[Automated voice] Conference will now be recorded.

[Deana Crouser] To improve understanding of ecosystem dynamics and applications of that understanding to the management of living marine resources. Since 1986 the seminar has provided an opportunity for research scientists and practitioners to meet, present, and provoke conversation on subjects pertaining to the physical and fisheries oceanography, or regional issues in Alaska's marine ecology, ecosystems.

Visit the EcoFOCI web page for more information at ecofoci.noaa.gov. We sincerely thank you for joining us today as we transition into a hybrid virtual and in-person fall series. Look for our speaker lineup via the OneNOAA seminar series and on the NOAA PMEL Calendar of Events. If you miss a seminar catch up on PMEL's YouTube page. It takes a few weeks to get these posted, but all seminars will be posted.

I ask that you please make sure that your microphones are muted and that you are not using video during the talks. Please feel free to type your questions into the chat and we will be monitoring the questions to address them at the end of the talks.

Okay, first speaker is Silvana. Silvana is currently a postdoc at the Alaska Fisheries Science Center in Seattle. She earned her bachelor's degree in biological science from the University of the Republic of Uruguay and has a master's in oceanography, from University of Cádiz in Spain. She completed her PhD at the School of Aquatic and Fishery Sciences at the University of Washington in 2021.

Thank you so much.

[Deana Crouser] So stand up here.

[Emily Lemagie] It is up to you.

Okay, here we go.

[Silvana Gonzalez] Thank you everyone for joining. So I'll be presenting a synthesis of planktonic data

that I've been working on, in collaboration with
all these authors here that actually know about

plankton. And in particular we are looking at
the associations among different components

of the plankton and the environment

during this work period of 2017 through 2019, in the northern Bering and Chukchi Sea. So, we know that both the northern Bering and Chukchi Seas have been undergoing increasing warming and sea ice loss, in the last decades. And in particular during 2017 through 2019. So here there are some plots of water temperature. On top we have sea surface temperature for the Chukchi Sea.

And we see high positive anomalies, from the, for this 2017 through 2019 period. And here on the bottom we have sea surface temperatures for the northern Bering Sea in the dashed line here, and in 2017, 2019 we see the highest surface temperatures on record, for the northern Bering Sea.

Now looking at sea ice concentrations for the Chukchi Sea here on top, in gray we have values for the 1978 through 2015 period. And the gray line here indicates the average for that historic period. And in, orange, green, and red we have sea ice concentrations, for 2017 through 2019. And we can see how sea ice concentrations have been much lower than the mean average for the historic period.

And in many cases for the extreme values from that period. And similar things we see for the Bering Sea, here we have sea ice extent. The average for 1981 through 2010 in black here, and in yellow.

Don't know if you can see well. And in purple we have the sea ice extents for 2017, 2019 and they have been much lower and lowest than even the extrema for the period, the previous period. So all these warming and sea ice loss is expected to have really direct and rapid impacts on planktonic communities.

We've seen changes in primary production. Changes in the timing of the spring bloom and other biological processes. Shifting the distribution of some species and decreasing trends in the size of phytoplankton and zooplankton. Taxa and individuals.

So just to show a few examples of these shifts in phytoplankton and zooplankton, with warming, we have one example is from Li and collaborators from the Canadian Arctic showing how, with increasing warming and freshening of sea water we see that even though total chlorophyll hasn't changed much, we see a decreasing trends in the, in the amount of large phytoplankton cells. And increasing trends in the concentrations of smaller, phytoplankton cells.

Now moving to zooplankton, this is work by Dave, looking at anomalies in the abundances of Calanus copepods. These are bigger and lipid rich copepods, and here for small copepods, for a series of warm and cold periods, in the northern Bering Sea. And what we see is negative anomalies during warm periods for Calanus copepods, and positive anomalies for cold periods.

And, kind of the opposite trend for small copepods, with negative anomalies. So lower abundances than average in cold periods and positive anomalies during the warm period. In particular in 2017 and 2019. And we are not only seeing this shift in terms of the relative abundance of smaller versus larger copepods, but we're also seeing, kind of decreases in the size of individuals within one species. And here we have again some examples from Dave, showing the mean length for many life stages of Calnus copepods, for cold period and warm period. And we see how the mean sizes typically are lower and smaller, for the warm period.

So, these changes in species composition and size structure of planktonic communities, are expected to affect traffic interactions, energy pathways, Benthic-pelagic coupling and have the potential to alter the entire structure of Arctic marine ecosystems, and the services they provide to human communities.

