m 4 *' 1~' '1~ * 1~.~ WORLD BANK GROUP GFDRR Glkbn FKi ity for D'snste, Reductiun -nd Peco-ey 02016 The World Bank The International Bank for Reconstruction and Development The World Bank Group 1818 H Street, NW Washington, D.C. 20433, USA May 2016 DISCLAIMER This document is the product of work performed by GFDRR staff, based on information provided by GF DRRs partners. The findings, analysis and conclusions expressed in this document do not necessarily reflect the views of any individual partner organization of GFDRR, including, for example, the World Bank, the Executive Directors of the World Bank, UNDP, the European Union, or the governments they represent. Although GFDRR makes reasonable efforts to ensure all the information presented in this document is correct, its accuracy and integrity cannot be guaran- teed. Use of any data or information from this document is at the user5 own risk and u nder no circumstances shall GFDRR or any of its partners be liable for any loss, damage, liability or expense incurred or suffered which is claimed to result from reliance on the data contained in this document. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denomination, and other information shown in any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. RIGHTS AND PERMISSIONS The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. Table of Contents 9 Azerbaijan 107 Slovenia 13 Belarus 111 Tajikistan 17 Bosnia and Herzegovina 115 Turkey 21 Bulgaria 119 Turkmenistan vi Figure 1. A chronology of significant disasters affecting the ECA region. 25 Croatia 123 Ukraine 29 Cyprus 127 Uzbekistan vii Box 1. Return Periods 33 Czech Republic ix Table 1. Annual Average Affected Population and 37 Estonia GDP for Floods and Earthquakes. 41 Georgia 133 Table 2. Climate Models Used for Flood Risk x Figure 2. Annual Average GDP Affected by Floods. 45 Greece Estimates. 49 Hungary x Figure 3. Annual Average GDP Affected by Earth- 133 Table 3. RCP and S5P Scenario Combinations quakes. 53 Kazakhstan Used to Estimate Future Flood Risk. 57 Kosovo xi Figure 4. GDP Affected by a 100-year Flood. 61 Kyrgyz Republic xi Figure 5. GDP Affected by a 1o-year Flood Relative 65 Latvia to the Country's GDP. 69 Lithuania xii Figure 6. GDP Affected by a 250-year Earthquake. 73 Macedonia, FYR 77 Moldova xii Figure7. GDP Affected by a 250-year Earthquake Relative to the Country's GDP. 81 Montenegro 85 Poland 89 Romania 93 Russian Federation 1 Albania 99 Serbia 5 Armenia 103 Slovak Republic ACKNOWLEDGEMENTS Wc u'wuld likect akoleg tl emsad1nCiuasta mere nyvåmvd lI ic PrDInum omhi påu mmåston Fise w l like to ,ý)ank GJ,DI j" , lep c ialEy FjanileCiS- G l CesquIicrc, AlajlìI na SM>son, Sänmne Hallg Brnden pirngnma, Kesko Saom,and An1W1,1 uea Wc ývudalolke -,o tIlaik thie WVorld Ba,11s E,CA DIM Tcanr: David Sijslcil, EfAyhIan, Vkta Bget,Tl uulearnd Ko lakuuUhi. WVe thank Hesseý,l W,insenmjus fron Delt,ares, andl Philip WVard from »,he VUIves: nsedn lor the 1lood risk, modeling. WVe alsýo thank Jajnies Daniell an d Anid]reas SCliafer for1 the carj quk rilnk 1n ddeling. WVe thnk lý ManC Jålråt må Fernalnda Sernìa MNIr 0=1, dior Csntutiolns Läsa FunTa ParMccl [vii ýicr edikorral gudjJ,A.iMp or t11C (La.a Viual liol and ' t 112coreteau, [o jni . Ko,liCk Mm mnn and An~egil Tånsen fKr pieprång Ise pubåicka-oc, iv  Introduction T he Europe and Central Asia' (ECA) region is affected regularly by natural hazards, such as floods, earth- G reece Croatia quakes, droughts, landslides, and wildfires. Although Crete earthquake 365, Alexandria destroyed. Zagreb earthquake 1880, The Great Zagreb Earthquake." catastrophic events are not as frequent as in other parts of the Turkey Moldova world, in the last three decades alone, this region experienced Constantinople earthquake 1502. Rorania earthquake 1940, Chisinau partially destroyed close to 500 significant floods and earthquakes that caused 50,000 fatalities, affected nearly 25 million people, and re- Yerevan earthquake 1679, city destroyed. Skopje earthquake 1963, city destroyed. sulted in US$80 billion in damage. In fact, close to 30 percent of the capitals of the ECA countries have been at one time or another devastated by earthquakes and floods. The historical record of large disasters goes back to ancient Gec,where an earthquake in Crete destroyed Alexandria DhonkHnayUbksa Gre ece, weraneatqakinCeedetoedAexnriaDbvikPest floods 1838. Tashkent earthquake 1966, city destroyed. (Figure 1). Much more recently, in 1838, massive floods earthquake 1667. destroyed Pest, Hungary, and in the 1960s, earthquakes Romania Turkmenistan devastated Skopje, the former Yugoslav Republic of Macedo- Bucharest earthquake 1802, 'The Big Ashgahat earthquake 1948,10 percent of country's nia (1963), and Tashkent, Uzbekistan (1966). The impacts of Earthquake of Gods Friday. population died. these disasters are pervasive. They displace and kill people, Azerbaijan destroy property, incapacitate industries, disrupt day-to-day Shamakhi earthquake 1667, 80,000 people died. life, and often affect the economic development of countries Turkey for years after the event. Antioch earthquake 526, 250,000 people died. Figure 1. A chronology of signoficant disasters affecting the EMA region. 1. In this publication, we follow the World Bank classification of ECA cou ntries; The impacts of natural hazards will most likely become even Previous country-level risk estlmates in the ECA region were Baltic States; Belarus, Estonia, [atvia, Lithuania, Poland larger in the future due to changes in climate and growth based in part on records in the EM-DAT International Di- Caucasus States;Macedonaia,eeMhaiAanawGeorgia CentralAsian States; Kazakhstan, Kyrgyz Republic, Takista, TUrkmear bs n stan, Uzbekistan expected to experience economic growth in the near future. historical events. This information provides little insight into European Union States; Bulgaria, Croatia, Cyprus, Czech Republic, Greece, The combination of this growth with old building stock, the probability of event occurrence, however, as the data are Hungary, Moldova, Romania, Slovakia, Ukraine (unplanned) urbanization, and increased exposure creates incomplete and of uneven quality and they lack homogeneity. Russia South East European States; Albania, Bosnia and Herzegovi na, Kosovo, FYR conditions conducive to natural disasters. If governments do More robust risk estimates require additional data. Macedonia, Montenegro, Serbia, Slovenia, Turkey not act to reduce their exposure and vulnerability to them, their countries'Brisks will increase dramatcally. The country risk profiles for floods and earth quakes present- 2. D. Guha-Sapir, E. Below, Ph. Hoyois, EM-DAT: International Disaster ed in this publication are based on quantitative risk assess- Database, Universit6 Catholique de Louvain, Brussels, Belgium, www. ments derived using global flood and earthquake models. emdat.be. The objectives are to inform governments of the levels of risk mary at the top of the first page, however, are based on the provincial3 and country levels for different return periods (2, their countries face and facilitate discussions on how they can unrounded risk estimates. 5, 10, 25, 50, 100, 250, 500, and 1,000 years) and as annual av- become more resilient to both current and future risk. erages. The earthquake results also include information on re- Thetabe n te frs pae how th 1 prvinesor ll tu rn period s o f 1, 20, and 2 00 yea rs. Affected p opul ation an d provinces, if the country has fewer than 10-most at risk, as affected GDP are defined as those experiencing flood water ranked by the annual average GDP affected by floods or earth- at any depth or ground motion intensities equal to or greater quakes, as a percentage of each province's GDP'. The accompa- than Modified Mercalli Intensity (MMI) VI. Estimates are cal- T he profiles for the ECA countries presented here indicate nying map uses color to display the GDPs of all the provinces. culated in absolute numbers and relative to the population or Sthe flood and earthquake risks to which the countries are Note that a map of population at risk will look very similar to GDP of the province or country. In this publication, we mainly exposed on national and provincial levels. Annual avenages are that of GDP at risk as the correlation between the two char- portray the relative numbers, since they are better indicators often used to convey the gross domestic product (GDP) and acteristics is high. The sizes of the colored discs represent the for comparing the impact of a disaster on communities. In population at risk. The annual average affected GDP is defined relative amounts of the provinces'GCDPs affected annually by addition, the earthquake model was used to estimate fatalities as the affected GDP per year, averaged over many years. Since floods (blue) and earthquakes (orange). and capital losses. Capital losses represent the estimated cost the annual average does not represent the impact of a single The middle pages of each profile assess the country's flood of replacing and repairing damage to gross capital stock-that event, it is important to realize that much larger impacts can and earthquake risks in greater detail. The map shows the is, the fixed assets in a country. be caused by less frequent, more intense events, annual average affected GDP of the country relative to its Any consideration of major investments to reduce risk or provinces' cDs, which can be used to identiy the provinces increase resiliency needs to take into account changes in risk with high percentages of DP at risk annually. The vertical caused by climate change and socioeconomic developments. bars on the maps represent the percentages of GDP affect- For this reason, risk information has been generated not only ted by a 10- and 100-year flood or earthquake, respectively for current conditions and exposure, but also for conditions A 100-year flood means that in any given year there is a 1 This risk information can be used to identity the provinces in 208B0, ac co rding to two cl im ate s cenanrios defin ed by percent chance of a flood of a large magnitude occurring, expected to be vulnerable to more extreme, but less frequent, Representative Concentration Pathways (RCPs) for climate It does not mean that a 100-year flood will occur every events. In addition, a horizontal line across the bars indicates change and two socioeconomic conditions defined by Shared 100 years. It is possible to have two 100-year floods in the the percentage of annual average affected GD. Socioeconomic Pathways (SSPs) for socioeconomic trends. same or concurrent years. The same is true for a 250-year The final page displays information on the current annual Climate is presumed to have no impact on seismic risk, so the earthquake; there is a 0.4 percent chance of a 250-year average risk of fatalities and capital loss from earthquakes. earthquake model considers changes in 2080 exposure only earthquake occurring in any given year. Like the table on the first page, rose diagrams show these data by modeling all five SSPs. for the top 10 provinces, or for all provinces if the country has fewer than 10. In addition, exceeaance probability curves are The first page of each profile provides an overview of the provided for flood and earthquake risk An exceedance curve country and its risk. At the top is a short summary listing pop- represents the probability of exceeding any given amount 3. Model results have been calculated at administrative level land at ulation and GDP and, in terms of both absolute numbers and of affected GD. The line depicts the exceedance probability country level. In this publication, we will refer to administrative level 1as percentages, the population and GDP affected by, respectively, curve for 2015 conditions, whereas a striped band spans the provincial level. a 100-year flood and a 250-year earthquake. Also provided range of exceedance probabilities consistent with a set of are the capital loss and fatalities, in absolute amounts and socioeconomic and climate scenarios projected for 2080. 4. For more information on the RCPs, seeM. Meinshausen, S. J. Smith, percntaes,expetedfro a 20-yar arthuak. TeseK. V. Cal[vin, 1, S. Da niel, M. L. T. Kain uma, I.-F Lamarque, K. Matsu moto, percentages, expected from a 250-year earthquake. These et at., "The RCP Greenhouse Gas Concentrations anid Their Extension figures give an immediate indication of how much risk a coun- from 1765 to 2300." Climatic Change 109 (2011; special issue): 213-41, try is facing in case of an extreme, but less frequent, event. As DOlJO.1007/s10584-011-0156-z. For information on the SSP scenarios, explained in the methodology and limitation sections of this see a special issue of the journal Climate Change on Shored Socioeconomic introduction, the uncertainties in the absolute risk estimates or all the countries in the ECA region, population and Pathways (2014), e.g., N. Nakicenovic, R. J, Lerpert, and A. C. Janetos, are large. Therefore, the absolute risk estimates have been IGDP affected by floods and earthquakes are estimated at "A Framework for the Development of New Socio-economic Scenarios for Climate Change Research Introductory Essay,' Climatic Change 122 (2014; rounded to one significant digit. The percentages in the sum- special issue): 351-61. DOac10.1007s10584-013-0982-2. The uncertainty in projections of climate and socioeconomic ed once they experience flood water at any depth or ground The earthquake risk results include estimates of fatalities and conditions is large and becomes larger as the projections motion intensities equal to or greater than MMI VI. In reality capital loss, in addition to the affected population and GDP. reach farther into the future. The spread in outcomes for the effect of a flood or earthquake on a population or an The additional information is generated using vulnerability the risk estimates reflects the uncertainty due to changes in economy depends on depth of the water or the intensity of the functions that convert ground motion into fatality and damage climate change scenarios and socioeconomic development ground motion. The actual impact of 2 meters of flood water, estimates. and to the variability of the different climate models used to for example, is likely to he significantly greater than that of 10 estimate flood risk. centimeters of flood water, hut the model results will show the same amounts of affected GDP for both 2 meters and 10 events and their associated fatalities, affected population, and A detailed description of the methodology and models used in centimeters of water. damage inflated to 2015 dollars, which provides context for this publication is given in the technical annex at the end. interpreting the modeled impacts. When comparing historical Another limitation is associated with the Global Flood Risk events, however, one should not only inflate the dollars, but with IMAGE Scenarios (GLOFRIS) model, which can be used also account for the growth of a country's population and to assess large-scale river flood risks as well as global risks, wealth over time; to do so, one can either use risk models or although it does not assess coastal, flash, or urban floods, normalize historical records. T he information presented in the profiles is meant to in- Because information on flood defenses is sparse on a global form governments of the levels of river flood and earth- scale, the version of GLOFRIS used for this publication does Afew of the profiles also include modeled estimates of the quake risk in their countries and to facilitate discussions with not account for flood protection measures and will therefore fatalities and damage that would be caused by a historical them on the need to reduce these risks and increase resilience overestimate the affected population and GDP for return pe- earthquake if it were to occur today The 1969 Banja Luka to natural disasters. Estimates given in terms of fatalities, af- riods lower than the design protection level of existing flood earthquakes in Bosnia and Herzegovina, for example, caused fected population, affected GDP, and capital loss provide a first defenses. This in turn leads to an overestimation of the annual 14 deaths and $50 million in damage. If one corrected just for impression of the risk in each country and the risk ranking average affected population and GDP. inflation, the damage today would be more than $300 million of its provinces. As the information is produced using global in 2015 dollars. Because of population growth, urbanization, flood and earthquake risk models, it is important to be aware In general, uncertainties in absolute flood risk estimates are and increased wealth, however, both the damage and fatalities of the limitations of the methodologies used for both hazards. large, while estimates of relative changes in risk under differ- would actually be much greater. Model estimates suggest over ent scenarios or variability across space are more robustbh 400 fatalities would occur, and the damage would be approx- The national decision makers for whom these risk assess- imately $4 billion-over 20 percent of Bosnia and Herze- ments are intended can use them to focus attention on areas govina's GDP. These figures could be even larger with future of their countries at high risk and support the prioritization growth in population and wealth if appropriate efforts are not of studies for further quantifying it. The assessments should 5. H. Ape[, B. Merz, and A. H. Thieken, Quantification of Uncertainties in made to increase resilience to earthquakes. Similar arguments not e usd fo thedesin ofriskredutionmeasres,such Fltood Ri sk Ass ess m ents, "In ternational Journ a/ of River Ras in Man agemen t can be made for flood damage and mitigation of future flood not be used for the design of risk reduction measures, such(2008):149-62, DI 101080/15715124.200&9635344; H. De Moe as flood protection, retrofitting of buildings, or risk-informed and 1. C. 1. H. Aerts, 'Effect of Uncertainty in Land Use, Damage Models risk. urban planning. Such measures require more detailed and and Inundation Depth on Flood Damage Estimates," Natural Hazards 58 calibrated models that include vital information on local (2011): 407-25, DOI: lO.10071s11069-010-9675-6; B. Merz, H. Kreihich condtios, uchas ive prfils, crret fooddefnse, lcal R. Schwarze, and A. Thieken, "Assessment of Economic Flood Damage," conditions, such as river profiles, current flood defenses, local10(2010) 1697-1724, building standards, and soil characteristics, as well as infor- DOI:lo.sl94Inhess-1o-1697-201o. mation on exposure, such as the occupancy and construction 11 the regions in ECA are significantly exposed to both of local structures and the vulnerability of structures to forces 6. P. Bupeck, H. De Moe[, L. M. Brouwer, and J. C. 1. H. Aerts, "How Reliable floods and earthquakes. Floods pose the highest risk for generated by a peril. They also require the extensive engage- Are Projections of Future Flood Damage?" Natural Hazards and Earth Sys- the Baltic States, the European Union States, and the Russian ment of local experts and stakeholders. tenm Sciences 11 (2011): 3293-3306, DOI: 105194/nhess-11-3293-2011. The population and GDP affected by floods and earthquakes 7. A detailed description of the limitations of the LOFRIS model can be have been assessed only as a function of hazard and exposure; found in H. Winsemius and P. Ward, "Flood Risk Profiles Europe-Central 8. A detailed description of the methodology can be found in .E. Daniell, Asia Region," Final project report to the Global Facility for Disaster Reduc- "Development of Soda-economic Fragility Functions for Use in Worldwide tion apd Recovery (GFDRR), Wor id Bank, no. 1209814-n00-ZWS-0002, Rapid Earthquake Loss Estimation Procedures," doctoral thesis, Karlsruhe indicated above, population and GDP are considered affect- 2014. Institute of Technology Karlsruhe, Germany 2014. more significant.0 For example, the GDP affected by a 100- year flood is estimated at $60 billion in Russia (Figure 4), and Annual average affected Annual average affected that of a 250-year earthquake is estimated at $300 billion in Annual average GDP Annual average GDP Turkey (Figure 6). affected population (million US$) affected population (million us$) The view of risk is very different if one ranks countries by the percentage of affected GDP rather than absolute GDP. From Baltic States 800,000 9,000 50,000 600 this perspective, the country to which a 100-year flood poses the greatest risk is, by far, FYR Macedonia, with nearly 20 percent of its GDP affected (Figure 5). The countries at great- Caucasus States 300,000 800 600,000 2,000 est risk from earthquakes are Armenia, Albania and Georgia, in that order, with over 88 percent of their respective GDPs affected by 2 50-year events (Figure 7). Central Asian States 1,000,000 4,000 2,000,000 5,000 European Union Strat U2,000,000 20,000 1,000,000 20,000 States Russian Federation 2,000,000 20,000 200,000 1,000 South East European Sthas 1,000,000 9,000 2,000,000 20,000 States Table 1. Annual Average Affected Population and GDP for Floods and Earthquakes. SOURCE: VALUES FROM MODEL RESULTS OF THIS STUDY. Federation, and while the Caucasus States, the Cental Asian ($20 billion), followed by Poland and Turkey. For earthquakes, States and the South East European states are more affected it is Turkey ($10 billion), followed by Romania and Greece. by earthquakes (see Table 1). by e rthu aes (ee able1).The annual average GDP affected by floods and earthquakes A ranking of the ECA countries by the annual average GDP' af- seems small relative to the total GDP of each country; it is fected by floods (Figure 2) and earthquakes (Figure 3) shows generally less than 5 percent. Figures 4-7, however, show that that average annual affected GDP among the countries varies the impact of more intense and less frequent events, such as over an order of magnitude. For floods, the country with the 100-year floods or 250-year earthquakes, quickly becomes highest annual average affected GDP is the Russian Federation 10. A 100-year flood has a return period of 100 years, which means the probability of a flood's occurring is 1 percent per year. A 250-year earth- 9. Annual average affected GDP is estimated by averaging the affected GDP quake has a return period of 250 years, which means the probability of an by individual floods or earthquakes over a long period of time. earthquake's occurring is 0.4 percent per year. Russian Federation Turkey - Poland Romania Turkey Greece -m m Czech Republic Uzbekistan Slovak Republic Slovenia Kazakhstan Turkmenistan Romania Russian Federation Hungary Croatia Slovenia Bulgaria Croatia Kazakhstan Serbia Hungary Ukraine Azerbaijan Lithuania Georgia Uzbekistan Albania Turkmenistan Poland Greece Slovak Republic Latvia i Armenia BIH Tajikistan a Belarus Serbia Macedonia, FYR Macedonia, FYR Bulgaria Kyrgyz Republic Georgia BIH Azerbaijan Ukraine Albania Czech Repbulic Moldova Kosovo Armenia Cyprus Estonia Montenegro Talikistan Moldova Montenegro Estonia Kyrgyz Republic Lithuania Kosovo Latvia Cyprus Belarus Million US$ 0 5,000 10,000 15,000 20,000 Million LIS$ 0 2,000 4,000 6,000 8,000 10,000 Figure 2. Annual Average GDPAffected by Floods. Figure 3. Annual Average GDP Affected by Earthquakes. SOURCE VALUES FROM MODEL RESULTS OF THIS STUDY. SOURCE: VALUES FROM MODEL RESULTS OF THIS STUDY. Russian Federation Macedonia, FYR Poland Slovak Republic Turkey BIH Czech Republic Serbia Slovak Republic Kazakhstan Kazakhstan Slovenia Romania Montenegro Hungary i Georgia Slovenia Latvia Serbia Lithuania Croatia Croatia Ukraine Czech Republic Lithuania Tajikistan Uzbekistan Moldova Turkmenistan Kyrgyz Republic Greece Hungary Latvia Romania Belarus Turkmenistan Bulgaria Armenia BIH Albania Macedonia, FYR Poland Georgia Russian Federation Azerbaijan Uzbekistan Albania Ukraine Armenia Belarus Estonia Bulgaria Tajikistan Turkey Moldova Kosovo Montenegro Estonia Kyrgyz Republic Azerbaijan Kosovo Greece Cyprus Cyprus Million US$ 0 20,000 40,000 60,000 0% 5% 10% 15% 20% Figure t. GDP Affected by a 100-year Flood. Figure 5. GDP Affected by a 100-year Flood Relative to the Country's GDP. SOURCE: VALUES FROM MODEL RESULTS OF THIS STUDY. SOURCE: VALUES FROM MODEL RESULTS OF THIS STUDY. Turkey Armenia i -.-.-.-.-.-.-.- Romania Albania Greece Georgia Hungary Azerbaijan Azerbaijan Tajikistan Bulgaria Romania Kazakhstan Montenegro - - -...-.-.-. - Turkmenistan Moldova M22 Slovenia Kyrgyz Republic Croatia m Turkmenistan Uzbekistan Bulgaria Slovak Republic Slovenia Georgia u Macedonia, FYR Serbia im Greece Russian Federation = Kosovo Albania a Croatia Poland Hungary Armenia Turkey Cyprus Cyprus Ukraine Uzbekistan Macedonia, FYR Serbia m Talikistan BIH Czech Republic Kazakhstan Moldova Slovak Republic BIH Ukraine Kyrgyz Republic Czech Republic Kosovo I Poland Montenegro Russian Federation Estonia Estonia Latvia Latvia Lithuania Lithuania Belarus Belarus Million US$ 0 100,000 200,000 300,000 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% Figure 6. GDP Affected by a 250-year Earthquake. Figure 7. GDP Affected by a 250-year Earthquake Relative to the Country's GDP. SOURCE VALUES FROM MODEL RESULTS OF THIS STUDY. SOURCE: VALUES FROM MODEL RESULTS OF THIS STUDY. WORLDBANKGROUP EI GROPE AND CENTRALA5IA (ECA) AFFECTED AFFECTED CAPITAL LOSS Q GFDRR RI5K PROFILES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $11.6 billion* Albamia 0.ouain29mlin A lbania's population and econo- remainder Albania's per capita KOSOVO my are exposed to earthquakes GDP was $3,990. S ER"A and floods, with earthquakes posing the greater risk of a high impact, This map displays GDP by prow BULrTRIA lower probability event. The model ince in Albania, with greater color results for present-day risk shown in sut i g this risk profile are based on population within a province. The blue circles and gross domestic product (GDP) esti- i t s f e mates for 2015. The estimated damage floods and the orange circles the caused by historical events is inflated to r of er 2015 US dollars. normalized annual average of affected GDI. The largest circles F MCED I Just over half of Albania's population represent the greatest normalized lives in urban environments. The coun- risk The risk is estimated using A v ed % try's GDP was approximately US$11.6 flood and earthquake risk models. billion in 2015, with close to 70 percent T tabT d0 the p derived from services and with in- T e ss dustry and agriculture generating the at greatest normalized risk for k .-4 I hARTHQUAKE each peril. In relative terms, as 'L24 shown in the table, the province a I o Negligible TOP AFFECTED PROVINCES at greatest risk of floods is Shko- der, and the one at greatest risk Gas - ~~~~of earthquakes is Pier In absoluteGP(blinof) EARTHQUAKE terms, the province at greatest ANNUAL AVERAGE OF ANNUAL AVERAGE OF risk of floods is also Shkoder, and AFFECTED GDP (%) AFFECTED GDP (%) the one at greatest risk of earth- Shkader Fier quakes is Tirane.0 0 0B Shkoder Fier 9 T eALC Mirdite Lushnje 8 Permet Tirane 7 Tepelene Durres 7 Tepeene Dures 7There is a high Correlation Lushnje Kucove 7 (r=0.95) between the Fier Delvine 7 population and GDP ala Diber Kavaje 6 province. Kurbin 2 Gjirokaster 6 Gin r Tropoje 2 Vlore 6 Masaakaster i Kurhin 6greterGD Aa ia WORLDBANKGROUP ELGF RI R AND EENTRAL A51A (ECA) T he most deadly flood in If the 10- and 100-year bars are the Albania since 1900 occurred same height, then the impact of a 10- S ER B IA in 1992. It killed 11 Albanians year event is as large as that of a 100- and caused close to $12 million in year event, and the annual average of damage. Flooding in 2002 caused one affected GDP is dominated by events BUt GARIA fatality but about twice the damage that happen relatively frequently. KOSOVO ($23 million) of the 1992 flood. Dam- If the impact of a 100-year event is aging flooding also took place on the much greater than that of a 10-year Drina River in 2010. event, then less frequent events make a larger contribution to the annual aeEdk This map depicts the impact of flood- average of affected GDP. Thus, even MaasE,ah ing on provinces' GDPs, represented if a province's annual affected GDP - as percentages of their annual aver- seems small, less frequent and more Puke age GDPs affected, with greater color intense events can still have large ukes saturation indicating higher percent- impacts. ages. The bar graphs represent GDP affected by floods with return periods The annual average population affect- of 10 years (white) and 100 years ed by flooding in Albania is about Fr (black). The horizontal line across the 50,000 and the annual average affect- bars also shows the annual average of ed GDP about $200 million. Within GDP affected by floods. the various provinces, the 10- and Durres *Tlrana 100-year impacts do not differ much, Tirane When a flood has a 10-year return so relatively frequent floods have Librahd Affected GOP (%) for period, it means the probability of large impacts on these averages. oava t and 100-year return periods occurrence of a flood of that magni- Lushnje in Elbasan ,-b tude or greater is 10 percent per year. A 100-year flood has a probability ADRIATIC SEA Kucove G 20 of occurrence of 1 percent per year. rli This means that over a long period of erat _.._ Annual avorage - 5 4 Krce time, a flood of that magnitude will, Ma ISkrapar P Devoll F7 on average, occur once every 100 10-year 100-year years. It does not mean a 100-year flood will occur exactly once every VIore 100 years. In fact, it is possible for a PC, Annual Average of Affected GDP(% flood of any return period to occur G rokaster more than once in the same year, or to appear in consecutive years, or not De to happen at all over a long period of a C P time. Sarande Al i WORLDBANKGROUP E GRO AND EENTRALA5IA(ECA) A lbania's most deadly earth- to occur more than once in the same quake since 1900 took place year, or to appear in consecutive in 1920 in Tepelene, with years, or not to happen at all over a SERBIA a magnitude of 6. The earthquake long period of time. and the tsunami that followed caused about 600 fatalities. Since If the 10- and 100-year-bars are the KOSOVO then, Albania has experienced many same height, then the impact of a 10- earthquakes of varying severity. A year event is as large as that of a 100- BULGARIA significant earthquake that occurred year event, and the annual average of in 1967 caused 18 fatalities and $140 affected GDP is dominated by events million in damage. that happen relatively frequently. If the impact of a 100-year event isHa This map depicts the impact of much greater than that of a 10-year earthquakes on provinces' GDPs, event, then less frequent events make represented as percentages of their larger contributions to the annual av- annual average GDPs affected, with erage of affected GDP. Thus, even if a greater color saturation indicating province's annual affected GDP seems higher percentages. The bar graphs small, less frequent and more intense Kurbin R MACEDON A represent GDP affected by earth- events can still have large impacts. years (white) and 100 years (black). The annual average population affect- quake wit returnake perod ofbni 0 Jbou The horizontal line across the bars e a k i n b also shows the annual average of GDP 200,000 and the annual average affected by earthquakes. 10 and 100-year retur periods The annual averages of fatalities and One block -10% 100 When an earthquake has a 10-year capital losses caused by earthquakes return period, it means the probabil- are about 50 and about $100 million, ity of occurrence of an earthquake respectively The fatalities and capital of that magnitude or greater is 10 losses caused by more intense, less Anmy2 aerage 20 percent per year. A 100-year earth- frequent events can he substantial- quake has a probability of occurrence ly larger than the annual averages. 10-year 100-year of 1 percent per year. This means For example, an earthquake with that over a long period of time, an a 0.4 percent annual probability of earthquake of that magnitude will, on occurrence (a 250-year return period Au A o A GOP ( average, occur once every 100 years. event) could cause neariy 3,000 It does not mean a 100-year earth- fatalities and $2 billion in capital loss quake will occur exactly once every (about 20 percent of GOP). 