Earthquake Vulnerability Mapping using Binary Comparison Matrix and Analytical Hierarchical Process (AHP) in Lalmatia, Dhaka

From the historical records, geological evident and recent trends in earthquake, it is evident that Bangladesh is in a high-risk zone of earthquake hazard. The recent results of the CDMP, if a huge earthquake greater or equal to seven magnitude happened in this country, would lead to a serious human tragedy due to the defective structure. In Lalmatia, Dhaka study area there are various types of structures as like as RCC, masonry, semi-pucca etc. and the area has different old and newly filled soil development. This study considered the characteristics of RCC building elements and their behaviors to assess the risk against earthquake vulnerability in Lalmatia using the Turkish method. Next building vulnerability from Turkish Method used as one total vulnerability factor. In the method of this research, a pair wise comparison matrix for a numerical relationship between two elements and the Analytical Hierarchical Process (AHP) model has been applied to decide in weight and to get rank of the vulnerability factors. By weighted sum vector among the six factors and calculating vulnerability index (VI), the overall vulnerability were identified in Lalmatia. Using Geographic Information System (GIS) and defining an ordinal scale of calculated result, the vulnerability status of single buildings of the study area is presented here. This research tells that 8.23% buildings are highly vulnerable and 14.24% buildings are moderately vulnerable to earthquake. About 77.53% buildings are less vulnerable. As the study area is in a major urban center (Dhaka City), the scenario of unplanned urbanization increases the overall vulnerability to higher scale.


Introduction
Dhaka with estimated population in 2020 is roughly 2.1 million (UN, 2020) and density of population is 44,500 per sq.km. (UN Habitat, 2020) which is a muddy plain including of a shallow water table (Islam et al., 2010). Unplanned urbanization has been going on rapidly for the last few decades without appropriate planning. Consequently, fast urbanization has resulted in hasty population growth and unplanned building to accommodate Dhaka city dwellers. Dhaka is said as one of the 20 most risky cities in the world in the perspective of the earthquake risk and infrastructure protection disaster risk index (Ahmed and Ahmed, 2010). Though Bangladesh has not evidenced any serious damaged earthquake for more than 100 years but the unplanned situation of the infrastructure and the geological situation may happen that Bangladesh is at high risk of damaging earthquake in upcoming future (CDMP, 2009a(CDMP, , 2009b(CDMP, , 2009c. Earthquake seismic problems faced by the undulating ground and cities (Choudhury 1993;Paul and Bhuiyan 2010;Rashed and Weeks 2003;Sarraz et al. 2015;Sattar et al. 2009;Ziauddin 2016) in Bangladesh (e.g., Chittagong and Dhaka city) and the problems are also enlarging the spatial vulnerability as like as movement of plate, neo-tectonic depression (Khan and Hoque, 2002), geological development (Alam et al., 2008) and weekend in earth structure (Hossain, 2014). The city with rapid population and unplanned urbanization (Hossain, 2014), without seismic capacity multi-storied buildings (Islam et al., 2016) and old aged building which is non-regulated with seismic capacity (Ansary and Ali 2004;Hossain 2014;Jahan et al. 2008;Kamal 2009;Reja 2008) also may added dangerous problems in earthquake. In Last 100 years Bangladesh and adjacent areas faced 1000 earthquakes taking a magnitude equal or more than 4 scale (Alam et al., 2011) and earthquakes are expected equivalent about 60% of total natural disaster in world (Agarwal and Shrikhande, 2007).
Without appropriate planning Dhaka is developing very fast and as a result we can see more incidents like the collapse of the Begunbari building on June 1, 2010. The buildings were built on wet land and in earthquake the soil liquefaction may happened on this building (Rahman, 2010). The Meteorological Department and BUET has identified about 90 earthquakes were occurred in Bangladesh through May 2007 to July 2008. Among the identified history of earthquakes, nine of which are above five on the Richter scale and 95 % of which were within a radius of 600 km of Dhaka city (Ferdous and Rahman, 2015). CDMP (2009aCDMP ( , 2009bCDMP ( , 2009c assessed that the Madhupur fault generates 7.5 Mw magnitude for Dhaka city. According to this assessment, out of total 3,26,000 buildings, approximately 270,604 buildings will be at least moderately damaged which comprises over 89% of total building stock. Besides 238,164 buildings will be damaged outside restoration. Around 260,788 and 182,450 people will die respectively for an earthquake taking place at 2:00 AM and 2:00 PM. Around 1,527,668 people will be displaced aftermaths an earthquake (CDMP, 2009a(CDMP, , 2009b(CDMP, , 2009c).
An assessment was piloted by ADPC from February 2008 to August 2009 in Bangladesh Government initiatives titled on Comprehensive Disaster Management Program (CDMP). According to this study, a 7.5 magnitude earthquake originated from the Madhupur fault could have killed at least 1,30,000 people if the earthquake had been attacked in daytime in Dhaka. An earthquake of 8 Richter scale created close to the Chittagong of plate boundary fault 2 may kill about 69,900 people living in the capital if the earthquake had been attacked in daytime. There may 13,600 people need to be hospitalized and 61,288 people may need first aid treatment (CDMP, 2010). The planning interference in emergency evacuation preparedness (Chakrabarty, Rahman and Ubaura, 2020) may reduce the hazard impact (Rahman, Tariq and Sharmin, 2020). Thus, the capital city Dhaka and Bangladesh both are extremely vulnerable to earthquake and considering these aspects the earthquake vulnerability in the study area of Lalmatia was conducted. The Study was also conducted to realize the public's awareness about the willingness of building owners to retrofit the existing building against earthquake vulnerability.

