The Collection, Synthesis, and Quality Assessment of Data Describing the Response to Human Needs Following the 1999 Turkey Earthquakes

John R. Harrald*, Irmak Renda Tanali** and Jeanne B. Perkins***

*Professor, The George Washington University, Department of Engineering Management and Systems Engineering and Director of GWU Institute for Crisis, Disaster, and Risk Management, 707 22nd St. Staughton Hall, Room #208, NW Washington, DC 20052; PH 202-994-7153; jharrald@gwu.edu

**Research Associate, The George Washington University, Department of Engineering Management and Systems Engineering and GWU Institute for Crisis, Disaster, and Risk Management, 707 22nd St. Staughton Hall, Room #205, NW Washington, DC 20052; PH 202-994-7528; rendatan@gwu.edu

***Earthquake Program Manager, Association of Bay Area Governments, 101 8th St. Oakland, CA 94607; PH 510-464-7934; Jeannep@abag.ca.gov

Abstract

Services that met the human needs of persons affected by the magnitude 7.4 Kocaeli earthquake of August 17, 1999 and the magnitude 7.1 Düzce earthquake of November 12, 1999 were provided by agencies of the Turkish central and provincial governments, municipal governments, the Turkish military, Turkish non-governmental organizations, international governmental and non-governmental organizations.  The purpose of this project was to collect and synthesize the data describing the service delivery activities of these organizations before these data were lost or degraded.  The analysis of these data will enhance the ability to anticipate the scale of human needs (medical, sheltering, feeding, supplies) following future earthquakes in Turkey, the U.S. and elsewhere.  The data describing the impact of the 1999 Turkish earthquakes compiled by this effort are essential to the development of models that may be used to support future disaster planning, training, and response in Turkey since response and relief capabilities must be based on estimates of service delivery demands.  The availability of the data will also encourage a range of collaborative research activities related to earthquake loss estimation and will be available to individual and institutional researchers working on projects such as the World Bank-funded Marmara Earthquake Emergency Reconstruction (MEER) project.   This report describes the sources of data collected, the process used to collect, synthesize, and assess data and a preliminary description and interpretation of representative data products.  The complete products of this research have been posted on a web site maintained by The George Washington University Institute for Crisis, Disaster, and Risk Management and includes a set of annotated data tables constructed from multiple sources, a listing of web-based data sources, and a description of print references that may be used by future researchers. 

Introduction

Two massive earthquakes struck Turkey in 1999 - the magnitude 7.4 Kocaeli earthquake on August 17th and the magnitude 7.1 Düzce earthquake on November 12th.   Both earthquakes were caused from ruptures of the North Anatolian fault, with the Kocaeli earthquake lasting 45 seconds and rupturing 126 km of that fault, while the smaller Düzce earthquake produced a surface fault rupture of only 39 km

The Kocaeli earthquake resulted in 17,480 deaths, 43,953 injuries, and 66,441 collapsed or were heavily damaged housing units.  Estimates of losses range from $7 billion to $40 billion.  An additional 763 deaths, 4,948 injuries, and 26,704 collapsed or heavily damaged housing units occurred as a result of the Düzce earthquake (Ozmen, 2000).

The August 17, 1999 Kocaeli earthquake tragically illustrated the inability to cope with the result of poorly controlled development and rapid population growth in disaster-prone regions. In Turkey, the price of increased vulnerability and inadequate preparedness was the loss of thousands of lives, the displacement of hundreds of thousands of people from their homes, and the economic impact of over $20 billion.  Government and non-government organizations faced with meeting the human needs following the Kocaeli earthquake were overwhelmed by the demand for their services.  The purpose of this project was to identify, collect, synthesize and quality assure data describing the response to human needs following the 1999 Turkey earthquakes before these data were lost or the ability to interpret data degraded.  The resulting databases have been made available on the Internet for researchers and emergency planners and can provide the basis for developing models capable of predicting the service delivery capability required to meet human needs following future earthquakes in Turkey.    

 The premise of this research is that analysis of these data will enhance the ability to anticipate the scale of human needs (medical, sheltering, feeding, supplies) following future earthquakes in Turkey, the U.S. and elsewhere, and will support the development of adequate plans, procedures, and service-delivery capabilities.  These service-delivery needs are strongly determined by demographic and socio-economic factors in addition to the sustained physical damage.  The estimation of service-delivery demands requires a linked set of modeling activities, data to populate the models, and expert judgment to interpret the quality, meaning, and limitations of available data.  This scenario-based needs estimation is an essential precursor to the development of adequate response and recovery strategies, plans, and organizational structures. 

Obtaining comprehensive and consistent data describing the services delivered following the August and November earthquakes was complicated by the fact that emergency and relief services were delivered by a large number of Turkish and international organizations. The Deputy Governor of Yalova stated to the research team during a December 1999 visit to his city that, at the peak of activity, over 200 government and non-government relief organizations were actively delivering services in Yalova.  The Mayor of Izmit and the Governor of Kocaeli, during interviews in June 2000, described similar levels of organizational activity in their jurisdictions.

The George Washington University Institute for Crisis, Disaster and Risk Management (GW ICDRM) collaborated with the Association of Bay Area Governments (ABAG) in this effort to identify, collect, synthesize, and assess the quality of perishable data from the Turkey earthquakes, in cooperation with Turkish scientists, governmental officials, and emergency responders. The project team took advantage of existing collaborative partnerships in the U.S. and between U.S. and Turkish academic, non-governmental, and governmental organizations. 

 Project Goals, Research Questions, and Objectives

The goal of the project was to enhance the capability of government and non-government organizations for emergency management, humanitarian relief, and recovery planning for potential major earthquakes in Turkey. The research questions to be addressed were:

 ·        Will the seismic, structural, and demographic data available in Turkey adequately support existing loss estimation models?

