Premium
Model identification by space–time disaggregation: a case study from eastern Australia
Author(s) -
Wooldridge S. A.,
Kalma J. D.,
Franks S. W.,
Kuczera G.
Publication year - 2002
Publication title -
hydrological processes
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.222
H-Index - 161
eISSN - 1099-1085
pISSN - 0885-6087
DOI - 10.1002/hyp.329
Subject(s) - surface runoff , scale (ratio) , drainage basin , environmental science , representation (politics) , identification (biology) , hydrology (agriculture) , el niño southern oscillation , climatology , temporal scales , computer science , meteorology , geography , ecology , geology , cartography , geotechnical engineering , politics , political science , law , biology
Abstract In this paper, a disaggregation approach is suggested for the task of modelling hydrological responses within a spatially and temporally variable environment. With such an approach, large‐scale environmental characteristics are tested for their ability to provide insight into the dominant physical mechanisms responsible for observed catchment responses. Using a regional‐scale catchment in eastern Australia as a case study, the approach is firstly used to determine the utility of physical catchment data, and its organization in space, to provide insight into the compartmentalization of soil water storage within the catchment. In a second application, temporal disaggregation of the rainfall‐runoff record into the cold‐wet and warm‐dry phases of the El Niño/Southern Oscillation (ENSO) phenomenon is utilized to provide an objective comparison between alternative model structures, based on the ‘consistency’ of model parameters in describing the effect of ENSO phase on water yield. Finally, combining the improved spatial representation of hydrological response with the model structure identified by temporal analysis is shown to result in a predictive framework whose level of uncertainty is lower than either of the individual strategies, and whose responses are consistent with the available evidence. It is noted that such modelling insight is unlikely to have been gained with traditional modelling strategies that seek to force a predetermined model structure to ‘fit’ the observed data. Copyright © 2002 John Wiley & Sons, Ltd.