EC O L O G IC A L E C O N O M IC S 6 3 ( 2 0 07 ) 27 3 2 84

a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m

w w w. e l s e v i e r. c o m / l o c a t e / e c o l e c o n

Coastal disasters from the perspective of ecological economics

O. Prez-Maqueoa,b,, A. Intralawana , M.L. Martneza,c
a

Gund Institute of Ecological Economics, University of Vermont, 590 Main Street, Burlington, VT 05405-1708, USA Departamento de Ecologa Aplicada, Instituto de Ecologa A.C., km 2.5 antigua carretera a Coatepec no. 351, Congregacin El Haya, Xalapa, Ver. 91070, Mxico c Departamento de Ecologa Funcional, Instituto de Ecologa A.C., km 2.5 antigua carretera a Coatepec no. 351, Congregacin El Haya, Xalapa, Ver. 91070, Mxico
b

AR TIC LE I N FO
Article history: Received 16 February 2006 Received in revised form 10 November 2006 Accepted 10 November 2006 Available online 11 January 2007 Keywords: Coast Hurricanes Ecological economics Complex system

ABS TR ACT
Natural hazards are recurrent events that frequently result in high death tolls and large economic losses. Because of their large impact, they have concerned the international community for a long time. In spite of the efforts, the impact of natural hazards has increased. Oftentimes, the role of natural ecosystems and the ecosystem services they provide to human societies are not considered in risk reduction programs. How relevant are ecosystems? What are the consequences of the depletion of natural ecosystems and the loss of ecosystem services provided by them? Would the alternative vision of ecological economics, in which development and economy are seen as the Whole the ecosystem (Daly, H.E. and Farley, J., 2004. Ecological Economics: principles and applications. Island Press, Washington, DC., 454 pp.), be helpful in reducing disaster risk? In this paper we are set to test whether a holistic approach from the perspective of ecological economics is helpful to clarify and reduce the impact of natural hazards. We focused on hurricanes because they are the most frequently reported events of all natural disasters (Hewitt, K., 1997. Regions of Risk. Longman, Edinburgh Gate, 389 pp.). We analyze the relationship between the components of Human, Built, Social and Natural capitals with the damage caused by hurricanes in terms of mortality rate. We then generate a conceptual model to help envision the complexity of the system. A stepwise (back and forth steps) linear regression analysis revealed that mortality rate was significantly and positively affected by hurricane frequency (P b 0.01) while area covered by semi-altered ecosystems (a mosaic of natural and human-altered ecosystems) (P b 0.01) and GDP (P b 0.05) negatively affected mortality rate (R2 = 0.81). The proportion of natural/altered ecosystems yielding the best protection results needs to be determined yet. Natural capital alone does not decrease number of deaths. Rather, its complex interactions with the other capitals and the many feedback loops that are involved need to be considered to achieve effective disaster risk reduction. No single capital is enough to reduce the impact and intensity of natural hazard. A balance between Human, Built, Social and Natural capitals and an increasing awareness of the consequences of different development decisions, will help human societies to live with rather than cope with coastal hazards. 2006 Elsevier B.V. All rights reserved.

Corresponding author. Departamento de Ecologa Aplicada, Instituto de Ecologa A.C., km 2.5 antigua carretera a Coatepec no. 351, Congregacin El Haya, Xalapa, 5 Ver. 91070, Mxico. E-mail address: octavio.maqueo@inecol.edu.mx (O. Prez-Maqueo). 0921-8009/$ - see front matter 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.ecolecon.2006.11.011

274

EC O LO GIC A L E CO N O M ICS 6 3 ( 2 00 7 ) 2 7 3 2 84

1.

Introduction

Natural hazards such as earthquakes, tropical cyclones, floods and droughts are common features of our dynamic planet, and humans have been exposed to them since their very origins. However, we have not fully learned yet, how to cope with these recurrent natural events. Oftentimes, natural hazards create such economic damage and the loss of so many human lives that they become disasters (UNDP/BCPR, 2004). Globally, natural disasters exert an enormous toll on development with ever-increasing annual economic losses: in the 1960s the average economic loss associated with natural disasters was US$ 75.5 billion; in the 1970s this damage added to US$ 138.4; it almost doubled in the 1980s (US $ 213.9 billion) and by the 1990s economic damage owing to natural disasters had risen to US$ 659.9 billion. Furthermore, in the last two decades, more than 1.5 million people have been killed by natural disasters (UNDP/BCPR, 2004; EM-DAT, 2005). Today, 26.7% of the countries of the world are exposed to earthquakes, 17.6% to tropical cyclones, 65.2% to floods and 12.8% to droughts (UNDP/BCPR, 2004). Of these, earthquakes, tropical storms and flooding caused by storms in coastal areas are the events with the highest death tolls and economic damage (Hewitt, 1997). In spite of their relatively high frequency, the high economic costs as well as the loss of human lives that are involved, disasters are still usually perceived as exceptional natural events that interrupt normal human development and require humanitarian actions to mitigate loss (UNDP/BCPR, 2004). Certainly, they are not exceptional events. Because of their impact on human societies, natural disasters have concerned the international community for a long time. Almost twenty years ago, serious international efforts were aimed at developing policies to better deal with these recurrent events. On December 11, 1987, the United Nations General Assembly passed the resolution 42/169 to designate the 1990s as the decade of natural disaster reduction (http://www. un.org/documents/ga/res/42/a42r169.htm). The Economic and Social council recommended that the General Assembly take actions on developing an appropriate framework toward international cooperation in a risk reduction program. The objective of the International Decade for Natural Disaster Reduction was to reduce the loss of life, property damage and social and economic disruption caused by natural disasters. The impact of natural disasters was not ameliorated after this well-intentioned decade. In fact, after 19 years of growing international awareness, there has nevertheless been a significant increase in loss of life and economic damage due to natural catastrophes (Pielke, 2005). For example, in China the Yangtze River flood in 1998 caused more than 30 billion US dollars in damage, 2000 deaths, and the relocation of 13.8 million people from their destroyed homes (Brown and Halweil, 1998). Hurricane Brendan struck the Chittagong region of Bangladesh in 1991, causing more than 138,000 deaths and 1.5 billion dollars in economic losses (http://www.noaanews.noaa.gov/stories/s334b.htm). Hurricane Mitch swept through Honduras and Nicaragua in late October 1998 with an estimated death toll of more than 10,000 and economic losses of 8.5 billion US dollars (Pielke et al., 2003). And the most recent 2005 US hurricane series, Katrina, Rita, Stan and Wilma, imposed great damage to the US and Central America. Katrina alone caused over a million people

displaced, 1383 deaths and damage from $100 to $200 billions, making it the most expensive natural disaster in US history. Thus, it is obvious that current policies have not been enough to reduce the impact of natural hazards. Why? What is missing?

