Support the Arctic Sea Ice Forum and Blog

Author Topic: Recovery mechanisms of Arctic summer sea ice Tietsche et al 2011 Discussion  (Read 6014 times)

crandles

  • Nilas ice
  • Posts: 2391
    • View Profile
  • Liked: 52
  • Likes Given: 42
The thread is also to discuss other similar papers.

Let's start with the beginning.

https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2007GL030253
D. Schröder  W. M. Connolley
Impact of instantaneous sea ice removal in a coupled general circulation model


There are either two things seriously wrong with this model or I have a serious misunderstanding of how the arctic works.
1.

The melting season over the Arctic starts somewhere in April.  By May the periphery is melting fast and the inside forming meltponds. By June, CAB melting begins. This melting has one principal culprit, the sun. Of course most of that sun energy is returned back to space due to the high albedo of ice, but even then, there is enough heat to melt the ice.

This paper removes the ice on July 1st, because if they remove the ice before then refreezing begins. How? How in the world they remove all the ice during solstice and refreeze begins but if they do it 10 days later refreeze doesn't?

From the paper:

Quote
To examine the recovery mechanisms of Arctic summer sea ice, we simulate the consequences of an ice‐free Arctic Ocean during summer. We set up experiments to start on 1st July from initial conditions that are taken from the reference run, but are perturbed by converting the entire Northern Hemisphere sea ice to water with the same properties as the sea surface water below the ice. Such conversion of relatively fresh sea ice to salty sea water has the advantage of leaving the properties of sea surface water unchanged. The start date is chosen such that the effect of the perturbation is maximal: starting from ice‐free conditions earlier in the year leads to immediate re‐freezing,

This makes no physical sense. By june 21st, with the arctic full of meters thick, cold ice with high albedo, there is more than enough melting to spread across the whole arctic. However this model says that if all ice was removed by the 21st of june refreeze would immediately begin?

What am I missing?

2.

From the article and very important because all the paper I've seen are counting on negative auto correlation:

Quote
For SAT a large positive anomaly occurs between October and February after the initial perturbation, with a peak of almost 11 K in November (Figure 2). After February, there are no further SAT anomalies stronger than natural variability.


This does not match the reality observed during winter since 2016. We have seen how the Chukchi remains very warm, way past November, even with a giant mass of ice to the North.

We are also starting to see the accumulation of heat are having long term effect in the Bering and Chukchi.

Quote
While it is true that the most energy being received by the earth from the sun is at noon, the thermal response means that all that energy isn’t felt in the air for at least a few hours. So, if you add three or four hours to noon, you get 3 or 4 p.m.

The thermal response of the Arctic to the sun begins when the melting season begins, April in the outer periphery, May in the inner Arctic. By June melting is everywhere, that means the thermal response is everywhere. We don't see it in temperatures because the ice must melt before temperatures can rise. 

The only other source for cold is the snow of the continents. May most of the snow is already gone and the ASI is the only source of cold that must be overcome by thermal inertia.

This study has a source of cold that makes the ice return during solistice. I imagine the same source keeps the ocean of almost 0 albedo at barely 2k warmer, even when we have seen the anomalies that can form.


https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2010GL045698
Tietsche et al 2011
Recovery mechanisms of Arctic summer sea ice


Let me start with the carefully worded abstract, my bold:

Quote
Our results suggest that anomalous loss of Arctic sea ice during a single summer is reversible, as the ice–albedo feedback is alleviated by large‐scale recovery mechanisms.

Please correct me if I'm wrong, but what I read here is that if the ice disappears but everything else stays about the same, the ice will recover. What if the ice disappears because conditions got so bad that the ice melted? And what if the conditions persist the following year with a Arctic ocean full of thin, mobile, transparent ice?

This conclusion does not follow:

Quote
Hence, hysteretic threshold behavior (or a “tipping point”) is unlikely to occur during the decline of Arctic summer sea‐ice cover in the 21st century.

It is unlikely that the ice disappears anomalously. It is likely that the ice will disappear when the conditions for melting are met.

Quote
In this reference run, annual mean surface air temperature in the Arctic rises from −14°C in the 1900s to −4°C in the 2090s. Arctic sea‐ice extent declines, and the Arctic Ocean is typically ice‐free by the end of summer from 2070 onward

This is terrifying to me. Why was this model so wrong?  2070 sounds like a terrible under estimation after seeing the 2010's.

Quote
The start date is chosen such that the effect of the perturbation is maximal: starting from ice‐free conditions earlier in the year leads to immediate re‐freezing, and hence both earlier and later start dates imply shorter exposure of open water to sunlight, and a less pronounced ice–albedo effect.

This again. If all of the Arctic ocean is ice free, maximum exposure of open water to sunlight is on June 21. June 10th? would have the same exposure to the sun as July 1st. This is a simple astronomical fact. Why are they using it as an explanation?

Quote
[14] All our experiments start from sea‐ice free conditions on 1st July. As expected, the Arctic Ocean remains ice‐free for several months, and significant sea‐ice cover does not develop before November.

I believe the November start day for  freezing onset. It will be dark and cold regardless of BOE. I also believe fast freezing in extent. Volume will rise fast, but thickness won't go much more than 2m for the same reason. Thin ice thickens fast, but thick ice thickens slow and the time for thickening was cut short by the November start day. Come the melting season it will be thin, warm, first year ice across the whole arctic.

At this point I'm done with this paper until the start date for the simulation is correctly explained.

Now I'm going to look for the other papers crandles posted.

Maybe my answers are there.

Let's start with the beginning.

https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2007GL030253
D. Schröder  W. M. Connolley
Impact of instantaneous sea ice removal in a coupled general circulation model


There are either two things seriously wrong with this model or I have a serious misunderstanding of how the arctic works.
1.

The melting season over the Arctic starts somewhere in April.  By May the periphery is melting fast and the inside forming meltponds. By June, CAB melting begins. This melting has one principal culprit, the sun. Of course most of that sun energy is returned back to space due to the high albedo of ice, but even then, there is enough heat to melt the ice.

This paper removes the ice on July 1st, because if they remove the ice before then refreezing begins. How? How in the world they remove all the ice during solstice and refreeze begins but if they do it 10 days later refreeze doesn't?

From the paper:

Quote
To examine the recovery mechanisms of Arctic summer sea ice, we simulate the consequences of an ice‐free Arctic Ocean during summer. We set up experiments to start on 1st July from initial conditions that are taken from the reference run, but are perturbed by converting the entire Northern Hemisphere sea ice to water with the same properties as the sea surface water below the ice. Such conversion of relatively fresh sea ice to salty sea water has the advantage of leaving the properties of sea surface water unchanged. The start date is chosen such that the effect of the perturbation is maximal: starting from ice‐free conditions earlier in the year leads to immediate re‐freezing,

This makes no physical sense. By june 21st, with the arctic full of meters thick, cold ice with high albedo, there is more than enough melting to spread across the whole arctic. However this model says that if all ice was removed by the 21st of june refreeze would immediately begin?

What am I missing?

2.

From the article and very important because all the paper I've seen are counting on negative auto correlation:

Quote
For SAT a large positive anomaly occurs between October and February after the initial perturbation, with a peak of almost 11 K in November (Figure 2). After February, there are no further SAT anomalies stronger than natural variability.


This does not match the reality observed during winter since 2016. We have seen how the Chukchi remains very warm, way past November, even with a giant mass of ice to the North.

We are also starting to see the accumulation of heat are having long term effect in the Bering and Chukchi.

This is Tietsche et al paper not Schroeder & Connolley.

With 1. I agree that seems very surprising. If they were converting sea ice (pretty fresh) and simply dumping that as water in liquid form of the same salinity as the sea ice and at the temperature of the salt water below the ice then yes fresh water at -1.7C would freeze. However they say they are making the water as salty as the water below it so that shouldn't freeze. Would a slight pressure difference make the difference between surface water freezing while just below the sea ice it stays liquid?

This does feel like there is a mistake of some sort here. Perhaps the best way to resolve this is to email  steffen.tietsche at zmaw.de and ask.

Schroeder & Connolley paper does things differently;
Quote
Four sensitivity experiments were performed in which all the initial sea ice was removed on December 1st (starting time of Ctrl run), March 1st, June 1st and September 1st, respectively.

They also did
Quote
[8] The preceding experiments were somewhat unrealistic because although the ice was removed, the ocean was still in a state compatible with ice cover. Hence, in the following experiments the ocean temperature will be modified to examine an ice‐free situation in the real world where an ice anomaly is connected with an ocean heat anomaly. Where and how should the ocean temperature be modified to simulate a realistic ice‐free situation? To answer this question the annual cycle of the ocean is analysed in the transitional zone with seasonal sea ice (not shown). The uppermost 200 m of the polar oceans are affected by seasonal changes in the HadCM3 run. A maximum ocean temperature of about 3°C is reached in areas which are seasonally covered by sea ice. Based on these findings two further sensitivity experiments are performed in which the initial global sea ice is removed and the ocean temperature of the uppermost 200 m (10 model levels) is artificially increased to a minimum value of 3°C on March 1st and September 1st, respectively. The initial salinity and ocean circulation remain the same as in the Ctrl run.

Here the anomalies persist a few more years, but the size of the sudden forced departure is clearly much larger.

I am feeling that Tietsche et al could have done more to make the departure larger by starting earlier and adding .1C or .2C to the temperature to stop an immediate refreeze. If that is correct and they could have done more, does it matter to the conclusions of the paper? It is still a huge departure from normal conditions and it still recovers pretty rapidly. And the Scroeder and Connolley have done similar but larger departures with similar results.

Re 2:
I am not sure I am understanding your point.
>"We have seen how the Chukchi remains very warm, way past November"
and
"peak of almost 11 K in November (Figure 2). After February, there are no further SAT anomalies stronger than natural variability"

Seems pretty sensible to me: the peak of the anomaly is in November as you are pointing out. The higher temperature then radiates more heat to space resulting in the temperature anomaly disappearing over the next month or so. After that there is still a difference in the thickness of the ice. The anomalously thin ice ice results in extra heat loss (there is less insulating ice) which means more ice forms at a faster rate so the ice thickness anomaly largely disappears over the following month or so. If there is less snow supported by the newer ice, that newer ice may grow thicker.

If the sun was the only source of heat, you might expect the peak of the anomaly to be around end of September/ early October when the sun sets. However there is also downwards longwave radiation from the atmosphere, and with large areas largely ice free, it doesn't seem unreasonable for this to go on increasing the size of the peak anomaly a bit longer via the temperature dropping more slowly such that the size of the anomaly increases into November.

