations were acquired from Historic
England and Cadw Welsh Assembly Government.
Table 4.2 Impacted property types and source data sets.
Property Type Source
Residential (Detached, Semi-detached, Terraced, Flats)
Shop/Store
Vehicle Services
Retail Services
Office
Distribution/Logistics
Leisure
Sport
Public Building
Industry
Miscellaneous
Unclassified
NRD Property Points
Listed Buildings Listed Buildings in Wales GIS Point Dataset (Cadw Welsh Assembly Government)
Listed Buildings in England GIS Point Dataset (Historic England)
22
Infrastructure
Infrastructure types and datasets used are listed in Table 4.3. The majority of the
infrastructure sites are located in properties and were therefore represented as points.
Roads and railways were represented as lines. Individual infrastructure types were
grouped into broad infrastructure categories as detailed in Table 4.3. Infrastructure
categories are listed below:
Emergency services are the emergency response providers. This includes police,
ambulance, fire and coastguards. These features are important as the effectiveness of
their response to the consequences of flooding may be adversely affected by the flood
hazard itself.
Key sites are identified as core public sites that either provide essential services or might
create significant societal problems if disrupted by flooding. As such, there is a priority for
the sites to be open and accessible. Key sites include hospitals, schools, doctor’s
surgeries, care homes and prisons.
Utilities provide important services for water or provision of energy. Major outages of
power or water are already present in the NRA as separate risks in their own right, but
these are still significant features in flood impact assessment. They include water
treatment works, sewage pumping stations, power stations and electrical installations.
Potentially hazardous sites are locations that have the potential to cause further harm if
disrupted by flooding. This may be through diffusion of waste or pollutants into the
environment or through emission of dangerous substances into the atmosphere. Such
instances could have serious consequences for danger to life and the environment. These
sites include major hazard sites and industrial sites that produce radioactive or waste
materials that require specific licenses and management.
Transport infrastructure includes the road and railway network as well as transport hubs
such as bus and rail stations. Disruption of the transport network could have serious short-
term consequences during a flood when evacuation routes or emergency service routes
require diversion. In the longer term, impacts on the transport infrastructure may result in
increased traffic and longer journey times with consequences on services, society, costs
relating to lost working hours and other indirect business costs.
Road networks were populated with estimates of vehicle and passenger numbers.
Average Daily Flow data from the Department for Transport provided information on the
type, number and average speed of different types of vehicle passing along each node-to-
node segment of major road in a 24 hour period. The length of the road segment in km
was multiplied by the number of vehicles / passengers / lorries on that segment to produce
metrics for passenger km, vehicle km and lorry km (ITRC, 2013). Larger values indicate
busier and more important road segments. The vehicle km provides a measure of physical
road busyness, the passenger km populates road segments with people and the lorry km
provides a proxy measure for commercial traffic.
23
Table 4.3 Infrastructure types under consideration and data sources.
Infrastructure Category
Infrastructure Type Data Source
Emergency Services
- Fire Stations - Ambulance Stations - Police Stations - Coastguard Facilities
OS Addressbase Premium / VOA
Key Sites - Doctors Surgeries and Health Centres
Care Quality Commission GP practice data
- Hospitals - Care Homes - Schools - Prisons
NPD
Transport - Bus Stations NPD
- Train Stations NPD / National Rail station data
- Roads (including Primary/Trunk roads)
NRD Roads
- Railway (km) NRD Railways
- Ports DfT Transport Statistics PORT0101
Utilities - Electrical Substations - Large Electrical Substations (>100m2) - Power Stations
OS AddressBase Premium / National Grid DECC DUKES 5.10 dataset
- Nuclear Sites HSE library
- Waste Water Treatment Works (WWTW) - Sewage Pumping Stations - Water Treatment Works (Clean water)
EA/NRW Consented Discharge to controlled waters EA/NRW Consented Discharge to controlled waters EA/NRW Consented Discharge to controlled waters
- Petrol Stations OS AddressBase Premium / VOA
Potentially Hazardous
- Major Hazard Sites HSE library
- Waste and Recycling facilities EA (Environmental Permitting Regulations – Waste sites)
- IPPC - RAS Authorities - RAS Registrations
EA (Environmental Permitting Regulations – Industrial sites) EA (Radioactive Substances Register 2011) EA (Radioactive Substances Register 2011)
24
Core infrastructure components
In the previous coastal flooding assessment (Aldridge et al. 2015), HSL was asked by
stakeholders to incorporate the influence that core infrastructure components may have on
the entire infrastructure network. To highlight the importance of these core sites, agreed
infrastructure types were filtered by site size or capacity to identify more significant sites in
the infrastructure networks:
Railway stations. The NPD railway stations layer was enriched with station category data
from National Rail. Categories A and B represent National Hubs and National
Interchanges and were selected to represent major railway stations.
Electrical Substations. The base substation layer derived from OS AddressBase
Premium was joined to National Grid data, which provides data on the largest substations
in the national network (Supergrid and Bulk substations). These substations transfer
energy cross-country to smaller, local substations. The largest Supergrid substations (400
kV) were chosen to represent significant substations in the network.
Power Stations. Although all energy generation sites are important, this project
considered sites that generate above 1000 MW to be large sites. This followed work
conducted by the Climate Change Committee (CCC) (HR Wallingford, 2014) and was
completed using the DECC Digest of UK Energy Statistics (DUKES) database.
Waste Water Treatment Works (WWTW). The EA / NRW controlled discharge to
consented waters dataset was used to subset large WWTW following the CCC report (HR
Wallingford, 2014), using sites that process more than 30,000 cumecs (cubic metres per
second as a unit of flow of water) dry weather flow as a threshold.
Major Hazard Sites. UK major hazard sites are regulated by HSE, EA, SEPA, and NRW
under the European Seveso Directives. The UK implements these directives through the
Control of Major Accident Hazards (COMAH) regulations, which includes categorisation of
sites by the type and volume of hazardous substances stored, and the methods of storage.
Major hazard sites identified as ‘Top Tier’ under COMAH were selected to represent large-
scale sites.
Roads. Trunk roads and motorways are routes of strategic importance in the road
transport network. For this research, classification data in the NRD roads layer were used
to identify trunk roads and motorways as important routes.
Agriculture
Flooding of agricultural land can cause damages with severe consequences for both
arable and pasture farming (Penning-Rowsell et al. 2013). Agricultural land data was taken
from the Agricultural Land Classification (ALC) included in the 2010 version of the NRD.
The layer itself was created in 1988, and remains the most recent version available. The
25
ALC covers England and Wales and separates the entire landscape into 5 grades of
agricultural land (1 highest value – 5 lowest value), urban, and non-agricultural land. For
this research, only the graded agricultural land was used. To highlight damage to the
highest quality land, grades 1 and 2 were also included separately as an additional impact
metric.
Reporting areas
Stakeholder discussion identified that reporting flood hazard impacts at administrative
boundaries will provide summary information that is easier to disseminate and more
relevant to local planners and decision makers. The East Coast update (Aldridge et al.