So to illustrate a little bit, what we would expect for the energy flow under different conditions here I have some very general, conceptual model showing, what we would expect to see under cold years. Or also what we hypothesize to see in spring. When we have lower temperatures and higher availability of nutrients, these conditions are expected to favor large phytoplankton cells that can be directly consumed by large zooplankton.

And be readily available to upper trophic levels. Whereas, in warm years, what we expect to see with higher temperatures and in general lower availability of nutrients, is, more smaller phytoplankton cells that are typically consumed by microzooplankton. So we have microzooplankton as an intermediate here between phytoplankton and zooplankton.

And then, available to upper trophic level. So during warm years what we expect to see is more trophic links, a reduction in the trophic transfer efficiency, which results in less energy available for upper trophic levels. And for the 
[indistinct]. So understanding these associations among planktonic communities, and how will they respond to changes in the environment becomes critical to be able to understand more of the potential ecosystem were responses to climate change.

So, that's kind of the objective of the synthesis. So, to look at these associations among different components of the plankton, I'm using data from the four Arctic IRP surveys which included the two ASGARD surveys conducted in spring here, in the northern Bering and southern Chukchi Sea, in 2017 and 2018. And the two IES surveys conducted in late summer early fall in 2017 and 2019 in the Chukchi Sea. And this service had a very comprehensive sampling of different components of the plankton, and environmental variables. And for this work I'm using just data on temperature, salinity, nutrients, total chlorophyll, and chlorophyll in two size fractions, less than five and greater than five, representing phytoplankton cells of different sizes, the microzooplankton biomass including being a flagellate and ciliate.

And abundances and biomass of small and large zooplankton taxa collected with these two mesh sizes, respectively. So just first with to start with the physical characteristics, during these surveys, all of these surveys again were conducted during a period of extreme work conditions. But we still observe some differences in terms of the oceanographic conditions.

So here I dumped a lot of maps but I'm going to highlight some of the main differences across surveys. So here on the left we have water column temperature, salinity. And here surface nitrate concentrations, integrated for the layer above the mixed layer depth. And here we have water masses, surface water masses, and bottom water masses on the right. So focusing first on this spring, we see how overall spring of 2017 was slightly warmer than 2018.

In particular mean bottom temperatures were, like one degree Celsius higher, compared to 2017. And if we look at bottom water masses here we see how in 2018 we have a predominance of nutrient-rich Anadyr water here, and cold shelf water here, whereas in 2017 we see a greater presence of warm shelf water in the bottom. And regarding nutrient concentrations we see high nutrient concentrations throughout most of the region for both surveys, as we expect to have in the spring.

Now comparing the two summer surveys, summer of 2019, was overall much warmer and fresher than 2017. Mean surface temperatures were, over 2 degrees Celsius higher and salinities were more than one unit of salinity lower, compared to 2017. And here we see how in 2019 we have a predominance of warm coastal water throughout most of the Chukchi Shelf.

Whereas in 2017, the predominant water mass was warm shelf water. And in 2019 this, warmer and fresher surface layer resulted in a shallower mixed layer depth, and lower nutrient concentrations, again compared to 2017. Now looking at the phytoplankton and microzooplankton communities here, we have total chlorophyll, the proportion of chlorophyll in the small size fraction, and microzooplankton biomass combining both kinds of flash latency.

What we see if we compare the season as expected, spring had higher total chlorophyll, a greater predominance of large phytoplankton cells and lower microzooplankton biomass as compared to summer surveys. Now if we just look at the two springs, we see how during the relatively colder spring of 2018 we see higher total chlorophyll and specifically a higher biomass of diatoms that year.

Dinoflagellates, so in general microzooplankton biomass didn't change between these two surveys. But we had shifts in terms of which group was the dominant. The predominant each year. Now looking at summer we see how total chlorophyll was very similar between the two summers, but we see a greater predominance of phytoplankton in the small size fraction.

So smaller phytoplankton was predominant in summer and that was observed, particularly following the distribution of warm coastal water, in 2019. And here I'm just bringing up some observations by Mike Lomas, where he saw, really high contribution of these photosynthetic bacteria Synechococcus, to the total phytoplankton biomass in 2019. Also following the distribution of warm coastal water. So that predominance of small phytoplankton that we were seeing associated with warm coastal water is probably, mainly a result of this Synechococcus species.

And also in summer of 2019 what we see is a lower overall microzooplankton biomass, that is a result of really low biomasses of ciliates that year. So now moving to the mesozooplankton community, to look at the variability in the distribution of this whole plankton community across four surveys, what I used was a combination of cluster analysis, with the indicator species analysis.

