The Fallacies Of Risk

Some thoughts on their application to climate policy, in response to Judith Curry’s take.

As I noted on Twitter, two points. First, there really is no good physical case to be made that warmer global temperatures results in more extreme weather events. Storms are heat engines, driven by temperature differences, not total enthalpy. Also, I wrote about the fallacy of the precautionary principle as applied to climate policy four years ago.

24 thoughts on “The Fallacies Of Risk”

  1. A week or so ago Judith Curry had an entry about reliability and uncertainty which I thought was very good.

    At times, the statistical notion of reliability, or ‘statistical uncertainty’, dominates uncertainty communication. One must, however, seriously question whether the statistical uncertainty adequately captures the ‘relevant dominant uncertainty’ (RDU). The RDU can be thought of as the most likely known unknown limiting our ability to make a more informative scientific probability distribution on some outcome of interest; perhaps preventing even the provision of a robust statement of subjective probabilities altogether.

    You could have a group of experts saying they’re 95% certain of X, while another group of experts says there’s a 95% chance that the first group of experts doesn’t know what the f*** they’re talking about, which would be the relevant dominant uncertainty.

    1. BTW, that would have application to some of your writing on risk, NASA, etc. The uncertainty that you don’t understand something significant going on inside the engine, or with the O-rings, or with foam strikes, and thus can’t make a robust statement of risk no matter how much paperwork you’ve generated.

  2. Let’s talk about the Concorde supersonic jet.

    That design, in hindsight after a fatal accident, just didn’t have enough design margin in it.

    Owing to the kind of wing it had and the kind of turbojet engine, both required to cruise at Mach 2, it couldn’t quite sustain flight right after takeoff with two engines out on one side. All of this is intrinsic to the design and something one would know in foresight — I am sure this was known as part of the certification process for things like engine-out-on-takeoff.

    Also, there was probably some reason to put pairs of engines in the same nacelle, heck, the XB-70 Valkyrie had 6 engines in one nacelle, and the engineers had a reason, that design was supposed to lick the drag problem at supersonic speed using a kind of wave-riding technique called Compression Lift.

    I suppose pairing the engines opens up the possibility that if one engine self destructs from blade failure, blade failure from a bird strike, foreign object ingestion from the runway, the other engine may be wrecked too. Again, there must have been a design reason for this, but I recall both the Lockheed and Boeing SST proposals had individually podded engines, and both these companies had substantial civil airliner experience to know about such things, so the question could not have been an “unknown unknown” in the broader community.

    So what happened? A tire blows out on takeoff roll, maybe from striking a sharp object on the runway, and the pieces wreck the two engines on one side. That this started a fire may have meant the situation was hopeless, even if the pilots could have stabilized the flight to make the neighboring airport. But it was apparent that they could not maintain enough altitude being “behind the power curve” on that delta wing and with only two remaining engines.

    The Concorde was uneconomic and largely kept going as a matter of national prestige for a long time. I also read that the 9-11 terrorist attack killed a large number of their repeat customers, a different kind and unrelated aviation disaster, so they just pulled the plug on Concorde operations.

    Now if this were NASA in charge, they would continue to operate the Concorde, knowing about these serious design — not really flaws, just design attributes that leave very little margin for the things you cannot control very well. So they would start inspecting tires, increase the “FOD” sweeps of the runway — maybe clear a special anti-FOD truck to drive the runway right before clearing Concorde for takeoff. Maybe they would step up inspections of the engines and they would find more blade cracks, and they would mandate tear-down engine inspections before every flight.

    At the point where the unknown unknowns aren’t such unknowns anymore, and the O-ring concerns and the shedding foam were known in advance of the two accidents, NASA takes an unfixable design and attempts to fix it, using statistical models of temperature vs earlier O-ring near failures, statistical models of foam strikes. And all of the other workarounds of teardowns, refurbishments, and inspections regarding a whole list of other design flaws.

    It isn’t the unknown unknowns, it is a list of known problems with the design, and NASA practicing “safety” by putting a ton of money into workarounds.

    1. Most 4 engine planes can’t handle losing 2 engines on the same side.
      El Al dropped a 747 into Amsterdam when #3 failed and took out #4.

      the control margin gets real sporty after that.

      As for 2 engine pods, it was fairly successful on the B-52. You need some
      blast protection between the engines.

      Uncontained failure on any engine usually does a lot of bad things wether it’s a DC-10 losing
      control lines, a 747, bleeding along the wing, etc…

      The Concorde died from High fuel prices and aging airframes. They had a million feet of cable in the
      cockpit, that was just hard to maintain. Modern aircraft use a serial bus and fiber optic.

      1. The B-52 is a military aircraft where different trades are made between safety and performance.

        Also, the B-52 has 4 separate engine pods.

    2. The Concorde was uneconomic and largely kept going as a matter of national prestige for a long time.

      If I remember correctly, the British Airways Concorde service was operationally profitable, but never repaid its development costs because so few were sold after the US government banned supersonic overflights. As you say, the reduction in passenger traffic after 9/11 was what finally killed it.

