Dynamics of urban fire correlations with detrended fluctuation analysis

Abstract

To explore and understand the intrinsic dynamic mechanisms of urban fire system is an important prerequisite for the fire risk analysis and management. By analyzing the time series of urban fire system, it is helpful to achieve this goal. In this paper, the detrended fluctuation analysis (DFA) is applied to study the long-range correlations of urban fires in both small and large spatial scales with a case of Chinese city. It is found that the time series of direct economic loss caused by urban fires presents stable anti-power law correlations, while the time interval series of urban fire occurrence presents persistent long-range power-law correlations with two scaling exponents. Based on the characteristics of these scaling exponents, the intrinsic dynamics of urban fires is discussed. It is found that the effect of human activities on the dynamics of urban fire system may become prevalent at small timescales, while at large timescales the effect of human activities is not obvious in influencing the long-range correlation patterns of urban fire system, which may be controlled mainly by some general characteristics of a city such as the meteo-climatic conditions.