Basically what we're doing is grouping this, the stations, based on their similarity and the relative abundance of zooplankton taxa. And identifying which, species or combination of species, are more representative of each group. So what we got here are four different clusters; one cluster in gray here. That is restricted to the spring and is characterized by this species of copepod Neocalanus that typically is present in spring

and then disappears from the region. Then we have this other cluster here in orange that is restricted to the northern Chukchi Sea and characterized by the combined occurrence of Calanus glacialis and Calanus hyperboreus is to big copepods, lipid-rich and very desirable prey for upper trophic levels.

And then we have the distribution of these two other clusters in green and blue here, whose distribution shifted following the distribution of warm coastal water versus warm shelf water. So in 2019 when we had this predominance of warm coastal water over the Chukchi Shelf, we see an expansion both northward and offshore of these, clustered here in blue. That is characterized by the occurrence of these two copepod species that typically have a neritic distribution and now have a really wide distribution during this year.

And then the distribution of this group being green here is very much restricted to the north, compared to 2017. And is characterized by the occurrence of meroplanktonic taxa, in particular a Echinodermata larvae. So now to put all these pieces together, and look at the correlative associations among these different groups and how those associations change between the two seasons and across the four surveys, I'm using
Structural Equation Models. These models can be visualized as this path diagrams. And I think they have, like two advantages for these applications.

One is that, each path here, represent a hypothesized or a pre-assumed causal or directional relationship between the two variables. And also that each variable here can be both a predictor and a response variable. So we can test for indirect effects or cascading effects in our system.

So once we define our path basically we translate this into a series of structural equations which are basically just, multiple linear regressions. And then we see if our data supports this hypothesized structure. So for this model I started very complex but ended up very simple for it to work. So, at the end I included biomass of large copepods and small copepods. Large copepods including Calanus, Neocalanus, and Eucalanus species. And small copepods including; Pseudocalanus, Acartia, and Oithona species.

Then I chose ciliates to represent the microzooplankton community because they track pretty well, the total biomass of microzooplankton. And two size fractions of chlorophyll; less than five and greater than five. And temperature as our environmental driver. So basically, I got one of these models for each of the first surveys, so we can compare across surveys and between the two seasons. So this is what we have here. So here are the path diagrams for the two spring surveys on the left.

And for the two spring surveys, for the two summer surveys on the right. And the black arrows here indicate significant path and the gray dash arrows here indicate paths that we tested for but they didn't come as significant. And these coefficients here is for the significant path represent kind of the strength of the effect of one variable on the other.

So the greater value the stronger the effect, so we can compare within each year what variables had a stronger effect on the others. So coefficients for 2017 are shown in blue, for 2018 in green, and 2019 in red. So I just want to highlight a couple of things here, from this model. So the first thing is that we see that the effects of temperature on large copepods is negative, strong and negative for surveys both in spring and summer.

And then in the summer we see that the effect of temperature on small copepods is positive. So that's suggesting that colder conditions tend to favor large copepods, and warmer conditions favor small copepods, as we expect. And we know from literature as well.

And in particular the effect of temperature on small copepods in 2019 was much stronger than 2017. That was when we had the really warm conditions and also freshening of sea water. And also that year we saw much higher abundances of Pseudocallanus, the small copepods, in the area. And the second thing I wanted to highlight from here is that in spring we see a direct significant effect of chlorophyll on large copepods that is not present in the summer. Instead in the summer what we see is significant effects of ciliates on copepods.

So in summer, so this might be indicating some sort of predation pressure of copepods on microzooplankton, instead of directly on phytoplankton. And we know from work from Campbell and collaborators that, zooplankton tend to prefer to eat microzooplankton in the summer. So this is kind of expected. And also even though these are just correlative associations, we can kind of infer that in the summer we might be seeing a higher number of trophic levels than what we see in spring.

So what? So we see that warming and the increasing inflow of warm coastal water into the Chukchi Sea, that we observe in summer of 2019, was associated with smaller size phytoplankton, in particular this kind of novel high contribution of Synechococcus, to the total phytoplankton biomass in the region. A mesozooplankton community characterized by high biomass of small copepods, as we saw for Pseudocallanus in particular and also the occurrence of these two species of copepods that are typically neritic.

And have different feeding strategies. They're basically carnivores and we also saw lower microzooplankton biomass. And that can be due to a really high predation pressure because this microzooplankton in summer of 2019, not only had kind of all the microzooplankton that actually prefers to eat microzooplankton, but also these other two species that usually are not there, in those abundances.

