 |
| Titel |
Long-range correlations in the fire sequences with Detrended Fluctuation Analysis |
| VerfasserIn |
H. Zheng, W. Song |
| Konferenz |
EGU General Assembly 2009
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 11 (2009) |
| Datensatznummer |
250020331
|
|
|
|
|
|
| Zusammenfassung |
| Forest fires have been found to exhibit good power-law relation in the frequency-size
distribution over many orders of magnitude in different countries, which identifies that forest
fires behave as self-organized criticality (SOC). And in the temporal aspect, it is also found
that the frequency-interval distributions of fires obey power-law with periodic fluctuations.
The fire sequences cannot generally be described as Poisson point process, because the
distribution of the occurrence times is not homogeneous and shows a clustering behavior. So
the power-law distributions, the scaling behavior of the parameters are usually used to
describe the sequence. Inter-event time series, the waiting-time between consecutive events,
were studied in the similar earthquakes system in recent years, focusing on the distributions
and the intrinsic mechanism. In order to find the long-range correlations of fire sequences, we
analyzed the scaling behavior of the fires occurred in some places of Asia by means of the
detrended fluctuation analysis (DFA), which provides the information of the scaling
behavior and long-range characteristics in non-stationary time series. The scaling
exponents, larger than 0.5, indicate the presence of persistent long-range correlations,
while it performs white noise at 0.5. The detail fire data were investigated in several
places, and with the different thresholds of the burned areas or losses. The result
reveals the existence of long-range correlations in the fire interval sequences, and
the scaling exponents are quite constant over several orders of magnitude. But the
exponents are different from each other, possibly due to the orientation of the places we
analyzed and other local influencing factors: human activity, weather, economic etc.
Besides, the fire sequences of different types were studied in the same way, to find out
the possible different long-range behaviors and their possible reasons. The results
seem to be helpful to understand the underlying dynamics of the fire sequences. |
|
|
|
|
|
|