[Recording] This conference will now be recorded.

[Emily Lemagie] Good morning everybody and welcome to another EcoFOCI seminar series. I'm Emily Lemagie, I'm co-lead of the seminar series with Deana Crouser, and this seminar is part of NOAA's EcoFOCI biannual seminar series focused on ecosystems of the North Pacific Ocean, Bering Sea and U.S. Arctic to improve understanding ecosystem dynamics and application 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 promote conversation on subjects pertaining to physical and fisheries oceanographer... oceanography, or regional issues in Alaska's marine ecosystems.

Visit the EcoFOCI Webpage for more information, and we sincerely thank you for joining us today as we transition to this hybrid virtual and in-person fall seminar series. We look forward to the speaker lineup. You can find us at the One NOAA Seminar Series and on the NOAA PMEL Calendar of Events.

And if you missed a seminar you can catch up on PMEL's Youtube Channel. It takes a few weeks to get these posted, but all the seminars that are recorded will be available there, and I ask you to please make sure that your microphones are muted and you're not using video. And during the talks please feel free if you're online to type your questions into the chat. We'll be monitoring for questions but we will wait to address them until the end.

And today I'm excited to introduce three speakers: Deana Crouser, Alison Deary, and Ned Cokelet. This is a trial of a new five-minute lightning talk format for the EcoFOCI Seminar; we also have full seminars and paired 15 to 20 minute talks that we'll be trying this fall series. Today, we'll have the opportunity to hear from multiple speakers about recent projects, ecosystem indicators and measurement techniques and more.

In this session we'll begin with an introduction to the ABT Communication Method and end with an open discussion session with our speakers as a panel. We're going to begin with Deana Crouser.

Deana is a Zooplankton Ecologist and a contractor with Lynker Technologies in support of NOAA's Alaska Fisheries Science Center. Her research focuses on how environmental conditions affect zooplankton size and distribution.

[Deana Crouser] Thank you Emily. Alright, so as Emily said my name is Deana Crouser, I'm a Zooplankton Ecologist here, and um, here at NOAA, and I have spent the last two years at a communications training program specialized for climate change. Climate Ambassadors are passionate and dedicated aquatic scientists that want to reach a variety of audience and expand their thinking on the impacts that humans have on our aquatic environments, and we do this through use of communication [recorded message interruption].

Oops - and we do that through a use of a variety of communication tools learned and practiced over two years time. Um, to learn more about the program just search AFS Climate Ambassador Program, and before our lightning talks I just want to spend about 10 minutes teaching you a little bit about what I learned in hopes that I can contribute to your science communication skills and add new tools to your toolbox.

So one of my favorite tools that I learned in the Climate Ambassador Program was this ABT Method, which is a storytelling structure developed by a Marine Biologist turned filmmaker Randy Olson. Though its origins reached back thousands of years to Aristotle and the beginning of narrative culture, the ABT was used in the Gettysburg Address and in MLK's I Have a Dream speech. Even the scientific method is built around the ABT, which underlies the IMRaD Narrative template for scientific papers. I found the ABT to be one of the most compelling arguments for using a good narrative form while communicating our science.

So I'm going to start this talk by talking about storytelling. So there was a researcher - his name was Uri Hasson at Princeton's Neuroscience Institute and he looked at brain activity while subjects watched four different types of movies. The first movie is in green to the far left, and it was a suspense murder mystery, Bang You're Dead, and it was jam-packed with plot twists and foreshadowing. The second movie subjects watched was in blue, The Good, the Bad and the Ugly.

It was a less suspenseful western movie. The third, in red, Curb Your Enthusiasm, was a sitcom, and the fourth, in orange down at the bottom right corner, was a Washington Park reality segment, so it was just people randomly going about their day. To quantify the differences we can look at the ISC Score for each movie and the ISC score is how they quantify brain activity. So you can see in the green it's... the ISC Score is much higher for the murder suspense mystery movie than compared to orange, to the far right, where subjects were just watching random people walk around in a park.

So what is the difference between the four movies? Well, the answer is narrative. Narrative is a series of events that occurs in search for the solution to a problem and the 'but' is the powerful word of contradiction that changes the direction. It fires up the brain. You identify the problem when you say 'but'... so if you look at the difference between these two scans the one on the left is the reality segment: it had a non-narrative, and the one on the right is the high suspense murder mystery, and it had narrative.