2015) made use of Local Resilience Forum Areas (LRFs) and their composite Local
Authority (LA) boundaries. As a requirement of the Civil Contingencies Act (2004), LRFs
are multi-agency partnerships created to plan and prepare for localised incidents and
catastrophic emergencies (Cabinet Office, 2013). LRFs are composed of front line
Category 1 responders who include local emergency services, local authorities, the NHS,
EA and others. LRF boundaries align with local police force areas for easier management
of local emergencies. Use of LRFs can potentially promote co-operation between
neighbouring LRFs when flood impacts cross boundaries. A list of LRF areas and their
component LAs is provided in Appendix I.
Development of Impact Assessment Metrics
Population
Following previous studies (Aldridge et al. 2011; Aldridge et al. 2015), the population
impacts of the flood scenarios were based on the Flood Risks to People (FRTP)
methodology, implemented as outlined in FRTP Phase II guidance document (HR
Wallingford, 2006) and supplementary note (EA and HR Wallingford, 2008). As well as
information on flood hazard intensity (depth, velocity, debris), FRTP was created in a UK
context and takes into account receptor-specific factors including personal and physical
vulnerability to flooding, and vulnerability associated with local influences. An additional
requirement for estimating evacuation was also addressed.
In this analysis, population impacts were presented as counts of:
1) People within the flood area;
2) People who are more vulnerable to the flood hazard (calculated as a proportion of the
total impacted population);
3) People who are injured;
4) People who lose their lives2, and;
2 The injuries and fatalities metrics presented in this research are better considered as the extent to which
contributing physical factors combine to present a danger to life or of injury (See below)
26
5) People requiring evacuation, including those requiring assistance, or identified as
needing a priority evacuation response (within 24 hours).
Table 4.4Table 4. presents the impact metrics selected for each of the population impact
calculations. These metrics were calculated for each of the populations in Table 4.1. In all
cases, people are considered to be impacted by flooding if they are located inside the
flood extent, at locations where flood hazard reached 0.575 or higher. This is in line with
outcomes of Defra capacity building workshops for the uFMfSW and based on agreement
with project partners.
Table 4.4 Population impact metrics.
Impact name Impact Metric
Impacted population Count of all population that are located within flood extent (minimum flood hazard rating = 0.575, minimum depth = 0.005m)
Impacted Vulnerable population
Count of impacted population identified as more vulnerable to flooding
Injuries Count of injuries sustained. This is based on FRTP, and applied using Equation 4.3
Fatalities Count of fatalities sustained. This is based on FRTP and applied using Equation 4.4.
Evacuation and priority evacuation
Count of people requiring priority evacuation. This is based on statistics from the Winter 2013/2014 flood review (EA, 2016) and a count of vulnerable people impacted.
The number of people that might be more vulnerable to flooding is a key statistic, which
can help assist with prioritisation of flood management action. Vulnerability is defined as
the characteristics and circumstances of a community, system or asset that make it
susceptible to the damaging effects of a hazard (UN/ISDR, 2009). This research followed
the approach for vulnerability described in the Flood Risks to People (FRTP) methodology
(HR Wallingford, 2006; EA and HR Wallingford, 2008). Application of the FRTP
methodology requires measurements for two types of vulnerability:
People vulnerability
Area vulnerability
People Vulnerability
The FRTP methodology defines people vulnerability as the ability of those affected to
respond effectively to flooding. Two population groups are considered to be vulnerable to
flooding, based on physical attributes:
People suffering from limiting long term illness,
People aged 75 or over.
27
The two populations above were calculated using 2011 Census tables listed in Table 4.5.
The variables considered were: age, long-term limiting illness and economic activity.
These variables provide information on whether an individual should be deemed
vulnerable and whether they are likely to be at home or at another location during the day,
for assessment of day time population scenarios. The geographical units reflect the size of
the catchment that the population is drawn from (discussed above). Populations in
hospitals and care homes were attributed 100% vulnerability based on the assumption that
all patients/residents would be elderly or ill and therefore less mobile and more vulnerable
to the physical effects of flooding. The full breakdown of the specific vulnerability and the
calculations for each population type are detailed in Appendix II.
Table 4.5 Census data used for People vulnerability calculation.
Census table code Census table name Geographical Units
KS102EW Age Structure OA, LSOA, LA
QS601EW Economic Activity OA, LSOA
L3302EW Long-term health problem or disability
by general health by sex by age
OA, LSOA, LA
LC3101EWLS Long term health problem or disability
by sex by age
LSOA
LC6302EW Economic activity by hours worked by
long-term health problem or disability
LSOA
Area Vulnerability
The FRTP describes area vulnerability as the characteristics of an area that affect the
chance of people in the floodplain being exposed to the hazard. The area vulnerability is
composed of three elements:
Scope and effectiveness of EA flood warnings,
Speed of flood onset,
Nature of area with regard to the physical characteristics of individual receptor
locations.
In FRTP, these three elements are each assigned scores between 1 and 3, which are
summed together using Equation 4.1, to provide a score (AV), between 3 - 9, where 3
indicates the areas least vulnerable to flooding, and 9 represents areas most vulnerable to
flood impacts:
28
Equation 4.1. 𝐴𝑉 = 𝑓𝑙𝑜𝑜𝑑 𝑤𝑎𝑟𝑛𝑖𝑛𝑔 𝑠𝑐𝑜𝑟𝑒 + 𝑠𝑝𝑒𝑒𝑑 𝑜𝑓 𝑜𝑛𝑠𝑒𝑡 𝑠𝑐𝑜𝑟𝑒
+ 𝑛𝑎𝑡𝑢𝑟𝑒 𝑜𝑓 𝑎𝑟𝑒𝑎 𝑠𝑐𝑜𝑟𝑒
The flood warning score uses three EA flood warning measures, which are based on three
key targets in relation to flood warning:
P1. The percentage of Warning Coverage Target met (Percent of at risk properties
covered by flood warning system). (Target 80%)
P2. The percentage of Warning Time Target met (Target 100%)
P3. The Percentage of Effective Action Target Met (Percent of people taking effective
action). Target 66%. Based on the Public Flood Survey 2013/14 Flood warnings (795 post
flood interviews) - 66% of people who received a warning took action.
The flood warning score is calculated using equation 4.2:
Equation 4.2. 𝑓𝑙𝑜𝑜𝑑 𝑤𝑎𝑟𝑛𝑖𝑛𝑔 𝑠𝑐𝑜𝑟𝑒 = 3 − (𝑃1 𝑥 (𝑃2 + 𝑃3))
Equation 4.2 produces scores from 1 to 3, where a value of 3 indicates a weak / no flood
warning while 1 indicates a good flood warning and action system.
The second component is the speed of onset score. A value of 1 indicates onset of several
hours, 3 indicates flash flooding occurring in minutes. This report follows the previous East
coast report by using a value of 2, indicating a flood onset of approximately one hour
(Aldridge et al. 2015).
The final component is the nature of area score. This component describes the
vulnerability of a receptor in terms of the physical attributes of its location. For example,
multi-storey buildings are considered less vulnerable because residents are more likely to
live in higher storeys, while bungalows and campsites are considered more vulnerable.
This report follows the previous East coast report (Aldridge et al. 2015) and the FRTP
methodology as detailed in Table 4.6. All other populations were given a nature of area
score of 2.
29
Table 4.6 Nature of Area vulnerability modelling descriptions.
Population
type
How modelled Nature of
Area score
Populations
near busy
roads
Primary routes and trunk roads were extracted from the
NRD roads layer (based upon OS ITN data). Population
locations within 40m of these were allocated as high risk.