1.1. Natural coastal disasters from the perspective of ecological economics

What are the appropriate development policies and practices that contribute to the reduction of disaster risks? There are many interesting and potentially useful ideas. Dealing with hazardrisk issues entails the understanding of geographical as well as sociological matters. It covers interdisciplinary interaction pertinent to risk mitigation, preparedness, response, and recovery issues. Traditionally, hazardrisk analysis and risk reduction programs tend to be reactive and deal more with preparedness measures for disaster response, with the common variables of identifying the dangerous locations, and improving warning systems. Frequently, risk reduction of the impact of natural catastrophes focuses on issues like a) the analyses of social processes and underlying causes which may ultimately be quite remote from the disaster event itself (Blaikie, 1994, Alexander, 2000; Holzmann and Jrgensen 2004; b) the economics of catastrophe risk insurance and strategies for financing postdisaster infrastructure rehabilitation (Kunreuther, 1996, 2001; Andersen, 2002; Ermolieva et al., 2003; Freeman and Pflug, 2003; Berz, 2004; Gollier, 2005; c) the role of the Government, private sector, and NGOs for disaster preparedness and risk prevention (Benson et al., 2001; Daniels et al., 2006); or on d) a modeling approach to analyze the interplay between investment in mitigation and risk-sharing measures (Amendola et al., 2000). Occasionally, the role of natural ecosystems and the ecosystem services they provide to human societies are considered in development and risk reduction programs. For instance, the ecosystem service of protection provided by natural ecosystems such as forests, coastal mangroves, coral reefs, riparian habitats has been identified as a priority for hazard mitigation and risk reduction (Hook, 2000; Kreimer and Arnold, 2000; Bronstert, 2003). However, it is recognized that the interaction of wild and managed ecosystems with extreme events and their changed vulnerability as a result of human interference remain largely uninvestigated and unknown (Hook, 2000). Some studies have shown that Natural capital plays an important role on damage reduction. Hiraishi and Harada (2003) used fluid modeling to determine that 30 trees per 100 m2 in a 100-m wide belt would reduce by 90% of the flow pressure on the coastline. Danielsen et al. (2005) also found that areas covered by mangroves as well as tree shelterbelts were significantly less damaged from the Tsunami wave that hit the Cuddalore District, India in 2004. Certainly, risks can be reduced but natural hazards cannot. Risk reduction requires a deeper understanding of cultural, political, social, economic, psychological, technological, physical and natural conditions (McEntire, 2004) because a single perspective is limited. No doubt, the number of interactions and feedbacks between natural and socioeconomic variables demand a systemic vision of the problem. Here we are set to test whether a holistic approach from the perspective of ecological economics is helpful to understand and generate new ideas to reduce the impact of natural hazards. Ecological economics recognizes that human societies are

EC O L O G IC A L E C O N O M IC S 6 3 ( 2 0 07 ) 27 3 2 84

275

Fig. 1 Total number of people killed in storms since 1900 on the coastal areas (EM-DAT, 2005).

complex, adaptive systems, embedded within even more complex, natural ecosystems (Limburg et al., 2002). According to this point of view it is suggested that to understand and better deal with natural hazards it is necessary to include the Natural capital (the means of production provided by nature) (Daly and Farley, 2004) in addition to the capitals produced by humans (Human, Built and Social). Consequently, from the perspective of ecological economics, the protective role of natural ecosystems against natural disasters should be considered. Thus, we first aim at analyzing the relationship between the components of Human, Built, Natural and Social capitals and the damage caused by hurricanes. Secondly, we deal with the problem of complexity. We present a conceptual model that integrates the role of natural,

social and economic systems in response to the impact of hurricanes. We chose to only focus on coastal areas because statistically, tropical storms and flooding caused by storms are the most frequently reported events of all natural disasters (Hewitt, 1997).

2.
2.1.

Methods and results

The hurricanes of our planet

The risk of hurricane landings is unevenly distributed in the world (Hewitt, 1997). Some places are more risk-prone than others. Both developed and developing countries are at risk

Fig. 2 Total economic damages due to storm-related events since 1900 (EM-DAT, 2005).

276

Table 1 Data used in the multiple linear stepwise regression analyses Country Coastal Number Hurricane Life Adult GDP HDI Natural Semi- Croplands population of frequency expectancy literacy index (km2) altered (km2) (thousands) deaths index index (km2)
274 35 155 102 1341 210 7332 2022 82,232 72 426 444 7108 66 18,821 39 57,519 17,668 295 415 1179 20 650 1 1 1 1 1 2 2 4 4 4 4 4 4 5 5 5 5 5 7 8 8 10 11 3 3 2 6 1 1 2 1 1 3 2 4 1 5 1 4 1 1 5 2 5 7 3 0.78 0.91 0.89 0.73 0.83 0.84 0.90 0.82 0.72 0.82 0.88 0.74 0.44 0.77 0.79 0.78 0.89 0.81 0.68 0.90 0.78 0.85 0.75 0.97 0.98 0.99 0.99 0.99 0.97 0.97 0.86 0.94 0.90 0.81 0.63 0.88 0.95 0.94 0.86 0.82 0.90 313 338 221 4704 2601 431 0 0 0 0 315 0 5

Urban and builtup (km2)

401 160

Total (km2)

ESP Inverse Press Modified Liberty freedom equality index index index
0.85 0.95 0.96 0.91 0.85 0.96 0.76 0.95 0.82 0.85 0.87 0.89 0.80 0.88 0.77 0.92 0.83 0.84 0.89 0.87 0.87 0.89 0.97 1.00 0.43 0.57 0.84 0.43 0.43 0.84 0.47 0.68 0.69 0.39

