Rainfall Design Intensities in Southern Ontario - Long-term Station Data Show Overall Minor Decreases Since 1990

The previous post reviewed trends in observed annual maximum rainfall amounts at Canadian climate stations. These are provided in Environment and Climate Change Canada's Engineering Climate Datasets. Annual series are used to derive design intensities by first 'fitting a curve' to the data and creating a probability distribution to represent the frequency of observations. By selecting a design return period (i.e., the inverse of the probability of a rainfall intensity occurring in a given year), the corresponding design rainfall intensity can be calculated using the best-fit probability distribution of the observations. Design intensities across a range of durations (typically 5 minutes to 24 hours) create an intensity-duration-frequency curve (or IDF curve) for a given return period (typically 2, 5, 10, 25, 50 and 100-year).

Thirty years ago engineering consultants received 5 1/4 inch floppy disks with new IDF data whenever there were updates to Environment Canada data. A simple program (called "IDF87") was used to select stations of interest from a list and load the annual series data and IDF tables for printing. Today the data is online in "IDF Files" - link:  https://climate.weather.gc.ca/prods_servs/engineering_e.html   

In southern Ontario observed data records have increased since 1990 at long-term stations. In today's Version 3.20 datasets, selected stations in southern Ontario have an average of 49 years of record with a total of 1026 station-years of observations. Since 1990, 494 station-years of data have been added, adding 23.5 years of data, on average, to the observed records.

The following is a review of IDF data comparing design intensities using data up to 1990 and data up to 2017 at many stations. The stations assessed include:     Sarnia Airport, Chatham WPCP, Delhi CS, Port Colborne, Ridgetown RCS, St Catharines Airport, St Thomas WPCP, Windsor Airport, Brantford MOE, Fergus Shand Dam, Guelph Turfgras CS, London CS, Mount Forest (Aut), Stratford WWTP, Waterloo Wellington Airport, Bowmanville Mostert, Hamilton Airport, Hamilton RBG CS, Oshawa WPCP, Toronto City, Toronto International Airport (Pearson).

These are average changes in design intensity at the above 21 stations since 1990:

There is a balance of decreases and increases in all return periods and durations. Green cells show average decreases and red show average increase. As shown, overall all design intensities have decreased by an average of 0.24 % in southern Ontario since 1990. This is good news as infrastructure built over the past few decades, and that has 50+ years of service life remaining, is not "under-designed" for today's rainfall extremes - these design intensities are generally not greater than those used for decades.

The following chart shows these average trends at each return period and duration:

The red and orange bars represent the shortest durations of 5 to 10 minutes - intensities at those durations have decreased the most with only minor average increases for some return periods. The intensities across longer durations of 12 and 24 hours (dark blue and purple bars) have increased at higher return periods (extreme events). The trends for 1 to 6 hour durations are mixed for higher return periods.

The range of trends are at each return period are shown below:

The intensities used to design minor systems, i.e., local sewer and drainage works, are predominantly lower (i.e., 2, 5 and 10-year intensities). Extreme rainfall intensities used to design major overland drainage systems and storage facilities have both increases and decreases.

Many municipalities have reviewed their design IDF values over recent years and have found no increase in local values that would warrant updating design values based on past observations. A previous post presented examples in southern Ontario - link:  https://www.cityfloodmap.com/2018/03/extreme-rainfall-and-climate-change-in.html

Despite past observations, it is good practice to account for future projected changes in rainfall design intensities, given the long service life of today's drainage infrastructure. In Ontario the 2014 provincial policy statement (2014 PPS) includes explicit reference to Climate Change: "Planning authorities shall consider the potential impacts of climate change that may increase the risk associated with natural hazards."

Annual Maximum Rainfall Trends in Canada - Environment and Climate Change Canada's Updated V3.20 Datasets Show Few Significant Trends

Previous posts have presented overall trends in annual maximum observed rainfall amounts at Canadian climate stations (link: https://www.cityfloodmap.com/2020/12/design-rainfall-trends-in-canada.html).

Environment and Climate Change Canada (ECCC) periodically updates trend analysis on the annual maximum series (AMS) for each station, and these series are used to derive design intensities, i.e., Intensity-Duration-Frequency curves. Data area available here (link):  https://climate.weather.gc.ca/prods_servs/engineering_e.html

In May 2021 additional analysis has been included in these Engineering Climate Datasets. The "Whats_New_EC_IDF_v3-20.pdf" file includes a summary table that provides counts of stations and additional station years added with recent update:

The v-3.20 update adds 490 station-years of data. There are currently 17,133 station-years of data. The average length of record is 25.3 years. This represents a slight decrease in record length compared to 25.5 years in v-3.10, as new short record stations are factored in.

Trends in annual maximum rainfall have not change all relative to the 2020 v-3.10 data. The majority of station data show no statistically significant trend. The table below compares trends in earlier datasets, averaged across all durations. 

Trend in Maximum Rain    v3.20       v3.10       v3.00         v2.30

Significant Increase              4.16%     4.28%       4.18%        4.09%

Significant Decrease             2.25%     2.24%       2.33%        2.30%

No Significant Trend          85.73%    85.80%     85.55%      86.37%

No Calculation                      7.86%      7.68%       7.94%        7.24%


The v-3.20 datasets  have the following trends within various durations:

Excluding 'No Calculation" data, annual maximum observations with 'No Significant Trend' represents 92-94% of series, with an average of 93.0%. Excluding the 'No Calc' data results in an average of 4.5% significant increases and 2.5% significant decreases.

