Bad in simulations but even worse in real life

[Iceberg]

I am trying to tie a few loose ends together here so please forgive this thread.

There is another thread to follow the recording breaking ice loss over on the main summer forum, but I am more interesting in the why and the possibilities here.

I've picked a random paper that discusses the predicted effects on arctic ice taken from GCM ensemble work.

http://www.cgd.ucar.edu/oce/mholland/abrup...olland_etal.pdf

Basically all the models show up the concept of tipping points for arctic sea ice, but there is little agreement about when it might occur. Although again the ensemble average gives us a clue.

I've also included the current state of arctic ice. My best guess is that it could be as low as 3.7m this year, down approx 10% from the previous min.

Now this doesn't have to be an AGW, pro or anti thread, because I am more than willing to hear views from the likes of Polyokov et al where ice loss and tipping points are still factored in, and is caused instead of AGW by natural cyclical ocean variations (as long as there is some evidence).

But what are peoples thoughts on the whole subject, are we currently hitting a tipping point ?, why are the models underestimating ice loss currently ?. when will the ice start to reverse it's negative slide ?. We've seen a rapid reduction in ice thickness as well as extent with the thickness falling from 3.5m to less than 2m in March.

Matt

[Gray-Wolf]

There used to be an acronym for computer Model's R.I.R.O (rubbish in , rubbish out) and so I feel you need look at the models parameters to see why the shortfall occurs.

You could plainly see how fast the 'single season' ice ablated over July and in there lies and issue. We know the impact the mechanical action of wave and swell are on ice edge ablation but what of the areas around the Polynya's from last season? (facing the bearing straights). Obviously the ice thickness varies and as such stresses are not equally distributed leading to fracturing within the ice sheet.

Do the models allow for polar ice drifting? When the ice cap is no longer 'grounded' on continental shelf then it will run with winds and currents effectively 'ramming' multi year ice into warm water shallows allowing for increased ablation.

As I witness things the end of polar summer ice will be a very rapid affair as multi year ice gives way to single year ice and 'warm waters' project further under the sheets ablating from below leading to greater Polynya's formations and further fragmentation. In turn the 'free floating' ice will drive onto warm shelf regions and so 'central areas' of multi year ice will decline.

Not long now until we have moist air masses from the north watering/flushing out the far north of Eurasia further compromising the permafrost regions and releasing it's methane/CO2 .

[eddie]

In my opinion people put far to much faith in the accuracy of current climate models. You just can not accurately model something you do not fully understand.

In current climate models if a process isn't fully understood, and therefore can't be modelled based on an understanding of the actual physical processes involved, or is too complex to be modelled with current computing power then it is added to the model as a parameter. That parameter is essentially a number (or maybe a time series of numbers or whatever) that is made up to represent that process and fed into the model.

That number is based on the best science available at the time. If the level of understanding of that science isn't great then neither is the accuracy of the number. Add in a load of these parameters and the scope for error is large.

Since we have never seen (and I mean with accurate satellite measurements) the arctic melt before, our level understanding of how that happens and any feedback processes that might occur as a result is low. The same can probably be said for lots of other processes/cycles that affect climate such as ocean currents, El nino, solar variance, cosmic flux. These processes in climate model are really just a best guess.

I really think the most accurate thing we can expect from climate models at the moment is 'The earth is going to get warmer if we release more CO2 and here is the range that the temperature could fall within'. When they start saying there is going to be X less rainfall in this region, Y more in this region, the antarctic ice will melt by Z it makes me laugh. Maybe in 10 or 20 years when the science and computing power has advanced considerably.

[Devonian]

I really think the most accurate thing we can expect from climate models at the moment is 'The earth is going to get warmer if we release more CO2 and here is the range that the temperature could fall within'.

I agree (though I'd use 'are likely to' rather than 'could'), I'd never say anything else.

When they start saying there is going to be X less rainfall in this region, Y more in this region, the antarctic ice will melt by Z it makes me laugh. Maybe in 10 or 20 years when the science and computing power has advanced considerably.