SI
EC O LO GIC A L E CO N O M ICS 6 3 ( 2 00 7 ) 2 7 3 2 84

Barbados Cayman Islands Guam Tonga TrinidadTobago Antilles, Netherlands Canada Bermuda Brazil Dominica Guadeloupe Solomon Islands Tanzania Antigua and Barbuda Sri Lanka St. Kitts and Nevis United Kingdom Venezuela Bahamas Martinique Mauritius Vanuatu Virgin Islands

714.14 502.68 220.7 4848.05 5074.1 431.38

0.65

0 6 0 2817 0 16,790 317 183 5515 18,434 40 24,380 7 93,813 2856 3124 97 0 1826 0

144 2152

0.57 0.57

0.64 0.70

0.60

0.62

0.41 0.78 0.89 0.92 0.97 1.00 0.93 0.96 0.98 0.86 0.98 0.99 0.23

0.89 3,193,058 14,502 0.82 511 0 0.74 436,060 314,210 0.88 738 0 0.89 3784 1 0.57 24,265 337 0.48 0.83 0.74 0.82 0.89 0.78 0.82 0.94 0.82 0.91 0.73 64,618 2511 9031 201 22,244 158,767 90,253 1139 28,243 25,839 210 10,944 0 27,264 0 58,228 38,942 0 2 0 415 0

56,161 13 68,100 62 1151 67 998 277 3806 143 59,802 26,866 815 896 1095 281 0

3,266,537.6 524.16 835,160.26 1117.1 5119.45 30,183.9 94,993.74 2827.75 64,481.08 350.79 234,087.49 227,430.6 94,191.57 2133.8 29,338.07 28,361.19 209.78

0.84 0.55 0.84

0.76 0.51 0.62 0.52

0.52 0.70 0.79 0.59

0.43

0.37

0.66

0.48

0.71 0.29

0.82 0.56

0.63 0.51

0.72 0.45

0.46

0.14

0.16

0.55

0.29

Australia Runion & Eparses China Russian Federat Jamaica Somalia Samoa Grenada Cuba Belize Papua New Guine Costa Rica Fiji Mozambique Myanmar (Burma) Dominican Rep Guatemala El Salvador Mexico Pakistan Madagascar United States Haiti Nicaragua Honduras India Bangladesh Number of observations Mean Standard deviation

16917 721 306,312 21,783 2587 6318 179 89 11,201 242 2944 3798 847 11,538 24,178 8498 7480 6243 28,379 14,197 9585 120,298 7786 3361 4248 264,781 70,314 50 23,065 59,011

14 16 19 21 25 30 31 40 42 44 47 59 60 252 253 385 392 488 507 840 1161 1400 2961 3391 14,639 16,828 140,248 50 3686 19,956

8 3 2 3 7 1 4 3 10 3 2 2 6 3 2 5 3 4 20 2 17 28 5 5 3 27 15 50 5 6

0.90 0.81 0.77 0.68 0.85 0.40 0.75 0.66 0.86 0.82 0.55 0.86 0.75 0.22 0.52 0.70 0.68 0.76 0.80 0.60 0.48 0.88 0.47 0.74 0.68 0.65 0.60 50 0.73 0.15

1.00 0.89 0.86 1.00 0.88 0.38 0.00 0.98 0.97 0.94 0.66 0.96 0.94 0.48 0.83 0.85 0.71 0.80 0.92 0.46 0.69 0.97 0.53 0.68 0.76 0.60 0.43 48 0.83 0.21

0.80

0.90 1,548,979 0.85 1265

202 0

204,108 0 266,057 78,767 557 546 0 58 43,923 5904 5266 17,119 0 14,666 61,047 8582 14,180 5254 110,253 6164 5695 100,850 11,269 13,839 17,462 255,824 36,739 48 30,299 62,058

37,013 1032 75,119 38,371 2998 192 0 119 7477 664 1221 5059 1722 1029 2797 5588 3161 3404 52,515 3485 217 250,009 490 2207 2655 51,829 5311 48 16,231 40,242

1,790,302 2297 566,874.82 2,971,644 12,167.41 268,858.4 1081.52 1089.55 113,634 24,710.37 325,020.4 51,612.3 12,186.14 253,150.83 195,562.32 48,407.05 44,605.27 19,666.7 729,409.3 89,942.7 364,133.3 1,777,174

0.87 0.86 0.72 0.83 0.84 0.72 0.93 0.88 0.86 0.91 0.88 0.84 0.95 0.78 0.84 0.86 0.82 0.79 0.82 0.75 0.82 0.79

1.00

0.90

0.65

0.85

0.53 0.64 0.52

0.72 96,786 128,913 0.77 2,780,678 73,828 0.75 0.39 0.38 0.76 0.91 0.79 0.56 0.82 0.75 0.33 0.68 0.71 8612 250,047 1082 913 62,234 17,953 315,534 28,214 10,464 209,742 129,828 34,237 0 18,073 0 0 0 189 2999 1220 0 27,714 1890 0 178 63 25,957 120 45,281 99,475 0 132 1112 81,637 2836 50 19,540 51,128

0.14 0.29 0.57 0.00 0.71 0.00 0.71 0.57 0.71 0.57 0.43 0.00 0.71 0.43 0.57 0.57 0.29 0.43 1.00 0.14 0.57 0.57 0.57 0.43 36 0.49 0.26

0.20 0.40 0.83 0.12 0.77 0.84 0.04 0.76 0.74 0.83 0.67 0.52 0.04 0.70 0.51 0.65 0.60 0.43 0.69 0.84 0.28 0.68 0.57 0.58 0.37 38 0.58 0.24

0.60 0.51 0.62

0.31 0.40 0.67 0.77 0.78

0.63 0.59 0.46 0.64 0.56 0.31

0.49 0.54 0.60

0.60 0.70
EC O L O G IC A L E C O N O M IC S 6 3 ( 2 0 07 ) 27 3 2 84

0.52 0.02 0.65

0.59 0.53 0.55 0.65 0.43 0.30 0.84

0.53

0.64 27,086 0.70 10,946 0.79 540,684 0.49 80,174 0.49 312,940 0.90 1,326,840 15,314 61,921 47,195 171,122 26,543 50 243,748 639,193

0.48 0.47 0.69 0.62 0.59

0.57 0.55 0.47 0.58 0.81

0.39 0.46 0.46 0.62 0.46 0.64 0.46 0.57 0.39 0.47 32 50 0.55 0.15 0.73 0.17

27,072.9 0.89 78,099 0.86 68,423.8 0.85 560,412 0.78 71,428.82 0.88 50 50 307,956 693,668 0.86 0.06

0.40 0.44 0.62 0.66 25 0.57 0.09

0.55 0.53 0.59 0.49 29 0.58 0.18

Descriptive statistics for all used variables are shown at the bottom.

