Meta and Longitudinal Analyses of High Death Rates of Some Particular Municipalities in GEJET ICHARM International Centre for Water Hazard Ris Management under the auspices of UNESCO CTI Engineering Co., Ltd. Japan 2011.７.
IRDR (Integrated Research on Disaster Risk) Working Groups Forensic Investigations Meta-Analysis Longitudinal Analysis Critical Cause Analysis Scenario Analysis Risk Interpretation and Action Disaster Loss Data
Purpose Utilizing Meta and longitudinal Analyses, we try to examine: Why some municipalities had high human losses and the others less? Where those differences come from? Identify the Uniqueness of High Human Losses of Particular Municipalities
Over view of GEJET in 2011 M 9.0 Death Toll 15,690 Missing 4,735 Completely Destroyed House 112,901 A Partly Destroyed House 144,598 (2011.10.12) Tsunami Height North of Iwate Prefecture to Ozika Peninsula (220km) : Ria Coastline 20-30 m Max.40m South of Ozika Peninsula to Iwaki (150 km) : Straight Coastline 10-20 m Ibaraki Pref. and Chiba Pref. 5-10 m Aomori 5m Non Tsunami Affected Area 94 (0.5%) Tsunami Affected Area 20,329 (99.5%) Epicentral Area (Wave Source Area) : Extending 200 km east to west by 500 km north to south < From off the coast of Hachinohe to off the coast of Kashima > Tsunami Arrival : Approx. 30min. to 2.5hr (From north <South Sanriku> to south <South Boso Peninsula>)
Over view of GEJET in 2011 Death Number Human Loss Ratio (IA) <(Death Number / Population) in each Municipalities Inundated Areas> Dead or Missing 1~9 10~99 100~499 500~999 1000~ Death Ratio 0~1 % 1~3 % 3~5 % 5~10 % 10 % < Source : FDMA(9Sep.2011)
Human Loss Ratio : Sanriku Ria Coastal Area Dead- HLRatio Missing (IA) HLratio IA (km 2 ) Sendai Miyagi Fukushima Morioka Iwate Kuji 4 0.06% 0.01% 4 Noda 38 1.2 0.8 2 Fudai 1 0.9 1 Tanohata 33 2.1 0.9 1 Iwaizumi 7 0.6 0.1 1 Miyako 544 3.0 1.0 10 Yamada 853 7.5 4.6 5 Otsuchi 1449 12.2 10.6 4 Kamaishi 1180 9.0 3.1 7 Ofunato 449 2.4 1.1 8 Rikuzentakata 2115 12.7 8.6 13 Kesennuma 1414 3.5 1.9 18 Minamisanriku 987 6.9 5.7 10 Onagawa 949 11.8 9.2 3 Total/Average 10023 6.0% 2.8% 88
Methodology (How to investigate) Meta-Analysis : Literature survey and statistical analysis Longitudinal Analysis : Comparative studies of a set of target areas by temporal and historical perspective. Data Source for this Study <Open Source Data> 1. Central Gov. / 2. Regional Gov. / 3. Organizations / 4. Libraries internet source, major newspapers, statistical data, raw data historical documents, historical statistics, local newspapers, vernacular magazines <Closed Source Data> Field survey data mainly can be utilized for critical cause analysis Major Source for this Study (Mainly in Japanese) General info : Cabinet Office, Ministry of Internal Affairs and Communications, Fire Defense Agency Victims related info : National Police Agency Geographical info : Geospatial Information Authority of Japan Local info : Iwate Prefectural Gov., Iwate fukko net, Iwate Statistics Data Base Historical info : Japan Destructive Earthquake General Survey(Usami, 2003) Japan Maps Encyclopedia(Yamaguchi et al., 1980)
Methodology (What do we investigate?) Investigate Human Loss Ratio (IA) in Some Municipalities We need following related info. to investigate Past Experiences Compared to the Past Tsunami Disasters Age Structures in IA Age Structures of the Death People Social Change Evacuation Conditions Distance from Higher Ground Places Evacuation Methods, Warnings, Education/Training, Health However, it is not possible to get those direct data, therefore following approaches were taken Approach 1Past Experiences : try to estimate HVR (will be explained) and compared to the past tsunami (Meiji<1896>,Showa< 1933>) disasters. 2Age Structures : try to find elderly population rates, social participation rates including recent trends by prefectural statistical data. 3Evacuation Conditions: Growing population, infrastructures, and evacuation center s safety in IAs.