3
Multi-storey
apartments
Residential buildings within the area of interest with more
than 10 households present were looked at with OS
MasterMap data to gauge whether they might be described
as multi-storey. Those that fit were classified as low risk.
1
Campsites Population locations within campsites were determined as
high risk.
3
Single Storey
Schools
Schools for young children or for those requiring special
care were considered to have a high likelihood of being
single storey, and so were classified as more vulnerable.
Classifications taken from the NPD and NRD for infant,
junior, primary and special schools were used to set the
nature of area risk for schools meeting this description.
Although this is not always the case, it is a reasonable
assumption.
3
Road
Populations
Those in cars classified as high risk. 3
Injuries and fatalities
Following the FTRP methodology, the number of injuries and fatalities sustained in a flood
are calculated as functions of the vulnerable population. The number of injuries is
calculated using the number of vulnerable people at a location, the area vulnerability and
the flood hazard rating at that location (Equation 4.3).
Equation 4.3. 𝑁𝑖𝑛𝑗 = 2 ∗ 𝑁𝑧 ∗ 𝐻𝑅 (𝐴𝑉
100) ∗ 𝑃𝑉
Where Ninj is the number of injuries, Nz is the number of people at risk, HR is the hazard
rating, AV is the area vulnerability and PV is the people vulnerability (proportion of
vulnerable people).
The number of fatalities (Nf) is calculated as detailed in Equation 4.4.
30
Equation 4.4. 𝑁𝑓 = 2 ∗ 𝑁𝑖𝑛𝑗 (𝐻𝑅
100)
Population impact metrics were calculated at the level of individual receptor point, and
then aggregated to reporting level.
Injury and fatality estimates should be treated with caution. The multi-dimensional nature
of the impacts of flooding on people presents a high level of uncertainty. Complicating
factors relate to the nature of the hazard and the behaviour of the receptor. For example,
the nature of the flooding within the cell and the effect of features in urban areas could
alter localised flood depths and velocities, while the assumed location and behaviour of
people within and around a flooded property could change how they are impacted.
Consequently, the metrics used here are better considered as the extent to which
contributing physical factors combine to present a danger to life or of injury until these
factors can be better understood and accounted for.
Evacuees
The number of impacted residents (including those residing in hospitals, prisons and care
homes) provided a baseline of the population who may require evacuation. Those
requiring assistance with evacuation can be estimated based on the vulnerable population.
Priority evacuees represent a further subset of the vulnerable population, who may not be
easily identified using the NPD or census tables. These are people who require the most
urgent evacuation assistance. Priority evacuees were identified under the assumption that
they would have an immediate health requirement that presents a risk to wellbeing if care
is not available at short notice. People in care homes and hospitals would be in this group.
In addition, the UK Homecare Association Ltd (2016) estimate that approximately 512,000
people received state funded domiciliary care (care in the home) in England and Wales in
2013/14, with a further 228,000 people receiving care that is privately funded in the UK.
Adjusting the UK private estimate based on the state-funded figure produces a total of
690,000 receiving some form of domiciliary care in England and Wales. This represents
1.2% of the total population according to ONS mid-year population estimates. This
percentage was used in addition to the numbers of residents in care homes and hospitals
to estimate priority evacuees.
31
Property
The property point datasets were overlaid with the flood hazard extent dataset. A property
was considered flooded if the property point was located within the flood hazard extent
using a base flood depth threshold of 0.00m3. This aligns with the NaFRA approach.
Property damage was estimated by considering depth-related impacts for individual
properties and aggregating to reporting areas. Damage calculations for different property
uses and types (Table 4.2) were based on flood depth information and damage curve
calculations published in the Multi-Coloured Manual (MCM) (Penning-Rowsell et al. 2013).
The MCM is an established and comprehensive framework for assessing the economic
impacts of flooding. The MCM offers flood damage information for a range of different
flood types (salt / fresh water, short, medium, long durations) and includes a break-down
of cost components including domestic clean-up, household inventory damage and
building fabric damage. These values are provided in the MCM in the form of depth-
damage curves for different types of property at a range of flood depths, with the
component costs summarised to produce total damage and total damage per square
metre for a given depth.
MCM damage calculations require information on building use and footprint, which can be
found in the NRD, speed of flood onset and nature of the flood water (‘fresh’ and salt).
Different curves were applied for different property uses and types. For each impacted
property, a value of damage per m2 was calculated from the depth-damage curves based
on the flood depth and property type, which was then multiplied by the footprint of the
building (m2) to provide an estimate of the damage in pounds. An example is shown in
Figure 4.1Figure 4.1, which demonstrates the damage calculation for a 300m2 retail
property flooded to 1.5m depth.
Flood warnings have been shown to reduce damage to property due to the opportunity
presented to take action, primarily by protecting or moving personal possessions, stock, or
moveable equipment (Penning-Rowsell et al. 2013). The MCM estimates the reduction in
costs as a proportion of the total damages. For residential properties, the cost in damages
can be calculated for a warning of less than 8 hours and a warning of over 8 hours. For
non-residential, the damages can be calculated for a warning time of 4 hours. This can
provide useful best and worst-case scenarios.
Where specific building type information was not recorded, the curves for the ‘average’
category were used for residential properties and the ‘unclassified’ category for non-
residential properties. To accurately model the damage to properties below entrance
thresholds (doorsteps), the modelled flood depth was reduced by 0.25m when calculating
property damage following the NaFRA approach.
3 This was implemented as a depth threshold of 0.005 m to eliminate artefacts in the flood hazard data that modelled
very low depths over large areas.
32
Figure 4.1 Property damage calculation example for a 300m2 retail
property flooded to 1.5m.
Infrastructure
Disrupted infrastructure metrics were based on the potential exposure to different levels of
flood hazard ratings (Table 3.1). As for population, flood hazard ratings less than 0.575
were ignored. Metrics were produced as counts of flooded sites and the percentage of that
infrastructure type flooded in the Local Authority. These metrics provide information on the
absolute magnitude of impact and indicate the pressure on local resilience and
contingency. This was completed for all infrastructure assets listed in Table 4.3.
Additionally, to evaluate the disruption of key sites (which are typically buildings requiring
access by the public), additional metrics were calculated to count the number of key sites
inundated to a depth greater than 0.2m. This depth corresponds to EA advice on sandbag
usage (EA, 2009b), assumed here as a minimum level of protection that might be
expected at these sites.
Road and Rail networks are considered to be impacted if they are inundated to a depth of
0.15 m or greater, based on typical vehicle ground clearance4 as an indicator for roads
becoming impassable. Disrupted journeys are evaluated based on impacts on the mean
average total kilometres travelled each day by vehicles, passengers, and lorries (as
4 http://www.autoevolution.com
0
200
400
600
800
1000
1200
1400
1600
1800
0 0.5 1 1.5 2 2.5 3
Dam
age
(£ p
er m
2)
Flood Depth (m)
Example: Retail Property Floor area = 300m2
Flood depth = 1.5m Damage per m2 = £1,153 Property damage = 1153 * 300 Property damage = £345,900
33
indicators of domestic and commercial traffic). Additional metrics for Trunk roads
(including motorways) are reported as a subset of the full transport network.
