Urban centres and flood risk

A recent BBC article (Shanghai ‘most vulnerable to flood risk’) reports on a paper published in Natural Hazards by a team of researchers. The paper ‘A flood vulnerability index for coastal cities and its use in assessing climate change impacts’ in the journal Natural Hazards by Balcia, Wright and van der Meulen, follows in a tradition of trying to quantify risk using a set of key variables. (I think the paper is on open access so you should be able to read it via the link). The authors develop what they call the Coastal City Flood Vulnerability Index (CCFVI) that is composed of three parts: the hydro-geologic, the socio-economic and the politico-administrative. These parts represent the three key interacting subsystems that affect coastal flooding, the natural subsystem, the social-economic subsystem and the administrative and institutional subsystem. Within each of these the authors identify variables that indicate the degree of exposure to hazard, the susceptibility to the hazard and the resilience to the hazard. The hydro-geologic part only has indicators of exposure whilst the ‘human’ parts have indicators of all three.

Exposure is defined as the predisposing of a system to be disrupted by a flood event due to its location. Susceptibility is defined as the elements exposed within the system that influence the probability of being damaged during the flood event. Resilience is defined as the ability of a system, community or society to adapt to a hazard. This term is assessed through political, administrative, environmental and social organisational evaluation. Variables selected include sea-level rise, storm surge, number of cyclones in last five years, river discharge, foreshore slope, soil subsidence for the hydrogeologic subsystem. For the socio-economic subsystem the population close to the shoreline, the growing coastal population as well as cultural heritage are included as exposure factors whilst uncontrolled planning zones are an exposure variable for the political and administrative subsystem.  Susceptibility variables include the percentage of the population disabled or young or old and flood hazard maps. Resilience variables include shelters, level of awareness, institutional organisations and flood protection.

The paper carries out a detailed analysis of each subsystem and then combines the indicators into a single equation to determine overall vulnerability.  The selection of variables is well argued and the complexity and issues of using such indexes is discussed well, so the authors do not have a simplistic interpretation of hazards and vulnerability. Any paper that tries to squeeze and freeze the complex and dynamic concept of risk into a single index will always have the problem of simplification. Simplification, not only of the subsystems but also of the interpretation by others of the index itself.

The variables selected may reflect the data readily available plus a particular view of how the flood hazard should be alleviated. The focus on institutional organisations as resilience does imply a rather hierarchical view of hazard management and prevention (maybe a valid argument with a set of large urban areas with low social cohesion). Interpretation of the index, as in the BBC report, tends to focus on the final product rather than on the variables used in its construction and the ratings of the subsystems. Discussions could be made as to the appropriateness of the same variables for cities across the globe or for the selection of those variables anyway. Looking in detail at the breakdown of the index, it is clear that Shanghai is the most ‘vulnerable’ city on variables used to determine the hydro-geologic subsystem because of its high length of coastline and high river discharge (plus high soil subsidence). Manila, however, is ranked second because of its exposure to tropical cyclones and flooding – can both the same index combine both types of exposure? Does this mean that Manila is a more vulnerable location, as tropical cyclones are more frequent than high discharges? Can degrees of difference in vulnerability or rather exposure be assessed using a combined index? Shanghai is not, however, the top ranked city for all subsystems. For the economic variables, Shanghai is ranked fourth meaning that it is likely to recover quickly, economically at least, from the affects of a flood event.

An index like the one presented in this paper are very, very useful. They can be used, as the authors have done, to try to predict how changes in climate could impact on hazards and as such can be of great use in planning and management. A single index should, however, be used with caution, particularly if the choice of variables reflects a particular view of hazard management. Similarly, understanding how the index is constructed and how different parts of the index contribute to the whole is vital in understanding where vulnerability (and resilience) lie and how these might be improved.

