Mapping flood disadvantage in Scotland 2015: report

5. Conclusions

5.1. Summary of the project findings

This report has presented the research carried out into the assessment of social vulnerability to flooding and flood disadvantage, based on the assessment framework developed by Lindley et al (2011). It is an update of an assessment of flood disadvantage carried out by Lindley and O'Neill (2013).The main changes are the incorporation of surface water flooding and defended flood extents, and the use of a reviewed set of indicators of social vulnerability to flooding. Therefore, the report moves the understanding of social vulnerability to flooding and flood disadvantage in Scotland forward by taking account of the most current data and stakeholder views.

The investigation into the flood hazard-exposure index confirms that flooding is a substantial risk in Scotland: just over 108,000 residential properties are estimated to be exposed to one or more sources of flooding of low probability (1 in 200 years including the impacts of climate change), with a minor number constructed since 1st January 2009.

The residential properties that may be exposed to flooding are spread across Scotland and nearly half of all data zones are exposed to flooding. Nonetheless, some of the local authorities have higher proportions of data zones exposed to flooding. For example Falkirk, the Orkney Islands and West Dunbartonshire have the highest average proportion of residential properties exposed to coastal flooding. Stirling, the Scottish Borders, and Perth and Kinross and Moray have the highest average proportion of residential properties exposed to river flooding. In Aberdeen City, Highland and Moray surface water flooding is likely to affect the highest proportion of residential properties.

The assessment of the levels of social vulnerability has revealed that just below 8% of all data zones are classified as having an extremely high or acute vulnerability to flooding. These are mainly located in large urban areas (Glasgow, Edinburgh, Dundee and Aberdeen). Social vulnerability has a strong urban focus: 73% of extremely or acutely vulnerable data zones are located in large urban areas and further 23% in other urban areas. However, extremely low vulnerability is also mainly present in cities: 79% of data zones classed as having extremely low social vulnerability to flooding are in urban areas. This emphasises the need for spatially detailed investigations into vulnerability to flooding and flood disadvantage, going below local authority or ward level, as larger spatial units may mask the differences in vulnerability and disadvantage.

Accessible countryside and accessible small towns (within a 30 minute drive time of a settlement of 10,000 or more people) were found to have the lowest levels of vulnerability, which may be explained by higher proportions of relatively wealthy, young and healthy people living within commuting distance from cities compared to those living within the urban areas themselves. In contrast, remote small towns and remote rural areas emerge as having potential issues with social and physical isolation and mobility of people, which may raise concerns regarding the response to flood events in those areas. Coastal areas have also emerged as having higher levels of vulnerability than inland areas.

With regard to any type of flooding (1:200+cc), 3.6% of all data zones in Scotland can be classified as extremely (138) or acutely (98) disadvantaged. Extreme and acute flood disadvantage (from any type of flooding) may affect an estimated 100,000 people; over 28,000 of these people may be extremely or acutely disadvantaged in relation to coastal flooding. Over 60,000 people may be extremely or acutely disadvantaged in relation to river flooding, and 14,000 people with regard to surface water flooding. The scale of flood disadvantage suggests urgent action is needed to address the risks to highly vulnerable communities exposed to flooding.

When the distribution of flood disadvantage among local authorities is considered, Falkirk, West Dunbartonshire and the Orkney Islands have the highest percentage of data zones classified as extremely or acutely disadvantaged with regard to coastal flooding. Considering river flooding, Stirling, the Scottish Borders and East Ayrshire have the highest percentage of extremely/acutely disadvantaged neighbourhoods. However, the highest number of data zones with acute and extreme flood disadvantage with respect to river flooding are found in Edinburgh, Stirling and Highland, followed by Falkirk and Aberdeen. Glasgow presents the highest concentration of surface water flood disadvantage, with one-third of the extremely disadvantaged neighbourhoods (Figure 15). This is followed by City of Edinburgh and Aberdeen City. Flood disadvantage tends to be concentrated in urban areas; smaller urban areas (10,000 to 124,999 people) particularly contain a high proportion of extremely and acutely flood disadvantaged neighbourhoods.

