The Study on a Disaster Prevention/Mitigation Basic Plan … · Prioritization of bridge retrofitting work was evaluated through assessments of damages to bridges and major ... damages.

13th World Conference on Earthquake Engineering Vancouver, B.C., Canada

August 1-6, 2004 Paper No. 1361

THE STUDY ON A DISASTER PREVENTION/MITIGATION BASIC PLAN IN ISTANBUL

PART 2 - EVALUATION OF URBAN VULNERABILITY -

Yutaka KOIKE1, Osamu IDE2, Ryoji TAKAHASHI3, Mahmut BAŞ4, Mustafa Özhan YAĞCI5

SUMMARY This paper discusses evaluation of risks of urban infrastructures, especially road network system and lifelines, in a congested city. Securing road network system is one of the most important issues in emergency activities and disaster recovering processes. Prioritization of bridge retrofitting work was evaluated through assessments of damages to bridges and major emergency facilities, and assessments of effects to traffic and transportation caused by these damages. Conditions of road blockage, which were caused by collapse of buildings, were also estimated and possible areas of isolated residential blocks were identified and shown to exist widely. As a result, development and improvement plan for alternative transportation routes; facilities for maritime transportation were evaluated. Natural gas distribution pipelines are one of the major and important lifeline utilities and have been rapidly developed recently for reducing air pollution. Results of damage estimation for these newly constructed pipelines showed high earthquake resisting capacity. On the other hand, it was identified that building damage by earthquake motion would cause damage to gas service boxes, which are installed in each natural gas-utilized building and that possibility of gas leakage from unbroken gas distribution pipelines would be high over a wide area. This indicates that just strengthening of lifeline facilities and pipelines is not necessarily sufficient for securing safety and that development of total system down to end supplier should be fully considered.. These conclusions and results are applicable for consideration of earthquake vulnerability improvement in other congested urban areas.

1 Manager, Oyo International Corporation, Tokyo, Japan. Email: [email protected] 2 Manager, Oyo Corporation, Saitama, Japan. Email: [email protected] 3 Engineer, Pacific Consultants International, Tokyo, Japan. Email: [email protected] 4 Director, Istanbul Metropolitan Municipality, Istanbul, Turkey. Email: [email protected] 5 Engineer, Istanbul Metropolitan Municipality, Istanbul, Turkey. Email: [email protected]

1. INTRODUCTION The authors had carried out seismic microzonation study in Istanbul during “The Study on A Disaster Prevention/Mitigation Basic Plan in Istanbul” [1]. In this study, damage estimation was implemented covering a wide variety of fields. Evaluation of urban vulnerability helps to minimize damage by earthquakes by indicating where to start initial action to strengthen urban structures. In this study, evaluation of urban vulnerability including infrastructure and lifelines were analyzed to locate vulnerable areas to earthquakes. For infrastructure, evaluation of road network importance was analyzed to determine the effective and appropriate emergency road network and prioritize bridge reinforcement. In addition, probable road blockage was estimated to define areas where isolation risk is high. For lifelines (consisting of water, sewage, gas, electricity, and telecommunications), damage estimation of each lifeline was gauged to determine vulnerability of the network. Advanced understanding of vulnerability of infrastructure and lifelines will aid administrations to prepare for possible earthquakes and develop a strengthening plan aimed at minimizing unnecessary damage properly and efficiently, and with minimum efforts.