·        Are the human needs and service delivery data describing the impacts of the 1999 earthquakes available and adequate to populate modeling the human needs from future earthquakes?

·         Can the conditional probabilities necessary to link loss estimation and service delivery models be assessed using available data and reliable expert judgment?

 The project had four research objectives:

 1.      To identify the sources of data necessary for estimating potential damage and for determining the attributes of the potential impacted population for a selected set of future Turkey earthquake scenarios.

2.      To build a database describing the medical, feeding, sheltering, material distribution services delivered after the August 17 and November 12 earthquakes.

3.      To assess the quality and limitations of this data.

4.       To develop a preliminary conceptual model for the estimation of service delivery demands. 

 Loss prediction models are designed to predict the physical damage to structure and infrastructure.  This damage may then be used to predict the resulting human impacts and service-delivery needs (Perkins et al. 2000).  The estimation of response requirements must be based on estimation of service-delivery demands.  These demands are determined by human needs.  Estimating the conditional probabilities implied by this relationship requires viable data in four distinct areas: housing damage and functionality, infrastructure damage and functionality, human impacts, and service delivery demands.  Data describing the impact of 1999 Kocaeli and Izmit earthquakes provide a unique opportunity to provide the basis for populating and calibrating models that may be used to predict the service-delivery needs for future earthquakes in Turkey.

Background

 Causes of Building Damage Associated with Casualties

Most of the deaths, injuries, displacement of families in the Kocaeli earthquake were due to collapse of residential housing units, typically in 3 - 6 story reinforced-concrete buildings with masonry infill walls. 

 Although most damage was due to severe ground shaking, additional damage was due to:

fault rupture (causing the partial or total collapse of approximately 100 concrete frame buildings), liquefaction (when granular or sandy materials saturated with water can behave like a liquid, instead of like solid ground), coastal failures (including the failure at Degirmendere, where a large coastal slide carried a hotel into the bay, killing several people), and a small tsunami. (EERI, 2000)

 Sakarya Province, with a population of 731,800, and its capital, Adapazari, with a population of 183,000 were sites of severe liquefaction and amplification of ground shaking due to soil conditions.  A total of 5,078 buildings (27% of the total building stock of the city) were either severely damaged or destroyed.  19,043 housing units collapsed or were heavily damaged in Sakarya Province.  3,891 people were killed in the Province.

 Surface Transportation and Response

Overall response was limited somewhat by highway, road, and rail line damage.  The Istanbul-Ankara highway (E80, or Trans-European Motorway) was closed at several locations by surface fault rupture causing buckling of the road surface, and, in one instance, collapse of an overpass near Arifiye.  The bridge was removed and the highway reopened after three days.  Landsliding in inland areas caused many secondary roads to require clearing, and damaged the highway north of Lake Sapanca.   Shaking damage closed a bridge on a local road near Arifiye. 

The larger concern was the massive traffic jam for the first 24 hours after the Kocaeli earthquake that extended from Istanbul to the impacted area.  The traffic was due, in large part, to people who decided to drive to the impacted area when the phone system was not operational.  As a result, the Governor of Kocaeli placed immediate restrictions on travel into the impacted area following the Düzce earthquake. 

 Similarly, fault rupture closed the rail lines between Izmit Bay and Arifiye.  Although one of the lines was repaired enough to allow limited rail traffic the following day, and the second line was partially repaired after five days, the rail lines did not return to normal for several weeks.   

 Water and Air Transportation and Response

Damage to the Port of Derince, as well as ground failure surface fault rupture damage to the military port at Gölcük, limited the role of these major facilities in earthquake response and recovery. The Ataturk International Airport in Istanbul remained open subsequent to the earthquake.  It was undamaged and handled extensive additional traffic in the week following the earthquake due to international relief efforts. (EERI, 2000)

 Water Supply Damage and Response

Water storage dams experienced few problems, in part because the earthquake occurred in late summer when the reservoirs were relatively low.  Water treatment facilities also experienced minimal difficulties that could be rapidly repaired.  (EERI, 2000)

 The Izmit Water Project services the municipalities along the northern and western shores of the Bay of Izmit.  The Sakarya River main water transmission pipe (2.2m diameter) survived over 3 meters of fault offset without rupturing.  The leak in one kink of the pipeline was repaired approximately one month after the earthquake.   The water system remained operational, although demand increased, probably due to leaks.

 The strong shaking and liquefaction failures in the Adapazari area removed 70% of the pipelines from service, with the remaining 30% suffering leaks. One of the principle researchers has intimate knowledge of the water distribution system servicing the city of Adapazari and is aware that the system in place at the time of the Kocaeli earthquake was old and in need of pipe replacement and rehabilitation and was strained in terms of capacity. The extensive damage led to the decision to completely replace the entire pipeline system. Lack of water aggravated delivery of services for human needs, and  water pipeline replacement was delayed due to building demolition and construction activities. 

 Finally, the Yalova-Gölcük system serves the municipalities on the southern shore of the Bay of Izmit.  Damage to the transmission line led to up to one million people being without water for up to three weeks.  In addition, severe damage to the water distribution system occurred in Gölcük, with 45% of the system being destroyed and 25% damaged.   (EERI, 2000)

 Gas and Electric Power Disruptions and Response

Natural gas pipelines service only a portion of the Izmit area.  This system was installed in the late 1990s and experienced minimal problems. 

 Most of the electric power distribution system was restored within 11 hours of the Kocaeli earthquake, with all repairs being completed within 2 weeks.  However, as might be expected, the most heavily impacted areas were without power for over a week.  (EERI, 2000)

 Telecommunications Disruptions and Response

The telecommunication system was functional within three hours to three days.  Most disruption was due to disruption in the power supply and lack of, or failure of, backup power. (EERI, 2000)

Research Procedure  

The damage and impact of the 1999 Turkey earthquakes are described in a number of Turkish and organizational publications and web sites.   These publications served as the project’s initial data sources.  Of particular use were the Prime Minister’s Crisis Management Center’s publication (Turkish Prime Ministry, 2000), Prime Minister’s Crisis Management Center’s daily press releases on the Izmit and Düzce Earthquakes, the International Federation of Red Cross/Red Crescent daily situation reports, and Survey Results conducted by the State Institute for Statistics (SIS, 1999). Attempts were made to retrieve disaggregated data from the original sources and data describing the response to human needs were extracted and entered into an Access database.   