100 years. In fact, it is possible for Tepelene an earthquake of any return perihod Gj kaster a nWORLDBANKGROUP GF DP AND CENTRALA1A (ECA) ' )EARTHQUAKE EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS (MILLIONS $) ANNUAL AVERAGE FATALITIES he rose diagrams show the provinces with the potential for greatest annual average capital losses and highest annual average numbers of fatalities, as determined using an earthquake risk modeL The potential for greatest capital Kavaje er Kava E an loss occurs in Tirane, which is not surprising, given the economic importance of the province. ---------- ---- ------ ---------------------------- --- - - - - - - - - - - - - - - - - - - ----------------------- --------------------------------------------- ----------------------------------------- ---------- EARTHQUAK EXCEEDANCE PROBABILITY CURVE 2015 AND 2080 EXCEEDANCE PROBABILITY CURVE 2015 AND 080 he exceedance probability curves display the GDP E affected by, respectively, floods and earthquakes for 3.0 varying probabilities of occurrence. Values for two different 60 time periods are shown. A solid line depicts the affected 2DP for 2015 conditions. A diagonally striped band depicts 2080 the range of affected GDP based on a selection of climate 2.0 40 and socioeconomic scenarios for 2080. For example, if Al- bania had exterienced a 100-year return period flood event 30 in 2015, the affected GDP would have been an estimated $700 million. In 2080, however, the affected GDP from the 2015 d- same type of event would range from about $2 billion to 0.5 2015 abo $2.5 billion. If Albania had experienced a 250-year earthquake event in 2015, the affected GDP would have been about $10 billion. In 2080, the affected GDP from the Re year)Rtupodysame type of event would range from about $30 billion to . .about $60 billion, due to population growth, urbanization, in 1 and the increase in exposed assets. Probability '%)Probability % All historical data on floods and earthquakes are from, respectively, D. Guha-Sapir, R. Below, and Ph. Hoyois, EM-DAT: International Disaster Database (Uiveirsit6 Catholique de Loiuvain, BruIsel, Belgium), www.emdat.be, and the National Geophysical Data Center/World Data Service (NGDC/WDS), Significant Earthquake Database (National Geophysical Data Center, NOAA), doi10 7289/V5TD9V7K. Damage estimates for all historical events have been inflated to 2015 US$. WORLDBANKGROUP EI GEROPE AND CENTRALA5IA (ECA) AFFECTED AFFECTED CAPITAL LOSS Q GFDRR RI5K PROFILES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $10.7 billion* Arm emia 0 Pouato .0 milin IL I N J I I , RUSSIAN' A rmenia's population and economy remainder. Armenia's per capita GDP was are exposed to earthquakes and $3,550. floods, with earthquakes posing the greater risk of a high impact, but lower This map displays GDP by province in probability event. The model results for Armenia, with greater color saturation present-day risk presented in this risk pro- indicating greater GDP within a province. file are based on population and gross do- The blue circles indicate the risk of expe- mestic product (GDP) estimates for 2015. riencing floods and the orange circles the The estimated damage caused by historical risk of earthquakes in terms of normalized events is inflated to 2015 dollars. annual average of affected GDP. The largest circles represent the greatest normalized More than 60 percent of Armenia's pop- risk. The risk is estimated using flood and Koty - I EB I ulation lives in urban environments. The earthquake risk models. reva country's GDP was approximately US$10.7 6ergarkunik billion in 2015, with most derived from The table displays the provinces at greatest Ari services and industry (together about 80 normalized risk for each peril. In relative percent) and agriculture generating the terms, as shown in the table, the province WS raa at greatest risk of floods is Gergharkunik, and the one at greatest risk of earthquakes V zor is Armavir. In absolute terms, the province T UKN TOP AFFECTED PROVINCES at greatest risk of both floods and earth- quakes is Yerevan. EARTHQUAKE ANNUAL AVERAGE OF ANNUAL AVERAGE OF AFFECTED GDP (%) AFFECTED GDP (%) LMIC REPLEBLC OF IRAN r Gergharkunik Armavir 5 Kotayk Ararat 5 Syunik 2 Yerevan 3 Annual Average of Affected GDP (%) GDP (billions of $), Armavir 2 Shirak 3 Yerevan I Aragatsotn 3 5 foThere is a high correlation [on Koayk 3 I I I(r=0.95) between the Ararat I Gergharkunik 2 1 oEARTHQUAKE o O q population and GDP ofa Vayots Dzor 0 Vayots Dzor 1 EAP 1A A province. Tavush 0 Lori 1 p Shirak u Syunik 1 0 Negligible Arm enia F O DWORLDBANKGROUP Q GFDRR RSK" AROFIENTAA1AEA he most devastating flood year event, and the annual average of R in modern Armenia since it affected GDP is dominated by events gained its independence in that happen relatively frequently. - GE ORG IA 1991 occurred in 1997. It killed four If the impact of a 100-year event is people, affected about 7,000, and much greater than that of a 10-year caused $12 million in damage. event, then less frequent events make a larger contribution to the annual This map depicts the impact of flood- average of affected GDP. Thus, even ing on provinces' GDPs, represented ifa province's annual affected GDP as percentages of their annual aver- seems small, less frequent and more age GDPs affected, with greater color intense events can still have large saturation indicating higher percent- impacts. 5ak ages. The bar graphs represent GDP affected by floods with return periods The annual average population of 10 years (white) and 100 years affected by flooding in Armenia is (black). The horizontal line across the about 40,000 and the annual average bars also shows the annual average of GDP about $100 million. For most Aragatsotn GDP affected by floods. provinces, in which the impacts from 10- and 100-year floods do not differ Y6 evan When a flood has a 10-year return much, relatively frequent floods have rh n k period, it means the probability of large impacts on these averages. occurrence of a flood of that magni- For the few in which the 100-year Ararat tude or greater is 10 percent per year. impacts are much greater than the A 100-year flood has a probability 10-year impacts, less frequent events Vay Dzor of occurrence of 1 percent per year. make a significant contribution to the This means that over a long period of annual average of affected GDPTRE time, a flood of that magnitude will, on average, occur once every 100 years. It does not mean a 100-year flood will occur exactly once every 100 years. In fact, it is possible for a ISAIC RL I OF !RAN flood of any return period to occur -Affected GOP (%) for Annual Average of Affected GDP (%) more tha n on ce in th e s am e year, or 1 n 0-errtr eid 10 and 100-year return periods to appear in consecutive years, or not O Orne block =2 20 to happen at all over a long period of time. ovrg fj 10 If the 10- and 100-year bars are the Annua[ average same height, then the impact of a 10- year event is as large as that of a 100- 10-year 100-year ArmeiaWORLDBANKGROUP GFDR ROPE AND ECENTRAL A5IA(ECA) Arm enia RS[K PROFI A rmenia's worst earthquake If the 10- and 100-year bars are the since 1900 took place in 1988 same height, then the impact of a 10- in Spitak with a magnitude of year event is as large as that of a 100- CERU I 6.8. It caused about 25,000 fatalities year event, and the annual average of and more than $30 billion in damage. affected GDP is dominated by events A 1931 earthquake in Zangezur, with that happen relatively frequently. a magnitude of 6.3, killed over 2,800 If the impact of a 100-year event is people. much greater than that of a 10-year event, then less frequent events make This map depicts the impact of larger contributions to the annual av- earthquakes on provinces' GDPs, erage of affected GDP. Thus, even if a represented as percentages of their province's annual affected GDP seems annual average GDPs affected, with small, less frequent and more intense greater color saturation indicating events can still have large impacts. hirak higher percentages. The bar graphs represent GDP affected by earth- The annual average population quakes with return periods of 10 affected by earthquakes in Armenia is rK years (white) and 100 years (black). about 90,000 and the annual average erevan The horizontal line across the bars affected GDP about $300 million. The also shows the annual average of GDP annual averages of fatalities and cap- affected by earthquakes. ital losses caused by earthquakes are k about 400 and about $200 million, When an earthquake has a 10-year respectively. The fatalities and capital return period, it means the probabil- losses caused by more intense, less ity of occurrence of an earthquake frequent events can be substantial- Yeoa ofthat magnitude or greater is 10 ly larger than the annual averages. TURK percent per year. A 100-year earth- For example, an earthquake with quake has a probability of occurrence a 0.4 percent annual probability of of 1 percent per year. This means occurrence (a 250-year return period that over a long period of time, an event) could cause neariy 10,000 earthquake of that magnitude will, on fatalities and $6 billion in capital loss ISL - U L F IR average, occur once every 100 years. (about 60 percent of GDP) It does not mean a 100-year earth- Affected GDP ( for Annual Average of Affected GOP (%) quake will occur exactly once every 10 and 100-year return periods One block = 10% 100 100 years. In fact, it is possible for an e arthqu ake o f any retu rn pe ri od V 6 6r to occur more than once in the same 50 year, or to appear in consecutive Annua[ average 20 years, or not to happen at all over a long period of time. 10-year 100-year WORLDBANKGROUP G D ROP^ A ENTRAL A5IA (ECA) EARTHQUAKE EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS (MILLIONS $) ANNUAL AVERAGE FATALITIES T he rose diagrams show the provinces with the potential 1for greatest annual average capital losses and highest annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital Syunik 5 Armavir 20 Lo 2 rarat 0 loss occurs in Yerevan, which is not surprising, given the economic importance of the province. EXCEDANE PRBABLIT CURE, 015 ND 080EARTHIQUAKE EXCEEDANCE PROBABILITY CURVE 2015 AND 2080 he exceedance probability curves display the GDP 1 affected by, respectively, floods and earthquakes for 35 '50 varying probabilities of occurrence. Values for two different time periods are shown. A solid ]ine depicts the affected 30 40 GDP for 20-15 conditions. A diagonpllystriped band depicts 2080 the range of affectéd GDP based on a selection of climate 0 2and socioeconomic scenarios for 2080. For example, if 2.0 30 Armenia had experienced a 100-year return period flood 1.5 event in 2015, the affected GDP would have been an esti- 20mated $700 million. In 2080, however, the affected GDP t 2015 tO from the same type of event would range from about $2 0.5 - -10 2015 billion to about $3 billion. If Armenia had eperienced a 250-year earthquake event in 2015, the affected GDP would 10 50 100 2501 5 100 20 have been about $9 billion. In 2080, the affected GDP from Returnperiod (years)Rturn period (years) the same type of event would range from about $30 billion .. r . . .. .... . - - to about $50 billion, due to population growth, urbaniza- 0P bi (%) tion, and the increase in exposed assets. ALL historical data on floods and earthquakes are from, respecriveLy, D uh-api, R Below, and Ph. Hoyois, EM-DAT: International Disaster Database (Université CathoLique de Louvain, Brussels, Belgium), www emdat.be, and the National Geophysical Data Center/World Data Service (NGDC/WDS), Sigiicant Eaithqiake Dataase (National Geophysical Data Center, NOAA), doi:107289/V5TD9V7K. Damage estimates for all historical events have been inflated to 2015 US$. WORLDBANKGROUP GR EROPE AND CENTRALA5IA (ECA) AFFECTED AFFECTED CAPITAL LOSS Q GFDRR RI5K PROFILES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $54.6 billion* 5 APopulation 9.7nillion* D A zerbaijan's population and econ- Azerbaijan's per capita GDP was I H I FDRTION omy are exposed to earthquakes $5,630. and floods, with earthquakes posing the greater risk of a high impact, This map displays GDP by prov- Kh!cas lower probability event. The model re- ince in Azerbaijan, with greater sults for present-day risk shown in this color saturation indicating Mingetchaur / risk profile are based on population and greater GDP within a province. 'Divitc1 gross domestic product (GDP) estimates The blue circles indicate the risk for 2015. The estimated damage caused o e f by historical events is inflated to 2015 i. am r US dollars. quakes in terms ofnormalized Emally Just over half of Azerbaijan's population The largest circles represent the flKedabek AhhG a lives in urban environments. The coun- greatest normalized risk- The 0 hn try's GDP was approximately US$54.6 risk is estimated using flood and Tartar Bard a billion in 2015, with close to 60 percent earthquake risk models. N * L derived from industry, most of the reanergnrae y evce,ad The table displays the provinces Sederek KbajrAd AgdAabFiLa1%% ra remainder generated by services, and 9a ibd .L agrcutue aknga sal cntibtin, at greatest normalized risk for Nagorno-karaba kh agriculture making a small contribution.I'P1Ii ____________________ each peril. In relative terms, aslatchinAima shown in the table, the prince -P- TOP AFFECTED PROVINCES at greatest risk of floods is Zard- fterur Bi[Esuvar ob, and the one at greatest risk IBek o . I,#d1 of earthquakes is Ali Bajramly, KUIhat[Y DiebraiI NeAftethal[a EARTHQUAKE In absoluteterms, the province DLJtfa DjaLiabad ANNUAL AVERAGE OF ANNUAL AVERAGE OF at greatest risk of floods is Ali 4M Oruad Nkhitchevan/ F AFFECTED GDP (%) AFFECTED GDP (%) Bajramly, and the one at greatest I Zardob Ali Bajramly 6 ijufa City * O RAN ffikkan Ali Bajramly Astara 5 Ordubad City Sabirobad Salyany 5 Kurdamir Cherur 5 AnnualAverageof Affected GOP GDP (billions of) star Neftetchala Utchar 5 Salyay20 There is a high correlation Saatly 0 Geoktchay 4 Agdach Neftetchala 4 U EARTHQUAKE population and GDP ofa Ordubad 3 Lenkoran 4 C 1 province. Akstafa 3 Sahirohad 4 0 Negligible a ijanWORLDBANKGROUP |F RE AN EENTRALA51A(ECA) he most devastating floods If the 10- and 100-year bars are the in modern Azerbaijan since same height,then the impactofa 10- it gained its independence yeareventaslargeasthatofa1 - in 1991 occurred in 2003, affecting year event, and the annual avenge of more than 30,000 people and causing affected GDP is dominated by events / IL TO over $70 million in damage. Floods that happen relatively frequently. in 1995 affected over 1.5 million If the impact of a 100-year event is people and caused about $30 million much greater than that of a 10-year in damage. event, then less frequent events make Khatchmas This map depicts the impact of flood- avagerontiutet h en Anca I ak / CASPIAN SEA ing on provinces' GDPs, represented if a province's annual affected GOP EaKait as percentages of their annual aver- Z as erenags f her nnalavr- seems small, less frequent and more Tauz -cx Choki Ofuz - age GDPs affected, with greater color intense events can still have large - b a - Siya5an saturation indicating higher percent- impacts. r Chamkhor - 'hemakhauKhyiy ages. The bar graphs represent GDP Gvnaj W C Khyz affected by floods with return periods The annual avenge population affect- Geranboy oktchay of 10 years (white) and 100 years ed by flooding in Azerbaijan is about Kedabek - a (black). The horizontal line across the 100,000 and the annual average Tartar UB -< arard Utphareron bars also shows the annual average of affected GDP about $300 million. AR MEN I'A GDP affected by floods. Within the various provinces, the 10- hadji6bu Whn foo hsa 0-ea rtun and 100-year impacts do not differ Sederek Keladjar ,, Aqdjabedi Ali Bajram[y When a flood has a 10-year return __ much, so relatively frequent floods Nagorno-karabakh - - I period, it means the probability ofImcl peio, t ensth poabliy f have large impacts on these averages. 130 Bjadjan occurrence of a flood of that magni- tude or greater is 10 percent per year. Cherur Fizaib A 100-year flood has a probability A Bile5uvar of occurrence of 1 percent per year. Kuhatty Djebrai This means that over a long period of f Djaitabad eft ala time, a flood of that magnitude will, Nakhitchevan Onubd Zang6lan on average, occur once every 100 years. It does not mean a 100-year Yardam[y flood will occur exactly once every And G0e rtr period Cii k 100 years. In fact, it is possible for a 10 ac 10 r 0 flood of any return period to occur I LAM P REPUBLIC Astara tor apan onectivsae year, orno 40 Annual Average of Affected GDP (%)OFAN more than once in the same year, orj to appear in consecutive years, or not to happen at all over a long period of Annua[ average -u 10 time. r- ,%_ 10-year 100ae -year Aze ba ja WORLDBANKGROUP EL GF R 'ROP AND EENTRAL A5A(ECA) A zerbaijan's worst earth- of any return period to occur more quake in recent decades than once in the same year, or to took place in 2000 in the appear in consecutiveyears, or not U I'I FEDERATION capital city of Baku, with a magni- to happen at all over a long period tude of 6.8. It caused more than 30 of time. fatalities and over $10 million in damage. A 1999 earthquake caused If the 10- and 100-year bars are the one death and nearly $7 million in same height, then the impact of a Evlkh MinetchX'J damage. The most deadly known 10-year event is as large as that ofa g mabelaSEh earthquake in Azerbaijan's history 100-year event, and the annual av- occurred in 1667 or 1668 and erage of affected GDP is dominated caused around 80,000 fatalities. by events that happen relatively fre- quently. If the impact of a 100-year5 This map depicts the impact of event is much greater than that of earthquakes on provinces' GDPs, a 10-year event, then less frequent A represented as percentages of their events make larger contributions to Gandla annual average GDPs affected, with the annual average of affected GDP. greater color saturation indicating Thus, even if a province's annual higher percentages. The bar graphs affected GDP seems small, less fre- DAhk'sT represent GDP affected by earth- quent and more intense events can "_' quakes with return periods of 10 still have large impacts. IaIata years (white) and 100 years (black). Sederok The horizontal line across the bars The no-kara kh also shows the annual average of fce b e In A GD afctd yeathuke.is about 200,000 and the annualav-Saiod GDP affected by earthquakes. erage affected GDP about $1 billion. When an earthquake has a 10-year The annual averages of fatalities return period, it means the prob- and capital losses caused,by earth- ability of occurrence of an earth- quakes are about 800 and about quake of that magnitude or greater $200 million, respectively. The fatal- If Zardab is 10 percent per year. A 100-year ities and capital losses caused by Nakhitchevan earthquake has a probability of more intense, less frequent events A e % occurrence of t percent per year. can be substantially larger than This means that over a long period the annual averages. For example, of time, an earthquake of that mag- an earthquake with a 0.4 percent Djulfa City Ordubad City ae:aJ? nitude will, on average, occur once annual probability of occurrence Afc every 100 years. It does not mean (a 250-year return period event) Annual a 100-year earthquake will occur could cause nearly 40,000 fatalities ISLAMIC REP R LI exactly once every 100 years. In and $6 billion in capital loss (about OFI?,AN fact, it is possible for an earthquake 10 percent of GDP). 10-year 100-year Aze ba ja WORLDBANKGROUP E|G DR ROPE"AND EENTRAL A51A(ECA) EARTHQUAKE EARTHQUAKE - ANNUAL AVERAGE CAPITAL LOSS (MILLIONS $) ANNUAL AVERAGE FATALITIES Ze 0he rose diagrams show the provinces with the potential ,04 for greatest annual average capital losses and highest annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital Lenkoran 3 Al Bairamly 5 Gandia 20 herur 30 loss occurs in Baku, which is not surprising, given the eco- naomc iportance of the province. 28 2U EARTHQUAKE EXCEEDANCE PROBABILITY CURVE, 2015AND2080 EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 The exceedance probablity curves display the GDP T affected by, respectively, floods and earthquakes for 4300 varying probabilities of occurrence. Values for two different time periods are shown. A solid line depicts the affected Z: 250 GDP for 2015 conditions. A diagonally striped band depicts the range of affected GDP based on a selection of climate 2080 0and socioeconomic scenarios for 2080. For example, if 150 20Azerbaijan had experienced a 100-year return period flood event in 2015, the affected GDP would have been an esti- 2015 di 100 mated $1 billion. In 2080, however, the affected GDP from 1 0d the same type of event would range from about $2 billion to < 50 2015 about $3 billion. If Azerbaijan had experienced a 250-year earthquake event in 2015, the affected GDP would have 10 , 50 100 250 been about $40 billion. In 2080, the affected GDP from the Retu rn p eripot(y ea rs Retin period (years) same type of event would range from about $80 billion to about $240 billion, due to population growth, urbanization, P b (% 1 0-' 10 2 1 0.4, Probability Probability () 1and the increase in exposed assets. All historical data on floods and earthquakes are from, respectively, D Guha-Sapir, R Below, and Ph Hoyois, EM-DAT International Disaster Database (Universite Catholique de Louvain, Brussels, Belgium), www.emdat.be, and the National Geophysical Data Center/World Data Service (NGDCWDS), Significant Earthquake Database (National Geophysical Data Center, NOAA), doi107289/V5TD9V7K Damage estimates for all historical events have been inflated to 2015 US$. WORLDBANKGROUP GLI FEROPE ANDCENTRALA5IA (ECA) AFFECTED AFFECTED CAPITAL LOSS Q GFDRR RI5K PROFILES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAK(E EARTHQUAK(E GDP $56.8 billion* Be la rus 0.Population 9.2 million* LATVIA elarus' population and econo- a small contribution. Belarus's per my are exposed to earthquakes capita GDP was $6,160. and floods, with floods posing RUSSIAN FEDERATION the greater risk. The model results map displays GDP by province in for present-day risk shown in this Belarus, with greater color saturation LITHUANIA risk profile are based on population indicating greater GDP within a prov- 1 and gross domestic product (GDP) ince. The blue circles indicate the risk Virehsk estimates for 2015. The estimated of experiencing floods and the orange damage caused by historical events is circles the risk of earthquakes in inflated to 2015 US dollars. terms of normalized annual average of affected GDP. The largest circles Just over 75 percent of Belarus's pop- represent the greatest normalized ulation lives in urban environments. risk. The risk is estimated using flood in k The country's GDP was approximate- and earthquake risk models. Minsk ly US$56.8 billion in 2015, with close to 90 percent derived from industry The table displays the provinces at Gro1noMin-k and services, and agriculture making greatest normalized risk for each peril. In relative and absolute terms, the province at greatest risk of both floods and earthquakes is Vitebsk. TOP AFFECTED PROVINCES Brest- EARTHQUAKE ANNUAL AVERAGE OF ANNUAL AVERAGE OF PL AD AFFECTED GDP (%) AFFECTED GDP (%) Vitebsk Vitebsk 0 UKRAINE Gomel 2 Grodno 0 Grodno 2 Mogilev 0 Annual Average of Affected GDP (%) GDP (billions of $) srest I Minsk City 0 3 There is a high correlation Mogilev I Minsek 0 LOD(r=0.95) between the Minsk 0 Brest 0 1 population and GDP of a Minsk City 0 Gomel 0 3 EARTHQUAKE C KP province. 0 Negligible Bel rusWORLDBANKGROUP E|G DR ROPE AND EENTRAL A51A(ECA) he most damaging flood in If the 10- and 100-year bars are the L AT IA Belarus since it gained its inde- same height, then the impact of a 10- pendence in 1991 occurred in year event is as large as that of a 100- 1993, affecting approximately 40,000 year event, and the annual average of RUSS AN FEDERATI0N people and causing at least $150 affected GDP is dominated by events million in damage. Flooding in 1999 that happen relatively frequently. killed two people, affected more than If the impact of a 100-year event is 2,000, and caused over $5 million in much greater than that of a 10-year damage. event, then less frequent events make a larger contribution to the annual This map depicts the impact of flood- average of affected GDP. Thus, even ing on provinces' GDPs, represented if a province's annual affected GDP Virehsk as percentages of their annual aver- seems small, less frequent and more age GDPs affected, with greater color intense events can still have large saturation indicating higher percent- impacts. ages. The bar graphs represent GDP affected by floods with return periods The annual average population affect- of 10 years (white) and 100 years ed by flooding in Belarus is about a Minsk City (black). The horizontal line across the 100,000 and the annual average A bars also shows the annual average of affected GDP about $600 million. Ix GDP affected by floods. Within the various provinces, the 10- Grodno and 100-year impacts do not differ Moilev When a flood has a 10-year return much, so relatively frequent floods Minsk period, it means the probability of have large impacts on these averages. occurrence of a flood of that magni- tude or greater is 10 percent per year. A 100-year flood has a probability of occurrence of 1 percent per year. This means that over a long period of B r time, a flood of that magnitude will, - on average, occur once every 100 years. It does not mean a 100-year Affected GOP (%) for flood will occur exactly once every 10 ard 100-year return periods 100 years. In fact, it is possible for a One black = 1% 1 flood of any return period to occur more than once in the same year, or 5 Annual Average of Affected GDP (%) to appear in consecutive years, or not to happen at all over a long period of Annua[ average 1 2 time. 10-year 100-year 0 1 BelrusWORLDBANKGROUP EL GFR ROP AND ECENTRAL A5IA(ECA) elarus's worst earthquake affected GDP is dominated by events since 1900 took place in 1908. that happen relatively frequently. LATV IA If the impact of a 100-year event is much greater than that of a 10-year RUSS AN FEDERATION This map depicts the impact of event, then less frequent events make earthquakes on provinces' GDPs, larger contributions to the annual av- represented as percentages of their erage of affected GDP. Thus, even if a annual average GDPs affected, with province's annual affected GDP seems greater color saturation indicating small, less frequent and more intense higher percentages. The bar graphs events can still have large impacts. L represent GDP affected by earth- quakes with return periods of 10 The annual average population affect- Vitebsk years (white) and 100 years (black). ed by earthquakes in Belarus is about The horizontal line across the bars 100 and the annual average affected also shows the annual average of GDP GDP about $800,000. The annual av- affected by earthquakes. erages of fatalities and capital losses caused by earthquakes are less than & When an earthquake has a 10-year one and about $400,000, respective- return period, it means the probabil- ly. The fatalities and capital losses Minsk ity of occurrence of an earthquake caused by more intense, less frequent of that magnitude or greater is 10 events can be substantially larger Grodno percent per year. A 100-year earth- than the annual averages. For exam- Moqilev quake has a probability of occurrence pie, an earthquake with a 0.4 percent of 1 percent per year. This means annual probability of occurrence (a that over a long period of time, an 250-year return period event) could earthquake of that magnitude will, on cause about $20 million in capital average, occur once every 100 years. loss (less than 1 percent of GDP). It does not mean a 100-year earth- quake will occur exactly once every Brest Gomel 100 years. In fact, it is possible for an earthquake of any return period to occur more than once in the same I L N J year, or to appear in consecutive E years, or not to happen at all over a long period of time. GDP (%) not affected for 10 and Annual Average of Affected GDP (%) If the 10- and 100-year bars are the 100-year return periods same height, then the impact of a 10- Annual average = 0 year event is as large as that of a 100- year event, and the annual average of a / / S C

Gabrovo 10 0-ye ar im pacts do not d iffer much, oi When a flood has a 10-year return so relatively frequent floods have s9fact BLCKSE period, it means the probability of large impacts on these averages. BLAC SrA occurrence of a flood of that magni- tude or greater is 10 percent per year. A 100-year flood has a probability of occurrence of t percent per year. This means that over a long period of time, a flood of that magnitude will, on average, occur once every 100 YO MAEDONIA e'vogvrad years. It does not mean a 100-year TURKEY flood will occur exactly once every 100 years. In fact, it is possible for a Affected GP for flood of any return period to occur 10 and 100-year return periods more than once in the same year, or to appear in consecutive years, or not GREECEU to happen at all over a long period of tm.Annual Average of Affected GOP ()10 time. i Annrua[ average 4 If the 10- and 100-year bars are the same height, then the impact of a 10- o10-year 100-year Bulg riaWORLDBANKGROUP EL GF RROP AND ECENTRAL A5IA(ECA) ulgaria's worst earthquake years, or not to happen at all over a since 1900, with a magni- long period of time. tude of 7, took place in 1928 9 2 8MAAI A in Plovdiv. It caused over 120 fa- If the 10- and 100-year bars are talities and left more than 260,000 the same height, then the impact people homeless. An earthquake in of a 10-year event is as large as 1977 caused 20 deaths. that of a 100-year event, and the annual average of affected GDP is( This map depicts the impact of dominated by events that happen earthquakes on provinces' GDPs, relatively frequently. If the impact represented as percentages of of a 100-year event is much great- their annual average GDPs affect- er than that of a 10-year event, Vidin ed, with greater color saturation then less frequent events make indicating higher percentages. The larger contributions to the annual bar graphs represent GDP affected average of affected GDP. Thus, even by earthquakes with return peri- if a province's annual affected GDP ods of 10 years (white) and 100 seems small, less frequent and years (black). The horizontal line more intense events can still have across the bars also shows the large impacts. annual average of GDP affected by earthqakes.The annual average population af- earthquakes. n fected by earthquakes in Bulgaria ri oi; When an earthquake has a 10- is about 100,000 and the annual year return period, it means the avenge affected GDP about $1 probability of occurrence of an billion. The annual averages of earthquake of that magnitude or fatalities and capital losses caused greater is 10 percent per year. A by earthquakes are about 100 and 100-year earthquake has a prob- about $100 million, respectively. ability of occurrence of 1 percent The fatalities and capital losses per year. This means that over a caused by more intense, less fre- long period of time, an earthquake quent events can be substantially of that magnitude will, on average, larger than the annual averages. Affected GDP (%) for occur once every 100 years. It does For example, an earthquake with not mean a 100-year earthquake a GA percent annual probability 10 100 will occur exactly once every 100 of occurrence (a 250-year return CREECI years. In fact, it is possible for an period event) could cause nearly earthquake of any return period to 5,000 fatalities and $4 billion in Annual Average o A occur more than once in the same capital loss (about 8 percent of A 20 year, or to appear in consecutive GDP). ofth 1- nd10-ya10-year 100-year BulgariaWORLDBANKGROUP ER "AD ENTRALA51A(ECA) EARTHQUAKE (27 EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS (MILLIONS $) ANNUAL AVERAGE FATALITIES T he rose diagrams show the provinces with the potential for greatest annual average capital losses and highest annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital Haskovo 2 arna 4Kustndil BIc grad 10 loss occurs in Sofia-city, which is not surprising, given the economic importance of the province. ----------------- -- - -- --------- -------------------------- - ------------------ ----------------------------------------------------------------------------------------------------------------------------------------------------- EARTHQUAKE EXCEEDANCE PROBABILITY CIRVE, 2015 AND 2080 EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 he exceedance probability curves display the GDP E affected by, respectively, floods and earthquakes for 9 180 varying probabilities of occurrence. Values for two different g 160 time periods are shown. A solid line depicts the affected 7 140 GDP for 2015 conditions. A diagonally striped band depicts 2080 the range of affected GDP based on a selection of climate 6 o120 and socioeconomic scenarios for 2080. For example, if Bul- 5 100 2080 garia had experienced a 100-year return period flood event 4 80 in 2015, the affected GDP would have been an estimated $2 2015 3 60 billion. In 2080, however, the affected GDP frmm the same --- 40 2ols type of event would range from about $4 billion to about 1i 20 $8 billion. If Bulgaria experienced a 250-year earthquake event in 2015, the affected GDP would have been about 10 2 10 So 1oo 250 $30 billion. In 2080, the affected GDP from the same type eturn period (years) Return period (years) of event would range from about $70 billion to about $160 04 10 2 1 billion, due to population growth, urbanl4'ation, and the robability Probability (%) increase in exposed assets. All historical data on floods and earthquakes are from D. Guha-Sapir R. Below, and Ph. Hoyois, EM-DAT: International Disaster Database (Universit6 Catholique de Louvain, Brussels, Belgium), www.emdat.be; the National Geophysical Data Center/World Data Service (NG0C/WDS), Significant Earthquake Database (National Geophysical Data Center, NOAA), doi107289/V5TD9yK; and . Daniell and A. Schaefer, "Eastern Europe and Central Asia Region Earthquake Risk Assessment Country and Province Profilng," final report to GFDRR, 2014. Damage estimates tor all historical events have been inflated to 2015 US$ WORLDBANKGROUP EI GROPE AND CENTRALA5IA (ECA) AFFECTED AFFECTED CAPITAL LOSS Q GFDRR RI5K PROFILES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $47.6 billion* Cr0atl*a 0opuain42mlin HUNG AR C roatia's population and economy 70 percent derived from ser- are exposed to earthquakes and vices, most of the rest generated floods, with earthquakes posing by industry, and agriculture a ar in the greater risk of a high impact, lower making a small contribution. probability event. The model results for Croatis per capita GDP was present-day risk shown in this risk pro- $11,300. file are based on population and gross domestic product (GDP) estimates for This map displays GDP by Virovitica podravin 2015. The estimated damage caused by in C wit g li baah historical events is inflated to 2015 US Pomga-slavonija dollars. greater GDP within a province. isak-moslavina The blue circles indicate the risk Primorj StavnDk ro pota Vukoarovijc Nearly 60 percent of Croatia's popu- of experiencing floods and the lation lives in urban environments. orange circles the risk of earth- The country's GDP was approximately quakes in terms of normalized SERBIA US$47.6 billion in 2015, with close to annual avenge of affected GOR BOSNI AND The largest circles represent theHEREGIN, greatest normalized risk. The 10 TOP AFFECTED PROVINCES risk is estimated using flood and EARTHQUAKEE earthquake risk models. m ER IUK ZAaknin The table displays the provinces 0 Neg ibin FLODEARTHQUAKE at greatest normalized risk for A DRIATIC SEA ANNUAL AVERAGE OF ANNUAL AVERAGE OF ec prl I AFFECTED GDP (%) AFFECTED GDP (%) shown in the table, the prov- Medimurje 12 Grad Zagreb 6 ince at greatest risk of floods Vukovar-srijem 10 Zagreb 4 is Medimurje, and the one at DiIa I a Sisak-moslavina 7 Krapina-zagorje 3 g r Zagreb Varazdin 3 r tu Slavonski Medimurje 2 Grad Zagreb. In absolute terms, Brod-posav Dubrovnik- 2 the province at greatest risk of There sa high co relation Osijek-baranja neretva both floods and earthquakes is (r0.95) hetween Fe Lika-senj Zadar-knin 2 Grad Zagreb. population and GDP ofa Varazdin 4 Karlovac 1 Dubrovnieretva Karl0vac d vSibenik edpfrmin Sibenik m tKoprivnica 1 -mkrizaevkli Croatia Jft F L 0 0 D WORLDBANKGROUP Q|GFDRR ISK AIEE In the last 15 years, Croatia was hitaffected GDP is dominated by events by several floods, most ofthem that happen relatively frequently with relatively minor impacts. Ifthe impact ofa 100-year event is Flooding in 2014 killed three people much greater than that of a I0-year 1J and affected over 9,000. event, then less frequent events nake Modimurje HUNGARY a larger contribution to the annual This map depicts the impact of flood- average ofaffected ODE Thus, even ing on provinces' GDPs, represented ifa province's annual affected GDP SlVEN d n as percentages of their annual aver- seems small, less frequent and more age GDPs affected, with greater color intense events can still have large saturation indicating higher percent- impacts. ages. The bar graphs represent GDP affected by floods with return periods The annual avenge population affect- of 10 years (white) and 100 years ed by flooding in Croatia is about (black). The horizontal line across the 100,000 and the annual average bars also shows the annual average of affected GDP about$1 billion. Within GDP affected by floods. the various provinces, the 10- and KPr jorkK 100-year impacts do not differ much, When a flood has a 10-year return so relatively frequent floods have period, it means the probability of large impacts on these averages. SERBIA occurrence of a flood of that magni- tude or greater is 10 percent per year. A GOseno A 100-year flood has a probability On block = tr 40 of occurrence of 1 percent per year. This means that over a long period of BOSNIA ANO HERREGOINA time, a flood of that magnitude will, Zadar-knin I 0 on average, occur once every 100 Annla[ average years. It does not mean a 100-year ADRIATIC SEA flood will occur exactly once every 10-year 100-year 100 years. In fact, it is possible for a flood of any return period to occur Annua[ Average of Affected GOP more than once in the same year, or Split-dalmatija to appear in consecutive years, or not to happen at all over a long period of time. If the 10- and 100-year bars are the same height, then the impact of a 10- - Dubrovnik-n a tha thatE NE a yeear event is as large as that of a 100- year event, and the annual avenbge of CrO tiaWORLDBANKGROUP EL GF RROP AND ECENTRAL A5IA(ECA) roatia's worst earthquake took an earthquake of any return period place in 1667 in Dubrovnik, to occur more than once in the same with an estimated magnitude year, or to appear in Consecutive of 7.2. More than 3,000 people were years, or not to happen at all over a killed, and Dubrovnik (with 5,000 long period oftime. Medimurje HUNGARY homes at the time) was completely destroyed. If the same earthquake If the 10- and 100-year bars are the were to occur today, its estimated same height, then the impact of a death toll would be more than 1,500 10-year event is as large as that of and its damage over $7 billion. Other, a 100-year event, and the annual Grad Zagreb more recent earthquakes included average of affected GDP is dominat- one in 1927 in Slovac and another in ed by events that happen relatively Bjereb itogora 1962 in Podgora. frequently. If the impact of a 100- virI i a p a year event is much greater than that oie aaj This map depicts the impact of of a 10-year event, then less frequent earthquakes on provinces' GDPs, events make larger contributions to represented as percentages of their the annual average of affected GOPP annual average GDPs affected, with Thus, even if a province's annual Pio. o tir4osa Vakovar-srijem greater color saturation indicating affected GDP seems small, less fre- higher percentages. The bar graphs qunadmoeitsevnscn represent GDP affected by earth- still have large impacts. quakes with return periods of 10 A O years (white) and 100 years (black). The annual average population 10 and 100-year return periods The horizontal line across the bars affected by earthquakes in Croatia also shows the annual average of is about 100,000 and the annual av- GO fetdb atqae.erage affected GDP about $1 billion.BO AN'H7REOVA GDP affected by earthquakes.averages of falities and When an earthquake has a 10-year capital losses caused by earthquakes return period, it means the probabil- are about 20 and about $300 million, 10-year 100 year ity of occurrence of an earthquake of respectively The fatalities and cap- that magnitude or greater is 10 per- ital losses caused by more intense, cent per year. A 100-year earthquake less frequent events can be substan- has a probability of occurrence of tially larger than the annual averag- 1 percent per year. This means that es. For example, an earthquake with over a long period of time, an earth- a 0.4 percent annual probability of quake of that magnitude will, on occurrence (a 250-year return pen- average, occur once every 100 years. od event) could cause nearly 1,000 ML L HNEGR It does not mean a 100-year earth- fatalities and $5 billion in capital loss Dubrovnik-nee quake will occur exactly once every (about 10 percent of GOP). 100 years. In fact, it is possible for lf at aWORLDBANKGROUP E|GFDRRE AN CENTRAL A51A(ECA) ( EARTHQUAKE ,3 EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS (MILLIONS$) ANNUAL AVERAGE FATALITIES 47. he rose diagrams show the provinces with the potential .7 for greatest annual average capital losses and highest a- 0S annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital Dubrovniik-neretva Pnimorje-gorski S Varazdin 9 ora1 razdin 2 loss occurs in Grad Zagreb, which is not surprising, given the economic importance of the province. 6 EARTHQUAKE EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 he exceedance probability curves display the GDP T affected by, respectively, floods and earthquakes for 20 100 varying probabilities of occurrence. Values for two different time periods are shown- A solid line depicts the affected GDP for 2015 conditions. A diagonally striped band depicts 12 the range of affected GDP based on a selection of climate 2080 a 08. and socioeconomic scenarios for 2080. For example, if Cro- 73 0 o50 atia had experienced a 100-year return period flood event in 2015,the affected GDP would have been an estimated $4 billion. In 2080, however, the affected GDP from the same 2015 5 25 2015 type of event would range from about $9 billion to about $16 billion. If Croatia had experienced a 250-year earth- quake event in 2015,the affected GDP would have been 250 0 50 10 about $20 billion. In 2080, the affected GDP from the same Return period (years) Return period (years) type of event would range from about $50 billion to about 2 1 0.4 -4 $100 billion, due to population growth, urbanization, and tDbability 0)Pmobability( the increase in exposed assets. All historical data on floods and earthquakes are from D. Guha-Sapir, R. Below, and Ph. Hoyois, EMDAT International Disaster Database (Universit6 Catholique de Louvain, Brussels, Belgium), www.emdat.be, and J. Darietland A. Schae- ter, "Eastern Europe and Central Asia Region Earthquake Risk Assessment Country and Province Profiling," final report to GFDRR, 2014. Damage estimates for all historical events have been inflated to 2015 US$. WORLDBANKGROUP G D R ELUROPE AND CENTRAL AS1A(ECA) AFFECTED AFFECTED CAPITAL LOSS RIKPROF]LES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $20.3 billion*Gn Cyprus 0.ouain12 nlin ' C yprus's population and econo- by industry and agriculture making The table displays the provinces at my are exposed to earthquakes a small contribution. Cyprus's per greatest normalized risk for each per- and floods, with earthquakes capita GDP was $16,600. ii. In relative terms, as shown in the posing the greater risk of a high table, the province at greatest risk of impact, lower probability event. The This map displays GDP by province in floods is Gazimakusa (Famagusta), model results for present-day risk Cyprus, with greater color saturation and the one at greatest risk of earth- shown in this risk profile are based indicating greater GDP within a prov quakes is Famagusta. In absolute on population and gross domestic ince. The blue circles indicate the risk terms, the province at greatest risk product (GDP) estimates for 2015. of experiencing floods and the orange of floods is Gazimausa (Famagusta), The estimated damage caused by circles the risk of earthquakes in and the one at greatest risk of earth- historical events is inflated to 2015 terms of normalized annual average quakes is Nicosia. US dollars. of affected GD. The largest circles represent the greatest normalized Just over 70 percent of Cyprus's pop- risk The risk is estimated using flood ulation lives in urban environments. and earthquake risk models. The country's GDP was approximate- UN BUFFER ZONE Gazim4usa (Famagusta) ly US$20.3 billion in 2015, with over r 80 percent derived from services, Lefko a (Nicoiia Nicosia MEDITERRANEAN SEA most of the remainder generated TOP AFFECTED PROVINCES 0 Famagusta EARTHQUAKE m c b Annual Average of Affected GOP ANNUAL AVERAGE OF ANNUAL AVERAGE OF anc AFFECTED GDP (AFFECTED GOP UN BUFFER ZONE 0 Negligible GaziTagusa (Famagusta) d Famagusta by EARTHQUAKE Lefkosa (Nicosia) 0 Larniaca Girre (Kyrenia) C Gazimagusa (Famagusta) wt GOP (billions of $) Famnagusta D Paphos Larniaca i Lefkosa (Nicnsia) a There is a high correlation Limassol T Nicbsia 0 (r=.95) between the Nicosia Giroe (Kyre enia) a population and GtP oo r ag Paphts e Lionassol a province. aver WORLDBANKGROUP E|GDR ROPE AND ECENTRAL A51A(ECA) his map depicts the impact of event, then less frequent events make flooding on provinces' GDPs, a larger contribution to the annual represented as percentages of average of affected GDP. Thus, even their annual average GDPs affect- if a province's annual affected GDP ed, with greater color saturation seems small, less frequent and more indicating higher percentages. The intense events can still have large bar graphs represent GDP affected impacts. by floods with return periods of 10 years (white) and 100 years (black). The annual average population affect- The horizontal line across the bars ed by flooding in Cyprus is about 400 also shows the annual average of GDP and the annual average affected GDP affected by floods. about $4 million. Within the various provinces, little impact results from When a flood has a 10-year return floods with short return periods; period, it means the probability of thus, relatively infrequent floods have occurrence of a flood of that magni- large impacts on these averages. tude or greater is 10 percent per year A 100-year flood has a probability of occurrence of 1 percent per year This means that over a long period of time, a flood of that magnitude will, NGirne (Kyrenia) Gazimausa (Famausta) on average, occur once every 100 years. It does not mean a 100-year Lefkoga (Nicosia) Nicosia MEDITERRANEAN SEA flood will occur exactly once every 100 years. In fact, it is possible for a flood of any return period to occur more than once in the same year, or to appear in consecutive years, or not Nicosia Famagusta to happen at all over a long period of time. Paphos If the 10- and 100-year bars are the Larnaca same height, then the impact of a 10- UN BUFFER ZONE year event is as large as that of a 100- year event, and the annual average of Limasso Affected GOP I(%) for Annual Average of Affected GDP (%) affected GDP is dominated by events 10 and 100-year return periods that happen relatively frequently. One h[ock 1% If the impact of a 100-year event is Annual average 0 1 < r 69 much greater than that of a 10-year 10-year 100-year CypusWORLDBANKGROUP EL GFDRR EAND ECENTRAL A5IA(ECA) Cy ru @AkR RIS PROFILE yprus's worst earthquake since years, or not to happen at all over a about 5,000 and the annual average 1900 took place in 1953 in long period of time. affected GDP about $70 million. The Paphos, with a magnitude of annual averages of fatalities and cap- 6.5. The earthquake caused about If the 10- and 100-year bars are the ital losses caused by earthquakes are 40 fatalities. More recently, a 1995 same height, then the impact of a 10 less than one and about $10 million, earthquake caused two fatalities and year event is as large as that of a 100- respectively. The fatalities and capital nearly $7 million in damage. A major year event, and the annual average of losses caused by more intense, less earthquake occurred in 1222, causing affected GDP is dominated by events frequent events can be substantial- substantial damage and triggering a that happen relatively frequently. ly larger than the annual averages. tsunami. If the impact of a 100-year event is For example, an earthquake with much greater than that of a 10-year a 0.4 percent annual probability of This map depicts the impact of event, then less frequent events make occurrence (a 250-year return period earthquakes on provinces' GDPs, larger contributions to the annual av- event) could cause approximately represented as percentages of their erage of affected CDP. Thus, even if a $800 million in capital loss (about 4 annual average GDPs affected, with province's annual affected CUP seems percent of CUP). greater color saturation indicating small, less frequent and more intense higher percentages. The bar graphs events can still have large impacts. represent GDP affected by earth- quakes with return periods of 10 te l arae opulation years (white) and 100 years (black). The horizontal line across the bars U BUFFER ZONE Girne (Kyrenia Gazima usa (Famagusta) also shows the annual average of GDP affected by earthquakes. Lefko4a (Nicosia) *Nicosia MEDITERRANEAN SEA When an earthquake has a 10-year return period, it means the probabil- ity of occurrence of an earthquake of that magnitude or greater is 10 percent per year A 100-year earth- quake has a probability of occurrence of 1 percent per year. This means that over a long period of time, an Larnaca earthquake of that magnitude will, on UN BUFFER ZONE average, occur once every 100 ea rs. It does not mean a 100-year earth- quake will occur exactly once every GOP 100 years. In fact, it is possible for 0 ) rotuaffetedf an earthquake of any return period to occur more than once in the same year, or to appear in con ecutive 0 at al ove WORLDBANKGROUP E|G DR RO" AN CENTRAL A51A(ECA) EARTHQUAKE - P EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS Cs) ANNUAL AVERAGE FATALITIES OT he rose diagrams show the provinces with the potential f aT00for greatest annual average capital losses and highest Aa n n ut annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital loss occurs in Nicosia, which is not surprising, given the las0ca 0economic importance of the province. 4 )~L~ ~EARTHQUAKE EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 YJ'EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 T he exceedance probability curves display the GDIP .1 affected by, respectively, floods and earthquakes for 0.5 60 varying probabilities of occurrence. Values for two different time periods are shown. A solid line depicts the affected 0.4 soGDP for 2015 conditions. A diagonally striped hand depicts the range of affected GDP based on a selection of climate and socioeconomic scenarios for 2080. For example, if Cy- so prus had experienced a 100-year return period flood event in 2015, the affected GDP would have been an estimated 2 Jo$40 million. In 2080, however, the affected GDIP from the 20so 80same type of event would range from about $8 million to 2015< 10about $60 million. If Cyprus had experienced a 250-year earthquake event in 2015, the affected GDIP would have 1 ) 3 to 2 1) Too250 been about $7 billion. In 2080, the affected GDIP from the Eatun pr od(yers)Retun priod(yers)same type of event would range from about $20 billion to ....... about $50 billion, due to population growth, urbanization, i and the increase in exposed assets. All historical data on earthquakes are from the National Geophysical Data Center/World Data Service (NGDC/WDS), Significant Earthquake Database (National Geophysical Data Center, NOAA), doi:10289/V5TD9V7K, and . Daniell and A. Schaefer, "Eastern Europe and Central Asia Region Earthquake Risk Assessment Country and Province Profiling," final report to GFDRR, 2014. Damage estimates for all historical events have been inflated to 2015 US$. WORLDBANKGROUP G D R ELUROPE AND CENTRAL AS1A(ECA) AFFECTED AFFECTED CAPITAL LOSS RIKPROF]LES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE gGDP $183 billion*0 Qi Czech Repu b o . The Czech Republic's population small contribution. The Czech and economy are exposed to Republic's per capita GDP was earthquakes and floods, with $17,300. floods posing the greater risk. The mod- POLAND el results for present-day risk shown in This map displays GDP by this risk profile are based on population province in the Czech Republic, and gross domestic product (GDP) esti- with greater color saturation mates for 2015. The estimated damage indicating greater GDP within caused by historical events is inflated to a province. The blue circles nr 2015 US dollars. indicate the risk of experiencing floods and the orange circles Close to 75 percent of the population the risk of earthquakes in terms P of the Czech Republic lives in urban of normalized annual average of Vyhoocsk environments. The country's GDP was affected GDP. The largest circles StredoceIky approximately US$183 billion in 2015, represent the greatest normal- with 60 percent derived from services, ized risk. The risk is estimated Zapadoces , C r raky most of the remainder generated by using flood and earthquake risk industry; and agriculture making a models. The table displays the provinces TOP AFFECTED PROVINCES at greatest normalized risk for each peril. In relative terms, as hoceiky shown in the table, the province at greatest risk of floods is Stre- EARTHQUAKE docesky. and the one at greatest GERMANY ANNUAL AVERAGE OF ANNUAL AVERAGE OF risk of earthquakes is Severo-REPULIC AFFECTED GDP (%) AFFECTED GDP (%) moravsky. In absolute terms, Stredocesky f Severomoravsky 0 the province at greatest risk of AUSTR A Vychodocesky F Zapadocesky 0 floods is Praha, and the one at AnalAverage of Affected GDP (%) GOP (billions of $) Severocesky 3 Vychodocesky 0AnulAeaeoAfctdGPDP(iinsf lihocesky 3 Severacesky a greatest risk of earthquakes is Zapadocesky 2 Jihomoravsky 0 Severomoravsky. There is a high correlation lihomoravsky 2 Jihocesky 0 (r=0.95) between the Praha 2 Stredocesky 0 1 EARTHQUAKE 4r population and GDP of a Severomoravsky I Praha 0 province. o NegligibLe Czech R publicWORLDBANKGROUP Q |G DR I51K ANOFIEENRL5AEA he most devastating flood in more than once in the same year, or the Czech Republic since it to appear inconsecutive years, or not gained its independence in to happen at all over a long period of 1993 occurred in 2002. It killed 18 time. people and caused over $3 billion in damage. A 1997 flood caused 29 If the 10- and 100-year bars are the fatalities and almost $3 billion in same height, then the impact of a 10- damage. More recently, flooding in year event is as large as that of a 100- 2013 affected over 1 million people year event, and the annual average of and caused close to $850 million in affected GDP is dominated by events POLAND damage. Further floods in 2009 and that happen relatively frequently. 2010 caused over $150 million in If the impact of a 100-year event is damage per event. much greater than that of a 10-year event, then less frequent events make This map depicts the impact of flood- a larger contribution to the annual ing on provinces' GDPs, represented average of affected GDP. Thus, even as percentages of their annual aver- if a province's annual affected GDP age GDPs affected, with greater color seems small, less frequent and more saturation indicating higher percent- intense events can still have large ages. The bar graphs represent GDP impacts. affected by floods with return periods Z verom ravsky of 10 years (white) and 100 years (black). The horizontal line across the exposed to flooding in the Czech bars also shows the annual average of Republic is about 200,000 and the GDP ffeced b flods,annual average affected GDP about $4 GDP affected by floods. billion. For most provinces, in which When a flood has a 10-year return the impacts from 10- and 100-year period, it means the probability of floods do not differ much, relatively occurrence of a flood of that magni- frequent floods have large impacts tude or greater is 10 percent per year. on these averages. For the few in GERMA A 100-year flood has a probability which the 100-year impacts are much SL0VAK REPUBLIC of occurrence of 1 percent per year. greater than the 10-year impacts, less This means that over a long period of frequent events make a significant 10 and 100-year return periods time, a flood of that magnitude will, contribution to the annual average of AUSTRIA on average, occur once every 100 affected GDP. years. It does not mean a 100-year 00 flood will occur exactly once every 100 years. In fact, it is possible for a ArLalaverageI flood ohf any return pe ri od to occ ur 10-year 100-year Czec Rep blicWORLDBANKGROUP EL GF RROP AND ECENTRAL A5IA(ECA) az c Re umi * F R RIS[ PROFILES T he Czech Republic has expe- If the 10- and 100-year bars are the rienced several earthquakes same height, then the impact of of magnitude 7 in its history 10-year event is as large as that of a including one in 1786 in Tesin, one 100-year event, and the annual av- in 1872 in Gera, and one in 1901 in erage of affected GDP is dominated Trutnov. by events that happen relatively fre- quently. If the impact of a 100-year This map depicts the impact of event is much greater than that of earthquakes on provinces' GDPs, a 10-year event, then less frequent represented as percentages of their events make larger contributions to annual average GDPs affected, with the annual average of affected GDP. greater color saturation indicating Thus, even if a province's annual higher percentages. The bar graphs affected GDP seems small, less fre- represent GDP affected by earth- quent and more intense events can quakes with return periods of 10 years (white) and 100 years (black). The horizontal line across the bars The annual average population af- ae also shows the annual average of fected by earthquakes in the Czech * e Vychodocesky GDP affected by earthquakes. Republic is about 6,000 and the Pa t Stredocesky annual average affected GDP about . When an earthquake has a 10-year Whn neatquk hsa 0yer $100 million. The annual averag- $aaoesy~ everornoravsky return period, it means the prob- es of fatalities and capital losses Zapadocesky ability of occurrence of an earth- caused by earthquakes are less than quake of that magnitude or greater is 10 percent per year. A 100-year earthquake has a probability of losses caused by more intense, less lihomoravsky occurrence of t percent per year. frequent events can be substantially hhocesky This means that over a long period larger than the annual averages. of time, an earthquake of that mag- For example, an earthquake with nitude will, on average, occur once a GA percent annual probability every 100 years. It does not mean of occurrence (a 250-year return a 100-year earthquake will occur period event) could cause nearly Affected GP for AnnuaL Average of Affected GDP exactly once every 100 years. In $800 million in capital loss (about 1 10 and w0-year return periods fact, it is possible for an earthquake percent of GDP). AUSTRIA One block 4 of any return period to occur more than once in the same year, or to 2 appear in consecutive years, or not to happen at all over a long period Anua -Lh 1 of time. 10-year 100-year Cze h R pub icWORLDBANKGROUP E|GFDRR s"AND CENTRAL A51A(ECA) EARTHQUAKE EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS Cs) ANNUAL AVERAGE FATALITIES T he rose diagrams show the provinces with the potential for greatest annual average capital losses and highest e annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital loss occurs in Severomoravsky, which is not surprising, 00 oe ao,; given the economic importance of the province. on~Sk oo ooadsek EARTHQUAKE he exceedance probability curves display the GDP EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 Iaffected by, respectively, floods and earthquakes for 100 45varying probabilities of occurrence. Values for two different 100 oe45 e time periods are shown. A solid line depicts the affected S GDP for 2015 conditions. A diagonally striped band depicts 5 35 the range of affected GDP based on a selection of cli ate 208 a2 30 and socioeconomic scenarios for 2080. For example, if the 60 SCzech Republic had experienced a 100-year return period 20 flood event in 2015, the affected GDP would have been Ile10 2080QAK an eimaeda$20 pbillion. Inu2080dhowever the afetD EXCEE2015 10BA ILT GDPVE 2 1 AN 20 0E C E A C PR B BLT CU V , 0 5A D20 0afromcte b ,sam e t ey lofs v ntw ud erng ae s fo rb u 20 $40 billion to about $90 billion. If the Czech Republic had 5 4experienced a 250-year earthquake event in 2015, the 41;_ DP affected GDP would have been about $5 billion. In 2080, Ru 100iod0 10 p0 A00ars) the estimated affected GDP from the same type of event ~etun ferlo (yars)Retrn prio (yets)would range from about $10 billion to about $40 billion, 0 2 1 04 10 2 1 due populationgrowh, urbanization, andmthe increasein :Probablity (%) ..Probabilti (%) . - exoe ses 25 Czc Reu li had exposed assets yer eur ero All historical data on floods arnd earthquakes are from, respectively, D. Guhae-Sapir R? Belw, and Ph. Hoyois, EM-DATzInternationalfDisaterPDatabase (Uniiversite Catholique de Louvain, Brussels, Belgium), www.emndat.be, andl. Daniell arnd A. Schaefer, 'Eastern Europe and Central Asia Region Earthquake Rsk~ Assessment Country and Province Profiling," finalreport to GFJDRR, 2014. Darnage estimates for all historical eventslhave been inflated to 2015 US$. flo vn!i 05 h ffce D ol aebe WORLDBANKGROUP GR EROPE AND CENTRAL A5IA (ECA) AFFECTED AFFECTED CAPITAL LOSS Q GFDRR RI5K PROFILES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $23.6 billion* Estomia 0.ouain13mlin stonia's population and econo- and agriculture making a small con- my are exposed to earthquakes tribution. Estonia's per capita GDP and floods, with floods posing was $17,900. the greater risk. The model results for present-day risk shown in this risk pro- This map displays GDP by province in file are based on population and gross Estonia, with greater color saturation domestic product (GDP) estimates for indicating greater GDP within a prov- 2015. The estimated damage caused by ince. The blue circles indicate the risk historical events is inflated to 2015 US of experiencing floods and the orange dollars. circles the risk of earthquakes in terms of normalized annual average a Close to 70 percent of Estonia's pop- of affected GDR The largest circles 0 0 Raplamaa larvamaa ulation lives in urban environments. represent the greatest normalized The country's GDP was approximately risk The risk is estimated using flood agevmaa US$23.6 billion in 2015, with nearly 70 and earthquake risk models. percent derived from services, most of the remainder generated by industry, The table displays the provinces at o RU 5 5 IAN greatest normalized risk for each per- Saaremaa Parnumaa artuEDERATION il. In relative terms, as shown in the Viljandimaa table, the province at greatest risk of TOP AFFECTED PROVINCES floods is Parnumaa, and the one at 0 greatest risk of earthquakes is Tartu- o Potvamaa maa. In absolute terms, the province Vlgamaa EARTHQUAKE at greatest risk of both floods and ANNUAL AVERAGE OF ANNUAL AVERAGE OF earthquakes is Harjumaa. Vorumaa AFFECTED GDP (%) AFFECTED GDP (%) Parnumaa Tartumaa 0 Jogevamaa Raplamas 0 Tartumnaa Parnumnaa 0 Tda-rturaa Parrvumaa 0 Annual Average of Affected GOP ()GOP (billions of $ Ida-Virumas U Polvamnas 0 Harjumnaa 0 Hiiumaa 02 Hapjumaa Hari2rnaa 0 - -There is a high correlation ala amaa1 (r=0.95) between the La-ViruaaEARTHUAKE 9 population and GDP of a Laanrnaa 0 Idaa-Viruaa a EARTHQUAKEpovice Polvamaa 0 Laanemaa 0 Viljandimaa u Saaremaa 0 0 Negligible EstniaWORLDBANKGROUP GFDR ROPE AND ECENTRAL A51A(ECA) his map depicts the impact of event, then less frequent events make flooding on provinces' GDPs, a larger contribution to the annual represented as percentages of average of affected GDP. Thus, even their annual average GDPs affect- if a province's annual affected GDP ed, with greater color saturation seems small, less frequent and more indicating higher percentages. The intense events can still have large bar graphs represent GDP affected impacts. by floods with return periods of 10 years (white) and 100 years (black). The annual average population affect- The horizontal line across the bars ed by flooding in Estonia is about Ta also shows the annual average of GDP 6,000 and the annual average affect- affected by floods. ed GDP about $100 million. Within Harjumnaa the various provinces, the 10- and ZLaane-Virumaa :M When a flood has a 10-year return 100-year impacts do not differ much Ida-Virumaa period, it means the probability of so relatively frequent floods have occurrence of a flood of that magni- large impacts on these averages. r tude or greater is 10 percent per year - Raptamaa lr A 100-year flood has a probability HiiumaaL of occurrence of 1 percent per year. This means that over a long period of time, a flood of that magnitude will, on average, occur once every 100 L U S I A N years. It does not mean a 100-year Saaremaa Parnumaa Tartumaa FEDER AT ION flood will occur exactly once every Vi[jandima 100 years. In fact, it is possible for a flood of any return period to occur more than once in the same year, or Polvamaa to appear in consecutive years, or not Valgamaa to happen at all over a long period of time. Vorumaa Ifte 0 ad10-ea ar reteAffected GOP ()for Ann ualI Avera ge of Affecte d GODP() If the 10- and 100-year bars are the 1 n 0-errtr eid same height, then the impact of a 10- 1 n 0-errtr eid year event is as large as that of a 100- One block=1% 10 year event, and the annual average of 0 affected GDP is dominated by events that happen relatively frequently. Annual average 2 ATVIA If the impact of a 100-year event is much greater than that of a 10-year 10-year 100-year EstniaWORLDBANKGROUP GFDR ROPE AND ECENTRAL A5IA(ECA) E stonia has experienced to happen at all over a long period several modest earthquakes. of time. Its worst since 1900 took place in 1976 in Osmussaar. Earlier If the 10- and 100-year bars are earthquakes happened in 1602, the same height, then the impact 1670, and 1881, all in Narva. of a 10-year event is as large as that of a 100-year event, and the This map depicts the impact of annual average of affected GDP is earthquakes on provinces' GDPs, dominated by events that happen represented as percentages of their relatively frequently. If the impact Tallinn annual average GDPs affected, with of a 100-year event is much greater greater color saturation indicat- than that of a 10-year event, then Harjumaa ing higher percentages. The bar less frequent events make larger Laane-Virumaa graphs represent GDP affected by contributions to the annual average Ida-Virumaa earthquakes with return periods of affected GDP. Thus, even if a of 10 years (white) and 100 years province's annual affected GDP .... (black). The horizontal line across seems small, less frequent and Raplamanarvamaa lrm the bars also shows the annual more intense events can still have Hiiumaa Laanemaa average of GDP affected by earth- large impacts. Jogevamaa quakes. The annual average population af- When an earthquake has a 10-year fected by earthquakes in Estonia is Parnumaa return period, it means the prob- about 200 and the annual average T R USIAN Saaremaa TartumanO ability of occurrence of an earth- affected GDP about $3 million. The VLaandimaaaFEDERATI0N quake of that magnitude or greater annual averages of fatalities and is 10 percent per year. A 100-year capital losses caused by earth- -------- earthquake has a probability of quakes are less than one and about Polvama occurrence of t percent per year. $700,000, respectively. The fatal- Valgaimaa This means that over a long period ities and capital losses caused by of time, an earthquake of that mag- more intense, less frequent events Vorumaa nitude will, on average, occur once can be substantially larger than every 100 years. It does not mean the annual averages. For example, LATV A a 100-year earthquake will occur an earthquake with a 0.4 percent exactly once every 100 years. In annual probability of occurrence GDP (%) not affected for 10 and Annual Average of Affected GDP (%) fact, it is possible for an earthquake (a 250-year return period event) 1oo-year return periods of any return period to occur more could cause about $30 million in Annual average = 0 than once in the same year, or to capital loss (less than 1 percent of appear in consecutive years, or not GDP). Es on aWORLDBANKGROUP E|GFDRRE AN CENTRAL A51A(ECA) EARTHQUAKE EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS ($) ANNUAL AVERAGE FATALITIES 00 0 ihe rose diagrams show the provinces with the potential 0 4 s Zfor greatest annual average capital losses and highest annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital Searemaa 20,0 - i rumaa 7 000 Raplamaa 0 Laane-Virumaa 0 loss occurs in Harjumaa, which is not surprising, given the economic importance of the province. 