2. IMPORTANCE EVALUATION OF ROAD NETWORK Since various types of communication and supply lifelines are buried along roads, roads are the most important means of traffic and transportation supporting urban functions. Therefore, once roads are damaged by earthquake, individual structures buried along roads will be physically damaged and this leads to potential malfunction of total systems. Furthermore, roads play important roles in many activities, such as evacuation, information gathering, rescue, medical aid, etc., all of which are required immediately after earthquakes. Roads are also important roles in the transportation of relief goods and restoration activities that are critically required after earthquakes. It is essential to study the current situation of roads and their functions quantitatively in consideration of 1) preventive measures against earthquake damage of roads, 2) restoration plan of roads and also 3) implementation priority. This includes evaluation of importance degree of road networks, routes and sections. Based on the above viewpoint, 1) the importance of road networks, 2) the prioritization evaluation of reinforcement of bridges against earthquakes, and 3) the estimation of damage from road blockades caused by collapse of roadside buildings, are described in the following sections. 2.1 Evaluation Method In evaluating the importance of the road network in view of disaster prevention, it is necessary to study the relative importance of routes along sections of the entire network, and also to study impact caused by damaged bridges. These elements are directly reflected to designate emergency road networks and to prioritize bridges to be strengthened. Figure 1 shows the flow of the evaluation study of the importance of road networks. (1) Importance Evaluation of Attributes: The study area is divided into 500 m by 500 m grid to form sections, with roads in each grid. As shown in Figure 1, importance is evaluated according to the sum of the attributes with optional grades and weight based on each characteristic. (2) Importance Evaluation of Road Network Characteristics: Importance of Road Network Characteristics is evaluated from results of Shortest Path Analysis connecting facilities for disaster management in each emergency response phase, focusing on function of road just after earthquake and during recovery phase. (3) Importance Evaluation of Routes and Sections:

Importance of route and section is evaluated based on the importance of attributes and road network characteristics.

Disaster PreventionFactor

Traffic CharacteristicsFactor

Route CharacteristicsFactor

Crossing Large Bridgeand Viaducts

Importance of Route andSection (Attributes)

Disaster PreventionFactor

Traffic CharacteristicsFactor

Importance of Route andSection (NetworkCharacteristics)

Length of Bridge onMain Line

Conditions Under Bridge

Impact ofBridge Collapse

Probability ofBridge Collapse

Importance Evaluationof Attributes

Importance Evaluationof Network

Characteristics

Impact Evaluation ofBridge Collapse

Importance Evaluation of Route and Section

Importance Evaluation of Road and Bridge inEarthquake Disaster

Priority of BridgeReinforcement

Figure 1 Flow of Importance Evaluation

(4) Impact of Bridge Collapse: Scale of impact from bridge collapse is based on importance of road where bridges are located and conditions under bridge (importance of road under bridge). (5) Possibility of Bridge Collapse: For possibility of bridge collapse, the scoring method by Kubo and Katayama was applied. In this study, 480 bridges in total were evaluated. (6) Impact Evaluation of Bridge Collapse: Scale of impact from bridge collapse was evaluated based on impact on bridge collapse and probability of bridge collapse. (7) Importance Evaluation of Road and Bridge Based on importance evaluation of route and section and impact evaluation of bridge collapse, comprehensive necessity of road importance and reinforcement of bridges was evaluated, and priorities were established of the order for strengthening bridges. 2.2 Importance Evaluation of Route and Section Figure 2 shows result of importance evaluation of attributes. Evaluation of attributes is categorized into 3 levels, and generally in the study area, major belt highways and radial highways were selected as important roads. Figure 3 shows the results of importance evaluation of network characteristics. 103 locations “just after earthquake” and 502 locations “during recovery phase” are classified as base facilities for disaster management. As shown in Figure 3, importance was categorized into 3 levels by overlaying result of Shortest Path Analysis “just after earthquake” and “during recovery phase.” Result of importance of

(1)

(2)

(3)

(4)

(5)

(6)

(7)

network characteristics showed a similar trend as with the results of importance evaluation of attributes; however, results showed high importance on radial roads from center of urban area, which were not abstracted from the results of importance evaluation of attributes.

Figure 2 Road Priority Based on Road Attribute

Figure 3 Road Priority Based on Disaster Management

Figure 4 shows results of importance evaluation of routes and sections categorized into 3 levels by overlaying result of importance evaluation of attributes and network characteristics. For roads within the study area, belt highways in east to west direction and radial highways from the center of urban area were clearly evaluated as especially high importance roads. In addition, the road routes and sections evaluated as high importance correspond to the actual traffic demand on routes and sections that form major road network.

Figure 4 Road Priority: Conclusion

2.3 Impact Evaluation of Bridge Collapse In case of occurrence of bridge collapse, by overlaying bridges which effect on route and section, and bridges which have high probability of collapse, spatial relationships with high importance routes and sections were abstracted. Figure 5 shows bridges with high impact from bridge collapses, which are categorized into 3 levels. 2.4 Priority Evaluation of Bridge Reinforcement Based on the result of importance evaluation of routes and sections and impact evaluation of bridge collapse, priority of bridge reinforcement was set, and the results are categorized into 4 levels as shown in Figure 6. For bridge reinforcement, priority is evaluated as high if importance of bridge (high impact in case of collapse) and importance of route and section is also high.

