Arctic sea ice set to hit record low

Hoping that wxman can shed some light on these temporal extremes. Seems to me that daily/weekly/synoptic model divergence has a lot to do with too-coarse spatial scales. Also that you can't set the initial conditions correctly at fine spatial scale, because they are not known. I am thinking of heat and mass flow, but there may be better ways to describe initial conditions. Zeng does suggest they have done much better at that scale, and (a) I don't know how (b) not necessarily relevant to climate modeling. Anyway, that's the quick stuff, the weather models.

On the other extreme, climate models. Zeng says predictions over 10 to 20 years are poor, and that seems right to me because the ocean moves vast heat around at those time scales in ways that the models cannot yet capture. Step out to 30 years plus, and ENSO, PDO, AO (etc.?) get averaged out. Thus, their model hindcasted well at the longer temporal scales.

If all ocean sloshing (my inelegant term) happens at 20 year scales or less, I can see it all working. But I am not confident that longer oceanic oscillations are known to be absent. I guess there is evidence for a 60-year cycle in what we might call the 'minority' climate literature? I'd like to know more about that because it could put a hard ceiling on climate-model performance at that time scale.

 

Yes, I agree.

There is a new NWP model (NAM-WRF) that has a spatial (horizontal grid) resolution of 4 km over the CONUS (as opposed to the ~60 km horizontal grid resolution of the climate models per Zeng). That's the finest resolution for an operational model to date (some research models are down to 1 km as I understand it). It is also non-hydrostatic and has vertical resolution of 60 layers. The global medium-range models, GFS and ECMWF, have horizontal resolutions of 27 km and 16 km, respectively, and vertical resolutions of 64 layers and 91 layers, respectively. It's my understanding that climate models have vertical resolutions of something like 23 layers. Thus some important processes occurring between those vertical layers may be missed on top of sub-grid-scale horizontal processes.

Sub-grid-scale horizontal processes are typically "parameterized", which includes small-scale meteorological events (e.g., diurnal air mass thunderstorms). Parameterization is based on what "typically" occurs under a given set of synoptic and/or mesoscale conditions. However, it is actually a poor representation of actual atmospheric processes.

As far as initial conditions are concerned, I've seen operational NWP models initialize incorrectly! If you or anyone is interested, take a look at Animation using Javascript Animation Player, which is a "spaghetti" plot (ensemble) of the same model (GFS) run with slightly different initial conditions and/or different physics packages. Generally, by day 15 (360 hours), the ensemble member spread is so great as to preclude any help whatsoever, which is why I questioned Zeng's assertion that NWP models have skill up to two weeks.

It's my understanding from discussions with climate scientists from ORNL that climate model errors are assumed to cancel out over a very long period of time. However, I agree with you, are these cycles really well understood? I'm also skeptical that ingrained errors in the models (e.g., parameterization, poor terrain representation, relatively poor horizontal and vertical resolutions) will cancel out as opposed to propagate as they do over relatively shorter time scales (as in the "spaghetti plots" referenced earlier).

 

This struck home because of what happens when tornado conditions occur in the North Alabama area:

When tornadoes are around, the local TV stations that still have a news department turn over broadcasting to their weather departments. Each has their own radar systems and analysis with very fine detail on where specific radar echos are found, directions, velocities, lightning strikes, and precipitation. Some even have private, automated weather reporting units in the broadcast area at schools or other locations to report wind direction, temperature, and other automated recording and reporting data. Each TV station works hard to be the 'definitive source' down to citing neighborhoods. But as soon as the front moves on, they revert to standard programming and commercials. Perhaps this might be the way to bootstrap more accurate models?

I'm wondering if small regional weather models, ones based upon the dimensions of a typical broadcast range, might be the way to build distributed, the fine resolution models needed for more accurate, predictive models. To work on a larger scale (and handle the 15-30 knot change the wind blows in) to take data from adjacent models at boundaries. Thus the conditions in central and Northern Mississippi can feed the Huntsville model which in turn feeds the Eastern Tennessee and Western Georgia model(s).

The BEST project shows how local temperature records over 150 years soon enough became nearly a world-wide collection of stations. So perhaps the best approach is to encourage development of regional or local weather and climate models with significantly high resolution but NOT turned off when a local weather crisis passes. These micro-models with boundary interfaces could then feed adjacent models as they are developed.

Bob Wilson

 

Now matter how bad the projections from models might be, that should give one no confidence that the actual results will always deviate from the model in ways we approve of.

