Climate Catch 22: big bucks, big bets and big mess

Climate modelling is at a tipping point. Desperate to counter increasing assaults on its reputation, the discipline is making a controversial new pitch. It is promising (one day) to overcome the “scale challenge” and produce accurate “fine-scale (regional and local) climate projections” (not predictions).

But there is a “truly scary” catch:   millions of additional research dollars will be needed to fund soaring computational costs, deal with a looming data-storage nightmare, and reverse the “worrying” decline in quantitative and technologically capable graduates, more of whom prefer investment banking to trying to save humankind from itself – and alleged climate oblivion.

The kookaburras were laughing outside the University of Western Australia’s Alexander Lecture Theatre.  Inside, an audience of about 80 heard much ado about “the decisions we make as a species”, how the science of climate change “was nailed 50 years ago” and lamentation over the “complete disconnect between public understanding and the science”, as Professor Andrew Pitman, Director of The ARC Centre of Excellence for Climate System Science at the University of New South Wales, presented this year’s UWA Joseph Gentilli Memorial Lecture last month.

Pitman, a lead author on the Intergovernmental Panel on Climate Change 3rd and 4th assessment reports and Review Editor of the 5th assessment report (AR5), scheduled for release on 27^th September, kicked off with a riff on atmospheric carbon dioxide content. It was now “way above the natural level of 300ppm” at about 400ppm.

IPCC’s AR4 (2007) expressed “considerable confidence that climate models provide credible quantitative estimates of future climate change, particularly at continental scales and above”, 50 to 100 years away.

Pitman’s enthusiasm for this view was no surprise. ARC’s focus is developing “extraordinarily sophisticated” coupled climate models “because crystal balls don’t work very well”. The problem is the models are not working very well either.

One big elephant in the IPCC room is the (unpredicted) global temperature standstill since the mid-1990s; According to a recent post by Barry Brill, it relies on an array of models “without any curiosity about why they have been so wrong for so long.”

IPCC lead author Hans von Storch is, ironically, on the same page. He recently told Der Spiegel that ”if things continue as they have been, in five years, at the latest, we will need to acknowledge that something is fundamentally wrong with our climate models...”

There are other issues too. John McLean, is 95% certain the IPCC is untrustworthy. For him, scientists who claim they "95% sure" humankind is to blame for climate change are guilty of “a conglomeration of fraud, denial and lack of logic”; due to the way IPCC reports are “written, edited, re-written and then, at the very end, apt to be recast all over again in the interests of politics and political expedience.”

Pitman and his colleagues know that to remain relevant - and to justify the global billions of climate research dollars spent annually - they have to offer something new. Alarmist rhetoric about vulnerability to climate change in half a century and beyond will no longer do the trick. They have to convince governments that more grant money will deliver something useful, such as “much more detailed temporal [daily, hourly] information around extremes”.

For him, the basic science is “settled at the big picture scale”. The “phenomenally interesting” challenge now is to re-engineer global simulations to make predictions at the catchment, regional and paddock scales.(43.40min.)

According to other researchers, however, there is a long way to go. Extreme weather event (EWE) analysis is a tricky business. Reluctant to be caught out over-promising (again), the orthodoxy admits it also should “manage expectations”.

Even the UN-funded World Climate Research Programme describes the quest for a EWE science as a "Grand Challenge"; with its “climate information service” concept only at "first draft" stage.

Any improvement in predicting changes in EWE frequency and intensity would require “improved representation of key processes in climate models” and resolution of other complex issues; with “much work needed to take careful account of uncertainty when delivering forecasts of extremes [EWEs] to users” (Karoly, WGSP, 2012, white paper, I3).

Nevertheless, ARC is claiming success in its recent EWE research.

AP: “ARC has now got to the stage of being able to attribute specific extreme events to specific causes. For instance, the heatwave event that hit Australia in January, we can now show that it wasn’t possible without the additional warming associated with global warming [DAGW]” (35.41min).

People often say you can’t attribute a single event to DAGW. That is a myth. It is not true. It just takes a year or so of bloody hard work to do it. And you cannot do it when the media asks you fifteen minutes after the cyclone has hit whether that cyclone was linked to DAGW. You have got to do the year’s work.” (36.20min.)

I asked Pitman if there was any research documenting successful prediction. Could he expand on his closing remark: “We’ve done all that – prediction”?

Apparently not; he provided links to only these two research papers: “Anthropogenic contributions to Australia's record summer temperatures of 2013” and “Local sea surface temperatures add to precipitation in northeast Australia during La Niña”.

