backtop


Print 74 comment(s) - last by texbrazos.. on Dec 29 at 11:40 AM

Paper questionably assumes no add'l feedback to warming will occur, drops last five years of ocean data

Markus Huber and Reto Knutti, researchers at the ETH Zurich's Institute for Atmospheric and Climate Science, earlier this month published an interesting research letter in the peer-reviewed journal Nature Geoscience, entitled "Anthropogenic and natural warming inferred from changes in Earth’s energy balance".  This letter is significant, as it appears to be the first climatology work to try to extensively model warming from an energy balance perspective.  This represents a major step forward in trying to understand a model global warming.

Ultimately, the paper tries to assign a percent blame to mankind for the current warming.  However, there is good cause to debate the validity of these final conclusions, given some of the paper's rather naïve methodology.  It is here that the paper falls back on the mistakes of some of its predecessors in perhaps oversimplifying the system.

I. Mankind is Contributing to Warming the Earth

Whether it's one trillionth of a degree Celsius or one degree Celsius, there's one thing most scientific-minded observers of the warming debate can agree upon -- mankind is responsible for some part of the Earth's recent warming trend.

Basic physics tells us carbon dioxide is a greenhouse gas, as is methane.  Mankind has emitted a lot of carbon dioxide and methane in the last 100 years, and atmospheric monitoring has shown a discrete rise in levels of these greenhouse gases.

The greenhouse gases trap heat.  So mankind is clearly responsible for some quantifiable amount of the warming effect.

Energy balance
Global warming is symptomatic of either an excess incoming energy flow or an deficient outoging energy flow, thus the energy balance is a good means of examining warming.
[Image Source: Learner.org]

The important questions are:
  1. How much of recent warming (%) is mankind responsible for?
  2. How far can the system go before reaching equilibrium?
It is here, though, that the logical "skeptic" and some climatologists diverge.  Some researchers claim that both of these questions have been, more or less, definitively answered.  Scientifically-grounded skeptics, on the other hand, argue that much of the current work has been simply in fitting weak models to data sets, leading to potentially misleading conclusions.

The new work, for its part, does do an admirable job pointing out some of the shortcomings of the current modeling approach.  The authors write:

The optimal fingerprint detection and attribution framework provides a rigorous, statistical method to quantify the contributions of different external forcings and internal variability to the observed climate changes. In essence, it is based on a regression of the observations onto model simulated patterns and relies on the spatio-temporal response patterns from different forcings being clearly distinct. The assumptions are that climate models simulate the spatial patterns reasonably well and that regional responses from different forcings can be scaled and combined linearly. The global energy budget is not necessarily conserved and observed changes in the energy budget are not considered. Previous studies showed that observed patterns of surface air temperature provide a constraint on the human contribution to the observed warming.

In other words, past studies have taken temperature observations, assumed that we know exactly how much temperature increase is caused per unit forcing (e.g. per unit CO2), and linearly combined forcings to find a model that fits current temperature rises.

Forcing inputs
The new work attempts to combine forcings linearly and assumes no additional feedback, much like past works. [Image Source: ETH Zurich]

II. Looking at the Energy Balance

The new work alters this approach slightly, by using the energy balance, rather than the temperature, as the target to fit to.  But ultimately its conclusions may be skewed by the fact that it, like its temperature-based predecessor, uses a simplistic approach in which inputs' contributions:
  1. Scale linearly
  2. Are additive
  3. Are of equal efficacy (from the paper: "We assume that all forcing agents have equal efficacy.")
This is a rather simplistic model, although the analysis was complex with "thousands of model simulations".

The paper's ultimate conclusion is that:

Our results show that it is extremely likely that at least 74% (±12%, 1σ) of the observed warming since 1950 was caused by radiative forcings, and less than 26% (12%) by unforced internal variability. Of the forced signal during that particular period, 102% (90-116%) is due to anthropogenic and 1% (-10 to 13%) due to natural forcing.

The authors conclude that man is responsible for 74 percent of warming.

Temperature trends
The paper shows that the model was fit well with recent historic temperatures.
[Image Source: ETH Zurich]

The blame
The paper claims that mankind is responsible for 74 percent of warming via greenhouse gases,once the offsetting contribution of aerosols is considered. [Image Source: ETH Zurich]

This is a fascinating conclusion, because if correct it provides a much more straightforward assignment of blame to mankind, using what appears to be a more scientifically sound methodology (i.e. the energy balance).

