Climate models and climate past

[waitforufo]

Several of the topics in this forum have to do with climate modeling, particularly as it pertains to global warming. In these forum topics I and others have suggested that these models be verified by demonstrating an ability to closely track well know past climate events. So for example:

 

Can current climate models track and explain the little ice age from beginning to end?

 

Can current climate models track and explain the medieval warm period from beginning to end?

 

Can current climate models track and explain the Holocene climatic optimum from beginning to end?

 

Are such questions reasonable? Why or why not?

 

It seems to me that part of verifying any model would be to test the model with one or more climate forcings removed or held constant.

 

Proponents of current climate models often talk about data going back past several ice ages. They also seem to be able to pinpoint with some accuracy that date at which they believe anthropogenic greenhouse gas concentrations began to have an effect. Since this data seems to be available and often presented by proponents of current climate models, why not verify these models over the periods of time I have suggested above? Would mankind not benefit from the knowledge gained from such verification?

[iNow]

Since this data seems to be available and often presented by proponents of current climate models, why not verify these models over the periods of time I have suggested above?

Hasn't this already been done?

[Reaper]

It has, and it was demonstrated many times, including right at the beginning of the debate thread I set up.

 

It's best if you keep this debate and similar topics over here on my thread anyway: http://www.scienceforums.net/forum/showthread.php?p=381047#post381047

[waitforufo]

It has, and it was demonstrated many times, including right at the beginning of the debate thread I set up.

 

It's best if you keep this debate and similar topics over here on my thread anyway: http://www.scienceforums.net/forum/showthread.php?p=381047#post381047

 

Dear Lockheed,

I went to the beginning of your thread and did not find answers to the questions I asked. Perhaps you can provide quoted text from your input or from the input of others throughout the body of your thread. Even better, perhaps you can provide a peer reviewed paper the covers the questions I asked in post #1 of this thread.

 

Are my questions too ambiguous? Do I need to clarify them further?

[iNow]

Before everyone runs around finding references to address this, references which will likely be ignored anyway, can you please explain the relevance of these three particular epochs?

 

 

 

the little ice age from beginning to end?

...

the medieval warm period from beginning to end?

...

the Holocene climatic optimum from beginning to end?

 

 

I'm learning from experience, and don't care to waste any more of my time with nonsense. Thanks for any clarification you're willing to offer.

[waitforufo]

Dear iNow,

The topic of my thread is “climate models and climate past.” To determine the accuracy of any model one must verify the model against known data. In dynamic systems, this modeling verification would generally be done over time events that show significant deviation from normal or that have some otherwise unexplainable characteristic (e.g. not related to earth’s orbital variation in the solar system). I selected the little ice age, medieval warm period, and Holocene climatic optimum because each of these seem to fit such a definition. I also selected these three epochs because they included known warming and cooling climate events. Two warming events because of the current interest in warming trends. Better understanding these epochs would further our understanding of climate variation. Good science in and of itself. More important, verifying current models over these epochs would prove our understanding of climate variation without an anthropogenic increase of greenhouse gases. Modelers would be able to show that there models predict past events well and that our current situation can be explained best by anthropogenic forcings.

 

Such verification would go a long way toward dispelling climate model skeptics.

 

 

P.S. Happy New Year

[JohnB]

waitforufo, personally I think that your questions should be answered but there are other things to be done first.

 

As we cannot directly measure the past climate we are forced to rely on proxies. These proxies are calibrated by comparing them to the period of time when we do have instrumental records. However to do this properly we must be sure that the instrumental records are accurate. In the case of the USHCN a volunteer audit is currently occurring and the last figures I saw was that with 27% of the US network audited 85% of those did not meet the USHCN criteria for siting which effects accuracy (by the USHCN rules) by up to 5 degrees.

 

It does make one wonder about the siting of the other stations in the GHCN network. The situation has of course improved dramatically in the last 30 years with the increased use of satellites. From that we can deduce some sort of Global Mean Temperature. (And we need to be sure of our methodologies) For some reason the HADCRUT-3 figure seems to be consisently below the one from GISSTemp. I have no idea why but I'll keep reading.

