Snow and Ice Northern Hemisphere - Winter 23 /24

[Midlands Ice Age]

8 hours ago, WYorksWeather said:

Thanks @Midlands Ice Age - this is some interesting work. One other factor I thing may be worth considering is the slowdown in the AMOC. Interestingly, I did find a few sources that seemed to suggest that an AMOC slowdown would delay an ice-free Arctic.

 

This article below is a fairly accessible explanation as to why this might be the case. Essentially, there is somewhat of a negative feedback between the AMOC and sea ice. A weaker AMOC helps to preserve sea ice in several important regions.

 

Cryospheric Sciences | For Dummies: How Arctic sea ice and the AMOC interact

BLOGS.EGU.EU

---

In this post, I will talk about two famous characters of the climate system; I will define them and see how they have changed in the current context of climate change. I will also show you how these two characters interact...

 

Thanks WYW...

Yes I am totally aware of the AMOC.

 The thing is that its impact is mainly in the North Atlantic., not the North Pacific where this odd effect of early freeze up  is being currently mainly observed.

Looking at the seas involved,,,

The Baltic  is isolated from the Atlantic by about 1000 miles round the Skagerrak Sea, and we all are told that SST's are very high not low there...

The Bering , SOO, and Cooke Inlet (and probably the Yellow Sea when it freezes) are at the top of the north Pacific. -  About 5000 miles from the North Atlantic.

 

As for latest list of oceans I produced, showing these early freezes, then  they occured in the middle of the 2010's. 

Again the Oceans involved are not those in the middle of the Arctic where the AMOC has its major effects.

Hudson, the CAA and Baffin are 'hidden' away at the top of the Baffin Sea and are isolated by the Labrador Straits from the AMOC.   

They could not be further away from the AMOC.  

 A small sub branch of it does indeed move up the Baffin  Ocean, but it very rarely ever freezes(!) and when it does it is always last in the season to freeze , so any effect cannot be relevant and is very minimal.

I have included a very basic map of the Arctic Oceans, so that you can see the problem with trying to pin it on any AMOC effect. If we were talking about Greenland, Barents or Kara Oceans then I could see it having some effects. But no impact has been seen there.

If the AMOC had been based in the Pacific then it is fairly likely to be at least a part of the answer...  But...

So....     what did cause these effects that I have reported?

.                   Arctic Oceans and Masie  

MIA

Edited November 10, 2023 by Midlands Ice Age

[WYorksWeather]

4 minutes ago, Midlands Ice Age said:

Thanks WYW...

Yes I am totally aware of the AMOC.

 The thing is that its impact is mainly in the North Atlantic., not the North Pacific where the odd effect of early freeze up  is being mainly observed.

The Baltic  is isolated from the Atlantic by about 1000 miles round the Skagerrack Sea, and we all are told that SST's are very high not low there...

The Bering , SOO, and Cooke Inlet (and probably the Yellow Sea) are at the top of the north Pacific. -  About 5000 miles from the North Atlantic.

 

As for latest list of oceans showing these early freezes, they occured in the middle of the 2010's. 

Again the Oceans involved are not those in the middle of the Arctic where the AMOC has its major effects.

Hudson, the CAA and Baffin are 'hidden' away at the top of the Baffin Sea and are isolated by the Labrador Straits from the AMOC.     A small sub branch of it does indeed move up the Baffin  Ocean, but it very rarely ever freezes(!) , so any effect cannot be relevant and is very minimal.

I have included a very basic map of the Arctic Oceans, so that you can see the problem with trying to pin it on any AMOC effect. If we were talking about Greenland, Barents or Kara Oceans then I could see it having some effects. But no impact has been seen there.

So....     what did cause these effects that I have reported?

.                   Arctic Oceans and Masie  

MIA

Hm. It's an interesting one this in that case - I agree that based on what you've said any AMOC effect is likely to be modest. I would of course counter with the fact that the recent AMOC slowdown will have wide-ranging effects that are perhaps not entirely known - maybe there is some sort of teleconnection influence that produces an indirect effect? Since the slowdown is fairly recent I doubt the full effects are properly understood even by people far more qualified than you or I.

Without a clear cause, the next idea would be to look at internal variability. We essentially would need to quantify whether the recent changes represent a statistically significant change from the long-term trend which is then worthy of further investigation. A lot depends on how long any stabilisation or increase lasts for. Unfortunately, the Masie data only starts in 2006, so the power of any statistical tests would be limited. Therefore, I'm not sure it's worth the bother, as we'd probably just get 'no significant change'.

 

The issue with predicting ice extent is it's a tipping process, with a threshold - melting above 0C, freezing below. Once we reach a certain level of warming the extent might fall rapidly without much prior warning, as the melting season extends. I also think based on memory that 2012 was affected by an unusual storm track which disturbed the ice.

Whereas surface air temperatures are much more linear, and global averaging removes a lot of the internal variation, and that variation that does occur is very well understood.

The ice will probably continue to confound us! Personally, I'd go for a first BOE in 2050, but I'd put very low confidence on that estimate. 

[Midlands Ice Age]

Thanks again.

 Whilst I agree with most of your comments, a few points need to be taken into account,

Talking of statistics...

