SM1: Trends in Losses for Two Statistical Models

Supplementary material

The supplementary material (SM) with supplementary figures (SF) and tables (ST) is provided in 4 sections: SM1-4.

SM1: Trends in losses for two statistical models

Figure SF1 shows trends in losses for two statistical models as reported in Visser and Petersen (2012). The first model is a simple exponential fit, while the second one uses and integrated random walk. Visser and Petersen suggest that both statistical models are equally valid.

Figure SF1: Loss trends in different statistical models

Sources: Visser and Petersen, 2012

Note: the source of the loss data is both times Munich Re, 2011

SM2: Loss detection and projection studies

Table ST1 provides a comprehensive overview of loss trend detection studies. The table and our analysis show that the different approaches that are used for adjusting loss lead to different outcomes. On the other hand, the strength of the different studies is that different approaches and use of indices and different datasets lead to very similar outcomes. In general, many studies have studied loss databases with a limited length (periods shorter than 40-50 years), which reduces the possibility to determine impacts of anthropogenic climate change. Table ST1 provides a comprehensive overview of loss trend detection studies.

Table ST1: Disaster loss trend detection studies

Hazard / Region / Period / Normalization / Normalized loss / Reference
Bushfire / Australia / 1900-2009 / None / No trend / Crompton et al. 2010
Flood / USA / 1926-2000 / Wealth, population / No trend / Downton et al. 2005
Flood / China / 1950-2001 / GDP / Increase since 1987 / Fengqing et al. 2005
Flood / Europe / 1970-2006 / Wealth, population / No trend / Barredo 2009
Flood / Spain / 1971-2008 / Insurance premiums, dwelling values / No trend / Barredo et al. 2012
Flood / Korea / 1971-2005 / Population / Increase / Chang et al. 2009
Flood / Bangladesh / 1970-2007 / GDP, population, vulnerability / Decrease / Tanner et al. 2007
Hail / USA / 1951-2006 / Property, insurance market values / Increase since 1992 / Changnon, 2009a
Hail / Germany / 1974-2003 / Insured value, number of contracts / Increase / Kunz et al. 2009
Windstorm / USA / 1952-2006 / Property, insurance market values / Increase since 1952 / Changnon 2009b
Windstorm / Europe / 1970-2008 / Wealth, population / No trend / Barredo 2010
Thunderstorm / USA / 1949-1998 / Insurance coverage, population / Increase / Changnon 2001
Tornado / USA / 1890-1999 / Wealth / No trend / Brooks and Doswell 2001
Tornado / USA / 1900-2000 / None / No trend / Boruff et al. 2003
Tornado / USA / 1950-2011 / Wealth, population / No trend / Simmons et al. 2013
Tropical storm / Latin America / 1944-1999 / Wealth, population / No trend / Pielke et al. 2003
Tropical storm / India / 1977-1998 / Income, population / No trend / Raghavan and Rajesh 2003
Tropical storm / USA / 1900-2005 / Wealth, population / No trend / Pielke et al. 2008
Tropical storm / USA / 1950-2005 / GDP, population / Increase since 1970, no trend since 1950 / Schmidt et al. 2009
Tropical storm / China / 1983-2006 / GDP / No trend / Zhang et al. 2009
Tropical storm / USA / 1900-2008 / GDP / Increase since 1900 / Nordhaus 2010
Storm / USA / 1952-2006 / Property, insurance market values / Increase / Changnon 2009
Thunderstorm, hail / USA / 1970-2009 / GDP / Increase / Sander et al. 2013
Weather / Global / 1980-2009 / Wealth / No trend / Neumayer and Barthel 2011
Weather / Global / 1990-2008 / Wealth, insurance conditions / No trend / Barthel and Neumayer 2012
Weather / Germany / 1973-2008 / Wealth, insurance conditions / Increase / Barthel and Neumayer 2012
Weather / USA / 1980-2008 / Wealth, insurance conditions / Increase / Barthel and Neumayer 2012
Weather / Australia / 1967-2006 / Dwellings, dwelling value / No trend / Crompton and McAneney 2008
Weather / USA / 1951-1997 / Wealth, population / No trend / Choi and Fisher 2003
Weather / World / 1950-2005 / GDP, population / Increase since 1970, no trend since 1950 / Miller et al. 2008