Objectives and Methods
The main aim of this research is to assess the earthquake vulnerability of Lalmatia area, Dhaka city. To implement this aim, the following objectives have been taken into consideration: to assess the earthquake vulnerability in the existing buildings using Turkish Method and to investigate the earthquake vulnerability priority using Analytical Hierarchy Process.
Data for this research were collected through field observation and checklist survey. Primary data will be collected about number about stories, existence of a soft story, apparent building quality, existence of heavy overhangs, shape of the building, existence of short columns and pounding effect. For the study sample size is calculated through the following procedure: Total population (Building) size (N) = 1647, Error level (e) = 5%, Confidence level 95% and sample size was determined as 316. Secondary data were collected from different published and unpublished materials, journals, websites of different responsible organization etc.

Source: Developed by author, 2019
Map 1: Study Area Map

Turkish Method
The Turkish government and the Japan International Cooperation Agency (JICA) came forward to implement a regional earthquake assessment and rehabilitation program after the 1999 earthquake in Kocaeli and Duzce. The Turkish method Level-1 is used in this work. The first phase of the survey from the sidewalk was conducted by observers through a walkdown visit.

Survey Parameters
The parameters selected in the Level-1 survey to indicate building vulnerabilities are as follows: Yes = 1 or No = 0 The intensity of ground motion at a particular location depends mainly on the efficiency of the distance and the local soil conditions. There is a strong relationship between PGV (Peak Ground Velocity) and local soil shear wave velocity (Chowdhury, 1993). So PGV was selected to represent the intensity of ground motion in the study. Peak ground velocity (PGV) can be taken between 40 cm /sec to 50 cm /sec (Wu et al., 2003). Thus, Zone II (40 <PGV<60) is considered for calculating performance scores because our study area matched the same characteristics with Zone II. The different base scores described in Table 1 which were determined based on the number of stories and the earthquake risk level in the site building.  Sucuoglu and Yazgan, 2003 Building Seismic Performance At first the vulnerability factors are fixed by the walk down survey and then the location of the building is determined by its location (by GPS survey), the seismic Performance Score (PS) can be finding out by using Eq. Then, the vulnerability value is found which is equal to the PS divided by BS. If this computed value is low, the vulnerability of the building will be high. Decide the range of vulnerability levels (Table 2) in the study (very low, low, medium, high and very high) and develop a map of vulnerability of concrete buildings.  Sucuoglu and Yazgan, 2003 This assessment process of building vulnerability is only applicable for RCC (pucca) building. In this process tin shed and semi pucca buildings are not evaluate for the vulnerability assessment of the study area.