1) Each year we have seen, exactly the same effect in ((probably) - we are at 4 so far this year)  all 5 of the outer sea areas.   This at a time when most supposed experts are forecasting the ice to be slower to form and soon disappear in these areas. The effect may well NOT  continue..  It also has not resulted in a notable increase of ice in these areas.

2) It has happened in all these areas consistently for 3 years running. That is effectively 14/15 times in a row! Statistically that is fairly high .  >99.5%

3) For the same 3 years we have had La NIna conditions, and only those 3 years,  going back for about 10 years.. 

4) Although there is no known connection between the two systems, the odds cannot be ignored.

I am prepared to venture on the odds alone that one does exist, but we do not understand it yet. It is not a simple connection.

5) I suspect that we will not see the same effect next year when (and if) La Nina loses its power. 

MIA   

Edited November 11, 2023 by Midlands Ice Age

[WYorksWeather]

11 hours ago, Midlands Ice Age said:

Thanks again.

 Whilst I agree with most of your comments, a few points need to be taken into account,

Talking of statistics...

1) Each year we have seen, exactly the same effect in ((probably) - we are at 4 so far this year)  all 5 of the outer sea areas.   This at a time when most supposed experts are forecasting the ice to be slower to form and soon disappear in these areas. The effect may well NOT  continue..  It also has not resulted in a notable increase of ice in these areas.

2) It has happened in all these areas consistently for 3 years running. That is effectively 14/15 times in a row! Statistically that is fairly high .  >99.5%

3) For the same 3 years we have had La NIna conditions, and only those 3 years,  going back for about 10 years.. 

4) Although there is no known connection between the two systems, the odds cannot be ignored.

I am prepared to venture on the odds alone that one does exist, but we do not understand it yet. It is not a simple connection.

5) I suspect that we will not see the same effect next year when (and if) La Nina loses its power. 

MIA   

Actually, I was doing some reading, Turns out there may be a confirmed scientific link after all!

Apparently some simulations were run, and sea ice generally decreases more following an El Nino with a winter peak, and roughly the opposite for La Nina. Once again a bit like the AMOC link, the causation is both ways. Apparently less sea ice may also lead to stronger El Nino events per other studies!

The effect of ENSO on Arctic Sea Ice and Its Prediction

UI.ADSABS.HARVARD.EDU

---

The El Niño-Southern Oscillation (ENSO) has a substantial influence on regional patterns of Arctic sea ice thickness and concentration in simulations, especially in late summer and autumn following a large El Niño/La Niña...

The correlation is with the ENSO state in the preceding winter. So for this year, we're looking at DJF 2022-23, which was La Nina.

Of course, it is only a shift in probability. The best way to quantify this would be to look at the correlations - how much of the variation in sea ice extent can be explained by the previous winter's ENSO state?

Don't think I'll get chance today but if you could let me know which data file(s) you used for your analysis I could probably merge something in with the ENSO data and see if we can quantify it a bit better. Or of course if you'd prefer, you could do the analysis yourself.

[Midlands Ice Age]

Thanks again for the reply..

Been out with Grandkids all day and just got some time.

The current ice situation is not simple.

Let me recap.

I have found that during 2014 -2019 (mainly Neutral or slight El Nino?) the ice refreeze started earlier in the CAA, Baffin and Hudson. (North America). also In some areas the ice remained over the next melt season. There was no extra extent or sea  ice as they are enclosed areas.   Presumably there was extra depth and volume though. 

Volume and depth are not easy to measure, and I do not have access to this data  for the areas (yet). Although i may be able to get something by searching around..

In 2020's  (2021- 2023) it was definitely La NIna . My observations are that the 5 ocean areas (Bering, SOO, Yellow Sea, and Cooke Inlet. plus interestingly Baltic, have all started to refreeze between  2 and 3 weeks earlier than the previous 15 years. There is no evidence that more extent or area was produced, as they are so far outside the Arctic Circle that ice cannot survive once and  as soon as any melting starts.

It is not about the 'amount' of ice,   as it cannot survive.  So it is no use looking for any correlation with any ice quantity. It is simply a date significance change in the start up of freezing outside the Arctic Circle,. It may have happened inside it  also, but it is very difficult to spot there.

The fact that this seems to have happened in different places on 2 occasions now indicates that something is going on which has not been accounted for in any ice models. The models agree that oceans this far south will soon not contain any ice, and yet here it is freezing earlier - not later.  

My data source was the NSIDC 'Masie' dataset which is produced daily by and for the US Navy for their shipping (so it is accurate). They also produce the NOAA dataset, which is used by Climate Change people and which has been smoothed/modified over 5 days somewhat,  for their use.  It could have smoothed out this effect - I haven't checked.

The other dataset containing the extent information is  Jaxa - produced by the Japanese, and which is  which is more generally accepted these days..

I intend looking into this dataset to confirm my findings.

I would be grateful if you could perform the analysis, I am supplying  the address to download the data.  I load it into an Excel spreadsheet, and I have to  use judgement as to when the refreeze gets under way. It is not just a straight forward data analysis.

Remember though that it is these intuitive  dates that are of interest not the extent values.  It is  not a simple mathematical modelling exercise. I converted the dates to a base of 300  (after or before  the 300th day of the year)..

 MASIE-NH Daily Image Viewer | National Snow and Ice Data Center (nsidc.org)                                     https://nsidc.org/data/masie  

MIA

Edited November 11, 2023 by Midlands Ice Age

[WYorksWeather]

Will continue this tomorrow (or rather later today!).