Perhaps you could explain why you are seeing this as a problem?
« Last Edit: July 27, 2019, 12:57:20 PM by crandles »

crandles

  • Nilas ice
  • Posts: 2391
    • View Profile
  • Liked: 52
  • Likes Given: 42

Please correct me if I'm wrong, but what I read here is that if the ice disappears but everything else stays about the same, the ice will recover. What if the ice disappears because conditions got so bad that the ice melted? And what if the conditions persist the following year with a Arctic ocean full of thin, mobile, transparent ice?

This conclusion does not follow:

Quote
Hence, hysteretic threshold behavior (or a “tipping point”) is unlikely to occur during the decline of Arctic summer sea‐ice cover in the 21st century.

It is unlikely that the ice disappears anomalously. It is likely that the ice will disappear when the conditions for melting are met.

Yes I agree it is true that "It is unlikely that the ice disappears anomalously. It is likely that the ice will disappear when the conditions for melting are met."

However, the paper is looking at hysteresis and concluding hysteresis behaviour is unlikely so the sentence with the word hysteretic in it, is supported. Disturb it from equilibrium and it pretty well goes back to its equilibrium position.

>What if the ice disappears because conditions got so bad that the ice melted?
Is just a completely different question.


Quote
In this reference run, annual mean surface air temperature in the Arctic rises from −14°C in the 1900s to −4°C in the 2090s. Arctic sea‐ice extent declines, and the Arctic Ocean is typically ice‐free by the end of summer from 2070 onward

This is terrifying to me. Why was this model so wrong?  2070 sounds like a terrible under estimation after seeing the 2010's.

Those dmi graph temperatures are not properly averaged, they give far too much weight to temperatures very close to the poll. Not sure I have a very good handle on what these annual mean temperatures has been and how much they should be expected to change in order to judge whether it is a terrible understatement. You having that reaction doesn't surprise me. I'll have to think about whether there is anything appropriate I can say to provide some explanation.

Quote
This is Tietsche et al paper not Schroeder & Connolley.
You are correct. I got them mixed up. Typical for me. My apologies.

Quote
This does feel like there is a mistake of some sort here. Perhaps the best way to resolve this is to email  steffen.tietsche at zmaw.de and ask.

There is a mistake. Whatever the reason is, it is not because there is less sun hitting open ocean. 

>What if the ice disappears because conditions got so bad that the ice melted?
Is just a completely different question.

But isn't that the question we need answered?

Quote
Schroeder & Connolley paper does things differently;

ok. From the real Connolley paper this time:

Quote
Starting without sea ice in December (blue line) the “normal” sea ice area recovers in only one month in the Arctic

This makes perfect sense. If all the sea ice in the arctic is magically (technical term) removed while leaving the rest of the arctic as in the same state as any freezing season before 2016/17, the ice would indeed return extremely fast. There would be some evaporation and some clouds to retard heat irradiation, but it won't matter. The top layer of the ocean would freeze extremely fast.

The question then becomes, in this simulated scenario how thick will the ice become by April?

Let's assume no clouds form, there is nothing but clear skies, the ocean remains nice and calm, the rest of the NH doesn't exist and all the ice returns by January 1st. Then we assume the ice has 4 months of 2000's level winter. Very cold, no intrusions from the south.

Making a crude visual calculation:





How thick would the ice be come April, with winter like the 2000's?

I estimate about 1.5 meters thick.

Could such ice take a 2000's level summer?  How about a 2010's level summer? what about export?

At the very least you must admit that if the ice was magically removed in December,  the likelihood of a BOE would be very high. The thin ice would be very vulnerable to even an average melting season.

Next I was expecting to cover what happens when the ice is removed during summer but, alas, This is pretty much all it says:

Quote
A longer‐lasting impact is achieved by removing the sea ice in summer. This is because no sea ice can build up during summer, and if no sea ice is present the reduced surface albedo causes an increase in ocean temperature (up to 2.5 K for the area mean in the Arctic and 1 K in the Antarctic at a depth of 5 m) which delays freezing in the next autumn.

And looking at their graph I can see why they don't talk much about any of the other scenarios

Here it is:


 
Notice something interesting? Removing ice during either December or June result in a BOE the following year. And the year following that, but they it gets a bit better and it heals.

This defies logic and science. Even under very limited very favorable models used here, hysteresis happens for a few years and then it magically disappears.

This defies belief. I mean, not even a mention of this. They skipped right over the consecutive BOEs to arrive to the conclusion that in a few years the ice would recover. They could have at least mention it.

So what are these heat sinks that the scientists see but you don't?

They chose July 1st as the date for the first removal of the ice  and then they claim:

Quote
The start date is chosen such that the effect of the perturbation is maximal: starting from ice‐free conditions earlier in the year leads to immediate re‐freezing, and hence both earlier and later start dates imply shorter exposure of open water to sunlight, and a less pronounced ice–albedo effect.

The day that maximizes open water to sunlight is June 21st. That implication is obviously not true.   

How thick would the ice be come April, with winter like the 2000's?

I estimate about 1.5 meters thick.

Could such ice take a 2000's level summer?  How about a 2010's level summer? what about export?

At the very least you must admit that if the ice was magically removed in December,  the likelihood of a BOE would be very high. The thin ice would be very vulnerable to even an average melting season.

Yes I will admit that likelihood of a BOE would be very high and indeed that is what the model shows:



volume getting up to 1.3 * 10^13 m^3 instead of the normal model level of 1.5 * 10^13 m^3.
Not sure I can judge the dates accurately but next summer is ice free for roughly 3 months so the ice starts to form by November if not late October. This gives more time for ice to form so the volume gets up pretty close to the low edge of the normal model range.


sorry mangled that answer confusing blue(1 Dec removal) and black (1 March removal) lines. Model shows after 1 Dec removal volume practically back to normal by May. By September volume at minimum is near 1k km^3 below normal but not by much

1March removal leaves little time for volume build up. This is where volume reached 1.3 vs normal 1.5 *10^13 m^3 by a late maximum in May. This seems surprisingly high to me.

The later sentence
Quote
The preceding experiments were somewhat unrealistic because although the ice was removed, the ocean was still in a state compatible with ice cover.
may have some relevance here.

Anyway not surprising this date of removal has more effect with ice free conditions reached for about 3 months circa Aug-Oct.

1 Sept removal (red line) has least effect, it is straight back to a normal volume certainly by April and perhaps normal as soon as November.

Unsurprisingly, 1 April removal (green) has most effect: There is no significant ice formation til Dec. The next 2 minimum show the green line as lowest just below the normal range then it settles back into the normal range.

Quote
Next I was expecting to cover what happens when the ice is removed during summer but, alas, This is pretty much all it says:

Quote
A longer‐lasting impact is achieved by removing the sea ice in summer. This is because no sea ice can build up during summer, and if no sea ice is present the reduced surface albedo causes an increase in ocean temperature (up to 2.5 K for the area mean in the Arctic and 1 K in the Antarctic at a depth of 5 m) which delays freezing in the next autumn.

And looking at their graph I can see why they don't talk much about any of the other scenarios

Here it is:


 
Notice something interesting? Removing ice during either December or June result in a BOE the following year. And the year following that, but they it gets a bit better and it heals.

This defies logic and science. Even under very limited very favorable models used here, hysteresis happens for a few years and then it magically disappears.

This defies belief. I mean, not even a mention of this. They skipped right over the consecutive BOEs to arrive to the conclusion that in a few years the ice would recover. They could have at least mention it.


You have read the caption under that figure which says
Quote
Sensitivity experiments without sea ice and modified ocean temperature: time series of ice area (m2), ice volume (m3), and area mean (60°N to 90°N and 60°S to 90°S) ocean temperature at a depth of 5 m and 204 m. The orange area represents the climate of the Ctrl run ± twice the standard deviation. The green area represents the observed mean ice area for the period 1980 to 2000 [Comiso, 2003].

"modified ocean temperature" means this is where they did the extra large departures from normal:
Quote
the initial global sea ice is removed and the ocean temperature of the uppermost 200 m (10 model levels) is artificially increased to a minimum value of 3°C on March 1st and September 1st, respectively.

200m increased to a minimum of 3C, that is huge volume of water to pretty warm temperatures. That is a massive departure from normal conditions. The top 50m (mixed layer) may well cool to close to normal in the first winter but the rest is going to persist at warmer than normal for several years and provide extra heat to the system for several year.

Despite this massive departure from normal, the appropriate conclusion seems to me to be it recovers in just a few years (~8).

crandles

  • Nilas ice
  • Posts: 2391
    • View Profile
  • Liked: 52
  • Likes Given: 42
Quote
Despite this massive departure from normal, the appropriate conclusion seems to me to be it recovers in just a few years (~8).

If we assume 1980-2000 climate, ignore clouds, atmospheric disruptions and pretend the world will remain static to the disruption caused by an ice less, warming Arctic Ocean, like in the models, then sure. Hysteresis is only temporal and local.

That we are basing our risk assessment for climate change based on an idealized model with serious inconsistencies is plain wrong.

Quote
"modified ocean temperature" means this is where they did the extra large departures from normal:

Which they did only once, to simulate the missing heat that the models are not reflecting. On the BOE that followed the first BOE they didn't add the heat, even when the circumstances were the same. Thus the first BOE is faked and heat is added to better represent reality. This first BOE produced a natural BOE is not faked but the heat correction is not added. Instead the bias of the model is left to take over with 1980 climate and restore the ice.

I bet you that if the heat was added in the second BOE (to compensate for the model's flaw) the third year would also produce a BOE.

If a 2010's reference run was used instead of 1980-2000 they likely won't need to add extra heat.

Quote
Despite this massive departure from normal, the appropriate conclusion seems to me to be it recovers in just a few years (~8).

If we assume 1980-2000 climate, ignore clouds, atmospheric disruptions and pretend the world will remain static to the disruption caused by an ice less, warming Arctic Ocean, like in the models, then sure. Hysteresis is only temporal and local.

That we are basing our risk assessment for climate change based on an idealized model with serious inconsistencies is plain wrong.

You assumed 1980-2000 climate in your rough estimation, the model used didn't.
You ignored clouds, the model used didn't. (maybe rather parameterised but they are included in models)

Models are imperfect, they attempts to include all the important dynamics as far as they can. Yes there are possibilities that things considered unimportant might interact to create import effects that are then missed by the model or the necessary simplifications are just a bit too simple. However they are the best that we have got.

.

What "serious inconsistencies"?

1 July being stated as maximal effect? I agree this looks odd, but you are not yet at the point of being able to say it is a 'serious inconsistency'. We simply don't understand it and there are various possibilities:

It might be the paper pushing what has actually been done, or
it might be we simply don't understand something.

