The actual situation of population distribution in cities varies greatly depending on the date and time, and the earthquake occurrence time can make a big difference to the extent of damage. As estimated by the national government and the local municipality, the earthquake occurrence times follow three patterns: (1) winter at 5 p.m., (2) summer at noon, and (3) winter at 6 p.m. Such damage estimations consider data from a census conducted once every five years as demographics. Although the reliability is high, the problem is that the findings cannot reflect the recent population situation, consider the trend of population by the day of the week, or reflect a realistic picture of the daytime population dynamics because these are statistics pertaining to a night population. NTT DOCOMO has developed a new technique to determine population statistics called Mobile Spatial Statistics (MSS), which estimates the actual population based on mobile phone operation data. It can serve as a tool to understand the population distribution of cities with rapid fluctuations. However, there are few studies that focus on understanding the population dynamics of a single day at times of normalcy and clarifying issues on local disaster response using MSS. Therefore, this study indicates the possibility of the application of MSS in times of normalcy to earthquake damage estimation and disaster prevention plans.

 This study consists of four parts. The first part reveals the features and accuracy of MSS by comparing them with population census data. In Nagoya city area, MSS at 4 a.m. on a weekday exceeded the census population data in all ages because about 10% of the MSS’s population comprised visitors from other municipalities. However, the excess rate was 4% to 20%, thereby showing that the MSS maintains accuracy.

 The second part elucidates the local disaster characteristics of Nagoya city by combining the classification of daily population trends and building collapse risk on a weekday (Wednesday, July 12, 2017) and on a holiday (Saturday, June 3, 2017), Classifications into five characteristics based on daily population trends were performed using cluster analysis by MSS on the population within a 1 km square. By combining those classification results with building collapse risk, it is clear that there are differences in the areas and time zones in which the risk of disaster increases.

 The third part, with regard to population inflow areas by classification results, examines the risk of victims who are unable to return home. A past estimation used the Person Trip Survey (PTS), which has been conducted once every 10 years on weekdays in each city in Japan. This study estimates the number of victims unable to return home by using MSS. Therefore, it considers not only the latest population data but also that of the holidays. Using it together with the PTS could lead to formulating detailed measures.

 The fourth part analyzes changes in local disaster response capability based on changing trends in the aging rate and the composition of fire brigade volunteers in population outflow areas by classification results. In particular, during daytime on weekdays, there are areas and time zones in which disaster risk will increase because the proportion of elderly people who require assistance during a disaster increases, and the fire brigade often work outside of the area. A series of analysis results using MSS can effectively serve as basic data for earthquake damage estimation and disaster prevention plans considering the day of the week and time.