The first step of my analysis is below. I dropped Yellow Sea and Cooke Inlet since their ice totals are low and also way too subject to random variation and/or measurement error - the fluctuations were on the order of 50-100 days or more.

First step was to separate out year and day of year. Then I found the maximum extent for each year, and filtered for days at least 1% of that (fairly arbitrary, can change this criterion to something else easily if you prefer). I then got rid of days below 250 (to eliminate days before refreeze starts) and then filtered for the lowest day number in every year and area which met the criteria. Result was 54 values (3 areas x 18 years), which I've plotted below.

Didn't have chance to link up ENSO or do much checking to see how sensitive this is to that arbitrary selection criteria, but here's the plot:

Edited November 12, 2023 by WYorksWeather

[Midlands Ice Age]

 

@WYorksWeather

WYorksWeather

 

Thanks..

Didn't mean to keep you awake!!

The data looks reasonable as above. But the major effect that is striking to me, is that the further from the Arctic Circle the area is,  the stronger was the effect, so if poss you should include the 2 excluded areas. (that is why I said look at the dates rather than worry about the quantities).

Could you include them by  just taking the first appearance date for the ice?   (very rough I know as I actually looked in detail at when a proper refreeze took hold in these areas).  I did say that it might need some manipulation as it is clearly not one of the major extent drivers at the moment. But never the less it would appear that something is happening as eyeballing the data above there does seem to be a downward trend  (in the 2) towards the more recent times,  but not Bering, (possibly) - which I suspect is much more impacted by the weather patterns, and also being 'in contact' with the Arctic may well not show the same effect as consistently.

 It may be that you could include the 2 lesser areas, by just  looking at the first ice date, and maybe adjust positive or probably negative ? Since it is the trend we are looking at not the absolute values. The ice in these area does tend to come and go a bit , but it is purely down to the weather patterns affecting the temperature more in the short term.

Run the analysis then (on all 5)  and after that try excluding Bering, and see what happens.

Again - thanks for the action!. 

MIA

 

 

Edited November 12, 2023 by Midlands Ice Age

[Midlands Ice Age]

Hospital this pm for some tests...

Before I go, a quick update..

Masie increases have been subdued the last 2 or 3 days with increases of (+83K) , (-9K) and yesterday (+46K).

Viewing the National Ice Centre Coverage Charts for today  it looks to be better. with the Kara extension to Severnya Zemlya, now almost complete as the ice pack is moving to the  south of the islands.

General ice increases elsewhere (especially around Svalbard), though a slight retraction in the CAA, as the ice tries very hard  to get a hold of the mainland of Northern Canada. 

The ice in Beaufort really is creeping into Alaska. It cannot be much more than 10 miles now  along the hole coastline.

Snow remains steady with a renewed push west in Russia starting again.

Back later..

MIA

Edited November 12, 2023 by Midlands Ice Age

[WYorksWeather]

1 hour ago, Midlands Ice Age said:

 

@WYorksWeather

WYorksWeather

 

Thanks..

Didn't mean to keep you awake!!

The data looks reasonable as above. But the major effect that is striking to me, is that the further from the Arctic Circle the area is,  the stronger was the effect, so if poss you should include the 2 excluded areas. (that is why I said look at the dates rather than worry about the quantities).

Could you include them by  just taking the first appearance date for the ice?   (very rough I know as I actually looked in detail at when a proper refreeze took hold in these areas).  I did say that it might need some manipulation as it is clearly not one of the major extent drivers at the moment. But never the less it would appear that something is happening as eyeballing the data above there does seem to be a downward trend  (in the 2) towards the more recent times,  but not Bering, (possibly) - which I suspect is much more impacted by the weather patterns, and also being 'in contact' with the Arctic may well not show the same effect as consistently.

 It may be that you could include the 2 lesser areas, by just  looking at the first ice date, and maybe adjust positive or probably negative ? Since it is the trend we are looking at not the absolute values. The ice in these area does tend to come and go a bit , but it is purely down to the weather patterns affecting the temperature more in the short term.

Run the analysis then (on all 5)  and after that try excluding Bering, and see what happens.

Again - thanks for the action!. 

MIA

 

 

The issue with Yellow Sea and Cooke Inlet is that in some years they don't refreeze at all before the end of the year. I'd therefore have to do some hacking around to define refreeze seasons in order to get a value for every year (in other words, to count e.g. day 20 of 2007 as part of the 2006/7 refreeze) and so I can't plot a trend line as some of the data are missing.

The only other approach would be to pick an arbitrary day early in the refreeze (say, day 300) and compare extent at that point, so years that didn't show up would have an extent of zero, so you could still try to do some sort of trend out of that.

Since neither of those two methods are ideal, I decided to plot a scatter graph instead. This is based on first appearance of any ice at all.

Note that in 2006 and 2007 the Yellow Sea didn't refreeze before the end of the year. And in 2023, it hasn't frozen yet.

The Cooke Inlet didn't refreeze before the end of the year in 2014.

In any case, your idea still shows up quite well even if we ignore those two - here's my previous plot with a trend line:

Unfortunately, there isn't enough data, so when I add the standard errors you can see that all we can really say is no trend.