277

278

EC O LO GIC A L E CO N O M ICS 6 3 ( 2 00 7 ) 2 7 3 2 84

Fig. 3 Human development index based on life expectancy, education and standard of living among nations (Human Development Report, 2004).

(Figs. 1 and 2). According to the EM-DAT (2005) database, the number of people killed by hurricanes per country since 1900 shows that Central America and IndiaBangladesh are the most vulnerable regions on Earth (Fig. 1). Bangladesh alone accounts for more than 60% of the registered deaths between 1980 and 2000. On the other hand, total accumulated economic damage illustrates that the areas most affected by hurricanes are North America and India (Fig. 2). Some regions such as Europe and Africa (except for Southeastern Africa) have never been hit by a

hurricane. South America had never been hit until Brazil was hit by an un-named hurricane in 2004 (http://science.nasa.gov/ headlines/y2004/02apr_hurricane.htm). Although the category of this hurricane was low, damage was relatively large.

2.2.

Human and Built capital

Based on earlier analyses by Vemuri and Costanza (2006) we represented Human and Built capital together because of the

Fig. 4 Natural capital of the coastal ecosystems of the world. The calculations were based on the database from the World Resources Institute (Burke et al. 2001), several World atlases: World Atlas of Coral Reefs, World Atlas of Sea Grasses; World Atlas of Mangroves and FAO (Reaka-Kudla, 1997; Spalding et al., 1997, 2001; FAO, unpublished; Green and Short, 2003), and unit ecosystem-service values previously calculated (Costanza et al. 1997; Sutton and Costanza, 2002).

EC O L O G IC A L E C O N O M IC S 6 3 ( 2 0 07 ) 27 3 2 84

279

Fig. 5 Social Index based on national data on civil liberties, press freedom and income equality (Freedom House, 1999; World Resources Institute, 2003).

high correlation found by these authors when they tested a wide array of Human and Built capital variables. This information can be obtained from the UN's Human Development Index (HDI) and from the Human Development Report (2004). HDI is an indirect measure of human development and is derived from data on comparative calculation based on life expectancy, education and standard of living among nations (Table 1). The life expectancy index represents the longevity of the population and general health-related conditions. Education was evaluated based on adult literacy, which was considered in terms of the ability to read and write at an adult age. Standard of living index was calculated based on Gross Domestic Product (GDP). All these variables were standardized so that their values fell between 0 and 1. Finally, the three indices (life expectancy, education and standard of living) were averaged to calculate the HDI. A more detailed explanation of these calculations can be read in the technical notes of the Human Development Report (2004). World distribution of Human Development Index is shown in Fig. 3. From there it is easy to recognize that Africa and Southeastern Asia show the lower HDI values, while Europe and North America have the highest. HDI in Latin America is moderate to high (Mexico, Argentina and Colombia).

2.3.

Natural capital

Because the study focuses on the coastal regions, and the coasts are comprised by terrestrial and aquatic ecosystems, we considered these two types of ecosystems. Coastal terrestrial ecosystems were defined as those located within 100 km of coastline (Martnez et al., 2007). We based our calculations on the database from the World Resources Institute (Burke et al., 2001). These authors used the 1 km resolution Global Land Cover Characteristics Database (GLCCD, 1998) which was derived from the Advanced Very High-Resolution Radiometer (AVHRR) satel-

lite for the 19921993 period. Fifteen different land cover classes, including natural, semi-altered (mosaic of natural vegetation and cropland) and fully altered ecosystems (Burke et al., 2001), were distinguishable and served to provide the basis for calculating the total area covered by each one within the 100 km limit. The area covered by aquatic ecosystems such as mangroves, coral reefs and sea grasses was obtained from different databases: World Atlas of Coral Reefs, World Atlas of Sea Grasses; World Atlas of Mangroves and FAO (Reaka-Kudla, 1997; Spalding et al., 1997, 2001; FAO, unpublished; Green and Short, 2003). Once the total area covered by all the terrestrial and aquatic ecosystems was secured (Table 1), the Ecosystem Services Product (ESP) was calculated, based on unit ecosystem-service values previously calculated by Costanza et al. (1997) and Sutton and Costanza (2002). From this calculation, it was possible to obtain the ESP per square kilometer per ecosystem per country. The addition of ESP values per ecosystem per country yielded total dollar value of ecosystem services per country. The next step was to normalize ESP values by calculating the log of ESP values per square kilometer (Vemuri and Costanza, 2006), and finally, standardize these values as an index between 0 and 1. The Natural capital of the coastal ecosystems (adding both terrestrial and aquatic (Ecosystem Service Product (ESP) is shown in Fig. 4. The highest ESP values were calculated for Canada the US, Mexico, Brazil, the Russian Federation; Southeastern Asia and Australia. Lower ESP values elsewhere are due to intense degradation of coastal ecosystems (i.e. Europe) and also because the ecosystem-service values calculated for the natural ecosystems in these countries was not high in the first place (Costanza et al., 1997).

2.4.