Activities that responded to human needs were conducted by agencies of the Turkish central government, the provincial government, municipal governments, the Turkish Military, Turkish NGOs, international NGOs, and government teams.  Capturing data from all of these sources proved to be difficult and visits to key officials and researchers in Turkey were critical in increasing the ability to capture information from multiple sources.  In May 2001, the research team visited Turkey and interviewed senior officials in the Emergency Management Agency of Turkey, the General Directorate of Disaster Affairs, the Turkish Red Crescent, the American Red Cross and the Turkish Medical Assocation.  The team visited Kocaeli and met the Mayor of Izmit, the Governor of Kocaeli, and the Director of the Kocaeli Chamber of Industry.  The team visited the Disaster Management Center at Middle East Technical University in Ankara, as well as the Center for Disaster Management and Kandilli Observatory and Earthquake Research Institute at Bogazici University in Istanbul.  In the United States, interviews were conducted with officials of the American Red Cross, INTERACTION, the U.S. Office of Foreign Disaster Assistance, and the United Nations.    Appendix B provides a list of the persons interviewed for this project.   In all cases, these officials were exceptionally helpful and their staffs provided valuable information not available from publicly available sources.  In addition to formal visits, official letters were written requesting data pertaining to the response activities to Ministries and State Organizations who were active in the response operations, and to the governors and the mayors of the affected provinces and municipalities. Since the response efforts of the Ministries and all the other state organizations were coordinated by the Prime Ministry, representatives of these organizations responded to these letters by pointing to the Prime Ministry’s Crisis Management Center Report (Turkish Prime Ministry, 2000).  The survey results on the displaced population (SIS, 1999) were received from the State Institute for Statistics as a response to this letter.  

Results

 During the study period several important data and information sources have been obtained and were used as a basis for an initial effort to build a comprehensive database. The main sources of data that were obtained were either in hard copy report form or were on the Internet. Several reports were identified as main sources for this study. These reports were obtained during the study team’s formal visits to various State organizations as described in the research procedure section.

 Turkish Government Sources

One of the most important sources of event specific data was a report by the Turkish Prime Ministry describing the activities performed by the Ministries and Turkish State Institutions in the aftermath of 1999 earthquakes (Turkish Prime Ministry, 2000). Written in Turkish, this report covers both the August 17 and November 12th earthquakes and is a rigorous, introspective account of information pertaining to the scale of the disasters and the response and recovery efforts, particularly those involving state organizations. It effectively summarizes the data collected from various state organizations and provides overall figures on each topic. The following information can be found in the report:

 1.   Boundaries of the disasters; statistical figures of affected population, affected regions, impacts on housing, impacts on population (dead and injured), and economic loss figures

2.   Description of the organizational structure and responsibilities of Prime Minister’s Crisis Management Center

3.   Search and rescue efforts; activities of responsible government organizations in the aftermath of earthquakes

4.   Description of damage to civil infrastructure systems and emergency response efforts for repairing and resumption of telecommunication and transportation infrastructure

5.   Activities of Turkish Armed Forces

6.   Sheltering services; temporary and permanent housing

7.   Debris removal services

8.   Health and public health services

9.   Domestic and foreign aid

10. Social aid and housing aid

11. Education services

12. Social security services

13. Economic measures taken in the aftermath of the disasters

14. Post-earthquake organizational efforts and legal arrangements

 The data related to items 1 through 10 have been incorporated into a database that is integral to this project.  

A main source for human and housing impact statistics about the August 17 earthquake was an effort of an official of the General Directorate of Disaster Affairs (GDDA) and was published by the Turkish Earthquake Foundation (Ozmen, 2000). This report that was prepared in Turkish contains in-depth statistics on the damage to housing infrastructure according to provinces, districts and sub-districts. It also covers casualty data and damage to infrastructure facilities. Although there are occasional discrepancies between individual and total sum figures, these inconsistencies tend to be of only minor consequence and the data provided was relied upon heavily as a source of casualty and damage figures. A similar report published by GDDA on the November 12 earthquake was also useful. (Ozmen and Bagci, 2000). Tables A-1, A-2 and A-3 show sample data types obtained from these sources for both earthquakes.

 The State Institute for Statistics (SIS) has published the results of a survey that they conducted between September 11 and 19 of 1999 under the auspices of the Ministry of Housing and Resettlement (SIS, Sept. 11-19, 1999). Although that survey was conducted solely to gather information for planning future housing needs, it was found particularly useful for this project. The survey was conducted among all households (59,844 households) that were victims of the August 17 earthquake and were living in either tent cities, public guest facilities within the disaster region, or in detached tents next to their homes at the time of the survey. The survey did not include households who left the area, or who were staying at a hospital at the time of the survey. The survey results provide information relating to the damage state of the housing units (heavily, moderately or slightly damaged or not damaged), the number of people residing in the housing units, and the location of displaced people within 1 month after the disaster. This study is expected to be particularly useful for the purpose of building relational database models for estimating displaced population figures.  A similar survey conducted by SIS assessed the temporary housing preferences of the displaced population (SIS, Sept.23, 1999).  The survey maps the temporary housing preferences of the victims against the damage level and the ownership status (see Table A-5). Another survey result published by SIS attempts to assess the needs of displaced victims in terms of shelters and relief materials, as well as how these needs were being met at the time of the survey (SIS, Oct. 28, 1999). The third survey only provides overall percentage figures (conducted among 49,821 households), unlike the previous two surveys that provided information in greater detail.  These general data are useful for estimation purposes.

