Harvey attribution games begin at NOAA

The US National Oceanic and Atmospheric Administration, owner of the National Weather Service, has now posted some spectacularly speculative numbers claiming that Harvey’s record rainfall was so rare as to be unnatural. This is the attribution game; the new wave of alarmist pseudoscience.

The title of their article is “Reviewing Hurricane Harvey’s catastrophic rain and flooding.” Mind you they say that these funny numbers are not NOAA’s official guess, as that is yet to come thanks to some data problems. So in order to catch the news cycle they are reporting the wild guesses of other attribution groups. They do this as a front page story on Climate.gov, their student indoctrination website. The story, and that is all it is, is deliberately non-technical.

This is like the 100 year flood game, which anyone can play. You first find the historical data for the type of event in question, which can be a problem in itself. Then you extrapolate that data to a much longer period, to guess what happens at the frequency in question, such as the 100 year flood. The longer the period, the wilder the guess. Finally you express this frequency as a uniform probability that the event will occur in a given year.

For example, by definition the 100 year flood has a one percent chance of occurring in any given year. The trick is to guess what that flood looks like. If that guess is based on just 100 years of data then it is not likely to be correct. You would need at least 1000 years of data plus an unchanging natural climate, neither of which exists. If it is based on 10 years of data then it is a wild guess. If it is based on 1 year of data or less then it is just nuts.

The Harvey rainfall guesses fall into the nuts category. But in the attribution case the game is to guess the frequency, not the flood. It is the same math worked two ways. Given the frequency (once in X years), find the flood, or given the flood (Harvey), find the frequency.

Keep in mind what is going on here, which is trying to attribute the event in question to human caused global warming. This is why these are called “extreme event attribution” studies. The less likely the guess is for a natural occurrence, the more they can argue for a human cause. This is why NOAA asks “How likely is an event like this?” of the Harvey rainfall amounts.

In Harvey’s case the guesses that NOAA cites are absurdly small. The most generous is one in a thousand years for the heaviest single day of rain. But they range all the way up (or down) to an astonishing two chances in a million for the full five-day rainfall. It is the 500,000 year flood! Assuming we have 100 years of good rainfall data, which may be unlikely, this 5 day guess is like estimating the 100 year flood based on less than a days worth of flood data. It is truly nuts.

Clearly there is no science here, just unbelievable crank turning math — extrapolation without end. Arithmetic does not care about physics and this is baseless arithmetic pretending to be science. Pretty much all of so-called attribution science falls into this category of scams. So it will be interesting to see what NOAA’s official attribution estimates look like, if they ever produce them.

There is nothing unnatural about Harvey’s rainfall, which is well understood. A major hurricane ran into a huge high pressure system, called a Bermuda high. So it sat in one place for five days, producing a huge amount of rain. This may well be a natural 100 year flood, taking the entire USA into account, or maybe even several hundred. There is nothing unnatural about such numbers. The only unnatural thing here is these wacky guesses.


About the Author: David Wojick, Ph.D.

David Wojick is a journalist and policy analyst. He holds a doctorate in epistemology, specializing in the field of Mathematical Logic and Conceptual Analysis.