Social capital

A proxy for Social capital was calculated based on national data on civil liberties, press freedom and income equality (Table 1). These are arguably good Social capital proxies

280

EC O LO GIC A L E CO N O M ICS 6 3 ( 2 00 7 ) 2 7 3 2 84

Table 2 Results of the multiple linear stepwise regression Coefficients

(Intercept) Ln frequency Ln semi-altered Ln GDP Ln Liberty index

according to UNDP/BCPR (2004) so that values ranged between and + , as follows: ViV Vi =1Vi where Vi is the transformed variable (ranging from to + )

Value
8.9521 1.4308 0.9297 1.9042 1.2964

Std. error
0.9106 0.3894 0.2268 0.7863 0.6690

t value
9.8305 3.6748 4.1000 2.4218 1.9378

Pr (N t)
0.0000 0.0028 0.0013 0.0308 0.0747

The following variables did not significantly add to the ability of the equation to predict Ln r and were not included in the final equation: Ln natural, Ln croplands, Ln urban, Ln life expectancy index, Ln adult literacy, Ln press freedom and Ln equality index. Normality test: passed (P = 0.439), constant variance test: passed (P = 0.317).

Vi

is the original variable (ranging from 0 to 1).

because they represent the laws, political and economic factors as well as social equity. The level of civil liberties is rated on a 17 scale, with 1 representing the most free and 7 the least free. A rating of 1 indicates a country with an established and equitable rule of law. In contrast, countries with a rating of 67 experiences severely restricted dictatorship, political terror (political prisoners) and prevention from free association of peoples. Press freedom is an index developed by Freedom House as the degree to which each country permits the free flow of information, using a scale from 1 to 100 (Freedom House, 1999). A country with a press freedom equal to 1 has total freedom while a score of 100 indicates total lack of freedom. Income equality was measured through the Gini Index. A score of 0 implies perfect equality while a score of 100 implies perfect inequality (World Resources Institute, 2003). Transformation of the values of these three indices made the score match the direction of positive results of the other variables. Thus, the greater the civil rights, freedom of press and equality, the larger the value. Additionally, every value was standardized on a 01 scale to match the indices of the other forms of capitals. The Social capital index was the average between the three proxies used. When Gini Index was not available, the Social Index was calculated as the mean value of liberty rights and press freedom alone (Fig. 5). Social Index was highest in the US, Canada, Europe and Australia; medium in South America, South Africa, Southeastern Asia and the Russian Federation, and lowest in Africa.

All variables were transformed to a natural logarithm scale to standardize the data before statistical analyses. Given the normal distribution of the dependent variable we used ordinary least squares (OLS) to fit a multiple linear regression. A stepwise (back and forth steps) linear regression was performed in which Ln(r) was the dependent variable and natural logarithms of the components of the capitals were the explanatory (independent) variables. Pearson correlation coefficients were calculated in order to detect autocorrelation between variables in the final model. We did not use economic damage as a response variable because it is not equally distributed among the countries. High economic losses in developed regions (e.g. Europe, USA and Canada) reflect the high value of infrastructure. In less developed regions, in turn, decreased economic damage does not mean a reduced impact on infrastructure but rather a lack of infrastructure to be damaged. In these cases, even a small economic loss may represent critical damage. Economic damage was therefore not an adequate variable to perform our analyses with. The stepwise linear regression results revealed that mortality rate was significantly and positively affected by hurricane frequency (P b 0.01), while area covered by semialtered ecosystems (P b 0.01) and GDP (P b 0.05) negatively affected mortality rate (R2 = 0.81) (Table 2). Liberty index positively affected mortality rate (Table 2), but was not a

2.5.

Multiple regression analysis

The total number of deaths in countries where hurricanes produced at least 1 death from 1990 to 2004 (N = 50) were obtained from EM-DAT (2005). A mortality rate per country (r) was calculated according to Dilley et al. (2005) as follows: r M=P where M is the total number of fatalities produced by hurricanes from 1990 to 2004, in each country and P is the estimated total population per country that lived on the coast in 2000 (World Resources Institute, 2003). Results were expressed as number of deaths per thousand people. The components of Built, Social, Human and Natural capitals (life expectancy, adult literacy, GDP, land use cover, liberty, press freedom and equality) were used as explanatory variables. They were expressed as percentages, and latter transformed

Table 3 Pearson Correlation coefficients and significance of bivariate analysis of mortality rate (r) caused by hurricanes which is affected by the natural logarithms of hurricane frequency (Freq), percentage of semi-altered land use type, Gross Domestic Product (GDP) and Liberty index Ln frequency
Coefficient P value N Ln Coefficient frequency P value N Ln semiCoefficient altered P value N Ln GDP Coefficient P value N Ln(rh) 0.503 0.000198 50

Ln semialtered
0.366 0.0468 30 0.124 0.513 30

Ln GDP

Ln Liberty index

0.357 0.175 0.0451 0.356 32 30 0.0133 0.0802 0.942 0.674 32 30 0.176 0.131 0.400 0.560 25 22 0.415 0.0280 28

EC O L O G IC A L E C O N O M IC S 6 3 ( 2 0 07 ) 27 3 2 84

281

Table 4 Relationship of damage caused by hurricanes with different explanatory variables Type of event Region Dependent variabled
Ln of number of deaths (K)

Explanatory variables
Ln physical exposure (PhExp) Ln percentage of arable land (Pal) Ln human development index (HDI) Ln wind speed (g) Ln area of wetlands (w) Ln frequency (Freq)

Effect on dependent variable

Positive Positive Negative

Model

Reference

Tropical storm, Global cyclones, hurricanes and super typhoons

Ln(K) = 15.86 + 0.63Ln(PhExp) + 0.66Ln(Pal) 0.03Ln(HDI) R2 = 0.863

UNDP/ BCPR (2004)

Cyclones, hurricanes, tropical storms Cyclones, hurricanes, tropical storms

United States

Ln of economic damage/GDP (TD/GDP) Ln number of deaths/pop (r)

Positive Negative Positive

Ln/TD/GDP) = 10.551 + 3.878Ln(g) 0.77Ln(w) R2 = 0.60 Ln(r) = 8.95 + 1.43Ln(Freq) 0.92Ln (semi-altered) 1.90Ln(GDP) + 1.29Ln (Liberty index) R2 = 0.813

Costanza et al. (in revision) This study

Global

Ln semi-altered land use type (semialtered) Ln Gross Domestic Product (GDP) Ln Liberty index

Negative

Negative Positive

statistically significant variable (P = 0.07) in the model. Also, it was correlated with GDP (Table 3). Thus, its effect on mortality rate should be treated with caution. However, it is interesting

to notice that, when included in the model, 81% of the variance was explained and when eliminated this value decreased to 63%. The remaining potentially explicatory

Fig. 6 Conceptual model that explains the multiple relationships and feedbacks between damage to ecosystems, economic damage, number of deaths and hurricane impact. Arrows represent causal relationships and positive and negative signs the direction of the relationships. Reinforcing (R) and balancing (B) loops are also indicated.