Even if the paper is pushing what has actually been done, it doesn't mean it has any effect on the papers conclusions.

.

If models are the best we have got, why do you assert that using them is plain wrong?

Quote
"modified ocean temperature" means this is where they did the extra large departures from normal:

Which they did only once, to simulate the missing heat that the models are not reflecting. On the BOE that followed the first BOE they didn't add the heat, even when the circumstances were the same. Thus the first BOE is faked and heat is added to better represent reality. This first BOE produced a natural BOE is not faked but the heat correction is not added. Instead the bias of the model is left to take over with 1980 climate and restore the ice.

I bet you that if the heat was added in the second BOE (to compensate for the model's flaw) the third year would also produce a BOE.

If a 2010's reference run was used instead of 1980-2000 they likely won't need to add extra heat.

The reason they added the heat is explained:
Quote
The preceding experiments were somewhat unrealistic because although the ice was removed, the ocean was still in a state compatible with ice cover.

This isn't a model flaw that means the extra heat needs to be added every year. It is a massively overdone effect to ensure the ocean state being compatible with ice cover is not causing too much rapid ice formation that stops the albedo feedback that they want to introduce to see how much effect that sudden introduction of albedo feedback has.

Quote
This isn't a model flaw that means the extra heat needs to be added every year. It is a massively overdone effect to ensure the ocean state being compatible with ice cover is not causing too much rapid ice formation that stops the albedo feedback that they want to introduce to see how much effect that sudden introduction of albedo feedback has.

Crandles,  the bold part is wrong. It is not a massively exaggerated number.They took the values of the model runs of the time calculated a number for the whole Arctic. Bet ya peanuts to dollars that the number they used is too low relative with recent temperatures.

FTA:

Quote
The preceding experiments were somewhat unrealistic because although the ice was removed, the ocean was still in a state compatible with ice cover. Hence, in the following experiments the ocean temperature will be modified to examine an ice‐free situation in the real world where an ice anomaly is connected with an ocean heat anomaly. Where and how should the ocean temperature be modified to simulate a realistic ice‐free situation? To answer this question the annual cycle of the ocean is analysed in the transitional zone with seasonal sea ice (not shown). The uppermost 200 m of the polar oceans are affected by seasonal changes in the HadCM3 run. A maximum ocean temperature of about 3°C is reached in areas which are seasonally covered by sea ice. Based on these findings two further sensitivity experiments are performed in which the initial global sea ice is removed and the ocean temperature of the uppermost 200 m (10 model levels) is artificially increased to a minimum value of 3°C on March 1st and September 1st, respectively.

And then this:

Quote
The initial salinity and ocean circulation remain the same as in the Ctrl run.

Quite literally, we are betting our lives on weak assumptions.
 
Quote
You assumed 1980-2000 climate in your rough estimation, the model used didn't.

That's what they used.

Quote
A 20‐year run is used as the reference run (called “Ctrl” in the following) beginning from a standard HadCM3 state taken from the control run, which has greenhouse gas forcing appropriate to pre‐industrial levels. Thus, these 20 years are stable with respect to atmospheric and sea ice parameters.

Quote
The green area represents the observed mean ice area for the period 1980 to 2000

Had they used more recent, CO2 forced data the result would probably change.

Quote
You ignored clouds, the model used didn't. (maybe rather parameterised but they are included in models)

Bet ya the model parameters pale in comparison with the clouds experienced during the last few years.  I used the old average to give you the biggest benefit of doubt. If I had use more recent years, as I should have, your case becomes even worse.


Quote
What "serious inconsistencies"?

The fact that a BOE faked under 1980's climate without co2 forcing and with no variation of the rest of the climate produces 2 consecutive BOEs under these darling scenarios, yet the conclusion of the paper is that we can safely assume hysteresis won't happen.

Quote
If models are the best we have got, why do you assert that using them is plain wrong?

I'm not asserting such thing. I'm using them myself, but I'm using them not as a bible  with an ultimate truth, but as microcosm of reality that they are. All models are wrong, some are useful.

These models are right, but the conclusion of the authors are highly misleading. These models say  that when a BOE happens in a world that does not change and is incapable of supporting a BOE, hysteresis is temporary. That is a very bad sign for when the world actually does support a BOE.

These models give the first hints of hysteresis and warrant a much closer look. Yet they are being used to pretend there will be no hysteresis.


Quote
You assumed 1980-2000 climate in your rough estimation, the model used didn't.

That's what they used.

Quote
A 20‐year run is used as the reference run (called “Ctrl” in the following) beginning from a standard HadCM3 state taken from the control run, which has greenhouse gas forcing appropriate to pre‐industrial levels. Thus, these 20 years are stable with respect to atmospheric and sea ice parameters.


Sorry, yes you are right. I was confusing the two papers.

Quote
The green area represents the observed mean ice area for the period 1980 to 2000.

1980 to 2000 is observations.

They are using "greenhouse gas forcing appropriate to pre‐industrial levels", which I imagine you find rather alarming. Perhaps it would be rather concerning if Tietsche et al hadn't redone it with more modern & realistic climate following A1B emission scenario.

These models are right, but the conclusion of the authors are highly misleading. These models say  that when a BOE happens in a world that does not change and is incapable of supporting a BOE, hysteresis is temporary. That is a very bad sign for when the world actually does support a BOE.

These models give the first hints of hysteresis and warrant a much closer look. Yet they are being used to pretend there will be no hysteresis.

What first hints of hysteresis? They are being given enormous perturbation, after which some deviation from normal are expected but the question is whether they settle down into the same state or some different state. Do you see noticeably different states that the models settle down to?


1980 to 2000 is observations.

And the control runs match the observations. The model climate is one similar to the 1980 to 2000 climate.

Quote
They are using "greenhouse gas forcing appropriate to pre‐industrial levels", which I imagine you find rather alarming. Perhaps it would be rather concerning if Tietsche et al hadn't redone it with more modern & realistic climate following A1B emission scenario.

Realistic?

Quote
In this reference run, annual mean surface air temperature in the Arctic rises from −14°C in the 1900s to −4°C in the 2090s. Arctic sea‐ice extent declines, and the Arctic Ocean is typically ice‐free by the end of summer from 2070 onward

The first ice free arctic ocean will occur much sooner than 2070.

Quote
The start date is chosen such that the effect of the perturbation is maximal: starting from ice‐free conditions earlier in the year leads to immediate re‐freezing, and hence both earlier and later start dates imply shorter exposure of open water to sunlight, and a less pronounced ice–albedo effect.

This is a fundamental flaw of this model. With no ice over the Arctic, there is no reason for refreeze to commence during June. During June the ice dictates the sea and wind surface temperatures. Except for the deep ocean and Greenland, there is no other heatsink. However because of the nature of the perturbation, the appearance of a heat sinks is given.

Quote
As shown by Serreze et al. [2007], heat transport into the Arctic Ocean by advection of warm water and export of sea ice is only between 4 and 7 Wm−2 (March and August mean, respectively). Our model shows comparable results for oceanic heat transport into the Arctic.

...

Hence, we find that anomalies in oceanic heat transport into the Arctic are unimportant for the observed recovery of the Arctic energy budget.

[21] Consequently, the oceanic heat content anomaly is determined by the remaining two factors: (i) the latent heat anomaly induced by the initial conditions of the experiment and (ii) the surface heat flux anomaly.

Those are not the only two factors that remain, but the factor I'm about to mention was not evident at the time of this paper. Heat transport by air, facilitated by the breakdown of the atmospheric currents that we are already witnessing and accompanying clouds.

In fact the paper counts on the exact opposite.

Quote
Atmospheric heat content and lateral heat transport are not significantly affected

What we are seeing is the hemisphere bombard the arctic with heat at levels never seen before. Lateral heat transport is increasing tremendously. This model assumes lateral heat is insignificant.

We need newer models that better describe the changes happening in the arctic. Risk assessment and planetary level decisions are being made with unrealistic models, incomplete models.

crandles

  • Nilas ice
  • Posts: 2391
    • View Profile
  • Liked: 52
  • Likes Given: 42

Quote
In this reference run, annual mean surface air temperature in the Arctic rises from −14°C in the 1900s to −4°C in the 2090s. Arctic sea‐ice extent declines, and the Arctic Ocean is typically ice‐free by the end of summer from 2070 onward

The first ice free arctic ocean will occur much sooner than 2070.

Sounds entirely plausible. That is what that model says but that is only one model and we should treat models as a tool not reality. If every model said 2070 then maybe we should start to consider whether to believe the models. But they don't so we certainly shouldn't treat this model as gospel truth.

Having said this, what you wrote is a rather emphatic assertion that requires evidence to justify.


Quote
The start date is chosen such that the effect of the perturbation is maximal: starting from ice‐free conditions earlier in the year leads to immediate re‐freezing, and hence both earlier and later start dates imply shorter exposure of open water to sunlight, and a less pronounced ice–albedo effect.

This is a fundamental flaw of this model. With no ice over the Arctic, there is no reason for refreeze to commence during June. During June the ice dictates the sea and wind surface temperatures. Except for the deep ocean and Greenland, there is no other heatsink. However because of the nature of the perturbation, the appearance of a heat sinks is given.

It is something we don't understand. I have suggested you email the corresponding author. There are all sorts of answers you might get. For example, a simple one might be that they tried 4 different starting points 1 Jan, 1 April, 1 July and 1 Sept and found that the 1 July starting point had most effect. With such an answer you may be able to see that the paper might be pushing its interpretation of when the effect is maximum but that this does not indicate a fundamental flaw in the model. You need to understand it before you can go reaching conclusions like "This is a fundamental flaw of this model"

Quote
Quote
As shown by Serreze et al. [2007], heat transport into the Arctic Ocean by advection of warm water and export of sea ice is only between 4 and 7 Wm−2 (March and August mean, respectively). Our model shows comparable results for oceanic heat transport into the Arctic.

...

Hence, we find that anomalies in oceanic heat transport into the Arctic are unimportant for the observed recovery of the Arctic energy budget.

[21] Consequently, the oceanic heat content anomaly is determined by the remaining two factors: (i) the latent heat anomaly induced by the initial conditions of the experiment and (ii) the surface heat flux anomaly.

Those are not the only two factors that remain, but the factor I'm about to mention was not evident at the time of this paper. Heat transport by air, facilitated by the breakdown of the atmospheric currents that we are already witnessing and accompanying clouds.

In fact the paper counts on the exact opposite.

