Development of a solution based on GIS and fuzzy logic to rescue earthquake victims

doi:10.21203/rs.3.rs-4354024/v1

Download PDF

Research Article

Development of a solution based on GIS and fuzzy logic to rescue earthquake victims

https://doi.org/10.21203/rs.3.rs-4354024/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Earthquake is one of the natural disasters that can cause significant loss of life and property. The unpredictable nature of earthquakes distinguishes them from the other natural disasters, making them a primary topic of crisis management. There are some extensive and carefully-planned measures to deal with this crisis, one of which is the management of rescue forces to rescue people trapped in collapsed buildings. Research has shown that the chance of trapped people surviving after 72 hours is almost zero. In this study, the survivability of people trapped in collapsed buildings was modeled with the time elapsed and the population density. The model of Coburn and Spence was used as the base method for estimating the survivability of trapped people. In this study, buildings were classified into eight categories. A fuzzy inference engine was then built with the time elapsed and building’s type as input, and survivability as its sole output. The data related to district 11 of Tehran was given to this fuzzy inference engine to generate survivability maps for 8, 24, 36, 48, and 72 hours after earthquakes with an intensity of 7 and 9 on the Mercalli scale. The results showed that in the both Mercalli 7 and 9, buildings in the northern part of the area rapidly collapse, trapping a large number of people, which leads to decrease in survivability in this area. However, survivability decreases with a nearly constant pattern in other parts. According, urban planners can adopt preventive measures and build rescue and relief sites in these areas before the expected earthquake.

earthquake

survivability

Coburn coefficients

fuzzy logic

During the twentieth century, about 75% of the earthquake casualties were caused by buildings destruction and people trapped in collapsed buildings (Coburn & Spence, 2003); if the secondary phenomena are excluded, this amount would be almost 90% of earthquake-related deaths (Coburn & Spence, 2003). Earthquakes cause a wide range of crises. One way to overcome this crisis is to rescue the injured and transport them to the nearest medical center. Indeed, delays in rescue efforts, absence of medical facilities, and secondary disasters can lead to unavoidable loss of life in an earthquake-stricken area. The existence and implementation of proper crisis management strategies play a key role in avoiding these problems. One of the main components of crisis management is the management of rescue operations. After an earthquake, crisis management squads need to properly allocate the available manpower, facilities and resources to the affected areas to minimize the number of casualties. The majority of earthquake-prone countries have a range of pre-determined plans to deal with such crises (Zhou & Liu, 2003). The lack of such plans in those countries has caused huge damage to even the smallest earthquake. Earthquake damage in countries such as Iran and Turkey proves this claim. The average earthquake damage in these two countries is much higher than the leading countries in earthquake management such as China and Japan (Wang, 2021).

In general, rapid implementation of rescue operations in an earthquake-stricken area can prevent many post-earthquake casualties. Research has shown that time plays a key role in post-earthquake mortality. Since most earthquake victims cannot survive more than 72 under the rubble, the first 72 hours after an earthquake is called the golden time (Godschalk et al., 1999). Therefore, developing an effective method for optimized allocation of victims to medical centers can greatly help reduce earthquake casualties. In this operation, issues such as the number of rescue squads and their distribution, the capacity of medical centers, and determining the fastest routes to transport the injured to medical centers should be considered. The rescue operation can be more effective by considering these issue. This operation is well carried out when there is a proper estimate of the people trapped in the rubble.

On the other hand, there needs to be a criterion for measuring the extent of earthquake damage to structures. Over the years, researchers have proposed two measures for this purpose: the magnitude of the earthquake, and the intensity of the earthquake (Walker et al., 2014). Given the definition of earthquake magnitude and intensity, the former is proportional to the amount of energy released in the earth, but the latter is related to the amount of damage caused by the seismic effect. In other words, an earthquake of a certain magnitude can have different intensities depending on its depths, and can even induce different intensities at different distances from the epicenter. The magnitude of an earthquake, denoted by M, indicates the total amount of energy released. Basically, the magnitude of an earthquake is measured by its seismic waves, which are divided into surface waves and volumetric waves (Elnashai & Di Sarno, 2008). Several scales have been proposed for measuring earthquake magnitude, the best known among which is the Richter scale. The intensity of an earthquake, which is perfectly proportional to its destructive power, can be measured by the Mercalli scale (Duzgun et al., 2011).

In general, research on the modeling of earthquake impacts with Geographic Information System (GIS) can be divided into two categories: 1- pre-earthquake zoning and hazard assessment, 2- post-earthquake survivability modeling. Among the studies of the first category, one can mention the research carried out by Zhang et al., who stressed the importance of the demographic characteristics of the disaster-stricken area. They developed a general model based on fuzzy multi-attribute decision-making that takes characteristics such as age and education into account. After presenting their method for assessing vulnerability against natural disasters based on demographic characteristics, they implemented the model for the city of Helsinki (Zhang et al., 2014). In another study, Aghamohammadi et al. (2013), developed a vulnerability assessment model based on the artificial neural network using the data collected in the aftermath of the Bam earthquake in Iran. In this study, the number of casualties was used in each urban area as the model input. After training the neural network with Tehran city data, they used the model to predict the casualty rate of a hypothetical earthquake in Tehran. One of the strengths of this model is the use of real seismic data, which increases the accuracy of the results (Aghamohammadi et al., 2013).

Hashemi and Alesheikh (2011), developed a GIS-based model for earthquake damage assessment and implemented it also in Tehran. In their research, the earthquake simulation was carried out based on the parameters of the Masha fault near Tehran. Damage estimation was modeled based on ground shaking, attenuation, local geological amplification, earthquake intensity, and fragility curves of buildings. Damage assessment was divided into two categories: casualty and street assessments. In assessing the casualties, criteria such as the time of the earthquake, population, the amount of damage to buildings, and the mortality ratio were considered. In assessing street damage, criteria such as street width and building materials were considered. The simulation showed that the assumed earthquake would severely damage 64% of the district’s buildings, kill 33% of all residents, injure 27% of the population, and block 22% of the streets (Hashemi & Alesheikh, 2011).

Karimzadeh et al. (2014), also developed a model for the vulnerability assessment of buildings and people in the city of Tabriz (Iran) and implemented it in an earthquake simulation scenario. In this study, the maps of effective parameters including geological, geodetic, geotechnical, and geophysical factors were prepared in the GIS environment. Weights were then determined for the maps using the Analytical Hierarchy Process (AHP). In this study, a magnitude 7 earthquake in the fault north of Tabriz was used for simulation. The developed model was used to assess the vulnerability of buildings, people, and infrastructure of the city in two main areas. The results of this study showed if the assumed earthquake occurred at night, it would completely destroy 69.5% of the buildings and kill 33% of residents (Karimzadeh et al., 2014).

Ranjbar and Nekooei (2018) used the fuzzy TOPSIS method to identify the buildings vulnerable to earthquake damage in two scenarios day and night in Tehran city. Buildings and roads were the main criteria of this research. The results of the simulation showed that the Tehran Grand Bazaar has the highest level of vulnerability and the lowest level of accessibility for rescue operation (Ranjbar & Nekooie, 2018).

