Chapter 3: Risk
The term "risk" has different meanings: (a) as a synonym for probability of a harmful effect occurring and (b) as a synonym for the mathematical expectation of the magnitude of the undesirable consequence (even as a quasi-synonym of consequence, whereby risk has a similar meaning to undesirable outcome).
Ten years from the publication of this GAR, the world population is projected to exceed 8 billion, and by 2055, more than 10 billion. This growth in population has resulted in an increase in economic losses due to natural hazards from $14 billion annually to more than $140 billion between 1985 and 2014.
In the period since GAR15, the hazard community has shifted away from a focus on individual hazards and broadened its scope to examine more complex, real scenarios that acknowledge the likelihood of one hazard eventually leading to another (cascading hazard), or multiple hazards crossing in either time and/or space creating an even larger disaster. In addition, the Sendai Framework has expanded the range of hazards to be considered.
Most hazard sciences now use open source tools and are part of a larger movement promoting the widespread use of sharing open data. The democratization of risk information empowers individuals, communities and governments to draw conclusions and influence their own exposure and vulnerability. The shift towards open source and open data has provided a foundation for greater collaboration on a global scale within hazard communities and across hazard science.
The march towards openness, collaboration, interchange and cooperation has momentum. While there will be holdouts to this movement, trends in technology and data science suggest they will be increasingly in the minority. Openness solves many challenges, but there are still challenges to producing and communicating good risk information.
This part will outline developments related to understanding of risk since the publication of GAR15. In addition to expanding the scope of hazards under consideration beyond natural hazards, the Sendai Framework has called for recognition of the impact on and role to play for local, regional, national and global actors, and for a richer understanding of exposure and vulnerability. Furthermore, it considers an expanded list of hazards including human-made hazards and natural hazards that have been historically difficult to represent. In investigating the dynamic interconnected nature of risk, it calls for the imperative to develop new ways of thinking, living and working together that recognise the nature of systems.
New challenges call for novel solutions. While the GAR may never again produce individual risk metric figures for countries, this GAR is intended to give as true a picture of risk as possible. Facing that challenge, it must be acknowledged that: (a) the truth can be complicated and (b) some readers will be disappointed that the focus of this section is not on presenting probable maximum loss (PML) and average annual loss (AAL) figures. Furthermore, inasmuch as this GAR seeks to pay due respect to the expanded scope of hazards in the Sendai Framework there are hazards this report has previously covered that are not represented - notably, wind and storm. This GAR does include many hazards that have never been covered before, including biological risk, chemical and industrial, environmental, NATECH and nuclear/radiological. The GAR has never been exhaustive in its coverage of hazard and while GAR19 makes an effort to be comprehensive, there are and always will be sections that stand to be enriched in future iterations.
People and assets around the world are being exposed to a growing mixture of hazards and risks, in places and to an extent previously unrecorded. Heat-waves mixed with drought conditions can trigger intense wildfires that cause high levels of air pollution from burning forests and hazardous chemicals, such as the dioxins from burning plastics, as well as water pollution from the flame retardants used to fight the fires leaking into waterways, drinking water and marine systems. In other words, a perfect storm is created by the complex interlinkages of different natural and anthropogenic events and processes.
This part concludes with an exploration of drought hazard from a multidimensional perspective. Past GARs did not present drought risk partly because it is a highly complicated risk. The drivers are manifold, and the impact is felt more strongly in the secondary effects (lost livelihoods, forced migration, and top soil and nutrient erosion) than in primary effects. The chapter on drought will serve as an introduction to an off-cycle GAR special report on drought to be published in 2020.
The growth of accuracy and sophistication of risk assessment has been propelled by the hazard community. This is reflective of a past paradigm where disaster and hazard were used interchangeably. It also reflects the emphasis on empiricism in risk science. In many ways, that emphasis on scientific methods to understand hazards has led to a state in which disaster research is accorded a certain respect. Hazard research continues to dominate global research related to understanding risk.
The era of the Sendai Framework has opened the door for the inclusion of a broader community of research in understanding the true nature of risk. Social science researchers, economists, public policy specialists, epidemiologists and others who can contribute valuable information about the nature of vulnerability and exposure are finding a welcoming community whose main objective is to give increasingly clear and accurate risk information. There is no doubt that the nature of risk information is and will continue to be quantitative, but the focus on probabilistic modelling and homogeneous data sets is giving way to a future that is less definitive and more accurately representative of the world as it is.