The case studies for Dundee City, Dumfries and Galloway and Scottish Borders explored the results of the assessment of social vulnerability to flooding and flood disadvantage in more detail. The local authorities examined were supportive of the framework used and of the explicit links made between the vulnerability of communities and the hazard of flooding, as these issues tend to be considered in separation in local authorities' work. Also, the comprehensive approach to vulnerability assessment in the framework allows for identification of those who may not be in a difficult financial situation but have other issues that make them more vulnerable to flooding. In two out of three local authorities the data broadly reflected the participants' knowledge of where exposure and vulnerability coincide. However, fine-grained differences that were not picked up by the assessment at the data zone level were highlighted in the third local authority, emphasising the importance of using local knowledge against the maps developed using national-level datasets.

Nonetheless, the disadvantage assessment was thought to potentially support cross-departmental working on flood disadvantage; the results were also considered as useful to emergency services to identify areas where resources should be targeted. The maps were considered to be a useful tool to open up discussions with communities.

To utilise the social vulnerability to flooding and flood disadvantage maps and information to their fullest, the case study local authorities highlighted the need to provide concrete examples showing how the dataset has been used. In addition, more explicit connections should be made between this project's maps of flood disadvantage, SEPA's flood maps and the PVAs identified in NFRA (for example, identifying where they overlap and where disadvantage is present outside PVAs, see also Figure 24) in order to support work on LFRMPs. Finally, presenting the results in a manner that would allow displaying selected layers of information and would bring the maps and the underlying spreadsheets containing the data together was suggested.

5.2. Related research

The results of the flood disadvantage assessment are consistent with the NFRA ( SEPA, 2011) results. Nearly all of the acutely and extremely socially vulnerable and flood-disadvantaged data zones were located within PVAs, despite differences in the underlying data and methodology. Therefore, the results of the disadvantage assessment can be used to support Flood Risk Management Strategies developed for each of the 14 Local Plan Districts covering Scotland that take into account PVAs.

Also, the recent assessment of property level protection ( PLP) for Scotland ( JBA Consulting 2014) resonates with this study. PLP includes resistance measures, which aim to prevent water ingress (door guards and air brick covers for example), and resilience measures which reduce the damage costs should water enter a property (by elevating valuable goods or installing concrete floors). PLP was identified by JBA Consulting (2014) as a potentially fairer means of distributing scarce flood defence resources, which could be used:

• as an interim measure whilst a community is awaiting a larger flood defence scheme;
• in sparsely populated rural areas where it is difficult to justify the costs of capital works;

The JBA Consulting report identified the number of properties at risk of flooding in Scotland which might benefit from PLP as a cost-effective measure. However, the characteristics of communities and individuals may affect their ability to accept PLP in their properties or their implementation in a flood situation (Bichard and Kazmierczak, 2012).

When combined with the concepts underpinning this study, a number of targeted PLP policy initiatives may be identified:

• By identifying those areas where there are issues with the ability to prepare (particularly tenants and those who are not linked to the local area), there may need to be more targeted information and awareness raising campaigns amongst vulnerable groups as they may be less linked to social networks or have a poorer knowledge of local issues.
• In addition, private and social housing landlords may need to be identified and worked with in order to implement these measures. Dumfries and Galloway, for example, works closely with a registered social landlord in order to integrate PLP measures in social housing.
• In terms of the ability to respond, PLP (particularly manual measures) rely on individuals and communities to operationalise them. Thus, individuals in poor health and those who are mobility-impaired may require assistance with manually deploying PLP. In addition, adequate and timely flood warnings should be in place to allow timely deployment of PLPs.
• PLP is less of an issue in terms of ability to recover.
• Enhanced exposure can be lowered overall should a property be fitted with PLP yet become poorly maintained over time as other water ingress routes may appear that are not covered by the PLP.