282

EC O LO GIC A L E CO N O M ICS 6 3 ( 2 00 7 ) 2 7 3 2 84

variables did not yield any significant correlations and were excluded from the results. The multiple regression model was: Lnr 8:95 1:43LnFreq0:921Lnsemialtered 1:901LnGDP 1:29LnLiberty index

3.
3.1.

Discussion
Do ecosystems play a protective role against hurricanes?

There is empirical evidence showing that Natural capital plays an important role in storm protection (Badola and Hussain, 2005; Danielsen et al., 2005; Costanza et al., in revision). Several models have been elaborated to explain this ecosystem service in terms of different explanatory variables (Table 4). The calculations performed by UNDP/BCPR (2004) revealed that physical exposure and percentage of arable land (degraded Natural capital) are positively related with number of deaths, while HDI showed a negative effect (Table 4). In turn, Costanza et al. (in revision) found that coastal vegetation (preserved herbaceous wetlands) reduced significantly the economic damage caused by hurricanes and cyclones. In our regression model, mortality rate was negatively correlated with the cover of semi-altered ecosystems and GDP (Table 4). From these studies it becomes evident that the assessment of the protective role of ecosystems depends on the response variable used for the analysis. Apparently, Natural capital may reduce the economic damage caused by hurricanes while the level of degradation of natural ecosystems plays an important role in number of deaths. Thus, Natural capital alone does not decrease number of deaths. In fact, economic conditions within the country (represented by HDI or GDP) have an important effect on the number of fatalities caused by hurricanes (UNDP/BCPR, 2004). Our results suggest that a combination of infrastructure and relatively well preserved natural ecosystems (semi-altered ecosystems) seem to offer a good protection service against the impact of hurricanes in terms of human lives. Probably, semi-altered ecosystems are autocorrelated with other variables, such as, effective transportation systems, preparedness and effective evacuation programs (UNDP/BCPR, 2004). Based on the above, the most effective ratio of natural/developed land use and their spatial distribution need to be evaluated so that coastal protection services are improved. Current schemes of coastal development can no doubt be improved if the protective role of natural ecosystems and their interrelationship with Human, Social and Built capitals are considered. These schemes often assume that with development (Built capital) societies are better able to cope with natural disasters. Infrastructure, better communication systems, better health and education and better social organization (i.e. adequate evacuation programs) help people to cope with disasters. If Natural capital is added to these schemes a complex but more realistic picture can be elaborated. To achieve this, a conceptual model that helps to interpret the multiple relationships between the four capitals and the impact of hurricanes from a systemic point of view was generated. The model shows that with economic development (represented in Fig. 6 as GDP) vulnerability associated to mortality rate is decreased. However, the same urban and

industrial development that increases GDP could also have associated negative effects. For instance, urban and industrial development on the coast inevitably will expose more infrastructures and economic damage may be exacerbated. Likewise, disaster frequency of major hurricanes (categories 4 and 5) may be altered by human activities (Webster et al., 2005). The multiple interaction of our conceptual model has several feedback loops: Built capital favors (invites) migration to the coast, which in consequence increases Built capital. At the same time, migration to the coast increases population exposure to hurricanes, and thus, mortality rate. In extreme situations, migration to the coast should, in theory, decrease as the death toll increases. Nevertheless, human demographic trends do not always obey to logic but to economic and historical reasons (Burke et al., 2001). A consequence of the expansion of urban and industrial development at the coast and increased hurricane damage will likely damage natural ecosystems. Species evolutionarily adapted to these natural hazards (Lodge and McDowell, 1991; Tanner et al., 1991; Walker et al., 1992;) may be lost, possibly resulting in the depletion of ecosystem services of coastal ecosystems (gas regulation, climate regulation, disturbance regulation, water regulation, water supply, erosion control, soil formation, nutrient cycling, waste treatment, pollination, biological control, habitat, food production, raw material, genetic, recreation, cultural) (Costanza et al. 1997; Martnez et al., 2007), and the increased exposure to disasters (Abramovitz, 2001). Ideally, according to this and complementary studies (UNDP/BCPR, 2004, Costanza et al. in revision) a combination of natural and modified ecosystems, adequate built infrastructure and social awareness should yield the best results in living with hurricanes (Small et al., 2000, Adger et al., 2005, Baker and Refsgaard, 2007; Gaddis et al., 2007). Earlier studies (UNDP/BCPR, 2004) revealed that countries with low HDI and large rural populations (and greater Natural capital) tend to be more vulnerable to high winds, landslides and flooding because preparedness is low. In this sense, although Natural capital may reduce hurricane impact, for instance reduced wind speed (Danielsen et al., 2005), it is not sufficient to prevent the loss of human lives on its own. Additional capitals (such as Social and Built) are also necessary as well. For instance, countries like Cuba, with low HDI, high Natural capital and a large population exposed to hurricanes, report reduced numbers of deaths (UNDP/BCPR, 2004). Probably, their high Social capital is the key in dealing with hurricanes suggesting social paths (other than HDI) that may also reduce number of deaths. Countries with similar patterns of natural hazards have varying levels of damage, depending on development decisions made in the past and on the status of their four capitals. Because these capitals are very different in each country, an ideal perspective would be for each one to find its own balance among the four capitals with the ultimate goal of increasing standard of living and lowering risk and damage owing to hurricanes. An example of multiple factors determining hurricane impact was hurricane Mitch, which occurred in 1998 and largely affected Honduras and Nicaragua in terms of economic damage and losses of life. Much of the impact of this hurricane was due to the large number of floods and

EC O L O G IC A L E C O N O M IC S 6 3 ( 2 0 07 ) 27 3 2 84

283

landslides triggered by the hurricane, but not to the wind force per se (UNDP/BCPR, 2004). These secondary effects were aggravated by environmental degradation (deforestation). Furthermore, local populations were highly vulnerable in terms of social and economic aspects (weakness in early warming and disaster preparedness). Another and more recent example was hurricane Katrina, by far the deadliest and costliest hurricane in recorded US history although not the most powerful (it has been downgraded to category 3) (http://www.washingtonpost.com/wp-dyn/content/article/ 2005/12/21/AR2005122101960.html?referrer = email&referrer = email). In this case, environmental degradation, inefficient evacuation programs and probably poor performance of the levees resulted in large economic damage and the loss of many human lives (Whoriskey and Warrick, 2005). Again, this shows that one capital alone (Built in this case) was not enough to properly cope with hurricanes.