So, the other thing that we can get from these results is that this summer, open water type of production that is characterized by larger, by smaller phytoplankton cells, seems to be supporting a planktonic food web that has more trophic links that we see with a type of ice-edge, type of production, as we see in spring with large phytoplankton. So the question now is will summer conditions become kind of the norm for the Bering and Chukchi Seas, over longer periods of the year in the absence of sea ice?

Or during a much earlier sea ice retreat, and I guess the answer is going to depend on many other factors including what is going to happen with the mixing of the water column, and therefore the availability of nutrients. But what we can guess is that, this shift in in terms of at least the size of copepods.

So if we're replacing the really large copepods by smaller, less nutritious copepods, we are reducing the food quality that is available to upper trophic levels. And also if we have these higher number of trophic levels, we are reducing the energy that is available to those upper trophic levels. And also more energy, or carbon is going to be retained in the pelagic and less, less is going to be exported to the balances.

So overall what we see, expect is, this shift towards smaller planktonic organisms to have cascading effect throughout the food webs. And have the potential to impact commercial and subsistence fish and shellfish resources. And obviously benthic ecosystems. So, changing from one, from the typical type benthic-pelagic coupling to a weaker bethic-pelagic coupling in these systems. So that's all.

[Participant] Yay!
[Applause]

[Deana Crouser] All right. Do we have any questions for Silvana while we get Ed set up?

[Deana Crouser] Yeah Julie?

[Julie] I got very curious right at the end about the neritic carnivorous copepods.

[Silvana Gonzalez] Yeah, I don't know much about them. I found like one paper, like those typewriter papers and very obscure literature about those. So I don't know much about them. I just saw that, at least one of the species, I don't remember,

one is carnivores. So it's not even
omnivores. Just carnivores.

[Julie] Do you think they're being advected up? Or are they, is the food web more supportive of them?

[Silvana Gonzalez] I think they were advected because they are associated with that warm coastal water that is probably coming from the Alaska Coastal Current. But I don't know, Dave...

[Dave] They're definitely advected.

[Julie] But the conditions I guess are also.

[Dave] Well yeah, once they're advected, the conditions are favorable for them.

[Julie] Yeah.

[Dave] Because [?], for example, is a weird copepod. It's blue, it sits right in the surface. We don't see very much of it and it exploded over the last periods of low ice.

Why? We don't know and they're big time in salmon diets when they explode. So.

[Deana Crouser] Yeah Lisa?

[Lisa] I have a question. This is a little bit outside of your study but, with the warming your metabolic rates will also increase, so they require more energies. Is there a way you could also put that into your modeling?

[Silvana Gonzalez] Not in this type of model, like... This is more just correlative associations but in a different model for sure, maybe Jens can do that, and then it's probable.

[Deana Crouser] We have a question in the chat from Lauren.

What other structures of SEMs, S-E-M-S., did you try? You said the one, you presented was after some trial and error.

[Silvana Gonzalez] So I included, other groups of copepods like the neritic copepods that made it explicit in the model. I also included nitrate. I included salinity first, but yeah the model wasn't really performing well. So I was, I decided just to have the minimum things just to make some interpretations out of it. But at the beginning I started with multiple environmental variables, and more groups of copepods. And then the two groups of microzooplankton, so I put everything that I had.

[Deana Crouser] Ken.

[Participant] Yeah, my question is sort of related, pieces, I guess is it, is it possible that any of the reduction in a large zooplankton and under warm conditions could be due to a top-down effect of increased grazing by fish, as their metabolic demands are higher in the conditions?

[Silvana Gonzalez] Yeah and for sure and that applies even for chlorophyll as well. I don't know what is a bottom up driven, versus top down. So yeah, I'm not sure. If we had, yeah, we could add fish to the picture and see what the associations are, with zooplankton here. But definitely that can be everything.

[Deana Crouser] Colleen said, the spring and fall study areas are somewhat different. How do you think that this affects your results?

[Silvana Gonzalez] Yeah so that's the problem with these first surveys, that we don't know what is a seasonal difference versus a regional differences. But, yeah, so that's going to be always kind of trying to use literature to distinguish between the two effects. But yeah.

[Deana Crouser] Any other questions?

Awesome, well I guess we can transition, and get Ed set up.

[Applause]
[Silvana Gonzalez] Thank you so much.

[Deana Crouser] All right.

Ed is a Research Assistant Professor in Atmospheric Science at the University of Washington. And he studies high latitude climate and weather, with a focus on sea ice variability and predictability.

Welcome Ed!

[Edward Blanchard-Wrigglesworth] Thanks! And I'm hoping, you can see my screen?