Let me know if you want me to change the method to what I suggested at this stage - looking at extent at a fixed day of the year would save a lot of effort, and I think it would still give some useful insights.

[blizzard81]

1 hour ago, Midlands Ice Age said:

Hospital this pm for some tests...

Before I go, a quick update..

Masie increases have been subdued the last 2 or 3 days with increases of (+83K) , (-9K) and yesterday (+46K).

Viewing the National Ice Centre Coverage Charts for today  it looks to be better. with the Kara extension to Severnya Zemlya, now almost complete as the ice pack is moving to the  south of the islands.

General ice increases elsewhere (especially around Svalbard), though a slight retraction in the CAA, as the ice tries very hard  to get a hold of the mainland of Northern Canada. 

The ice in Beaufort really is creeping into Alaska. It cannot be much more than 10 miles now  along the hole coastline.

Snow remains steady with a renewed push west in Russia starting again.

Back later..

MIA

Many thanks for your continued updates on this and I hope everything goes well for you today

[Midlands Ice Age]

@WYorksWeather 

Thanks again...Must go now my apptmnt is at 14:30.

One or 2 more very quick thoughts though....

A) 2006 -2007 were very unreliable on Masie.     Drop them if it helps analysis.

B) Already I can see that  a trend to later ice melt in the 2015 -2020 period has also  appeared in this data. Maybe what is happening is we have gone back recently to the trends prior to 2013/2014?

Even that is interesting as this middle period  is the same period as the other sea ice areas that I looked into were freezing earlier. Maybe this effect  is simply the world wide changes which happen as the ENSO goes through its variations????.   

Back later

MIA

Edited November 12, 2023 by Midlands Ice Age

[WYorksWeather]

7 hours ago, Midlands Ice Age said:

@WYorksWeather 

Thanks again...Must go now my apptmnt is at 14:30.

One or 2 more very quick thoughts though....

A) 2006 -2007 were very unreliable on Masie.     Drop them if it helps analysis.

B) Already I can see that  a trend to later ice melt in the 2015 -2020 period has also  appeared in this data. Maybe what is happening is we have gone back recently to the trends prior to 2013/2014?

Even that is interesting as this middle period  is the same period as the other sea ice areas that I looked into were freezing earlier. Maybe this effect  is simply the world wide changes which happen as the ENSO goes through its variations????.   

Back later

MIA

@Midlands Ice Age - hope everything went well for you this afternoon.

Here's an update.

I thought I should try to come up with an objective measure that works for most or all of the regions, and doesn't require as much manual tweaking. Hence, what I thought of was to compare the extent reached by day 300 - in other words a measure of how much 'early freeze' we see. I know you said it might be best to avoid extent measures, but I think this only applied to maximum extent? This has a similar impact to measuring early refreeze, and doesn't require any awkward definitions of what refreeze means. I decided to include all areas in the chart , simply to see what it reveals. Probably need to zoom in to see everything, as it is a fairly complex chart!

Areas with zero area for all years are omitted. I've not plotted the confidence intervals - none of this is going to be statistically significant so we should take it all with a pinch of salt.

With that being said on the methodology, we see increases for Bering and SOO. A few others flat or roughly flat. All the areas with six digit extent numbers at this time of year are showing a declining trend.

I've also attached the same chart for day 260 and day 350 below for comparison. 260 being close to the minimum extent date in most years, and 350 showing what happens deeper into the season. Day 350 also allows for Yellow Sea and Canadian Archipelago to be seen on the chart.

  

The main conclusion I draw is that there is a downward trend overall, but it does appear fairly steady. There are excursions above and below the line, but I don't think they call for either unwarranted alarm or optimism. Much as with global temperatures, where fluctuations are superimposed on a longer-term trend, it is of course easy to cherry-pick in order to say that on a year below the trend we are accelerating towards a September Blue Ocean Event, or that based on a couple of better years the pattern is stable.

 

 

 

 

 

[Midlands Ice Age]

8 hours ago, Midlands Ice Age said:

@WYorksWeather 

Thanks again...Must go now my apptmnt is at 14:30.

One or 2 more very quick thoughts though....

A) 2006 -2007 were very unreliable on Masie.     Drop them if it helps analysis.

B) Already I can see that  a trend to later ice melt in the 2015 -2020 period has also  appeared in this data. Maybe what is happening is we have gone back recently to the trends prior to 2013/2014?

Even that is interesting as this middle period  is the same period as the other sea ice areas that I looked into were freezing earlier. Maybe this effect  is simply the world wide changes which happen as the ENSO goes through its variations????.   

Back later

MIA

A very good day today for Sea ice extent on Masie, with a double century gain of 236K Km2.. 

Increases  were spread throughout the regions and were between 20K and 40K KM2.  Interesting ones were Barents  (more of Svalbard),  Hudson, Greenland and Baffin.

 Chukchi ice also finally decided to move towards the Pacific. (for a winter holiday break?)

 

 

MIA

[Midlands Ice Age]

1 hour ago, WYorksWeather said:

@Midlands Ice Age - hope everything went well for you this afternoon.

Here's an update.

I thought I should try to come up with an objective measure that works for most or all of the regions, and doesn't require as much manual tweaking. Hence, what I thought of was to compare the extent reached by day 300 - in other words a measure of how much 'early freeze' we see. I know you said it might be best to avoid extent measures, but I think this only applied to maximum extent? This has a similar impact to measuring early refreeze, and doesn't require any awkward definitions of what refreeze means. I decided to include all areas in the chart , simply to see what it reveals. Probably need to zoom in to see everything, as it is a fairly complex chart!

Areas with zero area for all years are omitted. I've not plotted the confidence intervals - none of this is going to be statistically significant so we should take it all with a pinch of salt.

With that being said on the methodology, we see increases for Bering and SOO. A few others flat or roughly flat. All the areas with six digit extent numbers at this time of year are showing a declining trend.

I've also attached the same chart for day 260 and day 350 below for comparison. 260 being close to the minimum extent date in most years, and 350 showing what happens deeper into the season. Day 350 also allows for Yellow Sea and Canadian Archipelago to be seen on the chart.

  

The main conclusion I draw is that there is a downward trend overall, but it does appear fairly steady. There are excursions above and below the line, but I don't think they call for either unwarranted alarm or optimism. Much as with global temperatures, where fluctuations are superimposed on a longer-term trend, it is of course easy to cherry-pick in order to say that on a year below the trend we are accelerating towards a September Blue Ocean Event, or that based on a couple of better years the pattern is stable.

 

 

 

 

 

Unfortunately these are the 'normal' type curves produced by the usual extent analysis. 

Your choice of selection criteria  only displays the standard patterns of ice progress -  which I frequently report/ display  on here from the ASIF.

Everybody agrees that ice extent levels are less now than they were  in 2006.

What I am interested in is precisely the internal variation which you seem  to want to remove for the sake of having an 'overall' solution.

The internal variation you referred to is not random. I have detected a 'pattern' of earlier refreeze, (which is hidden by  extent) as it is only visible in the smaller ice areas outside the Arctic  basin  (and not in the  major 7 areas which reside within it) -  and these latter dominate any extent effect outside.. It is apparent only in the 5 smaller areas over the period 2021 - 23 and in the 3 N.A. based ones over the period from 2014 - 2019.

Using criteria to include these loses the variability signal in the outer areas. 

Extent is useless to evaluate in these  areas (as I described above) since the ice cannot last more than a short while whenever a slight rise in temps occurs. By its nature it is temporary , but nevertheless this earlier start to refreeze is occurring in these specific areas well away from the central Arctic. It is why this is happening that should interest most scientists as it has not been predicted by the models.

I notice that you earlier referred to the date 300, but then did not use it any further (apart from the y axis). I had no difficulties in getting figures in the new year using that as a base, and  by my mixture of automated and then overwriting when necessary later on, and seeing a trend. Could you not simply produce a list of day of year (reduced to zero if you like  ( -300)),  for each area and for each year where the extent reaches a predefined level? Then use that for your analysis. I also  think my suggestion of ignoring 2006 and 2007 will also help with getting a realistic/consistent  data set as a lot of data is missing from these earlier years.

Without this sort of data evaluation/analysis nothing will be added on your description of 'internal variability'. It does not seem to be random, and each area could  have its own cause.

As for the Yellow Sea - no it hasn't started to refreeze yet, but with temps of -25C just in land in Central China it wont be long!!!  (my forecast)

MIA

Edited November 12, 2023 by Midlands Ice Age

[WYorksWeather]

28 minutes ago, Midlands Ice Age said:

Unfortunately these are the 'normal' type curves produced by the usual extent analysis. 

Your choice of selection criteria  only displays the standard patterns of ice progress -  which I frequently report/ display  on here from the ASIF.

Everybody agrees that ice extent levels are less now than they were  in 2006.

What I am interested in is precisely the internal variation which you seem  to want to remove for the sake of having an 'overall' solution.

The internal variation you referred to is not random. I have detected a 'pattern' of earlier refreeze, (which is hidden by  extent) as it is only visible in the smaller ice areas outside the Arctic  basin  (and not in the  major 7 areas which reside within it) -  and these latter dominate any extent effect outside.. It is apparent only in the 5 smaller areas over the period 2021 - 23 and in the 3 N.A. based ones over the period from 2014 - 2019.

Using criteria to include these loses the variability signal in the outer areas. 

Extent is useless to evaluate in these  areas (as I described above) since the ice cannot last more than a short while whenever a slight rise in temps occurs. By its nature it is temporary , but nevertheless this earlier start to refreeze is occurring in these specific areas well away from the central Arctic. It is why this is happening that should interest most scientists as it has not been predicted by the models.

I notice that you earlier referred to the date 300, but then did not use it any further (apart from the y axis). I had no difficulties in getting figures in the new year using that as a base, and  by my mixture of automated and then overwriting when necessary later on, and seeing a trend. Could you not simply produce a list of day of year (reduced to zero if you like  ( -300)),  for each area and for each year where the extent reaches a predefined level? Then use that for your analysis. I also  think my suggestion of ignoring 2006 and 2007 will also help with getting a realistic/consistent  data set as a lot of data is missing from these earlier years.

Without this sort of data evaluation/analysis nothing will be added on your description of 'internal variability'. It does not seem to be random, and each area could  have its own cause.

As for the Yellow Sea - no it hasn't started to refreeze yet, but with temps of -25C just in land in Central China it wont be long!!!  (my forecast)

MIA

Hm - not sure if I made the graphs clear enough? The extent still shows mostly the same conclusions as you - there's a clear signal in SOO and Bering, and a slightly lesser signal in some of the other areas. I only included the other areas for completeness sake - can easily remove them.