III. The Blame Game -- Maybe Not Quite So Accurate

Unfortunately there are some significant issues with the paper which cause its conclusions to be brought into question.  

One notable is the issue is whether the scaled forcings are truly of "equal efficacy".  It's quite possible that different forcings could operate fundamentally differently, given where they put their heat (e.g. atmosphere v. ocean, etc.).  Problematically, the paper's authors fail to provide much explanation of why they think this approach is truly valid.

Secondly the study's methodology section states that another potentially problematic assumption was made:

It is assumed that the feedbacks are constant over time and the forcing uncertainty can largely be captured by a time-independent scaling factor.

This gets back to question 2 of the big questions previously raised -- "How far can the system go before reaching equilibrium?"

The model essentially is assuming that the Earth is going to sit here warming, without any significant cooling counter-effect being produced.  Geological record, as well as scientific common sense would cast doubt on this notion.  To give on example, past warming has altered sea currents, leading to a net cooling trend.

Finally, the paper admits that something strange is going on heat balance-wise with the ocean:

Ocean heat uptake for 3,000m depth is also larger, but the model is only constrained using data to 700m depth. Furthermore, uncertainties in ocean heat uptake are large and differences between various reconstructions are significant. The near constant ocean temperature over the past five years are not simulated by the model and its causes remain unclear.

For some reason the ocean temperatures have flatlined.  But the papers ignore this data, opting to only use the data that indicates nice, homogenous warming.

Observation comparisons
The energy balance results are compared with observational data, but the researchers questionably neglect comparisons for the last five years of ocean data.
[Image Source: ETH Zurich]

Thus, at the end of the day while this paper makes some progress in adopting a more sensible metric for quantifying warming (energy balance) and uses some impressive simulation techniques (e.g. the neural network trained to performing the BERN2.5D model Markov Chain Monte Carlo simulations), the uncertainty in its conclusions is likely understated.

Is mankind responsible for some amount of warming?  Certainly -- again, basic physics tells us this.  But is this warming permanent and linearly scaling with greenhouse gas emissions?  The answer there is far less certain.

It would be terrifically convenient to blindly accept the paper's conclusion and take it as dogma that mankind is responsible for 74 percent of warming.  Unfortunately taking a look at the paper's methodology from an analytical and scientific perspective reveals that doing so is likely folly.

The authors deserve praise for applying the energy balance and for providing explicit qualifications for some of the key assumptions and shortcomings of their work.  But the paper does not definitively pin what percent of warming mankind is to blame for. As The X Files states, "The truth is out there."

Hot summer day
[Image Source: Sustainable Blog]

It is important for politicians to bear this state of uncertainty in mind, when deciding on whether to adopt more costly means of "fighting" global warming by cutting greenhouse gas emissions.  Some methods like deploying nuclear energy or funding novel solar energy research represent lower cost, lower risk scenarios, in which mankind would still benefit in other ways, even if the warming threat proves overstated.  But when it comes to more extreme actions, such as adopting meat rationing or carbon credits (wealth redistribution) schemes, it seems a bit premature to resort to such expensive schemes given the lack of definitive understanding.

Source: ETH Zurich [PDF draft]



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

RE: Here is the problem
By Mint on 12/27/2011 1:57:19 PM , Rating: 3
The problem with your debunk is that you cherry picked 1998.

It's just like the people who, a couple of years ago, said the earth was cooling in the last decade. Well yeah, that 1998 El Nino episode and 2007 cool spell did wonders for your cherry picked regression line.

Fast forward to 2011, and the exact same methodology - one decade trendline - results in a warming trend. I could take a positively sloped function linear function, add some neutral perturbations, and you'd see the same regions of positive and negative slopes. Therefore 10 year slopes are useless.

The unusual thing about 20th century warming isn't the absolute temperature, but rather the rate of temperature increase. It is possible that our temperature proxies of the past simply mix data from adjacent decades and are simply incapable of resolving short, rapid rises in temperature. However, as far as we can tell, the current rate of warming is unprecedented.


RE: Here is the problem
By Tony Swash on 12/27/2011 2:34:37 PM , Rating: 1
quote:
The problem with your debunk is that you cherry picked 1998.

It's just like the people who, a couple of years ago, said the earth was cooling in the last decade. Well yeah, that 1998 El Nino episode and 2007 cool spell did wonders for your cherry picked regression line.