Approaches 1Past Experiences Following indicator was established to compare to the past tsunami disasters. (Meiji<1896>,Showa< 1933>) Human Vulnerability Ratio = Human Loss (HL) Ratio Completed House Destroyed(CHD) Ratio = HL numbers / Total Population CHD numbers / Total House numbers = HL numbers Total Population CHD numbers / Total House numbers Therefore, we could estimate HVR = HL numbers Affected Population
Approaches 1Past Experiences HVR (%) Iwate Miyagi Why do they reverse the historical trends? Did they learn from the experience?
Yamada town and Rikuzentakata city Rikuzentakata city Yamada town Approach 2 Age Structures Approach 3 Evacuation Conditions
Approaches 2 Age Structures Age Structures of Reported Death Over 60s in Reported Total Dead 3 Pref. Average 65% Rikuzentakata 61% Yamada 75%
Approaches 2 Age Structures Miyako,Ofunatos, and Kuji are included estimations by the impact of annexation. <miyako (2005.6.6) ofunato (2001.11.15)>
Approaches 2 Age Structures Percentage of HH with Elderly Persons over 65 Yamada Town Rikuzentakata City Coastal Area Ave. Iwate Pref. Present <Year> 57.4 <2006> 62.8 <2006> 54.3 <2006> 46.1 <2006> Change Ratio +26% (2004-2006) +16% (2004-2006) +25% (2004-2006) +19% (2004-2006)
Approaches2 Age Structures Aging society with less social participation
Yamada town and Rikuzentakata city Rikuzentakata city Yamada cho Approach 2 Age Structures Approach 3 Evacuation Conditions
Approaches Areas 3Evacuation Conditions Total Population Dead Ratio (%) Missing Ratio (%) Yahagi 1793 18 1 4 0 Yokota 1405 14 1 2 0 Takekoma 1291 38 3 4 0 Kesen 3480 194 6 65 2 Takata 7601 879 12 298 4 Yonezaki 2902 82 3 32 1 Otomo 2025 45 2 17 1 Hirota 3749 42 1 16 0 Total 24246 1312 5 438 2 Increasing Population in High Risk Area Rikuzentakata city 2011.7.11
Approaches 3Evacuation Conditions Takata area (high risk area)s population Year 1896 1940 1945 1950 1954 1995 2000 2005 Pop. 3,489 4,960 6,060 6,461 6,488 7,605 7,663 7,711 Meiji(1896) Showa(1933)
Approaches③ Evacuation Conditions 1913年 1933年 1952年 2000年 1968年 1960-1966 Tsunami Project
Approaches 3Evacuation Conditions Before After
Approaches 3Evacuation Conditions Affected Primary Evacuation Center Primary Evacuation Centers Iwate Prefecture Rikuzentakata city Total Numbers 411 68 Affected 48 35 Affected Ratio 11.7 % 51.5 % TOPICS : In the Takata area (high risk area), Rikuzentakata city, 70-80 evacuated and only 3 survived at Civic Gymnasium. (Source : Kahoku Shinbun)
Conclusion (Try to answer the first 3 inquiries) Why some municipalities had high human losses and the others less? Where those differences come from? Identify the Uniqueness of High Human Losses of Particular Municipalities Methods : Meta and Longitudinal Analysis - Development of the HVR. - Surveys of literatures, statistical data sets, and maps. Findings: Yamada town : Aging society with low social participation especially since 1980s Rikuzentakata city : Increasing pop. in high risk areas, especially after the last tsunami Combination with critical cause analysis should be implemented to investigate more detail
Suggestion for Decision Making from this study 1.Counter-measures for Aging Society should be considered the impact of elderly population. elderly people is difficult to evacuate. so many people tried to help elderly people and affected. 2. Land Use Regulation should be recognized past tsunami experience. land use is one of the most effective way for disaster management Not based on last tsunami but consider every historical tsunami 3. Education and Training should be noticed the importance of education and training School played a key role in GEJET Difficult to pass down the experience over generations. Tend to rely on the infrastructure as time passes. 4. Evacuation Planning should be remarked the location of evacuation centers and related planning These should be considered based on Worst Case Scenario
Ｍａｎｙ Ｔｈａｎｋｓ ICHARM International Centre for Water Hazard Ris Management under the auspices of UNESCO CTI Engineering Co., Ltd. Japan 2011.７.