[Deana Crouser] We can see a lot of stuff on your screen. [Laughs]

[Edward Blanchard-Wrigglesworth] What can you see now?

[Deana Crouser] I'm gonna try that. Just your presentation.

[Edward Blanchard-Wrigglesworth] Just the presentation?

[Emily Lemagie] Yeah that's perfect.

[Edward Blanchard-Wrigglesworth] Alright great. And you can hear me all right?

[Participant] Yes.

[Edward Blanchard-Wrigglesworth] Okay perfect.

Well thanks so much for inviting me to speak. And I'm sorry I can't be there in person. And today I want to, talk about a recent, study of a very recent case study.

That's been quite a fascinating case study. And, my title is that record low sea level pressure Arctic cyclone of January '22, how did it lead to record loss of sea ice and how forecasts fared.

And, this figure here shows the synoptics, synoptics situation on the 24th of January of 2022. So, about a year and a half. And, in black contours you can see the sea level pressure. You might be able to read those labels.

And in the background I show you the sea ice concentration, in the shading. So, the blue areas, where the sea ice and blue is, sorry the white areas where sea ice and blue is where there's open ocean. And we're focusing on the sort of a, I guess you could call it the European sect of the Arctic. So you can see Greenland on the left and Scandinavia to the bottom and sort of western, or northwestern Russia to the right. And, you can clearly see a very strong cyclone and...let me hide this.

Oop. No, sorry. I don't know how that happened.

Can you hear me?

Sorry, I don't know what... I was trying to hide the, the message on this, on this software.

So I'm hoping you can see this again.

Are we back to the first slide?

[Emily Lemagie] Yeah that looks great, thanks. Yes.

[Edward Blanchard-Wrigglesworth] Okay, sorry about that.

So anyway, so if you, this is by the way I'm going to be showing data from the ERA5 Reanalysis, for the atmosphere and the waves. And passive microwave for sea ice. And basically, the central pressure of the cyclone was 932 millibars and anyone in the room, that has experience, I'm a bit of a weather nerd. So anyone that has experience looking at weather maps, will immediately, notice or you know kind of realize what a low value this is.

And this is more typical of hurricanes, or extremely low, midlatitude cyclones in the storm tracks, right? This is not typical in the Arctic. And I remember seeing this, somewhat by chance on this day on the 24th of January, I was looking at weather maps of the Arctic and I remember quickly thinking this is, this is very atypical.

So with my colleagues that we kind of by chance working on a project of Arctic cyclones and impact of waves on sea ice, we quickly did, started doing some research. And, what I show you here is the time series of the lowest sea level pressure value every month, between 1979 and 2022.

So this is in the full ERA5 data set, in the satellite period. And on the left panel that is the lowest sea level pressure value every month, north of 70. So between 70 and 90 north. And on the panel on the right, is north of 80 north. So you know they're very high Arctic. And what I show then is in the asterisks are the 10 previous record values. Record low values of sea level pressure. And the dash values are the cyclone January 22 values.

And you can see that it beat the record, in both domains. Like if we just look north of 70 or north of 80. But what's quite remarkable is how far north you know, not just how strong the cyclone was, but how far north it was. So if you look at the panel on the right, it beat the previous record by five millibars. And the second place or the prior second place, by about 10 millibars. So you know, much, you know quite a bit bigger than the previous records.

The other thing I wanted to note is that if you look at these time series, there's you know, there's not an obvious trend to them. So, we don't see a climate change signal in terms of the depth of the lowest cyclones. If we look at the tracks of the cyclone. This basically originated near Greenland.

Kind of went overhead about Svalbard, and sort of then went into the high Arctic, kind of, through the Barents and the Kara, and that is shown in red for this cyclone. And then I'm showing in the other colors the previous 10 record cyclones that are all you know the lowest values north of 70. These are all Atlantic cyclones right, there's none, none of them coming in from the Bering. Of course there's strong and strong cyclones coming in the Bering but they don't have as low sea level pressure as the ones from the Atlantic.

Now, taking a little bit closer look at the cyclone, it basically developed in two stages. And this panel on the top right, this shows you the, the central pressure of the cyclone. So you know the eye of the storm, like what the pressure was, between the 21st of January and the 28th of January. And you can see that it kind of developed in two stages, right.

There's an initial, initial deepening of the sea level pressure from about 990 to 950 millibars on the 21st of January. And then it's kind of stable and then there's a second deepening, a secondary low actually develops on the 23rd. And it's sort of kind of, a little bit of a one-two punch. And, it's just quite interesting development, or evolution. And these panels here, they show the sea level pressure contours in black. In magenta is the sea ice edge. And the shading that kind of orange-red shading is the wind speed. And I show every 24 hours from the 20th of January, in the top left, and time goes kind of to the right and down. And, I do actually between the 23rd and the 24th of January, I show every 6 hours.

