HR: 10:35h
AN: H21G-02    [PDF]
TI: Multifractals and the Temporal Structure of Precipitation: Where are we Standing, What can we Expect
AU: * de Lima, M P
EM: lima@dec.uc.pt
AF: IMAR/Coimbra Polytechnic Inst., Bencanta, Coimbra, 3040-316 Portugal
AB: Precipitation is the driving agent of many other processes. Its temporal and spatial variability are important issues in many studies and areas of research. However, information on the amount and distribution of precipitation in space and time is often restricted because of its strong temporal and spatial variation. Therefore, hydrological models have usually to conceptualize processes based on simple, often homogeneous, approximations of nature (e.g. precipitation is expressed as depths over periods of a day). Such conceptualizations often lack sufficient temporal and spatial resolutions to permit a detailed modeling of complex hydrological processes. Moreover, empirical scale truncations are made often, and one scale is studied independently of the others. Thus, it is pertinent to know whether there are intrinsically different phenomena as one moves from one scale to the next; and whether results obtained on one scale can be "transported" to the other. Scale-invariant studies of precipitation are being successful in clarifying the spatial and temporal structure of precipitation and in quantifying the variability in this process. In particular, multifractal theory can be used to better understand the strongly irregular fluctuations of precipitation, which are manifested over a broad range of scales, because it has the potential to assess the full range of precipitation dynamics. The prototypical multifractal process is a cascade: the extreme small-scale variability is precisely the consequence of the wide range of scales over which it can build up. This study explores the invariance of properties manifested across scales and determines the multifractal behavior of precipitation in time, using data from different origins. The statistical properties of precipitation are characterized by empirical multifractal exponent functions describing the scaling of the probability distributions and moments of the precipitation intensity. The description of these functions using a multifractal model based on Levy random variables is investigated. Special attention is given to using multifractal innovative tools for assessing the probability of occurrence of extreme events. Results can contribute to improve data collection, with respect to the required temporal sampling resolution, and the generation of high-resolution synthetic precipitation data.
DE: 1800 HYDROLOGY
DE: 1854 Precipitation (3354)
DE: 3210 Modeling
DE: 3220 Nonlinear dynamics
DE: 3250 Fractals and multifractals
SC: Hydrology [H]
MN: 2003 Fall Meeting