Measuring Resilience

The recent BBC commissioned research of resilience in the England by Experian (http://publicsector.experian.co.uk/Products/Local%20Economic%20Resilience.aspx) factsheets are available at http://publicsector.experian.co.uk/Products/~/media/FactSheets/Strategy%20and%20research/Local%20economic%20resilience%20%20%20Sept10.ashx ) has produced some interesting results and highlights or confirms depending on your viewpoint, a clear north-south divide in terms the vulnerability and resilience of areas to the current economic climate, in particular to the expected government cuts in spending (http://www.bbc.co.uk/news/business-11233799) . The northeast is identified as the most vulnerable or least resilient area (http://www.bbc.co.uk/news/uk-england-11141264) with explanation of this focusing on the number of people employed in public sector jobs, the history of closures in large tradition industrial employers such as iron and steel and the lack of likely job creation in the private sector. The vulnerability of the northeast is contrasted with locations such as St Albans where the entrepreneurial spirit and the knowledge based industry focus are highlighted.

Any attempt to measure vulnerability and resilience is a good idea. Trying to put numbers to these issues is useful in debate and in identifying factors that may be important in the ability of an area to resist major events, which an economic downturn and a round of spending cuts are, even if they are economic and political rather than geophysical events. Care needs to be taken, however, to understand that these figures, any figures, that try to capture such a complex and slippery set of concepts will never succeed in illuminating every conceptual facet. It should also be borne in mind that merely having the figures can begin to set the agenda for the debate by focusing on the particular metrics used in compiling the figures. Any data collection will have some model of reality behind it, some idea of what determines resilience and vulnerability, and so will always collect data in the light of that model. Experian have published their methodology in detail on the BBC website (http://news.bbc.co.uk/nol/shared/spl/hi/uk/10/experian/doc/methodology.doc ) as well as the Excel data files upon which the analysis is based (http://news.bbc.co.uk/nol/shared/spl/hi/uk/10/experian/xls/resilience.xls ). This is, in my view, an excellent idea as it allows everyone to understand how the final indices are created and provides hints about the model of reality behind the numbers.

Experian view economic resilience as describing the ability of an area to withstand and respond to shocks in an external environment. Economic resilience is composed of four themes: business, community, place and people. Each is weighed differently in terms of it contribution to the overall index, with the business variables accounting for 50%, with the other themes accounting for about 17% each. Each theme is described by a series of question. Business, for example, contains questions such as how strong is the local business base? Is it dependent on sectors that have been hit by recession? Are businesses dependent on local markets, or do they export? People contains questions about age of population, their jobs and earnings. Community contains questions about life expectancy, do neighbours look out for each other and long-term unemployment. Lastly, place contains questions on house prices, green space and GCSE attainment rates. Experian use 33 variables to create the index and do state that the variables used were dictated, at least in part, by the data available at local authority level.

The actual variables used are listed for each theme and range from insolvency rates, to business density, to % of population who are corporate managers, deprivation and crime rates.

This is the basic outline of the analysis and its methodology, so from the viewpoint of environmental geography what can be asked and interpreted from this analysis? Firstly, is the unit of analysis used, the local authority area, the most appropriate for what they are trying to analyse, economic resilience? The LA may be a unit for which there is lots of data as it is an administrative unit but is it the unit within which businesses operate? Do businesses locate within local authorities or within business parks (maybe with incentives from local authorities) or where successful business or businesses like them are already located? Similarly, is the local authority really an appropriate area within which to judge community? Are communities this big? You all probably know your own local authority area, do you think it is that homogeneous, are all parts of it really that similar or are there ‘good’ and ‘bad’ bits? The appropriate unit of analysis is a continual problem in geography and, to be fair to Exeprian, they have taken a unit for which there is a lot of data and for which data is collected. This data is not collected fro the type of analysis Exeprian have undertaken so it doesn’t exactly match what they would like to be available but virtually every survey has this problem as well. The important thing to bear in mind is that whatever spatial unit you select to use there will always be problems and there will always be a geography below or above the unit you select. It is down to you to decide if these other geographies are more appropriate.