Source: Bouwer, 2011a, updated with recent papers, indicated in bold

SM3: Trends in drivers of risk

UNISDR for work on a global model for their bi-annual Global Disaster Assessment reports models exposure, vulnerability and risk for a number of hazards with a regional and decadal resolution. Figure SF2 shows trends in flood exposure, vulnerability and risk in relation for direct economic losses from 1980 to 2010 for South and South West Asia, and SF3 for the OECD region for 1990 to 2010 as reported in UNISDR (2011) and UNESCAP/UNISDR (2012).

Figure SF2: Trends in flood disaster risk, exposure and vulnerability between 1980-2010 for the SSW region

Note: In addition to Bangladesh the South and South-West Asian region comprises Afghanistan, Bhutan, India, the Islamic Republic of Iran, Maldives, Nepal, Pakistan, Sri Lanka and Turkey.

Source: UNESCAP&UNISDR, 2012

Figure SF3: Trends in flood disaster risk, exposure and vulnerability between 1990-2010 for OECD countries

Source: UNISDR, 2011

SM4: Modelling economic flood risk in Bangladesh

Table ST2 reports data detail on large riverine flood disasters in Bangladesh in terms of hazard (return period, flooded area), exposure (population, assets), and risk/impacts (fatalities, losses). As vulnerability cannot be observed directly, it is derived as normalized losses per GDP/flooded area (and fatalities per population/flood area), which can be considered a normalization procedure.

Table ST2: Selected impacts for the worst floods in Bangladesh over the last four decades

The empirical data form the basis for numerically modelling risk as a function of hazard, exposure and vulnerability as well as climatic and socio-economic, which is displayed graphically in figure SF4.

Figure SF4: Methodological approach for assessing risk driven by societal and climate changes for the case of riverine flooding in Bangladesh

Table ST3 tabulates key assessment modules including a discussion of functional relationship or drivers and the sources.

Table ST3: Modules, functional relationships and input data

Module and output / Functional relationship or drivers / Source
Climate (warming) / SRES scenarios A2 and B1 / Nakicenovic and Swart, 2000
Precipitation / Function of mean temperature change / PRECIS RCM for A2 and B1
(Tanner et al. 2007)
Maximum discharge / Function of precipitation / Based on
Conway et al., 2007 in Tanner et al., 2007
Economic vulnerability / Observed losses and vulnerability / Bangladesh loss statistics
(Based on CRED 2013)
Flooded area / Function of max discharge / Statistical model
(based on Mirza 2002)
Exposure / GDP, Population, assets / World Bank, 2013; SRES scenarios given by Nakicenovic and Swart, 2000
Risk: economic losses / Function of flooded area, economic vulnerability and exposure

The data to model the relationship between temperature change (over mid 20th century levels) and change in precipitation (%) is based on model results from 10 global circulation models). A simple relationship (polynomial of degree 2) between years and mean regional temperature change has been assumed which is based on the table above and the monsoon months. Modelled area flooded under projections of climate change (B1) scenario is shown in figure SF5.

Figure SF5: Projected change in frequency of severe instances with areas flooded

Source: Hassan and Conway, chapter 5 in Tanner et al., 2007.