Fast forward to 2011, and the exact same methodology - one decade trendline - results in a warming trend. I could take a positively sloped function linear function, add some neutral perturbations, and you'd see the same regions of positive and negative slopes. Therefore 10 year slopes are useless.


None of that addresses my actual point which was that the climate models predicted higher temperatures by now and those higher temperatures never materialised. A very substantial increase in CO2 has occurred and there has not been a year hotter than 1998 since 1998. The models said it would get warmer with higher CO2 - it didn't. To me that says we need to look at the models again. You say the models are OK even though they can't seemingly predict climate change.

quote:
The unusual thing about 20th century warming isn't the absolute temperature, but rather the rate of temperature increase. It is possible that our temperature proxies of the past simply mix data from adjacent decades and are simply incapable of resolving short, rapid rises in temperature. However, as far as we can tell, the current rate of warming is unprecedented.


There is absolutely nothing unusual about the rate of warming that occurred in the last two decades of the 20th century. Nothing. Similar rates of warming have occurred within recent history at various periods in the Holocene (our current climate period that started about 10,000 years ago) and this is clearly shown by the ice core reconstructions. Much of the Holocene was warmer than today. Nothing unusual has occurred with climate in the last 100 years. The most unusual event of the last 1000 years was the little ice age between 1550 AD and 1850 AD when the planet cooled to it's lowest point since the start of the Holocene and the end of the previous ice age. Since the end of the little ice age global temperatures have recovered in several distinct warming phases the last of which occurred in the last two decades of the 20th. That warming appears to have stopped over a decade ago.

With out reference to climate models there is nothing to indicate the need for any special forcing agent such as CO2. And those models got the prediction for the last decade very badly wrong.

See

http://a-sceptical-mind.com/unprecedented-part-2

and

http://a-sceptical-mind.com/the-hockey-stick-vs-th...


RE: Here is the problem
By kraeper on 12/27/2011 3:07:52 PM , Rating: 2
Well that's weird. I actually just up-rated a Swash post. Maybe the world really is ending!


RE: Here is the problem
By Tony Swash on 12/27/2011 8:00:43 PM , Rating: 1
quote:
Well that's weird. I actually just up-rated a Swash post. Maybe the world really is ending!


Nah - it's just the Christmas spirit :)

By the way the links I posted are to a web site I set up relating to climate change. I started out agnostic on the whole thing and just thought that the science must be sound but it was such an important issue and the climate change bandwagon was so big that I thought I would try to have a look at the evidence for myself. I was shocked at what I found, how flimsy the science was, how rigorous were the efforts to suppress dissent. So I collected some of the evidence that had pushed me towards being sceptical and put it on a web site.

If the evidence changes then I will change my mind.

quote:
When the Facts Change, I Change My Mind. What Do You Do, Sir?
John Maynard Keynes


RE: Here is the problem
By Dr of crap on 12/28/2011 12:54:29 PM , Rating: 2
Well, I have to agree with you on your points.

And why is it the climates scientists don't have any data to support the doomsday predictions?
That I don't understand. Why is it so vague?
And the problems of the past few years only make matters worse.

Of course we will not change anyones mind.
It's kind of like politics. If you be a Dem, all Reps are evil! and it's all their fault!


RE: Here is the problem
By Mint on 12/28/2011 12:46:48 PM , Rating: 2
It does address your point, and you repeated the fallacy:
quote:
A very substantial increase in CO2 has occurred and there has not been a year hotter than 1998 since 1998.

If you add noise to a baseline function, simply tracking the peaks does nothing to reveal that base function. Stop focusing on 1998. We're actually tracking fairly well to Hansen's 1988 prediction (his B scenario, which he called the most likely, and whose net forcing assumptions best matched what the earth received).

As for your links, I don't know where he got his data from. He points to a "Hockey stick" in the mid-1800's. He did the same cherry picking you did in choosing ranges. Here's what different reconstructions have found:
http://en.wikipedia.org/wiki/File:2000_Year_Temper...

In any case, the real problem with AGW isn't in the science but rather in the politics. There is really no justification for taking action against it other than nuclear energy, as the benefits are too small for the cost. The biggest contributer to AGW is Greenpeace when they killed appetite for nuclear in the last 40 years, leaving fossil fuels as the only other choice.


"Intel is investing heavily (think gazillions of dollars and bazillions of engineering man hours) in resources to create an Intel host controllers spec in order to speed time to market of the USB 3.0 technology." -- Intel blogger Nick Knupffer














botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki