What Are The Odds Of A “Thousand Year Rainfall?”

During the Colorado floods this summer, someone at NOAA started claiming that it was a thousand year flood. When that was shown to be utter nonsense, they changed the story to a thousand year rainfall, based on the odds of that much rain being received at a particular point location.

Suppose there 10,000 rain gauges in the country. The odds of any one of them receiving a thousand year rainfall in any given year are 1,000:1. So we would expect about ten “thousand year rainfall” events in the US every year (10,000/1,000.)

This is high school statistics, not climate change.

About stevengoddard

Just having fun
This entry was posted in Uncategorized. Bookmark the permalink.

8 Responses to What Are The Odds Of A “Thousand Year Rainfall?”

  1. R. Shearer says:

    Wouldn’t that be the case if each gauge were independent and rainfall events occurred randomly? But clearly, gauges near each other will cluster and rainfall events do not have to follow any pre-determined statistical rules, such as is the case with lotteries.

    I need a second cup of coffee to ponder this.

    • If there are 10,000 rain gauges in the US, that is an average of one per 25km x 25km square. Thunderstorm rainfall is much more localized than that. You can get huge differences within 1 km.

  2. Scott Scarborough says:

    I think it is about a 64% chance of getting 10 records or greater with your scenario. If the chance of something happening is 1/1000 per year, lets say, and 1000 years pass, your chances of it happening are about 64%. The odds of at least one 1000 year record being broken with your scenario are about 99.996%.

  3. I agree that the “thousand year” claims are false, in fact patent lies. However, the 10,000 (or however many) average annual rain gauge readings in your scenario are not, generally speaking, all mutually independent events (a given storm will affect more than one gauge, for example, even many gauges; and one would expect variations, from one region to another, in the amount of rain per storm that constitutes a thousand-year event in that region), so your simple calculation is false. Really, once you’ve shown the “thousand-year flood” to be nonsense, the “thousand year” rainfall claim is also negated. The simple logical proposition involved is that a “thousand-year rainfall event” is necessary for a “thousand-year flood event”, so if the latter has not occurred the former cannot have occurred either.

  4. R. Shearer says:

    An underlying point is that statistical events are being used to support specious arguments, e.g. “never let a crisis go to waste.”

    Here is a good discussion of the probability of rare events (applied to risk analysis in healthcare) that is relevant to this. http://gunston.gmu.edu/healthscience/riskanalysis/ProbabilityRareEvent.asp

  5. Phil Jones says:

    Given that records on rainfall have only been accumulated for slightly over 100 years… Well… Records would be broken far more often than if we had 1000yrs + worth of rainfall records…

    Basically … These tools have no idea what a 1000 year storm even is… Barely… Given the huge Boulders in Boulder and how they’ve moved in the past… It takes one hella storm …

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s