2.3.1 Weather-related disaster damage is increasing exponentially

Across the 21 countries and states (see Box 2.4), disaster occurrence and loss was down significantly in 2009. Given that most extensive risk is weather-related, its manifestations are closely related to climate variability, associated for example with the El Niño Southern Oscillation. As such it can be expected that both the number of events and losses increased again in 2010. Looking at the longer-term picture, the past 20 years have seen a significant increase in the number of local areas reporting losses, the number of houses damaged, the number of people affected and the damage to health and educational facilities associated with extensive disasters (Figure 2.24). This reinforces the view that the rapid increases in both population and GDP exposure described in Section 2.2 have not been addressed by commensurate reductions in vulnerability.

Extensive risk is also rising in relative terms. The number of houses damaged relative to population growth in all 21 countries and states has increased by approximately 600 percent since the early 1990s (Figure 2.25). The enormous difference between this increase and the increasing economic loss to major hazards, described in Section 2.2, reflects how extensive disaster loss is largely unaccounted for, disguising a transfer of risk within countries to low-income households and communities.

2.3.2 Extensive risk is expanding geographically

Spatially, the expansion of extensive risk mirrors urban and regional development and hence increasing population and asset exposure. Across all 21 countries and states, the number of local administrative areas reporting disaster losses has increased more or less continuously over the past 20 years (Figure 2.26). In Mozambique, for example, more local administrative areas reported losses more often between 1999 and 2009, than between 1989 and 1999 (Figure 2.27).

2.3.3 Mortality is still rising in the countries with the weakest risk governance capacities

These global trends in risk vary widely from country to country, indicating that risk accumulation processes that mirror development are as heterogeneous as development itself. However, confirming again the findings of Section 2.2, countries with stronger risk governance capacities appear better able to reduce mortality than to reduce the numbers of houses damaged and people affected (Table 2.7). The increase in extensive mortality risk reported in countries like Bolivia, Mozambique, Nepal and Yemen reflect low levels of development. In contrast, mortality risk in Chile and Costa Rica is falling while the rate of housing damage is rising. The heterogeneous nature of risk is further illustrated in Box 2.6, which explains that even in the world’s largest economy, the United States of America, there are major differences in risk governance capacities among wealthier and poorer states and counties.

2.3.4 Revisiting the underlying risk drivers

Improved reporting of disaster impacts and losses makes it difficult to determine with precision the cause of any increase in reports of disaster impacts and losses over time, even in the last 20 years. In the case of national disaster databases, there is certainly evidence of improved reporting in some countries such as Costa Rica and Sri Lanka, where new official data sources began to contribute to the datasets during the GAR11 analysis period. Nevertheless, improved reporting alone does not appear to explain the increase in damaged housing, for example, across the 21 countries and states used in the GAR11 analysis.

New case study evidence supports GAR09’s finding that increasing extensive risk is closely related to the challenges low- and middleincome countries face in addressing underlying risk drivers and reducing vulnerability. Risk is increasing most rapidly in small- and medium-sized urban centres with relatively weak capacities for managing urban growth (Table 2.8). Compounding this, landslide and flood risk at the local level is closely associated with poverty, and overall risk is magnified by deforestation and the destruction of coastal ecosystems.

