Abstract
Aim: This is the third in a series of four articles entitled “Assessment of Safety Level Based on the State Fire Service Statistics (SFS)”. The paper refers to the experimental research on spatial-temporal analysis of residential building fires that occurred in Warsaw. The temporal analysis means that the author described percentage distribution of residential building fires by categories of hours, days of week, months and quarters. Spatial characteristics shoud be understood as the techniques of residential building fires hazard maps development proven by the research and discussion on its results.
Introduction: Research on the frequency of fires, local emergencies and injuries in residential buildings class definitely demonstrates that the issue of risks in residential buildings is very important for the emergency services [1], [4]. The introduction includes the essential conclusions from the previous papers on the assessment of safety level in the aspect of fires [1] and local emergencies [4]. According to the State Fire Service statistics for 2007-2012, residential fires constituted 17% of all fires. Injury (75%) and fatality rates (81%) related to this class of buildings are a few dozen times higher than in case of other classes [1]. The percentage of local emergencies emerging in residential buildings amounts to 31% of all local emergencies with the coefficients of injuries and fatalities constituting 8% and 22% respectively [4]. In addition, the introduction contains the most important postulates from the literature review of 14 selected English language papers on fire hazards identification, estimation and assessment methods, that motivated the author to carry out the research and write this article.
Methodology: Research includes a selection of detached house fires and blocks of flats fires that had been registered in the National Headquarters of State Fire Service database between 2000 and 2012. The paper includes summaries on the quality of direct spatial reference (coordinates) and indirect spatial reference (address data) found in the fire reports. The methodology of geo-processing fire statistic into point class vector map that is suitable in geographical information system (GIS) is described too. GIS analysis of temporal characteristics of residential fires by the way of classification of the events by hours, days of the week and months is presented. The author compiled house fire hazard maps on the basis of 2 methods. The first one shows classification of cartography grid based on polyline estate vector map, where each cell’s grid has a parameter of the number of residential fires. The second one based on residential fire density vector map created by Kernel Density Estimation mapping method.
Conclusions: It is underlined that residential buildings fires are characterised by spatial and temporal dependencies. The number of residential buildings fires rises between 7-20 o`clock, and is peaking at 18-20. Then the number declines between 21-4. It is proved that the number of house fires is the highest in November, December, January and during the time of holidays such as June, July and August. When the density of dwelling estates rises, the residential fire hazard rises too. One of the hint in the field of residential building hazard maps is to implement KDE GIS method. The method is based on the density of point class vector map (e.g. fires) and it does not involve administrative boundaries. The better coordinates and address data, the better quality of the hazard maps. It is suggested to fill in the emergency calls forms and fire reports with the address data from the registry (if possible). There is an example of registry called TERYT [24] that is kept in SWDST, Polish State Fire Service system.
Keywords: fires, fire safety, fire hazards, housing fires, fire statistics, fire hazards maps, geographical information system, GIS, Kernel Density Estimation, KDE, temporal fire characteristic, spatial fire characteristic
Type of article: original scientific article