Jennifer Francis' and others work on jet stream meandering is to some extent done with models. So it isn't necessarily not in these models. It seems to remain rather contentious and while I tend to believe there is an effect, measuring the size of the effect seems to remain difficult. This should tell you that it is likely that it isn't a huge effect.

Quote
Quote
Atmospheric heat content and lateral heat transport are not significantly affected

What we are seeing is the hemisphere bombard the arctic with heat at levels never seen before. Lateral heat transport is increasing tremendously. This model assumes lateral heat is insignificant.

We need newer models that better describe the changes happening in the arctic. Risk assessment and planetary level decisions are being made with unrealistic models, incomplete models.

Why are you saying "This model assumes lateral heat is insignificant." That isn't the way the models work: Basic physics is put into models so anything like "Atmospheric heat content and lateral heat transport are not significantly affected" is a result coming out of the model not an assumption that is put into the model.

You can assert things like "we are seeing is the hemisphere bombard the arctic with heat at levels never seen before" fine. If you have evidence of a model failing to do what it should then you can start asserting the models are flawed. Until you show that, this looks more like your preconceived conclusion that there must be something wrong showing up in what you are saying.

oren

  • Young ice
  • Posts: 4172
    • View Profile
  • Liked: 670
  • Likes Given: 1190
Although the science is way above my level, my main beef with this important paper is the refreeze part.  Having seen the late refreeze seasons of 2016 in the surrounding seas (especially the Chukchi) and 2018 in the CAB, I think there must be a wrong assumption somewhere here regarding a stable autumn climate. The same beef I have with Chris Reynold's Slow Transition. But I am unable to pursue it further.

Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161
crandles, thank you for opening this topic and your insightful feedback.

I would like to add to the record some text that you posted as the IPCC justification for no hysteresis.

Quote
Regarding the behavior of Arctic sea ice under decreasing temperatures following a possible overshoot of a long-term temperature target, a substantial number of pre-AR5 studies have found that there is no indication of hysteresis behavior of Arctic sea ice (Holland et al., 2006; Schroeder and Connolley, 2007; Armour et al., 2011; Sedláček et al., 2011; Tietsche et al., 2011; Boucher et al., 2012; Ridley et al., 2012). In particular, the relationship between Arctic sea-ice coverage and GMST is found to be indistinguishable between a warming scenario and a cooling scenario.

We have already talked Tietsche and Schroeder, there is more to talk about yet, but to help set up this thread I've found links for what I believe to be the rest of the references used by the IPCC to claim  "there is no indication of hysteresis behavior of Arctic sea ice".


Holland et al., 2006

https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/2006GL028024

Armour et al., 2011

https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2010GL045850

Sedláček et al., 2011

https://journals.ametsoc.org/doi/full/10.1175/2011JCLI3904.1

Boucher et al., 2012

https://iopscience.iop.org/article/10.1088/1748-9326/7/2/024013

Ridley et al., 2012

https://iopscience.iop.org/article/10.1088/1748-9326/7/2/024013


I'll give you a reply to your latest feedback soon.
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

crandles

  • Nilas ice
  • Posts: 2391
    • View Profile
  • Liked: 52
  • Likes Given: 42
crandles, thank you for opening this topic and your insightful feedback.


Glad you find some useful. Sorry if some others are rubbish / gut reactions / overly harsh and probably other descriptions apply.

crandles

  • Nilas ice
  • Posts: 2391
    • View Profile
  • Liked: 52
  • Likes Given: 42
Although the science is way above my level, my main beef with this important paper is the refreeze part.  Having seen the late refreeze seasons of 2016 in the surrounding seas (especially the Chukchi) and 2018 in the CAB, I think there must be a wrong assumption somewhere here regarding a stable autumn climate. The same beef I have with Chris Reynold's Slow Transition. But I am unable to pursue it further.

late 2016 the ice grew slowly making 2017 max volume a record low by some margin. Is this an outlier or some new trend of rapid loss in max volume? 2018 and 2019 isn't much data to decide that yet but they were more back to previous normal.

I must say I tend to note what happens in the 2017 melt season more. A record low max volume should give more opportunity for open water formation so albedo feedback can get started earlier. Did it pull further away from the pack or even maintain the gap? No ended up back in the pack. One year is not much data and weather can beat the trends you might otherwise expect. Even so, sounds more like an outlier and reversion to normal to me. I guess we will see.


DrTskoul

  • Nilas ice
  • Posts: 1402
    • View Profile
  • Liked: 175
  • Likes Given: 59
There a few independent mechanisms:
1) heat storage in open water that slows ice re growth
2) location of open water in arctic - different areas experience slightly different fall/winter transition
3) NH weather setup for fall/winter
4) heat and water vapor release from open water

The exact transition from minimum to maximum depends on the interaction between those mechanisms, and the temporal and spatial coincidence of those interactions. So for most of the time it seems like a chaotic system. Every now and again the setup is just right and you get 2016/2017 winter.

However, the system is forced due to GHGs. 1, 2 and 4 are affected directly by the forcing, while 3 indirectly initially ( stronger in future ). As the open area distribution becomes consistent every year, a larger portion of the arctic only has FYI, and the heat in the system leads to hotter ocean, then at some point we will reach a tipping point where the maximum volume will come down consistently every year. Until it stabilizes when most arctic ice has FYI. Then it will plateau until portions of arctic start seeing year around ice free areas. Year to year still will be chaotic, but after that point as long as the forcing exists the maximum will start diminishing again.


crandles

  • Nilas ice
  • Posts: 2391
    • View Profile
  • Liked: 52
  • Likes Given: 42
There a few independent mechanisms:
1) heat storage in open water that slows ice re growth
2) location of open water in arctic - different areas experience slightly different fall/winter transition
3) NH weather setup for fall/winter
4) heat and water vapor release from open water

The exact transition from minimum to maximum depends on the interaction between those mechanisms, and the temporal and spatial coincidence of those interactions. So for most of the time it seems like a chaotic system. Every now and again the setup is just right and you get 2016/2017 winter.

However, the system is forced due to GHGs. 1, 2 and 4 are affected directly by the forcing, while 3 indirectly initially ( stronger in future ). As the open area distribution becomes consistent every year, a larger portion of the arctic only has FYI, and the heat in the system leads to hotter ocean, then at some point we will reach a tipping point where the maximum volume will come down consistently every year. Until it stabilizes when most arctic ice has FYI. Then it will plateau until portions of arctic start seeing year around ice free areas. Year to year still will be chaotic, but after that point as long as the forcing exists the maximum will start diminishing again.

Thanks for the interesting post.

Unfortunately I am not sure I am following.

>"then at some point we will reach a tipping point where the maximum volume will come down consistently every year. Until it stabilizes when most arctic ice has FYI."

So is this discussing 2000-2012 period or some future period?

Why a tipping point? Around the edges ice free for more of the year so this ice is thin. Further into centre of ice, it is ice free for less of the year so less heat built up in ocean, surrounded by more nearby ice and ice grows thicker. Why a tipping point as opposed to the ice ice gradually retreating year by year subject to weather variation?

>"Until it stabilizes when most arctic ice has FYI. Then it will plateau "
No clear about this either. If the GHG levels are going up and ocean and atmospheric temperatures of wind and currents coming into arctic is going up... Should we be talking of a slow decline in max volume (compared to faster rate 2000-2012 when MYI was rapidly disappearing)?

>"has FYI" is this rather ambiguous? Are you talking about at annual maximum or minimum or...?

Sorry if I m being thick in not understanding what you are conveying.

DrTskoul

  • Nilas ice
  • Posts: 1402
    • View Profile
  • Liked: 175
  • Likes Given: 59
You are not thick. I am not explaining it clearly.

* the annual volume maximum has been decreasing but with lots of rebounds after 2012, 2016 etc. The increasing heat is not enough yet to fully counteract the fall/winter negative feedbacks. The tipping point that I refer to is when the heat added will be large enough to overcome the negative feedback and the overall thickness of the arctic will reduce to the point that we will have mostly first year ice.

At that point every winter the ice will be growing to almost the same volume.  The heat wont be enough for a while for perennial open waters in the arctic.

And yes you are right regarding the melt rates. I added a quick phone sketch...

Edit: added a better graph
« Last Edit: July 27, 2019, 06:50:53 PM by DrTskoul »

Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161
The  model that (Schroeder and Connolley, 2007)  use represent a simple climate without industrial era CO2. Please see the control runs in orange relative to the "green" (1980-2000 climate). 

That isn't the world we live in. This model represents a different world that is cooler and more unchanging than ours, specially after 2007 and then again after 2016.

In this representation the ice is artificially removed while the rest of the climate is left unchanged. The Arctic is ice free for 2 years after the disruptions and then it eventually fully recovers. This makes perfect sense because in the world simulated by this model there is no global warming and there is no wide scale changes. But we are observing wide scale changes.(citation needed. Are we experiencing hemispheric wide changes in atmospheric and oceanic patterns?)

What can be safely inferred from this model is that without CO2, if the ice magically disappears but the rest of the world stays the same, the Arctic is supposed to freeze back up and recover. That's very good to know and it explains why there has been ice over the arctic for so long. The conditions supported it.

This model does not support the IPCC's conclusion that  "there is no indication of hysteresis behavior of Arctic sea ice" because the climate used in this paper is not a good analogue for our climate.

(Tietsche et al., 2011) attempts to fix that.

TBC
« Last Edit: July 28, 2019, 05:12:25 PM by Archimid »
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161
All models are wrong, some are useful.

(Tietsche et al., 2011) is certainly useful, and to maximize its use we must compare the model with reality and learn.

It is very significant that the model predicts 2070 as the date of the first "natural" BOE. It is significant for the same reason Tietsche et al is an improvement over Schroeder and Connolley,   because the climate it represents more closely resembles the observed reality.

In Schroeder and Connolley the magical disappearance of ice happens in a very cold arctic that was not shrinking. The ice magically disappears but the very cold world remains magically unchanged, thus the world pulls the arctic ice back into place. 

In Tietsche et al the magical disappearance of ice happens in a warming world with a shrinking ice cap, thus it more closely resembles reality.

Tiesche et al can be improved by learning where it doesn't resembles reality, why it doesn't resemble reality and finding a fix.

So where does  Tiesche et al does not resemble reality?

1. 2070

Lets see how more recent runs are doing:



https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2015GL066855

FTA:
Quote
The CMIP5 ensemble (selection based only on ocean area) shows that September is most likely to become ice free approximately between 2045 and 2070, with a median of 2050 (Figure 2a).


 Tietsche et al improves Schroeder and Connolley in the same way in the same way Tietche would be improved if their climate simulation more closely resembled the more likely reality. The more likely reality is that the Arctic ice will go much sooner than 2070. That is supported by the models and the daily reality most definitely supports it.

