![Hier klicken, um den Treffer aus der Auswahl zu entfernen](images/unchecked.gif) |
Titel |
Return interval statistics in precipitation and river flow records |
VerfasserIn |
M. I. Bogachev, A. Bunde |
Konferenz |
EGU General Assembly 2009
|
Medientyp |
Artikel
|
Sprache |
Englisch
|
Digitales Dokument |
PDF |
Erschienen |
In: GRA - Volume 11 (2009) |
Datensatznummer |
250031167
|
|
|
|
Zusammenfassung |
We study the return intervals between events exceeding a certain threshold Q in precipitation
and river flow records from 32 stations all over the world. Using multifractal DFA for the
original and for the shuffled data, we find that the precipitation records are only slightly
long-term correlated with Hurst exponents h(2) typically around 0.55, and exhibit only weak
nonlinear memory. Due to the weak linear and nonlinear memory, the distribution
of the return intervals is only slightly broader than the exponential distribution
characterizing a random Poisson process. For river flows, linear correlations are pronounced
and characterized by h(2)-0.8, and nonlinear memory is apparent. While in the
precipitation data the return interval distributions approximately scale for different
thresholds Q, in the river flows there is a slight violation of scaling with increasing
Q, due to the nonlinear memory. We model both the precipitation and the river
flows by a multiplicative cascade model, and obtain substantial agreement in the
return interval statistics between the observational and in the simulated data. In both
cases the distributions of the return intervals can be approximated by the gamma
distribution P = [λ α-Π(α)]rα-1exp(- λr)with α = λ -2/3 for the precipitation
records and α = λ -0.1 for the river flow records, which for the precipitation records
resembles the return interval distribution in seismic records proposed in [Corral 2004]. |
|
|
|
|
|