APPENDIX
Cost Effective ness of Disaster counter measure Overview and explanation Now Considering (Gilbert F.White, J.Eugene Haas et al., 1975) Net Benefit + : Land Use, Flood-Proofing, Protection Net Benefit 0 : Warning Net Benefit - : Relief, Insurance
Impact of Landform Condition : Sanriku Ria Coastal Area and Straight Coastal Area Coastal Landform Condition Impact to Exacerbation of Disasters Kuji in Iwate Pref. to Onagawa in Miyagi Pref.:Ria Coastline Ishinomaki in Miyagi Pref. to Minamisoma in Fukushima Pref.:Straight Coastline Ria Coastline Straight Coastline Kuji city Dead and Missing 10,023 9,587 Morioka Iwate Miyagi Sendai Ria Coastline Onagawa-cho Fukushima Ishinomaki Straight Coastline Minamisoma city Completely-Destroyed House(CDH) Inundated Area(IA) (km 2 ) Population in IA Household Numbers in IA Human Loss (HL) Ratio CDH Ratio Inundated Ratio (IR) in Urban Area Tsunami Height 35,526 88 167,568 60,245 6.0% 59% 24% 20 ~ 35m Human Loss Ratio 6 % Ria Coastline the open sea is deep Tsunami Height is so High Tsunami Arrival is fast 61,370 375 297,613 103,450 3.2% 59% 26% 10 ~ 20m
Tsunami History Meiji(1896) 37 Showa(1933) 27 Chili(1960) 51 GEJET(2011) 78 Meiji(1896): PM7:30 First Wave Arrival Approx.35 min After EQ Showa(1933): AM 2:30 First Wave Arrival Approx.30 min After EQ GEJ(2011):PM2:46 First Wave Arrival Approx.30 min After EQ
Yamada town and Rikuzentakata city Rikuzentakata city Yamada cho 2 Age Structure 3 Evacuation Conditions
Yamada town and Rikuzentakata city Rikuzentakata city Yamada cho 2 Age Structure 3 Evacuation Conditions
Yamada town and Rikuzentakata city Rikuzentakata city Yamada cho 2 Age Structure 3 Evacuation Conditions
Yamada town and Rikuzentakata city Yamada town 1 Victims Analysis 2 Trends Analysis and Evacuation Social Background Analysis
Yamada town and Rikuzentakata city Rikuzentakata city 1 Victims Analysis 2 Trends and Evacuation Social Background Analysis
Trend and Evacuation Social Background Analysis Population Trend in Rikuzentakata city
1 Victims Analysis Population s Age Structure of Reported Death Toll caused by GEJET (2011) in Yamada-cho and Rikuzen takata city Yamada Rikutaka Average Over 60s in Total Dead Yamada 75% Rikuzentakata 61% 3 Pref. Average 65%
How to compare the historical trends under the limitation of data set? この事実はこういうところにもあらわれている 特徴 Human Vulnerability Ratio Human Vulnerability Ratio = Human Loss Ratio / House Destroyed Ratio = HLN / CDHN 1/ Average Size of House Hold Human Loss Ratio : Human Loss Number (HLN) / Total Population (TP) House Destroyed Ratio : CDH Numbers (CDHN)/ Total House Numbers (THN) How many people die compared to house collapse in the areas. If many people evacuated successfully, the HVR become lower
1 Victims Analysis Reported Death Toll / Pre Disaster Population School plays a key role Working time 70s, 80s male ratio higher than female (diff. from past) Same trend with 3 Pref. average(teens, male ratio)
Death Toll in Non-Tsunami Affected Area and Tsunami Affected Area Non-Tsunami Affected Area Tsunami Affected Areas Hokkaido 1 Aomori 4 Iwate 4 6,673 Miyagi 17 11,697 Akita Yamagata 3 Fukushima 40 1921 Ibaraki 8 17 Tochigi 4 Gunma 1 Saitama Chiba 5 17 Tokyo 7 Kanagawa 4 Total 94(0.5%) 20,329(99.5%) Deaths caused by tsunami disaster covered almost all of deaths Source : JDMA
Impact of Landform Condition : Sanriku Ria Coastal Area and Straight Coastal Area Coastal Landform Condition Impact to Exacerbation of Disasters Kuji in Iwate Pref. to Onagawa in Miyagi Pref.:Ria Coastline Ishinomaki in Miyagi Pref. to Minamisoma in Fukushima Pref.:Straight Coastline Ria Coastline Straight Coastline Iwate Miyagi Sendai Morioka Kuji city Ria Coastline Onagawa-cho Dead and Missing Completely-Destroyed House(CDH) Inundated Area(IA) (km 2 ) Population in IA Household Numbers in IA Human Loss (HL) Ratio CDH Ratio Inundated Ratio (IR) in Urban Area Tsunami Height 10,023 35,526 88 167,568 60,245 6.0% 59% 24% 20~35m 9,587 61,370 375 297,613 103,450 3.2% 59% 26% 10~20m Fukushima Ishinomaki city Straight Coastline Minamisoma city Human Loss Ratio 6 % Ria Coastline the open sea is deep Tsunami Height is so High Tsunami Arrival is Fast