Figure 5 Bridge: High Damage Possibility: Priority for Road Network

Figure 6 Bridge with High Damage Possibility: Reinforcement Priority

Based on Damage Possibility and Priority of Road Network

2.5 Estimation of Probable Road Blockage Road blockage caused by building collapse from earthquake is estimated based on conditions of buildings, width of debris from collapse and conditions of roads and sidewalks. Figure 7 is a summary of related factors to estimate road blockage. Probability of road blockage is estimated by following formula:

)()()( cPbPaPPf ××=

Where, P(a): Probability of collapsed building debris falling onto road P(b): Probability of collapse of buildings on both sides of road P(c): Probability that remaining road width is less than 3 m Estimation of road blockage covering roads within the study area and isolation of areas caused by road blockage were assessed. For roads in 500 m by 500 m grid system, isolation risk was configured from assessment indices as shown in Table 1. Areas becoming isolated by road blockage are shown in Figure 8, based on the assessment indices shown in Table 1. According to the assessment, high risk of isolation areas is very much concentrated in South Western part of the study area. In such areas that are isolated by road blockage, severe difficulties will be encountered in evacuation and rescue activities, removal of collapsed buildings and transportation of commodities. Therefore, a new policy on road arrangement and improvement of land utilization will be required to reduce the risk of isolation.

Table 1 Relation between Index and Road Blockage Evaluation State of Road Blockage

Risk Level of Isolation Road 2–6 m Wide Road 7-15 m Wide

Case 1*1 Most of roads are blocked N/A

Case 2*2 Blockage probability is higher than 50% No road 7-15 m wide exists Very High

Case 3*3 Blockage probability is higher than 50% Blockage probability is higher than 50%

High Case 1 Blockage probability is higher than 50% Blockage probability is 30 to 50% or higher

Case 1 Blockage probability is 30 to 50% or higher No road 7-15 m wide exists Little High

Case 2 Blockage probability is higher than 50% Blockage probability is 10 to 20% or higher

Low Case 1 Other than mentioned above Other than mentioned above

Note: *1 If most of 2-6 m wide roads are blocked, no matter 7-15 m roads are also blocked or not exist, isolation risk is “very high”.

*2 If 7-15 m wide roads does not exist in a grid, and more than 50% of 2-6 m wide roads are blocked, isolation risk is “very high”.

*3 If more than 50% of both 2-6 m wide roads and 7-15 m wide roads are blocked, isolation risk is “very high”.

Figure 7 Flow of Road Blockage Estimation

Conditions of Building

Conditions of Roads and Sidewalks

Assumption of width of collapsed building

Total number of buildings within road linkage

Road Width

Probability of collapsed debris onto road

Probability of Building Collapse

Building Structure

Probability of Road Blockage

Figure 8 Isolation Risk Caused by Road Blockage

In the area where road blockage is concentrated caused by building collapse, it is extremely important to remove and dispose of disaster debris in order to promptly secure road transport and implement recovery activities quickly after occurrence of earthquake. Consequently, for such areas, important measures include appointment of emergency road network considering removable of disaster debris and development of stockyards, etc. Furthermore, to secure smooth transportation of commodities, it is important to formulate new transport routes using combination of marine and land transport, and not only rely on road transport.

3. SEISMIC RESISTANCE EVALUATION OF LIFELINE Lifelines in recent decades have become indispensable. As a result, people will have difficulty if any of lifelines goes out of service. Ironically, more developed lifelines create more complicated systems and they sometimes cause more serious damage to urban life. Consequently, the vulnerability of these stratified networks is raising major concerns worldwide. For instance, the normal operation of water and telecommunications is maintained only if there is a steady supply of electricity. On the other hand, the generation and delivery of electric power cannot be ensured without the provision of fuel and gas. Therefore, securing the safety of one lifeline cannot be considered separately from all lifelines. Understanding vulnerability of the lifeline network is very much important to reduce damages against possible earthquakes. In the study, damage estimation on lifelines covered water, sewage, gas, electricity, and telecommunications. In this section, due to limited space, only gas pipeline and service box damage will be explained, since it treats the most hazardous material among the lifelines. 3.1. Existing Condition of Gas Distribution Network in Istanbul In Istanbul, Istanbul Gas Distribution Company (İGDAŞ), subsidiary company of Istanbul Metropolitan Municipality, was established in 1986 in order to start service of natural gas distribution. Within Istanbul

