Climate model bias

20 views
Skip to first unread message

Stephen Salter

unread,
Sep 22, 2022, 10:41:17 AM9/22/22
to geoengi...@googlegroups.com, noac-m...@googlegroups.com

Hi All

A paper at https://link.springer.com/content/pdf/10.1007/s00376-022-2036-z.pdf

says that there are significant biases in simulated cloud physical properties over the Southern Ocean.

Section 5 mentions a mean bias of “more than 30 Watts per square metre” lots more than I thought was the problem.

However it is not clear, at least to an engineer, whether it means plus or minus 30 watts per square metre.

This does not increase my admiration for climate modellers. Please help.

Stephen

Professor of Engineering Design

School of Engineering

University of Edinburgh

Mayfield Road

Edinburgh EH9 3DW

Scotland

0131 650 5704 or 662 1180

 

The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. Is e buidheann carthannais a th’ ann an Oilthigh Dhùn Èideann, clàraichte an Alba, àireamh clàraidh SC005336.

Govindasamy Bala

unread,
Sep 22, 2022, 11:18:02 AM9/22/22
to s.sa...@ed.ac.uk, geoengi...@googlegroups.com, noac-m...@googlegroups.com
I am not at all surprised by the regional bias of this magnitude in the previous generation of models. The paper says this bias is reduced in CMIP6 models to about 10 Wm-2. These biases are related to how global models "represent" cloud properties such as cloud liquid water, liquid cloud fraction, and total cloud fraction which have biases of about 10-20%. I would never expect global models with a resolution of about 100 km to reproduce accurately these subgrid-scale variables.  GCMs were not designed to "simulate" clouds which are "represented" through parameterizations using various "assumptions". GCMs are designed to only simulate large-scale (~ 1000 km) features well. Models are only our attempt to explain the real world and no model exists today in any branch of science that can explain everything in that branch of science. 

There is nothing here to admire or find fault with modellers.  It is just that the problem is too complex with too many degrees of freedom. In fact, I am happy we have made unbelievable progress in the last 3-4 decades. It is a work in progress (like modelling in any branch of science) and I do not expect an end game anytime soon. 
Cheers,
Bala

--
You received this message because you are subscribed to the Google Groups "geoengineering" group.
To unsubscribe from this group and stop receiving emails from it, send an email to geoengineerin...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/geoengineering/DB7PR05MB56927D3815DCF39B23757FA8A74E9%40DB7PR05MB5692.eurprd05.prod.outlook.com.


--
With Best Wishes,

-------------------------------------------------------------------
G. Bala
Professor
Center for Atmospheric and Oceanic Sciences
Indian Institute of Science
Bangalore - 560 012
India

Tel: +91 80 2293 3428; +91 80 2293 2505
Fax: +91 80 2360 0865; +91 80 2293 3425
Email: gb...@iisc.ac.in; bala.gov@gmail.com
Google Scholar
-------------------------------------------------------------------

Stephen Salter

unread,
Sep 22, 2022, 12:44:25 PM9/22/22
to Govindasamy Bala, geoengi...@googlegroups.com, noac-m...@googlegroups.com

Dear Bala

That helps a bit.  But even if we do not know the starting point accurately, can we deduce the size of the change we have to make by doing some geoengineering?

I thought from Julia Slingo that the problem was 1.7 watts per square metre out of a mean solar input of 340 so we only had to do 0.5%.

Stephen

 

From: Govindasamy Bala <bala...@gmail.com>
Sent: 22 September 2022 16:26
To: Stephen Salter <S.Sa...@ed.ac.uk>
Cc: geoengi...@googlegroups.com; noac-m...@googlegroups.com
Subject: Re: [geo] Climate model bias

 

This email was sent to you by someone outside the University.

You should only click on links or attachments if you are certain that the email is genuine and the content is safe.

Peter Eisenberger

unread,
Sep 22, 2022, 1:10:34 PM9/22/22
to bala...@gmail.com, s.sa...@ed.ac.uk, geoengi...@googlegroups.com, noac-m...@googlegroups.com
I  think the problem of modelling climate change is more fundamental than it is too complex or even how large the biases are. 

Our understanding of complex systems like the earth's climate is that they involve feedbacks like the melting of the ice caps 
that expose darker surfaces that are more adsorbing that increase the warming or the release of methane . In addition complex systems are organized around chaotic
attractors and while deterministic one cannot determine how a variable, let's say sea level rise , in one chaotic attractor organization will map to another.
So from this perspective modeling the earth's climate within one chaotic attractor has to deal with the inherent nonlinear nature of the climate system 
but most important is that catastrophic climate change involves the transition from one chaotic attractor to another which while deterministic is not predictable. 
One the other hand, modeling can help identify the most important feedbacks that could drive the system to reorganize and focus our observations on them 
to determine whether they show  evidences of nonlinear growth. In fact the under prediction of the rate of climate change is in my opinion the result
of such nonlinearities which in turn means catastrophic climate change has begun. This is a very important, most important,  finding for decision makers.
By focusing our discussions instead on detail predictions and comparisons between models we are not only certainly going to fail but we weaken the most important finding 
that not only is climate change upon us but we are observing that catastrophic climate change has begun. 
The earths climate has an equivalent of inertia,the rate at a which it reorganizes, called the  Liapunov exponent. 
Our hope is that if we globally organize , mobilize, now we can act faster. Of course that in turn means we have to reorganize 
the complex human system to address it .  
As a side comment the above analysis would need to also be applied to the risks of geoengineering efforts because at a high level they all are changing many feedbacks. 
Reasons like solar variations do occur without catastrophic consequences have to be reconsidered in terms of the current climate already being reorganized.    



--
CONFIDENTIAL INFORMATION: This email message and all attachments contain confidential and privileged information that are for the sole use of the intended recipients, which if appropriate applies under the terms of the non-disclosure agreement between the parties.

Peter Eisenberger

unread,
Sep 22, 2022, 1:18:26 PM9/22/22
to bala...@gmail.com, s.sa...@ed.ac.uk, geoengi...@googlegroups.com

rob...@rtulip.net

unread,
Sep 23, 2022, 1:39:26 PM9/23/22
to Stephen Salter, Govindasamy Bala, geoengi...@googlegroups.com, noac-m...@googlegroups.com

Stephen

 

The article states “Absorbed solar radiation (ASR) is calculated as the downwelling minus upwelling shortwave radiation of all-sky conditions at the top of atmosphere. A previous study by Trenberth and Fasullo (2010) found that too large ASR over the SO was simulated by the CMIP3 models, showing a substantial bias of more than 32 W m–2. The serious overestimation of ASR and underestimation of cloudiness over the SO led to poor model performance in simulating the energy budget in the Southern Hemisphere in climate models (Marchand et al., 2014).”

 

They say the more recent CMIP6 models reduced this failure to accurately model Southern Ocean albedo to ±10 w/m2, compared to the CMIP3 ASR heating figure that was 30 watts too high due to underestimating cloudiness.

 

Robert

--
You received this message because you are subscribed to the Google Groups "NOAC Meetings" group.
To unsubscribe from this group and stop receiving emails from it, send an email to noac-meeting...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/noac-meetings/DB7PR05MB56921275E3F6E8B51335E7BBA74E9%40DB7PR05MB5692.eurprd05.prod.outlook.com.

Reply all
Reply to author
Forward
0 new messages