Kamranzad et al. (2020), assessed the earthquake hazard in Tehran. In this study, the effects of criteria such as geology, active faults, population distribution, land use, height and type of buildings were evaluated on the earthquake using GIS. The hazard, exposure and vulnerability maps were then obtained based on these criteria in terms of Peak Ground Acceleration (PGA) (Kamranzad et al., 2020).

Sen and Ekinci (2016), analyzed the level of post-earthquake vulnerability in Istanbul. In this study, fault line, geological structure, building structure, and the number of building floors were considered in post-earthquake vulnerability assessment. The building damage distribution map was then obtained by combining fuzzy logic and GIS (Şen & Ekinci, 2016).

Meshkini et al. (2013), used 11 spatial and physical criterion to assess the potential earthquake damage in the city of Zanjan (Iran) using fuzzy logic Inverse Hierarchy Process Weight (IHPW) (Meshkini et al., 2013). Fallah-zazuli et al. (2019) used GIS and Remote Sensing (RS) techniques to analyze explicit models of landslide in the Zagros Mountains, Iran. Logistic Regression (LR) and Multi-Criteria Decision Analysis (MCDM) were used to estimate the likelihood of landslide occurrences (Fallah-Zazuli et al., 2019). Mesbahi et al. (2020) prepared an earthquake vulnerability zoning map for the city of Tabriz with 10 environmental-geological criterion using fuzzy logic. The results showed a high level of risk in the northern and central parts of Tabriz, which have a high population density (Mesbahi et al., 2020).

Modeling of resource allocation and survivability in crisis management has also been the subject of several studies, like the one conducted by Dodo et al. in 2005 (Dodo et al., 2005). In another study, Aghamohammadi et al. (2013) focused on resource allocation after earthquake. They developed a heuristic method based on two nested genetic algorithms for allocating treatment centers to minimize post-earthquake relief time. In their method, the outer genetic algorithm is used to solve the location problem and the inner one is used to optimize the location-allocation of medical centers. The results showed that new places are needed for medical centers. The results also showed the good performance of the model in allocating the capacity of medical centers to individuals (Aghamohammadi et al., 2013).

Ren et al. (2016), studied the emergency planning to improve post-earthquake rescue operations. They selected 30 spatial and non-spatial criterion for evaluation of earthquake relief plans. They then used the Hesitant Analytic Hierarchy Process (HAHP) method and fuzzy logic to develop the earthquake. The results showed that HAHP is much more effective than AHP in terms of the effectiveness of earthquake rescue planning (Ren et al., 2016).

Ranjbar et al. (2017), examined the buildings damaged caused by an earthquake by using satellite images to determine the number of victims. In this study, criteria such as earthquake time, building materials, land-use, buildings age, and rough texture of the images were used. The introduced model reasonably predicted the number of casualties (Ranjbar et al., 2017).

Liu (2020), developed a new GIS-based model for quick rescue of affected people by an earthquake. This system seeks to reduce the cost and distance of transferring the injured to medical centers. In this model, seismic susceptibility map was prepared based on slope, altitude, lithology, fault, river, and road data. The result demonstrated the ability of the hierarchical model to reduce rescue time as well as casualties after the earthquake (Liu, 2020).

Yariyan, et al. (2020) developed a new hybrid learning model for seismic vulnerability mapping in Sanandaj City based on 15 factors (Yariyan et al., 2020). In addition to all these researches, various other studies were taken on resource allocation for optimum survivability in earthquake.

Various methods are used to estimate the number of injuries and casualties, including the methods developed by the US Federal Emergency Management Agency (FEMA), Shiono Coburn and Spence. The FEMA model provides a methodology for estimating the number of casualties due to the collapse of buildings and bridges, which is based on information about earthquakes and structures within the US (Molina et al., 2010). The models of Murakami and Shiono only estimate the number of casualties based on building failures (Shiono et al., 1991). The Coburn model is a fundamentally statistical method based on information about the number of casualties caused by the collapse of buildings in major earthquakes in different countries. This method can clearly estimate the casualties from the number of collapsed building and lethality ratio. Lethality ratio depends on different parameters, most of which are mechanism of collapse, building type, the time of earthquake occurrence, and rescue operations effectiveness (Coburn & Spence, 2003). Various studies have shown that the Coburn model can provide a relatively accurate estimate of casualties and injuries.

Japan International Cooperation Agency (JICA) has conducted a study to estimate the number of casualties in a potential earthquake scenario based on the Coburn method (JICA, 2000). However, the information and analyses need to be updated regularly to have a better rescue operation. This information includes: number of victims, number of rescue squads, capacity of medical centers, fastest routes to send the injured to medical centers, roads blocked by earthquake rubble and speed on the roads.

Relief for earthquake victims is very important in post-earthquake management, which has been mentioned in previous research. The most important difference between the present study and similar ones is the timeliness of the model. The studies that have been done estimate the casualties in general, while in this study, the survival expectancy of the injured is calculated in 72 hours after the earthquake for different structures according to the fuzzy function. This model provides the possibility of appropriate relief according to the chances of the injured being alive. On the other hand, studies conducted in Iran and neighboring countries are based on models used in the United States and East Asia, which are not compatible with the type materials and earthquake in these areas.

In this study, the survival rate of affected people is estimated using spatial analysis methods be optimally allocated to medical centers. Since the nature of the factors influencing the estimation of survivability is vague and uncertain, fuzzy logic is used for modeling. Therefore, this study provides a model for estimating the survivability of victims by combining the spatial information systems with fuzzy logic.

Like most studies, this research was carried out in two phases, one phase focused on the fundamentals of the subject and a second phase dedicated to model development and implementation. In the first phase, it was necessary to review the principles of earthquake management, earthquake scenarios and simulation, rescue and relief operations, and previous works in this field to develop a suitable model for survivability prediction.

2.1 Conceptual model

Accurate estimation of the number of earthquake victims and their survival is the most important goal of this study. This goal is achieved by the combination of spatial analysis and fuzzy logic methods. The flowchart of the research process is illustrated in Fig. 1. At first, the required data was collected from the study area and prepared to be used in Coburn tables and fuzzy functions for estimating the number of injured in different time periods. The outputs of this process provide a clear image of the locations where injured people are more concentrated in different time periods. This allows the rescue squads to give higher priority to the areas with a higher number of victims, and plan the rest of the rescue operation accordingly.

Figure 1 Flowchart of the developed model

2.2 Quantification of the impact of earthquakes on buildings

There are two methods for quantifying the impact of an earthquake. One method is to consider the effects of an earthquake on a specific site and location (earthquake intensity). The other one is to measure the characteristics of the earthquake at its source (magnitude of the earthquake). While the intensity of an earthquake varies from place to place and decreases with distance from the epicenter, the magnitude is always constant and does not change with distance from the epicenter and it is related to the total energy released during the seismic event (JICA, 2000). Given the direct relationship between the Mercalli scale and the collapse of structures and other features on the ground, this study used this particular earthquake intensity scale. Mercalli or seismicity scale is an observational and non-instrumental measure of earthquake damage that depends on the distance from the epicenter, the duration of the earthquake, the type of soil, and the depth of the bedrock. Basically, the intensity of an earthquake in a given place is determined by the observable effects of that earthquake in that place.