2. The July 1st anomaly.

From Tietsche et al:

Quote
The start date is chosen such that the effect of the perturbation is maximal: starting from ice‐free conditions earlier in the year leads to immediate re‐freezing, and hence both earlier and later start dates imply shorter exposure of open water to sunlight, and a less pronounced ice–albedo effect.



Quote
Arrow widths are proportional to size of anomalies. (top) Summer phase from July to August. (bottom) Winter phase from September to March.

April to June is ignored. The July 1st start date should have been explained here

Quote
The atmospheric energy budget anomaly is tightly coupled to the surface heat flux. During the summer phase from July to August, when the downward surface heat flux is amplified, the atmosphere only plays a passive role: the excess shortwave absorption of 30 AEU at the surface is balanced by an increase of net shortwave flux at the top of the atmosphere. Atmospheric heat content and lateral heat transport are not significantly affected

This should have been equally true for May and June, if in fact " Atmospheric heat content and lateral heat transport are not significantly affected".

This is cause for great concern because there is a heat sink that is not being accounted for, it is being ignored and an erroneous explanation given instead. If this heat sink that cancels out two months worth of sunshine is not accounted for the model can't be trusted  to plan for the fate of the world.


3.Winter Recovery mechanism

This is the huge one.

Quote
As expected, the Arctic Ocean remains ice‐free for several months, and significant sea‐ice cover does not develop before November. However, sea ice then grows very rapidly, since the growth rate for thin ice is much higher than for thick ice, which acts as a negative feedback on thickness during the growth season

0 ice in November. The freezing season begins in November to an Arctic atmosphere and NH that are likely at the warmest they have been since the eemian.

Yes, ice extent will grow extremely fast. Yes, thin ice thickens very fast, but  come march, the ice will not be 1.5 meters thick and it will not be very cold. The chances for a BOE are extremely high, specially if the world changed due to the BOE disturbance, or the world that Tiesche projects for 2070 is what cause the BOE in the first place.

Tiesche et al does not support the claim that "there is no indication of hysteresis behavior of Arctic sea ice" with enough confidence for decision makers to ignore the possibility of hysteresis.
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161
I forgot another huge one.

4.

Quote
Atmospheric heat content and lateral heat transport are not significantly affected (Figure 3b). However, during the longer winter phase from September to March, when the upward surface heat flux is amplified, the warming of the atmosphere leads to a decreased atmospheric heat transport into the Arctic Ocean domain by 70 AE

The opposite is happening. During the last few warm winters the hemisphere is advecting record amounts of heat and humidity into the Arctic. This experiment is blind about this fact.
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

crandles

  • Nilas ice
  • Posts: 2391
    • View Profile
  • Liked: 52
  • Likes Given: 42

Quote
The atmospheric energy budget anomaly is tightly coupled to the surface heat flux. During the summer phase from July to August, when the downward surface heat flux is amplified, the atmosphere only plays a passive role: the excess shortwave absorption of 30 AEU at the surface is balanced by an increase of net shortwave flux at the top of the atmosphere. Atmospheric heat content and lateral heat transport are not significantly affected

This should have been equally true for May and June, if in fact " Atmospheric heat content and lateral heat transport are not significantly affected".

This is cause for great concern because there is a heat sink that is not being accounted for, it is being ignored and an erroneous explanation given instead. If this heat sink that cancels out two months worth of sunshine is not accounted for the model can't be trusted  to plan for the fate of the world.

I am not sure I am fully following what you are saying.

I read the quoted passage as saying
"During the summer phase from July to August, atmospheric heat content and lateral heat transport are not significantly affected"

If you think it is saying "atmospheric heat content and lateral heat transport are not significantly affected throughout the year" then the diagram shows this to be wrong for Sept to March so I don't think that interpretation is tenable.

Therefore we don't know if this would be equally true for May and June.

>"This is cause for great concern because there is a heat sink that is not being accounted for"

Not following this at all. Why is there a heat sink which is not accounted for? Surely it is saying there is a flux of an extra 30AEU going into the ocean which is reflected as heat held in the ocean?

Later September and onward this gets released to atmosphere at a rate of 110 AEU which warms the atmosphere and being warmer radiates some of the extra (40) to space and there is also decrease in heat flux into Arctic by 70 AEU.


Quote
3.Winter Recovery mechanism

This is the huge one.

Quote
As expected, the Arctic Ocean remains ice‐free for several months, and significant sea‐ice cover does not develop before November. However, sea ice then grows very rapidly, since the growth rate for thin ice is much higher than for thick ice, which acts as a negative feedback on thickness during the growth season

0 ice in November. The freezing season begins in November to an Arctic atmosphere and NH that are likely at the warmest they have been since the eemian.

Yes, ice extent will grow extremely fast. Yes, thin ice thickens very fast, but  come march, the ice will not be 1.5 meters thick and it will not be very cold. The chances for a BOE are extremely high, specially if the world changed due to the BOE disturbance, or the world that Tiesche projects for 2070 is what cause the BOE in the first place.

>"Yes, ice extent will grow extremely fast. Yes, thin ice thickens very fast, but  come march, the ice will not be 1.5 meters thick and it will not be very cold."

To me, this just comes across as I know there is a negative feedback, but I just don't want to believe it will be that strong. My reaction is: Why not? Where is your evidence?

AFAICS, Paper doesn't show ice thickness or volume the following March. It does show the September extents. The model shows zero ice for the first Sept, and around 70% of normal for the next year when done in 1980, i.e. not a BOE the following year. When done in 2020 it is about 50% in the year following the ice removal. In 2040 it gets a bit erratic. But in each case it is getting back to normal after ~5 years.


Quote
Tiesche et al does not support the claim that "there is no indication of hysteresis behavior of Arctic sea ice" with enough confidence for decision makers to ignore the possibility of hysteresis.

If after perturbation some of the models were tending to settle down after a few year to a different level, that would show signs of hysteresis. However, every one of the models shown seems to settle down to same level as before the perturbation. The perturbation is supposed to be designed to maximise chances of hysteresis from different albedo supporting different levels of ice. There looks like there may be an issue with that maximisation. Even so, settling down at the same level each time supports a no hysteresis behaviour prior to a seasonally ice free arctic.

We might prefer a 1 May or 1 June sea ice removal in order to better support the no hysteresis conclusion, but the paper is what we have got.

crandles

  • Nilas ice
  • Posts: 2391
    • View Profile
  • Liked: 52
  • Likes Given: 42
I forgot another huge one.

Quote
Atmospheric heat content and lateral heat transport are not significantly affected (Figure 3b). However, during the longer winter phase from September to March, when the upward surface heat flux is amplified, the warming of the atmosphere leads to a decreased atmospheric heat transport into the Arctic Ocean domain by 70 AE

The opposite is happening. During the last few warm winters the hemisphere is advecting record amounts of heat and humidity into the Arctic. This experiment is blind about this fact.

No it is not.

Of course heat is flowing into Arctic from lower latitudes during Sept to March.

However as the Arctic is warmed up by massive amounts of heat being stored in the ocean during the summer and released to atmosphere during Sept to March this warms the Arctic. The temperature differencials are then less so less heat flows into Arctic. This is therefore equilalent to a net change of heat going out.

Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161

Quote
The atmospheric energy budget anomaly is tightly coupled to the surface heat flux. During the summer phase from July to August, when the downward surface heat flux is amplified, the atmosphere only plays a passive role: the excess shortwave absorption of 30 AEU at the surface is balanced by an increase of net shortwave flux at the top of the atmosphere. Atmospheric heat content and lateral heat transport are not significantly affected
...
Therefore we don't know if this would be equally true for May and June.

May to June is summer, not winter. The "the downward surface heat flux is amplified" as much during May and June as it is during July and August. The only reason the arctic is colder during May and June is because the thermal inertia of the ice must be beaten. Unicorn posted this wonderful video were the he phenomenom can be easily seen.

https://forum.arctic-sea-ice.net/index.php/topic,2591.msg216371.html#msg216371
Edit: Better link.

If you look at the temperature right above the ice that temperature quickly around april. Thats when "the downward surface heat flux is amplified", not July 1st. It is very difficult to tell because most of the heat goes to melt ice. If there was no ice it would immediately warm. Thus there is a heat sink in this model.

Quote
>"This is cause for great concern because there is a heat sink that is not being accounted for"

Not following this at all. Why is there a heat sink which is not accounted for? Surely it is saying there is a flux of an extra 30AEU going into the ocean which is reflected as heat held in the ocean?


Because if you remove the ice in June, maybe even May you remove the biggest heat sink in the NH. The Arctic should immediately warm. But this model, nor the author, understand that. Instead the author relies on an explanation that is wrong (he is not maximizing albedo feedback).

Quote
>"Yes, ice extent will grow extremely fast. Yes, thin ice thickens very fast, but  come march, the ice will not be 1.5 meters thick and it will not be very cold."

To me, this just comes across as I know there is a negative feedback, but I just don't want to believe it will be that strong. My reaction is: Why not? Where is your evidence?


Lets assume that, like Tiesche suggests, most of the heat has been vented out of the oceans by the end of November. Lets assume incredibly fast extent growth that covers the whole 8.4 *10^12 m2 Let's assume that the FDD's after growth start follow a path similar to 2016, yielding 3000 FDD's, and about 1.3 m ice thickness, then the melting season following this BOE starts with around 10.9 (1000km3).

The average melt for PIOMAS Arctic basin since 2010 is 11.4 (1000km3). 11.4-10.9 = -.6, thus the year following a BOE using the closest analogue year yields another likely BOE.



Quote
Tiesche et al does not support the claim that "there is no indication of hysteresis behavior of Arctic sea ice" with enough confidence for decision makers to ignore the possibility of hysteresis.

I stand by the prior statement. The uncertainties, observation disparities and errors make this paper ill suited as a justification to assume there will be no hysteresis.
« Last Edit: July 30, 2019, 04:47:54 AM by Archimid »
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

crandles

  • Nilas ice
  • Posts: 2391
    • View Profile
  • Liked: 52
  • Likes Given: 42
>"If there was no ice it would immediately warm. Thus there is a heat sink in this model."

Maybe there isn't a heat sink in the model and if ice is removed on 1 June, the ocean does start to warm. We don't know if that was tried.

If ice is removed on 1 May, maybe ice does form:

Firstly PIOMAS reaches maximum volume in mid April but it is perfectly possible that peripheral ice is melting but towards the centre the ice is still thickening until mid May. I believe I have seen this happening when I looked closely at the PIOMAS data.

