Bishop Hill An attack that is nothing of the sort
May 28, 2019 Climate: Models Yesterday, the New York Times got rather upset over changes to President Trump’s climate policy, which it represented a hardening of his “attack on climate science”.
Interestingly though, you have to read quite a lot of words before you actually get to the point – usually a sure sign that there is actually nothing much by way of news and quite a lot by way of hand waving. It turns out that Trump’s attempt to “undermine the very science on which climate change policy rests” is down to this:
[Director of the US Geological Survey,] James Reilly, a former astronaut and petroleum geologist, has ordered that scientific assessments…use only computer-generated climate models that project the impact of climate change through 2040, rather than through the end of the century, as had been done previously.
To describe this as an “attack” is obviously absurd. Reasonable people can question the ability of climate models to give us useful information about the climate in 20 years’ time, let alone 80. In a GWPF paperpublished last week, it was pointed out that climate models are overestimating warming in the tropical troposphere by a factor of three. With errors of that magnitude, how much trust can we really put in projections for the end of the century? You would have to be quite an innocent to take them at face value.
In another GWPF paper Professor Judith Curry points out that the climate may be fundamentally beyond our ability to predict it:
Arguably the most fundamental challenge with [climate models] lies in the coupling of two chaotic fluids: the ocean and the atmosphere. Weather has been characterised as being in state of deterministic chaos, owing to the sensitivity of weather forecast models to small perturbations in initial conditions of the atmosphere…A consequence of sensitivity to initial conditions is that beyond a certain time the system will no longer be predictable; for weather this predictability timescale is a matter of weeks.
To describe the President as “attacking” climate science when he doubts projections out to the end of the century is therefore clearly nonsense. Indeed, he should probably be congratulated for recognising the powerlessness of the field in the face of an overwhelmingly complex climate system.
Reader Comments (13)
Right now on Channel5 TV : Sinking Cities: London
The expert voices speak about the way the London land is sinking
..but the voice over goes on about Climate Change
\\ Documentary following the work of experts finding ways to protect cities from rising sea levels caused by climate change. Cameras focus on London, a metropolis that faces two rising environmental challenges - increasing levels of rainfall and being built upon a floodplain - while the Thames flood barrier may no longer be enough to protect the city from flooding. Plus, a look at the development of floating buildings. // Radio Times
Daily Mail:
Trump moves to de-fang global warming science with ‘climate review panel’
led by Princeton physicist who believes
'The ‘demonization of carbon dioxide’ is like ‘the poor Jews under Hitler’
Climate Scientists seem rather more sensitive than their delicately balanced computer models that always over react.
Is it known whether the New York Times required UN IPCC approved assistance, funded by the USA, to write this?
Trump already has questions to ask May on D-Day about British Intelligence and Dodgy Dossiers. He must have questions about the East Anglian intelligence that was processed into a Hockey Stick, particularly the dodgy Chinese stuff that mysteriously disappeared
I first started looking at projections a century ahead when I analysed the Limits to Growth model back in the 1970s as part of a small team commissioned by the Rothschild Think Tank. Among the many issues that arose was the limitation of predictive accuracy of fourth order Runge Kutta solutions to systems of differential equations that underpinned the DYNAMO language solver. This was a topic I had already looked at for A level maths (things were a bit different back then), and it was not too hard to show that the stability of solutions was in doubt due to the algorithm and rounding effects. Of course, there were other major failings of model design including incompleteness, over-simplification and parameterisation, and big issues with much of the data. It was an apprenticeship that has served me well ever since, and you could say a model for tackling climate "science".
Later, I encountered the work of the Santa Fé Institute on complexity theory and chaos, which put much more flesh on those early bones. It is perhaps surprising that climate modelling hasn't been properly subjected to the same kind of testing except largely by outsiders on all the issues that arose over LTG.
The GCMs are (and have always been) complete bosh.
This is not the first time a large (and clueless) swath of the populace has been deluded by unverified computer models of an immensely complex (possibly chaotic), non-linear, dynamic system where the independent variables and their very coefficients are unknown.
The flaws in predicting a chaotic system over a century do not require rigorous proof for most people. For policy purposes it isn’t worth the effort. After all, at first glance it’s obviously a bit silly for many people.
This seems like a disconnect between experts and the curious lay folk.
It’s just too much work for most people to prove the predictions don’t work. Not that it’s beyond them but rather that it’s not their core purpose in life so why bother? Instead they adopt one of three strategies:
A) Trust in the experts. The issue isn’t urgent and, so long as it’s not too costly, this is not as foolish as it sounds. You will only lose a little in scams but you won’t lose a lot in wasted time and effort.
B) Just ignore it and hope it goes away. This is not a callous as it sounds. Most things get fixed in a century and it isn’t everyone’s role in life to do the fixing.
C) Benchmark it against something you are more aware of. For most people that’s their bank account. If you think the market is predictable in 100 years’ time then you may believe that the CO2 emissions will roughly follow it and that the temperature will as well. But how much confidence does anyone put in any prediction over that time frame? It varies. But the condition is one that people are estimating for other reasons so there is no wasted effort.
None of these strategies actually involves knowing anything about climate science or much mathematics at all. But they are all coherent.
Judith Curry's criticisms are valid, but have been known, and ignored, for decades. Even the IPCC reports occasionally admit it. (The same problems also crop up in other scientific disciplines.)
In a valedictory address, Edward Lorenz, the Godfather of computational climatology, cautioned his peers to only work on tractable problems.
They ignored him, and went for the money instead of good and credible science.
It was an attack on their religion.
This statement: "In another GWPF paper Professor Judith Curry points out that the climate may be fundamentally beyond our ability to predict it". I am very aggree with that, although now we have modern technology but climate is so hard to predict. My opinion maybe with technology (now), truth of climate prediction is 40-60%
Unverified models are not competent to predict the climate next month, let alone 40 or 100 years.
The GCMs are (and have always been) complete bosh.
This is not the first time a large (and clueless) swath of the populace has been deluded by unverified computer models of an immensely complex (possibly chaotic), non-linear, dynamic system where the independent variables and their very coefficients are unknown.Also read this
This claim: "In another GWPF paper, Professor Judith Curry raises the possibility that our capacity to predict the climate may be fundamentally limited." Although we now have advanced technology, the weather is still highly unpredictable.
In my perspective, the accuracy of climate predictions may be 40–60% given current technologies.
My objective reality is that I experienced my first quantum leap back in 2013 using Hf 156/72 and OH* to unlock the alpha particle decay timing of 2.5 times the speed of light for 23 milliseconds to give me a leap of around 43 years up until 2055.