0 -- - - - - - - - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - - - - - - -- -- - - - - - - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - - - - - - - ------------- ------ ------------ --- - - 4 ,m (~EARTHQUAKE EXCEEDANCE PROBABILITY CURVE. 2015 AND 2080 KJ. EXCEEDANCE PROBABILITY CURVE 2015 AND 2080 he exceedance probability curves display the GDP E affected by, respectively, floods and earthquakes for 2.5 0.6 varying probabilities of occurrence. Values for two different time periods are shown. A solid line depicts the affected 2.0 0. GDP for 2015 conditions. A diagonally striped hand depicts the range of affected GDP based on a selection of climate 20801. and socioeconomic scenarios for 2080. For example, if Esto- 0.3 nia had experienced a 100-year return period flood event in 1.0 2015, the affected GDP would have been an estimated $600 M ~2080 0.2 million. In 2080, however, the affected GDP from the same 2015 1 0.5 /=w type of event would range from about $1 billion to about $2 200.1 billion. If Estonia had experienced a 250-year earthquake event in 2015, the affected GDP would have been about $80 10 50 100 million. In 2080, the affected GDP from the same type of turn period (yeas) Return period (years) event would range from about $200 million to about $500 10 2 1 4 billion, due to population growth, urbanization, and the robaPbihty o robability (%) increase in exposed assets. All historical data on earthquakes are from J. Daniell and A. Schaefer, "Eastern Europe and Central Asia Region Earthquake Risk Assessment Country and Province Profiling," final report to GFDRR, 2014. Damage estimates for all histori- cal events have been inflated to 2015 US$. WORLDBANKGROUP G D R EUROPE AND CENTRAL AS1A(ECA) AFFECTED AFFECTED CAPITAL LOSS GQ U M RI5K PROF] LES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $13.7 billion* G eo rgila 0.Population 3.9 million* eargia's population and econo- t 0pretdrvdfo evcs ntrso omlzdana vrg my are exposed to earthquakes mos fterm.in rgneaed o fetdGP h ags ice and floods, with earthquakes byidsranagiutrmaig rpeettegetsnomlzdrs.AnlAvaeoffecdGP posing the greater risk of a high asalcnrbto.Gogaspr Ters setmtduigfodad1 impact, lower probability event. The cpt D a 350 atqaers oes model results for present-day risk ThsmpdslyGDbypoicin Tetbedslsteprvcsat1ERHUK shown in this risk profile are based Goga ihgetrclrstr- gets omlzdrs o ahprl on population and gross domestic 0Ngiil product (GDP) estimates for 2015. tinindiaingraerGDPwtia Inrltvtrm,ssh ninheab, The estimated damage caused by prvneThblecrlsidctth thprvneagetstikoflods historical events is inflated to 2015 rikoexeinigfodanth Tbls,adteoetgetstikofGP(linsf US dollars. just over half of Georgia's population o ohfod n atqae sTiii lives in urban environments. The country's GDP was approximately US$13.7 billion in 2015, with closeAbhaautRe.USINFDATO G eorgia-to 70 percent derived from services, in terms of normalized annual aveage most of the remainder generated of affected GDP. The largest circles aL iby industy and agriculture making represent the greatest normalized risk. aal posig th greterriskofa igha small contribution. Georgia's per The risk is estimated using flood and 1 nowcapita G was $3,500. earthquake risk models. show in hisriskprofle re bsed This map displays GDP by province in Thtaldipysheroncst aGeorgia, with greater color satura- producttiD)estimate . tion indicating greater GDP within a I the atprovince. The blue circles indicate the hrisk of experiencing floods and the S sorange circles the risk of earthquakes e i lte term,le dpaysithe atreatnest rikTeeiathg r[t Zemo~~~~~1=.5 between thet Msht-Mint Kv Kl Just over hlf of Georga's populaton 4 orebthelst nlzd erthqkes Tahpeils. pp[to n O IIeni reI tee trmseasishownintheptabl country'sce aDPrewtesariprooifloodli US$13.7 billioniin2015, with closetStreatestfEriskTof eatqae sKvemo Kartly In abso-At.Re. KakheiAbkhaz . Rep. 0 h -LnnuahumieandeKofmAf(lower)GDPa(e) 1100 EARTHQUAKE AFFECTEDGDP (%) FFECTEDGlPli%)sLACK SE Thereiaihibehocorre[atio Racha-Leckhumi anopulation 0andtGDP$iofea MtskhRacha-irnehi 2iZeno (upper)lSwer)ti aneai Samtskhe-jvakhetir2Samtsahd-aZekhetup cS -Leche huand 2 Ka ti 61 B L A C KiSE A i M rsktianet Kvemo (pper) Svaneti Sha-tarti 6 Mketateti i ZIeet (upr va ib Kvm atiIAdjara Aut. Rep. GuiA Kakheti Geo giaWORLDBANKGROUP E|G DR ROPE AND EENTRAL A51A(ECA) he most devastating floods in When a flood has a 10-year return to happen at all over a long period of The annual avenge population affect- Georgia since it gained its in- period, it means the probability of time. ed by flooding in Georgia is about dependence in 1991 occurred occurrence ofa flood of that magni- 100,000 and the annual average in 1997. In that year, Georgia was hit tude or greater is 10 percent per year If the 10- and 100-year bars are the affected GDP about $400 million. by two floods, which together caused A 100-year flood has a probability same height, then the impact of a 10 Within the various provinces, the 10- 7 fatalities and over $40 million in of occurrence oft percent per year. year event is as large as that ofa 100 and 100-year impacts do not differ damage. Flooding in 2012 caused less This means that over a long period of year event, and the annual average of much, so relatively frequent floods damage ($3 million), but it affected time, a flood of that magnitude will, affected GDP is dominated by events have large impacts on these avenges. over 100,000 people. Flooding in on average, occur once every 100 that happen relatively frequenly. 2013 affected close to 25,000 people years. It does not mean a 100-year If the impact of a 100-year event is but also caused limited damage. flood will occur exactiy once every much greater than that of a 10-year Affected GOP (%) for Other floods occurred in 1995, 2004, 100 years. In fact, it is possible for a event, then less frequent events make 10 and 100year return periods 2005, and 2011, with fewer than flood of any return period to occur a larger contribution to the annual 2% 2,500 people affected and less than more than once in the same year, or avenge of affected GDP. Thus, even $4 million in damage per event. to appear in consecutive years, or not if a province's annual affected GDP average of Affected GOP According to a 2015 World Bank seems small, less frequent and more Annual Post Disaster Needs Assessment intense events can still have large (in press), the June 2015 flooding impacts. 10-year 100-year in Tbilisi caused 19 fatalities (in addition, three people are still miss- ing), affected over 700 people, and caused over $20 million in damages. Ti All these events highlight Georgia's vulnenebility to floods. They are not always devastating, but they follow eafch other quccly aned have a large 100 yers. Infact,itSisnossibllforn cumulative effect on the countthersa BLACK SEA This map depicts the impact of flood- ing on provinces' GDPs, represented t ht.,fini as percentages of their annual aver- age GDPs affected, with greater colorimrnhcKrti<-. saturation indicating higher percent- time.i ages. The bar graphs represent GDP Aa" affected by floods with return periods hi LML Kakheti of 10 years (white) and 100 years ae (black). The horizontal line across the bars also shows the annual avegge ofo GDP affected by floods. Gei WORLDBANKGROUP E GROP AE CENTRALA5IA(ECA) eorgia's worst earthquake years, or not to happen at all over a The annual average population affect- since 1900 occurred in long period of time. ed by earthquakes in Georgia is about 1991, with a magnitude of 300,000 and the annual average 7. It caused over 250 fatalities and If the 10- and 100-year bars are the affected GDP about $900 million. The close to $3 billion in damage. An same height, then the impact of a 10 annual averages of fatalities and cap- earthquake in 2002 affected nearly year event is as large as that of a 100 ital losses caused by earthquakes are 20,000 people and caused about year event, and the annual average of about 500 and about $500 million, $500 million in damage. The impact affected GDP is dominated by events respectively The fatalities and capital of earthquakes in 1992 and 2009 was that happen relatively frequently. losses caused by more intense, less less extensive. If the impact of a 100-year event is frequent events can be substantial- much greater than that of a 10-year ly large r than th e ann ual ave rage s. This map depicts the impact of event, then less frequent events make For example, an earthquake with Affected GDP for earthquakes on provinces' GDPs, larger contributions to the annual av- a 0.4 percent annual probability of 10 and 100-year return periods represented as percentages of their erage of affected GDP. Thus, even if a occurrence (a 250-year return period One h[ock= 10% 100 annual average GDPs affected, with province's annual affected GDP seems event) could cause about 20,000 greater color saturation indicating small, less frequent and more intense fatalities and $7 billion in capital loss Annual Average of Affected GDP higher percentages. The bar graphs events can still have large impacts. (about 50 percent of GDP). Annual average represent GDP affected by earth- quakes with return periods of 10 years (white) and 100 years (lack). The horizontal line across the bars also shows the annual average of GoP affected by earthquakes. domiate byD When an earthquake has a 10-year return period, it means the probabil- ity of occurrence of an earthquake oef that magnitude or greater is 10 n percent per year A 100-year earth-BLCSE quake has a p robability of ccu rren ce osfl pelrcent pfer year. Thmis mieans that over a long period of time, an earthquake of that magnitude will, on average, accuvr once evefry 100 yea rs. It does not mean a 100-year earth- quake will occur exactly once every 100 years. In fact, it is possible for an earthquake of any return period to occur more than once in the same Frieisi year, oer to appear in con ecutive GeogiaWORLDBANKGROUP E|GDR RO"" AND CENTRAL A51A(ECA) EARTHQUAKE EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS (MILLIONS $) ANNUAL AVERAGE FATALITIES T he rose diagrams show the provinces with the potential for greatest annual average capital losses and highest annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital Smk JkhAbkhazia Aut Re 0 Kakh 5 loss occurs in Tbilisi, which is not surprising, given the economic importance of the province. EARTHQUAKE EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 he exceedance probability curves display the GDP Taffected by, respectively, floods and earthquakes for 9 80 varying probabilities of occurrence. Values for two different 8 time periods are shown. A solid line depicts the affected 2080 GDP for 2015 conditions. A diagonally striped hand depicts 6 ;2 2the range of affected GDP based on a selection of climate 980 and socioeconomic scenarios for 2080. For example, if 5 40 Georgia had experienced a 100-year return period flood 4 o event in 2015, the affected GDP would have been an esti- 3 mated $1 billion. In 2080, however, affected GDP from the 2015 2 20 same type of event would range from about $6 billion to 1 _2015 about $8 billion. If Georgia experienced a 250-year earth- quake event in 2015, the affected GDP would have been 10 50 100 20 10 50 100 250 about $10 billion. In 2080, the affected GDP fmm the same Re turn period (years) Return period (years) type of event would range from about $50 billion to about ........ $70 billion, due to population growth, urbanization, and the Probability() Probability (increase in exposed assets. All historical data on floods, unless otherwise noted, and earthquakes are from, respectively, D. Guha-Sapir, R. Below, and Ph. Hoyols, EM-DAT: International Disaster Database (Universit6 Catholique de Louvain, Brussels, Belgium). www.emdat.be, and the National Geophysical Data Center/World Data Service (NGDC/WDS), Significant Earthquake Database (National Geophysical Data Center, NOAA) doi:10.7289/V5TD9V7K. Damage estimates for all historical events have been inflated to 2015 US$. WORLDBANKGROUP G D R EUROPE AND CENTRAL AS1A (ECA) AFFECTED AFFECTED CAPITAL LOSS RIKPROF]LES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $192 billion* E3 Greece 0 P IJ LG A AR I A G reece's population and economy the remainder generated by indus- are exposed to earthquakes and try, and agriculture making a small FYR OF MACEDONIA BLACK SEA floods, with earthquakes posing contribution. Greece's per capita the greater risk of a high impact, lower GDP was $17,800. probability event. The model results for present-day risk shown in this risk pro- This map displays GDP by provinc in reee, ithgrate coor atua-Annu al Averag of Affected GDP(% file are based on population and gross in Gec w g c r a domestic product (GDP) estimates for lion indicating greater GOP within FO 2015. The estimated damage caused by a province. The blue circles idi- Dytiki Makedonia historical events is inflated to 2015 US cate the risk of experiencing floods EARTHQUAKE dollars. and the orange circles the risk of earthquakes in terms of normalized oNgiil Just over 60 percent of Greece's pop- annual average of affected GOP The ulation lives in urban environments. largest circles represent the greatest There is a high correlation The country's GDP was approximately normalized risk The risk is estimat- (r=0.5) between the US$192 billion in 2015, with close to 80 ed using flood and earthquake risk population and GDP of a percent derived from services, most of models. province. The table displays the provinces at LGP (billions of $) TOP AFFECTED PROVINCES greatest normalized risk for each peril. In relative terms, as shown in ithe table, the province at greatest Dyti9don risk of floods is Anatoliki Makedonia, 000k DEARTHQUAKE and the one at greatest risk of earth- ANNUAL AVERAGE OF ANNUAL AVERAGE OFnole AFFECTED GaP %)gAFFECTED GOP (%)a d Te a terms, the province at greatest risk NoliPgaio Aatoliki Makedoria IDytiki Ellada 4 offloods isKentriki Makedonia, and Kai Thraki Ipeiros 3the one at greatest risk of earth - Thessalia r Voreio Aigaio 2 Ipeiros e lodsioi Nisoi 2 quake is Attiki. Kerntriki Makedonia f Kentriki Makedonia 2 Dytiki Ellada I Aniatoliki Makedornia 2 Peloporiisos Kai Thraki Dytiki Makedonia ofSterea Ellada 2 Sterea Ellada oThessalia 2 Attiki PeloponneeiSOS 2 P iE Ds lomIioi Nisoi o Attiki 1 GreceWORLDBANKGROUP EL GFDRPE" AND EENTRAL A51A(ECA) he most damaging floods ina ffected GDP is dominated by events Greece since 1900 occurred in that happen relatively frequently 1994 and 2003, causing over If the impact of a 100-year event is $700 million and $800 million in much greater than that of a 10-year damage, respectively. event, then less frequent events make U ARIA a larger contribution to the annual This map depicts the impact of flood- BLACK SEA ing on provinces' GDPs, represented if a province's annual affected GDP as percentages of their annual aver- seems small, less frequent and more age GDPs affected, with greater color saturation indicating higher percent- impacts. A ages. The bar graphs represent GDP affected by floods with return periods The annual avenge population affect- of 10 years (white) and 100 years ed by flooding in Greece is about (black). The horizontal line across the 50,000 and the annual avenge affect- Afftad 10e r bars also shows the annual average of ed GDP about $600 million. Within GDP affected by floods. the various provinces, the 10- and 6 100-year impacts do not differ much, Thealia . When a flood has a 10-year return3 Whn lod a a1-yarrtun so relatively frequent floods have Arnnua[I average period, it means the probability of large impacts on these averages. occurrence of a flood of that magni- -year -year tude or greater is 10 percent per year. VorcioAigaio A 100-year flood has a probability lonioi Nisoi Sterea Eada of occurrence of 1 percent per year. This means that over a long period of time, a flood of that magnitude will, Attik on average, occur once every 100 years. It does not mean a 100-year flood will occur exactly once every 100 years. In fact, it is possible for a flood of any return period to occur NotiaAigaic more than once in the same year, or to appear in consecutive years, or not to happen at all over a long period of time. If the 10- and 100-year bars are the I same height, then the impact of a 10- -0Kr- year event is as large as that of a 100- a year event, and the annual avengge of WORLDBANKGROUP EL GFR ROP AND ECENTRAL A5IA(ECA) reece's worst earthquake of time, an earthquake of that mag- events can be substantially larger since 1900, with a magni- nitude will, on avenge, occur once than the annual avenges. For exam- tude of 7.2, took place in every 100 years. It does not mean ple, an earthquake with a 0.4 percent 1953 in Kefalonia and caused over a 100-yearearthquakewill occur annualprobabilityof occurrence (a 450 fatalities. Many people left the exactly once every 100 years. In 250-year return period event) could U iG A R IA island after the event, reducing its fact, it is possible for an earthquake cause nearly 2,000 fatalities and $20 population to a mere 20 percent of any return period to occur more billion in capital loss (about 8 percent BLACK SEA of its size before the disaster. The than once in the same year, orto of GDP). same region was also hit by earth- appear in consecutive years, or not TURKEY quakes in 1867 and 2011. A 1999 to happen at all over a long period earthquake in Athens caused close of time. to 150 deaths and over $6 billion in damage. More recently, in 2014, If the 10- and 100-year bars are an earthquake in southern Greece the same height, then the impact Affected GOP for caused three fatalities and almost 10 and 100-year return periods $500 million in damage. that ofa 100-year event, and the One block = 10% annual average of affected GDP is 6 This map depicts the impact of dominated by events that happen earthquakes on provinces' GDPs, relatively frequently. If the impact Annual average represented as percentages of their of a 100-year event is much greater annual average GDPs affected, with than that of a 10-year event, then greater color saturation indicat- less frequent events make larger IonoiNiso ing higher percentages. The bar contributions to the annual average graphs represent GDP affected by of affected GOP. Thus, even if a Annual Average of Affected GDP (%) earthquakes with return periods province's annual affected GDP of 10 years (white) and 100 years seems small, less frequent and (black). The horizontal line across more intense events can still have the bars also shows the annual large impacts. average of GDP affected by earth- quakes. The annual average population affected by earthquakes in Greece -NotiRikigaic - When an earthquake has a 10-year is about 200,000 and the annual av- return period, it means the prob- enge affected GDP about $3 billion. ability of occurrence of an earth- The annual averages of fatalities quake of that magnitude or greater and capital losses caused by earth- is 10 percent per year. A 100-year quakesareabout50andabout earthquake has a probability of $700 million, respectively. The occurrence of 1 percent per year. fatalities and capital losses caused This means that over a long period by more intense, less frequent mg e WORLDBANKGROUP E|GDR ROPE AND CENTRAL A51A(ECA) EARTHQUAKE EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS (MILLIONS$) ANNUAL AVERAGE FATALITIES he rose diagrams show the provinces with the potential ;06.for greatest annual average capital losses and highest annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital riti 20 Dy ki Ellada 70 ilonioi Nisai 1 Krloss occurs in Attiki, which is not surprising, given the eco- nomic importance of the province. i ) EARTHQUAKE h xednepoaiiycre ipa h D EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 T affected by, respectively, floods and earthquakes for 12 800 varying probabilities of occurrence. Values, for two different Z, time periods are shown. A solid line depicts the affected 10 10 GDP fQr 2015 conditions. A diagonally striped band depicts 20800 2080 8 2the range of affected GDP based-on a selection of climiate 8 and Socioeconomic scenarios for 2080. For example, if 6 40 208 Z.J Greece had experienced a 100-year return period flood 1event in 201, the affected GOPwould have-been an esti 2015 4mated $3 bdEonXn 2080, however, the affected GP from 200 '~the samefi type of event would range from about $5 billionl /201 20r to about $10 billion. If Greece had experienced a 250-year earthquake event in 2015, the affected GDP would have been about $90 billion. In 2080, the affected GDP from the same type of event would range rom about $200 billion to about $800 billion, due to population growth, urbanization, to about$10and the inc rease in expos ened assets. been.... about.. $90 bilin In.. 208..heafecedGD.fomth All historical data on floods and earthquakes are from D. Guha-Sapir, R. Below, and Ph. Hoyois, EM-DAT: International Disaster Database (Universit6 Catholique de Louvain, Brussels, Belgium), www.emdat.be; the National Geophysical Data Center/World Data Service (NGDC/WDS), Significant Earthquake Database (National Geophysical Data Center, NOAA), doi:10.7289/V5TD9V7K; and L Daniell and A. Schaefer, "Eastern Europe and Central Asia Region Earthquake Risk Assessment Country and Province Profiling," final report to GFDRR, 2014. Damage estimates for all historical events have been inflated to 2015 US$ WORLDBANKGROUP EI GROPE AND CENTRALA5IA (ECA) AFFECTED AFFECTED CAPITAL LOSS Q GFDRR RI5K PROFILES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $126 billion* Hunga ry 0.Population 9.8 million* SLOVAK REPUBLIC ungary's population and to 70 percent derived from services, table, the province at greatest risk is Csongrad, and the one at greatest economy are eposed to most of the remainder generated by of floods is Csongrad, and the one at risk of earthquakes is Budapest. earthquakes and floods, with industry and agriculture making a greatest risk of earthquakes is Koma- R floods posing the greater risk. The small contribution. Hungary's per rom-esztergom. In absolute terms, model results for present-day risk capita GDP was $12,800. the province at greatest risk of floods shown in this risk profile are based on population and gross domestic This map displays GDP by provnce in A Borsodabauj-7emlen product (GDP) estimates for 2015. Hungary, with greater color satura- Buap & The estimated damage caused by Szabolci-s7atmar-bereg historical events is inflated to 2015 a province. The blue circles indicate US dollars. the riskofexperieningfloods and Heves the orange circles the risk of earth- 6yo -io n-5 on n s About 70 percent of Hungary's pop- quakes in terms of normalized annual Hdi ulation lives in urban environments. average of affected GDP. The largest The country's GDP was approximate- circles represent the greatest normal- -nagykun-szolnok ly US$126 billion in 2015, with close ized risk. The risk is estimated using Vas Veszpreme flood and earthquake risk models. N do a r TOP AFFECTED PROVINCES The table displays the provinces at greatest normalized risk for each per- Zalar i. In relative terms, as shown in the nagkunison Sonogy onac sk n EARTHQUAKE ANNUAL AVERAGE OF ANNUAL AVERAGE OF AFFECTED GDP (%) AFFECTED GDP (%)ROMA Baranya Csongrad 1 Komarom-esztergom 2 CR0ATIA lasz-nagykun-szolnok 13 Budapest I SERB A Gyor-moson-sopron Pest I Bekes Zala I Szabocs- Veszprem i Annual Average of Affected GOP (%) GDP (billions of $) szatmar-bereg Heves I Teves 2 No rad 1 20 There is a high correlation Borsod-abauj- lasz-nagykun-szolnok 1 EARTHQUAKE I (r=95) betwee the zemoplen Vas 1ATQAEl 1 d I d population anid GOP of a Bacs-kiskun province. Hajdu-bihar o Negligible Hun aryWORLDBANKGROUP E|G DR ROPE" AND EENTRAL A51A(ECA) POLAND T he most deadly flood in Hun- flood will occur exactly once every gary since 1900 took place in 100 years. In fact, it is possible for a 1970 and caused about 300 flood ofany return period to occur CZ ELH REPUBLIC fatalities and over $500 million in more than once in the same year, or damage. More recently, in 1999, two to appear in consecutive years, or not floods occurred that together caused to happen at all over a long period of at least eight fatalities, affected over time. 100,000 people, and brought over E $400 million in damage. A single If the 10- and 100-year bars are the flood in 2010 caused no fatalities same height, then the impact of a 10- but almost $500 million in damage. year event is as large as that of a 100- These statistics highlight the lives year event, and the annual avenge of being saved by disaster risk manage- affected GDP is dominated by events Budapest ment efforts but also the possibility that happen relatively frequently. that the damage associated with Ifthe impact ofa 100-year event is flooding will rise. much greater than that of a 10-year S Hee This map depicts the impact of flood-fevent,tthennlessaf This map ep astheimpacto fflod - a larger contribution to the annual ing on provinces' GDPs, represented average of affected GDP. Thus, even as percentages of their annual aver- if a province's annual affected GDP age GDPs affected, with greater color seems small, less frequent and more Vai Veszprem Fejer saturation indicating higher percent- intense events can still have large ages. The bar graphs represent GDP impacts. affected by floods with return periods of 10 years (white) and 100 years The annual avenge population affect- Zaa (black). The horizontal line across the ed by flooding in Hungary is about bars also shows the annual average of 200,000 and the annual average somogy GDP affected by floods. affected GDP about $2 billion. Within th e variou s provinces, th e 10 - an d ROMA When a flood has a 10-year return 100-year impacts do not differ much, Baranya period, it means the probability of so relatively frequent floods have occurrence of a flood of that magni- large impacts on these averages. CR0ATIA SERB tude or greater is 10 percent per year. A 100-year flood has a probability Affected GDP for Annual Average of Affected GDP of occurrence of 1 percent per year. 10 and 100year return periods One h[ock = 5% This means that over a long period of 50 time, a flood of that magnitude will, 30 0 F 6 & on average, occur once every 100 Annual averageA 10 years. It does not mean a 100-year itN iNs possibe f a 10-year 100-year HWORLDBANKGROUP EL GROPE AND EENTRALA5A (ECA) POLAND H ungary's worst earthquake to happen at all over a long period since 1900 took place in of time. 1911 in Kecskemet, causing CZECH REPUBLIC 10 fatalities. Others occurred in If the 10- and 100-year bars are 1599, 1763, 1783, and 1879, and, the same height, then the impact most recently, in 2011. of a 10-year event is as large as that of a 100-year event, and the This map depicts the impact of annual average of affected GDP is I F AN earthquakes on provinces' GDPs, dominated by events that happen represented as percentages of their relatively frequently. If the impact annual avenage GDPs affected, with of a 100-year event is much greater greater color saturation indicat- than that of a 10-year event, then ing higher percentages. The bar less frequent events make largera graphs represent GP affected by contributions to the annual aver- zabocs-szatmar-bereg earthquakes with return periods age of affected GDP. Thus, even if of 10 years (white) and 100 years a province's annual affected GP (black). The horizontal line across seems small, less frequent and Gyor-rnowon-sopron .ca Bdpt the bars also shows the annual more intense events can still have average of GDP affected by earth- large impacts. quakes. The annual average population When an earthquake has a 10-year affected by earthquakes in HungaryYO7PI Foj return period, it means the prob- is about 80,000 and the annual ability of occurrence of an earth- average affected GP about $1 quake of that magnitude or greater billion. The annual averages of is 10 percent per year. A 100-year fatalities and capital losses caused earthquake has a probability of by earthquakes are about one and w aongrad ccurrence of p ercent per year. about $200 million, respectively. This mean s th at over a l oIng p erio d The fatalities and capital losses ROMA of time, an earthquake of that mag- caused by more intense, less fre- nitude will, on average, occur once quent events can be substantially every 100 years. It does not mean Ilarge rtha n the an nu al averages. CROATI A S E RB A a 100-year earthquake will occur For example, an earthquake with exactly once every 100 years. In a . percent annual probability Af fectd greae () f rra f d is 100percent0perayear.uAn100-year fact, it is possible for an earthquake of occurrence (a 250-year return One block = 10% of any return period to occur more period event) could cause about than ne in the same year, or to $6 billion in capital loss (about S appear in consecutive years, or not percent of GP). Ariua[ average 20 BOSNI AND NAZCSO A 10-year 100-year HunaryWORLDBANKGROUP E|GDR ROPE AND CENTRAL A51A(ECA) EARTHQUAKE EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS (MILLIONS $) ANNUAL AVERAGE FATALITIES 00 6 0he rose diagrams show the provinces with the potential &6 T for greatest annual average capital losses and highest o annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital Komarom-esztergiom Komaram. Fejer 4 oBekes 0 loss occurs in Budapest, which is not surprising, given the economic importance of the province. 