 

"A powerful storm wreaked havoc on the Arctic sea ice cover in August 2012."
NASA - Multimedia - Video Gallery

 

Yes, and just how was the ice extent trending just before the storm hit??

 

Arctic sea ice went low, and this (multi) decadal change was underpredicted by models. As already stated here previously.
Also interesting is that N. hemispheric snow cover has diminshed faster over the same time period
Spring snow cover extent reductions in the 2008-2012 period exceeding climate model projections
and this was also underpredicted by models.
So, if one were to assert that such models suck, I would be hard-pressed to disagree, except that we haven't better tools available.
Models aside, this is just another 'things are changing' things to add to your list, if you are keeping one. The two are certainly related, at least to the extent that snow (as ice) has a higher albedo than anything else that could likely happen in the respective areas.
The darn reduction in albedo may trip us up, even if high-latitude, high carbon soils don't decide to exhale a bunch of stored carbon. Positive feedbacks can be such a pain.
So, three big cheers for Antarctica, for staying so icy and white. Keep doin' what your doin', baby.

 

"Antarctica gained 140,000 km² of ice overnight, to reach the sixth highest daily area ever recorded. Another day like today, and tomorrow will break the all-time record for most sea ice ever measured at either pole."

Antarctica Gains 2,400 Manhattans Of Ice Overnight | Real Science

 

Mojo, it is expected that water in the arctic will freeze this time of year. Since we had more open water than recorded before, it would make sense that we have more of it available to freeze.

 

ANTarcticA is the south .Arctic is north.
IMO its normal and natural for one pole to cool as the other warms.
At least that has often been the case for at least the past 800,000 years ,judging from both polar ice cores.
Thus Arctic warming could be ,and is perfectly normal and natural.
Now if both poles were to warm simultaneously , perhaps you can blame CO2.
But one pole warming is natural variation and only a bullshit artist ( or climate scientist)can claim otherwise.
For example how does CO2 create a greenhouse effect, only at the pole of one hemisphere?
Is there a greater concentration of CO2 at the top of the Earth ?
Is there more water vapor in the north due to CO2 ?
I doubt that very much.

 

The ice shelf breakups of Antarctica were of accumulated, layers that came from land and snow fall. The recent "freezing" is not going to have the same thickness and quantity as the long, gone, shelf ice.

What becomes more interesting is the degree that Antarctica ice melts in the Feb-April time range, the Southern hemisphere "summer." My expectation would be the thinner, 'new' ice would be more likely to melt and expose more ocean surface. But we won't know until the minimum, Antarctica ice maps come out in the next six months.

Bob Wilson

 

Looking, as usual, for hope among the noise. If temperature increases in both hemispheres is all that is required to convince skeptics of CO2 's role in global energy balance, the search is simple. Here's one
ei_reconsa
(about halfway down the page)
and there are many others. I believe that the BEST data can also be manipulated to compare N and S hemispheres.

If one actually needs simultaneous warming of both poles to accept CO2 's role in global energy balance, gosh, I don't know. You may have priced yourself out of the reality market.

Nevertheless, Antarctic temperature trends can be seen
New views of Antarctica - UCAR Quarterly
and perhaps few will be surprised that it varies across the continent. Looks to me that the peninsula is warming (it sticks out) and the big central (whacking cold and icy part) is not. If one actually requires the big central to warm, to accept CO2 's role in global energy balance, I hope that you never do accept it. Because (as I said before) I want that big cold chunk to stay really cold. Keep its water frozen. Keep its albedo really high. Looks like we may need it...

The natural variation of Arctic warming (over 1400 years as I recall), was studied in a link I posted earlier, and over that time scale, ice area was larger than current. Over that time scale, not seen before. I also linked a study that found it over longer time scales. Generally, I think it would be better to make 'natural variation' arguments based on such studies, as opposed to 'both poles simultaneously' or other personal hunches. There could be quite some space between 'could be' and 'is', so I'm not fond of conflating them as Mojo has done @#71

Arctic and antarctic sea ice compared (seems like a good introduction)
All About Sea Ice, Arctic vs. Antarctic :: National Snow and Ice Data Center

Now off topic, to link to a study:
Investment decision making under deep uncertainty -- application to climate change, Vol. 1 of 1
of infrastructure investments in the face of climate-change uncertainty. The economics-speak is hard for me to understand, but perhaps some here with more experience in that area could take a look and offer opinions on whether that group is saying something particularly sensible. Or not.