Agriculture
The MCM provides a damage cost per hectare for each of the five agricultural land grades.
Grade 3 land is given two values dependent on the proportion of livestock / arable crops
grown on the land. Consequently an average of the two values was taken (Table 4.7) for
this analysis. The method of measuring short-term and long-term impact to agriculture
follows the NaFRA methods by calculating:
1. The area impacted (above 0.00m of flooding) and;
2. The cost in damages where flood depths exceed 0.5m.
Impact and damage to Grade 1 and 2 agricultural land was included as a separate impact
metric. Estimates do not specifically account for damage costs associated with salinity.
Table 4.7 Cost of agricultural land by grade (£/ha) (1 highest quality, 5 lowest quality)
adapted from Penning-Rowsell et al. (2013).
ALC class Indicative land use Flood costs £/ha
1 Intensive arable (100%) 1320
2
Intensive arable (60%)
Extensive arable (35%)
Horticulture (5%)
1000
3a Extensive arable (70%)
Intensive arable (30%)
600
Mean = 470
3b Extensive arable (50%)
Intensive grass (50%)
340
4 Intensive grass (100%) 180
5 Extensive grass (100%) 100
Wider economic impacts
The wider economic costs use a selection of the impact metrics described above as inputs
to economic calculations. Economic costs to tourism and the environment have not been
included. Table 4.8 presents the calculation for each of the economic metrics, which
ultimately are summed together for the entire hazard.
34
Table 4.8 Calculations for wider economic cost impacts.
Wider economic
cost
Calculation Notes
Fatalities and
Casualties
(Worst case of night time or day
time fatalities * £1,836,054) +
(Worst case of night time or day
time injuries * £80,690)
Values relate to the cost of a
fatality and the cost of an
average injury is based on
DfT estimates using a
‘willingness-to-pay’ approach
(DfT, 2014b; HSE, 2011).
Lost Assets Total of property damage
(with warning)
Using MCM approach as
described above.
Lost Working hours –
employment impacts
Day time working population *
£11.61 * flood duration (hours)
£11.61 is the median hourly
pay (ONS, 2014).
Flood duration is assumed to
be 15 hours (2 days) based
on discussion within project
team.
Lost Working hours –
transport impacts
(Commuting and
business trips
(Total journey time (hrs) per LA /
percent of flooded transport
network) * £11.61
Total journey time is
calculated using DfT statistics
on average journey times and
multiplying up to the LA
population.
Shelter – Short term Night time impacted population *
flood duration * £35
Flood duration is assumed to
be 15 hours (2 days). £35 is
average cost of short term
accommodation per person
per night (DCLG, judgement-
based figure based on expert
consultation).
Shelter – Long term Impacted residential properties *
0.46 * £10,345
0.46 is the proportion of
impacted properties likely to
require extensive repair work.
£10,345 is the average per
property cost for this
relocation (From EA review of
2013/2014 flood impacts (EA,
2016)).
35
Implementation/Application
The assessment of flood impacts was separated into five steps as demonstrated in Figure
4.2Figure 4.. Step 1 intersects flood hazard data supplied by HR Wallingford with the
receptor database and identifies the receptors potentially at risk from the flood and assigns
flood attributes (depth, velocity and hazard rating). Step 2 uses the attributed receptor
information along with auxiliary lookup tables to calculate specific impact information for
each receptor using the methods described above, as summaries of danger to life,
economic damage to property, disruption of infrastructure and agriculture impacts. Step 3
aggregates individual impacted receptors into regional summaries by Local Authority and
Local resilience. Step 4 collects appropriate aggregate impact metrics to calculate wider
economic costs. Finally, Step 5 integrates impacts into a single spreadsheet-based results
template, which optimises presentation of results, focussing on 3 spatial scales:
1. National overview
2. LRF headline statistics
3. LA detailed statistics
Figure 4.2 Impact Assessment Implementation.
The process was largely automated using the statistical software package R. R is open
source statistical software capable of efficiently managing and manipulating multiple large
datasets. Processes were written as code, which can be rerun multiple times, increasing
efficiency of operation as well as providing a transparent methodology for future replication
Step 2
Step 1
Step 3
Step 4
Step 5
Boundary
data
Receptors exposed to
flooding
Calculation of impacts
Aggregation of impacts
to LRF/LA
Calculation of wider
economic impacts
Integration into a single
results template
Receptor
database
Flood hazard
inundation data
MCM curves
FTRP calculations…
Economic Valuation
Data
36
and research development. R also provides access to libraries with the capacity to handle
spatial queries, making it an appropriate tool for this research.
Quality Assurance
Quality assurance was completed throughout the impact assessment task:
1. Each receptor dataset was quality checked against established secondary sources
or OS base mapping. This involved manual verification of a random sample of
points to check that the coverage, location, function and other attributes of the data
were correct.
2. Automated Impact Assessment metrics, including MCM methods were checked
against manual methods to confirm that the processes were correct.
3. Visual checks and spot checks were performed on the final outputs
a. to capture extreme values and assess their sensitivity,
b. to ensure that calculations had been processed correctly,
c. to ensure that aggregation of results into Local Authority boundaries was
completed correctly
37
5. Discussion
Throughout this project, there has been a requirement to apply established methodologies
and information sources where possible. This has ensured that the approaches for both
hazard and impact modelling were robust, comparable to previous analysis and
reproducible in required timescales. The project has assessed a wide variety of impact
metrics, although some received more focus than others. This is largely related to the
output requirements, but also related to the availability and quality of established
methodologies.
5.1. Flood scenario generation and hazard modelling methods
The National Risk Assessment requires the definition of a likelihood (or probability)
associated with each of the flood scenarios. However, a single coastal scenario is
comprised of data on waves, sea levels and winds at multiple locations around the
country. It is complex to define the likelihood of the scenario in terms of these forcing
variables (i.e. winds, waves and sea levels) and there is no current agreed or established
method for doing this. The parallel H21 analysis sought to define the likelihood of a
scenario using a measure akin to taking the average of the Annual Exceedance Probability
(AEP) of all the forcing variables. There are a number of issues associated with this. In
particular the average AEP of a scenario defined in terms of the forcing variables does not
relate to the impact of the flood. For example, two separate scenarios can be defined that
have the same average AEP and one scenario can cause extensive flooding and high
impact and the other scenario can cause no flooding at all. In summary, there is no
agreed definition for defining likelihood for floods in terms of the multiple forcing variables,
and attempts to do so can lead to erroneous and anomalous interpretations and are best
avoided.
It is for this reason that the preferred approach for assigning likelihood to the scenarios is
risk-based. That is to say it is preferable for the likelihood to be determined in terms of
consequence or impact. Whilst this is certainly possible, it would have involved a
significant amount of further effort and resource than was available for the current project.
A risk-based analysis would involve translating the many thousands of floods output from
the statistical model into an impact metric. These impacts can then be ranked and a risk-
based likelihood of impact defined. Application of the models used in this study can be
computationally challenging and often, in practice, simplified or reduced complexity models
are used in their place (Gouldby et al. 2008). This is the case with NaFRA.
It is recommended that future analyses are risk-based and define likelihood in terms of
impact or consequence. These could use screening or simplified hydraulic models, if
computational resources associated with the simulations were constrained.