Not sure any climate modeller is saying such? 'Indications' maybe, 'likelyhoods' likewise, 'trends' likewise, but stating?

[Iceberg]

I think it's more an issue of resolution and not including all the feedback mechanisms, as you say GW the existance of Polynya etc probably isn't included nor are the effects of under ice current movements.

To save me getting to pessimistic anybody got any feedbacks that would help to keep the ice there ?

I am in general agreement Eddie the models are not good for general micro climate variations, I think they do well enough for general global patterns though.

Anybody think a tipping point has already been reached ?.

[eddie]

I can't think of any feedbacks that would stop all the sea ice from melting. Increased winter precipitation could perhaps slow down the melting of the polar ice cap itself.

It will be interesting to see what effect all that extra open water will have on the weather patterns seen at such latitudes.

Edited July 17, 2007 by eddie

[BLAST FROM THE PAST]

Anybody think a tipping point has already been reached ?.

Yes, and trouble lies ahead. Airtemps have not caused the melt, it has been warmer up there with good ice retention.

BFTP

[Gray-Wolf]

Yes, and trouble lies ahead. Airtemps have not caused the melt, it has been warmer up there with good ice retention.

BFTP

I feel high air temps were always just the 'icing on the cake'. As the NASA results of snow ablation on the mountains fringing the Ross embayment show the air temp is not the driving force and either solar radiation on 'dirty (particulate contaminated) snow' or warm (relative) waters under ice shelves/caps are the real drivers in the ablation process's.

Water takes a long while to heat up but also a long while to cool down (esp. when heated to the depths the Argo probes show) and so even if the warming stopped today we'd be losing ice for years to come (an' there ain't much more multi year ice to go!).

Of course once we get to the stage where fragmentation occurs then you can probably up the ablation rate by a factor of 5 (for all who have cleaned out a freezer LOL).

Edited July 17, 2007 by Gray-Wolf

[jethro]

I feel high air temps were always just the 'icing on the cake'. As the NASA results of snow ablation on the mountains fringing the Ross embayment show the air temp is not the driving force and either solar radiation on 'dirty (particulate contaminated) snow' or warm (relative) waters under ice shelves/caps are the real drivers in the ablation process's.

Water takes a long while to heat up but also a long while to cool down (esp. when heated to the depths the Argo probes show) and so even if the warming stopped today we'd be losing ice for years to come (an' there ain't much more multi year ice to go!).

Of course once we get to the stage where fragmentation occurs then you can probably up the ablation rate by a factor of 5 (for all who have cleaned out a freezer LOL).

If this is the case, then surely Polyakov's theory of natural drivers being the main cause of the problem, rather than Co2, actually makes more sense?

[Iceberg]

I am not sure it makes more sense, what is does do is introduce a lag time from AGW related ocean heat up.

Interesting the global dimming theory comes into play as well here, reduce particulate matter will remove cloud formation and lead to more radiative melting, hitting the Ice with a triple wammy of.

1 Increase radiative melting

2 Increase atmospheric warming leading to more metling.

3 Ocean heating and moving the point of NADW further north, leading to a change in ocean current patterns more melting.

My personal feeling is that the tipping point of ice melt in both poles will be the single biggest factor in AGW impact realisation.

[Iceberg]

A quick update since we are about to smash through the record minimum and the NW deep passage is due to open for the first time that we know off.

First presentation proves that the US Navy takes an ice free arctic seriously.

http://www.natice.noaa.gov/icefree/Gameboo...20picsFinal.pdf

The second is more interesting and details Wieslaw Maslowski's paper on arctic ice melt and how is might be current led caused by AGW as part of a Feedback mechanism.. If he's right then the date for an ice free arctic will be brought forward to 2020, probably meaning that the tipping point has been and gone.

http://trs-new.jpl.nasa.gov/dspace/bitstre...3/1/05-2376.pdf

Edited August 10, 2007 by Iceberg

[parmenides3]

Cryosphere Today has today confirmed that the Summer minimum sea ice area is now the lowest ever recorded: http://arctic.atmos.uiuc.edu/cryosphere/

As they point out, there is still up to a month of the melt season to go. Though the graph curve tends to level out in a curve for a few weeks prior to the last date of melt, there is no indication of this happening yet.