As it turns out, if I use current year, and drop 2006 and 2007, you're right that the data become a bit more intelligible. I've created a chart of the five interest areas (probably easier to interpret than all of them in hindsight!). This one has the first day of the year after day 260 (rough analogue for the mid-September minimum) when any sea ice at all is picked up. Cooke Inlet still had a weird anomaly for 2008, showing up as having some extent at exactly day 260 and then a very low extent number persisting for a long time, so I junked 2008 for Cooke Inlet.

NOTES:

1. Different y-axis scales for day of the year for each sub-plot, to better show the variation.

3. None of the trend lines reach statistical significance, so take them with a pinch of salt, and they're probably of less interest in any case since they average out the last 15 years.

I think this should be closer to what you're looking for?

I'd like to get a method you can agree with before I try to correlate with ENSO.

 

[Kirkcaldy Weather]

This tweet led me to check  snow depths through sweden and quite incredible that every station has AT LEAST 50cm HALF A METRE with many registering over 60cm

 

[Midlands Ice Age]

9 hours ago, WYorksWeather said:

Hm - not sure if I made the graphs clear enough? The extent still shows mostly the same conclusions as you - there's a clear signal in SOO and Bering, and a slightly lesser signal in some of the other areas. I only included the other areas for completeness sake - can easily remove them.

As it turns out, if I use current year, and drop 2006 and 2007, you're right that the data become a bit more intelligible. I've created a chart of the five interest areas (probably easier to interpret than all of them in hindsight!). This one has the first day of the year after day 260 (rough analogue for the mid-September minimum) when any sea ice at all is picked up. Cooke Inlet still had a weird anomaly for 2008, showing up as having some extent at exactly day 260 and then a very low extent number persisting for a long time, so I junked 2008 for Cooke Inlet.

NOTES:

1. Different y-axis scales for day of the year for each sub-plot, to better show the variation.

3. None of the trend lines reach statistical significance, so take them with a pinch of salt, and they're probably of less interest in any case since they average out the last 15 years.

I think this should be closer to what you're looking for?

I'd like to get a method you can agree with before I try to correlate with ENSO.

 

Thanks @WYorksWeather

Thats more like the data analysis that I came up with. I still think that the method of using a 'cutoff' (say 1KM2 for the above 5) would make the data even more useful, and might even then be useful for the other sea areas!.

(and it is definitely  not because most of your graphs show a downtrend!).. It is the variation that I am interested in....

#

OK then we can now either correlate/test  that (to make certain the method works) or move on to the 3 N.A. areas which show 'odd' behaviour in the 2013 - 2019 timescales.

Here you will find that you definitely  need a cut-off for extent as the numbers are hugely different and some had some ice most of the year - meaning your method will not pick up any realistic data..

Can I suggest my cut-offs of 1K for the above, 10K for Hudson, 60K for Baffin and 350/maybe 400K for the CAA. The reason is that  much ice(?)  was left in the upper reaches of the CAA (around Ellesmere  and Queen Elizabeth Islands - and a lot of it was iceberg's) during this period.  (Ice-bergs must be ruled out from any analysis!!).

Even then it may not be possible to get a consistent data file.

Sorry to be a bit sharp last night, it had been a long day.... though nothing too untowards was found.    - Just  the old Carpel Tunnel which means that I have lost my fingers dexterity -  not good for typing on here,  and golf - my major interests these days. It comes and goes a lot!!  One of the few  pleasures of getting old  (er) !!!! .

Seriously  though -  thanks for the time.             Maybe you too can become an avid ice and snow partaker?.

It is a fascinating scientific subject with so many complexities and which no one has come to grips with totally as yet. The models are improving, but as per the BOE discussions I refer to on here occasionally, there is a lot still missing.   Large amounts of research is still ongoing , and a lot still to be uncovered I think  at the detail level...  (Including during the thawing season from which I give myself a break!).

MIA

 

 

[Midlands Ice Age]

Whilst on here this morning a quickie from the Natice snow and ice  Coverage Charts -

           

Not a big change this morning, but still consistent gains in ice, though the snow has more or less stalled except for the slow creep north westwards in central and northern Russia towards Scandinavia. .  where  a lot more snow is forecast this week and particularly into next weekend.

Ice has increased in Baffin and Hudson and the CAA is now virtually filled. The slow creep towards Alaska has continued and closure can be expected early this week (watch the Barrow webcams! - cannot remember the Alaskan name)...

Elsewhere the big push yesterday towards the Novaya Zemlya Islands has stalled today  (probably a function of the 25% cutoff used for extent), whilst consolidation of the ice around Svalbard continues.  I must say that the push of ice this year into Kara and Barents has been the major biggy  this year. This compensates for the slower ice refreeze in North America.

A few graphics for you -

  snow now below average. 

 

 Ice 'steaming' ahead!!!

    and catching up! 

All for now..

MIA

 

 

 

 

[WYorksWeather]

8 hours ago, Midlands Ice Age said:

Thanks @WYorksWeather

Thats more like the data analysis that I came up with. I still think that the method of using a 'cutoff' (say 1KM2 for the above 5) would make the data even more useful, and might even then be useful for the other sea areas!.