38
5.2. Impact Assessment methods
Data
The breadth of impact metrics measured in this project has required collection and
collation of a wide variety of receptor data from a number of different sources. In some
cases, the raw source data was directly appropriate for use. In other cases, the raw data
required pre-processing to either filter or clean-up the data to produce desired formats and
specifications. Source dataset quality differs offering varying levels of confidence and
requirements for development before inclusion in the analysis. The spatial information in
the source datasets used was generally excellent, with references to spatial locations or
OS coordinate data for most datasets. As examples, OS AddressBase Premium and the
NRD (which is derived from OS AddressBase Premium) were key datasets for analysis of
property and infrastructure. They are well suited for this purpose and provide property
level locations, but the distributed local authority approach of the data capture suffers from
inconsistencies due to differing interpretations of property descriptions and classifications
(ONS, 2013).
Flood Risk to People
The estimation of life-loss and injuries arising from flood scenario modelling is complex.
Challenges arise as data relating to the number of deaths and injuries is sparse and
incomplete and the actual causes of injuries are not well-documented. There has
however, been a significant amount of research into methods that seek to provide
estimates. The outputs from all estimates are subject to substantial uncertainties
(Lumbroso et al. 2015). Estimation of injuries and fatalities in this research is based on
the FRTP Methodology which is implemented as outlined in the phase II guidance
document (HR Wallingford, 2006), and Supplementary Note (EA and HRW, 2008). The
FRTP method is implemented to calculate estimates at the level of individual property, and
then aggregated up to reporting areas (LAs, LRFs). Since the development of the FRTP,
more advanced approaches have been introduced and these include agent-based
modelling methods (Lumbruso et al, 2011). Site specific analysis has been conducted on
historic floods that do yield significant improvements. However, these have not to date
been applied and approved by the Environment Agency for use in wide-spread flood risk
assessments like NaFRA and national surface water flood risk estimates. It is therefore
possible that modelling methods of these types could be used in the future to providing
supporting evidence and aid calibration and verification of more simplistic models like
FRTP.
The FRTP model was developed to provide overall estimates of population impact based
on empirical data, and implements a standard risk modelling approach combining
indicators for flood hazard, vulnerability and exposure. It is not designed as a physical
model of floodwater inundation presenting a risk per individual property, nor to provide an
accurate estimate of the risk on a property-by-property basis. Although the FRTP method
is usually applied to zones of flooding (assuming single values for flood hazard,
vulnerability, population etc. to all properties within an area), additional benefits can be
39
gained from using locally-specific property level information (for example on vulnerability,
land use, flood intensity, population numbers, and building type) to refine the results, and
without compromising the integrity or structure of the model. Further, the results when
calculating using averages across areas are likely to be of a similar scale.
The report has emphasised that the injuries and fatalities metrics provided by the FRTP
model are better considered as the extent to which contributing physical factors combine
to present a danger to life or of injury. The method assumes that certain groups of people
are more vulnerable to flooding than others, and this is reinforced by evidence of higher
proportions of elderly fatalities in the 1953 Netherlands storm surge flooding and the 2005
New Orleans flooding (Jonkman and Vrijling, 2008). This definition of this vulnerable group
includes care home and hospital populations based on the residents physical
circumstances. It is likely that these locations have specific flood preparation plans in case
of an emergency, however it has not been possible to nationally model for these
circumstances due to the unique characteristics of each site location, each flood plan and
each responsible authority.
Key sites and Infrastructure
The methods used to measure the impact on infrastructure sites are relatively simple and
could be improved to more accurately reflect resilience in wider networks. The approach
does acknowledge some on-site flood defences in the flood hazard modelling, but, in a
similar way to public service sites, resilience actions as a result of individually site-specific
plans are not accounted for.
The research has measured the percentage of impacted infrastructure sites in an area to
provide a measure of local resilience. This approach is appropriate for some infrastructure
types where individual sites can be reasonably modelled to have a catchment of influence,
but it is less appropriate for infrastructure networks, such as water and energy supplies. In
these cases, asset managers have stated high levels of national resilience to flooding.
These statements are difficult to test due to the complexity of the gridded infrastructure
and the sensitivity of information relating to large-scale infrastructure; the approach taken
was chosen to fit within the bounds of the project and provide a simple and straightforward
set of results. The infrastructure measures provided do not take into account sites in
adjacent reporting areas, which may increase resilience and may more accurately reflect
the emergency response or post-event situation.
Communication of results is a critical feature of this project and there was a requirement to
ensure that results could be easily accessible by a range of different audiences. The
innovative multi-scale results template used provides a suitable format for presentation of
results that enable readers to review overall headlines or detailed local boundary data
depending on their needs. It should be noted that where the results provide absolute
counts or percentages which are relatively easy to interpret for wider audiences, they
suggest a certainty in measurement that may not be reflected by the methods used. This
may be a particular issue for population and infrastructure impacts, which may be
alleviated through individual and organisational response activities.
40
6. Conclusions and Recommendations
6.1. Conclusions
The data and methods applied in this research are suitable for the level of detail and the
spatial scale of the analysis. The methods are based on established flood risk science,
which means that results can be compared with previous UK assessments as well as
similar current projects. The research includes a broad range of physical and economic
impact metrics, which build on previous impact assessments by modelling the impact on
key infrastructure assets, increasing the recognition of resilience metrics and by
introducing more sophisticated economic cost measurements. Results are produced by
Local Authority and by Local Resilience Forum boundaries. These provide meaningful
aggregations of data, which can be more easily digested by the relevant audience. The
creation of a novel results template has enabled over 300 metrics across over 200
different boundaries to be represented in a clear fashion at different spatial scales and at
different levels of detail. This ensures that the results can be just as effectively
communicated to national decision makers and local emergency managers.
6.2. Recommendations
The following recommendations are suggested for future development of the impact
assessment:
The NRA requires the estimation of the likelihood of the flood occurring. It is
desirable to adopt a risk based approach to likelihood specification, whereby
likelihood of impact or consequence is the metric of relevance. This does, however,
require the hydraulic simulation and impact evaluation of many more scenarios, not
feasible on this project. Future work should however, consider the viability of
implementing a fully risk-based approach to scenario likelihood estimation.
Challenges present in the modelling of transport and infrastructure impacts have
highlighted the possibility for more sophisticated network resilience analysis. This
might include the analysis of diversionary or evacuation routes and related impacts
to commuting or emergency response. There may also be scope to improve
analysis of the impact on utility services including subsequent impacts on supply to
residential and commercial properties, including further economic or social impacts.
Collection of more detailed property data including features such as age or
construction materials could allow for a more sophisticated analysis of building
damage, allowing a deeper use of the Multi-Coloured Manual methodologies and
potentially more information on likely repair/rebuild times, which have an impact on
evacuation and shelter costs.
Modelling flood impacts on people remains a challenge and this research has
highlighted this further. The FRTP method uses relevant information to provide
useful indicators that help in the understanding of the risk but there are still
41
questions about the sensitivity of the fatality and injury estimates (particularly when
applied to a large scale analysis), and finding the best way of communicating the
results to highlight these sensitivities. Future research could evaluate the potential
to amend FRTP assumptions, equations and parameters as well as introducing
measures of uncertainty into the analysis. Comparison of FRTP with other methods
that provide smaller scale population impact estimates may help to understand
FRTP limitations and calibrate the model.
