WEBVTT Kind: captions Language: en 00:00:06.560 --> 00:00:12.400 Well, hello, everyone. Welcome to tonight's broadcast.   00:00:12.400 --> 00:00:17.200 I'm Chris Butler Minor, the community engagement  specialist for Olympic Coast National Marine   00:00:17.200 --> 00:00:23.760 Sanctuary. And I'll be facilitating today's live  presentation in addition to Rachel Brown with   00:00:23.760 --> 00:00:29.920 Feiro Marine Life Center and Cheyenne Palmo,  one of our AmeriCorps service members. She's the   00:00:29.920 --> 00:00:34.720 Environmental Education and Stewardship Specialist  who's also joining us to support the webinar.   00:00:35.360 --> 00:00:39.440 We're so pleased to have you join us for  our New developments for Passive Acoustic   00:00:39.440 --> 00:00:45.920 Monitoring of Sperm Whales in Hawaiian waters." This event is co-hosted by Olympic Coast National   00:00:45.920 --> 00:00:53.440 Marine Sanctuary and Feiro Marine Life Center as  the second installment of our 2021 Speaker Series,   00:00:53.440 --> 00:00:58.720 And the speaker series brings you marine focused  topics of interest from subject matter experts.  00:00:59.680 --> 00:01:04.000 During the presentation, all  attendees will be in listen only mode.  00:01:04.000 --> 00:01:08.720 But you're welcome to type questions  for the presenter in the question box,   00:01:08.720 --> 00:01:15.440 which is going to be on the lower part of your  control panel, probably docked on the right,   00:01:15.440 --> 00:01:20.560 right-hand side of your screen. Um, it's the same area that you   00:01:20.560 --> 00:01:25.360 can let us know about any technical issues that  you might be having. And so we'll be monitoring   00:01:25.360 --> 00:01:29.520 incoming questions and technical issues and  we'll respond to them as quickly as we can.  00:01:30.400 --> 00:01:35.600 We're recording this session and will share  the recording with registered participants   00:01:35.600 --> 00:01:40.880 through e-mail and then also on the Office of  National Marine Sanctuaries Webinar Series page   00:01:40.880 --> 00:01:48.080 once it's been closed captioned. And just the last little bit of housekeeping,   00:01:48.080 --> 00:01:50.720 we are really interested in any feedback that you   00:01:50.720 --> 00:01:54.160 might have that will help us improve  the experience. So if you don't mind,   00:01:54.160 --> 00:01:58.640 please take a few minutes to complete a  short evaluation after the presentation.  00:02:01.040 --> 00:02:07.120 So I'd like to kick us off by having Cheyenne host  a quick Slido poll. And if you're not familiar   00:02:07.120 --> 00:02:13.120 with it she’ll give you all the scoop on it. - Awesome! Yeah, So Chris, if you can go   00:02:13.120 --> 00:02:15.600 ahead and share your screen to disappear on us.  00:02:16.880 --> 00:02:22.240 And just re-share. There's going to be a QRC  code that's going to pop up on your screen.  00:02:22.240 --> 00:02:28.000 Um, and what you can do is just take  your smart phone and open up your camera   00:02:29.280 --> 00:02:37.920 and go ahead and scan that QRC code. And I'll bring you to be launched question.   00:02:38.560 --> 00:02:43.840 The first question is, have you ever heard  of NOAA's National Marine Sanctuaries before?  00:02:43.840 --> 00:02:46.240 So I'm going to go ahead  and activate that right now.  00:02:48.400 --> 00:02:53.920 And we'll see if we get some votes in.  We'll do, like, 30 seconds worth of 00:02:55.040 --> 00:03:03.840 figuring out, if you have heard of the  Office at National Marine Sanctuaries.  00:03:08.640 --> 00:03:12.160 And it's perfectly OK if you don't have  a smartphone, there's the option to   00:03:12.720 --> 00:03:16.720 join via a web browser at  slido.com and just click,   00:03:16.720 --> 00:03:21.200 and just type, then that number. There's only three questions and they're   00:03:21.200 --> 00:03:27.200 not crucial to the presentation itself. So, no worries if you can't participate.  00:03:29.600 --> 00:03:34.080 Alright. So I'm going to go ahead and stop this.  00:03:34.080 --> 00:03:43.920 You got 66 answers and an 85% of participants  said that they had heard of NOAA’s National   00:03:43.920 --> 00:03:46.720 Marine Sanctuaries before. Back to you Chris.  00:03:47.280 --> 00:03:54.960 -Cool, OK. So let's go on then.  00:03:56.720 --> 00:03:59.840 Let's see. We seem to have, there we  go. Just having a little delay there.  