In this section, there is still a focus on hazards first, but the interconnection among hazards and the connections of the hazard research community to other risk research is validation of the Sendai Framework.
This peril has been responsible for an average direct death toll of over 20,000 people per year in the last several decades and economic losses that can reach a significant fraction of a country's wealth. On average, earthquakes constitute 20% of annual economic losses due to disasters, but in some years, this proportion has been as high as 60% (e.g. in 2010 and 2011). In Central America and the Caribbean, the earthquakes of Guatemala (1976), Nicaragua (1972), El Salvador (1986) and Haiti (2010) caused direct economic losses of approximately 98%, 82%, 40% and 120% of the nominal GDP of each country, respectively.
While global earthquake models have not changed dramatically, many of the inputs have changed, as has the way in which earthquakes are being studied and understood. GAR15 focused on earthquakes as ground shaking and the impact of earthquakes as related to structural damage to buildings due to shaking. Nearly five years on, knowledge of earthquakes is being informed by new models, and by a better understanding of faults and thus movement within time and space. This has been facilitated by greater collaboration enabling local-level data to help inform the global level.
In general, earthquake models are heavily based on data from past earthquakes: magnitude, frequency, ground shaking and damage. Thus, models at the global level have been created mainly through statistical analyses of past events and empirical data on damage and mortality. Models are improving in several ways: increased understanding of how active faults accumulate seismic energy; greater availability of ground shaking recordings from damaging earthquakes; better understanding of the vulnerability of structures from field observations as well as computer simulations; and better descriptions of the human and built environment from a wide range of sources, including satellite imagery and crowdsourcing.
Global models now integrate local information about faults and microfaults as well as to reflect verified plate movement measurements. There is a growing emphasis on the use of geodesy (the branch of mathematics dealing with the shape and area of the Earth). Each factor affects ground shaking differently, thus the greater the level of detail, the more accurate forecasting can be.
The Global Earthquake Model (GEM) now includes nearly 10,000 fault lines. This level of comprehensiveness is available only due to the confluence of improved satellite capability, expanded availability of computing power and the inputs of hundreds of national and local seismic specialists.
As the level of available detail varies by location (by region, by country and sometimes even within countries), to ensure the most up-to-date data is incorporated into a global model, it is necessary to apply consistent methodologies and tools at all levels of analysis, from local to global. This information can then be combined into a homogeneous mosaic that allows comparisons of hazard among locations and regions.
Regionally, seismic models have extended such that there are now models for a larger part of the world at a better quality with improved catalogues and geological parameters than ever before. Risk modelling has progressed to include cascading hazards in the models. An example of this new capacity is the increasing focus on modelling contingent losses or indirect losses. Pilot efforts are showing that it could be possible to estimate the price increases for certain types of goods when disaster events of different scales occur in some contexts. For risk managers and planners, this will be useful in understanding the probable knock-on effects of the event, but also to inform emergency measures.
Tsunamis must be treated as a multidisciplinary hazard. They can be triggered by earthquakes, landslides, volcanoes or meteorological events, with large earthquakes being the most frequent trigger. Because their drivers require specific conditions to result in a tsunami, they are decidedly rarer than their triggering events. Tsunamis have a basis of historical evidence, but the data set is too sparse to characterize the tsunami risk on each specific coastline, especially in confined areas where there is a limited coastline section. Making this more challenging, over the last 100 years, only a handful of truly devastating tsunamis have occurred, contributing to most of the disaster tsunami losses across the globe. Large tsunamis occur with relatively low frequency but have potentially high impact. In the last two decades, this has been demonstrated, for instance, by the Indian Ocean (2004) and the Great East Japan (2011) tsunamis. The scale of these disasters far exceeded the previously perceived risk in these areas.
Assessing tsunami risk requires a comprehensive and multidisciplinary approach. It is a topic that includes a wide range of disciplines, such as geophysics (e.g. seismology, geology and faulting), hydrodynamics and flow modelling (e.g. landslide dynamics, volcanology, coastal engineering and oceanography), vulnerability and risk assessment (e.g. geography, social sciences, economy, structural engineering, mathematical and statistical sciences), in addition to disaster risk management and mitigation.