The response to Environment Agency flood warnings is integrated into the FRTP
methodology, but further acknowledgement of flood response may be useful for
improving the counts of impacted people or impacted sites. This may require more
detailed knowledge of local flood risk plans or individual infrastructure site flood
plans.
The hydraulic modelling has a temporal aspect but this is not included in the
scenario results or applied to the impact assessment. Temporal analysis of impacts
has the potential to provide added value and another aspect to response
prioritisation but challenges are present in applying this effectively and in
communicating the results.
The methodology outlined in this report and the associated code has been
developed for repeatability across different types of flooding, where the extent,
onset and composition of the flood waters are likely to differ. The concepts of the
impact assessment component are common to wider non-flooding contexts and the
authors would also encourage adaptation of the methodology for other applications
including other natural or man-made hazards. Further, the receptor database
created in Chapter 4 collates information for different types of property, key
infrastructure and service categories and a range of different population types.
Much of the information is not flooding-specific and it could be applied for use in
other impact assessments.
The impacts methodology could also be simplified and applied to statistical flood
scenario generation to produce a novel impact-based risk assessment. Applied in
this way, the model would allow for the estimation of risk, based on impact severity
and likelihood of occurrence. This could be a valuable tool for development of
evidence for the National Risk Assessment and for other flood impact applications.
Environmental impacts beyond those to agriculture are not currently included in the
assessment. These may include the impacts of prolonged salt water inundation or
the impacts of the release and diffusion of pollutants and other dangerous materials
by floodwater into the wider environment. Future work could aim to estimate
environmental impacts and costs based on the concept of natural capital.
The current model does not yet consider the social or psychological impacts of
flooding. Awareness is steadily growing of these chronic impacts (PHE, 2015),
which include stress, anxiety and depression. These impacts are amongst the most
challenging features to measure and quantify. It is anticipated that further research
into these areas could build on current indices based on community
characterisation by socio-economic data, there is also potential to build
collaborations with organisations of social scientists and psychologists to explore
alternative approaches.
42
7. References
Aldridge, T., Forder, K., and Balmforth, H. (2011). East Coast Flooding Impact Analysis,
Report produced for Defra by HSL, Reference MSU/2011/43.
Aldridge, T., Gunawan, O., and Rastall, P. (2015) East Coast Flooding Impact Analysis
Update Summary Report. produced for Defra by HSL, Reference MSU/2015/03.
Baxter, P.J. (2005). The east coast Big Flood, 31 January – 1 February 1953: a summary
of the human disaster, Phil. Trans R. Soc. A. 363 (1831), pp1293-1312.
BBC. (2013). Tidal Surge hits east coast UK coastal towns after storm. Accessed 12th
February 2015 http://www.bbc.co.uk/news/uk-25253080
Cabinet Office. (2013). National Risk Register of Civil Emergencies 2013 Edition.
Accessed 12th February 2015 https://www.gov.uk/government/publications/national-risk-
register-for-civil-emergencies-2013-edition.
Department for transport. (DfT) (2014a) 2011 National Travel Survey: England 2013, 2nd
June 2016.
https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/342160/nts2
013-01.pdf.
Department for transport. (DfT) (2014b). TAG UNIT A4.1 Social Impact Appraisal.
Environment agency and HR Wallingford. (2008) Supplementary note on flood hazard
ratings and thresholds for development planning and control purpose - Clarification of the
Table 13.1 of FD2320/TR2 and Figure 3.2 of FD2321/TR1, May 2008.
Environment Agency. (2009a) Flooding in England: A National Assessment of Flood Risk.
Environment Agency. Bristol.
Environment Agency. (2009b). Sandbags: how to use them properly for flood protection,
LIT3833.
Environment Agency. (2011a) The risk of widespread flooding – Capturing spatial patterns
in flood risk from rivers and coasts, SC060088/R1 Spatial Coherence of Flood Risk –
Technical Methodology Report.
Environment Agency. (2011b) National Receptor Dataset.
Environment Agency. (2013) What is the updated Flood Map for Surface Water? Report
version 1.0 November 2013. Environment Agency. Bristol.
Environment Agency. (2014) The updated Flood Map for Surface Water (uFMfSW)
Property Points dataset: Environment Agency. Bristol.
43
Environment Agency. (2016). The costs and impacts of the winter 2013/14 floods: Final
Report. Ref SC140025. Environment Agency. Bristol.
Geoplace.(2016). The UPRN. https://www.geoplace.co.uk/addresses/uprn. Accessed 4th
November 2016
Gouldby, B., Sayers. P., Mulet-Marti, J., Hassan, M. and Benwell, D. (2008). A
methodology for regional-scale flood risk assessment. Water Management, 161(3), pp
169-182.
Gouldby, B., Mendez, F., Guanche, Y., Rueda, A. and Minguez, R. (2014) A methodology
for deriving extreme nearshore sea conditions for structural design and flood risk analysis,
Coast. Eng 88, June.
Haigh, I.D., Wadey, M. P., Gallop, S. L., Loehr, H., Nicholls, R. J., Horsburgh, K., Brown, J.
M. and Bradshaw, E. (2015) A user-friendly database of 100 years (1915-2014) of coastal
flooding in the UK, Scientific Data (2), Article number: 1500 doi:10.1038/.
Heffernan, J.E. and Tawn, J.A. (2004) A conditional approach for multivariate extreme
values (with discussion). Journal of the Royal Statistical Society: Series B (Statistical
Methodology), 66(3), pp 497-546.
HR Walllingford, Middlesex University Flood Hazard Research Centre and Risk & Policy
Analysts Ltd. (2006). Flood Risks to People Phase 2: The Risks to People Methodology,
Defra / Environment Agency Report: FD2321/TR1.
HR Wallingford. (2014). Indicators to assess the resilience of infrastructure in England to
the projected impacts of climate change. Reference MCR5195-RT003-R01-00. Report
prepared for the ASC of the CCC. http://www.theccc.org.uk/wp-content/uploads/2014/07/5-
MCR5195-RT003-R05-00.pdf.
HR Wallingford. (2015). State of the Nation Flood Risk Analysis: Coastal Boundary
Conditions, Report RT025, produced on behalf of the Environment Agency.
HSE. (2011). The costs to Britain of workplace injuries and work-related ill health in
2006/07 – workplace fatalities and self reports, HSE Research Report: RR897.
ITRC. (2013). ITRC Second assessment of national infrastructure pilot results report
http://www.itrc.org.uk/wp-content/Second-assessment/ITRC-second-assessment.pdf,
(Accessed 4th January 2016).
Jonkman, S. N., and Vrijling, J. K. (2008) Loss of life due to floods. Journal of flood risk
Management 1(1), pp 43-56.
Kingston, G., Robinson, D. and Gouldby, B. (2008) “Reliable prediction of wave
overtopping volumes using Bayesian neural networks”, Proc. of FLOODrisk 2008, 30 Sep -
2 Oct, Oxford, UK In: Samuels et al. (eds). Flood Risk Management: Research and
Practice. Taylor & Francis Group, London.
44
Lamb, R. et al. (2010). A new method to assess the risk of local and widespread flooding
on rivers and coasts. Journal of Flood Risk Management, 3(4), pp 323-336.