Analytical Hierarchy Process (AHP):
The AHP approve the decision maker to create a model consisting complex problem in hierarchical way indicating the relevance of the goal, criteria, sub criteria and alternatives. It additionally permits the decision maker to incorporate each subjective and objective concerns during this method (saaty, 1980). The AHP method involves the following basic steps:  Construction of the hierarchy  Comparative judgments or executing data collection to achieve pair wise comparison data of the hierarchical structure on elements.  Overall priority rating construction (Harker & Vargas, 1987) At the first stage, decision maker's necessity to break down the complex multiple criteria decisions into its component. At each level of hierarchy, the criteria and sub criteria are not equally important to taking decision. In the decision-making task, AHP is able to consolidate and combine the evaluations of the criteria and alternatives by group or individual (Eastman et al., 1993). AHP and Multi criterion analysis is using in this study to explore the vulnerability of Lalmatia against earthquake torment. Table 3 explore that six parameters were selected to vulnerability appraisement and then six factors are further categorized into five sub criteria. Factors affecting the vulnerability against earthquake is shown in table 3.  Saaty (1980) developed the pair wise comparison method in the context of AHP (Analytical Hierarchy Process). Comparisons create a ratio matrix, as it takes the parameter in pair wise to produce the relative weights. Personal and subjective judgements can be taken in comparison (Chen, 2016). At a given time, two elements compared of this analysis can reduces the conceptual complexity (Muralidhar et al., 1990;Parvoti, 1992;Saaty, 1980). Three task involves in this analysis:  A comparison matrix development at each level of hierarchy  Relative weights calculation for each element of hierarchy  Consistency ratio estimating to check the judgment consistency (Li et al., 2006)  Equal rank Two events give equal judgment 3 Weak position of one over another Results are marginally favor one action over another 5 Strong importance Result strongly favor one activity over another 7 Confirmed importance Strongly favored an activity and its control is validated in practice 9 Entire importance Strongly favored an activity and is the maximum potential order of confirmation 2, 4, 6, 8 Middle values between the two-Negotiation is required neighboring decision

Reciprocals of Above Nonzero
If i shows the above nonzero numbers in the activity when comparing to activity j, then j will be common values in relating with i Source : Saaty, (1980) Analytical hierarchy used the 9-point Scale for ranging from 1 to 9 (indifference or equal importance to extreme preference or absolute importance) which is shown in table 4. In this comparison matrix elements are compared in pairs in each level with respect to importance. The decision maker evaluates the contribution of each factor in this pair wise comparison matrix. In the comparison matrix at a given level will be reduced to a number of square matrices M= [aij]n×n as in the following:  Source : Saaty, (1980) Discussions From the filed investigation (2019), it has been identified that several kind of land uses are presents in lalmatia like residential, commercial, educational, institutional, recreational etc. (Map 2). A very general picture in the study area are the first floor of buildings used for commercial purposes when the upper floors are completely residential. It was found that many mixed use activity like shops, service activity, residence are belonging the same building at a time. Lalmatia is now taken by developer and rapid development contributes to land use change in this area. Privatization is main factors of land use change. Residential land use increases day by day. Now the residential land use is 81.83% of the total land of Lalmatia area. Field survey reveals that Lalmatia is residential area, here most of the buildings use is residential. Second most use type is commercial, and then community service. Table 6 will present the number of building with their use. In case of response and recovery phase after a severe earthquake, road network especially road width has great momentous in moving the emergency vehicles. The road network of Lalmatia area is far better now. The road network has a deep relation with urban development because the Nowadays the structural development occurred and high-rise buildings are constructed rapidly. In Lalmatia area, 6 storied, 10 storied and 14 storied buildings are available. Different developer and private real estate firm constructed many high-rise buildings in this area. Different private hospitals, organization offices, commercial buildings, etc. are constructed at this time period. Arrangement of the buildings in the study area according to build height is shown on map 5.
The relation between numbers of buildings with the presence of short columns (Existence/ Nonexistence) in the building are displayed in the table 7. Table 7 shows that about 4% buildings have short columns. Due to the Presence of the short columns, building makes it subject to Torsional effect. For analyzing the current building vulnerability of the study area, 316 buildings (map 7) were selected based on preliminary survey. Such buildings were selected which are RCC, masonry and others type of buildings having one to above ten stories. The survey for seismic vulnerability analysis was mainly focused on earthquake issues such as identifying building type, number of building story, presence of soft story, presence of heavy overhang, apparent building quality, presence of short column, pounding effect, number of inhabitants, area of parcel, road width, building use, year of construction, etc.

