Assessing temporal-dependent correlations in the 2000–2008 Popocatepetl exhalation sequence by using detrended fluctuation analysis

Large eruptions may represent a serious threat for the people dwelling near a volcano. The analysis of volcanic time-series can provide important insights for the assessment of risk and hazards and for the proposal of civil protection strategies. Different statistical methods have been used to characterize time-series associated with broadband measures of volcanic activity, suggesting that eruptive events may not be independent. While most analysis efforts have focused on traditional measures, such as seismic signals and duration and intensity of large eruptions, little attention has been devoted to the time-series of small-size daily eruptions known as exhalations. Some studies have indicated that exhalations, as part of the background activity monitoring, are correlated in time with long-period seismic events and tremors. Hence, the analysis of exhalation dynamics should yield some insights on the activity of a volcano. This work uses detrended fluctuation analysis, a statistical method from nonlinear analysis, for quantifying correlations in the 2000–2008 exhalation time-series of the Popocatepetl volcano, Mexico. The results indicated (1) the presence of long-term correlations, suggesting that the exhalation events are not independent but evolve according to a fractal persistent behavior, (2) the existence of two quasi-periodic cycles (with approximate periods of about 1.2 and 0.22 years) in the correlations temporal dependence, and (3) the action of nonlinear effects measured in terms of a multifractality index. The results are contrasted and discussed in terms of two recent largest eruptive phases of the Popocatepetl.