• search hit 1 of 1455
Back to Result List

Towards Behavioural Model Parameterization: Integrated Evaluation of Hydrological Modeling and Parameterization Schemes for Improved Water Balance Simulations, Considering Pedo-transfer Functions and Spatial Patterns of Runoff Generation Processes

  • Physically-based distributed rainfall-runoff models as the standard analysis tools for hydro-logical processes have been used to simulate the water system in detail, which includes spa-tial patterns and temporal dynamics of hydrological variables and processes (Davison et al., 2015; Ek and Holtslag, 2004). In general, catchment models are parameterized with spatial information on soil, vegetation and topography. However, traditional approaches for eval-uation of the hydrological model performance are usually motivated with respect to dis-charge data alone. This may thus cloud model realism and hamper understanding of the catchment behavior. It is necessary to evaluate the model performance with respect to in-ternal hydrological processes within the catchment area as well as other components of wa-ter balance rather than runoff discharge at the catchment outlet only. In particular, a consid-erable amount of dynamics in a catchment occurs in the processes related to interactions of the water, soil and vegetation. Evapotranspiration process, for instance, is one of those key interactive elements, and the parameterization of soil and vegetation in water balance mod-eling strongly influences the simulation of evapotranspiration. Specifically, to parameterize the water flow in unsaturated soil zone, the functional relationships that describe the soil water retention and hydraulic conductivity characteristics are important. To define these functional relationships, Pedo-Transfer Functions (PTFs) are common to use in hydrologi-cal modeling. Opting the appropriate PTFs for the region under investigation is a crucial task in estimating the soil hydraulic parameters, but this choice in a hydrological model is often made arbitrary and without evaluating the spatial and temporal patterns of evapotran-spiration, soil moisture, and distribution and intensity of runoff processes. This may ulti-mately lead to implausible modeling results and possibly to incorrect decisions in regional water management. Therefore, the use of reliable evaluation approaches is continually re-quired to analyze the dynamics of the current interactive hydrological processes and predict the future changes in the water cycle, which eventually contributes to sustainable environ-mental planning and decisions in water management. Remarkable endeavors have been made in development of modelling tools that provide insights into the current and future of hydrological patterns in different scales and their im-pacts on the water resources and climate changes (Doell et al., 2014; Wood et al., 2011). Although, there is a need to consider a proper balance between parameter identifiability and the model's ability to realistically represent the response of the natural system. Neverthe-less, tackling this issue entails investigation of additional information, which usually has to be elaborately assembled, for instance, by mapping the dominant runoff generation pro-cesses in the intended area, or retrieving the spatial patterns of soil moisture and evapotran-spiration by using remote sensing methods, and evaluation at a scale commensurate with hydrological model (Koch et al., 2022; Zink et al., 2018). The present work therefore aims to give insights into the modeling approaches to simulate water balance and to improve the soil and vegetation parameterization scheme in the hydrological model subject to producing more reliable spatial and temporal patterns of evapotranspiration and runoff processes in the catchment. An important contribution to the overall body of work is a book chapter included among publications. The book chapter provides a comprehensive overview of the topic and valua-ble insights into the understanding the water balance and its estimation methods. Moreover, the first paper aimed to evaluate the hydrological model behavior with re-spect to contribution of various sources of information. To do so, a multi-criteria evaluation metric including soft and hard data was used to define constraints on outputs of the 1-D hydrological model WaSiM-ETH. Applying this evaluation metric, we could identify the optimal soil and vegetation parameter sets that resulted in a “behavioral” forest stand water balance model. It was found out that even if simulations of transpiration and soil water con-tent are consistent with measured data, but still the dominant runoff generation processes or total water balance might be wrongly calculated. Therefore, only using an evaluation scheme which looks over different sources of data and embraces an understanding of the local controls of water loss through soil and plant, allowed us to exclude the unrealistic modeling outputs. The results suggested that we may need to question the generally accept-ed soil parameterization procedures that apply default parameter sets. The second paper attempts to tackle the pointed model evaluation hindrance by getting down to the small-scale catchment (in Bavaria). Here, a methodology was introduced to analyze the sensitivity of the catchment water balance model to the choice of the Pedo-Transfer Functions (PTF). By varying the underlying PTFs in a calibrated and validated model, we could determine the resulting effects on the spatial distribution of soil hydraulic properties, total water balance in catchment outlet, and the spatial and temporal variation of the runoff components. Results revealed that the water distribution in the hydrologic system significantly differs amongst various PTFs. Moreover, the simulations of water balance components showed high sensitivity to the spatial distribution of soil hydraulic properties. Therefore, it was suggested that opting the PTFs in hydrological modeling should be care-fully tested by looking over the spatio-temporal distribution of simulated evapotranspira-tion and runoff generation processes, whether they are reasonably represented. To