GIS in Estimation of Human Damage in Earthquake


Zhexin HU*, Satoshi YOSHIDA and Satoru SADOHARA


Abstract
Human damage depends greatly on the time when earthquake happen. And the problem of local governments for the seismic damage assessment for Regional Plan is: the input earthquake time cannot be identified beforehand; therefore the damage that will occur might not necessarily be like the result. In this paper, we conducted a survey on time budgets, integrated data of earthquake motion, soil condition, urban structure and human distribution in an information system. With GIS we estimated the human damages in various cases such as fire occurring, building collapsing, transportation system destroyed and etc. in different time of a day.

Introduction
Regarding to earthquake, nowadays in Japan, estimation of human damage and resources allocation to mitigate the damages are conducted according to some fuzzy criteria. This paper aims to contribute in this field. Its major goals are to develop an information system that can identify human activities and distribution. Results of this study could contribute to the understanding of urban activities and the identification of population distribution pattern in space and time, which is essential in estimation of human damage and resources allocation for disaster prevention and preparedness. The applied methodologies include multivariate statistical analysis and the implementation of a Geographical Information System (GIS). Attention is devoted to the role that GIS currently and potentially plays in estimation of human damage and improving the efficiency of disaster prevention measures. It is our future aim to work out an information system with GIS in the area of human damage estimation mapping and visualization.

In most past study on estimation of human damage due to earthquake, earthquake was assumed to happen at PM 6:00 in winter when was thought to be the most dangerous time in the course of a day. However, the risk of human life differs dramatically according to "What they are doing" (activities) and "Where they are" (spatial distribution) at the moment when earthquake happen. Unlike those static factors (building, road, open space and etc.) that are established in urban, "Activity" and "Distribution" are dynamic factors that keep changing over time. Therefore, the risk of human damage differs in space and time. For example, if earthquake happens in time when people are preparing food, it's mostly possible to cause fire and its spread due to cooking. If it happens in commuting rush hour, damage to people who are in trains is severe. Even though the location is the same such as at a person's home, risks of damage depend greatly on whether they are sleeping or not. Thus, purpose of this paper is to develop an information system that can identify human activities and distribution. Referring to the estimation of human damage, we proposed a predictive and operational model that is embedded within a geographic information system (GIS). The contribution of GIS might be illustrated as figure 1.

Figure 1

Figure 1. A paradigm for geographic information science contribution to estimation of human damag

GIS and Human Damage Estimation
GIS has great potential in research of human damage estimation because of its three key functions: database management, spatial analysis and visualization. With the collective presence of these three components, GIS provide the capabilities of linking human activities with their geographic location. It also performs overlay functions that allow the association of population distribution with socioeconomic factors. Currently, no system has been developed to have the capacity to efficiently perform identification of time series and population distribution pattern. To show the usefulness of GIS technology in estimation of human damage, a case study was conducted in Minami ward of Yokohama city. Although developed for Minami ward, the methodology can be used in other urban area of Japan.

Methodology
The methodology to conduct this study involved the following two steps.
1 What they are doing
This step of study aimed to provide a method to calculate percentages and numbers of people who are doing various daily activities. Relationship between activity pattern and the differences of ages, genders and the status of employment are analyzed and results proved that activity pattern is affected mostly by the status of employment (employed or unemployed) instead of differences of genders and ages.
1-1 Survey & Cluster Analysis
A survey was implemented in order to identify human activities pattern in a day. Activities that related to human damage in earthquake are extracted as: activities at home (night sleeping, household working, cooking, and activities out of home (commuting, working, shopping, and etc.). The questionnaire contains the time budget of daily activities and respondents' age, gender and status of employment (employed or unemployed). Cluster analysis was conducted not only to search for homogenous groups (cluster) of people with similar activities' patterns, but also to study the association between activities' patterns and the difference of ages, genders and status of employments. Analysis results show that urban activities and the potential risk of human damage may differ greatly due to differences of population composition.
1-2 Visualization of the results
With almost 222,283 inhabitants and an average of 17,600 inhabitants per square kilometer, Minami ward is the most populated area in Yokohama city with the highest percentage of single aged person of 559 per kilometer. These facts justify the selection of Minami ward of Yokohama city as a case study for this project. A series of maps were generated to graphically display the percentage of people who could be at home and the damage risk of aged people at home.

2 Where they are This step of study aims to develop a system that can provide time series population distribution. It can show how many potential numbers of people there may be in various urban facilities such as buildings, road, train, and station. Therefore, overlaying functions can be used to combine population distribution layer with other layers as shown in Figure 1. At this moment, this phase of the study has not started yet. Ideas were proposed as follow:
2-1 Development of GIS database Overlaying functions can be used to link demographic features and socioeconomic data with their geographic locations. Data of daytime and nighttime population can be collected from nation census conducted every five years. Data regarding to the commuting and shopping population can be collected from urban transportation and economic census. Socioeconomic data include inhabitants, building, jobs, commercial establishments, accessibility and etc.
2-2 Visualization
Series maps can be produced to show the total number of population and the density of population per area. Because of the versatility of GIS, correlation between population distributions and the socioeconomic factors can be displayed and analyzed in a clear and concise format.
2-3 Application
It is expectable that there are many implications in regards to the information system of population distribution and its future uses in the issue of human damage. Not only can the system of human distribution benefit administrators in their efforts to target high population concentration area, but they also have the potential to help the public estimate risk with preventative measures. Besides, local governments or communities could be provided with better information for an immediate judgments and determine what is the best means to respond in the very beginning of earthquake. Also, urban planner agencies could benefit from the population distribution in determining the allocation of resources and initiating new programs.

Conclusions
This study shows that using GIS is a compatible mean in human damage estimation because of its geographic referencing capabilities. It can be made successfully in communication, risk and vulnerability assessment, study of life loss patterns, planning for disaster prevention and etc. These maps can be successfully used in establishing response priorities, developing actions plans, quick disaster location assessment for carrying out search and rescue operations effectively. GIS is an alternative system that could aid in human damage prevention and result in the decision-making process.

References
Nedovic-Budic, Z. and D.R Godschalk, 1996, Human Factors in Adoption of Geographic Information System: A local Government Case Study. Public Administration Review, 56(6), 554-567.
Monica Alexander and Wei-Ning Xiang, 1994, Crime Patern Analysis Using GIS. GIS/LIS, p1-3
Webster, C.J., 1993, GIS and the Scientific inputs to Urban Planning. Part 1: Description. Environment and Planning B, 20 (6), 709-728.


Author Information
*Graduate student
Graduate School of Environment & Information Sciences,Yokohama National University
79-7, Tokiwadai, Hodogaya-ku, Yokohama 240-8501, Japan
E-mail: d00je905@ynu.ac.jp
Tel: +81-45-339-4249
Fax: +81-45-338-1016