Date of Award

Spring 1-13-2015

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

First Advisor

Balaji Rajagopalan

Second Advisor

Martyn P. Clark

Third Advisor

Jeffrey R. Arnold

Fourth Advisor

Levi Brekke

Fifth Advisor

Joseph Kasprzyk

Abstract

The clear evidence of human-induced hydroclimatic shifts in the western U.S., and especially within the Colorado River Basin (CRB), has motivated the research community to pursue reliable estimates of future climate change impacts on hydrological processes. Although the large body of literature in this field shows consensus on the future drying of the CRB, there are large uncertainties in the magnitude of hydrologic changes because of differences in experimental approaches. Hence, this work focuses on the effects of methodological choices, in particular those related to hydrologic modeling, on the portrayal of climate change impacts at three catchments located in the headwaters of the CRB.

A commentary on the current development of complex process-based hydrologic models is first presented. It is argued that the relatively poor performance of such models may occur due to restrictions on their ability to refine their portrayal of physical processes, and improving hydrological models requires integrating the strengths of prior knowledge of hydrologic processes with the strengths of data driven inference. An assessment of the effects of hydrologic model choice and parameter calibration on projected hydrologic changes follows. Here, it is demonstrated that the subjective selection of model structures may introduce large uncertainties to hydrologic projections. Based on this, the third part of this study compares the effects of hydrologic model choice and parameter estimation strategies on projected climate change impacts. The main finding here is that the choice of parameter estimation methods can provide similar or larger uncertainties in some hydrologic processes when compared to uncertainties coming from model choice. The fourth component of this study evaluates the effects of regional climate model (RCM) configuration and forcing scaling on projected hydrologic changes. The results illustrate the implications of RCM configuration on projected changes of the water cycle, and provide an integrated view of the interplay between forcings and hydrologic model structures in the portrayal of climate change impacts. Finally, a commentary on the main results of this study and possible ways to move forward is provided. It is asserted that a greater role of expert knowledge is required to improve model selection and parameter estimation, and also for incorporating non-stationarity in climate change impact studies.

Share

COinS