And this is what you want. You want your audience to be engaged. You can tell that from the brain scans that people were sitting on the edge of their seats, they were actively listening, actively trying to follow the story, actively trying to predict the plot. That's what you want when you're communicating your science: you want everyone to be as engaged as you are. And that's what the power of narrative brings you and this is just a setup of why it's so important to use narrative when we're communicating our science.

So there are two types of bad narrative forms, which we have to try to avoid if we want our messages to stick when communicating our science. The 'and, and, and" or 'AAA' is one of them, and it has two little... too little content, too little narrative. There's no contradiction, which means that there's no narrative content. Now our inner circle can listen to this because they're just as interested in your work as you are, and it's just an information dump really. So you can do this to your inner circle but if you were to "and, and, and" an outer circle, no one's going to really be able to follow what you're saying, they're not going to really be involved, they're not going to understand.

And so this is another a type of bad narrative form. The other one is "despite, however, yet" which has too much narrative content. It has multiple contradictions, and this is something that we also want to avoid. This just causes confusion. A good narrative sits right in the middle: "and, but, therefore" -- it's the optimal amount of narrative content to keep your... the people who you're talking to, engaged. This is how you reach your outer circle, and this is when narrative is important -- when you're telling a story.

As scientists we're used to communicating our science, and we are familiar with using a narrative template like the IMRaD, but not everyone here is familiar with using a simpler storytelling structure for everyday use like outreach, presentations, proposals and more. Therefore, I would like to introduce you to the ABT Method.

So this is breaking down the ABT. 'And' is your setup. It's your agreement; it's literally the introduction for your scientific papers. Your 'but' is your tension, your problem, the contradiction - a confrontation. It's your methods and your results. And then your 'therefore' is your resolution. Your solution, your consequence, your action, your discussion.

So the... sorry I'm sick a little bit so bear with me... the IMRaD follows a basic ABT structure, but there are some major differences in how people outside academics absorb information. Therefore, we need to adjust our communication approach for the outer circle. So here's an ABT and you can see that the IMRaD follows a basic ABT structure: that's our setup. There are some major differences in how people outside academics absorb information: there's our problem. Therefore, we need to adjust our communication approach for the outer circle. So this is our solution - so it's... it's a simple three sentence setup in which you can... it could... really an ABT can be as long as you want it to be.

It could be three sentences, it could be an entire presentation, it could be a whole movie... You know, that's the best thing about an ABT is, you know, it could be as big or as little as you want to, but this is its simplest form, three sentences. And - But - Therefore. Set up a problem and a solution. Now one of my favorite tools that I learned for the ABT is using a past and present and future setup. So basically ABT goes like this: In the past, life was like this and all was well. But now we have this problem.

Therefore, in the future, we're going to do this. So this is how we set up our ABTs. But I'm sure you guys... I'm just giving you some examples and breaking it down for you, but sometimes it feels hard to start. So let's first... we'll have to see... the first step is identifying your problem. When you identify your problem you have to consider the following: Do you have a single narrative? When we're creating an ABT we don't want to have multiple narratives that are contradicting each other or confusing it, and we want it to be simple. We want one narrative. You also have to make sure that you have the correct source. Is the problem pollution killing fish in a river, or is it laws that allow for pollution? What's preventing the solution? If the problem is pollution, then why isn't it stopped?

Step two, the 'and'... The 'and' material has two main elements, the What and the Why. Your 'what' is a simple opening description in an ordinary world, like this: Management of caribou herds in Alaska has been studied for over a century. Now your 'why' is the why should we care element. What's at stake. And caribou are a major source of Alaska's $126 million annual income from tourism. Now you should end the opening of your 'and' statement as peaceful; it should be an end-of-story feeling.

So if you put it together, management of caribou herds in Alaska has been studied for over a century, and caribou are a major source of Alaska's $126 million annual income from tourism. I feel... I feel like that's the end of the story, I feel like it's peaceful, I feel like... I feel great. Here's a note always to keep in mind: the power of storytelling always rests in the specifics, so try to make the what's at stake element as specific as possible. So not just 'caribou are important' but actually how they are important: tourism. And if possible, at a dollar or an aesthetic value.

The 'but, and, therefore' - now these are two moments... the two moments of transition are the best chances to break through the noise and capture someone's attention. You want to maximize contrast by going from calm, when you felt like at your 'and' you're... you feel like the story is calm, peaceful, it's resolved, you heard the end of the story, everything's fine - to excited. You want to go from the 'what' to the 'how' and not in reverse, so the 'but' begins the contradiction with what the problem is and then follows with the 'how' of the problem. So the right way to do this would be 'Caribou are threatened with 90 percent population loss in the next year from habitat loss, pollution and hunting.' The wrong way would be: But habitat loss, pollution and hunting now threaten caribou with 90 percent of pollution in the next year. So you set it up with the overall statement and then you talk about the process that causes it.

