Changes in floods and flood regimes in Canada

Abstract

Flood events in Canada result in larger costs, in terms of property damage, than any other natural disaster. Changes in land use, and more recently the impacts of climate change, can lead to changes in the flood regime. This paper describes research examining changes in the flood regime for watersheds across Canada. To distinguish between changes resulting from land-use changes and those occurring due to changes in climate, the analysis of flood data is conducted only on watersheds that are either part of a formal Reference Hydrologic Network (RHN), or which exhibit RHN-like characteristics. A total of 280 gauging stations have been analyzed for changes to a number of attributes of the flood regime. Changes in the magnitude and timing of flood events as well as the frequency and duration of high-flow events are examined for all 280 sites. Using all 280 sites and different groupings of the sites, based on timing of floods, drainage area and the RHN designation, changes were determined in the annual number of sites at which the flow exceeded the 25-, 50-and 100-year return period events. An examination of the results leads to important insights into the nature of changes that are occurring in the flood regimes in Canada. The results have important implications for the design of hydraulic infrastructure, flood preparedness and the provision of adequate flood protection for populations at risk.Les inondations au Canada entraînent des coûts plus élevés, en fait de dommages matériels, que toute autre catastrophe naturelle. Les changements qui touchent l'utilisation des terres, et plus récemment les incidences du changement climatique, peuvent donner lieu à des changements dans le régime des crues. La présente communication décrit les recherches qui portent sur les changements touchant les régimes des crues pour les bassins versants à la grandeur du Canada. Pour distinguer entre les changements qui résultent du changement d'affectation des terres et ceux qui sont attribuables au changement climatique, l'analyse des données sur les inondations est menée uniquement pour les bassins versants qui font partie d'un réseau hydrographique de référence (RHR) officiel ou qui présentent des caractéristiques semblables à celles d'un RHR. En tout, 280 stations de jaugeage ont été analysées afin de déterminer les changements qui touchent un certain nombre de caractéristiques du régime des crues. Les changements liés à l'ampleur et au moment des inondations ou qui se rapportent à la fréquence et à la durée des débits élevés sont examinés pour l'ensemble des 280 sites. Grâce à l'étude de la totalité des 280 sites et à différents regroupements des sites, et en fonction du moment des inondations, de la zone de drainage et de la désignation RHR, les changements ont été déterminés dans le nombre annuel de sites où le débit a dépassé une période de récurrence de 25 ans, de 50 ans et de 100 ans. Un examen des résultats jette un important éclairage sur la nature des changements qui sont en train de s'opérer dans les régimes des crues au Canad...

Identification of changes in floods and flood regimes in Canada using a peaks over threshold approach

“…Using POT, the change in maximum magnitude for 132 sites for 1963 – 2012 was 1.5 (4.5) % significant increasing (decreasing) similar to what Burn and Whitfield () reported 1.9 (9.3) % for 107 sites for 1961 – 2010 for AMS. For 1933 – 2012 the changes were 4.2 (8.3) % for POT (24 sites) and 0.0 (5.3) % for 1931 – 2010 (19 AMS sites) (Burn and Whitfield, ). For the average day of occurrence for 1963 – 2012 the significant results were 3.8 (10.1) % while for AMS for 1961–2010 the corresponding values are 2.8 (17.8) %; for the longer time period these were POT 4.2 (20.8) % while Burn and Whitfield () reported 0.0 (26.3) %.…”

“…The results in Table can be loosely compared to the results for annual maximum series presented by Burn and Whitfield (). Using POT, the change in maximum magnitude for 132 sites for 1963 – 2012 was 1.5 (4.5) % significant increasing (decreasing) similar to what Burn and Whitfield () reported 1.9 (9.3) % for 107 sites for 1961 – 2010 for AMS. For 1933 – 2012 the changes were 4.2 (8.3) % for POT (24 sites) and 0.0 (5.3) % for 1931 – 2010 (19 AMS sites) (Burn and Whitfield, ).…”

“…For 1933 – 2012 the changes were 4.2 (8.3) % for POT (24 sites) and 0.0 (5.3) % for 1931 – 2010 (19 AMS sites) (Burn and Whitfield, ). For the average day of occurrence for 1963 – 2012 the significant results were 3.8 (10.1) % while for AMS for 1961–2010 the corresponding values are 2.8 (17.8) %; for the longer time period these were POT 4.2 (20.8) % while Burn and Whitfield () reported 0.0 (26.3) %. The similarity of these results is because of the predominance of nival watersheds in the grouped results.…”

“…Differences in flood occurrence and trends exist between reference and non‐reference hydrometric stations with the former showing decreasing trends in flood magnitude and the latter increasing trends. Burn and Whitfield () showed that changes in magnitude, timing, and the number and duration of high flow events exhibited both increasing and decreasing trends.…”

On the Relationship Between Flood and Contributing Area

“…However, documented trends in streamflow in this region from the second half of the 20 th century (Rasmussen, 2016;Burn and Whitfield, 2016) …”