Tuesday, December 12, 2006

By 2040: an Arctic with no ice?

An article on The Times Online by Lewis Smith entitled "By 2040: an Arctic with no ice?" discusses recent "calculations" concerning how quickly the "summer ice" of the Arctic ice cap may melt. Previously scientific models had suggested it would take another 65-75 years (not until 2070-2080) before the summer ice would be completed melted, but more recent simulation models are suggesting that the summer ice will be gone in 30 years.

How alarmed should we be?

Well, not so fast. Slow down, and consider what we have to work with.

These are simply "simulations" or "calculations" based on "models".

If the science had been all settled, why yet another new model?

And why should be believe that this new model is really the final, ultimate, end-all model?

The models are certainly interesting, but how well do they really forecast the future? And decades into the future at that?

According to Marika Holland who led this latest study:

We have already witnessed major losses in sea ice, but our research suggests that the decrease over the next few decades could be far more dramatic.

Wait a minute... how solid is this research if it merely "suggests" what "could" happen?

I am all in favor of funding the scientists to continue research into how the atmosphere and climate work, but it is simply not credible for any of these scientists to "suggest" that they truly fathom the complex mechanisms of planetary geology, continental land masses, foliage, oceans, the atmosphere, solar heating and cooling, climate, and weather systems with sufficient rigor to be able to produce relatively simple models that can forecast ocean, atmosphere, and climate effects decades into the future. It simply isn't credible.

By all means, let us encourage the scientists to produce ever more sophisticated models, but it is simply far too soon to base any governmental, economic, or environmental policies on any of the current, dubious "research" models.

-- Jack Krupansky

5 Comments:

At 6:30 AM , Anonymous Anonymous said...

Interesting comments. As an economist, I am interested in the realism or otherwise of models. Mainstream economists (following Milton Friedman) tend to assume that assumptions are unimportant, and that the quality of a model is decided by its ability to predict outcomes. I disagree with this approach, and think that assumptions do matter -- that complexity and the changing nature of society makes it difficult to capture reality within a simplistic model.

That said, many scientists will not see it as a damning criticism to be accused of inventing a model that is very simple. These particular scientists presumably believe that they have isolated the important causal factors in climate change, and they may have found that their new model adequately explains previous changes in the polar ice cap. Of course there is a degree of uncertainty about the future, but this model may constitute their best estimate.

 
At 6:35 AM , Anonymous Anonymous said...

Whoever said 'all the science has been settled'? Do you know any scientists who have ever claimed such a thing?

We need yet another new model' precisely because our understnading is not yet complete, and we are always discovering new things about the contemporary world which mean we need to refine our predictions of the future. Are you really suggesting that we should now stop trying to find out about how the world functions and building better models to reduce our uncertainty about the future?

Remember, the ice sheets are melting FASTER than models predicted 10 years ago. Obviously, this has no bearing on whether the predictions in question are accurate, but bear in mind that they may be under, as-well as over-estimates of the actual rate of change.

 
At 6:44 AM , Anonymous Anonymous said...

The question whether we should base policy on a model does not depend only on the reliability of the model. Krupansky says that the model is not reliable enough because (a) it corrects a recent model, and (b)the modeling covers some seriously complex factors and it is doubtful that the scientist can fathom them at this point. Let us suppose Krupansky is right, and the model is probably at least marginally wrong, the question whether we should base policy on it depends at least partly on how soon we are going to suffer damages from the consequences predicted by the model if we don't act. We could imagine the following situation: we cannot confirm the model as reliable before the 2030s, but only if we act before the 2020s are we going to be able to avoid the worst consequences. If the situation was of this kind, it is obvious that the most reasonable course of action would involve changing our policy concerning carbon pollution as soon as possible. This shows that prudence, and not the mere reliability of scientific models, should play a major role in our deciding policy.

 
At 2:26 PM , Anonymous Anonymous said...

How alarmed should we be? Very alarmed I'd say. 2040? 2080? Who cares exactly when. The point is that something needs to be done NOW to prevent global disasters such as this from occuring don't you think?

 
At 7:57 PM , Anonymous Anonymous said...

I think it is comfortable for you the be ignorant.
As an economist/or an opportunist you probably see resorts in north of Canada where is ice right now, but i'll tell you that you will be dead by then, and if not I'll make sure i will get there before any of you so called tday's economist.
the future jerk.
By the way, FUCK YOU.

 

Post a Comment

Subscribe to Post Comments [Atom]

<< Home