I mentioned earlier, that the ABT was used in MLK's 'I Have a Dream'. So I wanted to share that just to prove to you that this has been used and this is one of the most famous speeches that we have. Well I mean there's a lot of famous speeches, but this is a very very famous speech, and so we can go over and break it down, your "and" is in the blue your "but" is the red you're "therefore" is the green, I'm going to paraphrase it a little bit. You got your five score, years ago it opens with time which is one of the most powerful aspects of a narrative. A great American whose symbolic shadow

we stand today, signed the Emancipation Proclamation. 'And' this momentous decree came as a great beacon light of hope to millions of Negro slaves, who had been seared in the flames of withering injustice. It came as a joyous daybreak to end the long native captivity. So it's over and now we do our contradiction. 'But' 100 years later... again time is really specific, the contradiction begins with time, and what the problem is. The Negro is still not free. One hundred years later the Negro is still languished in the corners of American society and finds himself in exile in his own land. So 'therefore' we have come here today to dramaticize the shameful condition.

So you see how it set it up it's just a normal ABT, which is really, it's really cool, once you start seeing an ABT you get examples of what the ABT is you start recognizing it.  Cinderella has an ABT. The Gettysburg Address, I think, has an ABT. I Have a Dream has an ABT the IMRaD has an ABT. Once you start practicing it's not something that you can learn all in one day but once you start practicing it and recognizing it and using it on a regular basis it becomes easier. So this is just an introduction, and, I know using the words 'and' 'but' 'therefore'
are a great setup to narrative and to build a story but that can get a little repetitive, therefore there are other words to use. We have 'also', 'despite,' 'so'. 'Equally', 'however', 'thus'. 'Identically', 'yet', 'consequently'. So lots of different words that we can use, that still follow the agreement, contradiction, consequence, kind of, following way.

All right so that is what I have on the ABT. Now I want to talk to you a little bit about my current research and I will start with sharing with you my personal ABT with you. So zooplankton make up the base of the marine food web and they support the fish stock in Alaska's fisheries which generates 12.8 billion in economic output. But zooplankton may be becoming smaller and less abundant as climate changes and oceans warm. Therefore more resources need to be invested into using zooplankton as ecosystem indicators in order to predict the health of our fisheries.

So this is why my research is important and now I want to talk to you about what I'm specifically doing. Doctors David Kimmel and Jan Ohlberger recently received funding from NPRB to look at the effects of global warming on long-term changes in copepod size in the Gulf of Alaska, around line eight and the Bering Sea around M2. And today I'm going to share with you how we're going about answering this question. So first we started by identifying cold and warm years to begin our observations and obtain our preliminary results. We chose 2003 to 2005 to represent warm years and 2010 to 2012 to represent cold years.

The next step was the collection of images. I searched the archive room that we have for samples of interest and then bring them into the lab for imaging. For each year I collect images of 10 different species of copepod. Some are characteristically large and some are characteristically small. And for each species I'll identify an image of, I'll identify an image of 30 individuals at each of the seven stages so that's about 210 images each year for each species and then I save each image with a specific naming convention that's later used when determining size. I also manually measure about 10% of all images collected at the microscope and save it in the file name.

Uh, during an AI hackathon we participated in last year we met Senior Solutions Architect Ryan Simpson. And since then we've worked together to develop a standalone MATLAB Application. And this application automatically measures prosome length of copepods and from the images collected under a microscope.

The application starts by binarizing the image and turning it into black and white. And then it creates a mask or it detects the object of interest and in our case the prosome of the copepod, which excludes the tail, the legs and the antenna. And then filters the image which is a method used to reduce the noise or enhance the edges of an image and in machine learning. And then the application measures and identify prosome length in pixels. And then it converts the pixels to millimeters based off of a previously calibrated image of a micrometer at different magnifications and the application itself determines the magnification based on the file name.

So the two biggest obstacles in my opinion were getting the most precise calibration of image at each magnification and getting the algorithm to identify only copepods in the image. So this is a graph of an earlier iteration of the algorithm with outliers. The algorithm estimated copepod sizes incorrectly and then identified non-copepod targets. We observed an R-squared value of about 0.76, where each point represents a copepod and the dashed line is the one-to-one ratio line. And the blue line is the fitted linear regression line to the data. The human estimated length in millimeters was obtained from the 10% of the images that I manually measured at the microscope while collecting images. And the algorithm estimated length is just that. It's the length estimated by the algorithm.

