On characterizing the temporal dominance patterns of model parameters and processes

Björn Guse, Matthias Pfannerstill, Michael Strauch, Dominik E. Reusser, Stefan Lüdtke, Martin Volk, Hoshin Vijai Gupta, Nicola Fohrer

Research output: Contribution to journalArticle

21 Citations (Scopus)

Abstract

Diagnostic analyses of hydrological models intend to improve the understanding of how processes and their dynamics are represented in models. Temporal patterns of parameter dominance could be precisely characterized with a temporally resolved parameter sensitivity analysis. In this way, the discharge conditions are characterized, that lead to a parameter dominance in the model. To achieve this, the analysis of temporal dynamics in parameter sensitivity is enhanced by including additional information in a three-tiered framework on different aggregation levels. Firstly, temporal dynamics of parameter sensitivity provide daily time series of their sensitivities to detect variations in the dominance of model parameters. Secondly, the daily sensitivities are related to the flow duration curve (FDC) to emphasize high sensitivities of model parameters in relation to specific discharge magnitudes. Thirdly, parameter sensitivities are monthly averaged separately for five segments of the FDC to detect typical patterns of parameter dominances for different discharge magnitudes. The three methodical steps are applied on two contrasting catchments (upland and lowland catchment) to demonstrate how the temporal patterns of parameter dynamics represent different hydrological regimes. The discharge dynamic in the lowland catchment is controlled by groundwater parameters for all discharge magnitudes. In contrast, different processes are relevant in the upland catchment, because the dominances of parameters from fast and slow runoff components in the upland catchment are changing over the year for the different discharge magnitudes. The joined interpretation of these three diagnostic steps provides deeper insights of how model parameters represent hydrological dynamics in models for different discharge magnitudes. Thus, this diagnostic framework leads to a better characterization of model parameters and their temporal dynamics and helps to understand the process behaviour in hydrological models.

Original languageEnglish (US)
JournalHydrological Processes
DOIs
StateAccepted/In press - 2016

Fingerprint

catchment
parameter
hydrological regime
sensitivity analysis
time series
runoff
groundwater

Keywords

  • Catchment modelling
  • Diagnostic model analysis
  • Model understanding
  • Parameter characterization
  • Temporal parameter sensitivity analysis

ASJC Scopus subject areas

  • Water Science and Technology

Cite this

Guse, B., Pfannerstill, M., Strauch, M., Reusser, D. E., Lüdtke, S., Volk, M., ... Fohrer, N. (Accepted/In press). On characterizing the temporal dominance patterns of model parameters and processes. Hydrological Processes. https://doi.org/10.1002/hyp.10764

On characterizing the temporal dominance patterns of model parameters and processes. / Guse, Björn; Pfannerstill, Matthias; Strauch, Michael; Reusser, Dominik E.; Lüdtke, Stefan; Volk, Martin; Gupta, Hoshin Vijai; Fohrer, Nicola.

In: Hydrological Processes, 2016.

Research output: Contribution to journalArticle

Guse, Björn ; Pfannerstill, Matthias ; Strauch, Michael ; Reusser, Dominik E. ; Lüdtke, Stefan ; Volk, Martin ; Gupta, Hoshin Vijai ; Fohrer, Nicola. / On characterizing the temporal dominance patterns of model parameters and processes. In: Hydrological Processes. 2016.
@article{0c621ba149f0433082138f3caf21a1c5,
title = "On characterizing the temporal dominance patterns of model parameters and processes",
abstract = "Diagnostic analyses of hydrological models intend to improve the understanding of how processes and their dynamics are represented in models. Temporal patterns of parameter dominance could be precisely characterized with a temporally resolved parameter sensitivity analysis. In this way, the discharge conditions are characterized, that lead to a parameter dominance in the model. To achieve this, the analysis of temporal dynamics in parameter sensitivity is enhanced by including additional information in a three-tiered framework on different aggregation levels. Firstly, temporal dynamics of parameter sensitivity provide daily time series of their sensitivities to detect variations in the dominance of model parameters. Secondly, the daily sensitivities are related to the flow duration curve (FDC) to emphasize high sensitivities of model parameters in relation to specific discharge magnitudes. Thirdly, parameter sensitivities are monthly averaged separately for five segments of the FDC to detect typical patterns of parameter dominances for different discharge magnitudes. The three methodical steps are applied on two contrasting catchments (upland and lowland catchment) to demonstrate how the temporal patterns of parameter dynamics represent different hydrological regimes. The discharge dynamic in the lowland catchment is controlled by groundwater parameters for all discharge magnitudes. In contrast, different processes are relevant in the upland catchment, because the dominances of parameters from fast and slow runoff components in the upland catchment are changing over the year for the different discharge magnitudes. The joined interpretation of these three diagnostic steps provides deeper insights of how model parameters represent hydrological dynamics in models for different discharge magnitudes. Thus, this diagnostic framework leads to a better characterization of model parameters and their temporal dynamics and helps to understand the process behaviour in hydrological models.",
keywords = "Catchment modelling, Diagnostic model analysis, Model understanding, Parameter characterization, Temporal parameter sensitivity analysis",
author = "Bj{\"o}rn Guse and Matthias Pfannerstill and Michael Strauch and Reusser, {Dominik E.} and Stefan L{\"u}dtke and Martin Volk and Gupta, {Hoshin Vijai} and Nicola Fohrer",
year = "2016",
doi = "10.1002/hyp.10764",
language = "English (US)",
journal = "Hydrological Processes",
issn = "0885-6087",
publisher = "John Wiley and Sons Ltd",