Secondly, the method of ranking local authorities for different themes and then combining ranks can be misleading. A small, fractional percentage difference between two local authorities may be the difference between being ranked 5th adn 6th whilst a 10% difference may result being ranked 320th and 321st. You do not get any feel for the magnitude of the differences between areas. This is compounded once you combine ranks. So a drop between being 1st and 25th in the business themes may only reflect a slight difference in the absolute, numerical values of variables but is represented by a large number of places in the ranksing. A drop from 200th to 201st in the rankings may represent the same absolute differences in the magnitude of the variables as between the first 25 areas in the rankings. The raw data is needed to assess how the rankings and absolute changes match up. It shoudl be noted, however, that Experian do state that they have considered the correlation between variables in their analysis, another key issue in derving a meaningful statistic.

Thirdly, what do the variables used tell us bout the type of economy that Experian view as resilient and as vulnerable? The business variables depict resilience as being determined by a strong private sector and one that is not reliant upon its local market for its existence. Innovation, knowledge-based and able to draw on a reserve pool of skilled labour seem to be other important determinants of a strong business theme. Does this match your image of a strong, contemporary economy? What about the provision of infrastructure – or is that down to the private sector as well? Public services are viewed as a weakness in the local economy despite the need for provision of such services to support the private sector. If you have an aim to convert public to private then an area where private enterprise is strong are likely to be areas that rapidly take on public services once aspects of them are privatised. Likewise, people variable focuses on level of qualification and level of corporate managers in the population, once again highlighting the potential for a privatisation of the public. The focus on knowledge-based services may also suggest a preference for small firms as drivers of the local economy rather than large employers.

Thirdly, the community and place variables once again have a model of what is ‘good’ and ‘bad’ behind them. Indices such as long-term unemployment, claimant count and income vulnerability (likely to affect lower paid but increasingly the middle income as well) all point to a relatively inflexible work force in terms of skill development relevant for the sectors identified as important in the business variables, e.g. knowledge-based services. Place variables are a more mixed bunch ranging for educational attainment (related to skills again) through to crime rates and house prices. Do you feel these variables really reflect the feelings and strengths of communities and their networks of relations? The variables seem very static and unable to grasp the dynamic nature of such relations.

Overall, the Experian map is a great starting point in the debate over resilience and vulnerability to economic events and, potentially, other events as well. But it is only a starting point and like any analysis it is limited by the data available and will have underlying assumptions about what is ‘good’ and ‘bad’ based on a particular view of how the economy functions or should function. At least there is now a basis for discussion but the survey should not become an anchor point that sets the agenda without question.

Hazards and Vulnerability

Media reports and images are full of vulnerable people being struck by disasters. Film of families being rescued by inflatable boat in Pakistan has been common staple on recent news reports. When Hurricane Katrina smashed into New Orelans, it appeared the most vulnerable people in a developed country were being targeted by the disaster. It seems so clear but what do we actually mean by vulnerable? Leading on from this question is another important one. If we can define does this help us take steps to ensure these people are not affected by such hazards and disasters?

In a previous blog (Floods in Pakistan: Vulnerability) I began to discuss the complex nature of any definition of vulnerability and illustrated some of the issues using this ongoing disaster. If you want a simple definition then vulnerability can be defined as the potential for loss of life or property in the face of environmental hazards or environmental disasters (or indeed any hazard or disaster). Loss susceptibility is another term often used in relation to vulnerability. Other definitions include vulnerability as a threat to which people are exposed; vulnerability as the degree to which a system acts adversely to a hazard (whatever adverse might mean?!); differential risk for different social classes; interaction between risk and preparedness; inability to take effective measures; capacity of group to anticipate, cope with, resist and recover from impact of natural disaster. There are others and any one interested in the range of definitions used should have a look at Susan Cutter’s book, Hazards, vulnerability and environmental justice (2006, Earthscan). A key point to bear in mind is both the physical and human environments can be vulnerable. Physical systems can be fragile and susceptible to impacts as much as human systems. Outlining how these can be studied together will be the subject of a future blog. This blog will focus on social vulnerability, the vulnerability of the human part of the equation, rather than physical vulnerability.