4, 0 EARTHQUAKE he exceedance probability curves display the GDP EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 he ecedancepectily floods dieathe for T affected by, respectively, floods and earthquakes for 45 350 varying probabilities of occurrence. Values for two different 40 time periods are shown. A solid line depicts the affected 300 GDP for 2015 conditions. A diagonally striped hand depicts 2 250 the range of affected GDP based on a selection of climate 2080 -2 and socioeconomic scenarios for 2080. For example, if Hun- 25 I 200 gary had experienced a 100-year return period flood event 150 2in 2015, the affected GDP would have been an estimated $9 15 billion. In 2080, however, the affected GDP from the same 2015 100 2 10dtype of event would range from about $10 billion to about 50 $40 billion. If Hungary had experienced a 2 50-year earth- 2015 quake event in 2015, the affected GDP would have been 10 50 10 50 100 250 about $50 billion. In 2080, the affected GDP from the same Return period (years) Return period (years) type of event would range from about $80 billion to about 10 10 2 1 o4 $300 billion, due to population growth, urbanization, and Probability (%) Probability (%) the increase in exposed assets. All historical data on floods and earthquakes are from, respectively, D. Guha-Sapir, R. Below, arid Pih. Hoyois, EMDAT: International Disaster Database (Universite Catholique de Louvain, Brussels, Belgium), www.emdat.be, and I. Daniel[ and A. Schaefer, "Eastern Europe and Central Asia Region Earthquake Risk Assessment Country and Province Profiling,' final reportto GFDRR, 2014. Damage estimates for all historical events have been inflated to 2015 US$. WORLDBANKGROUP EI GROPE AND CENTRAL A5IA (ECA) AFFECTED AFFECTED CAPITAL LOSS QRI5KPROFILES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $120 billion* Kazakhstan '.Pplto 77mlin RUSSIAN FEDERATION K azakhstan's population and industry and agriculture making a peril. In both relative and absolute economy are exposed to earth- small contribution. Kazakhstan's per terms, the province at greatest risk quakes and floods, with floods capita GDP was $6,770. of floods is Atyrauskaya, and the one posing the greater risk. The model atgreatest risk of earthquakes is the results for present-day risk presented This map displays GDP by province in Almaty City area. in this risk profile are based on pop- Kazakhstan, with greater color satu- ulation and gross domestic product ration indicating greater GDP within (GDP) estimates for 2015. The esti- a provnce. The blue circles indicate -.idistanskaya mated damage caused by historical the risk of experiencing floods and events is inflated to 2015 dollars. the orange circles the risk of earth- Vt6darskaya quakes in terms of normalized annual Kustanayska isal Just over half of Kazakhstan's popula- average of affected GDP. The largest tion lives in urban environments. The circles represent the greatest normal- a country's GDP was approximately ized risk- The risk is estimated using US$120 billion in 2015, with close flood and earthquake risk models. Aktyubinskaya Vostochn kazachstanskaya to 70 percent derived from services, most of the remainder generated bysat greatest normalized risk for each izedyrik.dThekiskaisestimatdsusin TOP AFFECTED PROVINCES 4K Jamylkaya A Yujnokathtanskaya - ~ M N OLI Akty binskaya A Vostocn-0 hty a nrkay EARTHQUAKE ARMENIA ANNUALAVERAGEOF ANNUALAVERAGEOF AZERBAIJAN AFFECTED GDP (%) AFFECTED GDP (%) j 15UR MENISTAN fITAN CHINA Atyrauskaya 11 Almaty City area 4 Kyzylordinskaya 5 Almatinskaya 3 Zapadrio-kazachstanskaya 3 Jambylslkaya 2 Annual Average of Affected GDP (%) GDP (billions of$) Jambyls[kaya 2 Yujno-kazachstanskaya 1 Yujno-kazachstanskaya 2 Vostochno-kazachstanskaya 1 10 There is a high correlation Vostochno-kazachstanskaya 2 Mangistauskaya 0 5 (r=0.95) between the Severo-kazachstanskaya I Pavlodarskaya 0 1 EARTHQUAKE , - population and GDP of a Akmolinskaya I Kyzylordinskaya 0 1 E Q province. Aktyubinskaya 1 Severo-kazachstanskaya 0 Kustanayskaya I Akmolinskaya 0 0 Negligible Kaz khs anWORLDBANKGROUP E|G DR [ROPE AND EENTRAL A51A(ECA) he most deadly flood since more than once in the same year, or Kazakhstan gained its inde- to appear in consecutive years, or not pendence in 1991 occurred to happen at all over a long period of in 2010. It caused over 40 fatalities time. and close to $40 million in damage. RUSSIAN FEDERATION The most damaging flood took place If the 10- and 100-year bars are the in 2008, causing one death and over same height, then the impact of a 10- $100 million in damage. A 1993 flood year event is as large as that of a 100- caused approximately 10 fatalities year event, and the annual average of and close to $60 million in damage. affected GDP is dominated by events Flooding in 2011 caused only two that happen relatively frequently fatalities and damage close to $70 If the impact of a 100-year event is million. much greater than that of a 10-year event, then less frequent events make This map depicts the impact of flood- a larger contribution to the annual ing on provinces' GDPs, represented average of affected GDP. Thus, even as percentages of their annual aver- if a province's annual affected GDP age GDPs affected, with greater color seems small, less frequent and more saturation indicating higher percent- intense events can still have large Zapadno-kazachstanskaya ages. The bar graphs represent GDP impacts. affected by floods with return periods of 10 years (white) and 100 years The annual avenge population affect- (black). The horizontal line across the ed by flooding in Kazakhstan is about bars also shows the annual average of 300,000 and the annual average GDP affected by floods. affected GDP about $3 billion. For K sk most provinces, the 10- and 100- Mangistauskayalabsky When a flood has a 10-year return year impacts do not differ much, so period, it means the probability of relatively frequent floods have large occurrence of a flood of that magni- impacts on these avenges. For the SEA KYITI Z REPURLIC tude or greater is 10 percent per year. few in which the 100-year impacts A 100-year flood has a probability are much greater than the 10-year T ENISTAN of occurrence of 1 percent per year. impacts, less frequent events make a NACIHISNN Thsmen ta ve ln ero o iniian otrbtint teanulAffected GDP (%) for /Annual Average of Affected GOP(% INA This means that over a long period ofPers time, a flood of that magnitude will, averages of affected GD. One blck = on average, occur once every 100 years. It does not mean a 100-year 05 flood will occur exactly once every I verLge 100 years. In fact, it is possible for a Annual flood of any return period to occur 10-year 100-year Ka z khs a nWORLDBANKGROUP EL GF R ROP AND ECENTRAL A5IA(ECA) K azakhstan's worst earthquake year. or to appear in consecutive since 1900 took place in 1911 years, or not to happen at all over in Kemin, with a magnitude of long period of time. 7.7. The earthquake caused over 450 fatalities and more than $20 million Ifthe 10- and 100-year bars are the RUSSIAN FEDERATION in damage. Other earthquakes oc- same height, then the impact of a 10- curred in Aksu in 1716 and Alma-Ata year event is as large as that of a 100- in 1889. More recently, in 2003, an year event, and the annual avenge of earthquake caused three deaths and affected GDP is dominated by events affected close to 40,000 people. that happen relatively frequently. If the impact of a 100-year event is This map depicts the impact of much greater than that of a 10-year earthquakes on provinces' GDPs, event, then less frequent events represented as percentages of their make larger contributions to the annual average GDPs affected, with annual average of affected GDP Thus, greater color saturation indicating even if a provinces annual affected Ku A Paviodarskaya higher percentages. The bar graphs GDP seems small, less frequent and represent GDP affected by earth- more intense events can still have quakes with return periods of 10 large impacts. Zapadnokazachstanskaya years (white) and 100 years (black). - MCtGOLIA The horizontal line across the bars The annual avenge pop Aktubinskaya a also shows the annual average of affected by earthquakes in Ka- GO afete b e rt qu ke.zakhstan is about 200,000 and the Atruky GDP affected by earthquakes.affected GP about When an earthquake has a 10-year $1 billion. The annual averages of Kyzy[ordinskaya return period, it means the probabil- fatalities and capital losses caused Mangstauskaya ity of occurrence of an earthquake of by earthquakes are about 500 and that magnitude or greater is 10 per- about $400 million, respectively The CA Yujno-kazachstanskay A may City ara cent per year. A 100-year earthquake fatalities and capital losses caused SEA U RREPUIBLI has a probability of occurrence of by more intense, less frequent events 1 percent per year. This means that can be substantially larger than E N IS TA N over a long period of time, an earth- the annual averages. For example, NACIHISTNN quake of that magnitude will, onAffected GDP for A nnual Average of Affected G ( N quake o that agnitue will,on an arthquke witha 0.4 ercent10 and 100-year return periodIs ______________ average, occur once every 100 years. annual probability of occurrence It does not mean a 100-year earth- 250-year return period event) could [o 1 quake will occur exactly once every cause nearly 20,000 fatalities and 0 1 Imp 100 years. In fact, it is possible for $20 billion in capital loss (about 10 an earthquake of any return period percent of GOP). AnnOal to yccur more thnan once i the same 10-year 100-year Kaz khs anWORLDBANKGROUP EL|GFDRR "AND CENTRAL A51A(ECA) EARTHQUAKE EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS (MILLIONS $) ANNUAL AVERAGE FATALITIES Tahe rose diagrams show the provinces with the potential :5r elfor greatest annual average capital losses and highest annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital Kyzylordinskaya 0.4jno kazachstanskaya Karagandinskaya 0 ambylslkaya 30 loss occurs in the Almaty City area, which is not surprising, given the economic importance of the province. 'ae 0'TO EARTHQUAKE EXCEEDANCE ROBABILITY CURVE, 2015 AND 2080 EXCEEDANCE PROBABILITY CURVE 2015 AND 2080 he exceedance probability curves display the GDP T affected by, respectively, floods and earthquakes for 140 300 varying probabilities of occurrence. Values for two different time periods are shown. A solid line depicts the affected GDP for 2015 conditions. A diagonally striped band depicts 2080 100 200 the range of affected GDP based on a selection of climate 2080 and socioeconomic scenarios for 2080. For example, if 80 150 Kazakhstan had experienced a 100-year return period flood 60 event in 2015, the affected GDP would have been an esti- 10 lomated $10 billion. In 2080, however, the affected GDP from 40 r the same type of event would range from about $60 billion 2015 20 5i 2015 to about $100 billion. If Kazakhstan had experienced a ..... . -- 250-year earthquake event in 2015, the affected GDP would 10 500 1 50 100 250 have been about $20 billion. In 2080, the affected GDP from turn perod (years) Return period (years) the same type of event.would range from about $100 billion 1 1 10 2 1 04- to $300 billion, due to population growth, urbanization, and P robabilIty (%) Probability (%) the increase in exposed assets. All historical data on floods and earthquakes are from D. Guha-Sapir, R. Below, and Ph. Hoyois, EM-DAT: International Disaster Database (Universiti Catholique de Louvain, Brussels, Belgium), www.emdat.be; the National Geophysical Data Center/World Data Service (NGDC/WDS), Significant Earthquake Database (National Geophysical Data Center, NOAA), doi:10.7289/V5TD9V7K; and J. Daniell and A. Schaefer, "Eastern Europe and Central Asia Region Earthquake Risk Assessment Country and Province Profiling," final report to GFDRR, 2014. Damage estimates for all historical events have been inflated to 2015 US$. WORLDBANKGROUP G D R EUROPE AND CENTRAL AS1A(ECA) AFFECTED AFFECTED CAPITAL LOSS GQ U M RI5K PROF] LES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $6.3 billion* Kosovo 0.oplaio-.9milin K osovo's population and economy pecn)adgrulreee- S E are exposed to earthquakes and ating the remainder. Kosovo's floods, with earthquakes posing per capita GDP was $3,410. AnnualAverage of Affected GDP the greater risk of a high impact, lower probability event. The model results for present-day risk shown in this risk pro- province in Kosovo, with great- 1 EARTHQUAKE file are based on population and gross er color saturation indicating domestic product (GDP) estimates for greater GDP within a province. The blue circles indicate the0Nelibe 2015. The estimated damage caused by 0 Netligib[e historical events is inflated to 2015 US risk of experiencing floods dollars. and the orange circles the risk of earthquakes in terms of 4 Just over half of Kosovo's population normalized annual average of lives in rural environments. The coun- affected GDP. The largest circl try's GDP was approximately US$6.4 represent the greatest noral- Peja billion in 2015, with most derived from ized risk- The risk is estimat services and industry (together almost using flood and earthquake risk models. The table displays the proy- TO AFETE POVNCSinces at greatest normalized G1 TOP AFFECTED PROVINCES risk for ea ch p era. I n both - - relative and absolute terms, the province at greatest risk of FL EARTHQUAKE Gui) ARTHUAKE floods is Mitrovica, and thepone ANNUAL AVERAGE OF ANNUAL AVERAGE OF AFFECTED GDP (%) AFFECTED GDP (%) a g n o e,hu, is Prizreni. Mitrovica Prizreni 3 fR OF MACEDONIA Gjilani 0 Gjakova 2 LBA Prishtin6 Gjilani 1 Prizreni Peja 1 SubnatiheaKObounaries(on thensoso Peja 0 Ferizaji 1 Gjakova 0 PrishtinL 1 map represent p[anning areas for the There s a high correlation Ferizapr c Mitrovia 0 purposes of this analysis. A o 95) between he Sm a ppu aio a dGDP of a prvic province Kosov WORLDBANKGROUP EUGFDRR "AND CENTRAL A51A(ECA) Affected GDP (%) for N o reliable reports are available that happen relatively frequently. 10 and 100-year return periods on flood damage for Kosovo. Ifthe impact of a 100-year event is One block = 1% The country was, however, much greater than that of a 10-year 10 affected by floods in 2010 and 2014. event, then less frequent events make a larger contribution to the annual Annual average 2 This map depicts the impact of flood- average of affected GDP. Thus, even ing on provinces' GDPs, represented if a province's annual affected GDP S E RBIA 10-year 100-year as percentages of their annual aver- seems small, less frequent and miore age GDPs affected, with greater color intense events can still have large saturation indicating higher percent- impacts. Annual Average of Affected GOP ages. The bar graphs represent GDP affected by floods with return periods The annual average population of 10 years (white) and 100 years affected by flooding in Kosovo is (black). The horizontal line across the about 10,000 and the annual average o a 9 bars also shows the annual average of affected GDP about $50 million. For GDP affected by floods. most provinces, the 10- and 100- year impacts do not differ much, so When a flood has a 10-year return relatively frequent floods have large period, it means the probability of impacts on these averages. occurrence of a flood of that magni- Peja Prlshtlne tude or greater is 10 percent per yea r. A 100-year flood has a probability Pr,kne of occurrence of 1 percent per year. This means that over a long period of time, a flood of that magnitude will, At on average, occur once every 100 a vjilani years. It does not mean a 100-year flood will occur exactly once every 100 years. In fact, it is possible for a FenzaJI flood of any return period to occur more than once in the same year, or Prizreni to appear in consecutive years, or not to happen at all over a long period of MACOA time. LB A If the 10- and 100-year bars are the same height, then the impact of a 10- year event is as large as that of a 100- Subnational boundaries on the Kosovo year event, and the annual average of map represent ' planning areas" for the affected GDP is dominated by events purposes of this analysis. KosvoWORLDBANKGROUP EL GFR ROP AND ECENTRAL A5IA(ECA) K osovo's worst earthquake to happen at all over a long period And 10e r since 1900 took place in of time. On 10 1911. Its epicenter was in *80 Ochrida, now FYR of Macedonia. If the 10- and 100-year bars are 40 The same region was also hit by an the same height, then the impact AnLia[ average 20 earthquake in 1896. of a 10-year event is as large as that of a 100-year event, and the ;t E SRBI A 1-er10ya This map depicts the impact of annual average of affected GDP is earthquakes on provinces' GDPs, dominated by events that happen represented as percentages of their relatively frequently If the impact annual average GDPs affected, with of a 100-year event is much greater Annual Average of Affected GOP (%) greater color saturation indicat- than that of a 10-year event, then ing higher percentages. The bar less frequent events make larger graphs represent GDP affected by contributions to the annual average earthquakes with return periods of affected GDI. Thus, even if a mitrovica of 10 years (white) and 100 years province's annual affected GDP (black). The horizontal line across seems small, less frequent and the bars also shows the annual more intense events can still have average of GDP affected by earth- large impacts. quakes. The annual average population af- J When an earthquake has a 10-year fected by earthquakes in Kosovo is return period, it means the prob- about 30,000 and the annual aver- ability of occurrence of an earth- age affected GDP about $90 million. quake of that magnitude or greater The annual averages of fatalities is 10 percent per year. A 100-year and capital losses caused by earth- earthquake has a probability of quakes,are about two and about occurrence of 1 percent per year. $10 million, respectively. The fatal- This means that over a long period ities and capital losses caused by of time, an earthquake of that mag- more intense, less frequent events nitude will, on average, occur once can he substantially larger than every 100 years. It does not mean the annual averages. For example, a 100-year earthquake will occur an earthquake with a 0.4 percent exactly once every 100 years. In annual probability of occurrence L A fact, it is possible for an earthquake (a 25S-year return period event) of any return period to occur more could cause nearly $400 million in than once in the same year, or to capital loss (6 percent of GUD'). theatoaborclre sam hehht theothoimac appear in consecutive years, or not o map represent op[ar0ning areas for the purposes ofnthis analysis KosovoWORLDBANKGROUP E|GDR ROPE" AND EENTRAL A51A(ECA) EARTHQUAKE EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS (MILLIONS $) ANNUAL AVERAGE FATALITIES lhe rose diagrams show the provinces with the potential 0 ' for greatest annual average capital losses and highest annual average numbers of fatalities, as determined using Ferzji o,6 -akova 2 Feriza j 6aova 0 an earthquake risk model. The potential for greatest capital loss occurs in Prizreni, which is not surprising, given the economic importance of the province. --- --------- ~~~~~~----------------- --------------------------------- ---- - ----------------------------------------------------------------------------------------------- ' EARTHQUAKE EXCEEDANCE PROBABILITY CURVE, 2015 AND 280' EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 he exceedance probability curves display the GDP A" ITaffected by, respectively, floods and earthquakes for 08 1'6 varying probabilities of occurrence. Values for two different 07 14time periods are shown. A solid line depicts the affected 1GDP for 2015 conditions. A diagonally striped band depicts r0 A the range of affected GDP based on a selection of climate 2080 0. 0and socioeconomic scenarios for 2080. For example, if 04 200Kosovo had experienced a 100-year return period flood S event in 2015, the affectedQGDP would havebeen an esti- 03 d, x mated $200 million. In 2080,Jhowever, the affected GDP 2015 2015 fromhe me type of evet would range from about $300 0.1 2millon t about $700 million. If Kosovo had experienced a attar earthquke event in 2015, the affected GsP would have been about $3 billion. In 2 d080, the affected GDP from Retun pthe same type of event would range from about $6 billion to 10 2 10. 2O about $15 billion, due to population growth, urbanization, and the increerin exposed assets. mated $200 milhon-- ----In 2Q80 howeve the. - afce All historical data on earthquakes are fromJ. Daniel[ and A.Schaefer, "Eastern Eurcpe and Centl Asia Regio atuak Risk,Assessent Curitry and Province Profiling,"fria[ reportto2 GFeFaR, 2014 Dam ae estiatesfor allhis- torical events have been inflated to 2015 US$. WORLDBANKGROUP G D R EUROPE AND CENTRAL AS1A(ECA) AFFECTED AFFECTED CAPITAL LOSS RIKPROF]LES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $5.5 billion* Kyrgyz Repubili e Pplain .7mllo*- T he Kyrgyz Republic's pop- ture contributing 20 percent. The estimated using flood and earthquake ulation and economy are Kyrgyz Republics per capita GDP was risk models. exposed to earthquakes and $970. floods, with earthquakes posing the The table displays the provinces at greater risk of a high impact, lower This map displays GDP by province greatest normalized risk for each Annual Average of Affected GOP probability event. The model results in the Kyrgyz Republic, with greater peril. In relative terms, as shown in 5 for present-day risk shown in this color saturation indicating greater the table, the province at greatest risk profile are based on population GDP within a province. The blue cir- risk of floods is Talas, and the one at EARTHQUAKE and gross domestic product (GDP) ces indicate the risk of experiencing greatest risk of earthquakes is Osh. estimates for 2015. The estimated floods and the orange circles the risk In absolute terms, the province at AZAKHSTAN 0 Neg[igible damage caused by historical events is of earthquakes in terms of normal- greatest risk of both floods and earth- inflated to 2015 US dollars. ized annual average of affected GDP. quakes is Chuy The largest circles represent the Over 60 percent of the Kyrgyz greatest normalized risk. The risk is Republic's population lives in rural environments. The country's GDP was approximately UmS$5.5 billion in 2015, with nearly 50 percent de rived from services, most of the remainder generated by industuy and agricul- TOP AFFECTED PROVINCES FLOODEARTHQUAKE ANNUAL AVERAGE OF ANNUAL AVERAGE OFCHINA AFFECTED GDP ( ) AFFECTED GDP (4 Talas tablh 5 the provinces at Naryn 3 Chuiy 4 Osh 2 Ysyk-kol 4 nlal-abad is4 Chiiy a Batken 3(r=0.95) between the peril In relativ tem3asoni risknoffloodsispopulation and GP ofnd t rekats risk ofaa-a 2 provin Ink-o ab olt termss th provinceea Kyr yz epu licWORLDBANKGROUP EL GFD RPE" AND EENTRAL A51A(ECA) heyrgyz Relic hasO ex-DR T he Kyrgyz Republic has ex- If the 10- and 100-year bars are the avenage affected GDP about $70 Affected GOP (%) for perienced some floods since same height, then the impact ofa 10- million. For most provinces, the 10- 10 and 100-year return periods gaining its independence in year event is as large as that of a 100- and 100-year impacts do not differ One black = 2% 20 1991. Floods in 1998 and 2005 each year event, and the annual average of much, so relatively frequent floods caused over $3 million in damage. affected GDP is dominated by events have large impacts on these avenges. 10 More recently, in 2012, flooding in that happen relatively frequently For the few in which the 100-year Annua[ average41 Osh, Batken and jalalbad affected If the impact of a 100-year event is impacts are greater than the 10-year about 11,000 people. much greater than that of a 10-year impacts, less frequent events make a 1O-year 100-year event, then less frequent events make significant contribution to the annual This map depicts the impact of flood- a larger contribution to the annual avenge of affected GDP. ing on provinces' GDPs, represented average of affected GDP. Thus, even Annual Average of Affected GOP (%) as percentages of their annual aver- if a province's annual affected GDP age GDPs affected, with greater color saturation indicating higher percent- intense events can still have large ages. The bar graphs represent GDP impacts. 0 V 6 9 affected by floods with return periods of 10 years (white) and 100 years The annual avenge population affect- (black). The horizontal line across the ed by flooding in the Kyrgyz Repub- bars also shows the annual average of lic is about 80,000 and the annual GDaP aff cted by flGoo ds.isdominate When a flood has a 10-year return period, it means the probability ofo occurrence of a flood of that magni-a tud e o r gre ater is 10 p erc ent per year Ysyk-ko[ A 100-year flood has a probability/ ';> of occurrence ofi1 percent per year 1 This means that over a long period of time, a flood of that magnitude will, on average, occur once every 100 GP years. It does not mean a 100-year flood will occur exactly once everyR 100 years. In fact, it is possible for a flood of any return period to occur more than once in the same year, or to appaear in consccutive ye ars, ogr notth to happen at all over a long period of Ac G (f K R u iWORLDBANKGROUP E RO AND EENTRAL A5A ECA) T hyrgyz Reblicc worstR RISK RFIE T he Kyrgyz Republic's worst If the 10- and 100-year bars are the $200 million. The annual avenages of Affected GOP (%) for earthquake since 1900 oc- same height, then the impact of a 10- fatalities and capital losses caused 10 and 100-year return periods curred in 1911 in Pamir, caus- year event is as large as that ofa 100- by earthquakes are about 200 and One b[ock - 10% 100 ing over 90 fatalities. Earthquakes year event, and the annual avenge of about $100 million, respectively. The in 1992 caused over 50 fatalities affected GDP is dominated by events fatalities and capital losses caused s0 and close to $300 million in damage. that happen relatively frequently. by more intense, less frequent events Annua average More recently, in 2008, an earthquake If the impact of a 100-year event is can be substantially larger than the killed over 70 people. much greater than that of a 10-year annual averages. For example, an In-vear 10jea event, then less frequent events make earthquake with a 0.4 percent annual This map depicts the impact of larger contributions to the annual av- probability of occurrence (a 250-year earthquakes on provinces' GDPs, erage of affected GDP. Thus, even if a return period event) could cause Annual Average of Affected GOP (%) represented as percentages of their province's annual affected GDP seems nearly 8,000 fatalities and $4 billion annual average GDPs affected, with annua avene GDs affcted,with sm all, les s fr equ ent and more inten se in capital loss (about 60 percent of ____________ greater color saturation indicating events can still have large impacts. GDP). higher percentages. The bar graphs KAZAKHSTAN represent GDP affected by earth- The annual avenge population quakes with return periods of 10 affected by earthquakes in the Kyrgyz years (white) and 100 years (black). Republic is about 200,000 and the The horizontal line across the bars annual average affected GDP about also shows the annual average of GDP affected by earthquakes. When an earthquake has a 10-year return period, it means the probabil- ity of occurrence of an earthquake of that magnitudre or greater is 10 percent per yeari A 100-year earth- quake has a probability of ccu rren ce of 1 percent per year. This means that over a long period of time, an earthquake of that magnitude will, on avege, occur once every 100 years., ia It does not mean a 100-year earth- quake will occur exactly once every 100 years. In fact, it is possible forc an earthquake of any return perioda to cccur more thsan once ing the same year, oar to appear in cons cutive years, or not to happen at all over a long period of time. Kyr yz epu licWORLDBANKGROUP E|GFDRR NM5ENTRAL A51A(ECA) EARTHQUAKE EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS (MILLIONS $) ANNUAL AVERAGE FATALITIES he rose diagrams show the provinces with the potential for greatest annual average capital losses and highest annual average numbers of fatalities, as determined using atken ) bad t-6a an earthquake risk model. The potential for greatest capital loss occurs in Chuy, which is not surprising, given the eco- nomic importance of the province. EARTHQUAKE EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 EARTHQUNCE he exceedance probability curves display the GDP E D affected by, respectively, floods and earthquakes for 8a7 varying probabilities of occurrence. Values for two different. time periods are shown. A solid line depicts the affected 0 GDP for 2015 conditions. A diagonally striped band depts 6 the range of affected GDP based on a selection orlate and socioeconomic scenarios for 2080. Forempfe, if the 4 Kyrgyz Republic had experienced a r return period 3030 flood event in 2015, affected GDP wuld have been an estimated $400 miIlipn. n 2080, owever, the pffected GDP 2 20 from the sarnetpe of eventwould range fron about $4 2015 1 10 2011 billion to abou $7 billion. If the Kyrgyz Republic had expe- rienced a 250-year earthquake event in 2015, the affected 10 s0 100 250 10 so 100 250 GD0P wopd have been an estimated $4 billion. In 2080, the Return period (years) Return ped (years) affected GDP from the same type of event would range from A0 2 / ... .. . . 4 about $20 billion to about $60 billion, due to population 10 2 1 Probability (%) Poaitygrowth, urbanization, and the increase in exposed assets. All historical data on floods and earthquakeSare on D Guha-Sapir, R. Blow ard iW , M-D tintr isaster Datab ase (Universit Catholique de Louvain, Brussels, Belgium), www.emdat.be; the National Geophysical Data Center/World Data Servic(NGH/WDS), gificant Earthuake Database (inal n eophy i t NOAA), do 10 7289/V5TD9V7K ad I Daniell and A Schaefer, "Eastern Europe and Central Asia Region Earthquake Risk Assessment Country and Province,Profaling," lia report to >FQRR, 2014 Damage estimats for all historical eventS have been infated to 2015 l,I$ WORLDBANKGROUP EI GROPE AND CENTRALA5IA (ECA) AFFECTED AFFECTED CAPITAL LOSS QRI5KPROFILES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $28.2 billion* L3 L atvia's population and economy and agriculture making a small are exposed to earthquakes and contribution. Latvia's per capita floods, with floods posing the GDP was $22,000. greater risk The model results for pres- SALTIC SEA ESTONIA ent-day risk shown in this risk profile This map displays GDP by are based on population and gross province in Latvia, with greater domestic product (GDP) estimates for color saturation indicating 2015. The estimated damage caused by greater GDP within a province. aEiESA historical events is inflated to 2015 US T b c e ae dollars. risk of experiencing floods and the orange circles the risk ATk5e I Close to 70 percent of Latvia's pop- of earthquakes in terms of 0 ulation lives in urban environments. normalized annual average of The country's GDP was approximately affected GDP. The largest circles R US$28.2 billion in 2015, with nearly 70 represent the greatest normal- percent derived from services, most of ized risk The risk is estimated igas Madonas the remainder generated by industry using flood and earthquake risk models. GDPlu was $22000 Dobetes A7rul,Lda The table displays the prov- Rezekne5 TOP AFFECTED PROVINCES inces at greatest normalized risk for each peril. In relative terms, as shown in the table, U) EARTHQUAKE the province at greatest risk of ANNUAL AVERAGE OF ANNUAL AVERAGE OF floods is jekabpils, and the one AFFECTED GDP () AFFECTED GDP () at greatest risk of earthquakes L I T H U AN 1A Sis Ogres. In absolute terms, it Kuldigas Limbazu 0 is Rigas. L Ogres 3 Rezeknes 0 Annual Average of Affected GDP (GOP (billions of Preilu 3 Saldus 0 Rigas 3 Talsu 0 rs There is a high correlation Aizkraukles 2 Madonas 0 (r=O.95) between the auvas 2 EAirs 1 R UEARTHQUAKE population and GDP ofa eagkabs 2 BalOrs province. Valkas I Kuldigas 0 Valmieras 1 Aluksnes 0 0 Negligible Lat iaWORLDBANKGROUP E|G DR ROPE AND EENTRAL A51A(ECA) his map depicts the impact of event, then less frequent events make flooding on provinces' GDPs, a larger contribution to the annual represented as percentages of average of affected GDP. Thus, even their annual average GDPs affect- if a province's annual affected GDP ed, with greater color saturation seems small, less frequent and more indicating higher percentages. The intense events can still have large bar graphs represent GDP affected impacts. by floods with return periods of 10 years (white) and 100 years (black). The annual average population ESTONIA The horizontal line across the bars affected by flooding in Latvia is about also shows the annual average of GDP 30,000 and the annual average affect- affected by floods. ed GDP about $600 million. Within the various provinces, the 10- and VlirsR US S 1A N When a flood has a 10-year return 100-year impacts do not differ much, period, it means the probability of so relatively frequent floods have occurrence of a flood of that magni- large impacts on these averages. tude or greater is 10 percent per year. A 100-year flood has a probability of occurrence of 1 percent per year. Lff 'I) B This means that over a long period of time, a flood of that magnitude will, Madna on average, occur once every 100 years. It does not mean a 100-year flood will occur exactly once every a Dau LiepjasDobl,esAlzrauiesLudzas 100 years. In fact, it is possible for a flood of any return period to occur more than once in the same year, or to appear in consecutive years, or not to happen at all over a long period of time.Dagvps Ifa larger-contributionrtoatheaannual Affected GOP for Annual Average of Affected GDPPT%) same height, then the impact of a 10- 10 and 100-year return periods ye ar event is as large as that ofa 100- Ont block [a sl ha LAR S year event, and the annual average ofp affected GDP is dominated by events a d o L that happen relatively frequenty. If the impact ofa 100-year event is much greater than that of a 10-year 10-year 100-year LatViaWORLDBANKGROUP EL GFRROP AND ECENTRAL A5IA(ECA) T he worst earthquake to If the 10- and 100-year bars are affect Latvia since 1900 oc- the same height, then the impact curred in 1908. Other major of a 10-year event is as large as events have occurred in 1616 and that of a 100-year event, and the 1821. annual average of affected GDP is dominated by events that happen This map depicts the impact of relatively frequently. If the impact earthquakes on provinces' GDPs, of a 100-year event is much greater represented as percentages of their than that ofa 10-year event, then BALTIC SEA annual average GDPs affected, with less frequent events make larger greater color saturation indicat- contributions to the annual aver- ing higher percentages. The bar age of affected GDP. Thus, even if graphs represent GDP affected by a province's annual affected GDP Nb FEEERAT N earthquakes with return periods seems small, less frequent and Valks of 10 years (white) and 100 years more intense events can still have (black). The horizontal line across large impacts. VWntskilesTake the bars also shows the annual VentpiA average of GDP affected by earth- The annual average population Cesu Golbenes quakes. affected by earthquakes in Latvia is Rigas BalvU about 100 and the annual aver-age -Ri._ 2__ When an earthquake has a 10-yearTurn Whenan artquae hs a10-ear GOP about $2 million. The annual ',Kuldigas j m K3 Ors -- aoa return period, it means the prob- avrae o flsd ias ability of occurrence of an earth- quake of that magnitude or greater less than one and about $500,000, Jelgavas les is 10 percent per year. A 100-year respectively. The fatalities and epajas Rezeknes' e arthq uake ha s a probabil ity o f I eac rquae ha a probaltryeaof capital losses caused by more 7 lekabpils\ Prei lu occurrenceintense, less frequent events can This means that over a long period be substantially larger than the of time, an earthquake of that mag- annual averages. For example an Kras[avas nitude will, on average, occur once earthquake with a OA percent Daugavpits every 100 years. It does not mean annual probability ofoccurrence L_1HUANIA a 100-year earthquake will occur (a 250-year return period event) exactly once every 100 years. In could cause nearly $20 million in fact, it is possible for an earthquake capital loss (ess than 1 percent of GDP not affected for 10 and Annual Average of Affected GOP LA S of any return period to occur more GDP). 