Lumbroso, D. and Davison, M. (2016) Use of an agent-based model and Monte Carlo
analysis to estimate the effectiveness of emergency management interventions to reduce
loss of life during extreme floods, Journal of Flood Risk Management DOI:
10.1111/jfr3.12230
Lumbroso, D., Sakamoto, D., Johnstone, W., Tagg, A. and Lence, B. (2011) The
development of a Life Safety Model to estimate the risk posed to people by dam failures
and floods, Dams and Reservoirs, British Dam Society, 2011.
Met Office. (2014). Winter storms, December 2013 to January 2014. Accessed 12th
February 2015 http://www.metoffice.gov.uk/climate/uk/interesting/2013-decwind
Office of National Statistics (ONS). (2013). Emerging Thinking on a B2011 Address
Register. www.ons.gov.uk/ons/about-ons/get-involved/events/events/.../session-1-3.pdf.
Accessed 26th July 2016.
Office of National Statistics (ONS). (2014). ONS Annual Survey of Hours and Earnings.
Office of National Statistics (ONS). (2016), Output Area (OA),
http://webarchive.nationalarchives.gov.uk/20160105160709/http://www.ons.gov.uk/ons/gui
de-method/geography/beginner-s-guide/census/output-area--oas-/index.html, 2nd June
2016.
Penning-Rowsell, E., Priest, S., Parker, D., Morris, J., Turnstall, S., Viavattene, C.,
Chatterton, J., and Owen, D. (2013) Flood and coastal erosion Risk Management: A
manual for economic appraisal, Routledge, Oxon.
Public Health England. (2015). Flooding and health: national study.
https://www.gov.uk/guidance/flooding-and-health-national-study, Last accessed 25th July
2016.
Pullen, T., Allsop, N.W.H., Bruce, T., Kortenhaus, A., Schüttrumpf, H. and van der Meer
J.W. (2007). EurOtop: Wave overtopping of sea defences and related manual,
assessment manual.
Smith, G., Arnot, C., Fairburn, J. & Walker, G. (2005) A National Population Data Base for
Major Accident Hazard Modelling, HSE Research Report RR297.
Smith, G. & Fairburn J. (2008) Updating and improving the National Population Database
to National Population Database 2. HSE Research Report RR678.
UK Homecare Association Ltd. (2016). An overview of the Domiciliary Care Market in the
UK.
http://www.ukhca.co.uk/(S(qoday452pflmljtta1fwkehz))/pdfs/DomiciliaryCareMarketOvervie
w2015.pdf, Last accessed 2nd June 2016.
45
University of Southampton. (2014). SurgeWatch, available from
http://www.surgewatch.org/events, accessed 15th December 2015.
UN/ISDR. (2009). United Nations International Strategy for Disaster Reduction
Terminology on Disaster Risk Reduction. http://www.unisdr.org/eng/library/UNISDR-
terminology-2009-eng.pdf. Accessed 28/2/11. http://www.unisdr.org/eng/library/lib-
terminology-eng.htm.
Wyncoll and Gouldby. (2014) Application of a multivariate extreme value method to flood
risk analysis, J. Flood Risk Man. DOI: 10.1111/jfr3.12069.
46
Appendix I: Local Resilience Forums and constituent Local Authorities
47
Local Resilience Forum
Local Authority
Avon and Somerset Bath and North East Somerset Bristol, City of North Somerset South Gloucestershire Mendip Sedgemoor South Somerset Taunton Deane West Somerset
Bedfordshire Luton Bedford Central Bedfordshire
Cambridgeshire Peterborough Cambridge East Cambridgeshire Fenland Huntingdonshire South Cambridgeshire
Cheshire Halton Warrington Cheshire East Cheshire West and Chester
Cleveland Hartlepool Middlesbrough Redcar and Cleveland Stockton-on-Tees
Cumbria Allerdale Barrow-in-Furness Carlisle Copeland Eden South Lakeland
Derbyshire Derby Amber Valley Bolsover Chesterfield Derbyshire Dales Erewash High Peak North East Derbyshire South Derbyshire
Devon and Cornwall Plymouth Torbay Cornwall Isles of Scilly East Devon Exeter Mid Devon North Devon South Hams Teignbridge Torridge West Devon
Local Resilience Forum
Local Authority
Dorset Bournemouth Poole Christchurch East Dorset North Dorset Purbeck West Dorset Weymouth and Portland
Durham Darlington County Durham
Dyfed-Powys Carmarthenshire Ceredigion Pembrokeshire Powys
Essex Basildon Braintree Brentwood Castle Point Chelmsford Colchester Epping Forest Harlow Maldon Rochford Southend-on-Sea Tendring Thurrock Uttlesford
Gloucestershire Cheltenham Cotswold Forest of Dean Gloucester Stroud Tewkesbury
Greater Manchester Bolton Bury Manchester Oldham Rochdale Salford Stockport Tameside Trafford Wigan
Gwent Blaenau Gwent Caerphilly Monmouthshire Newport Torfaen
48
Local Resilience Forum
Local Authority
Hampshire Basingstoke and Deane East Hampshire Eastleigh Fareham Gosport Hart Havant Isle of Wight New Forest Portsmouth Rushmoor Southampton Test Valley Winchester
Hertfordshire Broxbourne Dacorum East Hertfordshire Hertsmere North Hertfordshire St Albans Stevenage Three Rivers Watford Welwyn Hatfield
Humberside East Riding of Yorkshire Kingston upon Hull, City of North East Lincolnshire North Lincolnshire
Kent Medway Ashford Canterbury Dartford Dover Gravesham Maidstone Sevenoaks Shepway Swale Thanet Tonbridge and Malling Tunbridge Wells
Lancashire Blackburn with Darwen Blackpool Burnley Chorley Fylde Hyndburn Lancaster Pendle Preston Ribble Valley Rossendale South Ribble West Lancashire Wyre
Local Resilience Forum
Local Authority
Leicester Leicester Rutland Blaby Charnwood Harborough Hinckley and Bosworth Melton North West Leicestershire Oadby and Wigston
Lincolnshire Boston East Lindsey Lincoln North Kesteven South Holland South Kesteven West Lindsey
Merseyside Knowsley Liverpool St. Helens Sefton Wirral
Metropolitan City of London
City of London Barking and Dagenham Barnet Bexley Brent Bromley Camden Croydon Ealing Enfield Greenwich Hackney Hammersmith and Fulham Haringey Harrow Havering Hillingdon Hounslow Islington Kensington and Chelsea Kingston upon Thames Lambeth Lewisham Merton Newham Redbridge Richmond upon Thames Southwark Sutton Tower Hamlets Waltham Forest Wandsworth Westminster
49
Local Resilience Forum
Local Authority
Norfolk Breckland Broadland Great Yarmouth King's Lynn and West Norfolk North Norfolk Norwich South Norfolk
North Wales Isle of Anglesey Gwynedd Conwy Denbighshire Flintshire Wrexham
North Yorkshire York Craven Hambleton Harrogate Richmondshire Ryedale Scarborough Selby
Northamptonshire Corby Daventry East Northamptonshire Kettering Northampton South Northamptonshire Wellingborough
Northumbria North Tyneside South Tyneside Sunderland Northumberland Gateshead Newcastle upon Tyne
Nottinghamshire Nottingham Ashfield Bassetlaw Broxtowe Gedling Mansfield Newark and Sherwood Rushcliffe