Some other terms borrowed from ecology also tend to be used when researching vulnerability. Adaptation refers to the ability of the actants in the socio-ecological system to find strategies to adapt to the hazard or disaster. Resistance is the ability of the actants to resist the impact of the hazard or disaster. Resilience is the ability of the system to absorb, self-organise, learn and adapt to the hazard or disaster. A useful resource for vulnerability can be found at the web pages of Neil Adger (http://www.uea.ac.uk/env/people/adgerwn/adger.htm) and at the Resilience Alliance website (http://www.resalliance.org/1.php) a site looking at research into resilience of socio-ecological systems and sustainability. As with most things borrowed, once you change the context the meaning changes as well, so the application and use of these terms does not necessarily match their original, potentially more limited definitions in ecological research.

An important aspect of vulnerability is that it evolves; it changes as the nature of the disaster or hazard unfolds and as the people who are vulnerable response and react to their situations. This also highlights the importance of scale for defining vulnerability. What scale is appropriate? The individual can be viewed as an important unit, but the individual usually operates within the context of a family or household, so is this a more appropriate unit for analysing vulnerability and resilience? What about larger entities such as communities and governments? As you change the unit of analysis would you expect the different units to have the same type of vulnerability, the same ability to resist or the same characteristics of resilience? Once these different spatial entities interact, such as the provision of aid by the government to individuals, does this cross scalar interaction affect vulnerability and resilience? In other words, what seems like a simple thing is very complex to unravel in detail.

At heart vulnerability is about the differential ability or power to access resources by individuals and groups in society. To escape a flood you need the power or ability to get out of the area. You need a car, you need early warning, you need a friendly policeman to wave you through and protect you from the other people trying to escape on foot. These material things require resources and access to them at the appropriate time. There are static and dynamic aspects to this access to resources. The static aspects of vulnerability, might be capable of identification before a disaster strikes. At the simplest level, mapping socioeconomic groups gives an indication to the availability of funds to gain access to resources. Likewise, mapping similar census data such as lone parent numbers or age (elderly and young are less able to escape floods for example) could also indicate the vulnerability of a place. A useful site that discusses such mapping and has developed a specific means of measuring it, the index of social vulnerability, can be found at the University of South Carolina at the Hazards and Vulnerability Institute (http://webra.cas.sc.edu/hvri/) of which Susan Cutter is the Director.

There is also a dynamic aspect to vulnerability; the manner in which relationships are organised and the manner in which they change through normal times and then during and after a disaster. Such flows could include the transport infrastructure; a key aspect that appears to have failed during this disaster and which has dramatically affected the ability of the institution of government to maintain an effective relationship with vulnerable groups. At a local level, however, if the transport infrastructure that remains intact sufficient for the local population to move to safety and then initiate community based activities that represent resilience at that level? Importantly this dynamic aspect is concerned with pathways and relations, both physical between locations and places and social and emotion between peoples and between individuals and organizations. From the above it is clear that trying to understand vulnerability also means trying to udnerstadd its geography; how it varies in space and time and how people succumb to, adapt or try to overcome this geography.

Two useful books on vulnerabiltiy are:

Measuring Vulnerabiltiy to Natural Hazards: Towards Disaster Resilient Socieites by the United Nations University (2007)

Hazards Vulnerability and Environmetnal Justice by Susan Cutter (2006)

Floods in Pakistan: Vulnerability

Disasters always seem to affect the most vulnerable people. Images flashed across the television screen always seem to be of the poor struggling with their few possessions to escape the onslaught of the physical events. But what is vulnerability? As you might expect it is not as simple as a single definition. Vulnerability is a complex term and only one of several that are essential to understanding how people respond and are able to respond to a hazard and a disaster (I will discuss the term more systematically in a later blog, only the bare bones of what I think is essential to this situation will be outlined here).

Vulnerability could be simply put as the potential for loss of life or property in the face of environmental hazards. The susceptibility for loss may be another useful phrase. Associated with vulnerability are other terms, often ‘borrowed’ from subjects such as ecology and used, with the various usual inappropriate translations, in geographic research. Adaptation refers to the ability of the actants in the socio-ecological system to find strategies to adapt to the hazard or disaster. Resistance is the ability of the actants to resist the impact of the hazard or disaster. Resilience is the ability of the system to absorb, self-organise, learn and adapt to the hazard or disaster. A useful resource for vulnerability can be found at the web pages of Neil Adger (http://www.uea.ac.uk/env/people/adgerwn/adger.htm) and at the Resilience Alliance website (http://www.resalliance.org/1.php) a site looking at research into resilience of socio-ecological systems and sustainability.