The accuracy of earthquake intensity estimations depends on the accuracy of the observer. Earthquake intensity estimation is a useful way of estimating the magnitude of historical earthquakes, especially in places like Iran, where there are enough ancient and historical buildings for extracting such information. To determine the intensity of an earthquake, an observer needs to rate the extent of damage caused by the earthquake in different structures based on the descriptions given in a series of tables developed for this purpose. The Mercalli earthquake intensity scale, originally defined in 1902, was between 1 and 10, but the modified version developed in 1932 is between 1 and 12 (JICA, 2000). The relationship between peak ground acceleration and earthquake intensity is shown Fig. 2.

Mercalli scale is used to measure the severity of earthquakes in terms of the structure collapse in the absence of seismic stations, which means it can be used to estimate the severity of historical earthquakes. The earthquake intensity data can also be used to draw earthquake intensity contour lines. One problem with the Mercalli scale is that it can be affected by unrealistic reports and exaggerations in earthquake descriptions. This method cannot accurately estimate the magnitude of an earthquake, and is only applicable to residential areas. To determine the intensity of an earthquake with the modified Mercalli scale, structures must be divided into four types based on their strength and build quality (Table 1) (JICA, 2000).

Table 1

Types of structure in terms of strength
Type	Description of structural strength
(A)	good design; good components; good combination of concrete, steel beams, and mortar; designed to withstand lateral seismic forces
(B)	good combination of concrete, steel beams, and mortar; designed by an expert but not to withstand lateral seismic forces
(C)	Ordinary design and components; cement is used in the structure but it is not resistant to lateral forces
(D)	Constructed without design; no cement or metal is used in the structure and it is resistant to lateral forces

Mercalli (earthquake intensity) scale is an observational and non-instrumental scale determined by expert knowledge based on different rules defined for measurements. In this study, the Mercalli scales 7 and 9 are used to assess survivability. Mercalli 7, for example, is characterized by terrified people fleeing to open spaces, very little damage to well-designed buildings, minor to medium damage to ordinary buildings, significant damage to poorly designed buildings, cracking and falling of plasters and loose bricks in type D buildings, appearance of cracks in type C buildings, difficulty to stand, breaking of some furniture, ringing of large bells, damage to cement drains, and strong shaking. Mercalli 9 however is characterized by significant damage to specially designed buildings, tilting of well-designed frame structures, shifting of buildings off their foundations, appearance of cracks in the ground, breaking of underground pipelines, public panic, destruction of type D buildings, severely damaged and sometimes completely collapsed type C buildings, severely damaged Type B buildings with major damage to the foundation, and discharge of sand and mud in alluvial areas.

2.3 Required data

At this stage, the required criteria such as the type of buildings and roads are collected from official organizations such as Tehran Municipality, Iran Crisis Management Organization and Geological Survey of Iran using the opinions of experts. The details of this data are discussed as follow:

2.3.1 Population

This index represents the size of the population that could be affected in the event of an earthquake. There is a close relationship between population density and earthquake casualties. Sheltering could be more difficult and rescue operations tend to be slower in higher population density. To emphasize the importance of this parameter, it should be noted that a high-intensity earthquake is likely to cause higher losses and damages in areas with low population density even if they consist of older structures. To incorporate this parameter, the study used the map of the population density of the region.

2.3.2 Type of structure

The type of structures in a region is one of the most important determinants of the extent of earthquake damage. Over the years, there have been many developments in the field of structural safety, especially against earthquakes. But older structures tend to have lower earthquake resistance than newer structures. For many centuries, Iranians have extensively used masonry materials and bricks in construction. Given the high cost of demolition and renovation of these buildings and the lack of a proper plan for this purpose, there are still many of these buildings in Iranian cities. Therefore, it is not uncommon to see a lot of diversity in the type of structures in different parts of a city, and this issue needs to be taken into account. Since there have been many studies on the impact of materials and construction methods and quality on earthquake resistance, it is not difficult to estimate the strength of different types of structures.

2.4 Estimation of structural damages and human losses

Given the limited number of rescue squads and equipment available immediately after an earthquake, allocating them to the affected areas can be vital. Therefore, the number of people killed and trapped in buildings in different areas needs to be estimated so that the squads can be sent with the highest concentration of trapped people. One of the methods developed for this estimation is the Coburn method. Using this method, the number of people trapped in collapsed buildings, the number of casualties due to the collapse of buildings, the mortality rate after the collapse can be estimated based on the type of rescue operation, the magnitude of the earthquake and the type of buildings.

In the next stage, fuzzy logic and spatial analysis are used to develop a model for estimating survivability based on the earthquake magnitude, the type of structures, and the time elapsed since the event. In this stage, the information related to earthquake magnitude (according to the defined scenario), roads, structures, and location of relief centers was collected. Having the survivability results and the information about road obstruction, speed on different roads, and location of relief centers, a model was then developed for optimal allocation of the injured to centers. This model is run every 4 or 8 hours and updated at each stage to decide which victims have a higher survivability rate.

Earthquake resistance characteristics of buildings tend to vary in each area. The relationship between earthquake intensity and fatality (damage) ratio is not always the same, even if the buildings are similar in appearance. These differences can be attributed to differences in construction methods. Therefore, one of the most important stages of damage assessment is to collect the records of multiple catastrophic earthquakes and extract the damage function based on local experiences.

In 1999, the seismic zoning map of the Tehran area was developed at the request of the Iranian government in collaboration with the Japanese government. The present work used the tables prepared in that study based on the characteristics, structural conditions, and statistical data of the city of Tehran. In this study, the Coburn table is used (Tables 2 and 3). The values of M3, M4d, and M5 are extracted from these tables. In this method, M1 to M5 are respectively the population per building, occupancy at the time of earthquake, the percentage of occupants trapped at collapse (which depends on the type of structure), the percentage of casualties at collapse (which also depends on the type of structure), and the mortality rate post-collapse (which depends on the rescue operation) (Fig. 3). The estimation of casualties. Each of these parameters is defined as:

Ks: Fatalities

D5: Number of collapsed buildings

M1: Number of people in each building

M2: Percentage of occupants present in the building at the time of earthquake

M3: A coefficient for determining the number of occupants trapped during the collapse, which depends on the type of structure

M4d: A coefficient for determining the number of occupants who die during or immediately after the collapse, which depends on the type of structure

M5: A coefficient for determining the number of people who die while (due to) being trapped, which depends on the rescue operation

Coburn coefficients M3, M4 and M5 should be adjusted to meet the relationship between earthquake intensity and mortality due to structural damage in previous earthquakes in Iran. In this study, the information used for this purpose include: the type of structure, the age of structure (year of construction), and the number of floors. Based on this information, buildings are divided into eight categories: adobe, brick and cement block, all-wood, type-2 reinforced concrete, type-1 reinforced concrete, brick-steel, type-2 steel, type-1 steel, and type-0 reinforced concrete.