Historical comparison of tsunami damage in ria coast area Human Loss Ratio House Destroyed Ratio Human Vulnerability Ratio Meiji (1896) Showa (1933) GEJ (2011) Meiji (1896) Showa (1933) GEJ (2011) Meiji (1896) Showa (1933) GEJ (2011) Kuji 4.7 0.7 0 23 13.2 0 20.4 5.3 Noda 10 2.5 0.8 19.5 31 19.5 51.3 8.1 4.1 Fudai 14.8 9.5 0 23 24.8 0 64.3 38.3 Tanohata 7.7 4.7 0.9 11.4 43.1 17.2 67.5 10.9 5.2 Iwaizumi 17.4 9.8 0.07 34.2 33.4 4.1 50.9 29.3 1.7 Miyako 15.5 3.5 1 23.4 11 16.3 66.2 31.8 6.1 Yamada 23.9 0.2 4.6 46.3 40.1 42.2 51.6 0.5 10.9 Otsuchi 9.2 0.5 10.6 57.4 55.5 57.8 16 0.9 18.3 Kamaishi 51.9 5.2 3.1 54 16.2 19.8 96.1 32.1 15.7 Ofunato 21.8 3.8 1.1 40.3 32.9 22 54.1 11.6 5 Rikuzentakata 5.4 1.7 8.6 10.3 28.7 40.5 52.4 5.9 21.2 Kesennuma 15.5 0.8 1.9 29.9 28.1 33.3 51.8 2.8 5.7 Minamisanriku 17.3 0.9 5.7 44.8 12.2 59.8 38.6 7.4 9.5 Onagawa 0.04 0.02 9.2 2.4 5.6 75.1 1.7 0.4 12.3 Human Loss Ratio : Human Loss Number (HLN) / Total Population (TP) House Destroyed Ratio : CDH Numbers (CDHN)/ Total House Numbers (THN) Human Vulnerability Ratio = Human Loss Ratio / House Destroyed Ratio = HLN / CDHN 1/ Average Size of House Hold = HLN / CDH HLN How many people die compared to house collapse in the areas. If many people evacuated successfully, the HVR become lower
Trend and Evacuation SB Analysis Historic Analysis of Death and CHD in Yamada-cho GEJET 51y 27y 37y Chili Showa Meiji Dead Building Collapse
2 Trend Analysis (Yamada ) Recent Population Change and Industry Situation Change in Yamada-cho and Rikuzen takata ( Exerpted from Iwate Prefecture Living Condition Indicators ) Population Economy Total Population Percentage of Elderly Persons Households Birth rate Municipality Grass Product Business Office Number Employee Number Present <Year> 18,745 <2010> 57.4 <2006> 6.34 <2006> 30,455,000 <2008> 932 <2006> 5,941 <2006> Change Ratio -25% (1980 2010) +26% (2004-2006) -12% (2004 2006) -7.9% (2004 2008) -11.1% (2001 2006) -8.3% (2001 2006)
Trend and Evacuation SB Analysis Funa koshi Bay Before Show(1933) Relocation After Showa Relocation A:Meiji(1896) B:Showa(1933) C:Chili (1960) D:GEJET (2011) Learn from the experience but could not protect completely
3 Evacuation Social Background Analysis Situation (what they did and how many minutes to evacuate) Meiji(1896): PM7:30 Showa(1933): AM 2:30 First Wave Arrival Approx.35 min After EQ First Wave Arrival Approx.30 min After EQ GEJ(2011):PM2:46 First Wave Arrival Approx.30 min After EQ They all have approx.30 min to evacuate. Chili(1960) PM3:00 First Wave Arrival and Approx.PM6:00 Maximum Wave Arrival Warning Begin at Approx.PM5:00
3 Evacuation Social Background Analysis (Yamada) Disaster Recovery Plan after the Showa (1933) Photo : Situation on Tanohama in Yamada-cho 1961 s Situation
3 Evacuation Social Background Analysis (Yamada) Summary : Why do they reverse the historical trend? Did they learn from the experience? Depopulation Economic Decreasing Declining Local Society In addition, concerning the evacuation, 1. Elderly social participation is low 2. NPOs numbers ratio is low 3. Elderly and singled population is high (Social systems strength maybe low, Maybe low coping capacity) These Impact evacuation activities. Learning from past experience, two areas (Funakoshi and Tanohama ) relocated to higher grounds. However, these areas are inundated in GEJET, it means the hazard scales are too large to protect completely.
