Sacha Ruzzante

Topic

Disentangling Complex Climate Change Impacts on Past and Future Streamflow with Statistical Analysis

Department of Civil Engineering

Date & location

• Tuesday, July 21, 2026

• 10:00 A.M.

• Engineering and Computer Science Building

• Room 660 & Virtual Defence

Reviewers

Supervisory Committee

• Dr. Tom Gleeson, Department of Civil Engineering, University of Victoria (Supervisor)

• Dr. Heather Buckley, Department of Civil Engineering, UVic (Member)

• Dr. Nandita Basu, Department of Civil & Environmental Engineering, University of Waterloo (Outside Member) 

External Examiner

• Dr. Bart Nijssen, Department of Civil & Environmental Engineering, University of Washington

Chair of Oral Examination

• Dr. Dr. Adam Monahan, School of Earth and Ocean Sciences, UVic 

Abstract

Climate change is changing the seasonal ebb and flow of rivers and making extreme droughts and floods more common. Many societies and ecosystems are struggling to adapt to the pace of these changes, with floods and droughts responsible for half of all deaths from natural disasters since 1970 and freshwater biodiversity declining faster than either marine or terrestrial biodiversity. However, the mechanisms by which climate change causes hydrologic change remain highly uncertain. Rising global air temperature is the most robustly predicted and observed climate change signal, but its effect on water resources remains uncertain because of complex interactions between air temperature, vegetation growth, evapotranspiration, snowmelt, and glacier retreat. The future trajectory of hydrologic change is also uncertain because current state-of-the-art hydrologic models struggle to simulate changes at the scale of individual streams (although they capture general trends at regional and global scales). Understanding the impacts of hydrologic change on people and the environment entails additional complexity arising from differences in exposure, vulnerability, and adaptive capacity, which can vary at fine spatial scales.

In this dissertation I present four studies which make contributions to three cross-cutting challenges: i) improving hydrologic process understanding, ii) improving hydrologic modelling, and iii) advancing understanding of climate change impacts.

In western North America, I show that rising temperatures are driving more extreme summer streamflow droughts primarily by increasing summer evapotranspiration, although declining snowpack and retreating glaciers also play a role. These droughts imperil the region’s aquatic ecosystems and agricultural sector, and set the stage for conflicts over water use. I develop statistical hydrologic models for hundreds of individual streams and show that these models perform better than state-of-the-art hydrologic models at predicting historical changes. These improved models project devastating outcomes over the next 75 years: by the end of the century, many streams could lose more than half their late-summer streamflow in an average year, but extraordinary droughts are likely to occur even within the next decade. By compiling life history data for a keystone species (Chinook salmon) throughout western North America, I demonstrate that the ecological impacts of this drying are likely to be severe but locally heterogeneous because of both hydrologic and ecological diversity. This is the first projection of the impacts of changing streamflow on Chinook salmon throughout their North American range.

I also provide some building blocks to improve hydrologic modelling of climate change impacts at the global scale. I show that current state-of-the-art hydrologic models struggle to simulate interannual variability and change in highly seasonal hydrologic regimes, which include highly climate-sensitive alpine, polar, and tropical regions. This result indicates there is considerable room to improve our understanding and modelling of past and future hydrologic change. The novel evaluation framework that I propose can guide research towards this goal. To further facilitate this research, I provide a large amount of freely available data, including bias-corrected climate model outputs for over 23,000 streams globally.