Risk driver	Outcome
Badly planned and managed urban development Disaster risk may be increasing faster in rapidly growing smalland medium-sized urban centres than in either rural areas or larger cities. Compared with small and medium urban centres, large urban centres and megacities generally have stronger risk governance and investment capacities along with slower growth, both of which facilitate planning and urban management.	Latin America In most Latin American countries, the number of disasters reported in small and medium urban areas is increasing at a faster rate than in large urban centres and megacities (Mansilla, 2010 x  Mansilla, E. 2010. Riesgo urbano y políticas públicas en America Latina: La irregularidad y el acceso al suelo. Background Paper prepared for the 2011 Global Assessment Report on Disaster Risk Reduction. Geneva, Switzerland: UNISDR. Click here to view this GAR paper. ).19  More than 80 percent of all reports of disaster loss in Latin America occur in urban areas. Although each country has a different urban structure, 40–70 percent of all nationally reported disasters occur in urban centres of less than 100,000 inhabitants, and 14–36 percent in small urban centres. This proportion is growing. In Mexico for example, small and medium urban centres accounted for 45.5 percent of total municipal disaster loss reports in the 1980s, and 54 percent since 2000.
	Colombia In Colombia, municipalities with the most rapidly growing urban population between 1995 and 2005 were also more likely to experience more disasters and have higher numbers of houses damaged (Serje, 2010b x  Serje, J. 2010b. Preliminary extensive risk analysis for GAR11. Case study prepared for the 2011 Global Assessment Report on Disaster Risk Reduction. Geneva, Switzerland: UNISDR. Click here to view this GAR paper. ).
Ecosystem decline Deforestation in tropical areas is a critical global driver of climate change. It also has important and often negative local feedbacks, leading to increases in mean temperatures and decreases in mean precipitation. Coastal ecosystems, including coral reefs, sea grasses, mangroves and other beach vegetation, play a key role in mitigating impacts of storm surges and coastal flooding. Unfortunately, coastal ecosystems in many areas are in decline, simultaneously increasing disaster risk while threatening the sustainability of local economies.	Peru In the Peruvian Amazon, deforestation at least partly explains why some watersheds experience greater disaster loss and damage as a result of floods and landslides than others. To establish this link, satellite images in selected watersheds of the upper Amazon were analysed to determine the rate of conversion of forest into agricultural land and other uses between 1986 and 1998. Statistical correlations suggest that those watersheds with the highest rates of deforestation are likely to experience greater disaster mortality and housing damage (Serje, 2010b x  Serje, J. 2010b. Preliminary extensive risk analysis for GAR11. Case study prepared for the 2011 Global Assessment Report on Disaster Risk Reduction. Geneva, Switzerland: UNISDR. Click here to view this GAR paper. ; Tonini et al., 2010 x  Tonini, M., Vega Orozco, C., Charrière, M., Serje, J. and Tapia, R. 2010. Relation between disaster losses and environmental degradation in the Peruvian Amazon. Lausanne, Switzerland:Institute of Geomatics and Risk Analysis, University of Lausanne. Click here to view this GAR paper. ). Note, however, that the clear link between deforestation and disaster loss does not mean that deforestation causes the loss directly. Deforestation usually occurs in areas with an expanding agricultural frontier and growing small urban centres, and other factors including increasing hazard severity, exposure and vulnerability, also shape risk.
	Jamaica In Negril, Jamaica, up to 55 metres of beach depth has been lost in some areas as a consequence of the degradation of coral reefs, the removal of sea grass meadows, the loss of mangroves, and increasing urban and agricultural pollution. Coral reefs, for example, provide ecosystem services that include shoreline protection, supply of beach material, tourism revenue and local fishing. In Negril, coral reefs have been degraded in numerous ways: damage inflicted by major storms (such as Hurricane Ivan in 2004); coral bleaching through increased sea temperatures; pollution from sewage and agricultural run-off causing algal growth that suffocates coral; invasive predators such as lion fish; and destructive fishing practices. Mangroves protect beaches and shorelines by dissipating near-shore waves and play a vital role as a breeding habitat for fish and shellfish, but they have been harvested for firewood and building materials. Sea grass meadows are also a significant natural source of beach material but are in decline mainly because of removal by the tourism industry. Other coastal ecosystems suffering degradation include wetlands and forests. This degradation of coastal ecosystems has increased storm surge risk in Negril. A 1-in-50-year hurricane has the potential to produce storm waves of almost 7 metres, affecting around 2,500 local residents, more than 60 hotels and their guests, and water and sanitation infrastructure (UNEP, 2010 x UNEP (United Nations Environment Programme). 2010. Linking ecosystems to risk and vulnerability reduction: The case of Jamaica. Risk and vulnerability assessment methodology development project (RIVAMP). Results of the pilot assessment. Geneva, Switzerland: UNEP (United Nations Environment Programme), Division of Environmental Policy Implementation and Division of Early Warning and Assessment/GRID-Europe. Available at http://www.unep.org/pdf/GEO_Haiti2010_IN(web).pdf. ).
Poverty Within countries, poorer areas tend to have higher disaster risk, illustrating the complex interactions between poverty and disaster risk analysed in detail in GAR09 (UNISDR, 2009 x UNISDR (United Nations International Strategy for Disaster Reduction). 2009. Global assessment report on disaster risk reduction: Risk and poverty in a changing climate. Geneva, Switzerland: United Nations International Strategy for Disaster Reduction. Click here to go to GAR09 page. ).	Indonesia In Indonesia, mortality risk from landslides is higher in areas with low levels of human development and higher levels of poverty. Detailed information on hazard factors, population exposure and a range of socioeconomic indicators was used to build a landslide risk model with a sub-national level of resolution, calibrated with disaster loss data from the recently developed DIBI information system (see Box 2.5). Landslide mortality risk correlated positively with physical exposure and the Human Poverty Index, and negatively with the Human Development Index. Poverty explained a significant proportion of the variance in landslide risk between provinces (Cepeda et al., 2010 x  Cepeda, J., Smebye, H., Vangelsten, B., Nadim, F. and Muslim, D. 2010. Landslide risk in Indonesia. Background paper prepared for the 2011 Global Assessment Report on Disaster Risk Reduction. Prepared by the International Centre for Geohazards, Norwegian Geotechnical Institute. Geneva, Switzerland: UNISDR. Click here to view this GAR paper. ). The poorer the province the greater the risk, and vice versa.
	Colombia A similar modelling exercise in Colombia showed that those municipalities with a greater proportion of unsatisfied basic needs and lower GDP per capita were more likely to see more people affected and more houses damaged during floods (OSSO, 2011b x  OSSO (Southwestern Seismological Observatory). 2011. Extensive risk analysis for 2011 global assessment report on disaster reduction: Análisis de manifestaciones de riesgo en America Latina: Patrones y tendencias de las manifestaciones intensivas y extensivas de riesgo. Background Paper prepared for the 2011 Global Assessment Report on Disaster Risk Reduction. Geneva, Switzerland: UNISDR. Click here to view this GAR paper. ).

2.4 Impacts on children and internal displacement

The mechanisms through which disaster losses contribute to poverty were explored in depth in GAR09 (UNISDR, 2009). This year’s report expands on the different and specific disaster impacts that affect child welfare and development.

Children are affected particularly severely by disasters and constitute an extremely large percentage of those who are most vulnerable (Bartlett, 2008). This is supported by a number of studies on how disasters affect children’s medium-term development (Baez and Santos, 2007; López-Calva and Ortiz-Juárez, 2009; Rodriguez-Oreggia et al., 2010). For example, destroyed or damaged schools together with the loss of household assets and livelihoods can force children out of school, and infant malnutrition caused by loss of food supplies may cause stunting and lead to poor educational achievement and greater propensity to disease.

Recent studies conducted in Bolivia, Indonesia, Mexico, Mozambique, Nepal, the Philippines and Viet Nam provide evidence of how extensive disasters negatively affect children’s education, health and access to services such as water and sanitation, though it was difficult to establish significant relationships between intensive disasters and child welfare (Tarazona and Gallegos, 2010; Seballos and Tanner, 2011). Given the importance of primary education for human and long-term economic development, these findings should serve as a warning to governments.

In areas in Bolivia that experienced the greatest incidence of extensive disasters, the gender gap in primary education achievement widened, pre-school enrolment rates decreased and dropout rates increased. Equivalent areas in Nepal and Viet Nam saw, respectively, reduced primary enrolment rates and a drop in the total number of children in primary education. Extensive disasters also led to an increased incidence of diarrhoea in children under five years of age in Bolivia, an increased proportion of malnourished children under three in Nepal, an increased infant mortality rate in Viet Nam, and an increase in the incidence of babies born with low birth weight in Mozambique. This study also found evidence of negative impacts in terms of access to water and sanitation in Mexico and Viet Nam. These impacts indicate a need for greater consideration of children’s vulnerability (Box 2.7).


Disasters also contribute to internal displacement (Box 2.8). Hazards such as floods, although causing relatively low mortality, destroy many houses and hence cause considerable displacement. Between 1970 and 2009 in Colombia, for example, 24 of the country’s 35 disaster loss reports detailed floods that killed fewer than 10 people but destroyed more than 500 houses. In total, around 26,500 houses were destroyed, potentially displacing more than 130,000 people. In the Indian state of Orissa, 265 floods with similar low mortality rates destroyed more than half a million houses.


Intensive disasters also lead to large-scale internal displacement. Pakistan’s 2010 floods have to date left an estimated 6 million people in need of shelter; India’s 2008 floods uprooted roughly 6 million people; Hurricane Katrina displaced more than half a million people in the United States of America; and Cyclone Nargis uprooted eight hundred thousand people in Myanmar and South Asia (IDMC, 2010).

Assuming a family size of five in the 21 countries and states included in the GAR11 analysis, the destruction of 5.9 million houses in intensive disasters between 1970 and 2009 would have displaced almost 30 million people. Although extensive disasters account for less than one-fifth (19 percent) of destroyed housing, this implies an additional 7.5 million displaced people, who are typically less visible than those displaced in intensive disasters subject to large-scale international humanitarian assistance.


Note