100-year return periods than once in the same year, or to AnnLal average = 0 appear in consecutive years, or not to happen at all over a long period of time. LatiaWORLDBANKGROUP E|GDR RO" AN CENTRAL A51A(ECA) EARTHQUAKE ( EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS ($) ANNUAL AVERAGE FATALITIES 0 he rose diagrams show the provinces with the potent i T for greatest annual average capital losses and highest annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital Madonas 20,000 Vamieras 30,000 Kuldigas 0 Saldus 0 loss occurs in Rigas. ,§~EARTHQUAKE EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 YJ'EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 he exceedance probability curves display the GDP .1 affected by, respectively, floods and earthquakes for 12 0.35 varying probabilities of occurrence. Values for two different 030 time periods are shown. A solid line depicts the affected 00 30 208025 the range of affected GDP based on a selection of climate 2080 and socioeconomic scenarios for 2080. For example, if Lat- 2015 i EARTHQUAKtype of eean e woldranebr abolycut $5 isllio to about 6$ ili I avia had experienced a 250-yearreunpiofld ehqaen 0XCEn 1 in 20 15 , the affected G P would have been a butaed $ 50 20 4illion. In 2080 e, th e affected G dP from the same ypeo 8 ;2 0.25type rnof eteud rDanged o au $5elion o abouat 0 05$ il.I avia had experienced a 250-year earthquake lod een 0 ~ ..0.**15ev in 2 0151 , th e affected G DP would h ave been a butiatd $70 1 0 100 2500 million. In 2080, oe, the affected GDP from the sameeo tn pend (years) event would range from about $100 million to about $300 10 2 1 0.4 million, due to population growth, urbanization, and the Probabi lity(% Probability (%) - increase in exposed assets. Damage estimates for all historical events have been inflated to2015 US$. WORLDBANKGROUP GR EROPE AND CENTRALA5IA (ECA) AFFECTED AFFECTED CAPITAL LOSS QRI5KPROFILES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $43.0 billion* Lithuamia o Lithuania's population and econ- and agriculture making a small omy are exposed to earthquakes contribution. Lithuania's per and floods, with floods posing capita GDP was $15,100. the greater risk. The model results for L A7,VIA present-day risk shown in this risk pro- This map displays GDP by prov- file are based on population and gross ince in Lithuania, with greater Teau o domestic product (GDP) estimates for color saturation indicating Siauli 'Panevezia 2015. The estimated damage caused by greater GDP within a province. historical events is inflated to 2015 US The blue circles indicate the risk dollars. of experiencing floods and the Kalpedes orange circles the risk of earth- Close to 70 percent of Lithuania's quakes in terms of normalized population lives in urban environments. annual average of affected GDR The country's GDP was approximately The largest circles represent the Ta rages US$43.0 billion in 2015, with nearly 70 greatest normalized risk. The percent derived from services, most of risk is estimated using flood and the remainder generated by industry, earthquake risk models. auno The table displays the provinces at greatest normalized risk for RUSS AN FEDERATION ilnia TOP AFFECTED PROVINCES each peril. In relative terms, as shown in the table, the prov- ince at greatest risk of floods is EARTHQUAKE Alythaus, and the one at greatest ANNUAL AVERAGE OF ANNUAL AVERAGE OF risk of earthquakes is Siauliu. In AFFECTED GDP (%) AFFECTED GDP (%) absolute terms, the province at greatest risk of floods is Vilniaus, Parevezio KlSipedos 0 and the one at greatest risk of Kauno 2 Telsiu 0 earthquakes is Siauliu. Annual Average of Affected GOP (%) BELARUS Vilniaus 2 Panevezio 0 5 Taurages 2 Marijampoles 0 5ODP (billions of$ Marijampoles 2 Utenos 0 There is a high correlation Klaipedos I Kauno 0 POLAvD 1 ATQAE(r=O.95) between the Utenos Taurages 0 EARTHQUAKE Siaulio Alytaus 0 population and GDP of a Teliu ViAniaus 0 0 Negligible province. Telsiui u VilniaUS 0 Lith aniaWORLDBANKGROUP E|G DR ROPE AND EENTRAL A51A(ECA) F looding in 2010 caused four much greater than that of a 10-year LATVIA fatalities in Lithuania. event, then less frequent events make a larger contribution to the annual This map depicts the impact average of affected GDP. Thus, even of flooding on provinces' GDPs, repre- if a province's annual affected GDP sented as percentages of their annual seems small, less frequent and more average GDPs affected, with greater intense events can still have large color saturation indicating higher impacts. percentages. The bar graphs repre- sent GDP affected by floods with re- The annual average population turn periods of 10 years (white) and affected by flooding in Lithuania is 100 years (black). The horizontal line about 60,000 and the annual average Te Siaul i across the bars also shows the annual affected GDP about $800 million. For average of GDP affected by floods. most provinces, in which the impacts _ from 10- and 100-year floods do not Klaipedos When a flood has a 10-year return differ much, relatively frequent floods period, it means the probability of have large impacts on these averages. occurrence of a flood of that magni- For the few in which the 100-year teno tude or greater is 10 percent per year. impacts are much greater than the A 100-year flood has a probability 10-year impacts, less frequent events Taurage of occurrence of 1 percent per year. make a significant contribution to the This means that over a long period of annual average of affected GDKa time, a flood of that magnitude will, on average, occur once every 100 years. It does not mean a 100-year v flood will occur exactly once every 100 years. In fact, it is possible for a nius flood of any return period to occur more than once in the same year, or to appear in consecutive years, or not to happen at all over a long period of time. Aftected GDP ()for If the 10- and 100-year bars are the and 100-year return periods same height, then the impact of a 10- One block = 2% 20 year event is as large as that of a 100- year event, and the annual average of 10 Annual Average of Affected GDP (%) affected GDP is dominated by events P 0 LAN D Annual average -L that happen relatively frequently. If the impact of a 100-year event is 10-year 100-year o 1 s C & 70ya 0-ea 6 : Lith aniaWORLDBANKGROUP EL GF RROP AND ECENTRAL A5IA(ECA) Lihua ia h RS[ PRO T he worst earthquake to affect If the 10- and 100-year bars are the LATVIA Lithuania since 1900 oc- same height, then the impact of curred in 1908 near the Be- 10-year event is as large as that of a larus border. A more recent, widely 100-year event, and the annual av- felt earthquake occurred in 1988; erage of affected GDP is dominated damage was minimal, however. by events that happen relatively fre- quently. If the impact of a 100-year This map depicts the impact of event is much greater than that of earthquakes on provinces'GDPs, a 10-year event, then less frequent represented as percentages of their events make larger contributions to annual average GDPs affected, with the annual average of affected GDP. greater color saturation indicating Thus, even if a province's annual TelsiuSauli Pan io higher percentages. The bar graphs affected GDP seems small, less fre- represent GDP affected by earth- quent and more intense events can quakes with return periods of 10 still have large impacts. KLaipedos years (white) and 100 years (black). The horizontal line across the bars The annual average population af- also shows the annual average of fected by earthquakes in Lithuania GDP affected by earthquakes. is about 100 and the annual average Taues affected GDP about $2 million. The K When an earthquake has a 10-year annual averages of fatalities and return period, it means the prob- capital losses caused by earth- Kauno ability of occurrence of an earth- quakes are less than one and about quake of that magnitude or greater $500,000, respectively. The fatal- is 10 percent per year. A 100-year ities and capital losses caused by RUSS AN FEDERATION J VilniaUS earthquake has a probability of more intense, less frequent events Marijampoles VIFiu occurrence of t percent per year. can he substantiall larger than This means that over a long period the annual averages. For example, of time, an earthquake of that mag- an earthquake with a 0.4 percent nitude will, on average, occur once annual probability of occurrence (a every 100 years. It does not mean 250-year return period event) could a 100-year earthquake will occur cause about $20 million in capital Alvtaus exactly once every 100 years. In loss (less than 1 percent of GDP). fact, it is possible for an earthquake BELARUS of any return period to occur more than once in the same year, or to GDP (%) not affected for 10 and Annual Average of Affected GDP (%) appear in consecutive years, or not i-U LALi-10O-year return periods to happen at all over a long period Annual average = 0 of time. Lit ua iaWORLDBANKGROUP EL|GFDRR "AND CENTRAL A51A(ECA) EARTHQUAKE 7 EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS ($) ANNUAL AVERAGE FATALITIES 7 6 he rose diagrams show the provinces with the potential 0o1S T for greatest annual average capital losses and highest annual average numbers of fatalities, as determined using Utenos 8,000 Tj u 40,000 T s 0an earthquake risk model. The province with the potential for greatest capital loss is Siauliu. . . ......-----0 EARTHQUAKE EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 he exceedance probability curves display the GDP 1 affected by, respectively, floods and earthquakes for 12 0.18 varying probabilities of occurrence. Values for two different 0.16 time periods are shown. A solid line depicts the affected 10 0.14 GDP for 2015 conditions. A diagonally striped hand depicts 2080 8 0.12 the range of affected GDP based on a selection of climate and socioeconomic scenarios for 2080. For example, if Lith- 6 u0.10 ania had experienced a 100-year return period flood event 0.08 in 2015, the affected GDP would have been an estimated $4 2015 2080 4 0.06 billion. In 2080, however, the affected GDP from the same .0 type of event would range fmm about $7 billion to about 0.02 2015 $10 billion. If Lithuania had experienced a 250-year earth- I I I _____ 2015quake event in 2015, the affected GDP would have been 10 50 100 250 10 50 100 250 about $60 million. In 2080, the affected GDP from the same Return period (years) Return period (years) type of event would range from about $100 million to about 10 2 1 0.4 10 2 1 074 $200 million, due to population growth, urbanization, and Probability (%) Probability (%) the increase in exposed assets. All historical data on floods and earthquakes are from D. Guha-Sapir, R. Below, and Ph. Hoyois, EM-DAT: International Disaster Database (Universit Catholique de Louvain, Brussels, Belgium), www.emdat.be, and 1. Daniell and A. Schae- fer, "Eastern Europe and Central Asia Region Earthquake Risk Assessment Country and Province Profiling," final report to GFDRR, 2014. Damage estimates for all historical events have been inflated to 2015 US$. WORLDBANKGROUP G D R ELUROPE AND CENTRAL AS1A(ECA) AFFECTED AFFECTED CAPITAL LOSS GQ U M RI5K PROF] LES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $10.5 billion* IVacedonia, FYR os- - T he population and economy of of the remainder generated by B U LG A I A the former Yugoslav Republic of industry and agriculture making Macedonia are exposed to earth- a smaller contribution. FYR of quakes and floods, with floods posing Macedonis per capita GDP was the greater risk. The model results for $5,040. present-day risk shown in this risk pro- file are based on population and gross domestic product (GDP) estimates for ince in the FYR Macedonia, wit Pti 0, DnlA%I' 2015. The estimated damage caused by greater color saturation indicating historical events is inflated to 2015 US greater GDP within a province. dolr.The blue circles indicate the risk Seti Nmikosn dollars. of experiencing floods and the About 60 percent of FYR Macedonia's orange circles the risk of earth- population lives in urban environments. quakes in terms of normalized The country's GDP was approximately annual average of affected GDP. BrRd US$10.5 billion in 2015, with over 60 The largest circles represent the percent derived from services, most greatest normalized risk The risk is estimated using flood and earth- Dbrr, Neon ________________________ quake risk models. TOP AFFECTED PROVINCES The table displays the provinces at greatest normalized risk for each Dernir4 N I'peril. In relative terms, as shownriGeqij EARTHQUAKE in the table, the province at great ANNUAL AVERAGE OF ANNUAL AVERAGE OF est risk of floods is Skopje, and the AFFECTED GDP (%) AFFECTED GDP (%) one at greatest risk ofearthquakes GREECE is Ohrid. In absolute terms, the Skopje Ohrid 6 province atgreatest risk of both Veles Struga 5 Valandovo Debar 4 floods and earthquakes is Skopje. Annua Average of Affected GOP Stip Resen 3 Negotino 2 Skopje 2 GDP (billions of 10 Brod 2 Berovo 2 There is a high correlation 5 F Kavadarci 2 Delcevo 2 b EARTHQUAKE Gevgelija Strumica 2population and DP of a Bitola 1 Gostivar 2 province. Debar o Virica 2 e0 Negligible M ace onia FYRWORLDBANKGROUP E|G DR ROPE" ANE ENTRAL A51A(ECA) he most devastating flood in fthe 10- and 100-year bars are the the former Yugoslav Republic same height, then the impact of a 10- of Macedonia since it gained year event is as large as that ofa 100- its independence in 1991 occurred in year event, and the annual average of SE 1995 and caused nearly $400 million affected GDP is dominated by events in damage. More recently, flooding in that happen relatively frequently. BULGARIA 2004 affected over 100,000 people If the impact ofa 100-year event is and caused almost $5 million in much greater than that of a 10-year damage. event, then less frequent events make Kriva Palanka a larger contribution to the annual Kumanovo/V This map depicts the impact of flood- average of affected GDP Thus, even ing on provinces' GDPs, represented if a provinces annual affected GDP as percentages of their annual aver- seems small, less frequent and more retiDl age GDPs affected, with greater color intense events can still have Large saturation indicating higher percent- impacts. ages. The bar graphs represent GDP affected by floods with return periods The annual average population affect of 10 years (white) and 100 years ed by flooding in FYR Macedonia is Gostivar S (black). The horizontal line across the about 70,000 and the annual averae bars also shows the annual average of affected GDP about $500 million. For GDP affected by floods. most provinces,the 10-and 100- year impacts do not differ much, so When a flood has a 10-year return relatively frequent floods have large Debar period, it means the probability of impact on te ans.oh occurrence of a flood of that magni- fhe tude or greater is 10 percent per year. are much greater than the 10-year Prilep a r a A 100-year flood has a probability A 10-ea foo ha aprbailty impacts, less frequent events make a Demtir Hisar of occurrence of 1 percent per year. b h a Ora id This means that over a long period of time, a flood of that magnitude will, e on average, occur once every 100 - years. It does not mean a 100-year Affected GDP (%) for flood will occur exactly once every 100 years. In fact, it is possible for a O ne block = 5% 40 flood of any return period to occur more than once in the same year, or to appear in consecutive years, or not Annual Average of Affected GOP (%) 20 to happen at all over a long period of Anniua[ average time.I T ~ T IftQ 0-an10-year 100-year M aced nia,FY RWORLDBANKGROUP EL GF R ROP AND ECENTRAL A5IA(ECA) T he former Yugoslav Republic appear in consecutive years, or not of Macedonia's worst earth- to happen at aD over a long period quake since 1900 happened of time. in 1963 in Skopje, with a magnitude S A of 6. It caused over 1,000 fatalities and close to $8 billion in damage, same height, then the Impactof a and Skopje was almost completely 10-year event is as large as that oh destroyed. A 1931 earthquake, also in Skopje, killed over 150 people. events that happen relatively fre- Kj,,nv This map depicts the impact of quently. If the impact of a 100-year earthquakes on provinces' GDPs, event is much greater than that of represented as percentages of their a 10-year event, then less frequent annual average GDPs affected, with events make larger contributions to greater color saturation indicating the annual average of affected GDP. higher percentages. The bar graphs Thus, even if a province's annual represent GDP affected by earth- affected GDP seems small, lessfre- quakes with return periods of 10 quent and more intense events can A years (white) and 100 years (black). still have large impacts. The horizontal line across the bars also shows the annual average of GDP ffeced b earhquaes.affected by earthquakes in FYR of GDP affected by earthquakes. Macedonia is about 40,000 and the When an earthquake has a 10-year annual average affected GDP about return period, it means the prob- $200 million. The annual averag- ability of occurrence of an earth- es of fatalities and capital losses quake of that magnitude or greater caused by earthquakes are about is 10 percent per year. A 100-year 10 and about $100 million, respec- 2 erthquake has a probability of eartqaehsapoaiiyo tively The fatalities and capital occurrence of t percent per year. losses caused by more intense, less This means that over a long period frequent events can be substantially of time, an earthquake of that mag- larger than the annual averages. nitude will, on average, occur once For example, an earthquake with every 100 years. It does not mean a 0.4 percent annual probability a 100-year earthquake will occur of occurrence (a 250-year return exactly once every 100 years. In period event) could cause about 400 fact, it is possible for an earthquake fatalities and $2 billion in capital Average of Affected GDP (%) of any return period to occur more loss (about 20 percent of GDP). Annua[ than once in the soame year, or to If t 10 10-year 100-year M a ed ni , YRWORLDBANKGROUP GFDRRP" AN CENTRAL A51A(ECA) EARTHQUAKE EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS (MllLiNS ANNUAL AVERAGE FATALITIES he rose diagrams show the provinces with the potential . oT for greatest annualfaverage capital losses and highest annual average numbers of fatalities, as determined using an earthquake risk model. The potentia for greatest capital Gostivar 2 truga 5 loss occurs in Skopje, which is not surprising, given the economic importance of the province. EARTHQUAKE he exceedance probability curves display the GDP EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 T affected by, respectively, floods and earthquakes for 14 >45 varying probabilities of occurrence. Values for two different time periods are shown. A solid line depicts the affected 12 GDP for 2015 conditions. A diagonally striped band depicts the range of affected GDP based on a selection of clmt 2080 and socioeconomic scenarios for 2080. For example, if FYR 8 25 of Macedonia had experienced a 100-year return period Co 20 flood event in 2015, the eced GDP would have been an 6 0 2 estimated $2 billin. , however, the affected GDP 4 t from the same type of event would range from about $7 d w0 2015 2 2015 billion totaot$11 billion. If FY of Macedonia had expe- T rienced at250-year earthquake event in 2015, the affected 0250 10 GDP would have been about $3 billion. In 2080, however, urn peod (years) the affected GDP from the same type of event would range .. .from about $20 billion to about $40 billion, due to popu- 0. laton growth, urbanization, and the increase in exposed Pro babi lity( )ases 1 7 assets. All historical data on floods and earthquakes are from D. Guha-Sapir, R. Below, and Ph. HoyDis, EN-DAT International Disaster Database (Universite CatholiqUe de Louvain, Brussels, Belgiurn), www.emdat.be, and J. Daniell and ASchae- fer, "Eastern Europe and Central Asia Region Earthquake Risk AsSessment Country and Province Priling," fial report to IFDRR, 2014 Damage estimates for all historical events have been inflated to 2015 US$. WORLDBANKGROUP EI GROPE AND CENTRALA5IA (ECA) AFFECTED AFFECTED CAPITAL LOSS Q GFDRR RI5K PROFILES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE Moldova '.Pplto 36mlin oldova's population and econ- remainder Moldovas per capita GDP UKRAINE Annual Average of Affected GDP omy are exposed to earth- was $1,760. quakes and floods, with floods posing the greater risk. The model This map displays GDP by province in Edinet EARTHQUAKE results for present-day risk shown in Moldova, with greater color satura- this risk profile are based on population ton indicating greater GDP within and gross domestic product (GDP) esti- a province. The blue circles indicate mates for 2015. The estimated damage the risk of experiencing floods and caused by historical events is inflated to the orange circles the risk of earth- 2015 US dollars. quakes in terms ofnornalized annual a G average of affected GDP. The largest Just over half of Moldova's popula- circles represent the greatest normal- tion lives in rural environments. The ized risk. The risk is estimated using country's GDP was approximately D ' 6 counry'sGDPwas pprximaelyflood and earthquake risk models. US$6.3 billion in 2015, with over 60 percent derived from services, and with The table displays the provinces at remainder.eMoldovahsgperarapitaion industry and agriculture generating the greatest normalized risk for each per- tur0 between-he il. In relative terms, as shown in the (r=pu[btwend CDPhe table, the province at greatest risk tphinri a~~~~f province Thrbueciclsnndcet. of floods is Dubasari, and the one at TOP AFFECTED PROVINCES greatest risk of earthquakes is Cahul. In absolute terms, the province atnuh greatest risk of floods and earth- rsmdl AEARTHQUAKE quakes is Chisinau. ANNUAL AVERAGE OF ANNUAL AVERAGE OF AFFECTED GDP () AFFECTED GDP(% Dubhasari 3 Cahuil 2Ggjzi Soroca i Sagauzia 1oo Edinet th Lappisna 1 Tighina 2 Ungheni 1 Balti 2 Chisinau 0 BLACK SEA Chisiflau 2 Ba[ti 0 W) Orheh 2 Tighina 0 Cahuf f Dubasari 0 UngheIn a Orhei 0 Lapusna 0 Soroca 0 M o d v WORLDBANKGROUP | D REAND EENTRALA51A(ECA) T he worst flood in Moldova on average, occur once every 100 since the country gained its in- years. It does not mean a 100-year ad GOP (%) for dependence in 1991 occurred flood will occur exactly once every One block = 1% 10 in 1994. It killed close to 50 people 100 years. In fact, it is possible for a and caused almost $500 million in flood of any return period to occur damage. In 1997, 28 of Moldova's more than once in the same year, or U N5 40 provinces experienced floods, to appear in consecutive years, or not Annua average 2 causing nine deaths and about $70 to happen at all over a long period of L, million in damage. Further flooding time. Edinet 10-year 100-year occurred in 1999, 2002, and 2005 with smaller impacts, ranging from If the 10- and 100-year bars are the $1 million to nearly $10 million in same height, then the impact of a 10- Soroca damage. This record highlights Mol- year event is as large as that ofa 100- Annual Average of Affected GDP (%) dova's vulnerability to floods. While year event, and the annual average of not always devastating, they have a affected GDP is dominated by events relatively large cumulative effect on that happen relatively frequently. the country when they follow one If the impact of a 100-year event is another quickly. much greater than that of a 10-year event, then less frequent events make This map depicts the impact of flood- a larger contribution to the annual Dubasari ing on provinces' GDPs, represented average of affected GDP. Thus, even U as percentages of their annual aver- if a province's annual affected GDP age GDPs affected, with greater color seems small, less frequent and more saturation indicating higher percent- intense events can still have large Ch nau ages. The bar graphs represent GDP impacts. affected by floods with return periods RO MANA of 10 years (white) and 100 years The annual avenge population affect- (black). The horizontal line across the ed by flooding in Moldova is aboutTighina bars also shows the annual average of 70,000 and the annual average affect- GDP affected by floods. ed GDP about $100 million. Within the various provinces, the 10- and When a flood has a 10-year return 100-year impacts do not differ much, period, it means the probability of so relatively frequent floods have Gagauzia occurrence of a flood of that magni- large impacts on these averages. tude or greater is 10 percent per year. A 100-year flood has a probability CahulBLACK SEA of occurrence of 1 percent per year. This means that over a long period of time, a flood of that magnitude will, MolovaWORLDBANKGROUP ELGDR ROPE AND ECENTRAL A5IA(ECA) M o d v isG DR RIS PROFILES 3 T he worst earthquake af- fact, it is possible for an earthquake fecting Moldova since 1900 of any return period to occur more Affected GDP (%) for * 10 and 100-year return periods occurred in 1986 in Vrancea, than once in the same year, or to Onebl[ock=10% 100 Romania, with a magnitude of 7.2. It appear in consecutive years, or not killed at least four people in Moldo- to happen at all over a long period va and caused over $200 million in of time. Uo damage. Other major earthquakes Annual average 20 affecting Moldova occurred in 1802, If the 10- and 100-year bars are the 1838, 1940, 1977, and 1990, and same height, then the impact of a Edinet 10-year 100-year all were centered in Vrancea. The 10-year event is as large as that of a 1802 event was one of the largest 100-year event, and the annual av- earthquakes on record to occur in erage of affected GDP is dominated Soroca Europe. by events that happen relatively fre- Annual Average of Affected GDP (%) quently. If the impact of a 100-year This map depicts the impact of event is much greater than that of earthquakes on provinces' GDPs, a 10-year event, then less frequent represented as percentages of their events make larger contributions to annual average GDPs affected, with the annual average of affected GDP. Orhei greater color saturation indicating Thus, even if a province's annual higher percentages. The bar graphs affected GDP seems small, less fre- Dubasari represent GDP affected by earth- quent and more intense events can quakes with return periods of 10 still have large impacts. *hsia years (white) and 100 years (black). * The horizontal line across the bars The annual average population Chisinau also shows the annual average of affected by earthquakes in Moldo- GDP affected by earthquakes. va is about 20,000 and the annual ROMANA average affected GDP about $40 When an earthquake has a 10-year million. The annual averages of fa- Tighina return period, it means the prob- talities and capital losses caused by ability of occurrence of an earth- earthquakes are about 20 and about quake of that magnitude or greater $50 million, respectively. The Fatal- is 10 percent per year. A 100-year ities and capital losses caused by earthquake has a probability of more intense, less frequent events occurrence of 1 percent per year. can be substantially larger than This means that over a long period the annual averages. For example, of time, an earthquake of that mag- an earthquake with a 0.4 percent auBLACK SEA nitude will, on average, occur once annual probability of occurrence (a every 100 years. It does not mean 250-year return period event) could a 100-year earthquake will occur cause about $4 billion in capital loss exactly once every 100 years. In (about 60 percent of GDP). M Ol ovaWORLDBANKGROUP EL|GFDRR "AND CENTRAL A51A(ECA) EARTHQUAKE EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS (MILLIONS $) ANNUAL AVERAGE FATALITIES he rose diagrams show the provinces with the potential 0 for greatest annual average capital losses and highest annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital Gagauzia Tighina 0 agauzia 1 loss occurs in Chisinau, which is not surprising, given the economic importance of the province. cz %r EARTHQUAKE EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 ,EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 he exceedance probability curves display the GDP 1 affected by, respectively, floods and earthquakes for 3.5 30 varying probabilities of occurrence. Values for two different time periods are shown. A solid line depicts the affected GDP for 2015 conditions. A diagonally striped hand depicts 2080 2.5 20the range of affected GDP based on a selection of climate a20 and socioeconomic scenarios for 2080. For example, if Mol- 2r 1dova had experienced a 100-year return period flood event 1.5 in 2015, the affected GDP would have been an estimated 1.0 cu 10 2080 $500 million. In 2080, however, the affected GDP from the 2015 same type of event would range from about $2 billion to 22015 --0.5 about $3 billion. If Moldova had experienced a 250-year 2015 earthquake event in 2015, the affected GDP would have 10 50 100 250 10 50 100 250 been about $4 billion. In 2080, the affected GDP from the Return period (years) Return period (years) same type of event would range from about $20 billion to 10 - 2 1 0.4 10 2 1 04 about $30 billion, due to population growth, urbanization, Probabihty (%) Probability(%) and the increase in exposed assets. All historical data on floods and earthquakes are from D. Guha-Sapir, R. Below, and Ph. Hoyois, EMDAT: International Disaster Database (Universiti Catholique de Louvain, Brussels, Belgium), www.emdatbe, and J. Daniell and A.Schae- fer, "Eastern Europe and Central Asia Region Earthquake Risk Assessment Country and Province Profiling," final report to GFDRR, 2014. Damage estimates for all historical events have been inflated to 2015 US$. WORLDBANKGROUP EI GROPE AND CENTRALA5IA (ECA) AFFECTED AFFECTED CAPITAL LOSS Q GFDRR RI5K PROFILES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKEt GDP $4.0 billion* G i Montenegro ROS NIA AND HERBZEFC A M ontenegro's population and and agriculture making a small economy are exposed to earth- contribution. Montenegro's per S ER BIA quakes and floods, with floods capita GDP was $6,470. posing the greater risk. The model results for present-day risk shown in This map displays GDP by prov- this risk profile are based on population ince in Montenegro, with great- P[jevlja and gross domestic product (GDP) esti- er color saturation indicating mates for 2015. The estimated damage greater GDP within a province. caused by historical events is inflated to he blue circles indicate the risk 2015 US dollars. of experiencing floods and the ora. nge ci rcl es the risk of earth - amke(ni Just over 60 percent of Montenegro's quakes in terms of noralized population lives in urban environments. annual avenge of affected GDP. Savnik 4ojkovac The country's GDP was approximately The largest circles represent the US$4.0 billion in 2015, with close to 90 greatest normalized risk- The Berane Ro7aj percent derived from services, most of risk is estimated using flood and the remainder generated by industry, earthquake risk models. The table displays the provincesAndrjvc at greatest normalized risk for /Annual Average of Affected GWN (% TOP AFFECTED PROVINCES each peril. In relative terms, as 10 -K> shown in the table, the province 00~I FLO Ar> at greatest risk of floods is BijeOH'rceg iNoERTvUK EATQAE PoIje, and the one at greatest 1ARHQ-K EARTHQUAKE ANNUAL AVERAGE OF ANNUAL AVERAGE OF risk of earthquakes is Budva. In AFFECTED GDP (%) AFFECTED GDP (%) absolute terms, the province at Negligible greatest risk of both floods andBuv Bijelo Polje 11 Budva 4 earthquakes is Podgorica. Danilovgrad Ulcinj 3 GOP (billions of$) Berane 6 Darilovgrad 2 Andrijevica 5 Kotor 2 There is a high correlation Podgorica 3 Bar 2 (r=0.95) between the Play 1 Podgorica 2 ADRIA[IC SEA populationand GDPoa Ulcinj 0 Tivat 2 province. Bar 0 Cetinje 2 Pluzine 0 Niksic 2 Cetinje ea Herceg Novi 