And that's very interesting period of the secondary development. So you can see that on the 21st and the 22nd, we see this kind of cyclone develops, in the lee of Greenland, kind of just to the east of Greenland. And it actually tracks along the sea ice edge for a couple of days, until the 23rd of January. And if you look at the top right panel you can see just east of Iceland. This kind of the secondary, low that starts developing in the warm sector. And if you look at the middle row, these are every six hours, between the 23rd and the 24th of January. You see this kind of secondary development, between Iceland and Norway. Just deepens extremely quickly and sort of then swings north towards Svalbard, and on the 20th, on the 24th of January at 12:56 just to the east of Svalbard.

And you can see all those really tightly packed contours. So very very high wind speeds. And then you know over the following few days it started weakening and, slowly, you know, traveling into the, into the Arctic and kind of pretty much loses it by the 28th or the 29th of January.

Now if we look at the winds and the ocean waves, it was extremely windy as you might expect from those very tightly packed, sea level pressure contours. And the record, the highest value was 100 kilometer an hour winds, and these are the sustained 1 hour winds. So the gusts would have been even higher. And again that is, that would be a hurricane, a Cat 1 
hurricane. So we have basically hurricane level winds, in the Arctic with this cyclone.

In the shading, that is the significant wave height. And you can see this you know massive waves, as you might expect from the winds. The blue arrows that is the direction of, of waves of ocean waves. And you can see we had wave heights of over 6, even 8 meters, in the open ocean. Especially in the Barents, traveling, these waves were mostly traveling towards the ice pack.

Right, or into the ice pack, as you can, the magenta there's the ice edge, right. So, you know, we expect that had a big impact on the sea ice. The other thing we did is, there's a new NASA, well relatively new NASA satellite, ICESat-2. Which is basically a very high precision laser altimeter, and myself and some colleagues have shown how you can use this to measure waves in the ice. And what I show you here is a transect from the satellite on January the 23rd. It's, it is, a day before the peak of the cyclone, but there was still very high winds and waves already on the 23rd.

And that bottom right panel shows you the surface height, along that transect and as you go from left to right, in that figure and that bottom right figure, is from south to north. And basically we saw or we can detect, pretty impressive waves in the sea ice. Like over 2 meters, in height at the southern, kind of sea ice area closer to the sea ice edge which is where you expect the bigger waves. And you can see how they dampen, as you go along the transect, but they still penetrate the sea ice quite a long, quite a long way. And why is this maybe important? 
This is a little, just to give you a sense of what waves in sea ice can be like. This is a little video I took myself, in the Beaufort Sea, from an ice breaker. And hopefully you can see, you know this is pretty much close to 100% or at least 95% sea ice concentration, right?

But the waves are still, getting into the, into the sea ice. And you can see this kind of swell. And the other thing to notice is how small the flow size distribution is, right? That ice is really broken up and that's, that's what waves can do to the sea ice. Okay in terms of the temperatures, of you know quite often these cyclones in the winter bring pretty warm temperatures and, and sort of heat and moisture fluxes into the Arctic.

This is, this figure I show you the daily map of temperature anomalies here. And there were some pretty big anomalies, up to 10, 15 degrees Celsius. You can see all that red shading especially on the 23rd, 24th, 25th, 26th of January. And it's normally to the east and north of the cyclone center. So this is where the warm sector kind of wraps around the cyclone, bringing warm air from the south.

So, it was a warm episode. But if we look at the temperature distributions the, these were not record warmth, right. So it was a record strong cycling in terms of the depth and the wind speeds. But not in terms of temperatures, right? So it was, it was warm but not record warm. And this is relevant when we see the, the impact on the sea ice of the of the cyclone.

These distributions on the right, they show the average temperature over the Barents domain and the Barents/Kara/Laptev domain. And the blue, the blue histogram is the historical temperatures for January between '79 and 2021.

And the black lines, that's the maximum of this storm. So you can see it's to the right of the distributions. You know so it's anomalously warm, but it's, it's not, they weren't record warm. Actually quite a few degrees, 3 - 4 degrees from, from the records for this time of the year.