fulfill the previous studies’ suggestions, the third paper then aims to focus on evalu-ating the hydrological model through improving the spatial representation of dominant run-off processes. It was implemented in a mesoscale catchment in southwestern Germany us-ing the hydrological model WaSiM-ETH. Dealing with the issues of inadequate spatial ob-servations for rigorous spatial model evaluation, we made use of a reference soil hydrologic map available for the study area to discern the expected dominant runoff processes across a wide range of hydrological conditions. The model was parameterized by applying 11 PTFs and run by multiple synthetic rainfall events. To compare the simulated spatial patterns to the patterns derived by digital soil map, a multiple-component spatial performance metric (SPAEF) was applied. The simulated DRPs showed a large variability with regard to land use, topography, applied rainfall rates, and the different PTFs, which highly influence the rapid runoff generation under wet conditions. The three published manuscripts proceeded towards the model evaluation viewpoints that ultimately attain the behavioral model outputs. It was performed through obtaining information about internal hydrological processes that lead to certain model behaviors, and also about the function and sensitivity of some of the soil and vegetation parameters that may primarily influence those internal processes in a catchment. Accordingly, using this understanding on model reactions, and by setting multiple evaluation criteria, it was possi-ble to identify which parameterization could lead to behavioral model realization. This work, in fact, will contribute to solving some of the issues (e.g., spatial variability and modeling methods) identified as the 23 unsolved problems in hydrology in the 21st century (Blöschl et al., 2019). The results obtained in the present work encourage the further inves-tigations toward a comprehensive model calibration procedure considering multiple data sources simultaneously. This will enable developing the new perspectives to the current parameter estimation methods, which in essence, focus on reproducing the plausible dy-namics (spatio-temporal) of the other hydrological processes within the watershed.
  • Die vorliegende Dissertation beschäftigt sich mit der Anwendung und Weiterentwicklung physikalisch basierter, räumlich verteilter Niederschlag-Abfluss-Modelle. Diese Modelle simulieren die hydrologischen Prozesse detailliert und berücksichtigen dabei die räumli-chen und zeitlichen Dynamiken hydrologischer Variablen. Die konventionelle Modellbe-wertung basiert überwiegend auf Abflussdaten, was zu einer eingeschränkten Darstellung der Einzugsgebietsdynamik führen kann. Die Arbeit unterstreicht daher die Notwendigkeit, eine umfassendere Bewertung der Modelle zu implementieren, die sowohl interne hydrolo-gische Prozesse als auch andere Wasserbilanzkomponenten einbezieht. Ein wesentliches Element der Dissertation ist die kritische Analyse der Auswahl und Anwendung von Pedo-Transfer-Funktionen (PTFs). Diese Funktionen sind entscheidend für die Parametrisierung von Bodenwasserretentions- und Hydraulikeigenschaften. Es wird gezeigt, dass die Wahl der PTFs erheblichen Einfluss auf die Modellsensitivität in Bezug auf die räumliche Verteilung dieser Eigenschaften und auf die Simulation der hydrologi-schen Prozesse hat. Durch den Einsatz verschiedener PTFs in einem kalibrierten und validierten Modell konnte die Variabilität der Simulationsergebnisse bezüglich der Bodenfeuchte, der Eva-potranspiration und der Abflusskomponenten deutlich dargestellt werden. 1. Manuskript: Evaluation des hydrologischen Modellverhaltens unter Nut-zung eines Multi-Kriterien-Bewertungsschemas, das sowohl quantitative als auch qualitative Daten integriert. Hierdurch konnten optimale Parameterkonfigurationen für Boden und Vegetation identifiziert werden, die zu konsistenten Simulationsresul-taten für Transpiration und Bodenwasser im Vergleich zu gemessenen Daten führten. 2. Manuskript: Es behandelt die Sensitivität des Wasserhaushaltsmodells be-züglich der Auswahl von PTFs in einem kleinen Einzugsgebiet in Bayern. Durch Va-riation der PTFs in einem kalibrierten Modell wurden deren Auswirkungen auf die räumliche Verteilung der Bodenhydraulikeigenschaften sowie auf die Wasserbilanz und die räumlich-zeitliche Variation der Abflusskomponenten aufgezeigt 3. Manuskript: Fokussiert auf die Verbesserung der räumlichen Darstellung dominanter Abflussprozesse in einem mesoskaligen Einzugsgebiet in Südwest-deutschland. Die Anwendung einer räumlichen Leistungsmetrik (SPAEF) ermöglich-te den Vergleich der simulierten Muster mit den Mustern, die aus digitalen Boden-karten abgeleitet wurden, und zeigte eine hohe Variabilität in Bezug auf Landnut-zung, Topographie und angewandte Niederschlagsraten. Die Ergebnisse der Dissertation tragen zur Lösung der in der hydrologischen Forschung identifizierten Probleme bei, insbesondere in Bezug auf die räumliche Variabilität und die Methoden der Modellierung. Sie bieten neue Perspektiven für die Kalibrierungsverfahren, die darauf abzielen, plausible Dynamiken (sowohl räumlich als auch zeitlich) der hydrolo-gischen Prozesse innerhalb des Wassereinzugsgebiets zu reproduzieren. Die weiterführen-den Untersuchungen, die in dieser Arbeit gefördert werden, sind von großer Bedeutung für die Entwicklung umfassender Modellkalibrierungsstrategien, die multiple Datenquellen simultan berücksichtigen und somit zu nachhaltigeren Wasserwirtschaftsentscheidungen beitragen können.

Download full text files

Export metadata

Author:Hadis Mohajerani
Referee:Markus Casper, Tobias Schütz, Thomas Kreiter
Advisor:Markus Casper
Document Type:Doctoral Thesis
Date of completion:2024/06/14
Date of publication:2024/06/14
Publishing institution:Universität Trier
Granting institution:Universität Trier, Fachbereich 6
Date of final exam:2023/11/22
Release Date:2024/06/14
Tag:Hydrological Modeling; Pedotransfer Functions
Number of pages:xii, 130
First page:i
Last page:130
Institutes:Fachbereich 6
Licence (German):License LogoCC BY-NC-ND: Creative-Commons-Lizenz 4.0 International

$Rev: 13581 $