 A significant source of information on response activities of the Turkish Government and State organizations is located on the web site of the Prime Minister’s Crisis Management Center (not available online anymore, 08/23/05). The Crisis Management Center (CMC) was formed within 48 hours after the August 17th Earthquake and stayed active until later stages of the recovery efforts. The type of information contained on their web site is daily (sometimes on an hourly basis as information became available) press releases about the situation assessment and response efforts by the CMC. Although the initial press release is dated Aug. 20, 1999 (more than 72 hours after the first earthquake), the information contained in this web site was found critical to the purposes of this project. The timeline aspect conveyed by the press releases provides a sense of revelation about the scope of the difficulties faced by response organizations in mobilizing their resources, especially during the initial stages of the disasters. The time sensitive nature of these accounts makes for an important resource for planning and for divining "lessons learned." The statistical information contained in the press releases of CMC that were of particular use for this project were the housing damage and casualty figures collected from each settlement, the statistics on the search and rescue efforts undertaken mostly by the Turkish Armed Forces (and victims themselves), the mass care efforts (number of persons that were sheltered and fed), and the medical services provided to the victims of the disaster. All of this information was subject to update and revision as more information became available. Figures 1 – 8 were produced from the Prime Ministry’s Crisis Management Center Press Releases illustrate the time dependence of disaster response. The graphs show the data regarding Kocaeli earthquake. Similar data also exist for the Düzce earthquake.

 Figure 1. Reported Death Figures by Day – Kocaeli Earthquake (source: Prime Ministry’s Crisis Management Center, 1999)

 Figure 2. Reported Injury Figures by Day – Kocaeli Earthquake (source: Prime Ministry’s Crisis Management Center, 1999)

 Figure 3. Reported Housing Damage Figures by Day – Kocaeli Earthquake (source: Prime Ministry’s Crisis Management Center, 1999)

Figures 1 through 3 show the evolution of data describing “facts” that were, in principle, knowable immediately after the disaster occurred.   Difficulties in accurately accounting for deaths, injuries, and damaged structures in these two earthquakes are illustrative of the difficulty of capturing an accurate assessment of the impact of a major disaster.  These difficulties also reinforce the case for using the results of predictive models of potential events for scaling initial response efforts.

 Figure 4. Reported (Cumulative) Number of Tents in the Disaster Region by Day – Kocaeli Earthquake (source: Prime Ministry’s Crisis Management Center, 1999)

 Figure 5. Reported (Cumulative) Number of Tents in the Disaster Region by Day – Düzce Earthquake (source: Prime Ministry’s Crisis Management Center, 1999)

Figure 6. Reported Number of Health Personnel in the Disaster Region by Day – Düzce Earthquake (source: Prime Ministry’s Crisis Management Center, 1999)

 Figure 7. Reported Number of Response Personnel in the Disaster Region by Day – Düzce Earthquake (source: Prime Ministry’s Crisis Management Center, 1999)

 Figure 8. Reported (Cumulative) Number of Relief Material Distributed to Displaced Persons by Day – Düzce Earthquake (source: Prime Ministry’s Crisis Management Center, 1999)

Figures 4 through 8 show the temporal evolution of selected aspects of the response effort.  Tents were still being erected and basic relief supplies such as blankets, sleeping bags, and stoves were still being distributed a month after both the August and November earthquakes.  In the United States, peak shelter populations are typically experienced 3-7 days after a major earthquake (Perkins et al., 2000).  It is apparent that peak populations of temporary shelter populations were experienced in tent cities months, not days, after the earthquakes in Turkey. 

 The Prime Ministry’s report (Turkish Prime Ministry, 2000) that was discussed at the beginning of this section is basically a collection of the overall figures and information contained in the press releases.

 Since future modeling and planning efforts would require the use of base (pre-earthquake) statistics regarding the settlement that are susceptible to earthquake damage, an effort was made to identify the sources of these types of data. Detailed historical population counts of the settlements affected by the 1999 earthquakes in the province, district, and sub-district (including municipalities and smaller settlements other than villages) level were obtained from the State Institute for Statistics (SIS, 1945-1997). However, no attempt was made to obtain location-specific data regarding land use and building inventory data because these data are too specific and obtaining them would be outside the purpose of this project. These types of data can be obtained from local governments. Hazard-related data including geologic maps, intensity exposure, liquefaction data and the like were not collected for similar reasons and these types of data are located in various state and academic institutions and can be obtained from those institutions upon request.

 Other useful sources of information pertaining to the Turkish government are listed under the Related Materials section. Some statistics on individual provinces and municipal settlements can be found on their web sites. The web sites covering damage and casualty-related information that were drawn upon during the writing of this document are listed in the section on Related Materials. These materials are useful for comparison purposes and present a documentary of evidence relating to the scale of the disaster and testimony to the activities that followed. Narrative accounts are also of use for the purposes of building predictive models for emergency response and relief operations.

Turkish NGOs

The Turkish Red Crescent is the most active of the Turkish NGO's and is very experienced in large scale disasters. The mass-care activities of the TRCS in the aftermath of the Kocaeli and Düzce Earthquakes were extremely challenging. The activities of the TRCS have been described both in the Prime Ministry’s Crisis Management Center report and in the online press releases that were described earlier.

