Evaluation of the impact of climate variability on modelling hydrological response /Hak-Soo Kim.
Date
2010
Authors
Kim, Hak-Soo
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Conceptual rainfall-runoff models are increasingly being proposed as important tools for water resource prediction, and in particular for assessing the impacts of climate and land use changes on catchment hydrology. But there are many issues that must be addressed in order to avoid adverse effects from model complexity, over-parameterisation and limited data. A model's ability to capture the effect of climate variability is obviously one of the most important issues for addressing modelling capacity to identify the impacts of climate and land use changes. The main objective of this thesis is to provide an assessment framework to evaluate model adequacy in representing the impacts of climate variability on streamflow. This includes various techniques for reducing predictive uncertainty arising from data and model calibration. The methodology employed in the thesis is primarily based on a combination of data- and model-based analysis. The information content in hydrological data is explored and results of data-based analysis (trend analysis, baseflow separation and estimation of non-parametric impulse response functions) are presented to give detailed information regarding the temporal and spatial variation in the hydrologic response characteristics for climatically and topologically different catchments. A conceptual rainfall-runoff model is applied to represent the impact of weather patterns on streamflow. The consistency (or variability) over time and space in model performance and parameter values are investigated to detect problems related to the temporal variability of the model. Several performance criteria are used to evaluate the performance of the model in each calibration period. In order to gain a comprehensive appreciation of the model, the investigation of the model includes rigorous examination of the behaviour of the model through comparison with alternative model approaches. A data analysis technique (non-parametric impulse response function estimates based on an average unit hydrograph) using different-sized flow events is incorporated into the assessment of model adequacy; and temperature dependency on model performance is investigated based on a new approach: incorporating bias corrections and temperature classifications, to identify model deficiencies in representing the impact of different climate conditions. Various techniques (rainfall correction, areal rainfall estimation and selection of an appropriate calibration length) are used to retrieve improved information from the available data to provide increased model reliability. Some issues related to complexity of multi-objective optimization problems are reduced in a new multi-criteria rejection approach where the capacity to generate a reasonable subset of the possible Pareto solutions is demonstrated using a graphical technique. Results from model assessment considering climate variability are presented for gauges in the upper Murrumbidgee Catchment. The consistency of the estimated catchment response and model performance is highly related to catchment characteristics. The rainfall regime is a main driver of different hydrological response characteristics. The effective rainfall and the quick flow volume for large events in the simulation period are underestimated if the calibration period includes high flows. Increases in systematic errors of modelled flow and resulting poor model performance with temporal variability in model parameters are strongly related to the large differences in impulse response estimates for small and large flows. The model has a potential limitation in its capacity to analyse the impacts of land use, or land use and climate variability simultaneously, especially in drier catchments that have a hydrological response dominated by quick flow. Model performance in peak flows is highly correlated with temperature in each season and a tendency towards underestimation of high flows is dominant in cooler conditions across all seasons. Guidance toward appropriate model improvements is provided to reduce model deficiencies. This will improve the ability to detect the influence of land use change on streamflow and increase the accuracy of the climate change impacts on catchment-scale river discharge for such catchments which are more vulnerable to land use changes.
Description
Keywords
Citation
Collections
Source
Type
Thesis (PhD)
Book Title
Entity type
Access Statement
Open Access
License Rights
Restricted until
Downloads
File
Description