2.4.1 Coefficients M1 and M2

M1 represents the number of people living in each building, and may vary from city to city as it is determined by the average household size. This parameter is higher in cities that experience rapid population growth or suffer from housing shortages.

Earthquake time (M2) is another determining factor in the number of casualties. Casualties will be different depending on the population concentrated in residential buildings or office buildings. For example, if an earthquake occurs when most people are inside the buildings, the casualties will be very high.

2.4.2 Coefficient M3 and related table

It is necessary to have statistical data from past earthquakes to determine the coefficient M3. Obviously, all occupants of a building will not be trapped once it collapses. Some occupants can flee before the collapse, and some may even escape the collapsed building on their own. The number of trapped people also depend on the type of structure, its strength, and the number of floors. Statistics show that people on the upper floors have almost no chance of escaping the building during an earthquake. The percentage of people who cannot escape the collapsing buildings needs to be estimated based on the number of people present in the building at the time of earthquake. Coefficient M3 is obtained based on earthquake intensity and the type of structure, which is prepared by JICA (Table 2).

Table 2

Number of people trapped in collapsed buildings (M3)
Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	0	0	0	0	0	0	0	0	0
4	0	0	0	0	0	0	0	0	0
5	0	0	0	0	0	0	0	0	0
6	0	0	0	0	0	0	0	0	0
7	0.5%	1%	1%	1%	1%	1%	2%	3%	5%
8	1%	2%	4%	4%	4%	4%	10%	15%	30%
9	3%	8%	12%	15%	15%	15%	50%	70%	60%
10	8%	20%	50%	60%	60%	60%	80%	90%	70%
11	15%	30%	65%	75%	75%	75%	90%	97%	80%
12	20%	35%	70%	80%	80%	80%	95%	100%	90%

2.4.3 Coefficient M4d and related table

People who are trapped in collapsed buildings can be injured in various ways. The injuries usually include wounds, dislocations, cracked or fractured bones, rupture of soft tissues, damage to the chest, and respiratory problems. The type of injuries highly depends on the type of building. Buildings made of heavy materials generally inflict heavier injuries on residents. Therefore, the type of building and the intensity of earthquake should be taken into account. M4 represents the number of casualties immediately after the collapse of the building. To calculate this coefficient, it is assumed that some of the people trapped in collapsed buildings would die immediately by collapsing floors or suffocate by the fumes of collapsed adobe and brick materials (Table 3).

Table 3

Mortality at the zero hour after building collapse (M4d)
Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	5%	30%	50%	30%	40%	40%	60%	80%	20%
4	5%	30%	50%	30%	40%	40%	60%	80%	20%
5	5%	30%	50%	30%	40%	40%	60%	80%	20%
6	5%	30%	50%	30%	40%	40%	60%	80%	20%
7	5%	30%	50%	30%	40%	40%	60%	80%	20%
8	5%	30%	50%	30%	40%	40%	60%	80%	20%
9	5%	30%	50%	30%	40%	40%	60%	80%	20%
10	5%	30%	50%	30%	40%	40%	60%	80%	20%
11	5%	30%	50%	30%	40%	40%	60%	80%	20%
12	5%	30%	50%	30%	40%	40%	60%	80%	20%

2.4.4 Coefficient M5 and related table

The speed of rescue operations is also important. Trapped people with serious injuries tend to die if left untreated. People with fewer injuries survive a little longer, and have a better chance of survival. Therefore, time is of the essence, and many survivors can be rescued if rescue squads act quickly. The success of rescue squads largely depends on the time taken for them to act after the earthquake. The survivability ratio reaches zero approximately 72 hours after the event. This means that most of the trapped people will die if not rescued within 72 hours. Therefore, the speed of rescue operation is the main determinant of the proportion of trapped people who do not die immediately at collapse but die afterward. Relief is the main and determining factor in the amount of casualties due to the collapse of buildings

Table 4

Post-collapse mortality, without rescue (M5a)
Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	95%	95%	95%	95%	95%	95%	95%	95%	95%
4	95%	95%	95%	95%	95%	95%	95%	95%	95%
5	95%	95%	95%	95%	95%	95%	95%	95%	95%
6	95%	95%	95%	95%	95%	95%	95%	95%	95%
7	95%	95%	95%	95%	95%	95%	95%	95%	95%
8	95%	95%	95%	95%	95%	95%	95%	95%	95%
9	95%	95%	95%	95%	95%	95%	95%	95%	95%
10	95%	95%	95%	95%	95%	95%	95%	95%	95%
11	95%	95%	95%	95%	95%	95%	95%	95%	95%
12	95%	95%	95%	95%	95%	95%	95%	95%	95%

after the earthquake. Earthquake mortality rates rise over time. This variable is applied by a M5 coefficient. The coefficient M5a is used for calculating the losses after the collapse of the building without any relief (Table 4). The M5b coefficient is used for calculating the losses after the collapse of a building in the case of public relief only (Table 5). The M5c coefficient is the loss in case of public relief as well as relief groups (Table 6). The M5d coefficient is used to calculate the losses after the collapse of a building in the event of public assistance with relief teams and a group of professional rescue experts (Table 7).

Table 5

Post-collapse mortality, with rescue squads (M5b)
Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	80%	70%	60%	70%	60%	40%	60%	70%	60%
4	80%	70%	60%	70%	60%	40%	60%	70%	60%
5	80%	70%	60%	70%	60%	40%	60%	70%	60%
6	80%	70%	60%	70%	60%	40%	60%	70%	60%
7	80%	70%	60%	70%	60%	40%	60%	70%	60%
8	80%	70%	60%	70%	60%	40%	60%	70%	60%
9	80%	70%	60%	70%	60%	40%	60%	70%	60%
10	80%	70%	60%	70%	60%	40%	60%	70%	60%
11	80%	70%	60%	70%	60%	40%	60%	70%	60%
12	80%	70%	60%	70%	60%	40%	60%	70%	60%

Table 6

Post-collapse mortality, with community rescue + rescue squads (M5c)
Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	35%	50%	50%	50%	50%	30%	50%	60%	50%
4	35%	50%	50%	50%	50%	30%	50%	60%	50%
5	35%	50%	50%	50%	50%	30%	50%	60%	50%
6	35%	50%	50%	50%	50%	30%	50%	60%	50%
7	35%	50%	50%	50%	50%	30%	50%	60%	50%
8	35%	50%	50%	50%	50%	30%	50%	60%	50%
9	35%	50%	50%	50%	50%	30%	50%	60%	50%
10	35%	50%	50%	50%	50%	30%	50%	60%	50%
11	35%	50%	50%	50%	50%	30%	50%	60%	50%
12	35%	50%	50%	50%	50%	30%	50%	60%	50%

Table 7

Post-collapse mortality, community rescue + rescue squads + experts (M5d)
Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	15%	30%	45%	35%	45%	25%	45%	55%	45%
4	15%	30%	45%	35%	45%	25%	45%	55%	45%
5	15%	30%	45%	35%	45%	25%	45%	55%	45%
6	15%	30%	45%	35%	45%	25%	45%	55%	45%
7	15%	30%	45%	35%	45%	25%	45%	55%	45%
8	15%	30%	45%	35%	45%	25%	45%	55%	45%
9	15%	30%	45%	35%	45%	25%	45%	55%	45%
10	15%	30%	45%	35%	45%	25%	45%	55%	45%
11	15%	30%	45%	35%	45%	25%	45%	55%	45%
12	15%	30%	45%	35%	45%	25%	45%	55%	45%

2.5 Fuzzification

Since the probability of survival drops significantly after the first 72 hours, the first hours after the earthquake are very important. These hours are commonly referred to as golden hours of post-earthquake rescue operations. In this study, fuzzy logic (Jalilzadeh & Behzadi, 2020; Zadeh et al., 1996) was used to model the effect of time on survivability.