Um, now this is the graph of a species Calanus marshallae with the outliers removed. It had an R-squared value of 0.92, so we could say that this algorithm is as good as a human and we're officially in business. We're now working with Brett Shoelson, a principal application engineer with MathWorks to refine and improve our image detection ability to reduce outliers and bump up that R-squared value. The plan is to ultimately create an algorithm that can detect and measure the length of copepod prosomes, with the accuracy of a human and export the corresponding size data to an Excel Sheet.

But wait there's more, in addition to length we're also trying to develop an algorithm that measures the size of a lipid droplet. So this data can be compared to the work that Rika Group is currently doing for us, which is exactly that, measuring lipid droplet size. The capabilities of our AI image detection algorithm is ever growing in theory for now.

So, goals for this project going forward are: producing a MathWorks blog post about a private/public partnership and developing this application, to continue to expand the capabilities of our AI Imaging algorithm, so it can be used across groups here at NOAA. We're hoping that we're able to beef this up and and pass it over to the Ichthyoplankton Team to do a little study on eggs. So yeah we're really excited about that and we plan on presenting our initial size results at the upcoming Alaska Marine Science Symposium, so I hope to see some of you there and we hope to get some future publications out of this research so this is all thanks to funding from NPRB. Thank you very much.

[Applause]
[Emily Lemagie] Alright, thank you, Deana. 

Our next speaker is Alison Deary. She's a Fisheries Biologist at NOAA's Alaska Fisheries Science Center. Her research focuses on early life history stages of fishes from the Northeast Pacific Ocean, Bering Sea and the Arctic.

[Alison Deary] Okay, good morning and thank you so much for joining me today. So, I will be dovetailing off of what Deana's already spoken about and we're still going to focus on zooplankton but narrow it down to a particular class of zooplankton, known as Ichthyoplankton. These are the planktonic stages of fishes, mainly their eggs, their larvae, but in some cases their juveniles as well. Similarly to what Deana talked about in her ABT talk, we also contribute Ichthyoplankton indicators to the fisheries management process here at the Alaska Fisheries Science Center.

We do this through three main deliverables, the first being the Ecosystem Status Report which has a regional focus. The second being the Ecosystem and Socioeconomic Profiles which have a species and more specifically a stock focus. And then finally we also are included as a set of considerations that all in a risk table that all stock assessment authors must fill out that details if and when changes are made to the acceptable biological catch that they recommend.

However, for today's story I want to focus less on how we apply Ichthyoplankton data to studies of the ecosystem and talk about how we, sorry, thanks, how we actually identify them in our samples.

So for those of you that aren't familiar when we collect plankton we bring it back to a lab and it comes into a plankton jar similar to what's pictured here. And unfortunately for us they don't come pre-sorted, that would be really lucky. So what we then have to do is process the plankton sample and remove out of the early stages of fishes and just to remind you that includes their eggs, larvae and juveniles are associated with plankton. So what we get from the sorting process are individuals that we've removed from the sample in these rough groupings. And we now need to classify them into taxonomic groupings, preferably into species but that's not always the case. The question is how do we actually do this. How do we assign these individuals into species, considering that they look very different from the adults due to their small size and adaptations to surviving at small sizes in their environment.

Well we use a combination of tools. Firstly we start with microscopes that allow us to magnify the individuals and highlight their features that we can use to identify them that distinguishes them from other individuals. We also use a combination of laboratory guides and illustrations, to assist us when we're looking at the microscope.

What's really neat about being in the Ichthyoplankton Team of EcoFOCI is that back in 1989, we literally wrote the book that sets the baseline information of all of the taxa that we currently are able to identify, in our large marine ecosystems. Within this guide our illustrations organized into what we call developmental series, which this example picture here is for Walleye Pollock and a developmental series are sets of illustrations that highlight progressively later stages, and they highlight the key characteristics that help us distinguish one individual from another individual within a different species.

Now for Walleye Pollock this is a pretty complete developmental series we have an illustration of a late stage egg, a newly hatched larvae through progressively larger larval stages and into the juveniles. This is not often the case for many species. We have what we call gaps in the Developmental Series, where certain stages we don't have yet descriptions for or illustrations. Therefore, the mission that we conduct as part of the Ichthyoplankton Team is to continue advancing our taxonomic resolution our ability to make reliable species level identification.

