Flooding is a major concern for Canadian society as it is the costliest natural disaster type in Canada. Southern Ontario, which houses one-third of the Canadian population, is particularly affected by early spring floods following snowmelt. During the last three decades, there has been a shift in flooding events from March-April to earlier months due to earlier snowmelt coupled with extreme rain events. Hydrological models run with different scenarios of climate change suggest further enhancement of this shift in the future. These projections of streamflow are associated with a cascade of uncertainties due to the choice of Global Climate Models (GCM’s), climate change scenarios, downscaling methods or hydrological models. A large part of the uncertainty is also associated with internal variability of climate due to the chaotic nature of the climate system. Despite these uncertainties, little is known about the impact of atmospheric circulation on past streamflow in southern Ontario and how the internal variability of climate is expected to impact the overall uncertainties in the projections of the future hydrological processes.
In this thesis, the Precipitation Runoff Modelling System (PRMS), a semi-distributed conceptual hydrological model, was established in four watersheds in southern Ontario to assess the impact of atmospheric circulation on the modulation of streamflow and number of high flows. Recurrent meteorological patterns (Or Weather regimes), based on 500hPa geopotential height (Z500), have been first identified in Northeastern North America using the k-means algorithm. The occurrences of these weather regimes patterns were used to create a regime-normalized hypothetical temperature and precipitation dataset that have been used as input in PRMS. Then, to investigate the future evolution of the hydrological processes, PRMS was forced with temperature and precipitation from the 50-members Canadian Regional Climate Model Large Ensemble (CRCM5-LE), a dynamically downscaled version of CanESM2-LE. The 50-members were classified into different classes of similar change in average temperature, precipitation and streamflow to identify the corresponding large-scale patterns. The specific focus of this analysis was on winter high flows, with the identification of a heavy rain and warm index, that can help to explain the generation of winter high flows in southern Ontario. The future evolution of these hydrometeorological extreme events, calculated for each member of CRCM5-LE, was analyzed with respect to the corresponding k-means weather regimes calculated for each member of CanESM2-LE. Finally, the uncertainties in the projections of the hydrometeorological extremes from the 50-members ensemble were compared to other sources of uncertainties using an analysis of variance applied to 504 simulations in the Big creek watershed. The high flows were projected using seven sets of PRMS parameters, 11 CMIP5 climate models forced with 2 scenarios of climate change and the 50 members of CRCM5-LE.
The results, focusing on the winter season, showed that weather regimes High-Pressure (HP) and southerly winds (South) are associated with a higher average streamflow volume and high-flows frequency in the historical period. Regime HP is characterized by high geopotential height anomalies on top of the Great Lakes region together with higher temperature and precipitation amounts. Regime South is characterized by high Z500 anomalies in the Atlantic east coast and is associated with stronger southerly winds and higher precipitation amount in southern Ontario. The temporal increase in HP in the past contributed more than 40% of the increase in average streamflow in winter. In the future, all 50 members of CRCM5-LE ensemble produce an increase in January-February streamflow. 14% of the ensemble depict a larger streamflow increase due to increase in Z500 anomalies in the east coast. This pattern, well defined by the regimes South, is expected to become a major contributor in the generation of hydrometeorological extreme events in Southern Ontario in the future. Regime HP is expected to contribute less to the high-flows due to the disappearance of snow. Overall, the contribution of internal variability of climate to high flows will be stable through the 21st century, primarily due to an increase in rainfall as generators of high flow events. The results suggest that the regional representation of rainfall in the GCMs-RCMs chains will be a critical area to improve with great societal implications for floods. / Dissertation / Doctor of Science (PhD)
| Identifer | oai:union.ndltd.org:mcmaster.ca/oai:macsphere.mcmaster.ca:11375/25325 |
| Date | January 2020 |
| Creators | Champagne, Olivier |
| Contributors | Arain, Muhammad Altaf, Earth and Environmental Sciences |
| Source Sets | McMaster University |
| Language | English |
| Detected Language | English |
| Type | Thesis |
Page generated in 0.0016 seconds