Metropolitan Area, the total length of pipeline network is 4,670 km. Pipe materials are Steel and Polyethylene, with share approximately 14% and 76% respectively.

Figure 9 Existing Distribution Network of Natural Gas

3.2. Evaluation Method Lifeline facilities are classified into two major categories: nodes and links. Nodes include facilities such as substations and purification plants. Links include facilities such as pipes or cables for supply and distribution purposes. A statistical approach for damage estimation of links (i.e., distribution pipes and lines) is applied in the study. Damage to node facilities is not estimated in this study, because such structures are different with respect to purpose and location and so a statistical approach is not applicable. Separate detailed surveys are required for the damage estimation of node facilities. 3.2.1. Pipeline Several researchers have proposed a correlation between pipeline damage and seismic parameters such as peak ground acceleration (PGA) or peak ground velocity (PGV). Based on these studies, PGV was selected as the seismic parameter to evaluate the damage of pipes in this study, because PGV showed a slightly better correlation compare to PGA. Figure 10 shows the damage function used in earthquake damage estimation study by the Disaster Prevention Council of the Tokyo Metropolitan Area [6] for welded steel gas pipes. This damage function was derived from the damage in Kobe City due to the 1995 Kobe Earthquake. Polyethylene pipes are treated as suffering no damage. The damage of gas pipes due to the 1999 Izmit Earthquake is reported in some papers. Tohma et al. [11] reported that there was no damage to gas distribution pipelines in the Avcılar area, which has polyethylene pipes, in spite of the heavy building damage. Kudo et al. [12] estimated the PGV in the Avcılar area during the Izmit Earthquake to be about 35 kine. O’Rourke et al. [13] reported the damage in Izmit City. There were 367 km middle density polyethylene (MDPE) pipes and 38 km steel pipes in Izmit City and no damage was found. There is a strong motion seismometer in Izmit and the record shows 40 kine in PGV, but the station is located at a stiff rock site, so the PGV in the city area might have been higher. Based on the

damage function by Disaster Prevention Council of the Tokyo Metropolitan Area [6], the damage to the pipeline in Izmit is estimated to be 0.14 points/km for steel pipes. This corresponds to the result of “no damage” in Izmit. If steel pipes would experience one break in Izmit, the damage ratio would become 0.026 points/km. Therefore “no damage” should be interpreted to be between 0.0 and 0.026 points/km from a statistical point of view. From the above consideration, the damage function by the Disaster Prevention Council of the Tokyo Metropolitan Area [6] is selected for use in the damage estimation in this analysis.

Buried G as Pipe dam age function - welded steel -

0

0.02

0.04

0.06

0 50 100 150

PG V (kine)

Damage ratio (points/km

)

Tokyo(1997)

Izm it (Izm it Eq.)

Figure 10 Relation between Damage Ratio of Welded Steel Gas Pipe and PGV

The damage function for Istanbul, based on Disaster Prevention Council of the Tokyo Metropolitan Area [6], is formulated as follows:

Rm(PGV) = R(PGV) x Cp x Cg x Cl

Where Rm(PGV): damage ratio (points/km) PGV: Peak Ground Velocity (kine = cm/sec) R(PGV) = 3.11 x 10-3 x (PGV-15)1.3 Cp: pipeline material coefficient 0.01 for Steel, 0.00 for Polyethylene Cg: ground condition coefficient 1.5 for Soft, 1.0 for Medium, 0.4 for Stiff Cl: liquefaction coefficient 2.0 for High, 1.0 for Low 3.2.2. Service Boxes Building Census Data conducted in 2000 includes information on natural gas installations. Based on this data, about 186,000 buildings (= 25.6%) have natural gas systems installed in Istanbul. The gas service box is installed on the ground floor of buildings or on the outer wall. If the building would collapse, the