Okay so now moving on to the, impact on the sea ice of the storm. Here is, that I show you the sea ice concentration on January the 21st. And January the 27th. So before and after the cyclone. In the panels on the left and then the panel on the right, shows the difference between the two dates. And I mean you can just visually pick up, just looking at the two panels on the left, but when you look at the difference it very clearly shows like there was a massive, decline in sea ice concentration. And there were two main areas. You know, one is the main sea ice edge along the Barents and the Kara. And that basically you know, that retreated, kind of into the Arctic.

And then there's the second area of sea ice or open ocean, that, or ocean that opened between northern Svalbard and northern Franz Josef. And, and yeah this is, this is quite atypical for the sea ice this far north to, to basically retreat and open up in January. And you know this is climatologically this is still a time of sea ice growth and expansion.

Now, if you integrate the sea ice concentration in this dashed box, on the right, this is the Barents/Kara/west Laptev. We see that this was a record loss of weekly sea ice area over like, over a six day period. And that is shown by this histogram in the, in the middle in the bottom. This histogram that is all six-day changes in sea ice area for that region from '79 to 2020.

The black line is the, the observed value with this cyclone. And it was, we lost almost half a million square kilometers, in this region over just six days. And this was 30% greater than the previous record. And if you look at the time series on the left, I show this. 2022 is shown in black, and all the other years, of '79 to 2020 are shown in the kind of orange to blue faded, time series.

And with the mean in magenta and you can see that, this cyclone actually meant that we went from slightly positive sea ice extent anomalies, sea ice area anomalies, in this region to the sixth lowest on record for that date. And you know generally sea ice concentration is a lot more, has a lot more persistence than say temperature or atmospheric variables, right?

So, for you know such a big change over such a small time period, is, is very very unusual. How did the forecast do? So this is a segue now into how the forecast performed. These are forecast from the ECMWF of sea level pressure in the black contours and the winds in the shading. And I show lead times from eight days to zero days. So the top left panel that's the 192 hour forecast, which is eight days. And as you go along each panel, you kind of get a day shorter to the 24th, of January which is the target forecast. And that is shown on the figure on the right. And if you're looking at this figure, especially from the back of the room and the middle and bottom rows, if they all look the same to you, and they look very similar to the target figure or the target panel on the right that is a sign that these were very skilled forecasts of the cyclone.

And that's what we found. The ECMWF model did a really good job at forecasting this cyclone up to five days in advance. But even if you look at the top panel, sorry the top row, those eight, seven and six day forecasts, it was still forecasting a really big storm, somewhere in the vicinity, right? It wasn't quite as strong and you can see how the forecast from one day to the next kind of, you know it's sort of it's bouncing around a little bit. But then from five days onwards, it's, you know it's a very good forecast, for such an extreme and unusual event.

And this is confirmed by looking at metrics like the central pressure error. So that is the error in the forecast of the cyclones lowest pressure in the middle. Or looking at the position error of the cyclone which is shown by the kind of more orangey time series. And, uh, these metrics are very high scales,
especially from 120 hours onward especially when we compare it with past cyclones. Or the scale of other cyclones in the past.

And finally how did the forecast of sea ice did? Well this is the, that same model, the ECMWF forecast model, also has a sea ice component. So it can also forecast sea ice variables like sea ice concentration and thickness, and these are the same figures I showed you earlier, but for the forecast, right. So on the top left is the sea ice concentration in the forecast when it's initialized on the 21st of January.

And then the middle panel is the forecast, the six-day forecast of sea ice concentration on the 27th of January. And then the panel on the right is the difference in the forecast. And just to remind you what the observed difference was, if you compare those two panels you can see that the forecasts got a lot less of the observed sea ice loss.

And that's quite interesting because the atmospheric forecast was very good, but the sea ice just didn't respond enough in the forecast model. And if you look at the total sea ice area, for that region in the forecast, that is shown in the colored lines. That kind of orange, so the red to blue faded lines that is for every initialization from the January, from different forecasts in January. And you can and the observations are in black. And you can see that, yes they do forecast the loss in sea ice, but it's about half to a third of what we observed.

Now, we see something similar in thickness. And I think I might skip this for time. But very quickly I mentioned that we saw quite a loss of sea ice thickness in observations. But the forecasts, didn't really, you know they underpredicted how much sea ice thickness was lost with the cyclone.

Now, I just, this is kind of a little schematic slide. It's my final slide, about how waves and sea ice, how what we think in reality they have an impact, and how they are for now, assimilated in forecast models.

So on the left, we see that you know in reality we know that waves, coming from the ocean, and they do penetrate into the sea ice. They do dampen and become smaller, as they go into the sea ice, but they do actually, you know they can travel quite a bit into the sea ice pack. And we know that they break the sea ice, so the flow size distribution becomes a lot smaller.