Thought you'd like to know. I'll post this on Carinth's page, too.

:)P

[Stratos Ferric]

...

Since we have never seen (and I mean with accurate satellite measurements) the arctic melt before, our level understanding of how that happens and any feedback processes that might occur as a result is low. The same can probably be said for lots of other processes/cycles that affect climate such as ocean currents, El nino, solar variance, cosmic flux. These processes in climate model are really just a best guess.

I really think the most accurate thing we can expect from climate models at the moment is 'The earth is going to get warmer if we release more CO2 and here is the range that the temperature could fall within'. When they start saying there is going to be X less rainfall in this region, Y more in this region, the antarctic ice will melt by Z it makes me laugh. Maybe in 10 or 20 years when the science and computing power has advanced considerably.

I think you underestimate slightly how far thinking has come, and also the broad understanding of climate (as opposed toi daily variation in the weather). It is perfectly easy to produce broad systems for the global climate at the macro level, and for some climatic regimes to be very clear about the long period impacts (e.g. drier / warmer). I do agree that projecting daily or yearly variation is impossible, but that will always be true.

Re your first point, nobody had ever seen an atom bomb explode before Hiroshima but that didn't mean people didn't know what would happen. It is perfectly possible to join pieces of understanding together to create a bigger picture, as yet unseen. For sure there will be times when the detail in practice varies, but I wouldn't discount the ability to have a good general idea of consequences. As I've mentioned previously, one test of all models is to retrogress them. I.e. feed in data from twenty years ago and let them run to the present day to see whether they seem to replicate the current system. It doesn't follow AUTOMATICALLY that this ensures robust future projection, but it helps a lot. In this regard much environmental modelling is getting better and better, not least because an increasing history provides ever more basis for fine tuning, all of this in additioj to year on your doubling of processing capacity and increasing understanding.

...

The second is more interesting and details Wieslaw Maslowski's paper on arctic ice melt and how is might be current led caused by AGW as part of a Feedback mechanism.. If he's right then the date for an ice free arctic will be brought forward to 2020, probably meaning that the tipping point has been and gone.

http://trs-new.jpl.nasa.gov/dspace/bitstre...3/1/05-2376.pdf

I don't think the tipping point has been and gone, but the point of no return has probably been passed.

[BrickFielder]

The cryosphere data has some problems with historical data (ice cover data too high during 1970's and 1980's) which we have discussed before. Other sources do confirm that the ice has melted quickly and is at a very low point this year however. The factors which tend to affect artic ice conditions are winds direction across the ice cap , the state of the artic oscillation and the position and size of the Hadley cell. These things are only coarsely modelled in the general circulation models used for climate modelling and to some extent can be influenced by the very top layers of the atmosphere which again are not or poorly modelled in climate models.

Once ice is largely melted during the summer then the reflective effects of the snow cap will be removed essentially changing the earths albedo (amount of outgoing longwave radiation) so the artic becomes warmer. I don't think we are at the tipping point yet and the tell tale sign I think will be a rapid reduction in northern hemisphere snow cover during winter months. We should remember that increased precipitation across the artic in recent years is actually causing the ice to be deeper in the greenland cap at high altitudes(where temperatures never rise above freezing). The net result of any polar ice cap melt is likely to be a little un predictable with large scale changes in circulations patterns ,such that some dry areas will become very wet and warm areas very cold and vice versa. Again something not modelled particularly well in climate models because of their granularity.

[parmenides3]

There is a point in 2002-3 when monthly mean anomalies (by standard deviation) went from a steady decline with some variability to an almost linear, steep decline. See: http://nsidc.org/sotc/sea_ice.html for a graphic. A similar pattern can be read from the CT hemispheric anomaly graph. I'm suggesting that this was the point at which the decline of Summer sea ice became unstoppable. Any thoughts?