And we do this through continued plankton monitoring of our large marine ecosystems, which include the Gulf of Alaska, the Bering Sea, and the Arctic. What gets really exciting for us is when we get to sample either a new ecosystem or an ecosystem that we monitor but a different time of year because that provides fresh material of potentially species we don't often encounter, as well as stages we don't see during our normal monitoring efforts.

And what I have here are a collection of illustrations that are part of a recently submitted manuscript, that put forth some advances that we've made into our developmental series of several species. And we've been able to fill gaps in the developmental series of these species either by replacing poor quality illustrations with newer better quality illustrations, as well as in the case of Arctic Cod, adding a stage that was previously undescribed and not illustrated.

Now once we have these gaps filled in my developmental series we then contribute them to our laboratory guide but unlike back in 1989 with a static guide, we have migrated this to an online format known as our Ichthyoplankton Information System. And what that does is it allows us to disseminate widely these taxonomic advances, as well as in a timely manner. And that contributes to our ability to present reliable species level data in our area in a timely manner that produces robust data sets that then can be contributed to support sustainable fisheries in our area and examine questions of community resilience, under climate change scenarios and with that I'd just like to thank the co-authors, the Plankton Sorting and Identification Center in Poland, for all their hard work, and you for your attention.

[Applause]

[Emily Lemagie] Next we have Ned Cokelet, also known as the Saildrone Data Guru. He is a Physical Oceanographer at NOAA's Pacific Marine Environmental Laboratory. Ned advises on Saildrone instrumentation and the variables they measure among other topics.

[Ned Cokelet] Thank you. So this is not much of an EcoFOCI topic, it's about wave measurements from saildrones. I volunteered to be one of the guinea pigs for the ABT, the first, so far this year. So we have these saildrones, right here. We've been developing them in 2015, they used to look like this: taller than the present one, but in 2021 we, working with Saildrone Incorporated made smaller saildrones, with lower wings, and more rugged and we put them out in hurricanes in the Atlantic.

So, this is a camera on a saildrone during Hurricane Sam, in Atlantic in September 2021, September 30th. And you can see it's a place where you really wouldn't want to be yourself, if you're out there because it's pretty rough.

Not a great place, you know I always thought maybe we should just send some NOAA Corps officers out in Zodiacs
but that's probably not a good idea. [Laughter]

So, we put them out, we measured the... this is a map here, this is the hurricane track right here, and this is a a circle here of the closed pressure contours. 
[Interrupted by loud beeping]
[Beeping stops]

And various counters of velocity as you finally get into the center of the hurricane and this is at the time when the hurricane was closest to the saildrone. The saildrone is that little symbol right here into the center of the hurricane, at the same time, well, before that time, there was a NOAA data buoy over here, the hurricane passed by
[indistinct] NOAA data buoy [indistinct] saildrone.

What I'm showing here is the wave spectrum first measured at the NOAA data buoy in green, and the saildrone in red. What this really means is this is, time here in September, October 2021. This is the height, the wave height, of the significant wave height of the waves.

And then, try here, so this significant wave height, it's the height of the average height of the highest

one-third waves, but the saildrone itself measures individual waves. These are wave profiles here of 15 minutes, over here of the mouse (there we go) 15 minutes of record here waves and you can see that this wave right here is 27 meters high, that's a pretty big wave.

And if you looked last summer at the literature and newspaper there were some people in British Columbia who put out a measurement buoys to measure waves, rogue waves, of the Pacific and they measured what they thought was the highest wave ever measured. Well we also measured that was 27 meters and we also measured a 27-meter-high wave here in the hurricane. There's a difference though they were looking at rogue waves which just appear out of nowhere but not in the storm. Well these are measured in a storm so there's a difference in what's going on.

Okay, great. So we measure these waves, but how good, how good are the measurements? So a normal wave measuring device like this here is a wave buoy. This is the same, same wave buoy that was in Europe that was hurt, the saildrone, and you'll notice
that they're circular or discus-shaped. But some wave buoys are actually pointy, they look like ship hulls, and our saildrones are also pointy they look like ship hulls. So this is the standard way you measure waves but this is the way we're measuring waves with a hurricane, so how good is that?

Well, one way to determine that is to look at, what's called the wave spectrum here. I'm going to show you a series of plots. This is the spectrum of the waves measured at the saildrone and this is the spectrum of the waves measured by NDBC buoy.