(and it is definitely  not because most of your graphs show a downtrend!).. It is the variation that I am interested in....

#

OK then we can now either correlate/test  that (to make certain the method works) or move on to the 3 N.A. areas which show 'odd' behaviour in the 2013 - 2019 timescales.

Here you will find that you definitely  need a cut-off for extent as the numbers are hugely different and some had some ice most of the year - meaning your method will not pick up any realistic data..

Can I suggest my cut-offs of 1K for the above, 10K for Hudson, 60K for Baffin and 350/maybe 400K for the CAA. The reason is that  much ice(?)  was left in the upper reaches of the CAA (around Ellesmere  and Queen Elizabeth Islands - and a lot of it was iceberg's) during this period.  (Ice-bergs must be ruled out from any analysis!!).

Even then it may not be possible to get a consistent data file.

Sorry to be a bit sharp last night, it had been a long day.... though nothing too untowards was found.    - Just  the old Carpel Tunnel which means that I have lost my fingers dexterity -  not good for typing on here,  and golf - my major interests these days. It comes and goes a lot!!  One of the few  pleasures of getting old  (er) !!!! .

Seriously  though -  thanks for the time.             Maybe you too can become an avid ice and snow partaker?.

It is a fascinating scientific subject with so many complexities and which no one has come to grips with totally as yet. The models are improving, but as per the BOE discussions I refer to on here occasionally, there is a lot still missing.   Large amounts of research is still ongoing , and a lot still to be uncovered I think  at the detail level...  (Including during the thawing season from which I give myself a break!).

MIA

 

 

I'm going to leave the question of the other areas / using extent cut offs at this stage. Two main reasons. Firstly to save my time - adding the cutoffs moves the data into a different year. Once you add a cutoff, the Yellow Sea and Cooke Inlet often don't freeze until the new year, and that would therefore require a lot of slightly painful hacking around with the data. I can definitely do it, I'm just not sure it's a valuable use of my time when we have a solution that works reasonably well for the first five areas of interest.

It is here that the differences between a coded solution and a non-coded solution become apparent. In Excel, it's trivial to just pick out the next value in the new year, record it somewhere and move on. But in a coded solution, the criteria have to be robust enough to apply to everything you're interested in, otherwise it doesn't really work.

Of course, the flip side is that once you've designed the analysis, it's usually trivial to make minor alterations to it -  look at extent instead of dates, change the look of the graph, import a different dataset and with a few tweaks get the same graph as a comparator, and so on.

That being said, I've gone ahead and imported the ENSO data. I matched to the DJF data. For example, 2023 sea ice data is matched to the 2023 DJF reading, which was -0.7, or La Nina.

The correlation between day of first ice and El Nino was 0.07, which is very weak (and like everything else - not statistically significant as there's not enough data!). Of course, in this sort of case, we shouldn't expect much variation to be explained by a single variable. Since the correlation isn't statistically significant we should probably say no trend.

Possibly a better way of doing this might be to get sea ice extent data from further back so we have more chance of getting statistical significance with more years to compare. I'm hesitant to say anything about non-significant results, it could well be just noise.

Conclusion

Taking the correlation as it is, and ignoring uncertainty, we could say that there is a tendency for ice to refreeze later following an El Nino, and vice versa following a La Nina. Of course that's probably not a surprising conclusion. But again don't take that too seriously as it's not a significant result.

[Midlands Ice Age]

As I wont have to much time over the next  48 hrs and the snow and ice situation looks to be in in a slow state at the moment with not much changing, I thought I would look  at the weather possibilities for the outlook period for our northern continents.

The possibility of a 2010 style freezing outlook for the UK has been thoroughly aired on the MOD thread, and currently seems likely to be watered down as the Atlantic anticyclone cannot get far enough North. The charts below are reflective of that situation.

So what does this actually mean for the Arctic in general?

Well todays temperatures  are depicted in the DMI chart (ECM based)-  which has been falling rapidly over the last week.

  Now down to an average below -20C. (about a week later than longer term averages).

and this is also shown in the C.R. temperature charts for today (GFS  based) -

 

     Min                                                                                   Average anomaly                                                          Max           

                                      

 

Points of interest wrt longer term are Scandinavia (cold still building), Atlantic (cooling down),  East Pacific (cooler), Russia (quite a lot warmer) , a warm corridor form Alaska thru to Laptev over the Arctic, but cold on the western side,  and East Siberia into China - cold.   Pretty much average overall.

Step forward 6 days and the surface Level pressure patterns are changing (from this).

   

 

Gone are the low pressure systems strung out over the North Russian coastline (shown above), to be replaced by a slackish area of low pressure  over the pole -  (To this) -

 

    

 

Remember that slack low pressure is good for ice (MIA's theory), so with widespread snowfall shown in the central Arctic it looks OK for ice as the temperatures continue to stay low and with the help of the snowfall shown -

               

However notice the much colder air moving into Northern Russia with daytime temps around -30C and min of below -40C. Also the sliver of 'milder' air associated with snowfall form Canada to Eastern Siberia.. 