}

TY - JOUR

T1 - On characterizing the temporal dominance patterns of model parameters and processes

AU - Guse, Björn

AU - Pfannerstill, Matthias

AU - Strauch, Michael

AU - Reusser, Dominik E.

AU - Lüdtke, Stefan

AU - Volk, Martin

AU - Gupta, Hoshin Vijai

AU - Fohrer, Nicola

PY - 2016

Y1 - 2016

N2 - Diagnostic analyses of hydrological models intend to improve the understanding of how processes and their dynamics are represented in models. Temporal patterns of parameter dominance could be precisely characterized with a temporally resolved parameter sensitivity analysis. In this way, the discharge conditions are characterized, that lead to a parameter dominance in the model. To achieve this, the analysis of temporal dynamics in parameter sensitivity is enhanced by including additional information in a three-tiered framework on different aggregation levels. Firstly, temporal dynamics of parameter sensitivity provide daily time series of their sensitivities to detect variations in the dominance of model parameters. Secondly, the daily sensitivities are related to the flow duration curve (FDC) to emphasize high sensitivities of model parameters in relation to specific discharge magnitudes. Thirdly, parameter sensitivities are monthly averaged separately for five segments of the FDC to detect typical patterns of parameter dominances for different discharge magnitudes. The three methodical steps are applied on two contrasting catchments (upland and lowland catchment) to demonstrate how the temporal patterns of parameter dynamics represent different hydrological regimes. The discharge dynamic in the lowland catchment is controlled by groundwater parameters for all discharge magnitudes. In contrast, different processes are relevant in the upland catchment, because the dominances of parameters from fast and slow runoff components in the upland catchment are changing over the year for the different discharge magnitudes. The joined interpretation of these three diagnostic steps provides deeper insights of how model parameters represent hydrological dynamics in models for different discharge magnitudes. Thus, this diagnostic framework leads to a better characterization of model parameters and their temporal dynamics and helps to understand the process behaviour in hydrological models.

AB - Diagnostic analyses of hydrological models intend to improve the understanding of how processes and their dynamics are represented in models. Temporal patterns of parameter dominance could be precisely characterized with a temporally resolved parameter sensitivity analysis. In this way, the discharge conditions are characterized, that lead to a parameter dominance in the model. To achieve this, the analysis of temporal dynamics in parameter sensitivity is enhanced by including additional information in a three-tiered framework on different aggregation levels. Firstly, temporal dynamics of parameter sensitivity provide daily time series of their sensitivities to detect variations in the dominance of model parameters. Secondly, the daily sensitivities are related to the flow duration curve (FDC) to emphasize high sensitivities of model parameters in relation to specific discharge magnitudes. Thirdly, parameter sensitivities are monthly averaged separately for five segments of the FDC to detect typical patterns of parameter dominances for different discharge magnitudes. The three methodical steps are applied on two contrasting catchments (upland and lowland catchment) to demonstrate how the temporal patterns of parameter dynamics represent different hydrological regimes. The discharge dynamic in the lowland catchment is controlled by groundwater parameters for all discharge magnitudes. In contrast, different processes are relevant in the upland catchment, because the dominances of parameters from fast and slow runoff components in the upland catchment are changing over the year for the different discharge magnitudes. The joined interpretation of these three diagnostic steps provides deeper insights of how model parameters represent hydrological dynamics in models for different discharge magnitudes. Thus, this diagnostic framework leads to a better characterization of model parameters and their temporal dynamics and helps to understand the process behaviour in hydrological models.

KW - Catchment modelling

KW - Diagnostic model analysis

KW - Model understanding

KW - Parameter characterization

KW - Temporal parameter sensitivity analysis

UR - http://www.scopus.com/inward/record.url?scp=84959432211&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84959432211&partnerID=8YFLogxK

U2 - 10.1002/hyp.10764

DO - 10.1002/hyp.10764

M3 - Article

AN - SCOPUS:84959432211

JO - Hydrological Processes

JF - Hydrological Processes

SN - 0885-6087

ER -