And basically, what this says is we've got time down here lower on the x-axis and energy of the waves here so the more energy inside frequency of the waves here and the energy is in color bars here. And the more energy there is the bigger the waves are and you can see here that here we the spectral peak of saildrone, of the cloud saildrone and the spectral peak at the NDBC buoy and this is the difference between the two.

So first off I just want you to notice that they look pretty similar to each other, so the saildrone and the buoy, basically
you're seeing the same kind of wave field here in a statistical sense.
So how good is that in general?

Well, we had times where other saildrones sailed near NDBC buoys kind of fortuitously but sometimes we actually did an experiment. So I'm just going to show you some of those results here. So always will be the saildrone spectrum on the left and the buoy spectrum on the right and it's just a visual comparison that's really not much to see.  The difference is down here but that's not so important. So you can see that in this instance here the saildrone was seeing much the same thing that the NDBC buoy is seeing there.

It's giving me good confidence and the difference areas I can't quite see this, the water depth is like 30 meters here and the water depth was 30 meters and the saildrone and the NDBC buoy were a kilometer apart but the catch is the saildrone is out in deep water because you couldn't bring the saildrone in at 30 meters of water because it could run aground.

So it's a comparison but it's not quite the right comparison and I should have said in the earlier one where the saildrone was measured with the NDBC buoy, they were about 200 kilometers apart so it's not a good check on the ability to measure these two things.

So if we put on the list here, here's a case where the saildrone out, it's only a couple kilometers apart but still in shallow water in this buoy and you can see though visually they're much the same thing. The two separate peaks are about the same, the buoy is seeing a little bit higher waves than the saildrone.

And again in that case now we're still closely showing four kilometers but they're still in fairly shallow water so the buoys and shallow water but the saildrones are in the deep water, the waves change as they come into shallower water so it's really not a fair apples to apples comparison.

But you can still see the spectral peaks are about the same and the look of the spectrum is not the same. Now also within three to ten kilometers and still shallow water but again we're seeing the same phenomena in both places.

Here also a few kilometers away but now we're in deep water so now the saildrone is seeing this and the wave buoy is seeing a little bit different but still visually much of the same compression of the wave spectrum.

And now here's a case where they look basically the same but these waves are really tiny here in August 2021. In deep water but there was no real storm going on during this time but the saildrone happen to be by the buoys
so we used these measurements.

In the last case here again visually very much the same. The wave buoys saw a little bit higher waves than the saildrone did. One of our case again doing the same. So the conclusion is how well can saildrones measure waves?

Well, reasonably well but the precise validation requires longer observations at different wave lengths and direction. So if you're really going to use a saildrone as a substitute for a wave measuring device at NOAA, you'd really want to do some tests where you put the saildrone down for several months at a time, sail around a buoy with all kinds of different wave conditions. Storms, no storms, a saildrone going in different directions because it is pointy that is, it doesn't respond the same way. It's something yet to be worked out in the future.

Let me just go back to this is the slide that spurred this whole thing going and you can see by the way to subtract these two spectrum and you'll see that the waves of the saildrone is higher at this frequency and lower at that frequency, and that's because you see the spectrum has moved out of the waves, are actually changed in this 200 kilometer shift from one place to the waves from the buoy here to the saildrone they got lower frequency and that's because the way the wave shape is developing. Anyway, that is the end of my topic.

[Applause]

[Emily Lemagie] Yeah, thank you to our speakers. And now we have some time for questions and discussion. The speakers in the room, why don't you come to the front. Sorry Ned. And Deana is going to join us online.

[Deana Crouser] Hello.

[Emily Lemagie] We have one question from the chat, for Ned. What is the benefit of a saildrone
versus a buoy?

[Ned Cokelet] The benefit of a saildrone versus
a buoy, first off the buoys are maintained by the National Data Buoy Center and they are rather expensive to keep out there and to maintain and sometimes they fail and this year two of them failed in the Gulf of Mexico. They use our saildrones as substitutes for their buoys to see if they could use a saildrone to make measurements. So there's other things besides that the buoy doesn't do. Base measurements, more measurements in the water and the atmosphere a buoy does, it has the ability to move around, which we were using in the hurricane to move out just to intercept the hurricane which the buoy doesn't do.

That's the difference between them. We are thinking about perhaps

using saildrones in some places where there are currently buoys now.

[Emily Lemagie] Do we have any other questions in the room?

[Participant] I have a question for Ned.

