Direct consequences to total arctic sea ice loss?

[wanderer]

Hey!
Maybe someone posted something like that before, but I wondered if there are any studies or scenarios, how weather or climate zones would be influenced and changed, if the arctic is really ice-free in summer (or later totally ice-free all over the year), what would that mean for the rest of the world?

[Pmt111500]

I'm not on the computer right now so cannot seek the few I've seen. The deniers drivels and more serious scientific challenges to modelling have no doubt made this kind of necessary study harder to get funded.

The wa/cc system is pretty much guaranteed during autumns early winters, of course. But I undersrand you ask what this does to the rest of the world, and on this there aren't too many studies. i guess you'd need to check what the GCMs say for 2060-80 and 2080-2100... Or the latest models for 2040-2060... No pointers on that either, sorry.

[wanderer]

How careless is that?

Even if it is not happening before 2035 (that would be early enough), how are we going to prepare for this, when there are no studies or simulations?
It must have a huge effect on climate zones, at least in the northern hemisphere...
I just found this article:
http://www.scientificamerican.com/article/arctic-sea-ice-loss-implications/

[crandles]

I would suggest it is complicated and there are quite a lot of recent studies which I would suggest haven't yet found all the effects let alone agreed about those effects. See video by Dr Francis linked from
http://forum.arctic-sea-ice.net/index.php/topic,750.msg68965.html#msg68965
 first 25 minutes is fairly basic stuff that I think most people here would know about global warming.
Then it gets onto more weather impacts of arctic amplification and loss of sea ice.

To attempt short summary:
Land and arctic warms faster than rest of globe.
Same pressure is higher in tropics than by pole and this drives jet streams.
This pressure gradient is reduced by arctic warming causing a greater increase in height of a given pressure at the Arctic than in tropics or mid latitudes.
This weakens jet stream.
Can cause more wavy, slower moving and increased frequency of blocks.
Causes more persistent weather. So more intense flooding, longer droughts, longer heatwaves, longer cold spells etc.

But it can be more complicated than that for instance interactions with PDO/warm blob/cold blob/El Nino.

This is probably all happening now with just fairly small regions of summer retreat of sea ice. May get more significant with further ice retreat or maybe more persistent changes if it requires interactions to really see an effect or maybe there will be other effects that we haven't realised yet becoming more apparent.

[Pmt111500]

The stuff with QBO and ENSO are, I guess, two of the items taking much of the scientists time. Then there is the glacier modelling and the sea ice stuff, much talked here too, that are difficult. There really should be some easily tunable GCM in existence which could be used to produce answers to questions like this.

[crandles]

There really should be some easily tunable GCM in existence which could be used to produce answers to questions like this.

Hmm. Ice edges seem one thing that GCMs do poorly with vastly different levels of sea ice from one model to another. Prescribed edges may be needed for this sort of work. However, are the effects modelled then a result of the model being out of balance with where it want to be or as a result of where the ice edges are being positioned? I would suggest this makes it far from easy to use such tunable GCMs.

(Even using GCMs normally seems far from easy. From what I have heard lots of runs set up and started only to find something wasn't set up as intended. Even I have found a model with snow accumulating on sea ice as in 1000s of tons per square metre - never losing any as such process(es) hadn't been set up.)

[Pmt111500]

There really should be some easily tunable GCM in existence which could be used to produce answers to questions like this.

Hmm. Ice edges seem one thing that GCMs do poorly with vastly different levels of sea ice from one model to another. Prescribed edges may be needed for this sort of work. However, are the effects modelled then a result of the model being out of balance with where it want to be or as a result of where the ice edges are being positioned? I would suggest this makes it far from easy to use such tunable GCMs.

(Even using GCMs normally seems far from easy. From what I have heard lots of runs set up and started only to find something wasn't set up as intended. Even I have found a model with snow accumulating on sea ice as in 1000s of tons per square metre - never losing any as such process(es) hadn't been set up.)

All valid points, I'd say. But would it be possible to f.e. run a gcm to a stable state say with 400 ppm (or what ever it takes to lose the sea ice in the model) of co2 in air and use that as a starting point for analyses for such times when there's only seasonal ice left? That would lose one degree of freedom (the multi-year sea ice) and the model would possibly run a bit faster afterwards. Of course this might lead to errors with glaciers, but at least the atmospheric and ocean components might run smoother and more accuratrly so Wanderer's question could be answered. I think I've seen one such study long time ago, but can't remember about anything more of that. Still not in the computer so leaving this here. Oops, looks like a double copy/paste in the beginning. (fixed)

[crandles]

The sea ice adjust pretty rapidly - only a couple of years. If the model has far too much ice area when allowed to run normally then you might have to run CO2 at 450 or 500 instead of the correct 400 to get the ice down to the right level and even then the total ice area might be correct but it might be distributed in the wrong places. The too high CO2 level is likely to make temperatures too high so then the weather effects that you see are likely due to high temperatures instead of ice edge effects. To me, this seem worse than using correct CO2 levels and prescribed ice edges. With prescribed ice edges you probably get wrong sst near the correctly located ice edges but at least it is only a limited area that is affected and probably not all that far wrong.

Maybe I didn't follow your comment. I doubt there is much speed difference whether you use prescribed ice edges or not.

I am not an expert but what I think you might want to do is:

Run a model for three scenarios each done scores (preferably hundreds) of times. Suppose the model has too much ice. One scenario would be prescribed ice edges from 1980s with current CO2. The second would be prescribed ice edges as last few years. The third would be prescribed ice edges as last few years but add a correction to sst near the ice edges. The model will be trying to freeze over the areas that it would normally cover with ice so making the sst too close to freezing point for too long. Apply some correction to such situations.

If the weather differences between first and second scenario is much larger than the differences between second and third scenario then you may be able to get away with claiming that the problems of using prescribed ice edges is not too severe and the weather effects of second and third scenarios compared to first scenario is likely to be a reasonable approximation of the weather effects of retreating ice edge.

Leaving such research to someone who has a model that happens to do pretty well with ice edges might be an easier option/cop-out. The research is limited to that one model but then stuff like this rarely uses more than one model unless you can get some scenarios into a model inter-comparison project.

In short, easy and tunable just don't go together when describing a GCM.

[Archimid]

What will happen to the arctic ocean circulation  once it has no ice? Does having a layer of ice constrains the circulation of the Arctic? I'm afraid that once the ice is gone something like a vortex could form centered on the north pole due to things like the coreolis effect.