An important question, and a very geographic one, is at what scale can you apply these concepts? What scale is appropriate? The individual can be viewed as an important unit, but the individual usually operates within the context of a family or household, so is this a more appropriate unit for analysing vulnerability and resilience? What about larger entities such as communities and governments? As you change the unit of analysis would you expect the different units to have the same type of vulnerability, the same ability to resist or the same characteristics of resilience? Once these different spatial entities interact, such as the provision of aid by the government to individuals, does this cross scalar interaction affect vulnerability and resilience? In other words, what seems like a simple thing is very complex to unravel in detail.
Another issue is at what point it is possible to identify a vulnerable people? Before a disaster is it possible to identify characteristics at an appropriate scale of a vulnerable individual, household, community or region? During a disaster are the characteristics that define vulnerable the same or do they alter and so does who or what is vulnerable alter? Likewise after the hazard do these characteristics change again? What I hope is becoming clearer is that vulnerability, resistance and resilience vary as the hazard or disaster varies and are in a dynamic relationships with the disaster as it unfolds.

So how can this set of concepts be used to analyse the floods in Pakistan? Taking into account that I have as much sketchy and incomplete information as everyone else that relies on selective media reports and selective web for information about the floods could I suggest the following. The static aspects of vulnerability, before the disaster strikes, could be analysed by looking at access to resources and power of different parts of the population. The poor, to generalize, have little access to resources such as funds for crops, for irrigations and the like. They also have little access to political power to ensure the infrastructure serves their needs. They also have little access to resources to escape the disaster (anyone else think it’s odd that media can hire helicopters and transport into and out of disaster zones but the victims can’t?!) Identifying low income areas may provide an indication of populations likely to be unable to cope with a disaster, a sudden disruption to their daily lives. Households not integrated into a wider community may not be able to resist a disaster as well as households who are well bounded within a wider community. This property, however, may not become clear until during or after a disaster, until the community responses to the event (indeed the community may be defiendby hte disaster such as the development of a community within refugee camps).

There is also a dynamic aspect to vulnerability; the manner in which relationships are organised and the manner in which they change through normal times and then during and after a disaster. Such flows could include the transport infrastructure; a key aspect that appears to have failed during this disaster and which has dramatically affected the ability of the institution of government to maintain an effective relationship with vulnerable groups. At a local level, however, is the transport infrastructure that remains intact sufficient for the local population to move to safety and then initiate community based activities that represent resilience at that level?

But vulnerability doesn’t need to be confirmed to the lowest entity you can identify and, as you might expect, the nature of that vulnerability might change as you change your scale or entity of analysis. The Pakistani government, for example, has come in for criticism in its handling of the disaster but you could argue it is vulnerable as well. It has an inadequate infrastructure for dealing with such a wide ranging disaster (although it does beg the question does any country have an adequate infrastructure for coping with such a spatially disperse disaster). The institutions of government respond using particular procedures, mechanisms and pathways that may be vulnerable if specific aspects of the infrastructure are lost. In addition, Pakistan could be viewed as having a lack of access to appropriate resources, both financial and material, (e.g. lack of reserve funds, lack of helicopters) to respond to the disaster. The country itself could be viewed as vulnerable because of its relative developing status compared to other countries and the uneven development, and so uneven access to resources and power, within the country.

There may be no answers in the above analysis but I do hope it points out some interesting and important questions about what vulnerability may mean and how that meaning changes as the nature of the disaster unfolds and, importantly, as the resilience of different the communities emerges.

Donations for the onoging disaster can be made within the UK via the Disaster Emergency Committee, DEC, go to http://www.dec.org.uk/ or to Islamic Relief UK http://www.islamic-relief.org.uk/Donate.aspx?gclid=CNeE4KSUz6MCFYT-2AodqALklg.