South Wales Swansea Neath Port Talbot Bridgend Vale of Glamorgan Cardiff Rhondda Cynon Taf Merthyr Tydfil
South Yorkshire Barnsley Doncaster Rotherham
Local Resilience Forum
Local Authority
Sheffield
Staffordshire Stoke-on-Trent Cannock Chase East Staffordshire Lichfield Newcastle-under-Lyme South Staffordshire Stafford Staffordshire Moorlands
Suffolk Babergh Forest Heath Ipswich Mid Suffolk St Edmundsbury Suffolk Coastal Waveney
Surrey Elmbridge Epsom and Ewell Guildford Mole Valley Reigate and Banstead Runnymede Spelthorne Surrey Heath Tandridge Waverley Woking
Sussex Brighton and Hove Eastbourne Hastings Lewes Rother Wealden Adur Arun Chichester Crawley Horsham Mid Sussex Worthing
Thames Valley Bracknell Forest West Berkshire Reading Slough Windsor and Maidenhead Wokingham Milton Keynes Aylesbury Vale Chiltern South Bucks Wycombe Cherwell Oxford
50
Local Resilience Forum
Local Authority
Thames Valley South Oxfordshire Vale of White Horse West Oxfordshire
Warwickshire Tamworth North Warwickshire Nuneaton and Bedworth Rugby Stratford-on-Avon Warwick
West Mercia Herefordshire, County of Telford and Wrekin Shropshire Bromsgrove Malvern Hills Redditch Worcester Wychavon Wyre Forest
West Midlands Birmingham Coventry Dudley Sandwell Solihull Walsall Wolverhampton
West Yorkshire Bradford Calderdale Kirklees Leeds Wakefield
Wiltshire Swindon Wiltshire
51
Appendix II. 2011 Census calculations for population vulnerability
1. Residential Night time (OAs)
Total population (KS102EW)
Total population suffering from a limiting long term illness aged 0-74 (LC3302EW)
Population aged 75+ (KS102EW)
=
2. Residential
Day time
term time
(LSOAs)
Total population suffering from a limiting
long term illness aged 0-4 (LC1301EWLS)
Total population aged 16-74 not at work
due to limiting long term illness (QS601EW)
Population aged 75+ (KS102EW)
=
Total population aged 0-4 (KS102EW)
Total population aged 16-74 at home
(QS601EW)*
Population aged 75+ (KS102EW)
3. Residential
Day time
non-term
time
(OAs)
Total population suffering from a limiting
long term illness aged 0-15 (LC3302EW)
Total population aged 16-74 not at work
due to limiting long term illness (QS601EW)
Population aged 75+ (KS102EW)
=
Total population aged 0-15 (KS102EW)
Total population aged 16-74 at home
(QS601EW)*
Population aged 75+ (KS102EW) * Total population aged 16-74 at home is calculated from the following categories in QS601EW: - Economically Active - Unemployed; - Economically Active - Full-Time Student; - Economically Inactive - Retired; - Economically Inactive - Student; - Economically Inactive - Looking After Home/Family; - Economically Inactive - Permanently Sick/Disabled; - Economically Inactive - Other.
52
4. Sensitive Schools and Childcare (MSOAs) Total population aged 0-15 (LC3302EW)
Total population suffering from a limiting long term illness aged 0-15 (LC3302EW)
=
5. Sensitive Hospitals and
Care Homes = All people in these populations are considered
to be vulnerable.
6. Places of
work
(LADs)
Total population suffering from a limiting
long term but at work (LC6302EW)
=
Total population at work (QS601EW)*
* Total population aged 16-74 at work is calculated from the following categories in QS601EW: - Economically Active - Employee
- Economically Active – Self-employed with employees
- Economically Active – Self-employed without employees
**Calculated from 2011 Census statistics as the overall proportion of the total England and
Welsh population aged over 75 or long term ill.
-
7. Roads,
Stadia,
Prisons
(sensitive),
and
Transport
(LADs)
Total population suffering from a limiting
long term illness aged 0-74 (LC3302EW)
Population aged 75+ (KS102EW)
=
Total population (KS102EW)*
8. Leisure
Caravans and camping, Other Accom.
(National average: England and Wales)
National proportion of
vulnerable people = 0.20** =
53
Appendix III. List of impact datasets and sources Dataset Date of
creation/ version
Source
National Receptors Dataset (NRD) Property Points
2014 Environment Agency (EA)
Listed Buildings in England 2015 Historic England https://historicengland.org.uk/listing/the-list/data-downloads/
Listed Buildings in Wales 2015 Cadw Welsh Assembly Government https://data.gov.uk/dataset/listed-buildings-in-wales-gis-point-dataset
National Population Database (NPD) - residential
2015 Health and Safety Laboratory (HSL)
NPD - workplaces 2015 HSL
NPD - prisons 2015 HSL
NPD - hospitals 2015 HSL
NPD - schools 2015 HSL
NPD - colleges 2015 HSL
NPD - care homes 2015 HSL
NPD - Child care 2015 HSL
NPD - roads 2015 HSL
NPD - bus stations 2015 HSL
NPD - train stations 2015 HSL
Labour Force Survey 2014 Office of National Statistics (ONS) https://discover.ukdataservice.ac.uk/series/?sn=2000026
Camping and caravan club sites data
2014 Camping and Caravan Club
Valuation Office Agency summary valuation dataset
2015 Valuation Office Agency
OS AddressBase Premium 2015 Ordnance Survey
Care Quality Commission GP practice membership
2015 Care Quality Commission http://systems.hscic.gov.uk/data/ods/datadownloads/ gppractice
NRD roads 2005 EA
NRD rail 2005 EA
Transport Statistics PORT0101 2014 Department for Transport https://www.gov.uk/government/statistical-data-sets/ port01-uk-ports-and-traffic
Electrical substation sites 2015 National Grid http://www2.nationalgrid.com/uk/services/ land-and-development/planning-authority/shape-files/
54
Dataset Date of creation/ version
Source
Digest of UK Energy Statistics (DUKES) database 5.10
2015 Department for Energy and Climate Change https://www.gov.uk/government/collections/ digest-of-uk-energy-statistics-dukes
Nuclear site locations 2015 Health and Safety Executive
Consented Discharge to controlled waters
2015 EA / Natural resources Wales (NRW) https://ea.sharefile.com/share?cmd=d&id=s36af9f1b6494efa8#/view/s36af9f1b6494efa8?_k=8eeypn
Major Hazard Sites database 2015 HSE
Environmental Permitting Regulations - Waste National Dataset
2011 EA
Environmental Permitting Regulations – Industry National Dataset
2011 EA
Radioactive Substances Register 2015 EA
Agricultural Land Classification 1988 Natural England (Supplied with NRD 2005)
LRF geographies 2013 ONS
LA geographies 2013 ONS
IDBR 2015 ONS
Output Areas 2011 ONS
Lower Super Output Areas 2011 ONS
Medium Super Output Areas 2011 ONS
Census 2011 tables: KS102EW QS601EW L3302EW LC3101EWLS LC6302EW
2011 ONS
Multi-Coloured Handbook 2015 Flood Hazard Research Centre, Middlesex university http://www.mcm-online.co.uk/handbook/
Check Wider Economic Costs