[Emily Lemagie] Can you come up to the front? Just, we're doing the hybrid thing. We want others to be able to hear. Thank you.

[Participant] Is that close enough?

[Emily Lemagie] Yeah.

[Participant] So my question was about is the difference between the buoy's ability and the saildrone's ability to record this data, the motion that the saildrone is under?

[Ned Cokelet] Probably. It's the fact that it's got - it's pointy, it's got a longer hull, is longer than it is wide. I don't exactly know the reason behind it and you know we don't really know which is the better measurement. Although the NDBC buoys with their discus shape have been around for a long time so that's kind of a standard.

So it's probably the fact that the elongated hull responding differently to the waves than the discus is and again that we really require a months-long measurement of the two interacting together the whole time. That still needs to happen. The very first time trying to get some idea that they measure waves at all, and the answer is: yes and pretty well.

[Emily Lemagie] We have a question from the chat. From Frank Hernandez. Regarding the MATLAB [indistinct], are there preservation issues that affect the lipid count?

[Deana Crouser] Um...That's a great question! We haven't gotten to doing the lipids yet, but the answer is yes. Thank you Dave. Ha ha!

[Emily Lemagie] Yes, Shaun?

[Shaun] Yeah, I'll come up.

[Emily Lemagie] Thank you. The camera's also that way.

[Shaun] My question is for Ali. So I hear all about the MATLAB facility that is Poland. But I have no idea what they're actually doing there, in order to take that lovely jar of slurry and to get some distribution of fishes, I'm assuming.

So it's a two, two point question.

One is can you elaborate just really quickly on what Poland's doing. And two is there an instance or an application for Deana's work with imagery, to help improve data
that's coming out of Poland's laboratory?

[Alison Deary] Yeah thank you, excellent question. So yeah as Shaun alluded to, so those of you who are not familiar, we are really lucky to have the Alaska Fisheries Science Center along with the Northeast Fishery Science Center, and the Southeast Fisheries Science Center, in that that plankton jar in that picture, we don't actually process that in-house which saves a lot of our manpower because that means we can focus on quality assurance and research when we get the data back.

So what happens in Poland is once we come back to the boat we have our boxes of jars, we send all of that over to Poland. And there is an enormous team of both zooplankton and ichthyoplankton ecologists and they go through and they sort through the plankton samples. And so from our side they first do, the processing for ichthyoplankton and they'll remove all of the ichthyoplankton that they find they will also do some resource on a subset and if they find more than 10% of the ichthyoplankton was missed they resort it again, so that we have as good as possible quantitative data from that net and we've also collected measurements of how long the net was in the water as well as how much water was filtered so we can get a quantitative assessment for that chunk of water we sampled.

How many fish eggs are being juveniles were collected. That same process gets repeated for zooplankton except they don't do the whole sample they do an hour clock because there's a lot of zooplankton and nobody wants to count 10 million copepods. We can extrapolate that so once they have the, ichthyoplankton separate from the sample they then run through and they do those steps that I detailed. They use laboratory guides that we generated, as well as auxiliary help with developmental series and illustrations we sent to them. They do all of the identifications.

They enter it into our database and it's this amazing contraption that comes back that, Kimberly our database manager, ensures everything looks good and then like I said we follow up with quality assurance and there's a lot of taxa that they do really really great, there's some taxa that they don't do really really great and that's partly because we don't yet have great diagnostic tools to help with that.

There's still questions on how species are related to each other. What they actually are to that level. Going back to the second part of your question, do we see imaging as contributing to that?
Yes.

The one issue that I see with that with ichthyoplankton is we rely on pigment characters. And so that gets a little bit difficult to look at at the resolution we want for a management perspective. Like, we can throw out an imaging system and I've worked, I see Frank's on the call with Frank Hernandez back in the Gulf.

We had what we call ISIIS, the Institute Ichthyoplankton Imaging System, it's towed behind the boat. It gets wonderful pictures that they also fish at Oregon State University and several other places. But the issue you have is you don't get the same taxa resolution. Some images you can classify to species, but what I fear for our ecosystem is the species we want to know a lot about are the gabots and we need to be able to identify Pacific Cod from Walleye Pollock.

An image won't necessarily get you that level of detail because you have to have the right orientation to know if the pigment patterns are that Pacific Cod or a Walleye Pollock.