This project included exploratory research into the number of properties that may be exempt from insurance under Flood Re, i.e. those that have been constructed after the 1st of January 2009 (see section 3.3).The methodology used has a number of limitations and only allows for broad-brush estimates of the recent development in flood risk areas. However, this is the subject of a project commissioned by ClimateXChange ( CXC), which aims to assess the rate of residential and non-residential property development in flood risk areas across Scotland over the past decade. The project is part of the overall Adaptation Indicators project. The project's remit includes the analysis of development rates in flood risk areas for different reference dates, considered also in relation to overall rates of development in Scotland over the period, to allow an understanding of the change in proportion of properties being developed in flood risk areas. The project will report its findings in August 2015.

Further, CXC has commissioned Land Use Consultants ( LUC) to carry out a study of how current and future flood risk is being accounted for in land-use planning decisions in Scotland. This research will improve understanding of the effectiveness of national and local planning policy in ensuring new development is avoided in areas at risk of flooding. The project is focussing on the two stages of land-use policy: development planning and development management. The project will report its findings in Spring 2016.

5.3. Notes on the methodology

Whilst the methodology applied in this assessment addresses some of the shortcomings of the first flood disadvantage assessment for Scotland, it still has certain limitations. The assessment at the data zone level, whilst providing a useful picture for strategic planning and responses, aggregates the households by finding an average rather than identifying individual vulnerabilities. This may be problematic in places where data zones are likely to contain communities of diverse characteristics, for example in rural areas, where the population density is low and data zones are large. Therefore, it is important that the outputs of this assessment are verified against and supplemented by local, up-to-date information in order to provide a finer-grain understanding of vulnerability, down to the level of individual households or people.

Also, certain groups which could be considered as very vulnerable are not included in this assessment due to data paucity, for example the homeless. Further, some of the data used is out of date and no alternatives are present: for example, census 2011 did not collect the information on the lowest level of dwellings, which is an important enhanced exposure factor. Thus, data from 2001 was used instead which may not accurately reflect the situation in areas that have undergone redevelopment in the last decade. More up-to date sources of information should be sought in future assessments. The Scottish Property Dataset, a national level property dataset commissioned by SEPA and used for the baseline appraisal of flood hazards was considered. However, it was not available to the project team within the timeframe of the project. In addition, whilst it provides information on the dwelling levels, a large proportion of properties are recorded as 'probable' rather than definite. Therefore, currently there seems to be no dataset available that would provide accurate information on the lowest property level.

Further, some of the properties located within flood risk areas may be equipped with PLPs and thus the impact of flooding may be reduced. Thus, we strongly encourage local authorities and service providers making use of the data to verify the dataset against locally sourced, up-to-date information on the property- or neighbourhood-level flood mitigation systems.

Nonetheless, to date, no methodological 'best practice' in assessment and mapping of social vulnerability to flooding and to climate-related events more broadly has been established (Preston et al., 2011). Therefore, this report makes a valid contribution to the understanding of social vulnerability to flooding and flood disadvantage in Scotland.

5.4. Recommendations

5.4.1. Policy and practice

This project has provided a strategic-level estimate of the numbers of people, residential properties and neighbourhoods associated with flood disadvantage, and the underpinning information on the factors influencing flood disadvantage. The high number of people (estimated 100,000) whose well-being may be adversely affected by flooding due to their personal, social and environmental circumstances, suggests that closer links should be made between policies related to flooding and health, aiming to reduce the impact of flooding on vulnerable communities.

The information on the concentrations of residential properties and neighbourhoods characterised by acute and extreme flood disadvantage can be used by SEPA to provide additional information supporting Flood Risk Management Strategies for Local Plan Districts. Further, the data on flood disadvantage can be fed by the Lead Local Authorities into the LFRMPs that turn FRMSs into Local Delivery Plans.

