The only thing we can be sure of about the future is that it will be absolutely fantastic. So if what I say now seems to you to be very reasonable, then I have failed completely. Only if what I tell you appears absolutely unbelievable, have we any chance of visualizing the future as it really will happen.

- Arthur C. Clarke (h/t Brin)

Monday, November 8, 2010

Curry Cannot Back Down

Judith Curry cannot easily take my advice and let her half-baked innovations about reasoning under uncertainty alone:
NG34A-03
308 (Moscone South)
Dec 15 4:30 PM - 4:45 PM
[note from mt -the above is to say, the annual Fall AGU meeting in SF]

Climate Surprises, Catastrophes and Fat Tails (Invited)
J. A. Curry

Low-probability, high-consequence events provide particular challenges
to developing robust policies to reduce vulnerability to climate
variability and change. Forced and unforced climate variability may
produce surprises that are abrupt or discontinuous, and extreme events
such as landfalling major hurricanes, floods, extreme heat waves and
droughts can have catastrophic impacts. Inadequacies of the IPCC
scenario simulations for conceptualizing surprises and catastrophes
arise from inadequate consideration of natural climate variability,
model inadequacies, and consideration of a range of scenarios of
natural and anthropogenic forcing that is much too narrow. Hence pdfs
constructed from climate model simulations have the potential to
seriously mislead, and a broader framework is needed to imagine future
abrupt climate change and extreme events. A method for evaluating the
possibility of future events is described that integrates the concepts
evidence support logic, modal falsification and possibility theory.
h/t James. She's got to say something then, whether she blogs about it between now and then or not. So she'd better get baking.


21 comments:

Lars Karlsson said...

I found the Italian flag!

Evidence Support Logic: A Guide for TESLA Users, p 6.

Michael Tobis said...

Lars, yes, Curry referred to that as a source document in one of her postings. But so far she is not using the tricolor in the same way as they do (or as she herself does for that matter).

Neven said...

AGU, isn't that just a union of ideologues?

Lars Karlsson said...

Yes Michael, for instance she confuses attribution (fraction) with uncertainty in her "Doubt" post as a number of commenters have pointed out.

Steve Scolnik said...

Might also want to point out for the peanut gallery that "pdfs" don't refer to Adobe documents.

Steve Bloom said...

"Inadequacies of the IPCC scenario simulations for conceptualizing surprises and catastrophes arise from inadequate consideration of natural climate variability, model inadequacies, and consideration of a range of scenarios of natural and anthropogenic forcing that is much too narrow."

Wowe. To paraphrase Al Gore, every bit of that is wrong.

Of course she can always just not show up.

But if she does, will she be ignored or will Big Names (or proxies therefor) show up to flame her?

Michael Tobis said...

The odd thing is that it sure seems to carry an alarmist spin. If it means anything it would seem to mean that extreme scenarios remain unaccounted for.

Will she say different things to different audiences?

Lou Grinzo said...

"Will she say different things to different audiences?"

As long as she can profit in some way -- monetarily or otherwise -- according to her personal utility function, yes.

Stephen said...

"A method for evaluating the
possibility of future events is described that integrates the concepts [sic] evidence support logic, modal falsification and possibility theory."

I'm curious if she's actually going to be presenting a "method", or just hand-waving.

Also, did she mean to say "model" falsification, or is "modal falsification" something I'm not familiar with?

Deech56 said...

Hopefully only slightly OT:

http://www.grist.org/article/2010-11-02-quest-to-keep-politics-out-of-climate-science-judith-curry

James Annan said...

Seems like she already gave this talk in Florida, which presumably means it's just vague waffle with no real content. The modal falsification stuff seems to be a rejection of Bayesian probability, with nothing usable to replace it with...

Hank Roberts said...

> Fat Tails

One tail or two?

http://scholar.google.com/scholar?hl=en&q="modal+falsification"

".... an informed decision would not only require the communication of all verified, falsified and articulated possibilistic hypotheses. It should, ideally, take into account an estimate of the potential to generate surprises for each of the respective options. On the one hand, this is obviously a weird idea, almost self-refuting and bordering on the paradoxical: We should assess, compare, and count statements we have not even articulated? Admittedly, I can't think of any detailed prescription for how to do that...." -- G. Betz
http://philsci-archive.pitt.edu/archive/00004490/01/paper_pluralpossibilities.pdf

"... The preferred methodology of the International Panel on Climate Change will be characterised as “modal inductivism”; it suffers from severe shortcomings which disqualify it for scientific policy advice. Modal falsificationism, as a more sound alternative, would radically alter the way the climate scenario range is set up. Climate science’s inability to find robust upper bounds for future temperature rise in line with modal falsificationism does not falsify that methodology, rather, this very fact prescribes even more drastic efforts to curb CO2 emissions than currently proposed...."