 The only data that were obtained in the form of an electronic database was the feeding tracking data from the Emergency Response and Relief Unit of the Turkish Red Crescent (TRCS). The TRCS kept an electronic record of the number of meals they served to the victims on a daily basis. Feeding hundreds of thousands of persons on a daily basis is a major undertaking that requires a significant amount of resources during times of a disaster. It appears from the data that TRCS set up their mobile kitchens one day after the August earthquake and increased their capacity as more victims populated the tent cities and temporary shelters. The following graph shows that the feeding activity starts on the 18th of August and reaches its initial peak (91,000 persons/day) 2 months after the date of the disaster and then declines until the November 12 earthquake strikes, at which time the trend shows an increase with the highest peak attained (226,000 persons served meals/day) approximately 5 months after the August 17 earthquake and 2 months after the November 12 earthquake. The main use of this database would be to map these results against displaced population figures and see how they correlate. The results should prove valuable as a resource for future planning estimates. Of peripheral interest might be its use for comparative studies involving behavior of displaced persons in other regions or countries.

 Figure 10. TRCS Feeding

 The Turkish Medical Association was one of the actively participating NGOs during the disaster relief operations. Immediately after the Kocaeli earthquake, emergency medical teams were formed quickly inside and outside the disaster region and dispatched to the staging areas. These units each had specific responsibilities for aiding Ministry of Health personnel (since the Ministry of Health is the responsible agency for coordinating the medical activities during the declaration of disasters according to the Disaster Law). The undertakings of the TMA units ranged from disaster medical, mental and public health services, to assisting and advising the responsible government organizations with feeding, sheltering, communication, and debris-removal activities. The TMA played a very active role in advising the Ministry of Health in formulating their disaster preparedness policies as well as in educating and recruiting volunteers for disaster services before and after the 1999 earthquakes. The TMA has an active website;  two of the most important reports providing an account of its activities during the response and recovery phase of the 1999 earthquakes can be downloaded from that website (www.ttb.org). The Preliminary Report that was submitted to the Ministry of Health is particularly useful for the purpose of obtaining a detailed and a critical account of the situation assessment before and after the August 17 earthquake (TMA, 1999). The health services statistics can be of use in developing a medical service delivery database. The type of statistics residing in the preliminary report is critical since they cover the situation assessment within the first few weeks following the disaster. Table A-6 shows the type of statistics collected by the TMA within three weeks after Kocaeli earthquake.

International Relief Organizations

The International Federation of Red Cross and Red Crescent Society (IFRC) published their daily situation assessment reports describing the activities of the participating national societies in great detail on the web (www.ifrc.org). A total of 48 sitreps provide daily statistics on damage and casualties together with the mass-care operations including logistics, sheltering, medical assistance, sanitary and health services, bulk distribution, and feeding. Since the damage and casualty figures were obtained from the media and the TR government sources, they are secondary data and are not of particular significance in general. However, it would be an interesting study to see how the information gets better with time and the response activities get improved accordingly. These sitreps provide the activities of some forty-five National Societies (around one-quarter of the International Federation’s total membership) who have dispatched relief supplies and equipment in the aftermath of the disasters. Although narrative in format, some of the statistical data were extracted and converted into a database format. The IFRC sitreps in general are extremely useful for research that would help improve future emergency response planning.

 Similarly, the U.S. Agency for International Development (USAID) and its Office of U.S. Foreign Disaster Assistance (OFDA) have published their on-scene efforts during the response and recovery time period. These can also be found on the web (www.usaid.gov). A total of 29 fact sheets on the Turkey earthquakes exist. These are not as comprehensive as the IFRC sitreps due to the smaller scale of operations of USAID-OFDA. The U.S. Government’s efforts included providing emergency response such as dispatch of a 70 person search-and-rescue team, setting up tent cities, medical and dental care, and sanitary health services, and coordinating and distributing materials from  U.S. donor organizations. The use of these factsheets are marginal for helping build databases since the information is usually in dollar figures or in bulk values, rather than itemized.

 Doctors Without Borders/Médecins Sans Frontières (MSF) was one of the other most active international relief organizations with medical expertise that were participating during the response and relief operations in the aftermath of Kocaeli earthquake. It sent four medical teams to the region a day after the disaster and actively worked in four locations: Izmit, Golcuk, Bursa and Ankara.  These doctors treated several hundred patients within the first week after the disaster. Brief accounts of the MSF team activities were located on their website (www.doctorswithoutborders.org)

 Other accounts on the disaster response and relief operations on the 1999 Turkey earthquakes by several international relief organizations are located on the ReliefWeb site which is a project of the United Nations Office for the Coordination of Humanitarian Affairs (OCHA) (www.reliefweb.org).

  Conclusions

 The research described above was motivated by four factors:

1.      The perishable nature of critical information collected by organizations responding to the 1999 Turkey earthquakes.

2.      The importance of ensuring quality data for current and future loss and impact modeling. 

3.      The ability to build on existing collaborative relationships.

4.      The similarity between U.S. and Turkey earthquake hazards.

Conclusions about interpretation of data

Although detailed analysis of the data collected was not the primary purpose of this project, several observations may be made using only basic descriptive analysis techniques:

·        Time-phased displays of data describing damage and displaced persons illustrated that information describing impact after the Kocaeli earthquake was initially very incomplete. Early estimates were inaccurate; in some cases it took a significant period of time (weeks and months) to obtain an accurate measure of the disaster’s impact.

·        Time-phased data figures also show that mass-care response to the Kocaeli earthquake (provision of tents and feeding) was slow but that the response to the November earthquake was more rapid and better documented.

·        The data describing the distribution of damage and the distribution of services provided should be analyzed for relationships and anomalies.  Descriptive analysis provides some indication of these relationships.

·        Not all the data needed for needs analysis are available.  The extent of the distribution of bulk food supplies by the Turkish Armed Forces and civilian sources during the early days of the relief effort is not clear from the data collected.  There is no way to ensure that the efforts of all non-governmental organizations that participated in the relief effort are captured in the statistics of the Turkish government, the IFRC, and the TRCS.