Figure 3.2. Selected tsunami wave heights (maximum wave heights recorded)
(Sources: National Oceanic and Atmospheric Administration National Geophysical Data Center/World Data Service Global Historical Tsunami Database 2019; National Centers for Environmental Information 2019)
The tsunami maximum wave heights in Figure 3.3 do not correlate with their level of damage. The largest known tsunami occurred in Lituya Bay, Alaska, in the United States of America in 1958. The massive scale of the wave caused relatively little damage due to the limited exposed stock in the area at the time. The Great East Japan tsunami in 2011 and the Indian Ocean tsunami in 2004 were far smaller than the Lituya Bay tsunami, but they caused far more losses.
Tsunami hazards are heterogeneous; smaller events can cause devastation, as evidenced by the events in Indonesia with the Palu tsunami in 2018 and the Mentawai tsunami in 2010. These events exemplify cases where unconventional mechanisms generate tsunamis that are unexpectedly large given the magnitude of the triggering event.
Due to their infrequent nature, tsunamis often catch coastal communities off-guard. Perhaps the most pertinent example is the 2004 Indian Ocean tsunami that hit a largely unprepared coastal population in nearly a dozen countries and resulted in more than 230,000 fatalities. Due to the enormous consequences of that tsunami, the need for more sophisticated and comprehensive methodologies to understand and manage tsunami risk in a wider range of locations immediately became obvious. The most obvious interventions were in risk mitigation activities such as construction of wave-absorbing sea walls, elevated facilities, evacuation routes and EWSs. After 2004, tsunami research and risk mitigation activities spread to many regions that previously had very little focus on tsunami risk - particularly South and South-East Asia.
The evaluation of landslide hazard should entail diagnosis of the geo-hydro-mechanical processes bringing about the landslides that eventually generate damage.
The assessment of landslide hazard based upon geo-hydro-mechanical analysis of slopes is generally recognized to be the planning basis for countries experiencing high landslide susceptibility (e.g. in Afghanistan, in Himalaya belt slopes in Asia, in Bolivia, Brazil and the Bolivarian Republic of Venezuela in South America, and in Italy and Spain in Europe). But the experienced losses from contemporary landslide events testify that these assessments, or the mitigation measures they should have precipitated, are not appropriately developed.
The Multiscalar Method for Landslide Mitigation is a new methodology for the assessment of landslide hazard at the local scale, based on geo-hydro-mechanical analyses. This method seeks to identify the geo-hydro-mechanical contexts most common in the slopes of the region, and for the corresponding landslide mechanisms, which are then recognized as the mechanisms typical to the region. Having as a basis the set of representative landslide mechanisms can make landslide risk management at the local scale more sustainable, since it can guide the selection of the mitigation measures based on awareness of the typical landslide features and causes.
Urbanization frequently extends over unstable slopes and ancient landslides. This is particularly true for informal settlements. Therefore, landslides often affect the poorest parts of urban areas, whose expansion is restricted to land that would not withstand simple engineering tests.
While seismic science has been able to move forward with a coordinated, collaborative approach to modelling the hazard, flood science faces several obstacles that make the process of reaching the same point more complicated. Floods are simply the presence of water on land that is usually dry. The causes of that flooding can be too much precipitation, snow melt that occurs too quickly, a dam break, a tsunami or storm surge, inadequate water management practices, etc. The dynamics that dictate flood risk are difficult to model - a key reason why not all flood causations can be modelled with contemporary resources. There are models for many different drivers of flooding, but not all, and the work of harmonizing the different drivers into a harmonized flood model remains a challenge for the flood community.
Several different flood models have been developed for riverine and coastal flooding. But the challenge in developing a more comprehensive global model is to combine these models together. A first step in this direction has been made by linking one hydrodynamics model with downstream boundary conditions from a tide and storm-surge data set. In doing this, the linked effects of flooding at river water levels and in estuaries have been mapped globally. Other initiatives are developing methods to nest local flood models within global models, thereby increasing computational efficiency and enhancing localized accuracy in those areas where the local models exist
When assessing flood risk, a key concern is related to triggering factors. There is no single source that causes a flood; it can arise from multiple drivers. Considering the challenges in accuracy related to short-term weather forecasts, where at least some of the dynamics can be modelled, the challenge of risk projection for precipitation drivers of flooding are orders of magnitude more complex. Precipitation patterns must consider multiple dynamic sources. Even in the same catchment area, the same precipitation distributed in different ways can lead to vastly different results. Other conditions must be factored in, including the soil conditions (very dry, partial saturation, snow melt, etc.), and all those elements must then be linked to local factors that are not always possible to project at the global level. The primary difference between global and local models is not the processes - those are effectively the same - but rather the ability to tailor them to a local context that can make the difference for producing a comprehensive understanding of risk.
Older hydrological models were focused on projecting probable discharge of rivers, creating a time series of the flow in the river and applying those discharge values to a hydraulics model that incorporated flood flow and depth. Now, with the ability to run calculations on far more powerful computers, the hydrological cycle can be resolved in a more accurate way, thus enabling improved simulation of hydrology and the production of more reliable values of discharge.
Using these tools, many probabilistic flood maps are now available. Recent work to combine them has highlighted the significant advances possible in recent years. Through the Global Flood Partnership (GFP), work is under way to compare the various existing models and identify gaps that will require further research and development. GFP is a multidisciplinary group of scientists, agencies and flood risk managers focused on developing efficient and effective global flood tools. Its aim is to build cooperation for global flood forecasting, monitoring and impact assessment to strengthen preparedness and response and to reduce global disaster losses. Much like seismic science, the ideal case is to use locally produced models, and a plan is required to collect these and figure out how to fill gaps. The result should provide a basis for other models and enable them to be mutually improved.
The increased number of intense heat-waves and wildfires that has been recorded during recent years on a global basis has raised great concerns. It is apparent that projected climatic changes may significantly affect such phenomena in the future. Each year, wildfires result in high mortality rates and property losses, especially in the wildland urban interface (WUI). These fires affect millions of people and have devastating global consequences for biodiversity and ecosystems. Wildfire disasters can rapidly change their nature into technological disasters (e.g. in mixed areas of forest and residential, in heavy industrial or in recycling zones). In such cases, there is a global concern because toxic components such as dioxins are released, as well as fine and ultrafine particles with transboundary effects. Even though international policies and fire safety legislation have resulted in effective prevention mechanisms, environmental and technological fire hazards continue to threaten the sustainability of local populations and the biodiversity of affected areas.
The year 2018 was reported as one of the warmest, affecting European Mediterranean countries such as Greece, Italy, Portugal and Spain, and also the countries of Central and Northern Europe. For example, Austria's June 2018 national temperature was 1.9Â°C above average and was one of the 10 warmest Junes on record. Higher temperatures have generally been correlated with extreme weather events such as prolonged droughts, heat-waves and flash floods. The short-term precipitation period that is spatially intensive usually causes flash floods and hence it more often occurs in drier climates. Under such circumstances, fire incidents in dry climate zones can easily be converted to megafires such as the Greek fires of August 2007, which destroyed huge forest areas, and even within the Arctic Circle, as seen in the Swedish wildfires of July 2018.
There is a general challenge surrounding the definition of fires. In the European Union (EU) the focus has been on forest fires. More frequent occurrences of wildfires have spurred an expanded definition into wildfire that does not require the fire at any point to affect a forest. A wildfire is a fire that is out of control. This excludes fires set for legitimate purposes such as crop burning but would include the same fires if they spread outside of the intended area.
Biological hazards cover a category of hazards that are of organic origin or conveyed by biological vectors, including pathogenic microorganisms, toxins and bioactive substances. Examples are bacteria, viruses or parasites, as well as venomous wildlife and insects, poisonous plants and mosquitoes carrying disease-causing agents. While biological hazards also cause diseases in plants and animals, this chapter focuses on those biological hazards that affect human health.
Like other hazards, biological hazards and their associated infectious diseases occur at different scales with varying levels of consequence for public health. Diseases may be categorized by the way in which they are spread and people are infected, namely: water and food-borne diseases, where the pathogen can enter the body via contaminated food or water; vector-borne diseases, which involve mosquitoes, ticks and other arthropod species, or other animals that transmit the disease from animals to humans (zoonotic diseases) or among humans; air-borne or respiratory infections, which are spread between humans by the respiratory route; and other infectious diseases involving contact with bodily fluids such as blood.
Biological hazards affect people at all levels of society. At the extreme, epidemic infectious diseases affect millions of people every year, with potentially severe consequences for individuals, communities, health systems and economies, especially in fragile and vulnerable countries where they are most common. However, no country is immune to the risk. New pathogens continue to emerge by mutating, re-assorting and adapting. Previously well-understood infectious agents change their behaviour or scale of impact as the world is getting warmer and more populated, with associated animal husbandry strategies, and with ecosystem changes, increasing speed of transportation and mass distribution systems.
As infectious diseases travel easily across administrative boundaries, the world's defences are only as effective as the weakest link in any country's efforts to anticipate and prevent emergence and outbreak at all scales. Biological hazards and their impact on global public health have brought to prominence the need for a collective and coordinated mechanism involving all sectors to prevent new risks, reduce and mitigate existing risks, and strengthen resilience. This approach is being promoted and reinforced by the integration of biological hazards in whole-of-society and all-hazard approaches to the management of risks, as reflected in the Sendai Framework, SDGs and the Paris Agreement, which are complemented by international health regulations (IHRs) and other relevant global, regional, national and subnational strategies and agreements.
Radioactivity and the radiation it produces existed on Earth long before life emerged. In fact, they have been present in space since the beginning of the Universe, and radioactive material was part of the Earth at its very formation. But humanity first discovered this elemental, universal phenomenon only in the last years of the nineteenth century. Most people are aware of the use of radiation in the nuclear power production of electricity or in medical applications, yet many other uses of nuclear technologies in industry, agriculture, construction, research and other areas are hardly known at all. The sources of radiation causing the greatest risk to the public are not necessarily those that attract the most attention (Figure 3.10). In fact, everyday experience such as air travel and living in well-insulated homes in certain parts of the world can substantially increase exposure to radiation.
There is no formal distinction between nuclear and radiological risks and thus between associated safety arrangements. However, it is a well-established practice to distinguish exposures related to nuclear power generation from other radiation exposures. From the physical point of view, both situations may result in the same kind of radiation exposure, so this distinction considers the different characteristics of the source of the risk. This GAR assumes that nuclear risks arise (or may potentially arise) from the uncertainties in the management of a nuclear chain reaction or the decay of the products of a chain reaction. Consequently, the radiological risks arise from uncertainties related to any other activities involving ionizing radiation.
Figure 3.10. Potential biological impacts of radiation damaging a cell
Industrial production is a central characteristic of the modern world economy. Industry creates jobs and provides a wide range of essential materials, products and services. However, authorities, in cooperation with industry, must ensure that industrial facilities producing, handling or storing hazardous substances such as tailings management facilities (TMFs), pipelines, oil terminals and chemical installations are safely located and operated, as accidents can have far-reaching and severe effects on people, environments and economies.
Industrial hazards originate from technological or industrial conditions, dangerous procedures, infrastructure failures or specific human activities. These include toxic releases, explosions, fires and chemical spills into the air, adjacent water courses and land. In many countries, industrial hazards are exacerbated by ageing, abandoned or idle installations. These problems are amplified by insufficient institutional and legal capacities to deal with technological risk reduction. Natural hazards - for example, storms, landslides, floods or earthquakes - can also cause industrial accidents by triggering the release of hazardous substances from industrial facilities that are located within their path of destruction (see section 3.1.9). The impact associated with industrial accidents relate to loss of life, injury, or destruction or damage of assets that could occur to a system, society or a community. Effective management of risks requires cooperation within and across systems, sectors, countries and scales.
Most industrial accidents entail the release of hazardous substances into water bodies with grave impacts on water resources, threatening the availability of safe water for drinking, household use and agriculture, as well as human safety.
For many decades, the issue of industrial accident prevention, preparedness and response has been of concern to governments, as well as industry. In the mid-1980s, the issue took on a new level of urgency and political importance in response to the Bhopal accident in India, which resulted in more than 15,000 deaths and more than 100,000 people affected. While regulation and new standards have driven significant progress in industrial safety in the past 40 years, major accidents still occur as countries face new challenges and emerging risks. In recent years, extreme weather-related events triggered industrial accidents with severe environmental and economic consequences, such as Hurricane Harvey in the United States of America.
A multidisciplinary and cross-sectoral approach to addressing industrial accident risk is required. The Sendai Framework promotes this across its four priorities in the systems-based approaches to risk management.
This section explores the trends in industrial risks and the underlying drivers of these risks (identifying the casual factors). It examines how progress in managing risks is measured, introduces industrial accident risk reduction approaches, and explores challenges and opportunities for effective risk management in the future.
Many of the goods and services upon which societies depend are provided by industrial activities. From refining, oil and gas production and transport, to nuclear power generation or the preparation of specialty chemicals, many of these activities have constructed inherent susceptibility to shocks, including those provoked by natural hazards.
Natural hazards have the potential to surpass safeguards, triggering negative impacts that may entail hazardous substance release, fire, explosion or indirect effects with wider repercussions than those felt in the immediate proximity. The cascading technological side effects of natural hazards are called NATECH accidents.