4.

Conclusion

It is troubling that disaster impacts have been increasing as global economic growth has. This suggests that economic growth has not been properly directed at mitigating disaster risks. The problem is that disasters are still usually (and wrongly) perceived as exceptional natural events. Our results revealed that a semi-altered landscape and GDP significantly reduce mortality rate. However, it remains to be determined the ratio of natural/transformed ecosystems that yield the best protection service. Balance between Human, Built, Social and Natural capitals and an increasing awareness of the consequences of different development decisions, will help human societies to live with rather than cope with coastal disasters.

Acknowledgments
MLM and OPM are grateful to Instituto de Ecologa, A.C. for financial support during their research stay at the Gund Institute of Ecological Economics (Vermont University). We are grateful to Robert Costanza, Joshua Farley, and the students in the Advanced Course on Ecological Economics (Vermont University) for their helpful comments on earlier versions of this text. Their insights largely improved the paper. Thanks to Valeria Prez-Martnez for her infinite patience and courage.

REFERENCES
Abramovitz, J.N., 2001. Unnatural Disasters. Worldwatch Institute, Washington, DC. Adger, W.N., Hughes, T.P., Folke, C., Carpenter, S.R., Rockstrom, J., 2005. Social-ecological resilience to coastal disaster. Science 309, 10361039. Alexander, D., 2000. Confronting Catastrophe: New Perspectives on Natural Disasters. Oxford University Press, USA. (288 pp). Amendola, A., Ermoliev, Y., Ermolieva, T.Y., Gitis, V., Koff, G., Linnerooth-Bayer, J., 2000. A systems approach to modeling catastrophic risk and insurability. Nat. Hazards 21, 381393. Andersen, T.J., 2002. Innovative financial instruments for natural disaster risk management. Inter-American Development Bank

Washington, D. C. Sustainable Development Department Technical Papers Series (42 pp). Badola, R., Hussain, S.A., 2005. Valuing ecosystem functions: an empirical study on the storm protection function of Bhitarkanika mangrove ecosystem. India. Env. Cons. 32, 8592. Baker, D., Refsgaard, K., 2007. Institutional development and scale matching in disaster response. Ecol. Econ. 63, 331343. Benson, C., Twigg, J., Myers, M., 2001. NGO initiatives in risk reduction: an overview. Disasters 25, 199215. Berz, G.A., 2004. Catastrophes and climate change: concerns and possible countermeasures of the insurance industry. Mitig. Adapt. Strategies Glob. Chang. 4, 283293. Blaikie, P., 1994. At Risk: Natural Hazards, People's Vulnerability and Disasters. Routledge, UK. 320 pp. Bronstert, A., 2003. Floods and climate change: interactions and impacts. Risk Anal. 23 (3), 545558. Brown, L.R., Halweil, B., 1998. The Yangtze Flood: The Human Hand, Local and Global. (http://www.worldwatch.org/press/ news/1998/08/13/). Burke, L., Kura, Y., Kasem, K., Revenga, C., Spalding, M., McAllister, D., 2001. Coastal Ecosystems. World Resources Institute, Washington, DC. (93 pp). Costanza, R., d'Arge, R., de Groot, R., Farber, S., Grasso, M., Hannon, B., Naeem, S., Limburg, K., Paruelo, J., O'Neill, R.V., Raskin, R., Sutton, P., ven den Belt, M., 1997. The value of the world's ecosystem services and natural capital. Nature 387, 253260. Costanza, R., Prez-Maqueo, O.M., Martnez, M.L., Sutton, P., Anderson, S.J., Mulder, K., in revision. The value of coastal wetlands for hurricane protection. Frontiers in Ecology and the Environment. Daly, H.E., Farley, J., 2004. Ecological Economics: Principles and Applications. Island Press, Washington, DC. (454 pp). National Symposium on Risk and Disasters, 2006. In: Daniels, R.J., Kettl, D.F., Kunreuther, H. (Eds.), On Risk and Disaster: Lessons from Hurricane Katrina. University of Pennsylvania Press (293 pp). Danielsen, F., Sorensen, M.K., Olwig, M.F., Selvam, V., Parish, F., Burgess, N.D., Hiraishi, T., Karunagaran, V.M., Rasmussen, M.S., Hansen, L.B., Quarto, A., Suryadiputra, N., 2005. The Asian tsunami: a protective role for coastal vegetation. Science 310, 643. Dilley, M., Chen, R.S., Deichmann, U., Lerner-Lam, A.L., Arnold, M., with Agwe, J., Buys, P., Kjekstad, O., Lyon, B., Yetman, G., 2005. Natural disaster hotspots: a global risk analysis. International Bank for Reconstruction and Development/The World Bank and Columbia University,. Washington, DC. (24 pp). EM-DAT, 2005. The OFDA/CRED International Disaster Database. Universit Catholique de Louvain, Brussels. (www.cred.be/emdat). Ermolieva, T., Ermoliev, Y., Fischer, G., Galambos, I., 2003. The role of financial instruments in integrated catastrophic flood management. Multinatl Financ. J. 7 (3 and 4), 207230. FAO, unpublished. Status and trends in mangrove area extent worldwide. By Wilkie, M.L. and Fortuna, S. Forest Resources Assessment Working Paper vol. 63. Forest Resources Division. FAO, Rome. (http://www.fao.org/docrep/007/j1533e/j1533e00. htm). Freedom House, 1999. News of the Century: Press Freedom 1999. Freedom House. (http://freedomhouse.org/pfs99/pfs99.pdf). Freeman, P.K., Pflug, G.C., 2003. Infrastructure in developing and transition countries: risk and protection. Risk Anal. 23, 601610. Gaddis, E.B., Miles, B., Morse, S., Lewis, D., 2007. Full cost accounting of coastal disasters in the United States: implications for planning and preparedness. Ecol. Econ. 63, 307318. GLCCD (Global Land Cover Characteristics Database), Version 1.2. 1998. Loveland, T.R., Reed, B.C., Brown, J.F., Ohlen, D.O. Zhu, Z., Yang, L., Merchant, J., 2000. Development of a Global Land Cover Characteristics Database and IGBP DISCover from 1-km AVHRR data. Int. J. Remote Sensing 21: 13031330. Data available on-line at: http://edcdaac.usgs.gov/glcc/glcc.html.