Then move forward another 5 days to the 10 day period -

      

Alaska is still cold, Canada is quite warm, but wide areas of Central Russia (except Eastern Siberia - still snowy)  has turned very cold with max at -30C and min around -45C. All the while Scandinavia is shown as being cold. The thing that really stands out here is the difference between min and max.  Remember that most of this area is under total darkness - so how comes the large difference particularly is under quite a bit of cloud?

       

Are yes the flat area of low pressure is now gone to be replaced with a extremely  flat Arctic SLP profile. What is happening to the strong Stratospheric Vortex running at record levels during this time? 

High pressure is dominating over CAA and Greenland,  with a filling low over Northern Russian coastline. Low pressure now has control over the Aleutian Islands and Bering Sea.  The  Atlantic high is succumbing to the pressure of the Atlantic jet and is being squeezed south eastwards. Though hopefully giving us some dry colder weather. The United Sates seems to be relatively 'calm' during the next 10 days with fairly average temperatures. 

So what does this mean for snow and ice?   Looks to me like a lot of snow for areas surrounding the Arctic Ocean.  Scandinavia and Central Russia look to add to their 50cms totals, and Eastern Siberia and Alaska (although not as cold (relatively)) , look prone to more snow, 

The UK  -   hopefully slowly drying up, as the weather turns a bit colder. 

Lets see how things turn out?

MIA

 

 

Edited November 14, 2023 by Midlands Ice Age

[Turnedoutniceagain]

Snow in the Alps !!

[Midlands Ice Age]

Well I have had my 3 days of hibernation , and although it started quietly  the Arctic has suddenly burst into life, possibly more quickly than forecast.

To start off with the Arctic temp map (and a comparison to last year   -

             

shows a similar profile so far (with the one exaggerated increase this year), but we have now reached a stasis at more or less the same as last year. at the same point.

Despite this slowdown in the last couple of days sea ice extent seems to have gone crazy...

Ice has leapt into action with Jaxa (ASIF based with thanks) showing gains of 164K Km2 and (178K Km2) in the last 2 days to move this year up by no less than 6 places (from 10th lowest   to 15th in the last 18years) in just 2 days.

          

Masie is showing similar gains -   of 228K Km2 yesterday, - and although todays data is not yet in, I expect a gain of at least 250)k KM2 -  It is clear that a very large increase has occurred today as ice extent total has leapt to 10.3K KM2. . 

 

In its own right yesterday's data was notable as it hit the 10,000K Km2 (was 9.97K Km2) -  and this at  the same date as the previous 2 years., also this year it means we have frozen an extra 3 -  500K Km2  - because of the lower start figure.

I am unaware  that this sort of date 'stability' in the sea ice extent figures has occurred previously. It must mean something , but not sure what. Although in about 3 weeks time (early Dec) , there is a long term trend that all the  yearly figures merge together.... probably when most of the 'easy' internal to the basin ice has  frozen out., and before the outer ice gets going.

In addition  ........   

I have managed a sneak preview of the weather charts to see how well the GFS predictions are turning out.

 So far they are almost spot on,  with the USA still undergoing  a snow retraction up to the 48th parallel, whilst Russia despite having a lesser amount compared to last year (see below) has suddenly, today, expanded southwards in the central and western areas (compared to previous day  (and also the last year). Heavy snowfall is indeed occurring in Siberia and Scandy, as predicted.

              Last year        

However -     Scandinavia is much more advanced this year.

 This is a picture provided by Carinthian (from his sister in Stockholm) who had just seen  her first snowfall of the winter (last year) on Nov 20th!!!  Whereas we have had most of Scandy covered for about 3 weeks now!!!.

  

The colder air has just started to appear in Northern Russia and it is forecast to really get cold their in the latest 10 day charts from GFS   

  

with temps down below -50c minimum and wide areas of Europe falling below freezing. (PS... this is today's 06Z run which turned cold even in the UK.)...

The latest GFS run also sees colder air crossing into the US again, as  a vigorous depression forms just to the east of Hudson Bay.  

We will see.....

MIA

 

  

 

 

[Metwatch]

Estonia seeing falls of snow in recent days.

[Midlands Ice Age]

On 17/11/2023 at 16:02, Metwatch said:

Estonia seeing falls of snow in recent days.

thanks Metwatch, I notice someone (Dorsetbred?) posted on the Nordic thread that the temp at  1200m was -22C today at 14:00 in Finland.

Today was a fairly average day for sea ice extent (+60K Km2), and snow actually spread unevenly into Poland and  Ukraine so the expected westward push has started. 

Sea ice although a normal increase consolidated in the right areas.

This included Chukchi and Bering (where the ice has filled bays on both the Alaskan and Siberian side) , as well as in the SOO. It also showed up on the Eastern shore of the southern portion of Hudson Bay..

So all looking very strong at the moment.

                 

MIA 

  

Edited November 18, 2023 by Midlands Ice Age

[Stravaiger]

Good to hear the ice is rebounding quickly up north this autumn, would be great to see ice extent hitting 1980-2010 mean levels again. It's been solidly below average for years now.

 

The situation across the cryosphere as a whole is pretty dire, though. NH snow extent is now back to being below average and Antarctic sea ice continues to chart a new course of record low extents for the time of year. 

The crumb of comfort for us in Blighty is that Scandinavia is having a cold and very snowy autumn, with persistent low temps since October now. If we can tap into that pool of cold air or, better still, draw it across the still warm North Sea, we might see some good amounts of snow and ice ourselves