The spatial resolution is country-level and for such analysis spatially-explicit hydrological modelling is not useful for informing hazard analysis. Instead, a statistical approach was applied for the monsoon months from June to August based on observed river flows in the three largest river basins in Bangladesh: The Ganges, the Brahmaputra and the Meghna. While rough, this appeared to be a valid approach, which may not be applicable for other countries with less homogeneous terrain. The relationship between precipitation change and (average) mean peak discharge is modeled based on Mirza (2002), who uses three global circulation models for each of the river basins. Discharge is modelled as the sum of maximum discharges (DMax_G, DMax_B, DMax_M in the 3 river basins (Meghna, Brahmaputra and Ganges), which are functions of precipitation change P.

(1)

(1’)

(1’’)

(1’’’)

A Gumbel model for event severity is used and the relationship between flooded area and the discharge levels is estimated with a nonlinear regression model based on past maximum discharge levels from 1950-2010. The coefficient of variation for the flooded area was kept constant over time, i.e. it was assumed that an increase in average flooded area due to climate change also would increase the standard deviation of flooded area.

(2)

A location and scale change re-parameterization (Fisher-Tippet) yields the following distribution:

(2’)

with

The Fisher Tippet distribution for maximum discharge and the non-linear function for flooded area are estimated based on past maximum discharge levels (1950-2010) and flooded areas. Overall, the relationship between flooded area and discharge level Dt is estimated as

(2’’)

Economic vulnerability is a function of flooded area Ft, and a time-dependent process VIt. Applying OLS regression for losses over time, we estimated the vulnerability index VIt in order to adjust vulnerability over the observed time period as well as future years.

(3)

Losses (risk) are finally calculated via multiplication of vulnerability with exposure in the respective year, and indicated in constant 2005 USD or as a percentage of GDP.

(4)

Additional references for the supplementary material

Barredo, J. I. (2009). Normalised flood losses in Europe: 1970-2006. Natural Hazards and Earth System Sciences, 9, 97-104.

Barredo, J. I. (2010). No upward trend in normalised windstorm losses in Europe: 1970–2008. Natural Hazards and Earth System Sciences, 10, 97-104.

Barredo, J., Saur, D, and M. C. Llasat (2012).Assessing trends in insured losses from floods in Spain 1971–2008 Natural Hazards and Earth System Sciences. 12, 1723–1729

Barthel, F., E. Neumayer (2012): A trend analysis of normalized insured damage from natural disasters. Climatic Change, 113, 215-237.

Boruff, B. J., J. A. Easoz, S. D. Jones, H. R. Landry, J. D. Mitchem, and S. L. Cutter (2003).Tornado hazards in the United States. Climate Research, 24, 103-117.

Bouwer, L.M. (2011a). Have disaster losses increased due to anthropogenic climate change? Bulletin of the American Meteorological Society, 92, 39-46.

Brooks, H. E., and C. A. Doswell (2001). Normalized damage from major tornadoes in the United States: 1890–1999. Weather and Forecasting, 16, 168-176.

Chang, H., J. Franczyk, and C. Kim (2009). What is responsible for increasing flood risks? The case of Gangwon Province, Korea. Natural Hazards, 48, 399-354.

Changnon, S. A. (2001). Damaging thunderstorm activity in the United States. Bulletin of the American Meteorological Society, 82, 597-608.

Changnon, S. A. (2009a). Increasing major hail losses in the U.S. Climatic Change, 96, 161-166.

Changnon, S. A. (2009b). Temporal and spatial distributions of wind storm damages in the United States. Climatic Change, 94, 473-483.

Choi, O., and A. Fisher (2003). The impacts of socioeconomic development and climate change on severe weather catastrophe losses: mid-Atlantic region (MAR) and the U.S. Climatic Change, 58, 149-170.

CRED (2013). EM-DAT: International Disaster Database, Centre for Research on the Epidemiology of Disasters, Université Catholique de Louvain, Belgium.

Crompton, R.P., K.J. McAneney (2008). Normalised Australian insured losses from meteorological hazards: 1967-2006. Environmental Science and Policy, 11, 371-378.