The spatial distribution of flood disadvantage can be used to support or evaluate decisions made on flood risk investment. The recent Joseph Rowntree Foundation report suggests that in England there is not a strong link between those local authorities which contain the most flood disadvantaged neighbourhoods, and levels of planned expenditure on flood protection (England and Knox, 2015). Therefore, the information on flood disadvantage can assist with targeting national investment into flood-risk areas in Scotland in a socially-just manner. Based on the findings, two areas require particular attention: the acute and extreme disadvantage in coastal urban areas and reducing the risks associated with physical and social isolation of communities in remote towns and rural areas (see section 4.3. for particularly flood-disadvantaged local authorities).

The flood disadvantage data could be used to increase preparedness for emergencies. Regional Resilience Partnerships can utilise the data to support the development of Community Risk Registers. The data can additionally assist community groups in the development of Community Emergency Plans.

Based on the results of the exploratory research into the number of recently built residential properties in flood risk areas (all caveats considered), it is recommended that the findings, in combination with the forthcoming reports commissioned by ClimateXChange (see section 5.3), are used as a basis to develop regulations and guidance for local authorities that would tighten development control at the local level and minimise the rates of development in flood risk areas.

There is currently no data available on the presence of PLP measures in either new developments or existing residential properties in flood risk areas. It is recommended that a project is commissioned to estimate the current levels of provision of flood resistance and resilience measures. A feasibility study with a view to developing a property flood resilience database for insurers has been funded by Innovate UK and is carried out by the Building Research Establishment (with Lexis Nexis and AXA Insurance). In addition, recording the presence of SUDS in residential developments by SEPA (in line with the requirements of the FRM Act) would help to understand better the levels of their exposure to flooding.

The increasing number of people renting rather than owning their houses in Scotland, and the high number of tenanted properties in some of the flood risk areas, calls for regulations on the provision of information on the risk of flooding provided by the landlord to the tenant. Also, tighter flood insurance regulations for rented properties are needed. In the meantime, local authorities with a particularly high number of rented properties should provide information on the risk of flooding to tenants and what measures can be taken.

For local authorities, mapped flood disadvantage provides a useful framework for planning actions in anticipation of the increased risk of flooding ( e.g. redevelopment that alters the use of the ground floor to minimise damage if a flood happens) and developing recovery strategies in the aftermath of flooding ( e.g. targeting financial assistance to groups least likely to have flood insurance).

Further, the extensive set of indicators compiled in the vulnerability assessment may be used by various departments to identify areas for action. For example, areas with high proportions of older people living on their own could be targeted for development or re-siting of day centres for elderly, whilst places with low mobility levels and poor physical access could be considered for 'ring and ride' services.

It is recommended that local authorities actively collaborate with third-sector organisations that may support local communities in the event of flooding. This is particularly important in remote rural areas and remote small towns, which have a relatively higher presence of locally-focused charities, to increase the self-help potential of the communities that may include physically or socially isolated individuals. Also, in inner-city areas, particularly within more deprived locations or those that are currently being regenerated, there are currently fewer locally-focused community organisations. It is important to ensure that some community resources are located in these areas to help social networks develop and provide a focal point for the community in the case of flooding.

SEPA can use the dataset developed in this project to assess where flood disadvantage is present outside their current flood warning target areas. The locations with high disadvantage should be prioritised as those where the flood warning service is needed the most. Also, the information about the characteristics of the community ( e.g. the number of people not speaking English, number of people who have moved in recently from outside the local area, number of people with limiting long-term illnesses) could guide the manner in which flood warnings are provided. The areas of high disadvantage falling outside the current PVA boundaries could be considered by SEPA against the candidate PVAs for the next cycle NFRA.

Some inconsistencies were found in the flood data provided by SEPA, whereby the higher probability, lower magnitude flood extents are not completely within the lower probability, higher magnitude flood extents. It is recommended that SEPA within its Flood Map development plan reviews the flood data in order to identify the reasons for these inconsistencies and address them in future revisions of flood maps.