Numbers change over time

One way that numbers change over time is that the number of people needing care (feeding, shelter, etc.) can vary from day to day.  For example, the Turkish Red Crescent set up their mobile kitchens one day after the August earthquake and increased their capacity as more victims populated the tent cities and temporary shelters. The feeding activity reached its initial peak (91,000 persons/day) 2 months after the August earthquake.   It then declined until the November 12 earthquake strikes, at which time the trend shows an increase with the highest peak attained (226,000 persons served meals/day) approximately 5 months after the August 17 earthquake and 2 months after the November 12 earthquake.

Another reason numbers can change over time is the quality and accuracy of data improves.  For example, the Turkish Prime Minister’s Crisis Management Center web site's press releases about the situation assessment and their response efforts provide insight into the scope of the difficulties faced by response organizations in mobilizing their resources, especially during the initial stages of the disasters.  In particular, the press releases document information on changing estimates of housing damage and casualty figures, search-and-rescue efforts, the mass-care efforts (number of persons that were sheltered and fed), and the medical services provided to the victims of the disaster. All of this information was subject to update and revision as more information became available. 

Assessment of form and completeness of data

The original data drawn upon in this project was collected and stored by organizations that were responding to an unexpected catastrophic event.  In the aftermath of any significant natural disaster, attention is focused on providing essential rescue, medical, and mass-care services to victims.  Government and non-governmental organizations do the best they can with what they have until adequate response resources can be mobilized.  Keeping track of who is providing what service is clearly of lower priority than responding to critical needs.  This means that if the responsibility and mechanism for collecting data are  not in place prior to the event, that data will be collected in a haphazard manner and that the data collection will be driven by the motivations of hundreds of unique organizations.  It is not surprising that the data available on the initial response to a catastrophic event may be incomplete and inaccurate.  In the case of the Turkey earthquakes, the completeness of the data describing the response to human needs were affected by the following:

·        Original data were   collected/stored by an individual organization. Efforts to integrate and reconcile data from different sources was made after a significant time delay.  Integrated data that has been reconciled and aggregated are typically available only in paper form. 

·        Data were collected by different organizations using different selection criteria (e.g. per person, per family, per unit of service delivered) and by differing geographic levels of decomposition (provincial, district, municipal)

·        Data definitions were determined by organizations collecting data and it is difficult to determine if definitions are compatible between sources. 

Services are not delivered in a vacuum

Basic human needs needed to be met after these earthquakes.  People needed to be rescued from collapsed buildings and to receive emergency medical care.  They needed water that was safe to drink, food, and shelter.  Later, they needed help to be reunified with their families and in moving forward with their lives.  Yet provision of these services had to occur in highly disrupted areas.  Roads were closed and jammed with traffic.  Some utilities, particularly water distribution lines, were not functional and frantic family members hoping to get through to their relatives jammed the phone system.  Given the magnitude of these disasters, the efforts of the service deliverers to collect data on what they were doing are to be commended.

 In addition, recovery occurs within the context of the economic and social fabric of the region affected.  In the U.S. it has been observed that when industrial operations are lightly impacted, employees of those companies are more likely to remain and require shelter for longer periods.  On the other hand, if companies go out of business, the employees no longer have jobs and are thus more likely to leave the area.  In the Turkish earthquakes, major manufacturing facilities survived and resumed operation relatively quickly. Thus, the need for sheltering was greater than it would have been if more damage to industrial facilities had occurred. 

Why are data currently collected?

Human needs data were not collected for earthquake researchers and people interested in modeling the potential impacts of future earthquakes.  For example, local governments were collecting data on structural damage to make decisions on safety related to occupying or reoccupying those structures.  Various non-governmental organizations (NGOs) were collecting data on the number of meals served or tents erected and compiling macro-data for use in fund raising.  Although many structural engineers focused their efforts on case studies, other researchers collected statistical data on structural damage that will be of use to those modeling future damage potential.  Similar techniques need to be employed by those doing research on human needs. 

What should be collected and by who?

Two organizations in Turkey appear to have both the organizational authority and responsibility to establish a coherent information management capability that would include standardizing data definitions, geographical boundaries, and collection procedures: the Turkish Emergency Management Agency (TEMAD) and the Turkish Red Crescent Society (Kizilay).   Funded by the World Bank MEER project, TEMAD will develop the Turkish Emergency Response Plan and Emergency Response Organization.  The development of data collection and information management capabilities is an integral part of the MEER project.  Kizilay, like the American Red Cross, has functioning units throughout the country that will be on scene for any earthquake or other major disaster.  Supported by the American Red Cross, the IFRC and other national societies, Kizilay is upgrading its ability to collect and manage information during major disaster operations.   A coordinated TEMAD/Kizilay effort to ensure that data describing the response to human needs is collected after future earthquakes will significantly improve the development of response and recovery plans and capabilities.  The database created in this project provides guidance on the type of data that could be collected in a coordinated effort. 

What opportunities exist for using these data?

The products of this research will enhance earthquake preparedness and planning in Turkey and will improve the risk and vulnerability modeling capability in the United States. 

Figure 11 shows the types of data and relationships governing them that are needed to populate damage models (Perkins, 1995, Perkins et al., 1996, Perkins et al., 2000).  Building inventory data, data describing actual ground shaking and liquefaction, and structural damage data for the two Turkish earthquakes has been collected by the Turkish government agencies and universities.  The focus of this project was on the perishable event specific data shown in Figure 12.  The estimation of response requirements must be based on estimation of service delivery demands.  These demands are determined by human needs and modeling this relationship requires viable data in four distinct areas: housing damage and functionality, infrastructure damage and functionality, human impacts, and service delivery demands as shown in Figure 12.  The most immediate challenge for future analysis is to investigate whether or not the data available are adequate to support this type of modeling.  Where data are  not adequate, methods of obtaining and utilizing expert judgment may be another promising research goal. 

