How Not To Argue

Today, google posted on my phone an article by one Dawn Stover writing a typical how stupid are you article on weather and the recent two hurricanes in the Bulletin of Atomic Scientists (“atomic” scientists of course know lots about climate). Her argument is that you trust weather reporting, which we don’t so therefore you should trust climate science because these forecasts are done by the same guys.  Notice those forecasts, most forecasts go out to 5-7 days, but only tomorrow’s forecast is worth a damn. 

Her article first tries to back-pedal from the “weather is not climate” because that’s been “oversold”. She points out the following:

For one thing, hurricanes get their energy from warm, moist air masses. Rising water temperatures in the Atlantic Ocean and Gulf of Mexico are creating more fuel for hurricanes. Irma, for example, encountered water approaching 90 degrees Fahrenheit as it made its way toward Florida.

Alas, this is an argument in the wrong direction. Because that’s a prediction that hurricanes should be more numerous and violent as the climate warms. But have hurricanes been more violent and more frequent? (hint: Hurricanes have not statistically become more frequent or more violent)

Since they haven’t this is a very good argument that the climate is not changing or at least the change, if occurring, is not statistically significant.  So it is rhetorically unsurprising that Ms Stover fails to address the frequency/strength question because alas, it would have been had the answer would like to admit. This sort of reasoning is not science. If data does not support your hypothesis it is not to be hidden or ignored. It needs to be addressed.

Also, Irma “encountered water approaching 90 degrees Fahrenheit as it made its way toward Florida” is this the norm for this season? What is the average in that area? How much does it normally fluctuate. Is 32 degrees C for the region within normal statistical variation? How was it measured.

I don’t know, Mr Stover intimates this is high but … gives no data or context.

Bad statistics and claims lack context are not the way to convince anyone of anything. For all I know, climate is changing and is caused by humans. But I’ve yet to see anything that resembles science that demonstrates that’s the case.

Leave a Reply

Your email address will not be published. Required fields are marked *


  1. says:

    Have oceans gotten warmer? yes

    Have storms in the No Atlantic become stronger and more numerous? yes

    It’s important to note that changes in frequency and intensity vary from basin to basin. In the North Atlantic Basin, the long-term (1966-2009) average number of tropical storms is about 11 annually, with about six becoming hurricanes. More recently (2000-2013), the average is about 16 tropical storms per year, including about eight hurricanes.

    Of course if each year is a sample then a sample of 13 (2000-2013) is very small. We also don’t have very good measurements of ‘storms’. First, we’ve only had weather satellites since the 60’s so before that there was no global count of hurricanes and other storms….only records made if they hit populated land. Second the counting of storms is qualitative rather than quantitative. A storm counts as a ‘hurricane’ if its wind speed hits a certain mph. That’s fine except you are really talking about energy. A hurricane with 110mph that’s about 250 miles in diameter is not the same as one with equal speed but covering 450 miles. The second storm is a hell of a lot more energy than the first (I suspect if we had a good way to measure it, it would vary with the square of the area rather than linearly). So simply counting them as ‘two storms’ is missing something very important. Suppose if we couldn’t count rainfall in an area but could only make reports on how many days of the year it rained? Such a qualitative way of measuring would certainly cause our statistical detection ability to be muffled.

  2. Mark says:

    Sorry, been busy.

    I didn’t say the oceans haven’t become warmer. What I did say, was that numbers without context are not how to argue.

    And .. not so much change. And you’re right, before the 60s … the numbers would probably be undercounted.

  3. says:

    OK but when the context is filled in it fits the theory so what’s your issue? The ‘context is missing’ thing only works if filling in the context changes the story rather than reinforces it!

  4. says:

    Speaking of stories, let me share something I’ve been thinking about.

    When dealing with intellectual dishonesty, there’s a bit of a vulnerability the intellectually honest has to contend with…a true theory can generate an unlimited number of failed predictions.

    What do I mean? Well imagine someone in your family became a crack addict. What would you say?

    No doubt you would say “this is horrible, you must stop”. “why?”….what happens next is what will get you in trouble.

    Rattle off some reasons why being addicted to crack is something someone should stop
    – You’ll lose your job
    – You’ll die
    – You’ll get arrested
    – You’ll end up in jail
    – You’ll end up getting hurt
    – You’ll end up stealing
    – You’ll end up prostituting yourself
    – You’ll end up broke, in debt, homeless

    If you dealt with addicts (I’ve experienced several), what you will discover is your predictions will end up betraying you. Every time you say “this can’t go on another week, she’s going to get arrested”…you’ll find she doesn’t….or the cop lets her go….etc. You say ‘he’ll lose his job’ but he’ll maneuver around the addiction and keep the job unaware.

    Why are you wrong? Well imagine you had a database of the full lifetimes of all crack addicts. Inside that database if you started analyzing all the ways it plays out you will find a lot of all those bad things…you aren’t wrong in assuming they are there. The dataset, though, will have some outliers. Some crack addicts did very well, at least up until they were done in (see Whitney Houston). And there will even be crack addicts who are functional all their lives, never get arrested, never end up on the street, manage to dodge all the potential pitfalls.

    But statistically you are safe saying crack addiction is likely to cause a lot of problems and end badly for someone. But how it will cause problems and end badly for a particular individual is probably impossible to predict. In contending with the addict whose intent on keeping your rationality at bay by distracting the issue, it is very easy to fall into making predictions that then don’t turn out which superficially will reinforce the addict’s ‘case’ that you are being alarmist. The 50th time someone buys crack and doesn’t get arrested and doesn’t overdose is 50 wrong predictions on your part….doesn’t that ‘prove’ you as wrong as wrong can be? Hmmmm.. (Cumulative probability is tricky here….the chance of getting arrested might be 1 in 1000. Getting crack 50 times and never getting busted can lead you to think you’re smart enough to be immune….but the reality is odds of doing it 50 times safely are high but if you do it every other day odds of NOT getting nailed within a few years are slim).

    This requires a more holistic view of things. If you take the set of things that can happen and are likely to happen they get scarier with a crack addiction versus no crack addiction. This alone then implies a reason to seek to avoid a crack addiction even though many people may get crack 1000 times and nothing happens in those 1000 times. Even though there might be some cases where things are better with crack!

    The global warming problem is somewhat similar. Raising the temps of the oceans implies trouble down the line. More hurricanes is a very plausible example of the potential trouble down the line. Exactly how many hurricanes, when and how strong? That misses the point, the unknowns might be less harmful than we fear but there’s no reason to assume they can’t be even more harmful. Not knowing exactly how harms will play out is even more reason not to mess with a system rather than an excuse to keep blindly messing with it.

  5. Mark says:

    Here’s the thing. Hurricanes and violent storms are not caused by “more warmth” but the temperature difference between hotter and colder climates. I’ve seen no data that show that temperature differentials would increase in a globally warmer world, in fact it might be the reverse, why do you think that there will be greater temperature differentials between the arctic and tropics? Raising ocean temps (and again the article stated the actual temperature but failed to note if this was unusual or the statistical variation of same which was my criticism), just means more water in the atmosphere. You won’t get high winds unless other areas are cold.

    That misses the point, the unknowns might be less harmful than we fear

    True. They might even be beneficial, which you seem to ignore. All change is bad? That’s an odd stance for a progressive.

  6. says:

    I can see two ways to get bigger temperature differentials. A cold spot could get colder or a hot spot could get hotter. If you warm something you’re going to probably start with a hot spot which will increase the temperature differential. You only avoid that if you manage to warm something by applying heat perfectly evenly to it.