The following chart shows that majority of station data show no statistically significant trend.

There has not been any appreciable change in the annual maximum rainfall trends considering earlier datasets (see v-2.30 to v-3.00 update https://www.cityfloodmap.com/2020/02/annual-maximum-rainfall-trends-in.html)

In Canada, the number of stations with annual series used to derive extreme rainfall statistics continues to increase. A previous post explored how manually-operated climate stations have been declining while this increase occurs (link: https://www.cityfloodmap.com/2020/06/do-we-have-enough-climate-stations-in.html). In 1990 there were only 11,268 station-years of data. Today with 17,133 station-years of data we have 52% more information to guide assessments of extreme rainfall.  

This table summarizes the rise in climate station count and rise in station-years as new data is added.

Some have confused the decline in manual stations with a decline in overall data (see the above post). This chart shows the rise in climate stations with IDF data used for engineering design (orange lines), along with the decline in manual station data (blue bars).

National Issues Report Identifies Growth, not Climate Change, as Major Cause of Increased Severe Weather Losses in Canada

Canada in a Changing Climate: National Issues Report by the Government of Canada has just been released and it reviews the cause of rising extreme weather costs (link: https://changingclimate.ca/national-issues). A previous post looked at the impact of growth on increasing losses in Canada over past decades (https://www.cityfloodmap.com/2018/06/catastrophic-losses-in-canada-have.html) - this new report does the same.

The National Issues Report speaks to the cause of increasing losses in Canada directly, considering growth and contradicting common media explanations for the rise, as indicated below:

Page 349 (my bold):

"Costs related to extreme weather events are increasing (see Section 6.4)

Costs associated with damage from extreme weather events in Canada are significant and rising, largely due to growing exposure and increasing asset values. The scale of costs suggests that households, communities, businesses and infrastructure are not sufficiently adapted to current climate conditions and variability."

Page 365:

"6.4 Costs related to extreme weather events are increasing

Costs associated with damage from extreme weather events in Canada are significant and rising, largely due to growing exposure and increasing asset values. The scale of costs suggests that households, communities, businesses and infrastructure are not sufficiently adapted to current climate conditions and variability.

The number of extreme events has increased since 1983, although the distribution of these events across Canada varies significantly, with Alberta being affected the most. Studies on the attribution of such events in Canada indicate that climate change is increasing the likelihood of certain types of extreme weather events, and may be playing a role in the trend of growing losses from such events. However, the majority of rising losses related to extreme weather events are the result of growing exposure and rising asset values. The scale of costs suggests that there is an adaptation gap or deficit, whereby households, communities, businesses and infrastructure are not sufficiently adapted to current climate conditions and variability."

And on page 372:

"Since 1983, the increasing trend in insured losses associated with extreme weather disasters in Canada has primarily been due to an accumulation of value (e.g., people, assets, wealth) year-on-year. But rising losses cannot entirely be explained by growing exposures, asset values and general price inflation—climate change may be playing a role. While a rise in normalized losses is not “proof” of climate change, it is certainly consistent with the anticipated intensification of some weather extremes with rising temperatures (Swiss Re Institute, 2020; Coronesea et al., 2019; IPCC, 2013)."

Opportunities for improvement?

The National Issues Report report does not speak to the effects of urbanization on increased flood risks. Reporting at the CBC has in the past identified a changing climate as the main driver for increasing damages, discounting the influence of other factors. See post: https://www.cityfloodmap.com/2020/06/yes-were-getting-more-extreme-rainfall.html

- if the cause of increased flood damages is not related to more extreme rainfall, then more focus could be given to adaptation as opposed to mitigation

The Radio Canada Ombudsman has reported numerous accuracy issues in past reporting that claimed climate change effects were driving increasing losses, and Radio Canada management has taken the rare step of deleting the article:  https://www.cityfloodmap.com/2020/11/radio-canada-ombudsman-finds-standards.html

- the new National Issues Report supports the Radio-Canada decision to delete its June 3, 2020 article that stated "Contributing factors to higher claims include higher property values, building onto known flood plains and building over green spaces and wetlands that help absorb rain mitigate flooding. Still experts say the extreme weather and flooding is the main factor says Natalia Moudrak, director of climate resilience at the Intact Centre."

- while higher property values were identified as a contributing factor in that Radio Canada reporting, extreme weather was cited as the main factor. In contrast, Canada's National Issues Report refers to the same data, normalized for growth, and finds: "Since 1983, the increasing trend in insured losses associated with extreme weather disasters in Canada has primarily been due to an accumulation of value (e.g., people, assets, wealth) year-on-year." The new report indicates "climate change may be playing a role", suggesting a lesser and uncertain influence on losses, contrary to the Radio Canada reporting.

Limitations in early losses estimates have not been considered and could affect normalized loss trends. The ICLR noted in its 2018 Cat Tales newsletter that prior to 2008 that data was 'cobbled together' and not very robust (see page 5 article "Big data? How about just starting with data?": https://issuu.com/iclr/docs/cat_tales_march_april_2018

- if lower early loss reporting is accounted for, the normalized increasing trends in losses in the new report would be lower, likely making Canada's trends fall in line with other studies with more robust early data that show losses normalized for growth actually can be decreasing over time.