 Figure 11:  Pre-Disaster Data Required For Loss Estimation

  Figure 12: Scenario-Specific Data for Determination of Human Needs and Service Delivery Demands

 Future Collaborative Research

 The work described in this report built on existing collaborative partnerships in the U.S. and between U.S. and Turkish academic, non-governmental, and governmental organizations.  The GW ICDRM and ABAG have collaborated for six years in the development of loss estimation and estimates of human needs and service delivery demands following potential earthquakes in the San Francisco Bay Area (Perkins, Chaqui, Harrald, and Jeong, 1996; Perkins et al., 2000).  The results of this work are currently being used by the American Red Cross Bay Area Chapter (ARCBA) as the basis for response planning (ARBCA, 1999). ARCBA disaster planners will assist in the interpretation of the data collected in this project.  GW, ABAG, and ARCBA all have strong, independent ties to Turkish organizations that will be key to the success of the proposed research.  GW researchers have worked with faculty from several Turkish Universities, including Middle East Technical University, Ankara University, Bogazici University, and Kultur University.  GW University has formed a partnership with Bogazici University in Istanbul and both Universities have committed funding from their endowments to promote the growth of joint programs.   A key element in the BW/BU partnership is a disaster management collaboration that involves the GW ICDRM and the GW Department of Engineering Management and Systems Engineering with the Bogazici Kandilli Observatory and Earthquake Engineering Department and the BU Center for Disaster Management.  The universities jointly sponsored a Disaster Management Workshop in Istanbul in December 1999, and the GW ICDRM assisted Bogazici to establish the BU Center for Disaster Management (CENDIM) in 2000.   For several years, the ABAG and the Union of Marmara Metropolitan Region (UMMR) established a formal relationship between council of governments.  ABAG and UMMR have an existing relationship for data exchange and network building. 

 The American Red Cross has made a major commitment to assist the Turkish Red Crescent Society, and a key part of this commitment has been building and maintaining a close relationship between the San Francisco Bay Red Cross and the TRCS.  The Red Cross has announce the establishment of a joint American Red Cross—Turkish Red Crescent Earthquake Research and Information Center for Humanitarian Assistance.  This center will initially have two locations—Ankara, Turkey and the San Francisco Bay Area, California.  GW and ABAG intend to collaborate closely with this new Center and with the Turkish Universities it supports.   These organizational and personal contacts will be essential to the successful future collaboration as the data developed in this project is used to support earthquake disaster planning in Turkey.

 Specific areas where the data collected and structured for this report will facilitate collaboration include:

Supporting development of loss estimation and impact models:  There is significant interest in Turkey in developing a loss estimation methodology similar to the US HAZUS system.    Turkish earthquake engineers, seismologists, and civil engineers have developed the data and relationships necessary to build such a modeling system.  As described above, the data obtained in this project will provide the ability to create the ability to extend the damage prediction models to provide the ability to estimate the human needs and demand for relief services following a major earthquake in Turkey.

Supporting contractors funded by the World Bank MEER project:  the response planning and training elements of the Marmara Earthquake Emergency Reconstruction (MEER) project cannot be successfully completed without estimates of requirements of services required to meet human needs.  The data will be essential to the development and testing of response strategies and plans.

Supporting individual researchers from U.S. and Turkey:  The GW ICDRM has already been approached by highly qualified individuals wishing to visit the Institute and work with the GW research staff on various projects relating to this data.  Mechanisms for enabling this exchange include Fulbright fellowships, post-doctoral awards, and sabbatical positions.  Similarly, GW faculty may visit Turkish educational institutions on similar arrangements

Contents of Project Website

The complete products of this research have been posted on a web site maintained by The George Washington University Institute for Crisis, Disaster, and Risk Management. The URL for the project website is www.seas.gwu.edu/~icdm/turkey. The website contains:

1)      A listing of web-based data sources constructed from multiple sources and a description of print references,

2)      A set of annotated data tables in Excel spreadsheet form:

a)      Building and human impact statistics for both Kocaeli and Düzce Earthquakes (Provincial, district and sub-district level) based on data published by the General Directorate of Disaster Affairs

b)      District level displaced population statistics relating the post-earthquake residency type (in tent city or outside tent city) of the displaced population to the damage state of the residences (collapsed/heavy damage, moderate damage, slight damage and no damage), and household size based on the survey results conducted between Sept.11-19 1999 and published by the State Institute for Statistics

c)      District level displaced population statistics relating the pre-earthquake home ownership type (own, rent/lease, employer provided or jointly owned) to the household size and damage state of the residences (collapsed/heavy damage, moderate damage, slight damage and no damage) based on the survey results conducted and published by the State Institute for Statistics on Oct.28 1999.

d)      Data tables describing the daily progress of :

·        Building and human impact statistics and statistics on the government response (search and rescue, sheltering, medical, feeding and other mass care activities) based on the daily press releases posted by the Prime Minister’s Crisis Management Center (from Aug.20 to Oct.19 1999 and from Nov.13 to Dec.20 1999)

·        Feeding tracking data (number of meals served per day) obtained from the Turkish Red Crescent (from Aug.17 1999 to Aug.31 2000)

3)      Relational Databases for both Kocaeli and Düzce Earthquakes in MSAccess20002 format. This database is constructed using the excel spreadsheet data described in item 2 above and relates the information in two different ways: a)according to impact location, and b) according to time (days)

List of References

 1.      Doctors Without Borders (Medecins Sans Frontieres) web pages on August 17 earthquake (www.doctorswithoutborders.org)

2.      Earthquake Engineering Research Institute, 2000. The 1999 Kocaeli, Turkey, Earthquake Reconnaissance Report:  EERI, Oakland, CA, Supplement A to Volume 16, December 2000, 461 pp.