Figure 3.17. Hurricane Harvey caused several oil spills and chemical releases in Texas, 2017
(Source: Union of Concerned Scientists 2019)
Disclaimer: The boundaries and names shown and the designations used on this map do not imply official endorsement or acceptance by the United Nations.
Evidence from the latest intergovernmental and global assessments shows that the planet is overheating and becoming increasingly densely populated. Climate change, food insecurity, rapid urbanization and growing levels of pollution are damaging human and ecosystem health. Growing inequalities in wealth and access to technology and resources are leading to malnutrition, conflicts and the displacement of millions of people.
Understanding of environmental hazards and associated risks and distributional impacts caused by these pressures has been enhanced through the assessments of various key international scientific bodies. The concept of interlinkages among environmental risks lies at the heart of the concept of planetary boundaries and dynamic systems. Four out of the nine planetary boundaries (climate change, loss of biosphere integrity, land-system change, altered biogeochemical cycles (phosphorus and nitrogen)) have now been crossed. Fifteen out of 24 categories of ecosystem services are in decline due to overuse of resources. The spread of zoonoses and invasive alien species is being exacerbated by climate change and global trade, and is already posing direct threats to native and endemic species and ecosystem functioning. Overharvesting, land-use change, unsustainable use of - and lack of fair access to - genetic resources, and climate change are key drivers of the decline in wild plant resources, including those used commercially for food and medicinal purposes. Approximately 15,000 species or 21% of global medicinal plant species are now endangered due to overharvesting and habitat loss.
Intense heat-waves, wildfires and storms occurred in 2018. The 20 warmest years on record have all occurred in the last 22 years. Meanwhile, GHG emissions keep rising (another 2.7% increase in 2018) and extreme weather-related events continue to spread and intensify globally.
In past GARs, the production of the Global Risk Model and standard risk metrics (AAL, PML and hybrid loss exceedance curves) relied on a global data set of standardized and homogeneous exposure data. Due to the heterogeneity of national reporting and the availability of data, model-based exposure calculations relied on an understanding of the constructed environment and used data from satellite observations. These satellite-based exposure layers were often validated locally through ground truthing. A team of on-the-ground analysts would visit a satellite-modelled site and verify if the model layer accurately depicted the extent of construction, building use, construction type, density, floors, materials, etc. The advantage of this approach was that the loss and replacement value of construction materials is relatively easy to describe country by country, even considering local market variability. A second advantage was that the use of built assets meant that in the cases of disaster events that affected areas that were more often insured, modelling data could be validated and corrected based on loss claims. Third, many of the hazards that were modelled were major natural hazards for which extensive engineering tests had been done to better understand their robustness faced with certain natural phenomena. For example, extensive testing has been done to understand the maximum ground acceleration due to earthquakes that the different types of building materials can withstand or the scales of modelled flooding a typical family home would be expected to experience.
3.2.1 Structural Exposure
There are several difficulties in relying on structural exposure. Huge regions of the world rarely experience seismic hazards. For example, much of Africa is at relatively low risk from a seismic perspective. Furthermore, the nature of construction materials, population densities and other elements of structural exposure as modelled for Africa dictate that the true risk of many African countries was not fully revealed. As past GARs have noted, the prevalence of extensive risk in many parts of the world have been historically underrepresented. When the predominantly extensive risk profile is coupled with relatively low rates of insurance penetration and very diverse construction types, it becomes evident how difficult it has historically been to reveal the true cost of risk in many countries. Droughts, epidemics, epizootics, agricultural infestations, etc., imply effectively no damage to structures, but their economic cost in direct and indirect terms could be devastating.
The Ebola virus outbreak in Guinea, Liberia and Sierra Leone in 2014-2015, which killed more than 11,000 people, is estimated to have cost 9.4% of GDP in Guinea, 8.5% in Liberia and 4.8% in Sierra Leone. Liberia lost more than 8% of its health-care workers. Surveillance, treatment and care of HIV/AIDS, malaria and TB were set back, and the entire region suffered economic effects of the stigma. An exposure model predicated on counting and categorizing buildings would have captured effectively none of the above exposed elements and thus failed to show the true risk faced by those countries.
Figure 3.19. Projected economic losses due to Ebola in Liberia, Guinea and Sierra Leone, 2010-2016
(Source: World Bank 2016)
3.2.2 Exposure related to growth
Leaving aside the above-mentioned challenges of keeping pace with the exposure drivers for the built environment, the exposure for people, infrastructure and systems implied in those growth rates represents an astronomically complicated computation.
