In a comprehensive manner, the term disaster means sudden calamities or damages resulting from droughts, conflagrations, epidemics, earthquakes, typhoons, and floods. Natural disasters, apart from man-made ones, are defined as the kinds of disasters caused by natural phenomena. They are generally characterized as unexpected in time and place, with numerous unspecified victims and extensive areas of damage. In this respect, the prevention of natural disasters, whose time and place of occurrence are unexpectable, is very critical.
The structure of cities has an inevitable vulnerability to natural disasters. Such vulnerability is attributed to artificial alterations in areas and terrains in natural conditions, where humans have built structures for residential purposes. Nature with altered conditions tends to result in more damages than altered nature does when under natural disasters of the same intensity. Thus, the recent trends of urban engineering are focused on the field of disaster prevention. In doing so, disaster prevention is highly considered in urban master plans. Prevention plans for river basins have been established. Under these circumstances, identifying and analyzing current situations in urban area disasters are crucial.
South Korea, as a highly urbanized country, has been managing all the regions in its territory with urban planning elements. So, both urban and non-urban regions in the country have some urban characteristics. In addition, most natural disasters in the country consist of wind and flood. Therefore, the government and the local governments are concentrating their relevant policies and budgets on preventing and restoring damages by wind and flood. Wind and floor cause both direct and indirect damages to the target area`s urban characteristics. Moreover, due to the increasing amount of weather anomalies, damages by natural disasters are becoming intensified. That is why it is necessary at this point of time to research on natural disasters; research fields need to be multifaceted including occurrence prediction in a meteorological manner, methods to mitigate natural disaster damages, and the amount of damages caused by natural disasters. In particular, to analyze the association between urban characteristics and natural disasters from a urban engineering perspective is worthwhile to prevent future disasters and minimize damages from them.
This study is to empirically analyze the factors that affect damages by natural disasters, considering urban characteristic in a comprehensive and macroscopic manner, on the areas of Gyeongsangnam-do. And, with the outcomes drawn from the empirical analysis, it aims to find certain urban characteristics vulnerable to natural disasters in the Gyeongsangname-do areas. Finally, it is to provide a proposal for policy implementation to reduce damages by natural disasters in those areas.
In order to achieve the research purpose, this study is analyzing the amount of damages by natural disasters in the 18 areas of Gyeongsangnam-do from 1999 to 2012, and the urban characteristics of each period and area. The research period of 14 years reflects time-sequential changes in natural disaster occurrences and urban characteristics.
Damages by natural disasters are divided into the amount of whole damages and the amount of public damages for analysis. The amount of whole damages by natural disasters includes damages of buildings, ships, farmland, farm produce, and public facilities. In other words, the whole damages are the indicator of the actual damages in one area including both public and private sectors. The amount of public damages subtracts the private sectors from the whole damages. Damages to public facilities are restored with public budgets. Damages to the private sectors are basically restored with private insurances or self-funds despite some assistance from the public sectors. The more damaged public facilities are, the more amount of budgets their restoration requires, because the majority of such damages are restored with public budgets. While reducing the whole damages is important, mitigating the public damages, which are supposed to be restored with the taxes of the people, is essential. Therefore, this study divides damages by natural disasters into the amount of the whole damages and the amount of the public damages for analysis.
Previous researches on natural disaster damages in regards to urban characteristics are classified as either microscopic or macroscopic. From the microscopic view, researches were conducted on such surrounding areas as river basins and lakes with GIS-oriented analyses, rather than empirical ones. Most macroscopic researches carried out empirical analyses on various areas based on time-sequential data. Although all the previous researches include factors regarding urban characteristics, microscopic researches are liable to only take land use aspects into consideration. Also, they fail to clarify the association between land use characteristics and natural disaster damages because they are not empirical analyses. While microscopic, GIS-oriented work needs to be done to the certain extent, it seems that more researches are to follow, since microscopic researches do not provide detailed characteristic factors. This study resembles previous macroscopic researches. Similar to the microscopic ones, however, they do not reflect various urban characteristics into analyses. Also they are mainly centered on Gyeonggi-do and the metropolitan area. Since the metropolitan area has its buildings and population very concentrated, its aspects of damages by natural disasters differ from ones in other regions. Therefore, the research outcomes from the metropolitan area are limited to be directly applied to other regions. That is why this study aims on the southern Gyeongsangnam-do area, which is the second most concentrated area in terms of buildings and population.
The spatial autocorrelation between the whole damages and the public damages that occurred during the research period is analyzed in global and local aspects. Analysis on the spatial autocorrelation shows how one area affects its surrounding areas. This study postulates that an area with natural disaster damages would have a spatial dependency to analyze the spatial autocorrelation. The global spatial autocorrelation is illustrated by employing Moran`s I, and the local one is drawn from the LISA Cluster Map. In spite of the postulation, the autocorrelation between the whole damages and the public damages by years is not well-drawn, nor are the time-sequential effects. The yearly deviation is very drastic and, moreover, years with severe typhoons draw less autocorrelation.
The purpose of this study is to empirically analyze urban characteristics that affect the amount of whole damages and the amount of public damages by natural disasters from 1999 to 2012 on the 18 areas of Gyeongsangnam-do. For doing so, empirical analysis employes the Panel Models. The whole damages and the public damages are defined as dependent variables for each model, and seven independent variables are categorized by their attributes: 16 variables in total. In addtion, comparative analysis is conducted to determine the most appropriate Panel Model for empirical analysis among the Pooled OLS Model, the One-way Fixed Effect Model, the One-way Random Effect Model, the Two-way Fixed Effect Model, and the Two-way Random Effect Model. As a result, the Two-way Fixed Effect Model turns out to be most appropriate for both the whole damages and the public damages.
When it comes to analysis on the whole damages, the surface area of residential zones, the surface areas of green zones, and the number of industrial employees are estimated to have a negative (-) effect, becoming damage-reducing factors. On the other hand, the number of apartment buildings, the number of building permissions, and the road density are estimated to have a positive (+) effect, becoming damage-adding factors. When it comes to analysis on the public damages, the budget scale, the surface area of residental zones, the surface area of green zones, and the number of industrial employees are estimated to have a negative (-) effect, becoming damage-reducing factors. On the other hand, the number of apartment buildings and the road density are estimated to have a positive (+) effect, becoming damage-adding factors.
As a result of empirical analysis, proper policy establishment and implementation about the damage-adding factors is needed in order to reduce the amount of damage increases by those factors in the future. At the same time, policies in response to each factor is necessary for mitigation in Gyeongsangnam-do. In addition, rivers and reservoirs are not estimated to have an effect to natural disaster damages, which means that the two types of prevention facility are not performing well; research needs to be done on this. Last but not least, in order to facilitate the methods of responding to natural disasters to be effective, strengthening the organizational and systemic connection of the local governments and the government is required.