gas box would also be damaged. Even if the gas pipeline is not damaged, gas leakage can occur from the service box, which may cause an explosion. In this study, it is assumed that all of the service boxes in heavily damaged buildings and half of those in moderately damaged buildings will be damaged. The following considerations support this assumption: According to O’Rourke et al. [12], there were 26,000 gas users in the city of Izmit before the Izmit Earthquake, and 860 service boxes were damaged. The mean number of housing units in one building in Izmit is assumed to be the same as in Istanbul-- namely, 4.2 housing units/building. Therefore, it is assumed that about 6,190 buildings have service boxes in them. Building damage estimates for Izmit are not available; therefore, the damage ratio in Izmit is assumed to be half of that of Gölcük and Değırmendere. Kabeyasawa et al. [9] reported 16% of buildings heavily damaged and 18% of buildings moderately damaged in these areas. According to these assumptions, it is estimated that 774 gas boxes were damaged in Izmit. 3.3. Result of Gas Pipeline and Service Box Damage Estimation The damage estimation definition is shown in Table 2.

Table 2 Definition of Gas Pipeline Damage Estimation Object Distribution, Service Pipes Service Box

Content of Damage Break of pipes or joints

Pull out of joints

Break of Boxes

Amount of Damage Number of damage points Number of damage points

The damage in each 500 m grid is calculated and illustrated in Figures 11 and 12. In numbers, only 13 points would be damaged; however, due to the great number of damaged buildings, it is calculated that 28,729 service boxes (= 16%) will be damaged. The damage of the gas pipeline system is slight. The main reason is that the gas pipeline in Istanbul was recently installed and İGDAŞ applied polyethylene pipes in the most of pipeline network, which have high flexibility and earthquake-resisting capacity, in accordance with the experience in past earthquake damage. However, the damage to service boxes amounts to over 25,000 because of the poor building structures. As a result, even though damage to network is very much limited, due to building collapse, service boxes will be collapse and serious gas leakage will occur in many places. Izmit Earthquake caused 860 Gas Service Box damages; however, explosions were not reported. Reason can be considered as 1) easy to stop operation due to limited size of the city, 2) number of damages were limited, and 3) did not occur explosions by chance. However, Istanbul as mega city, emergency response can not be taken into place promptly as small cities, due to delay of response, it can be estimated that many places will be filled with gas. Natural gas will not cause explosion if they are blowing outside, however, if closed area, such as inside building, is filled with gas, then caused by short-circuit from broken cables of electricity, and etc., explosion will occur. It is important to strike a note of warning against the danger. Therefore, measures are necessary to avoid gas leakage such as installation of central control system with automatic shut down function. For reference, in case of water distribution, total length of pipeline is 7,568 km, and number of damage points will be 1,577, which is a different result from that of gas pipeline damage. The main reason is the type of pipes used in water distribution network (such as Concrete, Steel, Ductile Iron, and Galvanized Iron pipes) which have rather less resistance against earthquakes.

Figure 11 Distribution of Gas Pipe Damage

Figure 12 Distribution of Gas Service Box Damage

4. CONCLUSION In the study, urban vulnerability in Istanbul for infrastructure and lifelines was evaluated. Understanding existing urban vulnerability will help minimize loss of unnecessary efforts and expenditures, because the vulnerable area is precisely defined from the results of evaluation, and so necessary efforts can be concentrated in a smaller area. The evaluation results should also be reflected in emergency operation manuals and regional disaster management plans to achieve the goal of disaster management. In other words, the combination of physical and administrative strengthening will create a disaster safe city. As a result of seismic resistant evaluation of lifelines, especially for gas distribution, it was defined that the network is resistant enough against possible earthquakes; however, critical service boxes have potential to trigger serious secondary disaster such as fires, and explosions. This indicates that just strengthening of lifeline facilities and pipelines is not necessarily sufficient for securing safety and that development of a total system down to the end supplier should be fully considered. Therefore, measures must be considered to avoid collapse of service boxes together with building strengthening efforts, and to have a systematic central control system with emergency response plan, which staffs are fully acquainted with. These conclusions and results are applicable for consideration of earthquake vulnerability improvement in other congested urban areas.