284

EC O LO GIC A L E CO N O M ICS 6 3 ( 2 00 7 ) 2 7 3 2 84

World atlas of sea grasses. In: Green, E., Short, F.T. (Eds.), UNEPWCMC. University of California Press, Berkeley (298 pp). Gollier, C., 2005. Some aspects of the economics of catastrophe risk insurance. CESIFO Working Paper, vol. 1409 (26 pp). Hewitt, K., 1997. Regions of Risk. Longman, Edinburgh Gate. (389 pp). Hiraishi, T., Harada, K., 2003. Greenbelt tsunami prevention in south-pacific region. Rep. Port Airpt. Res. Inst. 42, 123. Holzmann, R., Jrgensen, S., 2004. Social risk management: a new conceptual framework for social protection, and beyond. Int. Tax Public Financ. 8, 529556. Hook, W.H., 2000. U.S. participation in international decade for natural disaster reduction. Nat. Hazards Rev. 1, 29. Human Development Report, 2004. United Nations Development Program. (http://www.undp.org.in/hdr2004/Report_hdr2004/). Kreimer, A., Arnold, M., 2000. World Bank's role in reducing impacts of disasters. Nat. Hazards Rev. 1, 3742. Kunreuther, H., 1996. Mitigating disaster losses through insurance. J. Risk Uncertain. 12, 171187. Kunreuther, H., 2001. Mitigation and Financial Risk Management for Natural Hazards, Center for Risk Management and Decision Processes, The Wharton School. University of Pennsylvania, Philadelphia. (29 pp). Limburg, K.E., O'Neill, R.V., Costanza, R., Farber, S., 2002. Complex systems and valuation. Ecol. Econ. 41, 409420. Lodge, J., McDowell, W.H., 1991. Summary of ecosystem-level effects of Caribbean hurricanes. Biotropica 23, 373378. Martnez, M.I., Intralawan, A., Vzquez, G., Prez-Maqueo, O., Sutton, P., Landgrave, R., 2007. The ecological, economic and social importance of the coasts. Ecol. Econ. 63, 254272. McEntire, D., 2004. The Status of Emergency Management Theory: Issues, Barriers, and Recommendations for Improved Scholarship, Paper presented at the FEMA Higher Education Conference. (June 8, 2004, Emmitsberg, Maryland. 25 pp). Pielke Jr., R.A., 2005. Making Sense of Trends in Disaster Losses. Bridgesthe OST's Publication on Science & Technology Policy, vol. 7. 3 pp. Pielke Jr., R.A., Rubiera, J., Landsea, C., Fernandez, M.L., Klein, R., 2003. Hurricane vulnerability in Latin America and the Caribbean: normalized damage and loss potentials. Nat. Hazards Rev. 4, 101104.

Reaka-Kudla, M.L., 1997. The global biodiversity of coral reefs: a comparison with rain forests. In: Reaka-Kudla, M.L., Wilson, D.F., Wilson, E.O. (Eds.), Biodiversity II: Understanding and Protecting our Biological Resources. Joseph Henry Press, Washington D.C., pp. 83108. Small, C., Gornitz, V., Cohen, J., 2000. Coastal hazards and the global distribution of human population. Environ. Geosci. 7, 312. World mangrove atlas. In: Spalding, M.D., Blasco, F., Field, C.D. (Eds.), The International Society for Mangrove Ecosystems. Okinawa (178 pp). Spalding, M.D., Ravilious, C., Green, E.P., 2001. World atlas of coral reefs. UNEPWCMC. University of California Press, USA (424 pp). Sutton, P.C., Costanza, R., 2002. Global estimates of market and non-market values derived from nighttime satellite imagery, land cover, and ecosystem service valuation. Ecol. Econ. 41, 509527. Tanner, E.V.J., Kapos, V., Healey, J.R., 1991. Hurricane effects on forest ecosystems in the Caribbean. Biotropica 23, 513521. UNDP/BCPR, 2004. Reducing disaster risk. A Challenge for Development, UNDP/Bureau for Crisis Prevention and Recovery, New York (146 pp). Vemuri, A.W., Costanza, R., 2006. The role of human, social, built, and natural capital in explaining life satisfaction at the country level: toward a national well-being index (NWI). Ecol. Econ. 58, 119133. Walker, L.R., Votzow, J., Acherman, J.D., Fernandez, D.S., Fetcher, N., 1992. Immediate impact of Hurricane Hugo on a Puerto Rican rain forest. Ecology 73, 691694. Webster, P., Holland, J., Curry, G.J., Chang, H.R., 2005. Changes in tropical cyclone number, duration, and intensity in a warming environment. Science 309, 18441846. Whoriskey, P., Warrick, J., 2005. Report Revises Katrina's Force: Hurricane Center Downgrades Storm to Category 3 Strength. Washington Post. (Available at: http://www. washingtonpost.com/wp-dyn/content/article/2005/12/21/ AR2005122101960.html?referrer = email&referrer = email). World Resources Institute, 2003. World resources 20022004. Decisions for the Earth: balance, voice and power. United Nations Development Programme, United Nations Environmental Programme, The World Bank. World Resources Institute, USA (315 pp).