1,000 year floods

[Philly2034]

What would be a comprehensive list of all the 1,000 year floods in history?

Just curious

[dwagner88]

Local to me, some considered the 2010 Nashville flood to be close to a 5000 year event. I'm skeptical of that number, but I'm sure it was over 1000 at some gauging stations.

[ORH_wxman]

Gosh, this is a tough one since we really don't have much on floods past the last couple of hundred years...at least in the U.S. Floods with a return period of 1,000 years requires a lot of data...and we don't have that. There's paleodata which can give an idea in certain areas that have good paleo records, but this obviously depends on location.

[Philly2034]

 

[Windspeed]

Local to me, some considered the 2010 Nashville flood to be close to a 5000 year event. I'm skeptical of that number, but I'm sure it was over 1000 at some gauging stations.

 

I don't understand using 5,000 and 10,000 years as a reference point for floods. That's just too large of a period of time to source for localized flooding events. Variations in climate over such long periods would seem to bear more weight on large-scale regional flooding in the paleo record for those ranges of time versus what one could expect in the 100, 500 and 1,000 year range for localized events. As an example, a 5,000 year flood may be related to more severe episodic cycles of El Nino events that show up in the paleo record for a region versus a more common localized event such as an intense hurricane stalling out over the Tennessee valley or multi-MCS event occuring over the same location in a 72 hour period. As for the 2010 Nashville flood, the 1937 event was very similar. I'm not sold that was a 1,000 year flood and would expect such a flooding event on the Cumberland on a 100 to 500 year range than calling that a 1,000 year event. I should locate and study the flood map for the Cumberland again. I reviewed most of the TVA flood maps when I was in college, but that seems too low for what I would expect for a 1,000 year flood. If the 2010 event did surpass the 1,000 year mark, considering the severity of the 1937 flood, we may be underestimating what a 1,000 year flood should be for the Cumberland and Tennessee Rivers.

 

https://www.flickr.com/photos/nashvillecorps/5551008698/in/pool-nashvilleflood/

[illinois]

It is my understanding of the term is that it is more statistical than anything. A hundred year flood has a 1% chance of happening in a year, a 500 year event .5 and a 1000 year .1. So the return frequency would work out to be 1 time in a given period.

[sokolow]

These are the two guys i would ask:

http://blogs.agu.org/landslideblog/

http://washingtonlandscape.blogspot.com/

... either through their direct contact or in the comments of a relevant post. Like has been observed its a question that entangles a framework for thinking through present-day risk assessment vs. painstaking archival and survey fieldwork for reconstructing historic and prehistoric climate.

Basically you'd say that the criteria for a "1-in-a-1000-chance-per-year" event in a given floodplain has changed and is changing over time.

[Windspeed]

It is my understanding of the term is that it is more statistical than anything. A hundred year flood has a 1% chance of happening in a year, a 500 year event .5 and a 1000 year .1. So the return frequency would work out to be 1 time in a given period.

 

You are correct. It is based on the percentages of it happening in any given year. I.e., 1% represents a 1-in-100 chance event occuring in any given year. A 500 year flood has a .5% chance, 1000 has a .1%, etc.; however, the expectation for those events is set and based on how many times that flood level was attained over a period of time on average. That doesn't mean that flood stage can't be met or exceeded only once in a specific period of time. A 1,000 year event could have happened 20+ times in a specific 1000 years. But on average, it may only happen once every 1,000 years or so over a much longer period of time based on paleo data (if that data exists). The rest are just estimates. And it may very well be that our estimates are not exactly correct. For example, before the TVA built the dams and locks that make up the Tennessee river and its tributaries, there were some periodic epic floods in Chattanooga, Muscle Shoals, etc., and the flood maps show those. I think it may be an underestimate that flooding events in the upper and lower Tennessee valley system has less than a .1% chance of experiencing 10+ inches of rainfall in a 2-3 day period in any given year. That just seems too low for me. Furthermore, I'm not so sure I accept that is so rare for Nashville to receive 13 inches of rain in 48 hour period to say it only has a .1% of experiencing that kind of flood in any given year. I am curious if we have underestimated those percentages. In other words, I'd believe such rainfall rates and flooding seen in Nashville in 2010 and 1937, and so forth, has a greater than .1% chance of happening in any given year and are not 1,000 year floods. I'd think somewhere between .5 and 1% chance is more likely and the 2010 Nashville event is probably somewhere between a 1-in-100 and a 1-in-500 year event.

 

I am reminded of when Hurricane Camille stalled out over the Shenandoah Valley of Virginia in 1969. Some areas had localized 27" of rain in a 48 hour period. The flooding was catastrophic on the James River and is considered a 1000 year flood. However, I am doubtful if such an event like that is so rare that it can only be expected to occur at .1% chance in the same location, or really any general location in the southeastern portion of the United States, for any given year, based on the shear number of hurricane/tropical cyclone landfalls that occur on average. Having one stall out for several days over the same areas just doesn't lead me to believe it's such a rare event in any general local in the southeastern CONUS.

[sokolow]

For an extreme example, reframe the question more narrowly: list all the "thousand year" floods we have evidence of in Grant County, Washington State across the last 15,000 years.

Photo of Dry Falls, WA from Wikimedia

[Philly2034]

Thank you for all of your responses.