I agree with the concepts in the article in Source. I especially like the graph showing that there is about a 1-in-4 chance of flooding in a 30-year exposure at a location. You can also say there is about a 20% (1-in-5) chance of flooding in a 20-year timeframe (technically an 18% chance, but I kind of like having people expect that their odds of getting hit with a 1% annual likelihood event grow by 1% per year, rather than get into the details of cumulative statistical probabilities).
There are 2 other messages I think merit more consideration related to this topic. Though they were not really the point of the article, I think they are logical follow-ons to improving our flood risk messaging..
First is that the 1% probability event may not be the design standard we should be using. As I understand it, FEMA has already proposed (or made) the 0.2% event (500-year) a new target, but I don't know the details. Using the 0.2% annual probability event as a standard reduces the cumulative chance of being flooded to 4% in 20 years. I think that is much more consistent with what people might expect for their chance of being flooded in an area that their city or other zoning body says is suitable for development. I realize it can be very costly to retrofit to such a standard, but it may be reasonable as a standard for new approvals. (I can also see that 0.2% might be too stringent for levee design in rural areas). Some suburban communities in the Minneapolis-St. Paul area accepted standards like requiring a 1' freeboard above the 1% annual probability event flood level, or even 2' (or more) around landlocked basins. Where we were the engineers for stormwater management, we also made sure to account for runoff events of durations from 1/2-hour thunderstorms up to the 30-day snowmelt, the latter of which did not produce high flowrates, but produced the largest volume of runoff - which can be the most critical condition for flooding in ponds with limited outlet capacity. (A multi-year wet weather cycle can be even more critical for landlocked basins). I was kind of shocked to learn, when I lived through the June 2013 flood in Calgary, Alberta, Canada, that their design standard was a "50-year" flood (2% annual probability), rather than the "100-year" I was accustomed to from the US. As you can imagine, in the immediate aftermath, developers of new condominium buildings had to redesign so they would not be flooded in a 1% event, plus some freeboard - just in order to get people interested in considering buying a home in their building. The Dutch, of course, use a much more protective standard than 0.2% for keeping the ocean from inundating their country.
The second message is related to stormwater management design. Again, I think using the 1% annual return event is insufficient, and we should be discussing using the 0.2% event as a standard more in keeping with what residents around a pond, lake, or stream in an urban environment imagine as a reasonable level of protection from flooding. But my main concern is that we as designers need to be thinking about what happens when events greater than our design standard hit. We should, when possible, provide for a safe overflow pathway for the excess water, so that no catastrophic erosion or widespread damage will result. Where that is not possible, serious discussions with the responsible government body (City, County, State, or other) should be held to highlight the limitations and figure out what an acceptable mitigation strategy might be. In the mere 40 years of my engineering practice, portions of the Minneapolis-St. Paul MN area had 2 (maybe more) rainfall events that were well in excess of the nominal 100-year design standard, and one of those storms was close to the Standard Project Flood used in certain dam designs. In those huge events, debris blocking inlets to storm sewer pipes, catch basins, trash racks, etc. can be a critical factor in making flooding worse. A tree swept away by the flood plugged the inlet to a 42" pipe and nearly caused failure of a major embankment with a road and utilities. Only the presence of an overflow inlet saved the embankment. City workers in some places waded into ponded flood waters to clear branch-and-debris-blocked catch basins in order to protect surrounding homes from flooding.
There is a lot to consider when designing for stormwater and flood management. I think it's time to improve the risk communication, as described in the Source article. I think once we have communicated how frequently flooding can be expected with the existing design standards, it will be evident we need to talk about what our standard should be for protection from flooding. Then we must design so that our stormwater management is robust, even when events larger than the design storm hit.
------------------------------
James Langseth
Retired
Bethesda MD
------------------------------