03/07
What range of future scenarios should climate policy be based on? – Modal falsificationism and its limitations -- Gregor Betz
Contingency And Dissent in Science Project, Centre for Philosophy of Natural and Social Science, The London School of Economics and Political Science

http://www2.lse.ac.uk/CPNSS/projects/CoreResearchProjects/ContingencyDissentInScience/DP/BetzOnlineModalFalsification.pdf

Hank Roberts said...

quoting a bit more from the latter:

" section 5 proposes modal falsificationism as an alternative methodology for setting up future scenario ranges. How this principle can be applied in scientific climate policy advice and where its
limitations are to be found is discussed ....
.....
I will distinguish two types of climate scenarios: global mean temperature scenarios which project the global average temperature change in the 21st century, and sea-level rise scenarios projecting global mean sea-level rise for the same period...."

cagw_skeptic99 said...

Dr. Curry has no reason to back down. Replacing the proselytizing approach common to inhabitants of this blog with actual scientific research will provide major benefits to humanity.

Halting the massive diversion of productive resources into carbon dioxide suppression schemes, and ending the similar diversion into green energy nonsense like windmills, will allow those resources to be used to benefit humanity.

Millions of people could have access to clean drinking water and living conditions not fouled by burning dung to cook their food for a small fraction of the money wasted due to the recommendations of the CAGW crowd.

Of course, none of you are likely to accept any responsibility for the suffering the policies you support are causing. You have a mission to save mankind and anyone who suggests that your mission is based on a fantasy must be some kind of heretic.

Michael Tobis said...

"Replacing the proselytizing approach common to inhabitants of this blog with actual scientific research will provide major benefits to humanity."

It is amusing that you think this is what Curry is doing. I think she is making stuff up.

Stephen said...

"Millions of people could have access to clean drinking water and living conditions not fouled by burning dung to cook their food for a small fraction of the money wasted due to the recommendations of the CAGW crowd."

Oh look, it's the ghost of Bjorn Lomborg.

Lazar said...

the two cents of a non-statistician...

modal falsification...

in their example of weak evidence (ice sheet dynamics), all the information contained in a modal falsification approach should be available in a standard bayesian analysis, with the added advantage that the bayesian analysis allows calculation of probabilities (!) aside from unity... the modal falsification approach is throwing away useful information... it appears much like jc's confidence interval interpretation of the italian flag as the limits of a pdf...

in their example of stronger evidence (climate sensitivty)... their argument against modal inductivism that climate sensitivity may be non-linear seems like a restatement of the induction fallacy... their method doesn't allow placing an upper limit on climate sensitivity... nor assigning probabilities... sensitivity may as well be 20C as 2C... who knows... i think this is nuts... somewhere (can't remember) jc stated that she thinks a credible pdf cannot be constructed for climate sensitivity...

their argument against bayesian analysis is that subjective priors influence the posterior especially when the evidence is weak... that at least shows that the evidence *is* weak... and allows bracketing... whereas their approach gives no information on how strong or weak the evidence is for their various 'possible hypotheses which cannot be falsified'... and hence no way to calculate their probabilities...

David B. Benson said...

Michael Tobis --- Yes, she certainly does MSU.

Lars Karlsson said...

Curry asked by Republican minority to testify at U.S. House of Representative’s Committee on Science and Technology Subcommittee on Energy and Environment hearing.

GM said...

The lineup for the House hearing tomorrow has been announced:

http://science.house.gov/publications/ hearings_markups_details.aspx?NewsID=2947

Panel I : Dr. Ralph J. Cicerone, Dr. Heidi M. Cullen, Dr. Gerald A. Meehl, Dr. Richard Lindzen

Panel II: Dr. Benjamin D. Santer, Dr. Richard B. Alley, Dr. Richard A. Feely, Dr. Patrick J. Michaels

Panel III: Rear Admiral David W. Titley, Mr. James Lopez, Mr. William Geer, Dr. Judith Curry

Lindzen got himself on board.

DaveMcRae said...

And she did Lars - and as you probably thought, not alot of sense, bugger-all facts, big on cherry picks (she pulled the 2035 date as proof that it's all not the IPCC)
http://science.house.gov/publications/Testimony.aspx?TID=15568