      BTW, according to Wikipedia: “Post-accident investigation revealed that the aircraft was at or over the maximum takeoff weight for ambient temperature and other conditions, and 810 kg over the maximum structural weight.”

      So, you have an aircraft that’s at or over the maximum takeoff weight, hits debris on the runway, the tire explodes and starts a fire in the wing, it’s already travelling too fast to stop and has to take off, but it’s down to two engines working and one engine partially working and the undercarriage can’t be raised to reduce drag so it can accelerate to a safe flying speed, and the fire is melting the wing.

      I don’t think many airliners would survive that.

  3. “Storm are heat engines, driven by temperature differences, not total enthalpy.”

    Can you cite that?

    1. Surface winds on Venus, where the temperature is about 735 Kelvin, are about 4 mph. Wind speeds on Neptune, where the temperature is about 73 Kelvin and very little sunlight falls, are about 1,340 mph.

      Many of the most destructive hurricanes in the Atlantic were in the 1700’s, especially the great storm of 1780, which was during the Little Ice Age. Some of the worst winds on Earth are the katabatic winds of Antarctica and Greenland, where the extremely cold air can pick up monstrous velocities.

    2. What is it that you’re demonstrating your ignorance about? That storms are heat engines, or that heat engines are driven by temperature difference? This is basic meteorology and thermodynamics.

      1. ” heat engines are driven by temperature difference”
        ““Storm are heat engines, driven by temperature differences, not total enthalpy.””

        If you claim it’s Basic Meteorology, you should be able to cite that.
        It should be Textbook.

        1. It would really help reading your posts if you didn’t include random line breaks. As a benefit to you, it would make you appear more credible. You should ditch whatever program you are using to write your comments.

    3. As the temperatures increased over the past three decades, the Northern hemisphere wind speeds decreased. Nature article.

      If we take Venus, Earth, and Neptune as temperature and wind speed data points (I’ll halve Neptune’s top wind speed for its average) and do a simple curve fit, you get something like:
      average wind speed (m/sec) = 0.00204*T^2 – 2.1*T + 442, where T is in Kelvins. This predicts that the Earth’s average wind speed will drop about 15 or so percent for every 1C increase in temperature.

      It’s probably not valid, or perhaps there’s something to it in that wind is a response to pressure differences, which can result from temperature differences, and act to dissipate the differences by transporting energy, which is the sum of kinetic (wind speed), potential, and thermal.

      I bring this up because atmospheric scientists didn’t expect the outer planets to have much wind at all, because very little sunlight was available to drive weather phenomenon, yet what they found was the opposite, suggesting that we haven’t realized something extremely fundamental about wind speeds, which would help explain the Nature article where the scientists are all looking at real-world climate data and scratching their heads.

      1. Given the viscosity of the atmospheres on Neptune, and the gas giants, I wouldn’t want to
        scale them at all to earth.

        Venus, it’s 93 bar, so, if you despise Models of Earth’s climate, you are already 100X off
        any terrestrial data, and Neptune?

        I like your nature article, real science is full of surprises, however, the Nature article thinks it’s
        related to climate change.

      2. A 15 percent drop isn’t related to climate change, it is climate change, as in, the climate had one average wind speed and now it has a different one. But when they say it’s “related” to climate change, what they really mean is “it’s evidence that the gods are angry, but we just don’t understand why the gods are expressing their anger quite this way.”

        One of the amusements skeptics have is observing how “scientists” have to invoke supernatural spirits (Climate Change, son of Global Warming) in every article, making them read to us like scientific articles from the 1700 and 1800’s where the authors would work in a few words about how their findings reflect God’s majesty and the amazing wonderment of His creation. The bizarre references will forever and reliably date articles that make them, much the way an article written in the 60’s making reference to the Age of Aquarius, or the new consciousness, or peace and love as being obvious realities screams “this article was written in the 60’s – by a hippy.”

    4. Why would I waste my time looking up basic physics for an apparent ignoramus? If you don’t believe that storms are heat engines, or that heat engines are driven by temperature differences, provide your own citation.

    5. Can you cite that?

      Well this is meant for third graders, but here: Hurricanes as heat engines

      Now maybe a cite for 4 engine planes can’t handle a 2 engine failure? FAA and 14 CFR Part 135.373 and Part 135.383 states otherwise, but I’m sure someone as thorough and honest as you can provide citation as well?

  4. dn-guy: I’m not sure that the 2-engine pods on the B52 are relevant here. The reason is fairly simple; the B52 has a total of eight engines, so if one pod blows there is still thrust being applied to that wing. I would have thought that the relevant number is the number of distinct thrust sources (whatever the technical term is, I don’t know it but I think my meaning is clear) on each wing. A better comparison might be the comparison between Concorde and a two-engined airliner. Also relevant, I suspect, is the fact that the Concorde engines are embedded in the wing which is unlike any subsonic airliner currently in service.

