Climate models in perspective

8 02 2015
Mark Cochrane

Mark Cochrane

Another guest post from Mark Cochrane.

One of the favourite refrains from climate change ‘skeptics’ is saying that models are wrong, which is a bit like saying that water is wet. Models are simplifications of more complex systems and as such they are always going to be ‘wrong’ but the question is whether or not they are useful. Weather models are wrong too but we use them all of the time. Sometimes they over predict precipitation or temperature while at other times they under predict.  The Weather Channel actually uses ‘proprietary methods’ to over predict the chances of high precipitation (link), presumably because they’d rather be wrong in one direction than the other.

The ‘skeptic’ claim is that climate scientists do something similar with their projection of future climates for a range of probable emissions scenarios. Somehow hundreds of scientists across multiple groups all manage to collude in this mass delusion to make us worry more than we should about the future impacts of our fossil fuel dependent culture.

Favorite climate skeptic Christopher Monckton and colleagues recently published an article (link) saying as much in the somewhat obscure Chinese journal Science Bulletin. I’ve never heard of this journal before but they are making an attempt to be in the peer-reviewed literature so I won’t quibble. They make the case that the IPCC model scenarios are overstated by a factor of 2 or 3. If you want an in depth summary and critique of this paper see this blog post. Relative ranking of journals is often done using their respective Impact Factors; their likelihood to yield citations to any given work. Scientists aim to get their works into the best location for their work to be read and used by their peers. Science Bulletin has an impact factor of 1.4, Nature has one of 42.4.

Now an analysis has been done to test whether model misses of the so-called warming ‘hiatus’ are due to having incorrect model forcing (sensitivity to greenhouse gasses – e.g. CO2) or other random factors (mostly volcanoes and ocean modes – e.g. El Nino etc.). The paper in Nature by Marotzke and Forster (2015) does this by comparing model simulations used by the IPCC and observations of global mean surface temperature as functions of all possible 15-year (and 62) trends from 1900 to the present.

An in depth explanation of the paper can be found here.  The use of 15 years is somewhat arbitrary but it provides a test of using short time periods to get whatever trend is desired. If the nefarious climate scientists are really making their models over predict the rate of warming then when you plot all of the differences between the model outputs and the actual observations they would tend to be greater than zero. This is the pattern you see from 1998-2012 (middle below), the period of the so-called hiatus or pause. However, you see the exact opposite pattern from 1927-1941 when models would have been accused of seriously under predicting temperature increases (left panel below). When you plot all of the data (right panel) the results are not significantly biased either way.

Comparing model-simulated (brown bars) and observed global surface temperature (vertical black line) for the 15-year periods covering 1927-1941 (left) and 1998-2012 (middle), and for all 15-year periods between 1900 and 2012 (right). Source: Marotzke and Forster (2015)

The models are not systematically biased in either way, they just cannot account for chaotic or random factors within the atmosphere and ocean. Therefore they will seem to over predict and times and under predict at other times. Where are we now?

Still cruising along the lower portion of the 95% range of the model simulations. If it were a blood test you still wouldn’t have a little asterisk on your test but it would be close.

It is noteworthy that this apparent inaccuracy only pertains to ‘surface temperatures’ of the Earth. There are no indications that the rate of warming has actually slowed down in any way. The only change has been in the amount of energy being stored in the world’s oceans. Despite our terrestrial surface bias of measuring the climate, over 90% of the energy is piling up in ocean waters, with much of the heat being transported into deep reaches (>700 m) as chaotic processes lead to periods of ocean water turnover.

The wiggles in the ocean uptake (blue above) of heat end up as periods rapid warming or a ‘pause’ these days in surface temperature. Note, back before 1970 those down turns would cross beneath the land component and yield actual global cooling for short periods. Ocean processes call the tune of energy transfer and the atmospheric and surface temperatures dance to it on short time scales. None of this changes the steady ramp of ongoing energy storage in the long term trend due to green house gas accumulations that is driving global climate change.

Advertisements

Actions

Information

5 responses

8 02 2015
mikestasse

There’s a good reason why Mark hasn’t heard of Science Bulletin. It just (and by just I mean in 2015) changed its name from The Chinese Science Bulletin. Hat tip to Doug from Chris Martenson’s blog where he wrote this…..

http://en.wikipedia.org/wiki/Chinese_Science_Bulletin

Wikipedia quotes Journal Citation Reports as giving it an impact factor of 1.321. Under the new name the Monckton article appears to be its first publication.