Processes modeled in the real world are often inaccurate, and reality cannot be modeled with uncertainty as it is. There are also limitations to modeling uncertainty. Accordingly, fuzzy logic is used to solve complex problems with more effective solutions (Behzadi et al., 2019; Zhou et al., 2020). Basically, fuzzification transforms human logic into a mathematical formulation. Given the nonlinear behavior of earthquakes, it could be helpful to use fuzzy logic in the modeling of earthquake-related phenomena.

Since the first 72 hours after an earthquake are so important, this time interval is modeled based on fuzzy logic. For fuzzification, this time interval is divided into 5 classes: very low (VL), low (L), medium (M), high (H), and very high (VH) (Fig. 4). Each class determines the degree of membership of the person under the rubble at a particular time.

The purpose of fuzzification is to calculate the odds of survival of people trapped in collapsed buildings over time. In other words, the more time elapses from the event, the less likely people are surviving. It means that there is an inverse relationship between passing of time (up to 72 hours after event) and the probability of survival. Since the odds of survival vary depending on the type of structure and materials, each type of structure must have its own fuzzy function. For example, if a person is trapped under a concrete building, the odds of survival will be low because the material will be too heavy to remove and the injuries are also likely to be more serious. People injured in the collapse of this type of structure tend to die shortly after the incident. In contrast, the odds of survival and getting rescued will be higher if the structure is all-wooden. Therefore, different survivability functions are defined for different types of structures.

After the fuzzy inference procedure, defuzzification is performed to produce a survivability rate, which is the odds of survival given the time elapsed since the earthquake. This rate is multiplied by the values obtained from the Coburn method to obtain the number of people who survive under the rubble over different time periods. After performing the calculations, the maps related to different time periods are drawn.

3.1 Study area

This study is focused on the city of Tehran with 22 districts. This city is selected because of the presence of many critical infrastructures, the large number of people living in this city, and the existence of several faults in the region. Tehran has not experienced an earthquake for many years, and since this period has been longer than the return period of earthquakes in the region, many experts are concerned that the city is likely to experience a strong earthquake in the near future. Given the faults around Tehran, the return period of an earthquake with a magnitude of seven or more in this region is about 150 years (Fariborz, 2001). Since none of these faults have been active for about 150 years, there is a high probability that a devastating earthquake occurs in the coming years. District 11 of Tehran is selected for implementing the vulnerability assessment model (Fig. 5).

Figure 5 Greater Tehran area and district 11

3.2 Data of the study area

This section presents the data related to district 11 and explains the process of preparing these data.

3.2.1 Population

Population is the first data set of the study area. This data set is extracted from the database of the latest census (2016), which is available in the form of population blocks (Figurer 6).

3.2.2 Classification of buildings

The building data used in this study are the type of structure, the year of construction, and the number of floors. Based on this data, the buildings in the study area are divided into eight categories (Fig. 7). After calculating the population of each structure, the coefficient M1 was obtained. Then, tables related to M3 are used to estimate the percentage of people trapped in buildings, and the result is multiplied by M1 to determine the number of trapped people. In this study, M3 is calculated for earthquake intensities of 7 and 9 on the Mercalli scale. The coefficient M4d is also calculated for the same earthquake intensities.

3.3 Fuzzification and construction of fuzzy inference engine

In the fuzzification phase, five membership functions were defined for 8, 24, 36, 48, and 72 hours after the earthquake. The fuzzy membership function for the building type is also constructed based on the eight building types (Fig. 8). Finally, the fuzzy membership function for the survivability after the earthquake is modeled with five classes: very low, low, medium, high, and very high (Fig. 8). The building type and population data are fed to the engine as a raster. Then, the information of each parcel is placed in the relevant membership functions and the survivability output is calculated for two earthquake intensities and five time states.

3.4 Survivability results

The survivability output for each building is calculated for two different earthquake intensities on the Mercalli scale and five different time states. The results are then converted into a map. The green areas of the output map indicate the areas where there is still hope for survival. Over the time, some of these people lose their lives, which is reflected in the declining survivability. Thus, progressively larger parts of the district turn red as time passes. Between Mercalli 7 and 9 earthquakes, the latter is more severe, which leads to more collapse of buildings; therefore, more people getting trapped under rubble. As a result, the map of the Mercalli 9 earthquake has more areas with green color.

In the Mercalli 7 earthquake, the buildings in the northern part of the region rapidly collapse, trapping a large number of people, which results in declining survivability in this area. However, survivability decreases with a nearly constant pattern in the other parts. It can be concluded that planners and decision-makers need to pay more attention to crisis management in this part of the district. This increased attention can materialize as creating plans for establishing rescue facilities in this part based on the provided information or to implement seismic rehabilitation programs in this area. Figure 9 also shows the state of the district at 8 and 36 hours after the considered Mercalli 7 earthquake.

For better illustration of the changes, the state of the district at 24 and 72 hours after the Mercalli 7 earthquake is also shown in Fig. 10. Similar patterns can be observed in the results obtained for the Mercalli 9 earthquake. In this scenario, the northern parts of the region showed rapidly declining survivability. For comparison, the state of the district 8 and 36 hours after the Mercalli 9 earthquake is shown Fig. 11. As seen, the said pattern starts from the north of the region in the right figure (8 hours after the Mercalli 9 earthquake) and spread to the northwestern parts after 36 hours. Similar changes can be observed in the central parts at the same rate. The changing color in other areas also indicates a similar pattern.

Based on this trend of change, urban planners and decision-makers can find out which areas will experience a larger decline in survivability and plan the rescue process accordingly. The state of the district 24 and 72 hours after the Mercalli 9 earthquake is illustrated in Fig. 12. As seen, the pattern intensifies after 72 hours and survivability approaches 0 after this time. With a deeper analysis of figures, urban planners can also use the results to adopt preventive measures and build rescue sites in these areas before the expected earthquake. In other words, the results can be used for the site location of Red Crescent buildings, clinics, emergency rooms, and etc. for crisis situations. It may also be worthwhile to compare survivability at the same time instances after Mercalli 7 and Mercalli 9 earthquakes. The level of survivability in the study area 36 hours after the considered Mercalli 7 and Mercalli 9 earthquakes is shown in Fig. 13.