    Regarding Neptunian winds, there are two points of difference other than temperature and temperature differences between Neptune and Earth. One is that Neptune’s atmosphere has an average molecular weight of maybe 4, being mostly hydrogen and helium with a smattering of heavier gases. Most of the molecular gases normally found on other planets, such as methane, are liquid or solid at Neptunian temperatures. All other things being equal, the velocity of sound and the average velocity of individual molecules are both inversely proportional to molecular weight, IIRC. This rather large difference is bound to affect wind speeds.

    The other difference is even more obvious; Neptune has no solid or liquid surface, or if there is one it’s a very long way down underneath the bits we see. That’s even more likely to affect the weather.

    1. Some Airliners have used the Dual engine nacelle, such as the IL-62.

      Some Bombers used the Dual Engine Nacelle i.e B-52.

      My take is that most airliners don’t do well if they lose all the engines on one side, wether it’s a Concorde,
      a 747, a 707, a 757 or 787.

      1. My take is that most airliners don’t do well if they lose all the engines on one side, wether it’s a Concorde, a 747, a 707, a 757 or 787.

        Since the 757 and 787 are twin engined designs, any engine failure results in losing all the engines on one side. They perform well enough to meet safety requirements given no other failures such as hydraulics. Losing power on takeoff is the most critical issue, being that the power required is at the highest because the weight is the highest and the airspeed and altitude are the lowest. A four engined plane like the A-340, A-380 or 747 can survive losing both engines on one side at takeoff. It’s little different than a twin-engined plane losing one engine – you’re dealing with a 50% loss of power at a time when the power required is high. Losing 3 engines at once is, of course, a more serious issue and likely not survivable at takeoff unless the crew can dump fuel quickly.

        The Concorde accident involved a piece of debris causing the tire to come apart. That caused a rupture in the wing fuel tanks and damaged the engines on one side. When the flight engineer shut down the second engine, the minimum controllable airspeed under those conditions increased to (IIRC) something around 400 knots at a time of high drag due to the low aspect ratio of the wings and apparently the landing gear being stuck down. Had he kept that engine running a bit longer, it’s remotely possible the crew could’ve made it to another airfield. Once he cut that second engine, they were doomed.

    2. “Regarding Neptunian winds, there are two points of difference other than temperature and temperature differences between Neptune and Earth. One is that Neptune’s atmosphere has an average molecular weight of maybe 4 …”

      More like 2.2. The atmosphere is 85% H2, 15% He, and less than 1% of everything else, mostly CH4 at the altitudes we can see (methane is a gas in Neptune’s troposphere).

      A third difference is probably the biggest of all — Neptune generates about three times as much heat internally as it receives from the sun.

    3. Given that planetary wind speeds go up as you get farther from the sun, there’s some thought that the atmospheres further out have less drag because they have less turbulence. CS Monitor article. I’ve also seen planetary scientists wondering if cold temperatures help maintain laminar flow and suppress eddy formation because there are fewer molecular collisions across adjacent velocity gradients.

    4. Okay, I have a notion that might explain why winds speeds would decrease with temperature.

      You have a planet with a heat imbalance due to sunlight, and what the wind is doing is transporting heat from the hot part to the cold part. Given a heating imbalance in W/m^2, the gas transports heat to the cold spot where it radiates away, so the wind is moving thermal energy at some calculated rate. The energy dumped is a function of the mass of gas being cooled, the gases heat capacity, and the temperature difference delta T. So the absolute amount of heat being dumped depends on the mass flow rate, delta T, and Cp.

      But given two equal delta T differences (and PV=nRT), the temperature induced volume difference (which translates into a pressure difference due to inertia and gravity) becomes greater as the absolute temperature decreases. ie. a delta T of 50 Kelvins doubles the volume of a 50 Kelvin gas, but only increases a 300 Kelvin gas’s volume by 16 percent. So at low temperatures, the gas is going to start moving at a much smaller delta T, and if the delta T is smaller then the mass flow rate has to increase to compensate.

      1. I’m still diddling with this at a really crude level, and let me point out that the above says that dV/dT = nR((1 + dT/T)/P, where V is volume, but of course P is a function of the planet’s gravity, atmosphere mass, and surface area, not adjusting for altitude.

        So if you set up a thought experiment of something like an aquarium that contains the atmosphere, and heat one side, the volume on that side increases, swelling it and making the top surface taller (the height of the atmosphere increases) so that the air at the top of the atmosphere slopes down toward the cold side. Ignoring a ton of stuff that shouldn’t be ignored, you get something like dV/dT (where V is wind velocity) = k * sqrt(dT/T) as you convert potential energy to kinetic.

        I think some simple thought experiments and a bit of math and code could actually spit out a pretty iron-clad formula just based on simple gas physics, and I think it would lead to a pretty nifty paper.

Comments are closed.