Yet explaining away a EWE a year after the event, - as these papers do – is a different exercise from predicting a specific EWE. And “probabilistic” statements - such as “EWEs will be more prevalent in the future because of dangerous anthropogenic global warming (DAGW) and climate change (DACC)” are unhelpful.

Comparing “simulations with natural forcings” to “simulations with anthropogenic and natural forcings”, assumes both can be identified - and quantified – easily; and that climate models doing the simulating accurately describe reality. But is it, and do they?

Not for Robert Pindyck. As he recently explained in US National Bureau of Economic Research working paper 19244, Climate Change Policy: What do the models tell us?: “the physical mechanisms that determine climate sensitivity involve crucial feedback loops, and the parameter values that determine the strength of those feed-back loops are largely unknown.”

The night’s big surprise was Pitman’s admission that the orthodoxy is facing some “truly scary” challenges.

Professor Pitman:"What are the problems ahead? (37.20min.) Firstly, everything we do is extraordinarily computationally expensive. Not as expensive as the astronomers, but we are working really hard [to get there] (laughter).”

“Just as a guide, if you divide the globe into one degree-by- one degree cells – about 65,000 cells – and you run it out for 100 years, it means 40,000,000,000,000,000 calculations.

 “The second problem is we create lots of data. For AR4 in 2007 it was 35 terabytes. For AR5, it is about 30 petabytes. That’s getting to be really seriously worrying. For the next IPCC report, we are not quite clear how we are going to cope with it.”

“Here’s our third problem. We have a big demand for people. We desperately need people who can manage million-line versions of computer code, or learn how to slice-and-dice 50 pedabytes of data...Somehow, we have to bring lots more people through with those skills (41.20mins)

It takes a full-scale system now to hope to cope with the scale climate science demands of its science and what policy-makers and the public demand of the climate scientists. It is no longer a thing two or three people sitting in their offices can get away with doing.

What we really need is lot more kids getting into science and technology to solve the problems my generation is leaving them with. If we lose another generation to lawyers and accountants, we will seriously struggle to stay up to speed in these areas in the future.”

The orthodoxy had to lift its game too, for:

“the demands coming from the impact and adaptation communities to give them the information they need, puts immense pressure on the climate scientists to do what they have been doing much, much better and cleverer in the future” (48.23min.).

 When Pitman mentioned the 40,000,000,000,000,000 calculations, my head started spinning. What if there was a glitch and just one of them was wrong? What about all the dummy variables and assumptions in his 65,000 cells and so on?

Climate Catch-22 Syndrome (CCS) hit me half-way home, when I heard a voice. It was not the one that spoke to Apostle John in The Cave of the Apocalypse. The voice I heard was that of B-25 bombardier, Captain John Yossarian:

“There was only one catch and that was Catch-22...He would be crazy to fly more missions and sane if he didn't, but if he were sane he had to fly them” (Catch-22, p. 56, ch. 5).

For me, CCS arises whenever empirical data confirming cooling is proposed as evidence for – or “is consistent with” - warming, when no amount of contrary evidence can overturn the DAGW paradigm, when “missing” heat hides in the deep ocean, when guesstimates morph into predictions and so on.

One climate scientist, clearly suffering from CCS too, concluded that any “finding that human influence has not contributed substantially to the magnitude of a particular EWE may not be incompatible with a finding that human influence substantially altered the odds of such an event happening (especially a particular threshold exceedence).” Catch-22.

Another group (Perlwitz, et al) insisted that “some cold events are consistent with the inter-play of on-going global warming and internal variability”. Catch-22.

Want to play the climate game? Make sure you can explain away any meteorological or climatic outcome, even if you have to resort to ad hoc hypotheses. And discourage any attempt to falsify your claims.

CCS symptoms returned while reading the latest article by ARC affiliates, Ailie Galland and Sophie Lewis, also posted at The Conversation.

There are, they confided, “bits of the [climate] puzzle we don’t even realise are missing yet. These are “unknown unknowns”. There are bits of our puzzle at the limits of our reasoning or modelling, so their uncertainties can’t be quantified”.

Yet they went on to suggest probability analysis allegedly “allows “scientists to communicate findings more precisely and transparently”. Or is it merely an exercise in obscurum per obscurius?

Paradoxically, despite noting “that the only scientific certainty is uncertainty”, they were uncomfortable with scepticism because: “at times, these inevitable scientific uncertainties have been framed as synonymous with doubt and used to try to discredit findings.” Catch-22.

Socrates encouraged pursuit of truth by open discussion and free enquiry. In the rarefied world of climate computer games, however, the orthodoxy seems as keen to avoid scrutiny of its alarmist claims as it is to ignore the uncertainty monster - and its offspring - banging on the door.