1 M onte egroWORLDBANKGROUP E|G DR ROPE" AND EENTRAL A51A(ECA) F loods in 2010, declared by the If the 10- and 100-year bars are the Affected GDP (%) for country's government as the same height, then the impact of a 10- BOSNIA AND HER Z EGO VINA 10 and 100-year return periods "worst floods ever recorded,"1 year event is as large as that ofa 100- One block = 5% 40 affected over 5,000 people in Mon- year event, and the annual average of tenegro. In 2014, a devastating flood affected GDP is dominated by events 20 in the Balkans also affected Monte- that happen relatively frequently. Annua average 10 negro, although at the time of this If the impact of a 100-year event is S ER BIA publication its impact had not been much greater than that of a 10-year 10-year 100-year quantified. event, then less frequent events make a larger contribution to the annual This map depicts the impact of flood- average of affected GDP. Thus, even Pijevija ing on provinces' GDPs, represented if a province's annual affected GDP Annual Average of Affected GOP (%) as percentages of their annual aver- seems small, less frequent and more age GDPs affected, with greater color intense events can still have large saturation indicating higher percent- impacts. Z Plu7ab[jak ages. The bar graphs represent GDP affected by floods with return periods The annual average population --- of 10 years (white) and 100 years affected by flooding in Montenegro is Savnik Mojkovac (black). The horizontal line across the about 10,000 and the annual average bars also shows the annual average of affected GDP about $90 million. With- Rozaj GDP affected by floods. in the various provinces, the 10- and Niksic 100-year impacts do not differ much, Kotasin When a flood has a 10-year return so relatively frequent floods have period, it means the probability of large impacts on these averages. occurrence of a flood of that magni- osovo tude or greater is 10 percent per year. Kotor Cetinje A 100-year flood has a probability of occurrence of 1 percent per year. Herceg Novi This means that over a long period of time, a flood of that magnitude will, on average, occur once every 100 years. It does not mean a 100-year Budva flood will occur exactly once every 100 years. In fact, it is possible for a flood of any return period to occur ADRIATIC SEA Bar more than once in the same year, or ALBANIA to appear in consecutive years, or not to happen at all over a long period of time. Ulcinj EL ROP AND CENRA ASI (ECA) M onteneg ro WORLDBANKGROUP @ GFDRR RS ANOFIESETAA5AEA ontenegro's worst earth- year, or to appear in consecutive ad GOP (%) for quake since 1900 took years, or not to happen at all over a BOSNIA AND ERZEGOVINA 10 100 place in 1979. The earth- long period of time O quake caused over 120 fatalities and close to $13 billion in damage. Other earthquakes affecting Mon- the same height, then the impact Annua average tenegro during the twentieth cen- of a 10-year event a r e tury occurred in 1905 and 1968. thtoa10-ervn,adte trocurdi195ad16. annual average of affected GDP is 10-year 100-year This map depicts the impact of dominated by events that happen earthquakes on provinces' GDPs, relatively frequently. If the impact Pja jf represented as percentages of ofa 100-year event is much great- their annual average GDPs affect- erthan that of a 10-year event, 4 ed, with greater color saturation then less frequent events make indicating higher percentages. The larger contributions to the annual bar graphs represent GDP affected avenge of affected GDP Thus, even by earthquakes with return peri- if a province's annual affected GDP ods of 10 years (white) and 100 seems small, less frequent and ? kovac years (black). The horizontal line more intense events can still have across the bars also shows the large impacts. Berane Rozaj annual average of GDP affected by eathuke.The annual average population af-Wai earthquakes. fected by earthquakes in Montene- 'Andrijovica ~- When an earthquake has a 10- gro is about 9,000 and the annual year return period, it means the avenge affected GDP about $70 KOSOVO probability of occurrence of an million. The annual avenges of PLay earthquake of that magnitude or fatalities and capital losses caused greater is 10 percent per year. A by earthquakes are about eight Herr g 100-year earthquake has a prob- and about $10 million, respec- ability of occurrence of 1 percent tively. The fatalities and capital per year. This means that over a losses caused by more intense, less long period of time, an earthquake frequent events can be substantial- of that magnitude will, on average, ly larger than the annual averages. occur once every 100 years. It does For example, an earthquake with not mean a 100-year earthquake a GA percent annual probability will occur exactly once every 100 of occurrence (a 250-year return ALBANIA years. In fact, it is possible for an period event) could cause $400 earthquake of any return period to million in capital loss (about 10 occur more than once in the same percent of GOP). M onteneg-WORLDBANKGROUP ER "AD ENTRALA51A(ECA) EARTHQUAKE EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS ($) ANNUAL AVERAGE FATALITIES 1, co he rose diagrams show the provinces with the potential ofor greatest annual average capital losses and highest annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital Kotor 500,00 )novgrad Tivat 0 Danilovgrad 0 loss occurs in Podgorica, which is not surprising, given the economic importance of the province. Z4Sa Pn7erskaya, Samiarskaya affected by earthquakes.P0LANDBrasaaLptky Tmbvaa Whn neatquk hsa 0yer fatalities and capital losses caused BYnYaLetky Trbokvi' _ frnJgkyt Novo5ibrskaya When an earthquake has a 10-yearand Kurskayz iratovskaya, return period, it means the probabil- about $400 million, respectively. The Altayskiy ity of occurrence of an earthquake fatalities and capital losses caused [AKgradskaya K_AL AKHST_AN of that magnitude or greater is 10 by more intense, less frequent events ;R6bsiovskaYa,- percent per year. A 100-year earth- can be substantially largerthan the Amaklwnrka quake has a probability of occurrence annual averages. For example, an [A[rnkiya of 1 percent per year. This means earthquake with a 0.4 percent annual Krasnctlr.ki that over a long period of time, an probability ofoccurrence (a 250-year Adygeya earthquake of that magnitude will, on r peio e c S-AK average, occur once every 100 years. about 10,000 fatalities and $6 billion DagestaL It does not mean a 100-year earth- in capital loss (less than 1 percent of Sebarnino-balkariyaaAkiSevernaEa _AA quake will occur exactly once every GDP). T 100 years. In fact, it is possible for WORLIDBANK GROUP EL GFDRR AFND CEN TRAL A5]A (ECA) Taymyrskiy Sakha \kaChukotskiy Evenkiyskiy Maqadanskaya e - Koryaksk i Krasnoyarskiy Kemerovskaya r Khahirovskiy U st Burki Cht yAmurskaya Tyva Sakhali a Khakasiya Atay Affected GDP (%) for AnnuaL Average of Affected GDP (%) 10 and 100-year return periods One block =10% 100 Primorskiy 50 AnnLual average --L_ 20 10-year 100-year Rus ian Fed rat onWORLDBANKGROUP E|G DR RO" AN CENTRAL A51A(ECA) EARTHQUAKE K"\ EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS (MILLIONS $) ANNUAL AVERAGE FATALITIES he rose diagrams show the provinces with the potential for greatest annual average capital losses and highest annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital Chechnya Rep. D estan Rep. 40 Severnaya estan Re. loss occurs in Sakhalinskaya Oblast, which is not surprising given the economic importance of the province. EARTHQUAKE he exceedance probability curves display the GDP EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 T affected by, respectively, floods and earthquakes for 400 120 varying probabilities of occurrence. Values for two different time periods are shown. A solid line depicts the affected r 100 GDP for 2015 conditions. A diagonally striped band depicts 2080 300 the range of affected GDP based on a selection of climate 80 and socioeconomic scenarios for 2080. For example, if the 208 3 60Russian Federation had experienced a 100-year return pe- 200 r 60 riod flood event in 2015, the affected GDP would have been 0O an estimated $60 billion. In 2080, however, the affected 4o 0 2015 GDP from the same type of event would range from about 20 $200 billion to about $400 billion. If the Russian Federation had experienced a 250-year earthquake event in 2015, the affected GDP would have been an estimated $20 billion. 10~~~ 50 0 25 0 10 5 2eturn perd (years) Return period (years) In 2080, however, the affected GDP from the same type of - -- - event would range fromi about $50 billion to about $100 10. 10 2 1 0 billion, due to population growth, urbanization, and the Pro ba bility (h robabi[ity(% increase in exposed assets. All historical data on floods and earthquakes are from D. Guha-Sapir, R. Below, and Ph. Hoyois, EM-DAT: International Disaster Database (Universit6 Catholique de Louvain, Brussels, Belgium), www.emdat.be, and J. Daniell arid A.Schae- fer, "Eastern Europe and Central Asia Region Earthquake Risk Assessment Country and Province Profiling," final report to GFDRR, 2014. Damage estimates for all historical events have been inflated to 2015 US$. WORLDBANKGROUP ER EROPE AND CENTRALA5IA (ECA) AFFECTED AFFECTED CAPITAL LOSS Q GFDRR RI5K PROFILES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $36.4 billion* I Serbia HUN GAR Y erbia's population and economy agriculture making a small con- Annual Average of Affected GOP are exposed to earthquakes and tribution. Serbia's per capita GDP 0 floods, with floods posing the was rn nk40r greater risk The model results for pres- -his m EARTHQUAKE ent-day risk shown in this risk profile apadno-bac are based on population and gross ince in Serbia, with greater color domestic product (GDP) estimates for saturation indicating greater GDP 0 Negligible 2015. The estimated damage caused by within a province. The blue circles znj-atc historical events is inflated to 2015 US indicate the risk of experiencing dollars. floods and the orange circles the GDP (billions of $) ri sk of earth quake s in term s o f Senk zo n Just over 55 percent of Serbia's pop- normalized annual average of Bsk ulation lives in urban environments. affected GDP The largest circles The country's GDP was approximately represent the greatest normalized US$36.4 billion in 2015, with about 60 risk The risk is estimated using percent derived from services, most flood and earthquake risk models. ojai Bracvki Thers is a high correlation of the rest generated by industry, and 0 (r=095) between ths The table displays the provinces Kolubrk BorskipolaonndGPfa at greatest normalized risk for A pounaidGPos each peril. In relative terms, as -ymdlk TOP AFFECTED PROVINCES shown in the table, the pro.- P ince at greatest risk of floods is Moraiki '4 Severno-banatski, and the one 7~ ajecarski 4 EARTHQUAKE at greatest risk of earthquakes t< I alt i b'o. FLOO EARTHQUAKE N 0 a U4, ANNUAL AVERAGE OF ANNUAL AVERAGE OF Kayseri' EiVa AFFECTED GDP (%) AFFECTED GDP (%)Aksara platy' Koy S \Niqde Dyrai Bartin 8 Izmir 4 imay Batman Siit Adana Bursa 3 Amasya 4 Yalova 3 Mardi Srk Ha Bayburt 3 Istanbul 2 Kar 0 A n 7aI liurfa S LAMIC Tokat Canakkale 2 KiPUBLIC Agri Manisa 2 Karabuk Edirne 2 There sahighcorreation Igdir Erzurum 2 (r0.95) between he populatin Hatay 2 Van 2 and GDP at a province Sakarya 2 lgdir 2 TurkeyWORLDBANKGROUP E|GDR ROPE AND EENTRAL A51A(ECA) T he most devastating flood in Turkey event is much greater than that of a 10- have large impacts on these averages. For Affected GOP (%) for since 1900 occurred in 1998. It year event, then less frequent events make the few in which the 100-year impacts are 10 and 100-year return periods affected over 1 million people and a larger contribution to the annual average much greater than the 10-year impacts, One black = 5% 30 caused over $1 billion in damage. Flood- of affected GDP. Thus, even if a provinces less frequent events make a significant ing in 2006 caused almost $400 million annual affected GDP seems small, less contribution to the annual avenge of in damage, while further floods in 2009 frequent and more intense events can still affected GDP Annua[ average caused about $600 million in damage. have large impacts. This map depicts the impact of flooding on The annual avenge population affected by 10-year 100-year provinces' GDPs, represented as percent- flooding in Turkey is about 600,000 and ages of their annual average GDPs affected, the annual avenge affected GDP about $5 Annual Average of Affected GOP (%) with greater color saturation indicating billion. For most provinces, in which the higher percentages. The bar graphs rep- impacts from 10- and 100-year floods do resent GDP affected by floods with return not differ much, relatively frequent floods periods of 10 years (white) and 100 years (black). The horizontal line across the BLACK SEA bars also shows the annual average of GDP affected by floods. ..Bartin -EORGIA When a flood has a 10-year return period, utdak it means the probability of occurrence of K a flood of that magnitude or greater is 10 percent per year. A 100-year flood has a Istanbul b k qArtvin probability of occurrence of t percent per - a zon Sakrya Bol 5d _irpunAZ ERRBAl1A0 year. This means that over a long period ____ - Ankr Tokattr G-umvu5haneBybr of time, a flood of that magnitude will, on _ar average, occur once every 100 years. It Bur ik does not mean a 100-year flood will occur r ir, Y07gat Sivas exactly once every 100 years. In fact, it is K Kiriehir possible for a flood of any return period to occur more than once in the same year, or U Nfyhi Kayse i --latya to appear in consecutive years, or not to AkiBii V happen at all over a long period of time. D![i DiygrmAdir If the 10- and 100-year bars are the same T Brrardin ak Hakkari height, then the impact of a 10-year event KararAiaySanliurfanyeGaPaUe8SLlirC is as large as that of a 100-year event, Of IRAN and the annual average of affected GDP is dominated by events that happen relative- ly frequenty. If the impact of a 100-year Tu kyWORLDBANKGROUP GFDR sROP ANE CENTRAL A5IA(ECA) Tu ke ak L.R RISK PROILE T urkey's most deadly earthquake If the 10- and 100-year bars are the same The annual average population affected by Affected GOP (%) for since 1900 took place in 1939 height, then the impact of a 10-year event earthquakes in Turkey is about 1 million 10 and 100-year return periods in Erzincan, with a magnitude of is as large as that of a 100-year event, and the annual average affected GDP $10 One block = 10% 100 7.7. It caused more than 30,000 fatali- and the annual average of affected GDP is billion. The annual averages of fatalities ties and over $300 million in damage. A dominated by events that happen relative- and capital losses caused by earthquakes A0 1999 earthquake with a magnitude of 7.6 ly frequently. If the impact of a 100-year are about 1,000 and about $2 billion, re- caused nearly 18,000 deaths, affected over event is much greater than that of a 10- spectively. The fatalities and capital losses 1 million people, and caused close to $30 year event, then less frequent events make caused by more intense, less frequent 10-year 100-year billion in damage. larger contributions to the annual avenage events can be subs tantially larger than the of affected GP. Thus, even if a province's annual averages. For example, an earth-GDP (%) This map depicts the impact of earth- annual affected GDP seems small, less quake witha 0.4 percent annual probabili- Ana vrg fAfce O % quakes on provinces'GDPs, represented frequent and more intense events can still ty of occurrence (a 250-year return period Lt d as percentages of their annual avennge have large impacts. event) could cause about 80,000 fatalities GDPs affected, with greater color satura- and $60 billion in capital loss (about 8 ~ 6 tion indicating higher percentages. The percent of GP). bar graphs represent GP affected by eartARIA earthquakes with return periods of 10 BLACK SEA years (white) and 100 years (black). The horizontal line across the bars also shows CE OR CIA the annual average of GDP affected by earthquakes. When an earthquake has a 10-year return Edi period, it means the probability of oc- currence of an earthquake of that mag- nitude or greater is 10 percent per year. *Ankaa A 100-year earthquake has a probability n A ir of occurrence of 1 percent per year. This i ir means that over a long period of time, Kutah a 1(irsehi an earthquake of that magnitude will, on average, occur once every 100 years. It does not mean a 100-year earthquake will occur exactly once every 100 years. In fact, it is possible for an earthquake of any y return period to occur more than once in MardAn the same year, or to appear in consecutive Ice[ KamnOmni years, or not to happen at all over a long REiii period ofAtime. OF IRAN and $60 billion in capitallos aou TureyWORLDBANKGROUP |GFDRRP s"AND CENTRAL A51A(ECA) EARTHQUAKE EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS (MILLIONS $) ANNUAL AVERAGE FATALITIES T he rose diagrams show the provinces with the potential for greatest annual average capital losses and highest annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital Balikesir 30 B 100 Balikesir 3 li 100 loss occurs in Istanbul, which is not surprising, given the economic importance of the province. 0 /4 -1- EARTHQUAKE EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 he exceedance probability curves display the GDP Taffected by, respectively, floods and earthquakes for 160 2,500 varying probabilities of occurrence. Values for two different time periods are shown. A solid line depicts the affected 14 2,00GDP for 2015 conditions. A diagonally striped band depicts 120 20 the range of affected GDP based on a selection of climate 2080 .2 and socioeconomic scenarios for 2080. For example, ifTur- 100 1500 key had experienced a 100-year return period flood event in 2015, the affected GDP would have been an estimated 6o d$20 billion. In 2080, however, the estimated affected GDP from the same type of event would range from about $80 2015 500 2015 billion to about $140 billion. If Turkey had experienced 20 a 250-year earthquake event in 2015, the affected GDP would have been about $300 billion. In 2080, the affected tu pe d 0 (y50 ) st o i 0d a250' GDP from the same type of event would range from about ... .$1 trillion to about $2 trillion, due to population growth, 0.4 urbanization, and the increase in exposed assets. P ro ba bility I(%) Proba bi lity I(%) All historical data on floods and earthquakes are from, respectively, DG uha-Sapir, R Below, and Ph. Hoyois, EM-DAT: International Disaster Database (Universit6 Catholique de Louvain, Brussels, Belgium), www.emdat.be, and the National Geophysical Data Center/World Data Servite (NGDC/WDS), Significant Earthquake Database (National Geophysical Data Center, NOAA), doi:107289/V5TD9V7K. Damage estimates for all historical events have been inflated to 2015 US$. WORLDBANKGROUP GR EROPE AND CENTRALA5IA (ECA) AFFECTED AFFECTED CAPITAL LOSS QRI5KPROFILES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $50.0 billion* Turkmernistan o KAZAKHSTAN T urkmenistan's population close to 50 percent derived from and economy are exposed to industry, most of the remainder earthquakes and floods, with generated by services, and agricul- earthquakes posing the greater risk ture making a small contribution. UZBEKI5TAN of a high impact, lower probability Turkmenistan's per capita GDP event. The model results for pres- was $9,230. ent-day risk shown in this risk profile o are based on population and gross This map displays GDP by prov- Tashauz domestic product (GDP) estimates ince in Turkmenistan, with greater for 2015. The estimated damage color saturation indicating greater caused by historical events is inflated GDP within a province. The blue to 2015 US dollars. circles indicate the risk of expe- riencing floods and the orange Just over half of Turkmenistan's pop- circles the risk of earthquakes in ae s ulation lives in rural environments. terms of normalized annual av- The country's GDP was approxi- erage of affected GDP. The largest mately US$50.0 billion in 2015, with circles represent the greatest nor- malized risk The risk is estimated using flood and earthquake risk models. TOP AFFECTED PROVINCES M a ry The table displays the provinces at greatest normalized risk for each peril. In relative terms, as FLOO EARTHQUAKE lisA)EARTHUAKE shown in the table, the province at ANNUAL AVERAGE OF ANNUAL AVERAGE OF AFFECTED GDP (%) AFFECTED GDP (%) greatest risk of floods is Char- zhou, and the one at greatest risk ISLAMIC REPUBLC OF IRAN Chardzhou 2 Turkmenistan 4 of earthquakes is Turkmenistan A S N A N Turkmenistan 1 Territories Territories. In absolute terms, it is Annual Average of Affected GDP (%) GDP (billions of $ TerritoriesTurkmenistan Territories. Mary Mary 1 4 There is a high correlation Tashuz Tahati 0 000(r=0.9) between the ' U EARTHQUAKE population and GDP ofta province. o Negligible Turkm nista WORLDBANKGROUP Q |G DR EIsROP AND EENTRAL A51A(ECA) he worst flood in Turkmeni- affected GDP is dominated by events K AZAK HSTAN stan since it gained its inde- that happen relatively frequently. pendence in 1991 occurred in If the impact of a 100-year event is 1993 and caused about $200 million much greater than that of a 10-year in damage. event, then less frequent events make a larger contribution to the annual This map depicts the impact of flood- average of affected GDP. Thus, even ing on provinces' GDPs, represented if a province's annual affected GDP as percentages of their annual aver- seems small, less frequent and more age GDPs affected, with greater color intense events can still have large saturation indicating higher percent- impacts. ages. The bar graphs represent GDP affected by floods with return periods The annual average population affect- Tshauz of 10 years (white) and 100 years ed by flooding in Turkmenistan is (black). The horizontal line across the about 70,000 and the annual average bars also shows the annual average of affected GDP about $700 million. GDP affected by floods. When a flood has a 10-year return 4 period, it means the probability of Turkmenistan Territories occurrence of a flood of that magni- Chardzhou tude or greater is 10 percent per year. A 100-year flood has a probability Ashgabat of occurrence of 1 percent per year. This means that over a long period of time, a flood of that magnitude will, on average, occur once every 100 years. It does not mean a 100-year flood will occur exactly once every 100 years. In fact, it is possible for a ISLAMIC REPLI OF IRAN flood of any return period to occur more than once in the same year, or Affected GDP (%) for to appear in consecutive years, or not 10 and 100-year return periods to happen at all over a long period of One block = 1% time. 10 Annual Average of Affected GDP (%) If the 10- and 100-year bars are the 5 same height, then the impact of a 10- Annualaverage year event is as large as that of a 100- year event, and the annual average of 10-year 100-year o S C & Turk eni tanWORLDBANKGROUP EL GF RROP AND ECENTRALA5IA(EICA) T urkmenistan's worst earth- appear in consecutive years, or not K AZAK HSTAN quake since 1900 took place to happen at all over a long period in 1948 in Aschgabad, with a of time. magnitude of 7.3. It caused any- where from 50,000 to over 100,000 If the 10- and 100-year bars are the fatalities and almost $4 billion in same height, then the impact of a damage. Other significant earth- 10-year event is as large as that of a quakes affecting Turkmenistan 100-year event, and the annual aver- occurred in 1895, 1929, and 1946. age of affected dPQ s4 ominated by events that happen relatively fre- UZBEKI5TAN This map depicts the impact of quently. If the impact of a 100-year earthquakes on provinces' GDPs, event is much greater than that of represented as percentages of their a 10-year event, then less frequent annual average GDPs affected, with events make larger contributions to greater color saturation indicating the annual average of affected GDR higher percentages. The bar graphs Thus, even if a province's annual represent GDP affected by earth- affected GDP seems small, less fre- quakes with return periods of 10 quent and more intense events can years (white) and 100 years (black). still have large impacts. The horizontal line across the bars also shows the annual average of The annual average population GDP affected by earthquakes. affected by earthquakes in Turk- menistan is about 100,000 and the When an earthquake has a 10-year annual average affected GDP about return period, it means the prob- $2 billion. The annual averages of ability of occurrence of an earth- fatalities and capital losses caused quake of that magnitude or greater by earthquakes are about 300 and is 10 percent per year. A 100-year about $200 million, respectively. earthquake has a probability of The fatalities and capital losses occurrence of 1 percent per year. caused by more intense, less fre- I LAMICREPUBLEC OF IRAN This means that over a long period quent events can be substantially of time, an earthquake of that mag- larger than the annual averages. For Affected GDP (%) for nitude will, on average, occur once example, an earthquake with a 0.4 10 and 100-year return periods A every 100 years. It does not mean percent annual probability of oc- One block = 10% a 100-year earthquake will occur currence (a 250-year return period Af exactly once every 100 years. In event) could cause about 10,000 40 fact, it is possible for an earthquake fatalities and $5 billion in capital Annual average 20 of any return period to occur more loss (about 10 percent of GDP). than once in the same year, or to 10-year 100-year o Turk eni tanWORLDBANKGROUP |GFDRRE "AND CENTRAL A51A(ECA) EARTHQUAKE EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS (MILLIONS $) ANNUAL AVERAGE FATALITIES T he rose diagrams show the provinces with the potential for greatest annual average capital losses and highest annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital loss occurs in Turkmenistan Territories, which is not sur- prising, given the economic importance of the province. ( .EARTHQUAKE EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 he exceedance probability curves display the GDP T affected by, respectively, floods and earthquakes for 40 160 varying probabilities of occurrence. Values for two different 35 140 time periods are shown. A solid line depicts the affected ( 460 GDP for 2015 conditions. A diagonally striped band depicts 3 1the range of affected GDP based on a selection of climate 208025 100 2080 and socioeconomic scenarios for 2080. For example, if 20 80 Turkmenistan had experienced a 100-year return period 5 0flood event in 2015, the affected GDP would have been an 15 60estimated $3 billion. In 2080, however, the affected GDP 10 40 from the same type of event would range from about $10 2015 5 20 billion to about $30 billion. If Turkmenistan had experi- enced a 250-year earthquake event in 2015, the affected 10 50 100 '1250 10 50 100 250 GDP would have been about $30 billion. In 2080, the Retu period (years) Return period (years) affected GDP from the same type of event would range from 10 2 1 0.4 10 2 1 04 about $90 billion to about $140 billion, due to population Probability%) robability (%) growth, urbanization, and the increase in exposed assets. All historical data on floods and earthquakes are from, respectively, D Guha-Sapir, R. Below, and Ph. Hoyois, EM-DAT: International Disaster Database (Universit6 Catholique de Louvain, Brussels, Belgium), www.emdat.be, and L Daniell and A. Schaefer, "Eastern Europe and Central Asia Region Earthquake Risk Assessment Country and Province Profiling," final reportto GFDRR, 2014. Damage estimates for all historical events have been inflated to 2015 US$. WORLDBANKGROUP GR EROPE AND CENTRALA5IA (ECA) AFFECTED AFFECTED CAPITAL LOSS Q GFDRR RI5K PROFILES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $89.2 billion* U krai ne 0.ouain4. ilo* M- B0 E L A I RELARUS U kraine's population and econo- and agriculture making a small my are exposed to earthquakes contribution. Ukraine's per capita RUS N EDEATON and floods, with floods posing GDP was $1,990. the greater risk. The model results for present-day risk shown in this risk pro- This map displays GOP by prov- POLAN file are based on population and gross ice in Ukraine, with greater color domestic product (GDP) estimates for saturation indicating g G'P 2015. The estimated damage caused by within a province. The blue circles historical events is inflated to 2015 US indicate the risk of experiencing Zhytomyrs'ka dollars. floods and the orange circles the ri sk of earth quakes in term s o f nor-Khenyik Close to 70 percent of Ukraine's pop- malized annual average of affected Lvlty'kk Kk ulation lives in urban environments. GDP. The largest circles represent Cherkaska The country's GDP was approximately the greatest normalized risk. The VinnytSka u,nk US$89.2 billion in 2015, with about 60 risk is estimated using flood and ano-rankaft percent derived from services, most of earthquake risk models. the remainder generated by industry The table displays the provinces at gre atest norm alize d risk fo r each MyUA> 'Ikolayiv5'ka0 TOP AFFECTED PROVINCES ~ peril. In relative terms, as shown in0 dla - i10wI TOP AFFECTED PROVINCESthe table, the province at greatest Khersonka xirisk of floods is Kharkivs'ka, and the R one at greatest risk of earthquakes EARTHQUAKE is Krym. In absolute terms, the ANNUAL AVERAGE OF ANNUAL AVERAGE OF province at greatest risk of floods is AFFECTED GDP (%) AFFECTED GDP (%) Kharkivs'ka, and the one at greatest Kharkvs'k 3 Kym 2 risk of earthquakes is Odes'ka. Kharkivs'ka Krym 2 Zakarpats'ka Odes'ka 1 BLACK SEA Kirovohrads'ka Zakarpats'ka 1 Annual Average of Affected GOP ( GP (billions of $) Luhans'ka 2 Chernivets'ka 1 Chernihivs'ka Ivano 1 1 Theeisahghcorelatin Poltavs'ka -frankivs'ka (r=095) between the Vinnyts'ka 2 Vinnyts'ka 0 population and GDP Gf a Mykolayivs'ka L'vivs'ka 0 EARTHQpArE Khersons'ka - Ternopit's'ka 0 Zhytomyrs'ka 2 Khmel'nyts'ka 0 0 Neg[igib[e Khersonsnka 0 UIraneWORLDBANKGROUP E|GDR ROPE AND CENTRAL A51A(ECA) he most devastating flood in to happen at all over a long period of Ukraine since it gained its in- time. dependence in 1991 occurred in 2008, causing nearly 40 fatalities If the 10- and 100-year bars are the and about $1 billion in damage. same height, then the impact ofa 10- BEELARUS Flooding in 1993 caused about $300 year event is as large as that of a 100- million in damage, and a 1998 flood year event, and the annual avenge of Ididnal 0pol.Frhraffected GDP is dominated by events RUSSIAN FEDERATION killed nearly 20 people. Further flooding in 2006 and 2013 caused that happen relatively frequently no fatalities but over $20 million in If the impact of a 100-year event is damage in each year. much greater than that of a 10-year event, then less frequent events make WC'hernihivs'ka A This map depicts the impact of flood- a larger contribution to the annual ROLAND Simsnka ing on provinces' GDPs, represented average of affected GDP. Thus, even Zhytomyrs'ka as percentages of their annual aver- if a provinces annual affected GDP age GDPs affected, with greater color seems small, less frequent and more saturation indicating higher percent- intense events can still have large U 4 Khmel nts k ages. The bar graphs represent GDP impacts. Ternopits ka h affected by floods with return periods of 10 years (white) and 100 years The annual avenge population affect- (black). The horizontal line across the ed by flooding in Ukraine is about Zakarpats'ka K bars also shows the annual average of 600,000 and the annual average GDP Chnntstka GDP affected by floods. about $1 billion. Within the various provinces, the 10- and 100-year im- Mykoa0is'A When a flood has a 10-year return pacts do not differ much, so relatively period, it means the probability of frequent floods have large impacts on Odeska rsons'ka occurrence of a flood of that magni- these averages. R tude or greater is 10 percent per year. A 100-year flood has a probability of occurrence of 1 percent per year. Krym This means that over a long period of time, a flood of that magnitude will, on average, occur once every 100 years. It does not mean a 100-year Affected GOP for Annual Average of Affected GOP flood will occur exactly once every 10 and 100-year return periods 100 years. In fact, it is possible for a One h[ack = 1% 10 flood of any return period to occur 5 o 6B more than once in the same year, or Annual average 2 to appear in consecutive years, or not 10-year 100-year UIkr ineWORLDBANKGROUP EL GF RROP AND ECENTRAL A5IA(ECA) Ukr in @AkR RIS[ PROFILES U kraine's worst earthquake years, or not to happen at all over a since 1900 took place in 1927 long period of time. in Crimea, with a magnitude of 6.8. It caused about 15 fatalities If the 10- and 100-year bars are the and close to $200 million in damage. same height, then the impact ofa 10- Other major earthquakes affecting year event is as large as that of a 100- Ukraine occurred in 1170 in Kiev and year event, and the annual avenge of in 1751 and 1872 in Crimea. affected GDP is dominated by events RUSSIAN FEDERATION that happen relatively frequently. This map depicts the impact of If the impact ofa 100-year event is earthquakes on provinces' GDPs, much greater than that of a 10-year represented as percentages of their event, then less frequent events make Cumika annual average GDPs affected, with larger contributions to the annual av- POLAND greater color saturation indicating erage of affected GDP. Thus, even if a Zhytomyrska. higher percentages. The bar graphs province's annual affected GDP seems T K k represent GDP affected by earth- small, less frequent and more intense r quakes with return periods of 10 events can still have large impacts. - Psla4vska ) Kharkivsvkk years (white) and 100 years (black). 