ACKNOWLEDGEMENT The authors had an opportunity to implement “The Study on A Disaster Prevention/ Mitigation Basic Plan in Istanbul including Seismic Microzonation in the Republic of Turkey” supported by Government of Japan as one of Development Study Program of Japan International Cooperation Agency (JICA). In the course of implementing the study, fruitful discussions and meetings were held with members of JICA Advisory Committee, Turkish Technical Advisory Committee, Istanbul Metropolitan Municipality Government, and many other related organizations, the outcome of which included constructive and informative comments and advices. Thus, it is necessary to indicate that the study could not be completed successfully, if it has not been for the sincere support of the above-mentioned persons and organizations, and it is the authors wish to express deep gratitude and appreciation to all persons involved in the completion of the study.

REFERENCES 1. JICA. “The Study on A Disaster Prevention / Mitigation Basic Plan in Istanbul including Seismic

Microzonation in the Republic of Turkey”, Final Report Volume Ⅱ Main Report , 2002 2. JSCE. “Report on the Hanshin-Awaji Earthquake”, 1998 3. Disaster Management Council of the Tokyo Metropolitan Area. “Report on Seismic Damage

Estimation in the Central Area of Tokyo”, 1978 4. J.Sato, Y.Sinozaki, M.Saeki, R.Isoyama. “Importance Factor Evaluation of Existing Road Bridges for

A Seismic Disaster Mitigation Plan”, Journal of Structural Mechanics And Earthquake Engineering, No.513/Ⅰ-31, JSCE, 1995

5. H.Kawakami. “Evaluation of Earthquake Performance of Transportation Systems”, Journal of Structural Engineering , No.327 , JSCE, 1982

6. Disaster Prevention Council of the Tokyo Metropolitan Area. “Report on the Damage Estimation in Tokyo by the Earthquake Right Under the Area”, 1997

7. Erdik, M. “Report on 1999 Kocaeli and Düzce (Turkey) Earthquake”, http://www/koeri.boun.edu.tr/depremmuh/kocaeloreport.pdf.

8. FEMA. “Earthquake Loss Estimation Methodology”, HAZUS99 Technical Manual, National Institute of Building Science, Washington, D. C., 1999

9. Kabeyasawa, T., K. Kusu and S. Kono edited. “Inventory Survey in Golcuk and Degirmendere, Report on the Damage Investigation of the 1999 Kocaeli Earthquake in Turkey”, Architectural Institute of Japan, Japan Society of Civil Engineers and The Japan Geotechnical Society, 2001: pp. 200-251.

10. Kawakami, H., S. Morichi and M. Yoshimine. “Damage to Civil Engineering Structures, Damage Report on 1992 Erzincan Earthquake, Turkey”, Joint Reconnaissance Team of Architectural Institute of Japan, Japan Society of Civil Engineers and Bogazici University, Istanbul, Turkey, 1993

11. Tohma, J., R. Isoyama, S. Tanaka and M. Miyajima. “Damage to Lifelines, Report on the Damage Investigation of the 1999 Kocaeli Earthquake in Turkey”, Architectural Institute of Japan, Japan Society of Civil Engineers and The Japan Geotechnical Society, 2001: pp. 194-199.

12. Kudo, K., T. Kanno, H. Okada, O. Özel, M. Erdik, T. Sasatani, S. Higashi, M. Takahashi and K. Yoshida. “Site Specific Issues for Strong Ground Motions during the Kocaeli, Turkey Earthquake of August 17, 1999, as Inferred from Array Observations of Microtremors and Aftershocks”, Bull. Seism. Soc. Am., 2002: Vol. 92, No. 1

13. O’Rourke, T. D., F. H. Erdogan, W. U. Savage, L. V. Lund and A. T. Manager. “Water, Gas, Electric Power, and Telecommunications Performance, Earthquake Spectra, Supplement A to Vol.16, 1999 Kocaeli, Turkey”, Earthquake Reconnaissance Report, 2000: pp. 377-402.

14. Toprak, S. “Earthquake Effects on Buried Lifeline Systems”, Doctoral Dissertation Presented to Cornel University, 1998