Loading…

Partial spectral analysis of hydrological time series

► Ambiguity with respect to processes encoded in spectral density is resolved. ► Approach is based on spectral representation of partial correlation. ► Antecedent and subsequent influences of controlling processes are distinguished. ► Influences of air temperature and relative humidity on discharge...

Full description

Saved in:
Bibliographic Details
Published in:Journal of hydrology (Amsterdam) 2011-03, Vol.400 (1), p.223-233
Main Authors: JUKIC, D, DENIC-JUKIC, V
Format: Article
Language:English
Subjects:
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:► Ambiguity with respect to processes encoded in spectral density is resolved. ► Approach is based on spectral representation of partial correlation. ► Antecedent and subsequent influences of controlling processes are distinguished. ► Influences of air temperature and relative humidity on discharge are analyzed. Hydrological time series comprise the influences of numerous processes involved in the transfer of water in hydrological cycle. It implies that an ambiguity with respect to the processes encoded in spectral and cross-spectral density functions exists. Previous studies have not paid attention adequately to this issue. Spectral and cross-spectral density functions represent the Fourier transforms of auto-covariance and cross-covariance functions. Using this basic property, the ambiguity is resolved by applying a novel approach based on the spectral representation of partial correlation. Mathematical background for partial spectral density, partial amplitude and partial phase functions is presented. The proposed functions yield the estimates of spectral density, amplitude and phase that are not affected by a controlling process. If an input–output relation is the subject of interest, antecedent and subsequent influences of the controlling process can be distinguished considering the input event as a referent point. The method is used for analyses of the relations between the rainfall, air temperature and relative humidity, as well as the influences of air temperature and relative humidity on the discharge from karst spring. Time series are collected in the catchment of the Jadro Spring located in the Dinaric karst area of Croatia.
ISSN:0022-1694
1879-2707
DOI:10.1016/j.jhydrol.2011.01.044