Secondly, the atmosphere and ocean is in a state consistent with ice cover. Upwelling heat is getting delayed by ice cover keeping water near freezing point. Remove that ice cover and the heat can be lost rapidly and evaporation transfers latent heat to the atmosphere meaning the surface can freeze over.

Maybe the Authors just tried 4 or 6 dates to start from, saw the above effects starting to occur for 1 April or 1 May and decided to use 1 July for their main experiment.

.

Quote
Lets assume that, like Tiesche suggests, most of the heat has been vented out of the oceans by the end of November. Lets assume incredibly fast extent growth that covers the whole 8.4 *10^12 m2 Let's assume that the FDD's after growth start follow a path similar to 2016, yielding 3000 FDD's, and about 1.3 m ice thickness, then the melting season following this BOE starts with around 10.9 (1000km3).

The average melt for PIOMAS Arctic basin since 2010 is 11.4 (1000km3). 11.4-10.9 = -.6, thus the year following a BOE using the closest analogue year yields another likely BOE.

I am not sure what adjustments you are doing to get 1.3m thick. Red line is at about 800FDD on 1 Dec. Remove 800 from total leaves the thickness at about 1.5m thick.

« Last Edit: July 29, 2019, 10:15:40 PM by crandles »

Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161
Quote
Maybe there isn't a heat sink in the model and if ice is removed on 1 June, the ocean does start to warm. We don't know if that was tried.

It was tried.

Quote
The start date is chosen such that the effect of the perturbation is maximal: starting from ice‐free conditions earlier in the year leads to immediate re‐freezing, and hence both earlier and later start dates imply shorter exposure of open water to sunlight, and a less pronounced ice–albedo effect.

They did remove the ice earlier in the year and it lead to immediate refreeze, then without analysis or justification they just assume "that the effect of the perturbation is maximal" and earlier dates ends up in lesser exposure to sunlight.

Without ice on the Arctic in June the sun should quickly bring the top most layer above freezing point, but it doesn't.

What is this source of heat sinks that keeps the arctic below freezing during peak insolation but not 10 days after peak insolation?

Quote
I am not sure what adjustments you are doing to get 1.3m thick. Red line is at about 800FDD on 1 Dec. Remove 800 from total leaves the thickness at about 1.5m thick.

yeah, we are talking about the inner basin, not just N80. I thought 3000 FDD's was a good estimation.  We haven't even talk clouds or the extremely fast increasing winter temperatures of the Arctic.
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

crandles

  • Nilas ice
  • Posts: 2391
    • View Profile
  • Liked: 52
  • Likes Given: 42
Quote
Maybe there isn't a heat sink in the model and if ice is removed on 1 June, the ocean does start to warm. We don't know if that was tried.

It was tried.


We don't know how much was tried.

I very much doubt they started circa 360 models (yes many models use 12 months of 30 days for a 360 day year), one for each day of the year. Models write daily, monthly and quarterly files as they go along. Consequently easier to start model on 1st of the month. So starting 12, on 1st of each month might be a sensible approach and considered sufficiently comprehensive. However, if they didn't have the mainframe computer time necessary to do that, I could easily believe they might have done just 4 or 6 models. Or maybe they did have the computer time to do 12, but they are fiddly things to set up and quit easy to mess things up and by the time you realise something is wrong, the available computing time is rapidly disappearing. Or.....

Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161
Whatever the reason was, it was not ice albedo feedback as implied by the authors. A good explanation is required if claims like "maximal perturbation" are to be used. I bet they think the Arctic has an intrinsic heat sink other than the ice.

But that really is not the biggest problem. Winters are getting warmer much faster than summers and the NH hot air is intruding into the arctic more frequently and in a spectacular fashion. This model is counting on decreased heat and humidity from the hemisphere as a principal recovery mechanism but the exact opposite is happening.
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

crandles

  • Nilas ice
  • Posts: 2391
    • View Profile
  • Liked: 52
  • Likes Given: 42
Whatever the reason was, it was not ice albedo feedback as implied by the authors. A good explanation is required if claims like "maximal perturbation" are to be used. I bet they think the Arctic has an intrinsic heat sink other than the ice.

Seems a strange thing to be confident about to me, unless it is something real like:

But that really is not the biggest problem. Winters are getting warmer much faster than summers and the NH hot air is intruding into the arctic more frequently and in a spectacular fashion. This model is counting on decreased heat and humidity from the hemisphere as a principal recovery mechanism but the exact opposite is happening.

This perturbation is not a real world experiment. Therefore, it is just fine for heat to be being imported into the Arctic in the real world. Also for this heat import to reduce in a non physical perturbation that makes the Arctic much warmer: If the Arctic is much warmer, winds into arctic are still at least as warm, but the winds going out will be much warmer. The heat transfer depends on the temperature difference. How does this heat transfer compare with the normal situation? A: There is an effective net heat sink relative to the normal situation. I would be worried about the model if it wasn't doing this.

There being such a major problem with the model as you are suggesting would almost certainly be noticed and corrected.

Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161
Holland et al., 2006

https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/2006GL028024

Conclusion:

Quote
An analysis of additional climate models and future forcing scenarios indicates that abrupt transitions in the Arctic summer ice cover are not only present in the CCSM3 model but occur in numerous other projections of the future Arctic sea ice. Reductions in future greenhouse gas missions reduce the likelihood and severity of such events. A recent study [Winton, 2006] also indicates that under higher emissions scenarios some climate models exhibit abrupt transitions to completely ice-free conditions as first year ice is also lost. Abrupt transitions such as those exhibited by climate models would undoubtedly further strain adaptation of ecosystems and native peoples to climate change.

I have no clue how this paper supports the claim that  "there is no indication of hysteresis behavior of Arctic sea ice". This paper supports the exact opposite. Abrupt arctic collapse could leave us without ice by 2040. Yet that fact is completely disregarded  and the paper use as justification of "no hysteresis"
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161
Armour et al., 2011

https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2010GL045850

This paper does not even mention the ice. Yet it is used for justification for "no hysteresis".
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

crandles

  • Nilas ice
  • Posts: 2391
    • View Profile
  • Liked: 52
  • Likes Given: 42
Holland et al., 2006

https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/2006GL028024

Conclusion:

Quote
An analysis of additional climate models and future forcing scenarios indicates that abrupt transitions in the Arctic summer ice cover are not only present in the CCSM3 model but occur in numerous other projections of the future Arctic sea ice. Reductions in future greenhouse gas missions reduce the likelihood and severity of such events. A recent study [Winton, 2006] also indicates that under higher emissions scenarios some climate models exhibit abrupt transitions to completely ice-free conditions as first year ice is also lost. Abrupt transitions such as those exhibited by climate models would undoubtedly further strain adaptation of ecosystems and native peoples to climate change.

I have no clue how this paper supports the claim that  "there is no indication of hysteresis behavior of Arctic sea ice". This paper supports the exact opposite. Abrupt arctic collapse could leave us without ice by 2040. Yet that fact is completely disregarded  and the paper use as justification of "no hysteresis"

The paper is full of bits like:

Quote
[10] An analysis separating the contributions to the ice extent change from thermodynamics and dynamics indicates that the abrupt change is thermodynamically driven, with ice dynamic effects (i.e. transport or ridging) playing little direct role.

[13] Changes in ocean heat transport to the Arctic also play an important role in increasing the net melt rate. Over the 20th and 21st centuries, this heat transport exhibits a gradual upward trend overlaid by periods of rapid increase (Figure 3a). These rapid ‘‘pulse-like’’ events lead changes in the sea ice by 1 –2 years, which is evident from the timeseries of detrended heat transport and detrended ice thickness (Figure 3b). For Run 1, a rapid increase in heat
transport starts around year 2020, modifies the ice growth/ melt rates, and triggers positive feedbacks that then accelerate the ice retreat.

These transitions are associated with an increased open water formation efficiency for a given melt rate as the ice thins. The surface albedo feedback accelerates the ice retreat as more solar radiation is absorbed in the surface ocean, increasing ice melt. Additionally, rapid increases in ocean heat transport to the Arctic generally lead and possibly trigger the events.

They are saying the sudden changes are suspected to be forced by ocean heat transport and not the result of hyteresis.

Also note


There is more divergence than typical around 2020 between the 7 model runs. Hysteresis would be indicated by the models settling onto different tracks.

I agree that abrupt arctic collapse could leave us without ice much sooner than the trends and expectations suggest and this is concerning. But this isn't hysteresis if it is a forced change.


crandles

  • Nilas ice
  • Posts: 2391
    • View Profile
  • Liked: 52
  • Likes Given: 42
Armour et al., 2011

https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2010GL045850

This paper does not even mention the ice. Yet it is used for justification for "no hysteresis".

I think you have the wrong paper. It is referenced in Chapter 3 of the draft report as

Armour, K.C., I. Eisenman, E. Blanchard-Wrigglesworth, K.E. McCusker, and C.M.
Bitz, 2011: The reversibility of sea ice loss in a state-of-the-art climate model.
Geophysical Research Letters, 38(16), L16705,
doi:10.1029/2011gl048739.

so try

https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2011GL048739

Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161
Quote
They are saying the sudden changes are suspected to be forced by ocean heat transport and not the result of hyteresis.

Holland says:

Quote
[13] Changes in ocean heat transport to the Arctic also play an important role in increasing the net melt rate. Over the 20th and 21st centuries, this heat transport exhibits a gradual upward trend overlaid by periods of rapid increase (Figure 3a). These rapid ‘‘pulse-like’’ events lead changes in the sea ice by 1 –2 years, which is evident from the timeseries of detrended heat transport and detrended ice thickness (Figure 3b). For Run 1, a rapid increase in heat
transport starts around year 2020, modifies the ice growth/ melt rates, and triggers positive feedbacks that then accelerate the ice retreat
.

 Tiesche says:

Quote
As shown by Serreze et al. [2007], heat transport into the Arctic Ocean by advection of warm water and export of sea ice is only between 4 and 7 Wm−2 (March and August mean, respectively). Our model shows comparable results for oceanic heat transport into the Arctic. When we compare the reference run to the perturbed run, no significant changes of oceanic heat transport are visible.

The IPCC has no good justification for their no hysteresis claims.



Quote
Armour et al., 2011

Thanks, I'll take a look
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161
Sedláček et al., 2011

https://journals.ametsoc.org/doi/full/10.1175/2011JCLI3904.1


Before I post a more elaborate review of this paper, let me say that on first impression this is a great model. When I read their description of the climate changes as an effect of their perturbation experiments is like they were telling the story of the Arctic over the last few years.