:)P

[Stratos Ferric]

There is a point in 2002-3 when monthly mean anomalies (by standard deviation) went from a steady decline with some variability to an almost linear, steep decline. See: http://nsidc.org/sotc/sea_ice.html for a graphic. A similar pattern can be read from the CT hemispheric anomaly graph. I'm suggesting that this was the point at which the decline of Summer sea ice became unstoppable. Any thoughts?

P

I think you're right that there was an inflexion in the curve around 2001-2003. At present it's probably only just taken the mean trend line outside the previous corridor, and the current period is an important indicator of whether we're about to pass a tipping point. Historically, there'd have been some bounce, albeit still with a steady downward movement overall. The indications are that we may be about to cross an important line.

[parmenides3]

That's my point, SF; where's the 'bounce' gone? Okay, three-four years is too short to make a trend, but there does seem to be a visible difference in the patterns since '02-'03.

:)P

[Iceberg]

Completely agree with alot of things, we need to see three very important figures.

The actual minimum this summer ?

When the minimum is ?

The maximum reached during the winter ?.

IMO with the three figures we could draw a mini trend.

The two largest effects on ice melt are solar melting and current, with temps coming a little further down.

[trevw]

Cryosphere Today has today confirmed that the Summer minimum sea ice area is now the lowest ever recorded: http://arctic.atmos.uiuc.edu/cryosphere/

saw this in the other thread but can't find the text on the site.. has the confimation been withdrawn from the front page, is it buried in a sub menu I haven't found or am I just blind as a bat? :o

thanks

Trev

[parmenides3]

saw this in the other thread but can't find the text on the site.. has the confimation been withdrawn from the front page, is it buried in a sub menu I haven't found or am I just blind as a bat? :o

thanks

Trev

Try refreshing the page, trevw; it should then appear as if by magic.

:)P

Iceberg; we couldn't really draw a mini-trend from just the figures you suggest. Look at the seasonal averages charts on CT and see if you can spot any trends there.

[Iceberg]

Sorry I ment in conjunction with those trends, last year had a low winter extent, but other than that winters have been holding up fairly well.

[trevw]

Try refreshing the page, trevw; it should then appear as if by magic.

:)P

ahh, had to clear my cache - have now seen it! :-)

[trevw]

realclimate have just added a new item about it too - doesn't expand on cryosphere much though

http://www.realclimate.org/

[eddie]

It is perfectly easy to produce broad systems for the global climate at the macro level, and for some climatic regimes to be very clear about the long period impacts (e.g. drier / warmer).

I don't think it is that easy. Yes you can say that a particular area of the world will be warmer but once you try to predict what average rainfall over a particular area will be you are starting blur the line between climate prediction and weather prediction.

We simply don't know what effect climate change will be on ocean currents (we don't even fully understand why they change naturally nevermind when you factor in climate change) and, as the existance of this topic shows, what the effect will be on the melting of polar ice. Average rainfall depends *very* much on these two factors which is why I think that GCMs are really just guessing at the moment.

Re your first point, nobody had ever seen an atom bomb explode before Hiroshima but that didn't mean people didn't know what would happen.

You do know they exploded an atomic bomb in the desert (Trinity) before they dropped the one on Hiroshima? That fact aside, I don't think your analogy is valid anyway since they only knew it would go bang and not exactly how big that bang would be which is what we need to know here.

As I've mentioned previously, one test of all models is to retrogress them. I.e. feed in data from twenty years ago and let them run to the present day to see whether they seem to replicate the current system. It doesn't follow AUTOMATICALLY that this ensures robust future projection, but it helps a lot.

An average of many runs from a selection of climate models gives an fairly good fit to the observed historic temperature curve. This is a long way from having predicted what the the average rainfall was. The historic recreations also have the benefit of the ocean currents being in a known state.

[parmenides3]

Sorry, I must have missed something; what are you two disagreeing about?

:)P