 

One needs a better empirical understnding of how ice works, and how it breaks up before one can draw conclusions as to the cause and effect of "major storm" being the cause of record ice free conditions in the Arctic this summer. As someone who is admittedly no expert, I live in and make my living to some extent in winter fresh water ice. I depend on ice roads for vital transport from December through April. Spring "thaw" is referred to as " break up" for a very good reason. Large areas of ice in generally don't just melt a way. They break up, quite literally. In our case, breakup comes through a series of events. The first is a warming spell, where the bulk of the snow melts off the blue ice (the frozen lake ice that is very solid) as opposed to the slush ice, that forms as the weight of snow pushed down on an ice pan, forcing water int the snow forming,, slush ice. Depending on the snow cover in our case, the blue ice can be 3 feet thick by mid march. If there has been lots of snow, the blue ice tends to be much less, perhaps 18" due to the insulating quality of the snow. In those cases we may have 18" of slush ice, or even simply snow covered slush on top of the blue ice.

In the spring, the snow and slush and slush ice melt fairly rapidly as the ambient warms, often leaving vast quantities of water on top of the blue ice. I have actually run a boat on the water on top of the ice on occasion. This water tends to refreeze at night, so the draining process tends to be gradual, but if there is a rapid, prolonged warm spell, it can happen in a matter of days.the liquid on top of the ice, begins to drain through "air holes", while at the same time the ice begins to melt a way from south facing shores, leaving small leads of open water. The blue ice pan evenly frees itself from the shore, and begins to float, allowing the melt water to drain quickly. As long as temps remain a fair bit below freezing every night, this whole process takes about a month (April for us).

Now that the blue ice is exposed to the sun, it rapidly begins to turn black,(it actually looks like asphalt paving) further reducing its albedo. It absorbs vast amounts of heat, (you can see heat shimmring off the ice, just as you can on a highway)and the small melt water begins to drill small vertical holes in the ice, releasing trapped air. The ice then becomes what we call candle ice. It is as if you took an 18" bundle of candles and glued them together. The structure is all vertical. This is has very little strength and is very dangerous. While it can still be very thick, 12-18" but it has no strength. When it fails, it fails without warning. Unlike you image of people falling through the ice, with the ice cracking, in this case, it just fails with a load (a walking person, or a vehicle, and with out warning, you are in the water!

Now that the ice is essentials rotten, it needs one more factor to break up,, wind. (that said, rotten ice can stay in stasis for weeks, if it cools enough at night to refreeze a bit, or if it gets fresh snow to in case the Albedo again. On spring, our ice out was over a month late, because nearly every day we had a spit of snow, just enough to cover the ice and slow the process down). Once the ice pan is floating and there are leads against the shore, a fair bit of wind can start the whole pan moving. The more open water, the easier it is for the wind to get behind it and move the pan. The pan then encounters obstacles, like the opposite shore, or islands or shoals, and the force of the wind, coupled with the mass of the moving ice breaks it up. It can be, and is, very destructive to human structures in its path. We have a reach of lake, 10 miles long, 2 miles wide. (we are in an island in the middle!) When that 10 mile pan begins to move at 1/2 mph there is nothing you can do to stop it! The stronger the wind, the fast the break up. Typically our ice breaks up in a matter of about 24 hours,,,but I have seen it take several days, and also seen it sweep the lake clean in under an hour.

The whole process from earliest melt to open water usually takes about two months, from mid March to mid May. Last March, we had -33 on March 15, on March 17 we had +70, and did not see freezing again, even at night for three weeks. 4 feet of snow disappeared in 2 days, (putting 24" of water in top of the ice!) and we had complete open water April 2, a full month earlier than our previous reord early open water!

In the case of the arctic ocean, I would certainly posit, that open water begets more open water. I would also posit that the "large storms" that happened this summer resulted in more ice being swept away because there was more open water. The bottom line is, it doesn't matter how big a storm you get as long as the water is essentially sealed, the ice won't break up. Give ice a chance to move because of open water, and all bets are off.

I have read nearly every Polar exploration book from pre Franklin to Amundsun and more, and the terror of being on ship frozen in the ice come from whether the ice starts to move. My ice world is just smaller scale.

Icarus

PS. Let me also ad that I have absolutely no experience with sea ice, except for watching the ice bergs coming down Davis Strait, grounding on the northern Newfoundland coast.

I