Regionalization of hydrological model parameters using data depth

The parameters of hydrological models with no or short discharge records can only be estimated using regional information. We can assume that catchments with similar characteristics show a similar hydrological behaviour. A regionalization of hydrological model parameters on the basis of catchment characteristics is therefore plausible. However, due to the non-uniqueness of the rainfall/runoff model parameters (equifinality), a procedure of a regional parameter estimation by model calibration and a subsequent fit of a regional function is not appropriate. In this paper, a different procedure based on the depth function and convex combinations of model parameters is introduced. Catchment characteristics to be used for regionalization can be identified by the same procedure. Regionalization is then performed using different approaches: multiple linear regression using the deepest parameter sets and convex combinations. The assessment of the quality of the regionalized models is also discussed. An example of 28 British catchments illustrates the methodology.

[1]  Peter Rousseeuw,et al.  Computing location depth and regression depth in higher dimensions , 1998, Stat. Comput..

[2]  J. Tukey.  Mathematics and the Picturing of Data , 1975 .

[3]  Luis Samaniego,et al.  Simulation of the impacts of land use/cover and climatic changes on the runoff characteristics at the mesoscale , 2006 .

[4]  Günter Blöschl,et al.  A comparison of regionalisation methods for catchment model parameters , 2005 .

[5]  Alison L. Kay,et al.  A comparison of three approaches to spatial generalization of rainfall–runoff models , 2006 .

[6]  A. Bárdossy,et al.  Robust estimation of hydrological model parameters , 2008 .

[7]  Taha B. M. J. Ouarda,et al.  Statistical Models and the Estimation of Low Flows , 2008 .

[8]  R. Moore.  The probability-distributed principle and runoff production at point and basin scales , 1985 .

[9]  András Bárdossy,et al.  Comparison of four regionalisation methods for a distributed hydrological model , 2007 .

[10]  Günter Blöschl,et al.  Regionalisation of catchment model parameters , 2004 .

[11]  András Bárdossy,et al.  Modeling of the effect of land use changes on the runoff generation of a river basin through parameter regionalization of a watershed model , 2004 .

[12]  R. Serfling,et al.  General notions of statistical depth function , 2000 .

[13]  Taha B. M. J. Ouarda,et al.  Estimation of extreme flow quantiles and quantile uncertainty for ungauged catchments. , 2007 .

[14]  Hoshin Vijai Gupta,et al.  Regionalization of constraints on expected watershed response behavior for improved predictions in ungauged basins , 2007 .

[15]  Rob Lamb,et al.  Confidence intervals for a spatially generalized, continuous simulation flood frequency model for Great Britain , 2004 .

[16]  Regina Y. Liu,et al.  Multivariate analysis by data depth: descriptive statistics, graphics and inference, (with discussion and a rejoinder by Liu and Singh) , 1999 .

[17]  Soroosh Sorooshian,et al.  Toward improved streamflow forecasts: value of semidistributed modeling , 2001 .

[18]  Soroosh Sorooshian,et al.  A framework for development and application of hydrological models , 2001, Hydrology and Earth System Sciences.