Article

Uncertainty and multiple objective calibration in regional water balance modeling – Case study in 320 Austrian catchments

Institute of Hydraulic Engineering and Water Resources Management, Vienna University of Technology, Wien, Vienna, Austria
Hydrological Processes (Impact Factor: 2.68). 02/2007; 21(4):435 - 446. DOI: 10.1002/hyp.6253

ABSTRACT

We examine the value of additional information in multiple objective calibration in terms of model performance and parameter uncertainty. We calibrate and validate a semi-distributed conceptual catchment model for two 11-year periods in 320 Austrian catchments and test three approaches of parameter calibration: (a) traditional single objective calibration (SINGLE) on daily runoff; (b) multiple objective calibration (MULTI) using daily runoff and snow cover data; (c) multiple objective calibration (APRIORI) that incorporates an a priori expert guess about the parameter distribution as additional information to runoff and snow cover data. Results indicate that the MULTI approach performs slightly poorer than the SINGLE approach in terms of runoff simulations, but significantly better in terms of snow cover simulations. The APRIORI approach is essentially as good as the SINGLE approach in terms of runoff simulations but is slightly poorer than the MULTI approach in terms of snow cover simulations. An analysis of the parameter uncertainty indicates that the MULTI approach significantly decreases the uncertainty of the model parameters related to snow processes but does not decrease the uncertainty of other model parameters as compared to the SINGLE case. The APRIORI approach tends to decrease the uncertainty of all model parameters as compared to the SINGLE case. Copyright © 2006 John Wiley & Sons, Ltd.

Download full-text

Full-text

Available from: Juraj Parajka
    • "Several studies have shown that additional calibration criteria, such as mass balance data, lead to significantly better simulations of the additional water balance terms, but simultaneously to slightly poorer runoff simulations (Seibert 2000; Madsen 2003; Parajka et al. 2007; Schaefli and Huss 2011; Hagg et al. 2013; Mayr et al. 2013). In remote mountain areas, these data are sparse. "
    [Show abstract] [Hide abstract]
    ABSTRACT: We developed a spatially distributed conceptual melt and runoff model able to cope with debris covered glacier parts. Subdebris melt was calculated based on empirical equations derived from field observations. The model was applied at the Inylchek glaciers. While runoff from Southern Inylchek flows directly into Inylchek River, runoff from Northern Inylchek is blocked by Southern Inylchek and forms Lake Merzbacher. This lake is drained frequently by outburst floods. To calibrate the model, a multi-algorithm, genetically adaptive multi-objective method (AMALGAM) was used. Daily runoff observations from Inylchek River as well as glacier mass balance gradients from other glaciers in the Tian Shan were used as objective functions in three split samples. Runoff simulations yielded Nash–Sutcliffe coefficients mostly higher than 0.9, and the mass balance gradients were simulated in an acceptable range. The results of the different split samples vary considerably. Considering the used objective functions and additional data such as the equilibrium line altitude, we chose the most promising parameter set to calculate the filling of Lake Merzbacher. This analysis indicated answers to several questions concerning the hydrological system of the lake. Apart from calving at the ice dam formed by Southern Inylchek glacier, the lake seemed so be filled predominantly by meltwater from Northern Inylchek glacier. The lake has a permanent runoff, also during the filling process. After the outburst, the channels seemed to be closed quickly.
    No preview · Article · Sep 2014 · Geografiska Annaler Series A Physical Geography
  • Source
    • "Once these relationships are established for available gauged catchments, the values of the hydrologic model parameters in ungauged catchments can be estimated based on its physical/climatic characteristics. Strategies based on parameter regionalization have been criticized for ignoring the covariances among model parameter estimates (McIntyre et al., 2005; Bárdossy, 2007; Parajka et al., 2007; Oudin et al., 2008), and are restricted due to parameter identifiability issues (Beven and Freer, 2001) and model structural error (Wagener and Wheater, 2006). (ii) Constraining hydrologic model simulations by regionalized signatures – When regionalizing model parameters, one is invariably faced with problems such as the equifinality of http://dx.doi.org/10.1016/j.jhydrol.2014.06.030 0022-1694/Ó 2014 Elsevier B.V. All rights reserved. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Large uncertainties in streamflow projections derived from downscaled climate projections of precipitation and temperature can render such simulations of limited value for decision making in the context of water resources management. New approaches are being sought to provide decision makers with robust information in the face of such large uncertainties. We present an alternative approach that starts with the stakeholder's definition of vulnerable ranges for relevant hydrologic indicators. Then, the modeled system is analyzed to assess under what conditions these thresholds are exceeded. The space of possible climates and land use combinations for a watershed is explored to isolate sub-spaces that lead to vulnerability, while considering model parameter uncertainty in the analysis. We implement this concept using classification and regression trees (CART) that separate the input space of climate and land use change into those combinations that lead to vulnerability and those that do not. We test our method in a Pennsylvania watershed for nine ecological and water resources related streamflow indicators for which an increase in temperature between 3°C to 6 °C and change in precipitation between -17% and 19% is projected. Our approach provides several new insights, for example we show that even small decreases in precipitation (~5%) combined with temperature increases greater than 2.5ºC can push the mean annual runoff into a slightly vulnerable regime. Using this impact and stakeholder driven strategy, we explore the decision-relevant space more fully and provide information to the decision maker even if climate change projections are ambiguous.
    Full-text · Article · Apr 2014
    • "Once these relationships are established for available gauged catchments, the values of the hydrologic model parameters in ungauged catchments can be estimated based on its physical/climatic characteristics. Strategies based on parameter regionalization have been criticized for ignoring the covariances among model parameter estimates (McIntyre et al., 2005; Bárdossy, 2007; Parajka et al., 2007; Oudin et al., 2008), and are restricted due to parameter identifiability issues (Beven and Freer, 2001) and model structural error (Wagener and Wheater, 2006). (ii) Constraining hydrologic model simulations by regionalized signatures – When regionalizing model parameters, one is invariably faced with problems such as the equifinality of http://dx.doi.org/10.1016/j.jhydrol.2014.06.030 0022-1694/Ó 2014 Elsevier B.V. All rights reserved. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Summary Daily streamflow information is critical for solving various hydrologic problems, though observations of continuous streamflow for model calibration are available at only a small fraction of the world’s rivers. One approach to estimate daily streamflow at an ungauged location is to transfer rainfall–runoff model parameters calibrated at a gauged (donor) catchment to an ungauged (receiver) catchment of interest. Central to this approach is the selection of a hydrologically similar donor. No single metric or set of metrics of hydrologic similarity have been demonstrated to consistently select a suitable donor catchment. We design an experiment to diagnose the dominant controls on successful hydrologic model parameter transfer. We calibrate a lumped rainfall–runoff model to 83 stream gauges across the United States. All locations are USGS reference gauges with minimal human influence. Parameter sets from the calibrated models are then transferred to each of the other catchments and the performance of the transferred parameters is assessed. This transfer experiment is carried out both at the scale of the entire US and then for six geographic regions. We use classification and regression tree (CART) analysis to determine the relationship between catchment similarity and performance of transferred parameters. Similarity is defined using physical/climatic catchment characteristics, as well as streamflow response characteristics (signatures such as baseflow index and runoff ratio). Across the entire US, successful parameter transfer is governed by similarity in elevation and climate, and high similarity in streamflow signatures. Controls vary for different geographic regions though. Geology followed by drainage, topography and climate constitute the dominant similarity metrics in forested eastern mountains and plateaus, whereas agricultural land use relates most strongly with successful parameter transfer in the humid plains.
    No preview · Article · Jan 2014 · Journal of Hydrology
Show more