HR: 1330h
AN: H12B-0972 [PDF]
TI: Fractal structure of multivariate drought events recurrence analyses
AU: * Gonzalez Perez, J
EM: Javier.Gonzalez@uclm.es
AF: University of Castilla-La Mancha, Dept. of Civil Engineering
Avda Camilo Cela s/n, Ciudad Real, 13071
Spain
AU: Valdes, J B
EM: jvaldes@u.arizona.edu
AF: University of Arizona, Dept. of Civil Engineering and SAHRA, Tucson, AZ 85721-0072 United States
AB:
Droughts are events characterized by their prolonged persistence in time as a succession of low water availability indicator
values. Derived properties such as duration, severity, or maximum intensity are used to analyze and characterize the events.
However, the use of one or two of these properties is not always enough to fully perform a drought recurrence analysis over
the period of record. Each event may have different characteristics that describe its exceptional aspect.
In this work droughts are treated as multivariate events, using the direct sequence of recorded values in the analysis.
Therefore all drought aspects are taken into account in the analysis. The mean frequency of recurrence (MFR) of such sequence
is used to characterize each of the drought events. The MFR is computed by first estimating the mean frequency of recurrence
considering event of a fixed duration w (MFRw). Second the result is generalized by considering the occurrence possibility
of events of any duration by an embedding process. Analytical expressions of the MFRw were found in the form of an infinite
series. To facilitate the computation this analytical expression was simplified by an approximate 3-parameters function. The
parameters of this approximate function are related with the duration w by power law relationships. Furthermore, the
relationship between MFR and MFRw follows also a power law, suggesting a fractal structure of the phenomenon. This allows
explaining how recurrence properties of extreme events vary with the temporal scale used in the analysis.
The droughts characterization based on this approach generates a drought index, the Drought Frequency Index (DFI). The index
is compared with existing indexes in several case studies.
DE: 1812 Drought
DE: 1833 Hydroclimatology
DE: 1854 Precipitation (3354)
DE: 1869 Stochastic processes
SC: Hydrology [H]
MN: 2003 Fall Meeting