Exposure is not static, risk can increase with changes in exposure (e.g. a three-storey building can become five storeys over the course of a few weeks, populations can displace en masse very quickly or border crossings can be closed). In Africa, average GDP growth for 2018 was above 4%, with one third of African countries experiencing real GDP growth of more than 5% year on year. In developing countries and countries in transition, growing middle classes and expanded access to the global market are fuelling growth of exposed assets while regulatory structures and risk management capacity struggle to keep pace. The result is a compounded risk, as the scale of exposed assets and lower likelihoods of careful application of safety standards overtake public investment in risk management strategies. This applies equally to construction regulation as to food safety inspection, industrial facilities verification, disease surveillance, biodiversity preservation, etc.
Urbanization is one of the twenty-first century's most transformative trends, posing challenges in terms of exposure and vulnerability, with implications in housing, infrastructure and basic services. The developing world is experiencing 90% of this urban growth, and it is estimated that 70 million new residents are added to urban areas in developing countries each year; infrastructure development cannot keep pace with growth. Africa is the fastest urbanizing continent; between 1990 and 2015, the population in urban clusters increased by 484 million, while Asia has 89% of its population living in urban clusters. Low-income countries have seen a 300% increase in built-up areas and an 176% increase in population over the past 40 years. For example, the number of fire incidents in formal and informal dwellings per year are similar, but with approximately 18% of the population living in informal settlements, the informal settlement dweller is 4.8 times more likely to be affected by fire than someone residing in a formal dwelling. The propensity of informal settlements to fire indicates that the burden of fire disasters is often borne by the poor.
3.2.3 Environmental Exposure
Exposure in a global environmental sense takes into consideration systems for which individual quantitative figures do not exist. Over the last two decades, approximately 20% of the productivity of the Earth's vegetated surface has shown a persistent downward trend, due to climate change, biodiversity loss and poor management practices. With overharvesting of resources and land-use change remaining as key pressures, more than half of the world's ecosystems services are in decline.
The widespread loss of biodiversity and ecosystem health is evidence of a failure to account for and manage the breadth of exposed global assets. That loss also has a major effect on risk reduction and the mitigation of environmental hazards. This is because ecosystem services help to regulate climate, filter air and water, and mitigate the impact of natural hazards. There are other direct benefits such as availability of timber, fish, crops and medicines, all of which support human health. These are often lost in the immediate aftermath of a disaster and can take many years to restore. Freshwater biodiversity and ecosystem services are threatened more than any others. Rivers and wetlands the world over are distorted, dried and overwhelmed with waste, toxic pollution, invasive species, and are damaged by overfishing and overuse of irrigation water. Two thirds of all rivers are highly degraded, along with the freshwater habitat they support. This problem affects nearly 5 billion people living in high-water-threat areas.
Marine biodiversity is at risk from overfishing, ocean warming and acidification, melting of sea-ice with the loss of under-ice biota, oil and gas development, shipping, coastal habitat destruction, loss of coral reefs, eutrophication and pollution (including marine plastics, toxic algal blooms and invasive species). Terrestrial biodiversity is at risk from rising temperatures, loss of grasslands to deserts and drylands making them unsuitable for wildlife or agriculture, deforestation and degradation of tropical forests, and melting of glaciers in high mountain ecosystems and polar regions.
The impact of disasters encompasses more than just affected people or economic losses. While every society is vulnerable to risk, some suffer significantly more and recover more slowly than others when adversity strikes. Much of the existing literature on risk remains sector specific and treats vulnerability as people's exposure to risk. This section, building on the analysis offered in previous GARs and empirical evidence on the multidimensional aspects of risk exposure, reiterates the need for a more holistic and people-centred approach to vulnerability. It asks why some people do better in overcoming adversity than others by assessing the main obstacles that individuals, households and societies may face in managing risk, including challenges in terms of information, resources and incentives to build back faster and better.
Vulnerability is defined as the "conditions determined by physical, social, economic and environmental factors or processes which increase the susceptibility of an individual, community, assets or systems to the impacts of hazards". It occurs in connection with the incidence of disasters of varying magnitudes, which negatively affect the economic, social environmental/ecological profiles of countries over time. Implicit here is the notion of "differential vulnerability", referring to the different facets and variant levels of risk, to which populations are exposed, accounting for differentiated impacts and outcomes in disasters.