Their magical perturbation produced a disturbance that matches very well what has been happening in the Arctic. That, I like. That they did this in 2011 is a testament to the greatness of science.

That said, this paper does not at all support the conclusion of the IPCC that "there is no indication of hysteresis behavior of Arctic sea ice".  I'll try to make my case soon.
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

Richard Rathbone

  • Frazil ice
  • Posts: 496
    • View Profile
  • Liked: 19
  • Likes Given: 7
The abstract says,

"The sea ice area and thickness recover in both ensembles after 3 and 5 yr, respectively. "

For evidence of hysteresis the abstract would need to say something like

"The sea ice area and thickness settle at 50% after 3 years and 70% after 5 years respectively, and fail to recover further in the remaining 25 years of the model run."



Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161
For evidence of hysteresis of an ice free arctic, they should have tested on an ice free Arctic. They didn't.

The equilibrium run yielded a September minimum of  4.4 × 10^6 km2. When they magically increased albedo they got the ice down to 2.2 × 10^6 km2


In a world with increased CO2 their control run yielded a September minimum of  4.0 × 10^6 km2, when they magically increased albedo they got the ice down to 1.9 × 10^6 km2

They never tested an ice free Arctic.

Basically, the model predicted 2016. What we are seeing after 2016 is eerily similar to what this model predicted after a magical disturbance. In the case of 2016 the magical disturbance wasn't magic at all, but a very real confluence of factors over the Pacific, Atlantic and indeed the whole world that resulted in circumstances similar to what this paper magically created.

A true testament to good science.

However, a conclusion about hysteresis can be reached.

Moderate sudden reductions of Arctic sea ice can have planetary effects lasting years and a short term inter annual memory of the Arctic sea ice  appears.

That both the equilibrium and transient runs show a fading 3-5 years memory for a mild disruption of the ice is a serious indication of hysteresis.
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

Richard Rathbone

  • Frazil ice
  • Posts: 496
    • View Profile
  • Liked: 19
  • Likes Given: 7


That both the equilibrium and transient runs show a fading 3-5 years memory for a mild disruption of the ice is a serious indication of hysteresis.

You have this completely the wrong way round. That it fades is serious evidence of no hysteresis.

Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161
Hysteresis is the dependence of the state of a system on its history. Memory.

The arctic cycle is one year. After the initial perturbation on year 0, these models retain the memory of the event for 3-5 years even without additional perturbation. That's memory.  That's the hallmark of hysteresis.  No evidence of hysteresis would be a complete recovery the year following the initial perturbation, specially considering this is a model with a magical perturbation and an idealized, parametric, unchanging world that will pull the ice back.

That this model does not remove the ice, it merely reduces it to an area similar to 2012/2016, but still shows 3-5 years of memory is extremely concerning, specially in a rapidly warming world. If in that 3-5 year memory another perturbation happens, then the effects of the last perturbation accumulate with the new perturbation, further extending the impact on the ice.
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

petm

  • Grease ice
  • Posts: 580
    • View Profile
  • Liked: 254
  • Likes Given: 21
Perhaps RR is referring to rate-independent hysteresis and Archimid is referring to rate-dependent hysteresis?

https://en.wikipedia.org/wiki/Hysteresis#Types

Quote
Rate-independent: ... does not fade as the events recede into the past.

Quote
Rate-dependent: ... continues to respond for a finite time.
« Last Edit: August 02, 2019, 04:21:52 AM by petm »

Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161
 The fact that the Arctic is a decreasing system makes rate dependent hysteresis extremely relevant. The more perturbations there are the faster it will decrease. We can expect perturbations to increase as the climate changes.
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

DrTskoul

  • Nilas ice
  • Posts: 1402
    • View Profile
  • Liked: 175
  • Likes Given: 59
Why hysteresis?? Does the tank of a toilet display hysteresis ? It empties in one to 5 seconds and it might take a minute to fill. It is a linear mass balance energy balance system. What they did in the paper was to add a ton of energy in the system (remove the ice)  that took 3-5 years to lose.

No hysteresis required to explain that. Just a volume ( retained and stored amount ) and slow enough dynamics.

Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161
Quote
What they did in the paper was to add a ton of energy in the system (remove the ice)  that took 3-5 years to lose.

Just to be clear, Sedláček et al., 2011 didn't remove the ice completely, he merely added heat by lowering albedo values for 1 year. That left around 2.2 (insert unit) in the arctic by september minimum. After that the memory of the event remains for 3 years in area and 5 in thickness ( and the planet at large).

A common misconception around here is that the Arctic "vents the heat out to space" every winter and the next summer there is virtually no evidence of  the extra heat. That is simply false. There is memory. Extra heat outlives several winters, even without fully losing the ice cover.

This happens with or without CO2, so it is intrinsic to the Arctic, not the climate system.

If the perturbations weren't magical, but a series of ongoing events, then they would accumulate.

The IPCC claims that  "there is no indication of hysteresis behavior of Arctic sea ice". That is not true.  There are indications of hysteresis on the very models they quote as evidence for no hysteresis.

I'm not done reading all of the IPCC's sources for this claim.

I wish  Sedláček et al., 2011 would've performed a simulation where they removed all of the ice. I bet they did and the result was so ridiculous that they decided to do a partial albedo reduction instead.
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

DrTskoul

  • Nilas ice
  • Posts: 1402
    • View Profile
  • Liked: 175
  • Likes Given: 59
Quote
What they did in the paper was to add a ton of energy in the system (remove the ice)  that took 3-5 years to lose.

Just to be clear, Sedláček et al., 2011 didn't remove the ice completely, he merely added heat by lowering albedo values for 1 year. That left around 2.2 (insert unit) in the arctic by september minimum. After that the memory of the event remains for 3 years in area and 5 in thickness ( and the planet at large).

A common misconception around here is that the Arctic "vents the heat out to space" every winter and the next summer there is virtually no evidence of  the extra heat. That is simply false. There is memory. Extra heat outlives several winters, even without fully losing the ice cover.

This happens with or without CO2, so it is intrinsic to the Arctic, not the climate system.

If the perturbations weren't magical, but a series of ongoing events, then they would accumulate.

The IPCC claims that  "there is no indication of hysteresis behavior of Arctic sea ice". That is not true.  There are indications of hysteresis on the very models they quote as evidence for no hysteresis.

I'm not done reading all of the IPCC's sources for this claim.

I wish  Sedláček et al., 2011 would've performed a simulation where they removed all of the ice. I bet they did and the result was so ridiculous that they decided to do a partial albedo reduction instead.

Earth is not a closed system . Heat escapes from the outer layers of the atmosphere. And there is no "memory" just multiple reservoir with heat capacity that are interlinked with slow and fast dynamics.

In the same way if I try to flush the toilet before it is full again, you will claim the toilet has memory of the last flush....
« Last Edit: August 02, 2019, 01:40:37 PM by DrTskoul »

crandles

  • Nilas ice
  • Posts: 2391
    • View Profile
  • Liked: 52
  • Likes Given: 42
Quote
The perturbations are 50% larger than those used by Bitz et al. (2006) to create an ice reduction corresponding to that of doubling the CO2 concentration.


It is still a big perturbation even if it didn't remove all the ice. Also note it did affect water temperatures down to ~ 200m.

Top 50m is mixed layer and that can be stripped of excess heat in a winter. 50-200m is going to retain some of the heat and release parts of it over next few years. So of course there is a remaining effect of the perturbation over the next 3-5 years. But it disappears with the model settling down at the unperturbed level.

interstitial

  • Frazil ice
  • Posts: 142
    • View Profile
  • Liked: 24
  • Likes Given: 56
To me the fact that the ice in the model fully recovers is an indication the model is faulty. The arctic is not in equilibrium or even a stabile system. If we stop polluting today it will be many years before the ice stabilizes at a new normal. More ice only melts this year than last year when the system is out of balance. If the amount of imbalance is small it takes a very long time to notice the difference. We are changing the system quicker than we have records of it changing in the past.  A properly functioning model would restablish itself at some lower value. while I don't know the exact magnitutde of that change I expect to see it.

Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161
Quote
Earth is not a closed system. Heat escapes from the outer layers of the atmosphere.

We can agree on that. I don't understand how this is relevant.

Quote
And there is no "memory" just multiple reservoir with heat capacity that are interlinked with slow and fast dynamics.

There is memory. Perturbations have effects that can last many cycles so the cycles are not independent of each other. There is no hard reset every March. Memory exist in the water column, the amount of ice, the temperature of the ice and global climatic disturbances.

In a system in equilibrium this wouldn't be a problem. In a system under increasing forcing
this is a huge problem.

That the memory exists in both control and experiment runs makes the indication of hysteresis even stronger.

Quote
In the same way if I try to flush the toilet before it is full again, you will claim the toilet has memory of the last flush....

The toilet shows hysteresis if the next flush attempt is tried before the minimum volume needed to complete a flush is in the reserve (thick ice). During that time the system will remember the past and act accordingly, namely a floating spinning turd.  However outside that time frame, the system shows no hysteresis.

Interestingly, hysteresis can be hidden with a big enough tank ( thick ice). So you can flush to your heart's content and no memory will be shown by the system ( ignoring a stuck turd). That is until that one flush that depleted the reserves and the memory of the system shows up.
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161
It is still a big perturbation even if it didn't remove all the ice. Also note it did affect water temperatures down to ~ 200m.

It was a big perturbation. The way I understand it they created something similar to meltponds everywhere there was snow. That's a big deal. It wasn't enough to get the Arctic to a BOE  or an ice free state but it was enough to reveal 3 years of memory in area and 5 years of memory in thickness.

Quote
Top 50m is mixed layer and that can be stripped of excess heat in a winter
.

Heat lost as radiation must first pass through the atmosphere before it can leave for space. The more humid the atmosphere the warmer the atmosphere gets. The warmer the atmosphere is  the least ice is created.

Then there is evaporation...

It is going to be warmer above the ice restricting ice creation.


Quote
50-200m is going to retain some of the heat and release parts of it over next few years. So of course there is a remaining effect of the perturbation over the next 3-5 years. But it disappears with the model settling down at the unperturbed level.
[/quote]

Yep it will also be warmer than normal below the ice too. Even after most of the heat of the top layer is vented out to the atmosphere.
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

crandles

  • Nilas ice
  • Posts: 2391
    • View Profile
  • Liked: 52
  • Likes Given: 42

Quote
Top 50m is mixed layer and that can be stripped of excess heat in a winter
.

Heat lost as radiation must first pass through the atmosphere before it can leave for space. The more humid the atmosphere the warmer the atmosphere gets. The warmer the atmosphere is  the least ice is created.

It is going to be warmer above the ice restricting ice creation.