3.      International Federation of Red Cross and Red Crescent Organizations, Sitreps (1 through 48 on Turkey Earthquakes), IFRC, http://www.ifrc.org/

4.      Ozmen, B. and Bagci, G. (Eds.), Nov. 2000. Report on the 12 November Düzce Earthquake, T.R. Ministry of Construction and Resettlement, General Directorate of Disaster Affairs, Earthquake Research Department, 224 pp. (in Turkish)

5.      Ozmen, B., June 2000. Quantitative Damage Assessment of Izmit Bay Earthquake, Turkish Earthquake Foundation, Istanbul, 132 pp. (in Turkish)

6.      Perkins, Jeanne B., Harrald, John R., Mikulis, Kathleen, Kramer, Chris, Renda-Tanali, Irmak, Bettridge, Matt, 2000. Preventing the Nightmare – Post-Earthquake Housing Issue Papers:  Association of Bay Area Governments:  Oakland, 80 pp.

7.      Prime Ministry Crisis Management Center, Press Releases on 1999 Earthquakes

8.      T.R. Prime Ministry Crisis Management Center, (Turkish Prime Ministry) 2000. 1999 Earthquakes: Activities by The Ministries and State Institutions Following August 17 And November 12 Earthquakes, 296 pp. (in Turkish)

9.      T.R. State Institute for Statistics, (SIS) 1999. A Study on the Determination of Temporary Housing Needs in the Earthquake Region (Sept.11-19 1999), 61 pp. (in Turkish)

10.  T.R. State Institute for Statistics, (SIS) 1999. News Bulletin: A Study on the Determination of Temporary Housing Needs in the Earthquake Region, (Sept. 23, 1999), 15 pp. (in Turkish)

11.  T.R. State Institute for Statistics, (SIS) 1999. News Bulletin: Results of a Survey of Households In Disaster Region (Oct. 28, 1999), 28 pp. (in Turkish)

12.  T.R. State Institute for Statistics, (SIS). Population Statistics, 1945-1997  (in Turkish)

13.  Turkish Medical Association, 2000. Year 1 Evaluation Report, http://www.ttb.org/ (in Turkish)

14.  Turkish Medical Association, Sept. 1999. Preliminary Report on August 1999 Earthquake Submitted to Ministry of Health, www.ttb.org (in Turkish)

15.  Turkish Red Crescent, Directorate of Emergency Response and Relief, Meal Tracking Data (in electronic form: Excel spreadsheet) TRCS, August17 1999 through August 31 2000.

16.  United Nations Office for the Coordination of Humanitarian Affairs, Reliefweb Site (www.reliefweb.org), 2000.

17.  USAID, Fact Sheets on Turkey Earthquakes (1 through 29), www.usaid.gov 

 Related Material

 1.      Architectural Institute of Japan Reconnaissance Team (with BU, ITU and METU), Dec. 1999. Report on Damage Investigation after Kocaeli Earthquake by Architectural Institute of Japan (Draft), 50 pp.

2.      Barka, A. et.al (Eds.), 2000. The 1999 Izmit and Düzce Earthquakes: Preliminary Results, Istanbul Technical University, Istanbul. 350 pp.

3.      Daley, R., Karpati, A., and Sheik, Mani, May 2001. “Needs Assessment of the Displaced Population Following the August 1999 Earthquake in Turkey”, Disasters, 25(1): 67-75.

4.      Demirtas, R. (Ed.), 2000. Report on the 17 August Izmit Bay Earthquake, T.R. Ministry of Construction and Resettlement, General Directorate of Disaster Affairs, Earthquake Research Department, 295 pp. (in Turkish)

5.      Earthquake Spectra, 1999 Kocaeli, Turkey, Earthquake Reconnaissance Report, Supplement to Volume 16, EERI, Dec. 2000, 461 pp.

6.      EERI, Earthquake Research Bulletin, Dec. 2000, No: 69, pp.113-126, EERI,

7.      The Izmit (Kocaeli), Turkey Earthquake of August 17, 1999, Special Earthquake Report, October 1999.

8.      Erdik, M., 1999 Kocaeli and Düzce (Turkey) Earthquakes, Report no: 2000-12, Dec.2000.  Bogazici University, Kandilli Observatory and Earthquake Research Institute Department of Earthquake Engineering, Istanbul. 37 pp.

9.      Gendarmerie General Commandership, Head of Operations, 2000. Annual Publication, 10 pp.  (in Turkish)

10.  Istanbul Technical University, Preliminary Evaluation Report. (in Turkish)

11.  Izmit City General Assembly, Demographic, Economic and Social Situation Assessment following Kocaeli Earthquake, 233 pp. (in Turkish)

12.  Kocberber, E. and Erdem, A., 2000. 2000 Building Census, State Institute for Statistics, 11 pp.

13.  MCEER, The Marmara, Turkey Earthquake of August 17, 1999: Reconnaissance Report, Technical Report MCEER-00-0001, University of Buffalo, State Univ. of New York. 190 pp.

14.  Mitchell, William A, Social, Political Response: Kocaeli, Turkey Earthquake, August 17, 1999, Baylor University.

15.  Ozmen et.al., 1997. TR Ministry of Construction and Resettlement, General Directorate of Disaster Affairs, Earthquake Zone Investigation by the Use of GIS, 90 pp.

16.  USGS, Nov. 26, 1999. Initial Geotechnical Observations of the November 12, 1999, Düzce Earthquake, a report of the Turkey-US Geothechnical Engineering Reconnaissance Team, 70 pp.

17.  USGS, Sept. 3, 1999. Initial Geotechnical Observations of the August 17, 1999, Kocaeli Earthquake, A report of the Turkey-US Geotechnical Engineering Reconnaissance Team, 68 pp.

Web Resources

1.      http://www.duzce-bld.gov.tr/ (Web site of Düzce Municipality)

2.      http://www.izmit-bld.gov.tr/ (Web site of Izmit Metropolitan Municipality)

3.      http://www.kocaeli.gov.tr/ (Web site of Kocaeli province)

4.      http://www.sakarya.gov.tr/ (Web site of Sakarya province)

5.      http://www.sg.com.tr/kocaeli (Kocaeli Disaster Information Sharing System)

6.