 

Once we have an accurate record of temperature for the modern period we can then be confident that we can use this record to calibrate our proxies. We also need to choose our proxies with care. Tree ring proxies are very high resolution but tree growth is not just dependent on temperature, precipitation, methane and CO2 fertilization can all also play a part. Other proxies that are influenced by lesser factors are available but they are generally of lower resolution, decadal or centennial. Others vary, for example Ice Cores are high resolution for the modern period but move to decadal and centennial as you go further back in time due to the compacting of the ice. In short, all proxies have their plusses and minusses, it's an imperfect world.

 

To give an example from Briffa et al 2001;

(For those interested, this is not a comment on Briffa et al, I was looking for a spaghetti graph to link to and this was the first one I found that wasn't in a pdf.)

 

You will notice that while the reconstructions agree quite well during the instrumental period there is an increasing divergence as you go back past 1700 or so. Which of the 8 reconstructions are you going to use to calibrate your model?

 

Personally, I think that these areas deserve a lot more research than they are currently getting. A catastrophic warming event would be the greatest danger that the human race has faced in recorded history. Anything that helps us understand palaeoclimatology is vital to the understanding of the present climate systems. This means a synthesis of many disciplines, Dendrochronology, History, Archaeology, Physics, Ice Cores, Oceanography and others to combine into a complete (and hopefully reasonably accurate)picture of the past.

 

Even the tree ring cores that the climatologists have to work with seem woefully out of date with many ending in 1985 and few since then. I mean really, what would it cost to put two blokes in a 4 wheel drive and send them around the countryside for a year collecting cores? I'm sure both our govts waste more than that in a week just telling us about their latest you-beaut program.

Hasn't this already been done?

To be frank, I think so but I'm not sure. I have read in some papers comments along the lines of "We found our model to be accurate in modelling the Younger Dryas....." but the actual methodologies and results are not mentioned. I'm not sure what to think.

 

Happy New Year.

[waitforufo]

Dear JohnB,

Thank you for your well considered and polite reply. My personal opinion is that the concept of global climate, let alone global climate modeling is a young science requiring significantly more research before conclusions can be drawn.

 

The plot you present for example is of temperature smoothed with a 25 year filter (least-squares?). Temperature is of course a weather variable. I have often heard within several Science Forums threads that a 20 to 30 year average of weather defines climate, but I have seen no scientific justification for this averaging number. Assuming this number is years is correct, many should take issue with presenting data on a year to year basis. When bandwidth is reduced, it is generally best to also decimate the data as well. A more reasonable number would be to present climate data on at most a 5 year basis but more likely a 10 year basis depending on filter parameters. Since the last time global cooling was considered likely, in the 70's, we have 3 to 6 data points showing a warming trend. Something we should keep and eye on, but hardly enough data, by itself, to justify modifying human behavior.

 

With time, I am sure that climate scientist will get better at calibrating proxies and validating current climate models to explain past climate events. When this is accomplished I'm sure that current climate modeling skeptics like myself will be won over.

[JohnB]

Mate, I'm a climate model skeptic myself.

 

TBH, I think we're concentrating a bit much on the physics and not listening enough to what history has to tell us. Historical records can't tell us what the absolute temp was in prior times but it can tell us about major climate change in the past.

 

For example, we don't know the climate in North Africa 5,500 years ago but we do know that it suddenly suffered a major climate change. Saharan Oases that harboured cities of 10,000 people (and that's a lot of water) suddenly dried up leading to a population migration. If you're interested, many moved to the Nile Valley and became the Dynastic Egyptians.

 

We know this to be true because the pottery found in the earliest Egyptian towns matches the pottery found at the Oases and is found in much larger than "trade" quantities. By C¹⁴ datings we know when this occurred.

 

These Archaeological records can help us to refine our proxies in that if the proxy tells us that North Africa was reasonably wet 5,500 years ago, then the proxy is wrong (miscallibrated).

 

Similarly, if the reconstruction says that Greenland was somewhat colder during the MWP than today, then we can adjust the proxy to suit the observed facts. i.e. That there were working farms during the MWP in places that until a couple of years ago were unworkable permafrost.

 

"No man is an Island" goes the old saying and neither is a science unconnected to the other sciences. If a theoretical reconstruction is at odds with an observed Archaeological fact, then the model is inaccurate, not reality.