Crompton, R. P., K. J. McAneney, K. Chen, R. A. Pielke Jr., and K. Haynes (2010). Normalised Australian bushfire building damage and fatalities: 1925-2009. Weather, Climate, and Society, 2, 300-310.

Downton, M., J. Z. B. Miller, and R. A. Pielke Jr. (2005). Reanalysis of U.S. National Weather Service flood loss database. Natural Hazards Review, 6, 13-22.

Downton, M.W., R.A. Pielke Jr. (2005). How accurate are disaster loss data? The case of U.S. flood damage. Natural Hazards, 35, 211-228.

Fengqing, J., Z. Cheng, M. Guijin, H. Ruji, and M. Qingxia (2005). Magnification of flood disasters and its relation to regional precipitation and local human activities since the 1980s in Xinxiang, Northwestern China. Natural Hazards, 36, 307-330.

Miller, S., R. Muir-Wood, and A. Boissonnade (2008). An exploration of trends in normalized weather-related catastrophe losses. Climate Extremes and Society, H. F. Diaz and R. J. Murnane, Eds., Cambridge University Press, 225-347.

Mirza, M.M.Q. (2002). Global Warming and Changes in the Probability of Occurrence of Floods in Bangladesh and Implications, Global Environmental Change, Vol. 12, 127-138.

Nakicenovic, N. and R. Swart, (eds) (2000). IPCC Special Report on Emissions Scenarios. Cambridge University Press, Cambridge.

Neumayer, E. and F. Barthel (2011). Normalizing economic loss from natural disasters: a global analysis. Global Environmental Change, 22(1), 13-24.

Nordhaus, W. D. (2010). The economics of hurricanes and implications of global warming. Climate Change Economics, 1, 1-20.

Pielke Jr., R.A., J. Gratz, C.W. Landsea, D. Collins, M. Saunders, and R. Musulin (2008). Normalized hurricane damages in the United States: 1900-2005. Natural Hazards Review, 9, 29-42.

Pielke Jr., R.A., J. Rubiera, C. Landsea, M.L. Fernandez, and R. Klein (2003). Hurricane vulnerability in Latin America and the Caribbean: normalized damage and loss potentials. Natural Hazards Review, 4, 101-114.

Raghavan, S., and S. Rajesh (2003). Trends in tropical cyclone impact: a study in Andhra Pradesh, India. Bulletin of the American Meteorological Society, 84, 635-644.

Sander, J., Eichner, J.F., Faust, E., Steuer,M. (2013). Rising variability in thunderstorm-related U.S. losses as a reflection of changes in large-scale thunderstorm forcing. 5, 317-331.

Schmidt, S., C. Kemfert, and P. Höppe (2009). Tropical cyclone losses in the USA and the impact of climate change: a trend analysis based on data from a new approach to adjusting storm losses. Environmental Impact Assessment Review, 29, 359-369.

Simmons, K.M., Sutter, D., Pielke Jr., R.A. (in press). Normalized tornado damage in the United States: 1950–2011. Environmental Hazards, doi:10.1080/17477891.2012.738642.

Tanner TM, Hassan A, Islam KMN, Conway, D, Mechler R, Ahmed AU, Alam M. (2007). ORCHID: Piloting Climate Risk Screening in DFID Bangladesh. Detailed Research Report. University of Sussex, UK: Institute of Development Studies.

UNESCAP and UNISDR (2012). Reducing Vulnerability and Exposure to Disasters. The Asia-Pacific Disaster Report 2012. UNESCAP and UNISDR, Bangkok.

UNISDR (2011) Revealing risk, redefining development. Global Assessment Report on Disaster Risk Reduction (GAR). Geneva: United Nations International Strategy for Disaster Reduction.

Zhang, Q., L. Wu, and Q. Liu (2009). Tropical cyclone damages in China: 1983-2006. Bulletin of the American Meteorological Society, 90, 489-495.

World Bank (2013). World Development Indicators. World Bank, Washington D.C.

3