Hazard identification is only an initial step within a risk management strategy. While the intensity remains important, of greater importance is the profile of a population whose economic, demographic, environmental, institutional and social characteristics may place its members at greater risk before, during and after a disaster. Whereas evidence suggests that wealthier countries with more developed institutions or governance are better able to reduce disaster risk, several countries have witnessed rapid economic growth in the last few decades without a commensurable rate of vulnerability reduction.
3.3.1 Measuring Vulnerability
Disasters significantly interfere with daily life. They disrupt livelihoods, family and social networks, and interrupt schooling trajectories, access to health services, infrastructure networks, supply chains and connections of essential services, all of which are critical for people's well-being. Conceptually, the quantification of vulnerability has been surrounded by debate in recent decades about appropriate methodologies, metrics and indicators applied within quantitative, survey-based methods (single cross sections, panel surveys and community surveys) and qualitative ones. Empirical literature on risk and vulnerability is extensive. It is therefore inevitable that there would be differences in how analysts/organizations define and measure vulnerability in relation to disasters. However, considering the increasingly damaging impact of disasters, an improved ability to measure vulnerability - albeit incomplete and imperfect - should be a welcome step towards the promotion of a disaster-resilient culture.
3.3.2 Life-cycle Vulnerability
Risks and capacities to cope accumulate over lifetimes. The life-cycle approach has been commonly used to cluster different vulnerable groups and prioritize action among them. It is founded on a multidimensional concept of vulnerability, initially conceived by the World Bank, which allows the identification of risk factors for each group and thereafter forecasts the long-term consequences of those risks into next stages in life. Life trajectories are the result of investments made in preceding stages as the consequences of shocks may cascade into long-term consequences. A setback in early childhood has compounding effects throughout the rest of a person's life, in terms of growth, job and social status and the uncertainties involved with growing older and the transmission of vulnerability to the next generation. This GAR argues that the cumulative and cascading nature of vulnerability requires timely and continuous investment to effectively protect those groups whose vulnerability profiles - many structural and many tied to the life cycle - make them more susceptible to risks.
Once metrics for observation have been selected, the life-cycle approach can be used to rank various groups, by degree of destitution, by their numbers or a combination of both. As vulnerable groups are clustered according to their specific characteristics, poverty data can be extremely useful as a touchstone because it is well measured and relates to most of the other characteristics (age, gender, health and asset ownership). If such basic data is not available, the survey-based approach is preceded by a qualitative analysis to cluster population groups.
The advantages of a life-cycle approach to vulnerability is that it can forecast socioeconomic impacts for different population groups and thus prioritize risk-coping mechanisms but also develop policies to prevent these risks from cascading into the next stages in life. In other words, the analysis is not static; rather it adapts based on learning from the dynamic processes that perpetuate vulnerabilities over time.
In practical terms, when it comes to assessing such vulnerabilities this means that if a vulnerable group is identified at an early stage of analysis, analysts can better measure the elements of such vulnerabilities over time by tracking those indicators through longitudinal surveys. This type of information does not need to be collected in isolation. Rather, vulnerability analysis can inform the development of existing and future surveys and census data developed by national statistical offices (NSOs). In ideal cases, the inclusion of disaster-sensitive indicators offers improved measurements of disaster incidences, identifies linkages with other aspects of welfare and integrates those with risk management instruments.
3.3.3 Socioeconomic Vulnerability
An overreliance on asset losses to explain vulnerability obscures the relationship between risk and poverty. By definition, wealthy individuals have more assets to lose; therefore, their interests dominate in risk assessments that are limited to asset losses. But measuring asset losses misses a major dimension, particularly in the developing world; the poor are less likely to have assets to lose. Just as highly developed countries are more exposed to risk (by virtue of having more to lose), so too are wealthy people. But the losses felt by less-wealthy countries and less-wealthy people are not less important. In fact, they also lack the means and opportunity to smooth the impact of shocks while maintaining their consumption, and to recover and rebuild their assets.
To compensate for the bias towards asset losses as the key metric of vulnerability, the Unbreakable: Building the Resilience of the Poor in the Face of Natural Disasters report introduced the concept of well-being losses. In addition to traditional asset losses, well-being losses account for people's socioeconomic resilience, including:
- Their ability to maintain their consumption for the duration of their recovery
- Their ability to save or borrow to rebuild their asset stock
- The decreasing returns in consumption - that is, poorer people are more affected by a $1 reduction in consumption than richer individuals
Traditional risk assessments evaluate asset exposure and vulnerability to hazards to determine expected asset losses. The Unbreakable model additionally incorporates the socioeconomic resilience of the communities to predict well-being losses.