Research Projects

Quantifying the uncertainty in modelled estimates of future extreme precipitation events

Description: The project involves: (a) trend analysis of extreme precipitation events for historical data; (b) refinement and improvement of an existing weather generation model; (c) assessing the uncertainty in extreme precipitation events in the form of intensity-duration-frequency (IDF) curves; and (d) developing procedures for risk-based management of water infrastructure. Trend analysis is being conducted on annual extreme precipitation data of various durations using the Mann-Kendall non-parametric statistical test for trend. The trend test results provide an indication of the nature of any changes currently identifiable in the historical data.  An existing weather generator is being modified and will result in a weather generator capable of synthesizing daily and hourly precipitation data at either a single site or for multiple sites.  The model is capable of generating weather data either for current (historical) conditions or for conditions expected to occur as a result of climate change.  Possible future realizations of extreme precipitation will be obtained using the weather generator, driven by the combination of a particular GCM and a future emission scenario, to obtain a daily time series of precipitation values. These values will then be processed to obtain an annual time series of precipitation extremes for different storm durations. Annual extreme precipitation values for different storm durations will be used to develop an estimated intensity-duration-frequency (IDF) curve based on the projected future conditions.  Finally, procedures will be developed to incorporate information from the probabilistic IDF curves that are created into infrastructure management policy. 

Funding:  Canadian Foundation for Climate and Atmospheric Sciences (CFCAS)

Co-Investigator: Dr. Slobodan P. Simonovic, Civil and Environmental Engineering, University of Western Ontario 

Revision and regionalization of the MTO Intensity-Duration-Frequency (IDF) curves for extreme flow estimates

Description: The research involves building a new set of Ontario Ministry of Transportation (MTO) Intensity Duration Frequency (IDF) curves using the best contemporary tools for statistical analysis of station data, spatial interpolation of station parameters, and estimation of extreme flows. The station analyses is being conducted with the assistance of Meteorological Service Canada. Interpolation methods use regional analysis tools with a topography and landcover database. Refined modelling tools are being used to improve the estimates of return periods associated with extreme events.

Funding:  Ontario Ministry of Transportation (MTO)

Co-Investigator: Dr. Ric D. Soulis, Civil and Environmental Engineering, University of Waterloo

Estimation of Hydrological Extremes

Description: This reasearch work focuses on two main problems.  The first problem relates to the determination of flooding probabilities for rivers with a focus on locations for which there exists little or no streamflow data.  Flood information derived from streamflow records for similar locations is used in an attempt to transfer information from locations for which extreme flow information is available to the location for which the requisite data are not available.  This process is generally referred to as pooled (regional) flood frequency analysis.  An important process in pooled frequency analysis is the examination of the homogeneity of the collection of stations from which extreme flow information is pooled to estimate design quantiles at the site of interest.  This part of the research work seeks to develop improved approaches for evaluating the homogeneity of the available extreme flow information to ensure an effective transfer of extreme flow information when estimating design flow quantiles. The second research problem relates to the determination of the probability of drought events.  The analysis of drought events must typically consider the severity of the drought event as well as the drought duration.  The proposed research will adapt pooled flood frequency analysis techniques for the estimation of drought event probabilities.  In particular, the research will focus on the copula approach for conducting multi-variate frequency analysis that will allow the determination of joint and conditional probabilities that properly reflect the severity and duration of drought events and the correlation structure between drought severity and drought duration.

Funding:  Natural Sciences and Engineering Research Council of Canada (NSERC)