Source: Developed by Author, 2019
Map 7: Selected Building for vulnerability Assessment Lalmatia contain three types of building like RCC, masonry and semi-pucca building, but they are different in amount. Most of the building types are RCC. The relation between numbers of buildings with the number of stories (such as one-two, three-four, five-six and above stories) are shown in figure 1.

Number of Stories
In figure 3 represent the relation between numbers of buildings with the presence of heavy overhangs. In the study area, found that most buildings tend to have overhangs (mostly balcony). Here, almost 38% building having a heavy overhang. Buildings with heavy overhang was built in most of the upper floor from two to three feet. Figure  4 reveals that the percentage of soft story buildings are less than the buildings without soft story. Almost 38% buildings having soft story. Figure 5 shows the relation between numbers of buildings with the apparent building quality in the study area. It is found that the majority of the apparent building quality is good in Lalmatia area, and it is almost 59.2%.   Overall Vulnerability Scoring by AHP AHP (Analytical Hierarchy Process) is a flexible, effective, and simple method to decision-making process. Saaty proposed the AHP method in 1980. It is a familiar method that decomposes several level decision-making problems through creating hierarchical relationship between different levels. This method uses comparison as a pair to distribute weights of different factors that helps to measuring the relative importance through the suing Saaty's 1 to 9 level scales. Consistency ratio (CR) is also calculating to verify the judgmental coherence. The accepted consistency ratio must be about 0.1 or less. This method includes following three steps-

Generating of a Binary Comparison Matrix
According to Saaty, a pair wise comparison matrix is a numerical relationship between two elements that appreciate more important element. In the numerical representation the weight of each factor compared together. Matric n×n (in this case 6×6) record the results that also called binary comparison matric Aij = [a n×n]. In Analytical Hierarchy Process, all elements of the metric are positive and concerning the "reverse condition" (the weight of j in regard to i will equal to 1/k, if the weight of i in regard to j equals to k). In every binary comparison matric, we will have two numerical quantity of Aij and 1/Aij.

Calculation Step for Different Factors
A comparison matrix (table 8) has been developed to ascertain the each factors weight. Calculation step includes the following: Step 1: Computation of Weighted Sum Vector (WSV) table 8. Step 2: Calculating the Inconsistency Vector (IV) (table 9).
In our result the CR is estimated 0.085 which means there is a consent in result because we know that if CR is greater than 0.1 then the result should be reassessed and if CR ≤ 0.1 then it should be agreement in the result.

Overall Vulnerability Evaluation
The weights for the criteria are computed to evaluate the overall vulnerability using AHP method and afterwards vulnerability map of Lalmatia is prepared based on vulnerability level.

Findings
At the stage of Turkish Method, the range of vulnerability levels is depended on high rise, large, heavy overhang, huge soft story, have close change of pounding effect, short column existence is severe and apparent building quality is bad. Table 10 represents high vulnerable buildings score by Turkish method. With the value of Turkish Method, at the time of AHP model the range of vulnerability levels is depended on huge population, very high construction age, narrow road and building use is residential. AHP considers all multiple aspects that can affect any building vulnerability and weighted with seismic buildings related factors and present the vulnerability category. According to AHP method, by giving a priority on very high vulnerable and high vulnerable building (Turkish method) represents the three priority ranking. In case of very high vulnerable buildings (total buildings 5), 2 buildings get first priority, 1 get second priority and 2 gets third priority. In case of high vulnerable building (total buildings 21), 13 buildings get first priority, 5 gets second and 3 gets third priority. This priority list represents priority-based retrofitting. Figure 8 represents the priority list of high vulnerable buildings. AHP method is used only for very high vulnerable building and a high vulnerable building in Lalmatia area. Map 9 represents the very high vulnerable building priority and map 10 represent the high vulnerable building priority. It was concluded that (based on AHP results) 5 buildings of very high vulnerable get more priority for retrofitting. And then 21 buildings get priority for retrofitting. Study Area Map Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.

Figure 2
Land use Map of Study Area Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors. Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.

Figure 4
Building Types Map of Study Area Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.

Figure 5
Building Stories Map of Study Area Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.

Figure 6
Construction Year Map of Selected Buildings Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.

Figure 7
Selected Building for vulnerability Assessment Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.

Figure 8
Vulnerability Assessment Map by Turkish Method Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.

Figure 9
Vulnerability Assessment Map (1) by AHP Method Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.

Figure 10
Vulnerability Assessment Map (2) by AHP Method Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.