Certainly atmosphere gets warmer and this limits ice production, but this is the way warmer air temperatures radiates more heat to space.

So, the question is the timescale to lose the heat. The models are telling us that this is pretty fast. Even where ocean warms quite a bit, the mixed layer loses its heat and ice is being formed by December if not November. That allows most of the ice volume to form by the maximum and the next year is not as severe as the perturbation.

Quote
Then there is evaporation...

Yes, that transfers lots of energy from ocean to atmosphere via latent heat

The air doesn't hold limitless quantities of water. Clouds/fog/ warm moist air will create situations for convection which carry heat higher into atmosphere where the heat is more easily lost to space. Rain leaves the latent heat away from the surface.

So I don't see how this helps your cause when the issue is how fast the heat is lost from Ocean to space. This is a pretty fast method for ocean to lose heat.


Quote
It wasn't enough to get the Arctic to a BOE  or an ice free state but it was enough to reveal 3 years of memory in area and 5 years of memory in thickness.

So what? If the water below 50m is affected by the perturbation, I would expect some residual effects leaking out in declining fashion over the next few years. It didn't result in the model settling down to a different state.

The perturbation isn't meant to be a real effect, it is just to see if the history of what has happened affects where the model settles down. The answer for perturbations of up to 3 months ice free appears to be no. Maybe in 50+ years time we will need to do more testing of perturbations of 4, 5 or 6 months ice free, but for at least the next couple of decades the answer looks to be no hysteresis imminent. 



Richard Rathbone

  • Frazil ice
  • Posts: 496
    • View Profile
  • Liked: 19
  • Likes Given: 7


That the memory exists in both control and experiment runs makes the indication of hysteresis even stronger.


You have this completely wrong. There is no memory in the runs. There is a response, but its transient, it comes back to where it started. If there was memory it would come back to a different state.


DrTskoul

  • Nilas ice
  • Posts: 1402
    • View Profile
  • Liked: 175
  • Likes Given: 59
The article is talking about tipping points ( "hysteretic threshold behavior" ).

The oceanic + atmospheric system has no "memory". It might appear so but it does not. The equations governing the system are very simple - heat and mass transfer, momentum transfer, phase change, and there are several  heat reservoirs that are connected by those transfer mechanism ( deep ocean, 12m and 50 m, middle atmosphere and top atmosphere vertically ) and several heat inlets and outlets ( top of the atmosphere and ocean/atmosphere surrounding the arctic ocean). The dynamics of the reservoirs depend on the volume, the specific heat capacity, the salinity for the ocean and the temperature for the air.

Arctic ice is artificially removed on July 1st by converting it to water ( massive injection of energy as latent heat ). When the ice is removed it leads to the stratification of the top layer of the ocean that can break dow during the winter cooling. The latent heat takes several years to dissipate and the system returns to equilibrium. It tak
es more in the 80s because the extent is much larger.

Clouds are not modeled explicitly but that will lead to "richer" dynamics, there is no "iris" like effect.

Paper reasoning based on the solved model seems reasonable to me. The perturbations were on top of a hypothetical pathway to ice free arctic in the 2070s.

DrTskoul

  • Nilas ice
  • Posts: 1402
    • View Profile
  • Liked: 175
  • Likes Given: 59
The graph below is from Gerontocrat. It shows average annual extent. You will see that the "recovery" after 2007 and 2012 lasted about 2 years an indication of the dynamics being in the order of few years. So it seems is the case after 2016. The lag between minimum extent and minimum in graph is due to the averaging.


Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161

Certainly atmosphere gets warmer and this limits ice production, but this is the way warmer air temperatures radiates more heat to space.


Whatever that speed is, the fact remains the atmosphere will be warmer particularly the surface, this means less ice formation than before.

Quote
So, the question is the timescale to lose the heat. The models are telling us that this is pretty fast. Even where ocean warms quite a bit, the mixed layer loses its heat and ice is being formed by December if not November
.

From November to March, starting from virtually 0 ice, a warmer atmosphere, a warmer ocean, who knows what will happen with atmospheric and oceanic circulations. You say 1.5m of ice I say 1.3m of thin warm first year ice.

I say that in a world that is warm enough to actually cause the perturbation magically created in the models the chance that the following year there won't be a BOE are almost 0.


Quote
That allows most of the ice volume to form by the maximum and the next year is not as severe as the perturbation.

That is an extraordinary claim. In none of the models discussed so far the maximum could have been remotely close to normal. There was a memory of at least three years after the events, that means the maximum took several cycles to recover. "most of the ice volume" does not form by maximum. That's memory. That's most certainly indication of hysteresis. In a system that's being increasingly forced, this type of memory is lethal.

Quote
Yes, that transfers lots of energy from ocean to atmosphere via latent heat

Yes, lots. And an open ocean transfers a lot more than a cold dry Arctic desert.

Quote
The air doesn't hold limitless quantities of water.

No it doesn't. It has a minimum and maximum humidity. The maximum is irrelevant for our discussion, since it will never get that hot. What matter is the difference between 20th century humidity, today's humidity and humidity after a BOE. It will be warmer and more humid. Much more.

Quote
Clouds/fog/ warm moist air will create situations for convection which carry heat higher into atmosphere where the heat is more easily lost to space.

Sure, among other positive feedback cycles. But I doubt that during the winter night the sum of all positive and negative will overcome the sheer amount of heat that we've put into the system. There is no magic bullet that will make the heat "drain" faster, except being warmer. The system can only be obstructed.

Quote
Rain leaves the latent heat away from the surface.

Can you explain this to me? I have trouble imagining how liquid fresh water falling on salty cold ice helps ice formation. In my mind rains convects heat into the ice. I've seen what hard enough rains can do to the ice pack over at the melting season.

Quote
So I don't see how this helps your cause when the issue is how fast the heat is lost from Ocean to space. This is a pretty fast method for ocean to lose heat.

The issue is not how fast heat is lost from ocean to space, the issue is how much of the EXTRA heat hangs around longer and delays ice formation.

And we are once again ignoring convection from the hemispheres, which was disregarded by Tiesche but Sezlacek found to be an effect of the perturbation.


Quote
So what? If the water below 50m is affected by the perturbation, I would expect some residual effects leaking out in declining fashion over the next few years. It didn't result in the model settling down to a different state.

All the experiments discussed so far result on the models remaining in a different state for up to 8 years.

Quote
The perturbation isn't meant to be a real effect, it is just to see if the history of what has happened affects where the model settles down.

All the experiments discussed so far result on the models remaining in a different state for up to 8 years.

Quote
The answer for perturbations of up to 3 months ice free appears to be no

The answer for perturbations that cause a minimum area of 2 million km2 is yes, for 3 years in extent and 5 in thickness, in a world where the climate is cold and unchanging enough to produce an ice free arctic by 2070.

In a model that produced an ice free arctic by 2050 the memory would be even larger.

Quote
Maybe in 50+ years time we will need to do more testing of perturbations of 4, 5 or 6 months ice free, but for at least the next couple of decades the answer looks to be no hysteresis imminent.

We have the models today, they are failing today. If they are wrong about this we don't have 50 years. The IPCC conclusion is wrong about this. There are serious indications of hysteresis.
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

DrTskoul

  • Nilas ice
  • Posts: 1402
    • View Profile
  • Liked: 175
  • Likes Given: 59
 Btw hysteresis, really literally means delay! You both refer to a tipping point , a non linear  bifurcation.

petm

  • Grease ice
  • Posts: 580
    • View Profile
  • Liked: 254
  • Likes Given: 21
Maybe in 50+ years time we will need to do more testing of perturbations of 4, 5 or 6 months ice free, but for at least the next couple of decades the answer looks to be no hysteresis imminent.

I think I catch your drift, but I hardly think that today's models will be deemed sufficient in even 1 decade let alone 5...

Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161
Quote
You have this completely wrong. There is no memory in the runs. There is a response, but its transient, it comes back to where it started. If there was memory it would come back to a different state.

All the experiments discussed so far result on the models remaining in a different state for up to 8 years. Even when  Sedlacek didn't remove all the ice, Tiesche has a weird heat sink and  Schroeder and Connolley happenned with pre-industrial CO2, the Arctic retained memory.

It is likely a tipping point doesn't show in the models because the favorable climate in climate models that expect the ice gone by 2070 bring the ice back.

You are skipping information. The fact that the arctic retains a memory of up to 8 year in these models is an indication of that if the forcing continues the melt will accelerate.
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

Archimid

  • Nilas ice
  • Posts: 2171
    • View Profile
  • Liked: 196
  • Likes Given: 161
The article is talking about tipping points ( "hysteretic threshold behavior" ).

To be clear, I'm arguing against this quote from the IPCC excerpt that crandles posted:

"there is no indication of hysteresis behavior of Arctic sea ice"

There is. Strong one and uniform across three of the references the IPCC gave for no indication of hysteresis. If the arctic is perturbed it retains a memory for 3 to 8 years in models that predict the ice gone by 2070.

Maybe in more realistic models that predict the ice gone by 2050 or before a tipping point is found that is now hidden and only shows as memory.


Quote
The equations governing the system are very simple - heat and mass transfer, momentum transfer, phase change, and there are several  heat reservoirs that are connected by those transfer mechanism ( deep ocean, 12m and 50 m, middle atmosphere and top atmosphere vertically ) and several heat inlets and outlets ( top of the atmosphere and ocean/atmosphere surrounding the arctic ocean). The dynamics of the reservoirs depend on the volume, the specific heat capacity, the salinity for the ocean and the temperature for the air.

We've all seen the heat advected into the Arctic straight from the pacific. You must add the NH hemisphere to the system.

Quote
Arctic ice is artificially removed on July 1st by converting it to water ( massive injection of energy as latent heat ). When the ice is removed it leads to the stratification of the top layer of the ocean that can break dow during the winter cooling. The latent heat takes several years to dissipate and the system returns to equilibrium. It tak
es more in the 80s because the extent is much larger.

Since the system is so simple, and you seem to be reading Tiesche et al, can you explain why he expects maximum perturbation from albedo on July 1st and not June 21? He claimed the ice returned when removed earlier than July 1st. What  is this heat sink that makes the ice grow in June, during maximum insolation?

Quote
Clouds are not modeled explicitly but that will lead to "richer" dynamics, there is no "iris" like effect.

Clouds are the name of the game.

Quote
Paper reasoning based on the solved model seems reasonable to me. The perturbations were on top of a hypothetical pathway to ice free arctic in the 2070s.

Do you believe the first September without ice will happen after 2070?
I am an energy reservoir seemingly intent on lowering entropy for self preservation.