'TernopilVka Th hriona lneacos hebas The annual average population affect- / t - " Cherkas'ka',Lhask The horizontal line across the bars Vinnyts'ka Lu ha ns1a also shows the annual average of GDP ed by earthquakes in Ukrine is about affected by earthquakes. 1akarpati'ka l Dnipropetrovs!ka affected GDP about $100 million. fhriesaDonets'ka When an earthquake has a 10-year The annual averages of fatalities return period, it means the probabil- and capital losses caused by earth- 0ykolayivs'ka ity of occurrence of an earthquake quakes are about 20 and about $60 Zaporizka of that magnitude or greater is 10 million, respectively The fatalities Odes,ka Khersonska percent per year. A 100-year earth- and capital losses caused by more quake has a probability of occurrence intense, less frequent events can be R M A N oft1 percent per year. This means sbtnilylre hnteana that over a long period of time, an averages. For example, an earthquake earthquake of that magnitude will, on with a 0.4 percent annual probability average, occur once every 100 years. of occurrence (a 250-yearreturn It does not mean a 100-year earth- period event) could cause about 700 quake will occur exactly once every fatalities and $2 billion in capital loss Affected GOP for Annual Average of Affected GOP (6) 100 years. In fact, it is possible for (about2 percent of GDP). 10 and 100-year return periods an earthquake of any return period One block = 5% to occur more than once in the same 30 BLACK SEA year, or to appear in consecutive Annual average 10-yea r 100-year UI ra neWORLDBANKGROUP GFDRRP s"AND CENTRAL A51A(ECA) EARTHQUAKE EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS (MILLIONS $) ANNUAL AVERAGE FATALITIES he rose diagrams show the provinces with the potential for greatest annual average capital losses and highest annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital Donetstka karps'ka 5 Luhans ka 0 karpats ka 1 loss occurs in Krym, which is not surprising, given the eco- nomic importance of the province. Sc 0 EARTHQUAKE EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 he exceedance probability curves display the GDP T affected by, respectively, floods and earthquakes for 45 120 varying probabilities of occurrence. Values for two different 40 time periods are shown. A solid line depicts the affected -1 100 GDP for 2015 conditions. A diagonally striped band depicts the range of affected GDP based on a selection of climate 30 2 80 2080 and socioeconomic scenarios for 2080. For example, if 25 0 Ukraine had experienced a 100-year return period flood 20 event in 2015, the affected GDP would have been an esti- 15 2080 mated $4 billion. In 2080, however, the affected GDP from 10 rthe same type of event would range from about $20 billion 2015 0to about $40 billion. If Ukraine had experienced a 250-year 2015 earthquake event in 2015, the affected GDP would have 10 5o 100 been about $6 billion. In 2080, the affected GDP from the eturn perod (ars) Return period (years) same type of event would range from about $40 billion to 110 2 1 about $100 billion, due to population growth, urbanization, Probability Probabtity(%) and the increase in exposed assets. All historical data on floods and earthquakes are from, respectively, D. Guha-Sapir, R. Below, and Ph Hoyois, EM-DAT: International Disaster Database (Universit6 Catholique de Louvain, Brussels, Belgium), www.emdat.be, and L Daniell and A. Schaefer, "Eastern Europe and Central Asia Region Earthquake Risk Assessment Country and Province Profiling," final report to GFDRR, 2014. Damage estimates for all historical events have been inflated to 2015 US$. WORLDBANKGROUP ER EROPE AND CENTRALA5IA (ECA) AFFECTED AFFECTED CAPITAL LOSS Q GFDRR RI5K PROFILES BY 100-YEAR BY 250-YEAR FROM 250-YEAR FLOOD EARTHQUAKE EARTHQUAKE GDP $66.0 billion* Uzbeki*stan '.Pplto 02rilo*M Uzhekistan's population and econ- (together about 80 percent) and omy are exposed to earthquakes agriculture generating the remain- and floods, with earthquakes der. Uzbekistan's per capita GDP K AZA KHS5TA N posing the greater risk of a high impact, was $2,190. lower probability event. The model results for present-day risk shown in This map displays GDP by prov- this risk profile are based on population ince in Uzbekistan, with greater and gross domestic product (GDP) esti- color saturation indicating greater mates for 2015. The estimated damage GDP within a province. The blueKa pksn caused by historical events is inflated to circles mndicate the risk of expe- Krklasa 2015 US dollars. reng hlndsofnda thq seing Ta nt man1 More than 60 percent of Uzbekistan's terms of normalized annual av- ,T."e CtKYGZRPBI population lives in rural environments. erage of affected GDP The largesta The country's GDP was approximately circles represent the greatest nor- Khorezm Tskent US$66.0 billion in 2015, with most malized risk. The risk is estimated derived from services and industries using flood and earthquake riskd Th table displays the provinces at 5UK E I T NS R 'knd zay TOP AFFECTED PROVINCES greatest normalized risk for each CH I NA peril. In relative terms, as shown ]:kaAJKSA Adp- in the table, the province at great- EARTHQUAKE est risk of floods is Andijan, and ha ANNUAL AVERAGE OF ANNUAL AVERAGE OF the one at greatest risk of earth - AFFECTED GDP () AFFECTED GDP () quakes is Namangan. In absolute Andijn 3 amanan 7 terms, the province at greatest Fergana 2 Andijan 7 risk of fl ood s is Fergana, a nd the A F CHA N Karakalpakstan 2 Fergana 6 one at greatest risk of earthquakes An nu al Average of Affe cte d G DP I(%) GDP (billions of $) Namangan 1 Ta sh ken t city 3 is Namangan. Sirdarya 1 Samarkand 3 7 There is a high correlation Bukhara I Tashkent 35 (r 0.95) between the Tashkent 1 Surkhanidarya 3 1 EARTHQUAKE f _p 16 p60population anid GDP of a Jizzakh 1 51ird arya 2 I LAM b RE UB I OF IR Nprovince. Samarkand 1 Jizzakh 2 C Kashkadarya I Kashkadarya 2 0 Negligible 0@127 Uzb kis anWORLDBANKGROUP EL GFDRPE" AND EENTRAL A51A(ECA) A flood that occurred in Uzbeki- If the impact of a 100-year event is stan in 2005 affected around much greater than that of a 10-year 1,500 people. event, then less frequent events make a larger contribution to the annual This map depicts the impact of flood- average of affected GDP. Thus, even ing on provinces' GDPs, represented if a province's annual affected GDP as percentages of their annual aver- seems small, less frequent and more age GDPs affected, with greater color intense events can still have large saturation indicating higher percent- impacts. ages. The bar graphs represent GDP affected by floods with return periods The annual average population affect- of 10 years (white) and 100 years ed by flooding in Uzbekistan is about (black). The horizontal line across the 400,000 and the annual average bars also shows the annual average of affected GDP about $800 million. For GDP affected by floods. most provinces, the 10- and 100- Karakalpakstan year impacts do not differ much, so When a flood has a 10-year return relatively frequent floods have large period, it means the probability of impacts on these averages. For the Kity occurrence of a flood of that magni- fewin which the 100-year impacts RNavoiyY tude or greater is 10 percent per year are much greater than the 10-year Ka Tashkent A 100-year flood has a probability impacts, less frequent events make a orezm TaihkInt of occurrence of 1 percent per year. significant contribution to the annual An n This means that over a long period of average of affected GDAn time, a flood of that magnitude will, Bukhara on average, occur once every 100 Samarkand years. It does not mean a 100-year CHINA flood will occur exactly once every KashkJ-ISTA 100 years. In fact, it is possible for a flood of any return period to occur 1A SA Fergana more than once in the same year, or Surkhandarya to appear in consecutive years, or not Affected GOP (%) for to happen at all over a long period of 10 and 100-year return periods time. One block= 1% 10 ArCHA N If the 10- and 100-year bars are the 5 Annual Average of Affected GDP (% same height, then the impact of a 10- year event is as large as that of a 100- Annual average 2 year event, and the annual average of affected GDP is dominated by events 10-year 100-year 0 1 d < that happen relatively frequently. Uzbek stanWORLDBANKGROUP EL GF RROP AND ECENTRAL A5IA(ECA) Uzbekistan VGDR RS[ PRFIE U zhekistan's worst earthquake than once in the same year, or to since 1900 took place in appear in consecutive years, or not 1902 in Andizhan, with a to happen at all over a long period magnitude of 6.4, and caused nearly of time. 5,000 fatalities. More recently, earthquakes in 1992 and 2011 If the 10- and 100-year bars are the caused approximately 10 fatalities same height, then the impact of a per event. Other major earthquakes 10-year event is as large as that of a KAZAKHSTAN affecting Uzbekistan occurred in 100-year event, and the annual aver- circa 838, 1966, and 1984. age of affected GDP is dominated by events that happen relatively fre- This map depicts the impact of quently. If the impact of a 100-year earthquakes on provinces' GDPs, event is much greater than that of represented as percentages of their a 10-year event, then less frequent annual average GDPs affected, with events make larger contributions to Karakalpakstan greater color saturation indicating the annual average of affected GDP. higher percentages. The bar graphs Thus, even if a province's annual Tashkent represent GDP affected by earth- affected GDP seems small, less fre- Tashkent Cty K quakes with return periods of 10 quent and more intense events can avoi U years (white) and 100 years (black). still have large impacts. Khorezm The horizontal line across the bars also shows the annual average of The annual average population af- GDP affected by earthquakes. fected by earthquakes in Uzbekistan a Sirdarya is about 1 million and the annual ara When an earthquake has a 10-year average affected GDP $2 billion. The return period, it means the prob- annual averages of fatalities and TURK MEN I TAN CHINA ability of occurrence of an earth- capital losses caused by earthquakes II TAN quake of that magnitude or greater are about 200 and about $900 is 10 percent per year. A 100-year million, respectively. The fatalities earthquake has a probability of and capital losses caused by more Affected GOP %forrya occurrence of 1 percent per year. intense, less frequent events can 10 and 100-year return periods This means that over a long period be substantially larger than the One block = 10% 100 of time, an earthquake of that mag- annual averages. For example, an AF G H A N nitude will, on average, occur once earthquake with a 0.4 percent every 100 years. It does not mean annual probability of occurrence (a ISLAME REPUBLIC OF IRAN a 100-year earthquake will occur 250-year return period event) could Annual average 2 exactly once every 100 years. In cause about 10,000 fatalities and fact, it is possible for an earthquake $10 billion in capital loss (about 20 10-year 100-year I m p C r of any return period to occur more percent of GDP). Uzb kis anWORLDBANKGROUP E|GFDRR s"AND CENTRAL A51A(ECA) EARTHQUAKE EARTHQUAKE ANNUAL AVERAGE CAPITAL LOSS (MILLIONS $) ANNUAL AVERAGE FATALITIES he rose diagrams show the provinces with the potential 14 T for greatest annual average capital losses and highest annual average numbers of fatalities, as determined using an earthquake risk model. The potential for greatest capital Navoiy An 100 Tashkent Cit 1nan n 20 loss occurs in Namangan, which is not surprising, given the economic importance of the province. EARTHQUAKE EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 EXCEEDANCE PROBABILITY CURVE, 2015 AND 2080 he exceedance probability curves display the GDP 1 affected by, respectively, floods and earthquakes for 40 250 varying probabilities of occurrence. Values for two different 35 time periods are shown. A solid line depicts the affected 200 GDP for 2015 conditions. A diagonally striped band depicts 2080 30 the range of affected GDP based on a selection of climate 25150 2080 and socioeconomic scenarios for 2080. For example, if 20 SIUzbekistan had experienced a 100-year return period 15 flood event in 2015, the affected GDP would have been an estimated $4 billion. In 2080, however, the affected GDP 10 50 from the same type of event would range from about $20 2015 5 2015 billion to about $30 billion. If Uzbekistan had experienced a 2 50-year earthquake event in 2015, the affected GDP would 10 50 100 250 10 50 100 250 have been about $20 billion. In 2080, the affected GDP from Return period (years) Return period (years) the same type of event would range from about $100 billion .... to about $200 billion, due to population growth, urbaniza- 10 2 1~ 0.4 10 2 1 0.4 Probability (%) Probability (%) tion, and the increase in exposed assets. All historical data on floods and earthquakes are from D. Guha-Sapir, R. Below, and Ph. Hoyois, EM-DAT: International Disaster Database (Universiti Catholique de Louvain, Brussels, Belgium), www.emdat.be; the National Geophysical Data Center/World Data Service (NGDC/WDS), Significant Earthquake Database (National Geophysical Data Center, NOAA), doi:10.7289/V5TD9V7K; and J. Daniell and A. Schaefer, "Eastern Europe and Central Asia Region Earthquake Risk Assessment Country and Province Profiling," final report to GFDRR, 2014. Damage estimates for all historical events have been inflated to 2015 US$. Technmcal Annex T he country risk profiles presented in this document The population and GDP exposed to earthquake and flood In addition to GDP and population, the earthquake exposure were derived from complex flood and earthquake mod- risk were modeled at 30" x 30" resolution and the exposure includes gross capital stock. els developed for assessments on a global scale. This aggregated and presented at Level 1 and national levels. The annex provides an introduction to the technical details asso- administrative boundaries used to create the gridded data and Hazard component d!ate d with the exposure d ata u sed fo r the mod el ing an d to for aggregating results were based on shape files obtained The hazard component of the earthquake model consists of a the models themselves. Readers interested in more complete from the World Bank. While the Level 2 units in the shape files documentation should refer to articles referenced in the text. had very good correlation with ground features, such as rivers events of at least magnitude 5 and a suite of ground motion and border stations, these were adjusted, particularly within prediction equations (GMPEs). The hazard component for the Balkan region, to match census names and statistical data the ECA region contains 1,437 source zones and 744 faults. A with units and borders. Specifically, the GADM database for variety of publications was used to define source zones and Global Administrative Areas was used to define Level 3 units faults, and inconsistencies among the different source zones E ach model uses two major classes of exposure data: for Bosnia and Herzegovina. The adjustment resulted in the were wereiveconciledgaodaataagapsefilleddbaswnonaexeertoopiniin Lipopulation and gross domestic product (GDP). Both creation of 863 administrative units throughout the Europe For a given region, the maximum magnitude for earthquakes were derived from global data sets and downscaled to a grid and Central Asia (ECA) region. Subsequent changes aggregat- generated within a zone or by a fault was based on expert as- approximately 1 km by 1 km (30 x 30 arc seconds) in size, as ed the Bosnia and Herzegovina results data into three Level 1 sessment of the historical record. For more details, see Daniell described below. regions and updated the provincial boundaries forTurkey to (2014). include Diizce. The high-resolution population data were estimated by The earthquake zones were used to account for seismicity downscaling 2010 population data (Van Vuuren et a]. 2007) The GDP and population exposure data have been updated to and mapping them at 0.50 x 0.5' resolution, using population 2015 estimates using province- and country-specific correc- frequency and magnitude of the earthquakes within each scenarios for 26 world regions from the Integrated Model to tion factors. In most cases the changes were less than ±25%. zone of the catalog were specified using historical data and a Assess the Global Environment, or IMAGE (Bouwman et al. However, corrections for a few provinces and countries were Gutenberg-Richter (G-R) relationship between magnitude and 2006), and population maps for the year 2000 from LandScan more significant. For example, the country level correction number of occurrences. Specific characteristics (for example, 2010 (Bright et al. 2011). The 0.50 x 0.50 gridded population factor for Turkmenistan and Uzbekistan were 2.26 and 1.68, location or epicenter, fault motion, hypocentral depth, and data were further downscaled to 30" x 30" using LandScan respectively. fault length) of each earthquake were defined based on known population maps and corrected for differences between cen- faults and fault models, previously derived source regions, and sus Level 2 totals and LandScan totals over the same census geophysical knowledge. spatial units. For more information, see Ward et al. (2013). For earthquakes on known faults, fault source characteristics Most existing studies use a national avenage GDP per capita ' he stochastic earthquake model follows a standard risk (for example, the strike and dip of the fault and the type of to distribute GDP throughout a country. This study, however, T m odeling approach that uses exposure (see above), a haz- fault-strike-slip, thrust, or normal) were based on a random followed the approach of Daniel] (2014) and distributed GDP ard component representing earthquake events, and vulnera- sampling of seismicity from the period 1980-2010. Earth- at 30" x 30" resolution, using estimates of Level 2 GDP within bility functions to estimate the affected GDP, affected popula- quakes with magnitudes less than 5.8 and earthquakes within each country and the gridded population data. Nonetheless, tion, capital losses, and fatalities caused by an earthquake. The a zone representing background seismicity were modeled as a high correlation still exists between aggregated GDP and impacts of all the stochastic events are used to estimate risk point sources. Earthquakes with magnitudes equal to or great- population at Level 1 within a country typically r a0.95. er than .8 were modeled as finite faults. For a given ea rth- quake magnitude, the rupture length can vary significantly as in terms of GDP and population experiencing ground motion The first step in the flood hazard modeling is the simulation a function of fault characteristics. Rupture length along a fault intensities equal to or greater than MMI 6. of daily discharges at a horizontal resolution of 00 x 0.5 for a larger-magnitude earthquake was determined based using the global hydrological model PCR-GLOBWB (Van Reek on published relationships (see Daniell 2014). Hypocentral Risk calculations and Rierkens 2009; Van Reek et al. 2011). For the pres- depths were derived by randomly sampling the historical Once the impact for each event is determined, earthquake risk ent-day climate, this hydrological model was forced with daily record of earthquakes within each zone. Finally, the source meteorological data for the years 1960-99, provided by the model also accounts for the observation that hypocentrall Impcalcuts A ntsonnual averagesmo all thetumpacsiid EU-WATCH project (Weedon et al. 20 10). The second step is to depth tends to increase with magnitude. edmbycthe Anlenthaof ti (1000 ys) represteniptdiyith simulate daily time series of within-the-channel and overbank m lflood volumes at the 030 x 0. spatial resolution, using the Ground motion produced by an earthquake event is modeled catalog. DynRout extension (PCR-GLOBWB-DynRout), which simulates in terms of peak ground acceleration (PGA) using GMPEs Return period impacts are determined by rank ordering. The flood wave propagation within the channel and overbank. For specific to active and stable tectonic regions and estimates of top-ranked impact is the 10,000-year return period impact, as more information, see Winsemius et al. (2013) and Ward et al. local soil conditions. To estimate PGA at a location, a subset of GMPEs was selected from more than 300 (Douglas 2011) and impactuis the 5,0 00 er perio T, sto eve its applicabilityu combined using a logic tree approach. Two logic trees were oftat imt or0grear occur oerith 10,000 as sa developed, one for tectonically active regions, the other for The last step in the flood hazard modeling is the inundation inactive regions. These tectonic regimes were also used to es- lry, the tenth-ranked impact is the 1,000-year return pero modeling. Annual time series of maximum flood volumes were inmpact,easetenne.entssofttcaonicpactioesgreaterloocureover the timate local soil conditions frnm topographic slope, following extracted from the daily flood volume time series. Estimates Allen and Wald (2007). PGA-MMI relationships were used to of flood volumes pergrid cell (0.5' x 0.50) were derived for convert PGA to Modified Mercalli Intensity (MMI). The GDP and population model results have been updated to selected return periods (2, 5, 10, 25, 50, 100, 250, 500, and 2 015 estimrnates u si1ng provinc e- an d c ountry-s pe cific c orrec- 1,000 years), using extreme value statistics based on the Gum- Vulnerability component don factors. Inmost cases the changes were less than ±25%. bel distribution and the yearly non-zero flood volume time The vulnerability component for the earthquake model is However, corrections for a few provinces and countries were ries. These flood volume estimates were used as input for the based on surface intensity specified as MMI, not on earth- more significant. For example, the country level correction GLOFRIS downscaling module to calculate inundation depths quake magnitude in terms of PGA. MMI 6 is typically defined factor for Turkmenistan and Uizbekistan were 2.26 and 1.68, at the 30" x 30" level for the chosen return periods (Winsem- as the intensity at which buildings start to experience damage. respectively. ius et al. 2013). Under the assumption that flood volumes with Vulnerability was quantified using relationships that estimate two-year return periods would not lead to overbank flooding, capital loss and fatalities as a function of MMI in regions experiencing MMI 6 and greater. The vulnerability functions subtracting the two-year return period flood volume before used in this study relating intensity to capital loss account for changes through time in building typologies, seismic codes, he method for calculating flood risk in this study was this approach, such flood protection measures as the use of and the Human Development Index (HDI), as well as climate T different from that chosen for the earthquake risk. Where dikes or retention areas were not taken into account, which and the age of buildings. The HDI was used as a proxy for the earthquake risk was determined from 10,000 years of simulat- may have led to an overestimation of the actual flood extent. development of a nation and its expression thrnugh building ed events, flood risk was determined using 40 years of climate The quality of the modeled inundation depths depend on quality, bulding practices, and materials use. The vulnerabili- simulations and extreme value analysis. the quality of the elevation data used to project flooding on ty function for fatalities varied by region, time of day, and HDI. the terrain. This study used the Shuttle Radar Topography For more information on the development of these functions, Hazard component Mission (SRTM) elevation data, processed into a number of see Daniell (2014). Several modules of the GLOFRIS' giobal flood risk modeling The model also estimates the GDP and population affected cascade were used to derive the flood risk results in this study. by an earthquake. For these estimates, a step function that switches from zero to one at MMI 6 was used as a vulnera- bility function. In other words, the model calculated impact 1impoalFoo Riskh with00-yE enaprio 2.mpactterGlsbltwotreventse elevation-derived products available on the HydroSHEDS pages. CLIMATE MODEL DESCRIPTION Vulnerability and risk calculation GFDL ESM2M GFDL Earth System Model 2 with medium resolution In this study, affected GDP and affected population were considered as metrics for flood risk. Given the uncertain- ties in estimated flood depths, the annual average and HadGEM2-ES Hadley Global Environment Mode( 2-Earth System return period impacts were estimated in terms of GDP and population experiencing floodwater at any depth. The GDP and population affected by flood for each return MIRMIROC (Model for Interdisciplinary Research on Climate) Earth System CHASER-coupled Model period were represented by the counts of population or (Atmospheric Chemistry version) GDP in each grid cell that had non-zero flood depths at the selected return periods. The average annual values at each grid cell were derived by integrating over the nine return period impact estimates. The annual average and return period values for GDP and population affected by flood in the Level 1 administrative regions were determined by NorESM1-M Norwegian Earth System Model with medium resolution summing the impacts within each area. Future Earthquake and Flood Risk Table 2. Climate Models Used for Flood Risk Estimates. Estimates of the GDP and population affected by earth- quakes in 2080 were based on the 2080 socioeconomic and climate conditions associated with the five Shared Socioeconomic Pathways (SSPs) created by the Intergov- ernmental Panel on Climate Change (IPCC) for the Fifth RCP SCENARIO SSP SCENARIO SCENARIO CHARACTERIZATION Assessment Report (ARS). Estimates of future exposure data (GDP and population in 2080) were developed using the IMAGE model of the PBL Netherlands Environmental RCP4.5 SSP2 Cautiously optimistic Assessment Agency, forced by the five SSPs. The popula- tion numbers estimated with the IMAGE model were then further modified to be consistent on a Level 1 adminis- RCP8.5 SSP2 Present trends continue trative (province) level using the 2010 round of census data. The earthquake model was then used to calculate the affected GDP and affected population for the five projec- RCP8.5 SSP3 Worst case tions of the 2080 exposure. Estimates of the GDP and population affected by flood- ing in 2080 were based on both the socioeconomic and cimateconditios in 2080 . tthe 0climecondmitns Table 3. RCP and SSPScenario Combinations Used to Estimate Future Flood Risk. climate conditions in 2080. The 2080 climate conditions 3. http://hydrosheds,cr.usgs.gov. used in this publication were based on the Representative Concentration Pathways (RCPs) created by the Intergov- ernmental Panel on Climate Change (IPCC) for the Fifth Assessment Report (ARS). The RCPs were used to force the five climate models, listed in table 2, to simulate daily future climate data. The future flood hazard estimates were subsequently calculated using the same GLOFRIS model cascade described above but forced with the daily future climate data from the five climate models. The precipitation estimates for the climate models were bias corrected using the 1960-99 EU-WATCH data and a meth- odology developed by the Inter-Sectoral Impact Model Intercomparison Project (ISI-MIP). For details on the bias correction, see Hempel et al. (2013). Future flood risk for 2080 was then estimated for scenar- ios that combined the projections of the flood hazard and exposure data based on the RCPs and SSPs in table 3. For more information on the RCPs, see Meinshausen et al. (2011). For information on the SSP scenarios, see the special issue of the journal, Climate Change on Shared Socioeconomic Pathways (Nakicenovic et al. 2013). Allen, T I., and D. J. Wald. 2007. "Topographic Slope as a Proxy for Global Seismic Site Conditions Ward, P. J., B. Jongman, F. S. Weiland, A. Bouwman, R. van Beek, M. F. P. Bierkens, W. Ligtvoet, (VS30) and Amplification around the Globe." U.S. Geological Survey Open-File Report 2007- and H. C. Winsemius. 2013. "Assessing Flood Risk at the Global Scale: Model Setup, Results, and 1357. http://pubs.usgs.gov/of/2007/1357/index.html. Sensitivity." Environmental Research Letters 8, 044019, DOI:10.1088/1748-9326/8/4/044019. Bouwman, A. F., T. Kram, and K. Klein Goldewijk. 2006. Integrated Modelling of Global Environ- Weedon, G. P., S. Gomes, P. Viterbo, H. Oesterle, J. C. Adam, N. Bellouin, 0. Boucher, and M. mental Change: An Overview ofIMAGE 2.4. Bilthoven: Netherlands Environmental Assessment Best. 2010. "The WATCH Forcing Data 1958-2001: A Meteorological Forcing Dataset for Land Agency. Surface- and Hydrological-Models." WATCH Technical Report 22. Wallingford: Met Office Hadley Centre. Available at www.eu-watch.org. Bright, E. A., P. R. Coleman, A. N. Rose, and M. L. Urban. 2011. LandScan 2010 High Resolution Global Population Data Set. Oak Ridge, TN: Oak Ridge National Laboratory. Winsemius, H. C., L. P. H. Van Beek, B. Jongman, P. J. Ward, and A. Bouwman. 2013. "A Framework for Global River Flood Risk Assessments." Hydrology and Earth System Sciences 17: 1871-92. Daniell, J. E. 2014. "Development of Socio-Economic Fragility Functions for Use in Worldwide DOI:10.5194/hess-17-1871-2013. Rapid Earthquake Loss Estimation Procedures." Doctoral thesis, Karlsruhe Institute of Technol- ogy, Karlsruhe, Germany. Douglas, 1. 2011. "Ground-Motion Prediction Equations 1964-2010." BRGM/RP-59356-FR. Pacific Earthquake Engineering Research Center, University of California, Berkeley. Hempel, S., K. Frieler L. Warszawski, J. Schewe, and F. Piontek. 2013. "A Trend-Preserving Bias Correction: The ISI-MIP Approach." Earth System Dynamics 4: 219-36. DOI:10.5194/esd-4-219- 2013. Meinshausen, M., S. J. Smith, K. V. Calvin, J. S. Daniel, M. L. T. Kainuma, J.-F. Lamarque, K. Matsu- moto, S. A. Montzka, S. C. B. Raper, K. Riahi, A. M. Thomson, G. J. M. Velders, and D. van Vuuren. 2011. "The RCP Greenhouse Gas Concentrations and Their Extension from 1765 to 2300." Climatic Change 109 (1-2, special issue): 213-41. DOI: 10.1007/s10584-011-0156-z. Nakicenovic, Nebojsa, Robert Lempert, and Anthony Janetos, eds. 2013. "Special Issue on the Shared Socioeconomic Pathways (SSPs): A Special Issue of Climatic Change journal on the Framework for the Development of New Socioeconomic Scenarios for Climate Change Re- search." https://www2.cgd.ucar.edu/research/iconics/publications/ssps. Van Beek, L. P. H., and M. F. P. Bierkens. 2009. "The Global Hydrological Model PCR-GLOBWB: Conceptualization, Parameterization and Verification." Utrecht University, Utrecht. http://van- beek.geo.uu.nl/suppinfo/vanbeekbierkens2009.pdf. Van Beek, L. P. H., Y Wada, and M. F. P. Bierkens. 2011. "Global Monthly Water Stress: Water Balance and Water Availability." Water Resources Research 47, W07517, DOI:10.1029/2010WR009791. Van Vuuren, D. P., P. L. Lucas, and H. Hilderink. 2007. "Downscaling Drivers of Global Environ- mental Change: Enabling Use of Global SRES Scenarios at the National and Grid Levels." Global Environmental Change 17, 114130 DOI:10.1016/j.gloenvcha.2006.04.004. hec Europe and Central Asia (ECA) region experiences a variety of natural hazards, including floods, carthiquakes, droughts, landslides, and wildfir-es.'1The frequency and imipact of those events can be great. For examiple, close to 500 significant floods and carthiquakes have occurred in the region over the past thirce decades, affecting nearly 25 million people and causing 50,000 fatalitics and approximately US$80 billion in damiage. Disaster risk mianagemient could reduce the imnpact of su ch eventts app reciably. Maximizing developmient and mimnmizing the imi- pact of disasters requires understanding, managing, and miitigating disaster risk. Investments support- ing disaster risk management provide benefits beyond the miitigation of risk and the reduction of loss when disaster stri kes. Thecy als o stim ulate cco- n oiic activity throughi the creatio n of an improve d investing environment, provide additional bene-. ........... fits throughi social, environmiental, and economic synergics, and enhance social progress, even in the- absence of disaster' Intended as a catalyst for discussions of disaster risk managemient for countrics in the ECA region, thispublcatin povides high-level assessmients of thi pulcto pi 12-7 risk to gross domiiestic pro duct (G DP) an d p opula - tion fromi floods and carthiquakes. From it, national- decision makers can obtain an overview of the risk in each country, how the risk varics amiong a coun- try's provinces, and how countrics rank in regard to risk in theo ECA region. WOLDBAKGRUPGF R