I don’t know what the new name means, but under the old name it was apparently a reputable journal, perhaps just not circulated outside China. However, his co-authors, Willie Soon and David Legates set off alarm bells.

http://www.desmogblog.com/david-legates-asked-step-down-delaware-state-c

Quote:
David Legates announced this week that he was asked to step down as Delaware State Climatologist, a position he held for seven years. A long-time denier of the human contribution to climate change, Legates’ tenure as State Climatologist has always been a controversial one.

http://www.greenpeace.org/usa/en/campaigns/global-warming-and-energy/pol

Quote:
Of all the climate deniers, one scientist has been particularly closely involved in the campaign against the climate science consensus for the majority of his career: Dr. Willie Soon.
UPDATE Feb. 2013: New FOIA results show new funding to Willie Soon from secret Donors Trustnetwork, Koch brothers and Southern Company.

Click here to see the fossil fuel financiers of Willie Soon’s career.
Introduction:

This investigation shows that Dr. Soon has received substantial funding from the fossil fuel industry for most of his scientific career and heavy corporate funding in the last decade.

Quote:
Willie Soon’s Backers: Big Coal and Big Oil

Since 2001, Willie Soon has received direct funding for his research of $1.033 million from Big Coal and Big Oil interests. In contrast, he received $842,079 from conventional government or university funders in the same period. The last grant he received from a funder with no ties to dirty energy interests was in 2002 (a grant that carried through to 2006). Since then, he has been entirely funded by the fossil fuel industry.

8 02 2015
davekimble3

CMIP5 represents the combined output of 35 different models, each of which has been fine-tuned so it gives the closest fit to the actual temperature data up to 2013.

The four RCPs (forecasts for future ghg concentrations) are all calibrated to match the various ghg concentrations in 2013 (maybe 2010?), so by 2015 they haven’t had time to diverge very much, but they become very much different from each other as the decades pass. So it is a pity that the chart uses all RCPs combined – using just one RCP would make more sense, and the lowest one, RCP2.6, would make a lot more sense to peakists.

RCP2.6 forecasts rates of fossil fuel extraction that are still too high for peakists – by 2070 coal extraction rates would be 116% of 2010 levels, oil 32% and gas a totally unbelieveable 320%.

Nevertheless, it is surprising that the Hadley data is in danger of going outside (on the low side) of the 5 – 95% range of all models and all RCPs by 2020. It would be less divergent if RCP2.6 alone had been used, and even less divergent if a peakist RCP had been used.

While no one currently has any logical reason for eliminating any of the models, it is surely about time to create a new composite model (call it CMIP5*) that exclude the models that are giving consistently high predictions, and a more realistic RCP (call it RCP2.*). Then we would be able to see through the fog of uncertainty a bit better.

As it is, we can say that based on RCP2.6, one the models is predicting a temperature in 2037 of only +0.15 °C over 2013 levels and falling very slightly after that.

8 02 2015
rabiddoomsayer

In running multiple models, multiple time there is an assumption that we have a much greater reliability. Inherent in this assumption is the belief that the models are independent. This is not at all true, the models all share common understandings of known physics and expected responses and in some cases common subprograms.

The models are not in any way independent. Science of Doom has some interesting material this matter. http://scienceofdoom.com/

Climate Sight has some interesting materials on climate models http://climatesight.org/ Miss Kate has done some serious work on climate models, she would have to be considered a supporter of climate models but none the less you can still get a feel for the limitations

In both cases I have linked to the blog general and not specific articles.

Assemblages hide the extreme events that come out at different times in different models. Assemblages smooth the transitions despite the known clunkyness of climate change.

There are many unknowns in climate change, it would be foolish to believe models have captured the full range of possibilities. None the less they are useful, but accept the is a considerable error margin. Contrary to Dave, I suggest that the models tend to underestimating the degree of change. We have changed forcing well over an order of magnitude faster than anything in the paleo record and yet expect the same or slower pace of change. If you look at the paleo record upswings have been much faster than downswings.

11 02 2015
Idiocracy

Good old Lord Monckton… it’s amazing how hereditary priviledge make’s you an instant expert in anything of concern to you (i.e. sustaining the wealth/prosperity of the Peerage and their United Kingdom) and can even get you published in esteemed Scientific Journals of Communist China! 😛

I bet he was next on Abbott’s list to be a Knight of the Order of Down Under!

What a sorry state this world must be in for such nong’s to wield so much influence and get so much air time…

13 02 2015
MargfromTassie

Yes, it is ironic indeed that Monckton publishes his paper in a Chinese journal, when he has said so often that the ‘global warming hoax’ is a giant Communist conspiracy.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s