This study analyzes data on households'

No abstract

In this study, we examined setting up patterns of non-designated emergency shelters at the time of the Great East Japan Earthquake in Kamaishi City through analysis using GIS and interview surveys. Analysis result showed following patterns of non-designated evacuation shelters: 1. setting up around outside of the inundated area, 2. setting up in the area where the secondary evacuation were restricted by blocked routes , 3. setting up in the area where the ratio of vulnerable people with restricted ability to evacuate is high, 4. setting up near the intersection of secondary evacuation routes from other area, 5. setting up around inundated designated evacuation shelters, 6. setting up around an overcrowded designated evacuation shelter, 7. setting up using public interest facilities and 8. setting up using private facilities and houses in the area where the public facilities were not exist.

High levels of amplitude acceleration ( 300 to 800 cmis2 ) and velocity ( 80 cmis ) were observed over a wide area at the time of the 1995 Great Hanshin-Awaji Earthquake. The locations where the fires started are almost uniformly distributed in areas where buildings suffered severe structural damage, mainly areas where the seismic intensity reached seven on the JMA (Japan Meteorological Agency ) scale. Thanks to the weak winds at the time of the earthquake, the speed of the fire spread was far below that of past urban fires. Seven large fires that destroyed an area of more than 33,000 m2 each, broke out in areas with large concentrations of wooden houses. Most of the fireprotected wooden buildings that caught fire did so mainly because fire spread through the building's openings, not through their walls. Most of the cease-burning lines in urban areas were wide streets, railways, parks and similar large spaces, rows of buildings of fireproof construction, and fire proof buildings constructed on large building lots. This confirms that it is extremely important to provide more space between buildings and improve buildings' fireproof performance in order to prevent the spread of urban fires following earthquakes.

The smoke movement model developed at the National Research Council of Canada (NRCC) as part of its risk-cost assessment model is a simple computer model for calculating the probabilities of smoke hazard in multi-storey buildings. To validate this model, full-scale smoke movement experiments were conducted at the 10-storey experimental smoke tower at the National Fire Laboratory of the NRCC. The 10-storey tower was designed and built to study smoke movements in highrise buildings. Results of the experiments show good agreement between the data and the simple smoke movement model.