Monday, May 17, 2010

A 1000 Year Flood

NOAA has estimated that the Great Mayday Flood of 2010 in Tennessee exceeded the expectations for a once-in-1000-years flood in many places in the state. I'll duplicate the map of estimated recurrence intervals from the linked article here (click for full size version):

Our place is in the northwest corner of Lewis County, right on the 500-year contour. To our north large areas experienced a rarer-than-once-per-millenium rainfall.

So, what does this mean? Is it a sign of the apocalypse? Is this one of those beyond-the-range-of-historical-variation freak events that would be the vanguard of rapid global climate change? At this point, one would have to say "probably not." This may seem odd given that it shattered records across the area for rainfall totals and flood heights. But rare events must always be interpreted in the large scale.

First, these estimations of quincentennial and millenial extremes are based on extrapolations from roughly 100 years of real data. We have a general idea of the frequency of extreme rainfall events within regions based on data from many stations, but we don't have 500 years of real data from any single point to validate these estimates. So there is some room to question them. But more fundamentally, the occurence of extreme events, even hyperextreme events, is to be expected. If you have 1000 watersheds, on average you would expect one of them to experience a 1000 year flood every year. Someone wins the lottery, and someone gets hit by the meteorite.

In the case of this recent event, there's an important starting point. Tennessee is a rainy place. It is one of the rainiest of the 50 United States on average. At our homestead, my records for 2002-2009 average about 58" of precipitation per year, which is typical for highland rim locations. That's just shy of being a rain forest by some criteria. The Nashville Basin averages a bit drier, but it still runs well over 40" per year on average. Rainfall is also quite variable here. Droughts and heavy rains are common. Before this year, again just from 2002-2009, we have had months with over 15" of rain and months with no measurable rain at all. Daily totals over 3" occur every year; totals over 5" have happened a couple of times before in our brief experience. So, while a 48-hour storm total of 18" might be a millennial extreme occurrence just about anywhere, Tennessee is a good place to look for one.

As for the specifics, if you look at the map and look at the area enclosed within the 1000-year contour, you'll see that it is only a small fraction of the total area of Tennessee. It's actually probably only about 1/1000th of the total land area of the Lower 48. Which means (I'm sure you are ahead of me here already), a rain event this extreme might be expected to hit an area about this size every year, SOMEWHERE within the Lower 48. It's just not very likely to hit any single particular spot. This time it hit populated areas including a State capital. Next time it might be all rural areas and attract far less attention.

You can never infer a trend from a single incident. But even given that, it appears at this point that the 1000-year flood in Tennessee is still likely to be within the expected range of extreme climate variability for the region and continent as a whole. The red flag will be if these types of events become more common. There have been some studies suggesting that this is indeed happening, but you still can't come to a solid conclusion yet.

Personally, I remain suspicious, but will defer to the probabilities at this point pending further recurrences. I'm also still keeping an eye out for a second South Atlantic hurricane; this has not yet happened, leaving the 2004 storm categorized as an isolated event without long-term implications. I also derive considerable comfort in knowing that our own bottomland homestead evidently survived the 500-year deluge with only minor consequences.


At 3:18 PM, Blogger Unknown said...

The calculation of a 1000-year flood flow is fraught with peril, as it involves choosing a distribution function and then extrapolating it out to the ragged edge. While your flood may have exceeded previous estimates of a 1000-year event, it would take alot more than 100 years of data to make the error bars on such a calculation at all manageable.

I don't think we should expect extreme events like this to happen randomly with respect to each other, unfortunately. El NiƱo is famous for producing extreme events in California and affecting climate on a global scale. Certainly there must be other phenomena, perhaps not so easily characterized by ocean temperatures, which can push weather outside its usual boundaries across an entire continent for an extended period.

This brings the inevitable question of whether climate change plays a role. Climate models have difficulty projecting reliable changes in average precipitation, much less extreme events, in coming decades. It will take decades or centuries before we have enough data to know the effect of climate change on extreme events.

There's no reason to think your flood isn't a product of random events under the normalcy we've been used to over the last 100 years. We'd be in big trouble if the magnitude of the floods were made possible by climate change as our climate is going to change alot more than it already has.

At 9:46 AM, Blogger Jim R said...

We visited the Pigeon Forge area a couple years ago, as tourists (I was a bit surprised at what a Vegas-like vibe the place has). That year it was dry, dry, dry. Not quite like Texas, but the permanent streams were all down to a lazy trickle.

I think this is the numero uno symptom of AGW. Call it what you like, but it's not so much a dramatically hotter climate as it is a much more chaotic one. I have a feeling it will get much, much worse before it gets better, but it'll be impossible to predict whether any given place will get flood or drought.


Post a Comment

<< Home

Site Meter