Maps, Clarification on Eggos Vs. 10,000 Year Flood

I tweeted on Tuesday an ABC News article talking about how the historic* flooding last month in Atlanta, Georgia will delay the availability of Eggo Waffles until the middle of 2010.

A new article I read today clarifies things though, saying that, the plant was scheduled to reopen (from a sanitizing due to listeria) just before the floods. While the parking lot flooded, dirty water may have not entered the building at that time. Presumably because of the putrid conditions following the flood, Kellogg's voluntarily redid all the cleaning, which has considerably delayed the plant being ready to go back into production:

We also issued a news story today with a similar clarification, after getting Comments from Facebook users about where waffle shortages are (Indiana & Virginia) and are not (Pennsylvania and Georgia) showing up so far. Quotes from that are shown below:

I'm not 100% sure of the location of the plant... this article lists the location as "south of I-20 and west of I-285 near Thornton Road" which is presumably the Kelloggs plant listed at 5601 Bucknell Dr SW (google map and link to Street View photo below).

So how did the plant fare when you look at a total rainfall map? It's probably not relevant because the whole city of Atlanta got doused, and it depended more on what river or creek you were near, as to how much flooding you got. But for the record based on the map above I have placed an "X" at the approximate location of the plant on this NWS image showing 7-day rainfall (obtained from here). As you can see, it was well within the 6-9 inches range (as was most of the city):

*How historic was the rainfall? This NWS statement says:

"THE GREATEST AMOUNT WAS 16.7 INCHES JUST WEST OF DOUGLASVILLE. IT HAS BEEN DETERMINED THE CHANCES OF 10 INCHES OR MORE OCCURRING AT ANY GIVEN POINT ARE LESS THAN ONE HUNDRETH OF ONE PERCENT. THIS MEANS THE ODDS ARE 1 IN 10000 OR MORE OF RECURRENCE IN ANY GIVEN YEAR...OR A 10000 YEAR RAIN EVENT."

What this means (I think) is that, on average, it will be more than another 10,000 years before they see that much rainfall in Douglasville again (peace of mind for the residents, I am sure). That is... unless of course Climate Change is Increasing Extreme Precipitation Events (cue evil music).

The USGS says that some river stages (which depend on the rain in a larger area) in the area were on the order of 100 to 500-year floods, which means that their crests were the highest in USGS-recorded history. Twenty river gauges were destroyed by the high water! You can see graphs from all the gauges here.

Reader Comments (5): Keith Heidorn:

The statistics of return period depend on the length of record and the magnitude of annual extreme events. While this may have been a 1 in 10,000 year event in the past, if we recalculate the statistics now, it should be a one in X year event where X is the length of record (though in fitting the data to the equation, that might be a bit off.) Extreme events are based on the past events and expressed as a probability of occurrence. The "years" comes in by inverting the probability. As Crowley says the probability of such an event (1 in 10,000) is .01% which is the true meaning of the statistic. Engineers have inverted this to give probability a handle (years) that most can understand for design purposes. It does NOT, however, have anything to do with averages. Perhaps "return period" is best thought of as one of those confusing terms like "normal" that do not translate well for the public.

Posted by Keith Heidorn | November 23, 2009 11:07 AM r4i kort:

Floods is something which is a natural calamity and nobody can help it from that or nobody can fight against it.One has to keep faith in God and live. One day everybody day has to die.Can anybody solve the problem by moaning about it. Then its better to just to help people to come out from such situation who are facing difficulties.

Posted by r4i kort | November 22, 2009 10:51 PM Michelle:

My understanding is that I stand a better chance of watching Powder Springs flood again like it did in September than winning Mega Millions. I'd prefer those odds to be the other way around. :oP

As far as the Great Waffle Shortage goes, my local Kroger is down to the multigrain and a couple of boxes of blueberry. I guess everyone saw the news story on the local channels earlier this week and ran out to buy all the Eggos they could fit in their freezer.

Posted by Michelle | November 20, 2009 6:46 PM K:

The whole 100 year, 500 year, 1000 year etc. flood naming is a very big misconception. Like Pat said, a 10,000 year flood doesn't mean it will be another 10,000 years before something like that happens again, it means there is a .01% chance of it happening again next year (and any year after that).

The same misconception goes with precipation percentage changes. A lot of people I know, when they hear 60% chance of rain, they think it will rain 60% of the day when that isn't it at all. I have to tell them, no it means if you had 10 days that are exactly the same in terms of weather, 6 of those days would have rain.

You would think NOAA, USGS and all these other goverment agencies would explain those things better.

Posted by K | November 20, 2009 3:29 PM Pat Crowley:

Jesse, you make a common misinterpretation on the precipitation frequency. Douglasville will not have to wait 10,000 years to see another such event. They might see one tomorrow, but the probability is only .01%. Back to back 100 year events are not uncommon in the record. And since we have had no global warming in the last 10 years, you cannot blame the storm on that particular fallacy. At least they are over the recent drought.

FROM JESSE: Global Warming depends on who you ask, there's plenty of people who say it never stopped, but I don't want to get into that argument here. As to the 10,000 year question - doesn't it mean "on average" (though technically it could happen repeatedly) - otherwise why quote the years, they must mean something?

Posted by Pat Crowley | November 20, 2009 12:23 PM