How the dataset, maps and report are used by local authorities and other decision-makers should be monitored. The case study local authorities consulted within this project emphasised the need for examples of how the information can be used. It is recommended that the Scottish Government devises a method of collecting information from local authorities on their use of the datasets produced in this project.

Case study local authorities strongly recommended development of an online spatial portal to enable the display of selected layers of the information developed in this project. Such a spatial portal should bring together the underlying spreadsheets containing the data with the maps, together with comprehensive guidance materials. Climate Just ( www.climatejust.org.uk) contains an example of such a spatial portal. It would be optimal if the online maps allowed locally available (potentially sensitive and confidential) data to be incorporated into the assessment of social vulnerability to flooding and flood disadvantage in order to enhance the results. Also, the datasets developed in this project could usefully complement the information provided on SEPA's NFRA website ^[17] .

Further recommendations for actions on reducing flood disadvantage can be found in the Climate Just online resource: www.climatejust.org.uk.

5.4.2. Further research

It is recommended that future research on mapping social vulnerability and flood disadvantage in Scotland includes data that was not available within this project. Such data might include:

• Estimates of the number of homeless people and rough sleepers as one of the most vulnerable groups in relation to extreme weather events;
• Data on flood-related insurance claims in order to identify areas with probable high insurance premiums;
• Up-to-date information on the number of properties with the lowest dwelling level at ground floor or in the basement. This could for example be based on the Scottish property dataset, or similar;
• Information on the number of houses equipped with PLP measures (both resistance and resilience) for inclusion in the enhanced exposure index;
• More direct measures of social networks.

Further, whilst this research considered a number of indicators in the assessment of social vulnerability to flooding and flood disadvantage, they are considered independently of each other. Indicators identifying interconnected problems or the most disadvantaged groups ( e.g. identifying older people with health problems living in one-floor properties or private tenants that are on low incomes) would be helpful in carrying out the future assessments of vulnerability and disadvantage.

It is important to consider the fine-grained variability in levels of social vulnerability to flooding and flood disadvantage utilizing local knowledge. Even using relatively small Census units, e.g. compared to a similar analysis for England where geographical units were ten times larger, some of the case study workshop participants still felt that the spatial units and the use of national sources used led to a focus on the more urban areas within their local authority. In their view, national data sources often fail to pick up the disadvantages found in rural areas. Therefore, supplementing the national-level data with locally available information is crucial to progressing the understanding of flood disadvantage.

Exploring the future dimension of flood disadvantage is important for adequate planning of flood risk management. Therefore, it is advised that in further flood disadvantage assessments, the data on flood risk that includes climate change impacts is complemented by future projections of demographic and socio-economic characteristics of communities. Considering different scenarios of socio-economic development could be helpful to explore the potential futures. An analysis of ongoing and retrospective temporal changes in flood disadvantage would be helpful in identifying the direction of change and building scenarios.

In addition, the impact of flooding on businesses providing employment for local communities was considered to be important for future assessments by the case study workshop participants. This is because temporary closure of affected businesses, or even the withdrawal of employers from a flood-risk area, may have significant consequences for people relying on them for work. Easily laid-off casual workers or low-income groups may be amongst those most affected. Also investigating the impact of flooding on Small and Medium-size Enterprises ( SMEs) was considered an important angle to social vulnerability assessment, in particular in areas with a high proportion of people working in SMEs, as these businesses tend to be under-insured and rarely have contingency plans (Crichton, 2006).

Whilst the scope of this research was limited to assessing the levels of flood disadvantage, further research could usefully include the analysis of the resources available to manage flood risk. It is recommended that the distribution and resources of emergency services are reviewed in this respect. Mapping of rest centres and other social infrastructure that could be used locally in response to flooding and in the recovery phase would offer an additional layer of information. In addition, understanding of community responses, e.g. the distribution of flood groups or the presence of active flood wardens, could provide valuable information. Finally, it is important to check whether the investment in flood risk management follows the areas with high social vulnerability and flood disadvantage.