This study attempted to use fuzzy logic for more accurate and realistic modeling of survivability in collapsed buildings after earthquakes. In several similar studies, the number of victims has been estimated by the use of statistical methods exclusively based on previous information on the number of collapsed buildings. This study is compared to the one conducted by Mohanaselvi and Hemapriya in 2019. In this paper, an earthquake damage predicting was done using fuzzy logic techniques. The fuzzy logic toolbox in Matlab was used to implement the proposed fuzzy logic control system for predicting. In this research four variables was used for building damages such as: ground motion duration, types of soil, height of the building, and the maintenance (Mohanaselvi & Hemapriya, 2019). Compared to this article, in the present study the nine variables of building types are used for considering earthquake damages and outputs are more tangible and closer to predicts.

In another paper, Sadrykia et al (2017) developed a GIS-based model using a modified integration of AHP, fuzzy sets theory and TOPSIS for considering earthquake damages. In this article, six geological and physical factors were used to consider vulnerability of buildings. The applicability of the proposed method was tested in a district of Tabriz, which is in a near vicinity to the fault system (Sadrykia et al., 2017). According to results, just physical-spatial factors of building were used and population factors were not considered while in the present study, the fuzzified versions of two criteria, namely population and structure type were used to produce a more accurate estimate of survivability after the earthquake.

The Coburn and Spence method is the most well-known method based on Mercalli for estimating injuries, the coefficients of which must be adjusted for each country (Coburn & Spence, 2003). Mansouri et al. (2010) used the Hazus Earthquake Model, a general model provided by the Federal Emergency Management Agency (FEMA), to estimate the structures destruction and casualties (Mansouri et al., 2010). In their research, the coefficients of the method were adjusted according to the degree of destruction. While in the present study, the Coburn and Spence method was used. The coefficients of this model were adjusted according to previous earthquakes that occurred in Iran using JICA (JICA, 2000). Since the coefficients of the present research model are related to the study area rather the general ones, it gives us better accuracy in estimating the destruction and casualties.

In a similar study conducted by Aghamohammadi et al. (2013) for the city of Bam in Iran, the neural network method was used. Neural networks are good for modeling complex problems. However, it extremely depends on pre- and post-earthquake data. Such a model will not be applied to areas where an earthquake has not yet occurred. While the proposed model in this study is independent of post-earthquake data that can be used for any region.

One of the most important advantages of the present study is the timeliness of the model. In previous studies, the issue of time has not been considered, while in this study, not only the estimates of the injured but also the life expectancy of people under rubble is estimated in 0 to 72 hours after the earthquake.

In this article, buildings were theoretically divided into eight general categories, while it might be difficult to place a building in exactly one category in reality. Therefore, it is necessary to develop a more comprehensive model for the classification of buildings.

Future studies can also expand this research by adding the problem of optimal allocation of rescued people to medical centers using methods such as genetic algorithm, bee algorithm, or ant colony algorithm to solve this problem. It may also be possible to use the agent-based modeling approach, i.e. define a series of rules for agents (e.g. rescue squads, injured people, etc.), and simulate the rescue process accordingly to predict the potential outcomes of the crisis scenarios.

Competing interests

The authors declare no competing interests.

Funding

Not applicable. No funding

Aghamohammadi, H., Mesgari, M., Mansourian, A., & Molaei, D. (2013). Seismic human loss estimation for an earthquake disaster using neural network. International Journal of Environmental Science and Technology, 10(5), 931-939.
Aghamohammadi, H., Mesgari, M. S., Molaei, D., (2013). Development a heuristic method to locate and allocate the medical centers to minimize the earthquake relief operation time. Iranian journal of public health, 42(1), 63.
Coburn, A., & Spence, R. (2003). Earthquake protection. John Wiley & Sons.
Dodo, A., Xu, N., Davidson, R. A., & Nozick, L. K. (2005). Optimizing regional earthquake mitigation investment strategies. Earthquake Spectra, 21(2), 305-327.
Duzgun, H., Yucemen, M., Kalaycioglu, H., Celik, K., Kemec, S., Ertugay, K., & Deniz, A. (2011). An integrated earthquake vulnerability assessment framework for urban areas. Natural Hazards, 59(2), 917-947.
Elnashai, A. S., & Di Sarno, L. (2008). Fundamentals of earthquake engineering. Wiley New York.
Fallah-Zazuli, M., Vafaeinejad, A., Alesheykh, A. A., Modiri, M., & Aghamohammadi, H. (2019). Mapping landslide susceptibility in the Zagros Mountains, Iran: a comparative study of different data mining models. Earth Science Informatics, 12(4), 615-628.
Godschalk, D., Bohl, C. C., Beatley, T., Berke, P., Brower, D., & Kaiser, E. J. (1999). Natural hazard mitigation: Recasting disaster policy and planning. Island press.
Hashemi, M., & Alesheikh, A. A. (2011). A GIS-based earthquake damage assessment and settlement methodology. Soil Dynamics and Earthquake Engineering, 31(11), 1607-1617.
Jalilzadeh, A., & Behzadi, S. (2020). Flood Mapping and Estimation of Flood Water-Level Using Fuzzy Method and Remote Sensing Imagery (Case Study: Golestan Province, Iran). Forum Geografic,
JICA, C. (2000). The study on seismic microzoning of the Greater Tehran Area in the Islamic Republic of Iran. Pacific Consultants International Report, OYO Cooperation, Japan, 291-390.
Kamranzad, F., Memarian, H., & Zare, M. (2020). Earthquake risk assessment for Tehran, Iran. ISPRS International Journal of Geo-Information, 9(7), 430.
Karimzadeh, S., Miyajima, M., Hassanzadeh, R., Amiraslanzadeh, R., & Kamel, B. (2014). A GIS-based seismic hazard, building vulnerability and human loss assessment for the earthquake scenario in Tabriz. Soil Dynamics and Earthquake Engineering, 66, 263-280.
Liu, K. (2020). Post-earthquake medical evacuation system design based on hierarchical multi-objective optimization model: An earthquake case study. International Journal of Disaster Risk Reduction, 51, 101785.
Mesbahi, F., Akbari Baghi, M., & Nadiri, A. (2020). Assessment and mapping of the seismic vulnerability of Tabriz city using the Fuzzy logic. J Adv Environ Health Res, 8(3), 181-192.
Meshkini, A., Habibi, K., & Alizadeh, H. (2013). Using fuzzy logic and GIS tools for seismic vulnerability of old fabric in Iranian cities (Case study: Zanjan city). Journal of Intelligent & Fuzzy Systems, 25(4), 965-975.
Mohanaselvi, S., & Hemapriya, G. (2019). Application of fuzzy logic to earthquake damage predictions. AIP Conference Proceedings,
Molina, S., Lang, D. H., & Lindholm, C. D. (2010). SELENA–An open-source tool for seismic risk and loss assessment using a logic tree computation procedure. Computers & Geosciences, 36(3), 257-269.
Ranjbar, H. R., Dehghani, H., Ardalan, A. R. A., & Saradjian, M. R. (2017). A GIS-based approach for earthquake loss estimation based on the immediate extraction of damaged buildings. Geomatics, Natural Hazards and Risk, 8(2), 772-791.
Ranjbar, H. R., & Nekooie, M. A. (2018). An improved hierarchical fuzzy TOPSIS approach to identify endangered earthquake-induced buildings. Engineering applications of artificial intelligence, 76, 21-39.
Ren, P., Xu, Z., & Gu, J. (2016). Assessments of the effectiveness of an earthquake emergency plan implementation with hesitant analytic hierarchy process. International journal of information technology & decision making, 15(06), 1367-1389.
Sadrykia, M., Delavar, M. R., & Zare, M. (2017). A GIS-based fuzzy decision making model for seismic vulnerability assessment in areas with incomplete data. ISPRS International Journal of Geo-Information, 6(4), 119.
Şen, A., & Ekinci, V. (2016). Realization of Earthquake Vulnerability Analysis in Structure Scale with Fuzzy Logic Method in GIS: Kadikoy, Maltepe and Prince Islands Sample. International Journal of Environment and Geoinformatics, 3(3), 40-56.
Shiono, K., Krimgold, F., & Ohta, Y. (1991). Post-event rapid estimation of earthquake fatalities for the management of rescue activity. In Comprehensive Urban Studies (pp. 61-106).
Walker, B. B., Taylor-Noonan, C., Tabbernor, A., McKinnon, T. B., Bal, H., Bradley, D., Schuurman, N., & Clague, J. J. (2014). A multi-criteria evaluation model of earthquake vulnerability in Victoria, British Columbia. Natural Hazards, 74(2), 1209-1222.
Wang, Y. (2021). Multiperiod Optimal Allocation of Emergency Resources in Support of Cross-Regional Disaster Sustainable Rescue. International Journal of Disaster Risk Science, 1-16.
Yariyan, P., Avand, M., Soltani, F., Ghorbanzadeh, O., & Blaschke, T. (2020). Earthquake vulnerability mapping using different hybrid models. Symmetry, 12(3), 405.
Zhang, Z., Demšar, U., Rantala, J., & Virrantaus, K. (2014). A fuzzy multiple-attribute decision-making modelling for vulnerability analysis on the basis of population information for disaster management. International Journal of Geographical Information Science, 28(9), 1922-1939.
Zhou, J., & Liu, B. (2003). New stochastic models for capacitated location-allocation problem. Computers & industrial engineering, 45(1), 111-125.
Zhou, P., Zhang, L., Zhang, S., & Alkhateeb, A. F. (2020). Observer-based adaptive fuzzy finite-time control design with prescribed performance for switched pure-feedback nonlinear systems. IEEE Access, 9, 69481-69491.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Development of a solution based on GIS and fuzzy logic to rescue earthquake victims