So that's one of the limitations on the larval side but where I see the imaging really being advantageous for us is co-opting our egg database. We are the only center that identifies fish eggs or asks Poland to identify fish eggs, let me be honest, and it's really really cool. But we don't really do that much with the egg data. And part of that's because how we identify the eggs for a lot of the times is we have a diameter range and eggs falls into this diameter range instead of a particular season. Small Pollock doesn't follow that family range, it's a different species. What we're not tracking are individual lengths of the eggs and that's where I see the imaging coming in is that we can use that algorithm to assist in the sorting of the eggs we can take a picture of eggs and here's the quantitative data these ones are in this diameter range therefore they're Pollock.

But we can also then get lengths or diameters of those eggs and then we can start to ask ecological questions. Is egg size changing over time as climate warms? There's been work on gabots in the Atlantic that shows relative size of the egg relates back to the health of the spawning stock. We don't have the ability to ask those questions right now, but with advances in the imaging and having an algorithm to generate those diameters I think we can move that way. And that's really where I see the advantages for the imaging.

So thank you.

[Deana Crouser] Couldn't have said that any better. Thank you, Ali.

[Emily Lemagie]  So I have a question. For the imaging is that shipboard or are there any possibilities or technologies for doing that on mobile platforms?

[Deana Crouser] Yeah, the goal is that it will be used on live animals on board a ship but that is, we're, we're starting with in situ or ex situ, under the microscope images to train the algorithm. So it can identify it and then hopefully we'll be able to do, we have an RZA that we're trying, that we take (rapid zoopointed analysis) that we take on board and hopefully we'll be able to just take a picture of an aliquot and use the algorithm to automatically by, which species are being present based off of the images we've trained it on from Poland.

Who is collecting images for us as well as the images that I collect from our archive room and then it should be able to tell us what what's there and how big it is and everything like that so that way we don't need to send a zooplankton ecologist on every cruise. Anyone can do it, just snap a picture and this algorithm should be able to do it all for you.

[Emily Lemagie] Do we have any other questions from online? For any of our speakers and in the room?

[Emily Lemagie] What's the time outlook for these technologies? I guess for either of you, you know for being able to build up that a database and are is the technology development coming out of Deana's research that you would use for that, and do you have historical archives of aids that you'll be able to apply that to or is that going forward?

[Alison Deary] Yeah, so great question, so I'll jump in first, Deana. And please correct me if I'm missing something. So short answers yes. We are going to co-op all of the amazing work that Deana and Dave have been doing because they have already laid the foundation and the nice thing is eggs are circle so they're pretty easy to/for a computer to be like oh I found an egg!

So that part we're really excited about how I envision this moving forward is we're in a marathon it's not like a short race we're not in a 100 meter dash. But it's a marathon so the first step is that Poland typically puts all of the eggs at a station back together. The first step was requesting them to not do that and keep them separated by species, so that gives us the ability that they have the eggs pre-sorted and they stay sorted. And then we can start next year if we're going to add the imaging for the eggs is my goal.

So this year they're starting to take some pictures of ichthyoplankton, I want that so we can throw that onto our online Ichthyoplankton Information System and additional line drawings, have beautiful pictures. Going back to your question though about like what material we have available, those of us that do collections-type work tend to be hoarders.

So we actually have all of the ichthyoplankton we have ever collected and we have them maintained both in-house and our ichthyoplankton lab. We're always happy to show people what we have and then we've also had a long-standing collaboration with the University of Washington and they do the curation of all of our historical samples. So seven years and beyond.

So what I'd like to see is that become a thing that we start and then move forward and becomes a standard operating procedure, is part of the vision. But also we have the ability to go back and look at historical samples. I'm not necessarily thinking we're going to image every sample of egg, but do approach similarly to what Deana and Dave did and highlight certain years or certain surveys and get some of that coverage so we can look in the past but then make that this process moving forward.

[Deana Crouser] Yes, absolutely and the application is it's working it's generating data. So really all it is is imaging the, eggs or or plankton or copepods or whatever it is you're looking at specifically. So yeah it's it's working and we have two developers that are working on it so we're still continuously refining it and doing tweaks and bugs and stuff like that. So, it's something that's ever moving.

[Emily Lemagie] Questions? For one more question? Alright, thank you and thank you all, the speakers. We look forward to seeing everyone next week at the same time. We'll have two speakers Emily Hayden and Jens Nielsen.

And these two presentations in combination will provide an overview of ocean temperature and sea ice variability, paired with the discussion of its potential impact on phytoplankton bloom timing and zooplankton coming out of diapause
in the Bering Sea. So we'll see you next week.

[Deana Crouser] Thank you!

[Applause]