3 Evacuation Social Background Analysis (Rikuzentakata) Experience sometimes goes negative Chilean Tsunami (1960) Tsunami Height 2-4m Small damage GEJ (2011) Combined JMAs initial information, 3m Combined never come over beyond the railroad track 2000 年
3 Evacuation Social Background Analysis (Rikuzentakata) Summary : Why do they reverse the historical trend? Did they learn from the experience? Increasing population in high risk area (past inundated areas, maybe) Especially, this is after levees construction (from Interview, maps) Chilean tsunami experience works negatively Never though tsunami come over beyond the railroad track until they actually saw the tsunami in their own eyes. Evacuated and survived people were living near the higher grounds because they could evacuate after they actually saw the tsunami coming <interview> Initial 3m s Tsunami Height information from JMA combined with their experience of Chilean Tsunami (2-4m) +Altruism Activity (Helping Others) Against the experience of Tendenko Volunteers Fire Department 21%(27/127 ) Died <Interview> +Primary Evacuation Center Planning 51%(35/68) Inundated <Iwate: 12 %(48/411)> <Kahokushinbun>
3 Evacuation Social Background Analysis (Rikuzentakata) Evacuation Type Evacuated Not Evacuated Evacuated and Suffered Could not Evacuated Primary Evacuation Center 35/68 Inundated <Iwate: 48/411> Civic Gymnasium <Takata area> Only 3 / 70-80 Survived (Source : Kahoku Shinbun) Altruism Activity (Helping Others) Volunteers Fire Department data Died (Interview 20Jul.2011)
3 Evacuation Social Background Analysis (Rikuzentakata) 123 Difference リアス式海岸地域は地形の特徴から被害に遭いやすい 過去の災害との違いに大きな社会の変化があった山田町 極端な人口減少 経済衰退 高齢化の進行様々な社会的指標により地域社会の力が弱かったことが人的被害の誘因の一部となったと考えられる 陸前高田市 市全体の人口は 漸減傾向にあったが 沿岸部のリスクの高い市街地に人口が増加 脆弱性を高めていた 過去の経験から学んだか学んだところと学んでいない部分がはっきりとした 山田町では 2 カ所が昭和津波のあと 高台移転したが 今回の津波では浸水した 陸前高田では チリでの経験がかえってマイナスとなった 堤防ができると逆に人が移りすんだ チリの時の津波高 一定地区のみの被害が 今回の災害情報と結びつき避難行動がなかなか進まない原因となった Experience sometimes goes wrong, however we need to learn from the experience of experience goes wrong Tend to be influenced last disaster Social change should be considered to mitigate disasters
Levee Height and Simulated Tsunami Height in GEJET in 2011 Death Toll Under consideration However, water gate was shut down by local officer because of blackout by the earthquakes
Conclusion (Try to answer the first 3 inquiries, diff. and exp.) Compared to Meiji(1896), Showa(1933), so many municipalities could successfully mitigate the disasters. However, some are not. We examined two cases and indicated how and what we can investigate the disasters through meta- and longitudinal analyses We ve found the followings : 1. Social changes impact the disasters. diff. Yamada (Depopulation, high aging society, and social participation) Rikuzentakata (Increasing Population in High Risk Area) 2. Disaster reflects the reality of the local Societies. exp. Disasters are so much dependent on social characteristics and trends. Experiences are difficult to be cleared by meta- longitudinal- analyses. We still have lots of challenges : How can we investigate LPHC disasters by meta- and longitudinal- analyses? 1. From field interviews, so many things were realized. Tend to be influenced by last disaster. Difficult to pass down the experience over generations. Tend to rely on the infrastructure as time passes. Experience is easy to be disappeared, however infrastructure remains. An altruism action against the experience of tendenko is significant to be examined. Planning are also important points to be investigated. 2. A holistic approach (combination with various perspectives) is required. 3. Intensive and extensive investigations should be done to find more detail about the GEJET disaster.