Status:

Version 1

Abstract

Figures

1 Introduction

2 Materials and methods

2.1 Conceptual model

2.2 Quantification of the impact of earthquakes on buildings

2.3 Required data

2.3.1 Population

2.3.2 Type of structure

2.4 Estimation of structural damages and human losses

2.4.1 Coefficients M1 and M2

2.4.2 Coefficient M3 and related table

2.4.3 Coefficient M4d and related table

2.4.4 Coefficient M5 and related table

2.5 Fuzzification

3 Results and discussion

3.1 Study area

3.2 Data of the study area

3.2.1 Population

3.2.2 Classification of buildings

3.3 Fuzzification and construction of fuzzy inference engine

3.4 Survivability results

4 Conclusions

Declarations

References

Additional Declarations

Status:

Version 1

Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	5%	30%	50%	30%	40%	40%	60%	80%	20%
4	5%	30%	50%	30%	40%	40%	60%	80%	20%
5	5%	30%	50%	30%	40%	40%	60%	80%	20%
6	5%	30%	50%	30%	40%	40%	60%	80%	20%
7	5%	30%	50%	30%	40%	40%	60%	80%	20%
8	5%	30%	50%	30%	40%	40%	60%	80%	20%
9	5%	30%	50%	30%	40%	40%	60%	80%	20%
10	5%	30%	50%	30%	40%	40%	60%	80%	20%
11	5%	30%	50%	30%	40%	40%	60%	80%	20%
12	5%	30%	50%	30%	40%	40%	60%	80%	20%

Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	95%	95%	95%	95%	95%	95%	95%	95%	95%
4	95%	95%	95%	95%	95%	95%	95%	95%	95%
5	95%	95%	95%	95%	95%	95%	95%	95%	95%
6	95%	95%	95%	95%	95%	95%	95%	95%	95%
7	95%	95%	95%	95%	95%	95%	95%	95%	95%
8	95%	95%	95%	95%	95%	95%	95%	95%	95%
9	95%	95%	95%	95%	95%	95%	95%	95%	95%
10	95%	95%	95%	95%	95%	95%	95%	95%	95%
11	95%	95%	95%	95%	95%	95%	95%	95%	95%
12	95%	95%	95%	95%	95%	95%	95%	95%	95%

Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	80%	70%	60%	70%	60%	40%	60%	70%	60%
4	80%	70%	60%	70%	60%	40%	60%	70%	60%
5	80%	70%	60%	70%	60%	40%	60%	70%	60%
6	80%	70%	60%	70%	60%	40%	60%	70%	60%
7	80%	70%	60%	70%	60%	40%	60%	70%	60%
8	80%	70%	60%	70%	60%	40%	60%	70%	60%
9	80%	70%	60%	70%	60%	40%	60%	70%	60%
10	80%	70%	60%	70%	60%	40%	60%	70%	60%
11	80%	70%	60%	70%	60%	40%	60%	70%	60%
12	80%	70%	60%	70%	60%	40%	60%	70%	60%

Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	35%	50%	50%	50%	50%	30%	50%	60%	50%
4	35%	50%	50%	50%	50%	30%	50%	60%	50%
5	35%	50%	50%	50%	50%	30%	50%	60%	50%
6	35%	50%	50%	50%	50%	30%	50%	60%	50%
7	35%	50%	50%	50%	50%	30%	50%	60%	50%
8	35%	50%	50%	50%	50%	30%	50%	60%	50%
9	35%	50%	50%	50%	50%	30%	50%	60%	50%
10	35%	50%	50%	50%	50%	30%	50%	60%	50%
11	35%	50%	50%	50%	50%	30%	50%	60%	50%
12	35%	50%	50%	50%	50%	30%	50%	60%	50%

Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	15%	30%	45%	35%	45%	25%	45%	55%	45%
4	15%	30%	45%	35%	45%	25%	45%	55%	45%
5	15%	30%	45%	35%	45%	25%	45%	55%	45%
6	15%	30%	45%	35%	45%	25%	45%	55%	45%
7	15%	30%	45%	35%	45%	25%	45%	55%	45%
8	15%	30%	45%	35%	45%	25%	45%	55%	45%
9	15%	30%	45%	35%	45%	25%	45%	55%	45%
10	15%	30%	45%	35%	45%	25%	45%	55%	45%
11	15%	30%	45%	35%	45%	25%	45%	55%	45%
12	15%	30%	45%	35%	45%	25%	45%	55%	45%

Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	5%	30%	50%	30%	40%	40%	60%	80%	20%
4	5%	30%	50%	30%	40%	40%	60%	80%	20%
5	5%	30%	50%	30%	40%	40%	60%	80%	20%
6	5%	30%	50%	30%	40%	40%	60%	80%	20%
7	5%	30%	50%	30%	40%	40%	60%	80%	20%
8	5%	30%	50%	30%	40%	40%	60%	80%	20%
9	5%	30%	50%	30%	40%	40%	60%	80%	20%
10	5%	30%	50%	30%	40%	40%	60%	80%	20%
11	5%	30%	50%	30%	40%	40%	60%	80%	20%
12	5%	30%	50%	30%	40%	40%	60%	80%	20%

Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	95%	95%	95%	95%	95%	95%	95%	95%	95%
4	95%	95%	95%	95%	95%	95%	95%	95%	95%
5	95%	95%	95%	95%	95%	95%	95%	95%	95%
6	95%	95%	95%	95%	95%	95%	95%	95%	95%
7	95%	95%	95%	95%	95%	95%	95%	95%	95%
8	95%	95%	95%	95%	95%	95%	95%	95%	95%
9	95%	95%	95%	95%	95%	95%	95%	95%	95%
10	95%	95%	95%	95%	95%	95%	95%	95%	95%
11	95%	95%	95%	95%	95%	95%	95%	95%	95%
12	95%	95%	95%	95%	95%	95%	95%	95%	95%

Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	80%	70%	60%	70%	60%	40%	60%	70%	60%
4	80%	70%	60%	70%	60%	40%	60%	70%	60%
5	80%	70%	60%	70%	60%	40%	60%	70%	60%
6	80%	70%	60%	70%	60%	40%	60%	70%	60%
7	80%	70%	60%	70%	60%	40%	60%	70%	60%
8	80%	70%	60%	70%	60%	40%	60%	70%	60%
9	80%	70%	60%	70%	60%	40%	60%	70%	60%
10	80%	70%	60%	70%	60%	40%	60%	70%	60%
11	80%	70%	60%	70%	60%	40%	60%	70%	60%
12	80%	70%	60%	70%	60%	40%	60%	70%	60%

Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	35%	50%	50%	50%	50%	30%	50%	60%	50%
4	35%	50%	50%	50%	50%	30%	50%	60%	50%
5	35%	50%	50%	50%	50%	30%	50%	60%	50%
6	35%	50%	50%	50%	50%	30%	50%	60%	50%
7	35%	50%	50%	50%	50%	30%	50%	60%	50%
8	35%	50%	50%	50%	50%	30%	50%	60%	50%
9	35%	50%	50%	50%	50%	30%	50%	60%	50%
10	35%	50%	50%	50%	50%	30%	50%	60%	50%
11	35%	50%	50%	50%	50%	30%	50%	60%	50%
12	35%	50%	50%	50%	50%	30%	50%	60%	50%

Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	15%	30%	45%	35%	45%	25%	45%	55%	45%
4	15%	30%	45%	35%	45%	25%	45%	55%	45%
5	15%	30%	45%	35%	45%	25%	45%	55%	45%
6	15%	30%	45%	35%	45%	25%	45%	55%	45%
7	15%	30%	45%	35%	45%	25%	45%	55%	45%
8	15%	30%	45%	35%	45%	25%	45%	55%	45%
9	15%	30%	45%	35%	45%	25%	45%	55%	45%
10	15%	30%	45%	35%	45%	25%	45%	55%	45%
11	15%	30%	45%	35%	45%	25%	45%	55%	45%
12	15%	30%	45%	35%	45%	25%	45%	55%	45%

Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	5%	30%	50%	30%	40%	40%	60%	80%	20%
4	5%	30%	50%	30%	40%	40%	60%	80%	20%
5	5%	30%	50%	30%	40%	40%	60%	80%	20%
6	5%	30%	50%	30%	40%	40%	60%	80%	20%
7	5%	30%	50%	30%	40%	40%	60%	80%	20%
8	5%	30%	50%	30%	40%	40%	60%	80%	20%
9	5%	30%	50%	30%	40%	40%	60%	80%	20%
10	5%	30%	50%	30%	40%	40%	60%	80%	20%
11	5%	30%	50%	30%	40%	40%	60%	80%	20%
12	5%	30%	50%	30%	40%	40%	60%	80%	20%

Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	95%	95%	95%	95%	95%	95%	95%	95%	95%
4	95%	95%	95%	95%	95%	95%	95%	95%	95%
5	95%	95%	95%	95%	95%	95%	95%	95%	95%
6	95%	95%	95%	95%	95%	95%	95%	95%	95%
7	95%	95%	95%	95%	95%	95%	95%	95%	95%
8	95%	95%	95%	95%	95%	95%	95%	95%	95%
9	95%	95%	95%	95%	95%	95%	95%	95%	95%
10	95%	95%	95%	95%	95%	95%	95%	95%	95%
11	95%	95%	95%	95%	95%	95%	95%	95%	95%
12	95%	95%	95%	95%	95%	95%	95%	95%	95%

Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	80%	70%	60%	70%	60%	40%	60%	70%	60%
4	80%	70%	60%	70%	60%	40%	60%	70%	60%
5	80%	70%	60%	70%	60%	40%	60%	70%	60%
6	80%	70%	60%	70%	60%	40%	60%	70%	60%
7	80%	70%	60%	70%	60%	40%	60%	70%	60%
8	80%	70%	60%	70%	60%	40%	60%	70%	60%
9	80%	70%	60%	70%	60%	40%	60%	70%	60%
10	80%	70%	60%	70%	60%	40%	60%	70%	60%
11	80%	70%	60%	70%	60%	40%	60%	70%	60%
12	80%	70%	60%	70%	60%	40%	60%	70%	60%

Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	35%	50%	50%	50%	50%	30%	50%	60%	50%
4	35%	50%	50%	50%	50%	30%	50%	60%	50%
5	35%	50%	50%	50%	50%	30%	50%	60%	50%
6	35%	50%	50%	50%	50%	30%	50%	60%	50%
7	35%	50%	50%	50%	50%	30%	50%	60%	50%
8	35%	50%	50%	50%	50%	30%	50%	60%	50%
9	35%	50%	50%	50%	50%	30%	50%	60%	50%
10	35%	50%	50%	50%	50%	30%	50%	60%	50%
11	35%	50%	50%	50%	50%	30%	50%	60%	50%
12	35%	50%	50%	50%	50%	30%	50%	60%	50%

Intensity MMI	Type-1 steel - Type-0 reinforced concrete	Type-2 steel	Brick-steel	Type-1 reinforced concrete	Type-2 reinforced concrete	All-wood	Brick and cement block	Adobe	Masonry (according to Coburn)
3	15%	30%	45%	35%	45%	25%	45%	55%	45%
4	15%	30%	45%	35%	45%	25%	45%	55%	45%
5	15%	30%	45%	35%	45%	25%	45%	55%	45%
6	15%	30%	45%	35%	45%	25%	45%	55%	45%
7	15%	30%	45%	35%	45%	25%	45%	55%	45%
8	15%	30%	45%	35%	45%	25%	45%	55%	45%
9	15%	30%	45%	35%	45%	25%	45%	55%	45%
10	15%	30%	45%	35%	45%	25%	45%	55%	45%
11	15%	30%	45%	35%	45%	25%	45%	55%	45%
12	15%	30%	45%	35%	45%	25%	45%	55%	45%