And we also know that these waves, they can help mix the upper ocean, right? So you get turbulent mixing so you can bring up warm waters, from below in observations. And one of our hypothesis when we saw that huge loss of sea ice in the observations, with a cyclone that wasn't that warm. Because we looked at the surface fluxes and the atmosphere itself couldn't, cannot account for that much sea ice loss.

We hypothesize that basically this is a lot of melting from below, right? That the ocean with those very strong winds and waves is getting mixed and you're getting melting from below of that sea ice. Whereas in the model, we are as of today, these forecast models and even GCMs they don't simulate waves going into the sea ice, right.

They can simulate waves in the ocean, but once they get to the sea ice it's like a wall and they just, it's like they reach land. They just die. And the sea ice is not at all affected by the waves, in the model. And therefore you wouldn't be able to get this kind of ocean turbulent mixing process. And you probably wouldn't, and so you wouldn't be able to get enhanced melting from below.

Which is why we think that in the forecast, the atmosphere, you know even if you could forecast the atmosphere really well, you're not getting that sea ice response. And I will just leave some of the kind of summary highlights, here and if you're interested in the study we published this in JGR Atmospheres last year. Or you can send me an email or come and see me in Atmospheric Sciences on campus.

Thank you.

[Applause]
[Deana Crouser] Thank you so much, Ed.

Do we have any questions for Ed? That was a great talk!

[Participant] Sure. I'll ask a question.

So Ed, in the model comparison that you were pointing out, that didn't capture the sea ice loss, I also noticed in that, that it enhanced sea ice creation off the coast of Greenland

that was not in actual, in reality. You hypothesize that it's the same mechanism happening here that the turbulent mixing or whatnot is also not being accounted for on the cold sector side of the storm?

[Edward Blanchard-Wrigglesworth] That's a really good question and yes. I mean you could have the same mechanism even though on that side, of the storm you have more the offshore winds, right, or off ice winds rather than on ice winds. You still going to have, you're probably still going to have enhanced turbulent mixing, under the sea ice.

Even just from enhanced sea ice motion, you're probably still going to have some you know some waves maybe smaller on that side. But yeah that's quite likely, you know also having an impact on that side, of the cyclone.

In observations in the east Greenland, yeah we saw very little change in the sea ice concentration. Even if you're blowing these really strong, you know fairly cold winds on that side of the cyclone, kind of off ice.

But yeah, no, that's a good question. It's you know we focus more a little bit on, on the sea ice side of the, of the problem. But yeah I mean it's you know these mechanisms if there's a bias in the model in the mechanisms, it's going to be everywhere.

[Participant] What do you think it'll take to incorporate, waves and sea ice into forecast models?

[Edward Blanchard-Wrigglesworth] That's another very good question. And there's quite a few different research groups working on this. Including my colleague Cecilia Bitz. And, we are getting quite close and my goal or one of our goals in this project is to, run not with the ECMWF forecast model, because that's not something we can run.

Because you know it's you have to be at ECMWF to run that model, pretty much. But we've been using the NCAR GCM. The CSM model that some of you might be familiar with. And what we've been doing is that we've been using and it's a fully coupled model. And, Cece Bits is working on incorporating a wave sea ice coupling module to that model. And what we would love to do is to, well our aim is to run the model, with the observed atmosphere.

Right, so basically you replicate the observed cyclone in the GCM. You can do this with a nudging component and to basically switch on and off the wave sea ice coupling, right.

So you run a simulation where you have the observed cyclone but the, like the default model. So no wave, sea ice coupling. And then you repeat the experiment, but you allow or you have, you have the wave sea ice coupling active.

And we're very curious to see if, you know how big of a difference that might make. And in this model you know, you know, nice thing about working with the GCM, is that you pretty might have all the data output you want.

And so you can look at the ocean response. One of the issues with the observations is that although this is also something I'm working on with a colleague, we did find an Argo Float from the Barents for that week. And it does show a huge response of the ocean all the way down to the
bottom 200 meters depth, in terms of velocity and temperature difference, before and after the cyclone.

So you know we're starting to see some evidence that yes like the ocean responded quite strongly to the cyclone, and that it probably, you know fluxed a lot of heat to the sea ice.

[Emily Lemagie] Any other questions, in the room or in the chat?

[Deana Crouser] Looks like lots of compliments. Great talks. Lots of thank-yous!

[Emily Lemagie] All right. Thank you so much.

Let's just thank both of our
speakers one last time.

[Deana Crouser] Yes, thank you. Thank you.

[Applause]