Detection of trends in flood magnitude and frequency in Canada

Abstract

Changes in flood regimes in Canada were examined using a large-scale dataset of hydrometric stations from across the country. Significant trends were analysed in time series of both Annual Maximum (AMAX) and Peaks-Over-Threshold (POT) series of streamflow data. POT series were extracted from daily flows for each watershed using a semi-automated threshold selection method. Since flood regimes are complex by nature, a multi-temporal and multifaceted approach was employed to identify and properly characterize the types of changes. Common time periods of the recent 30-, 40-, 50-, and 60-years were studied. Trends were investigated both in terms of flood magnitude and frequency of these time series. Changes were examined using different groupings of sites based on dominant hydro-climatic regions, drainage area size, and land-use changes based on hydrometric reference stations. Trends were examined both at local and field scale. Notably significant trends were identified in the Atlantic, Central, Prairie, and Northern regions. Additionally, Reference Hydrometric Basin Network (RHBN) watersheds were found to have more statistically significant trends than the non-RHBN sites.New Hydrological InsightsAn increased number of threshold exceeding events was observed from this analysis. Flow magnitudes series showed more increasing trends in the most recent time windows while there are more decreasing trends in longer time periods. Different types of changes were observed for different hydro-climatic regions.