Thèse de Jonathan JALBERT soutenue le 30 octobre 2015
7 novembre 2016 ( maj : 22 décembre 2016 )
Encadrants : A.-C. Favre
Developing a non-stationary and regional frequency analysis method for intense precipitation events generated by the Canadian Regional Climate Model
The Intergovernmental Panel on Climate Change (IPCC) predicts an increase in the frequency and intensity of extreme precipitations will occur due to climate change. The impacts of these disruptions will have consequences in various fields such as urban drainage, the design and operation of hydraulic structures and the delineation of flood zones. In response to these vulnerabilities, the Ouranos consortium, a Canadian leader in climate science and climate change adaptation, has identified adaptation to extreme precipitations as a priority activity. This project is designed to improve our knowledge about changes in precipitation events and ultimately, lead to the development of adaptation plans for the Canadian climate change context.
The first requirement for this study is to simulate the evolution of the climate. Climate models are ideal as a basis for study since they make it possible to simulate probable future climate outcomes. To analyze precipitations, simulations of the Canadian Regional Climate Model (CRCM) are preferable to simulations from global models because the physical processes that produce precipitations also operate on a local scale. A regional climate model is thus more likely to generate adequate spatial patterns of precipitations. For this particular research, the Ouranos consortium will provide CRCM simulations for 1960 to 2100. The Figure 1 illustrates the sub-domain of the CRCM studied in this project.
Next, a frequency analysis will be carried out on the ensemble of data simulated by the model to estimate the probability of occurrence of intense precipitation events. Frequency analysis, a branch of statistics, makes it possible to predict the probability of occurrence of future events based on the analysis of observed events (or simulated events, in this case). The problem with this methodology is that it should be applied in a context of stationarity, which does not correspond at all to the reality observed with climate change. The first objective of this project is to develop a methodology of frequency analysis that could be applied in a context of non-stationarity.
The second objective of the research project is to reduce the uncertainty in the results obtained using frequency analysis. To do this, the statistical concept of regionalization will be integrated into the methodology developed. Since climate change will probably disrupt the intensity and frequency of precipitations at a regional scale, it seems natural to try to model this spatial dependency statistically. Not only is regionalization consistent with the physics of the problem, it makes it possible to make use of the homogeneous characteristics of a region in order to improve the accuracy of predictions.
In conclusion, the results of this study will provide a better knowledge of changes in extreme precipitations in the context of climate change in Quebec. This will help us to better anticipate the consequences of the extreme precipitations that may occur during the next century.
Dans la même rubrique :