00:04:00.800 --> 00:04:06.720 So. So it sounds like a lot of you are familiar with   00:04:06.720 --> 00:04:10.800 it. But for those of you that are not familiar  with the National Marine Sanctuary, in a nutshell   00:04:11.680 --> 00:04:19.360 there are 14 Sanctuaries plus Papahānaumokuākea  and Rose Atoll Marine National Monuments, all of   00:04:19.360 --> 00:04:24.880 which protect over 600,000 square miles of ocean.  So that's bigger than the entire state of Alaska.  00:04:25.920 --> 00:04:30.560 Activities at these sites include resource  protection, recreation and tourism,   00:04:30.560 --> 00:04:36.160 education and outreach, maritime heritage, plus  science and exploration. So NOAA’s national   00:04:36.160 --> 00:04:40.880 marine sanctuaries, are established and actively  managed by the agency under the National Marine   00:04:40.880 --> 00:04:46.080 Sanctuaries Act to sustain appropriate, sustain  appropriate ocean uses, including recreation.  00:04:48.720 --> 00:04:54.960 So, we're going to do another quick  Slido poll, same QRC code, and   00:04:56.400 --> 00:04:59.040 event ID. So take it away, Cheyenne.  00:05:00.400 --> 00:05:03.920 -All right. So, now that you kind  of saw a map, have you personally   00:05:03.920 --> 00:05:10.240 ever visited a National Marine sanctuary? Um, and if you'd like to, you can answer your   00:05:10.240 --> 00:05:15.360 question in the question box. If you don't  have access to the Slido, that's fine.  00:05:21.280 --> 00:05:26.080 Right. I like the “maybe” option here. A  lot of the time, you don't even realize that   00:05:26.080 --> 00:05:37.840 you're in the National Marine sanctuary. Alright, I'll give it like 10 more seconds.  00:05:40.640 --> 00:05:50.240 OK, so about 75% of everyone said, ya have been  to a marine sanctuary, 15 said no, and a strong,   00:05:51.280 --> 00:05:54.400 11% said maybe. Back to you, Chris.  00:05:56.560 --> 00:06:05.840 -Interesting, OK, well, hopefully we'll  inspire you to, to change that answer to, Yes!  00:06:07.600 --> 00:06:12.800 So, um, I just wanted to do that  by showing you some amazing photos.  00:06:13.360 --> 00:06:17.040 Like let's take a quick tour of some  of the treasures that are protected   00:06:17.040 --> 00:06:22.400 and National Marine sanctuaries,  such as these iconic sea giants or,  00:06:26.080 --> 00:06:29.680 the small reef fish and corals  found on the East Coast.  00:06:32.000 --> 00:06:36.080 National marine sanctuaries seek to preserve  the beauty and biodiversity of these special   00:06:36.080 --> 00:06:42.160 marine places while also providing a place for  charismatic marine creatures to seek shelter.   00:06:42.800 --> 00:06:46.480 And these special places also  protect some of the nation's most   00:06:46.480 --> 00:06:51.600 iconic marine and cultural heritage  sites by establishing national marine   00:06:51.600 --> 00:06:55.120 sanctuaries that protect shipwrecks  and other significant artifacts.  00:06:58.080 --> 00:07:04.960 They're also mandated to conduct research,  monitoring, resource protection, education   00:07:05.680 --> 00:07:07.600 and outreach, and, of course,   00:07:07.600 --> 00:07:11.600 management of America's Underwater Treasures  to preserve them for future generations.  00:07:15.280 --> 00:07:18.320 These places are protected for  your enjoyment and pleasure and   00:07:18.320 --> 00:07:26.080 sustainability for generations to come. We're going to go to our last Slido.  00:07:29.040 --> 00:07:30.480 And Cheyenne, if you don't mind,   00:07:32.080 --> 00:07:39.040 posing this question? -Of course. So just to  get a general idea for how large the sanctuary   00:07:39.040 --> 00:07:44.240 system actually is, how many square miles  do national marine sanctuaries encompass?  00:07:48.960 --> 00:07:57.840 So far, we're erring on the larger  numbers here, which is promising.  00:07:59.280 --> 00:08:04.880 I'm usually shooting for about 75 participants. Keep hitting it.  00:08:09.120 --> 00:08:19.440 OK, so a majority, a little more than 50%, shot  for the 1,425,000 square miles as large number,   00:08:20.720 --> 00:08:25.840 and then closely following, behind,  40%, said, 600,000 square miles.  00:08:26.800 --> 00:08:29.600 So what are we thinking Chris? What  is the answer to that question?  00:08:31.360 --> 00:08:37.920 -The answer is 600,000 square miles.  If you recall, I mentioned that it was   00:08:37.920 --> 00:08:41.760 a little bit bigger, the ocean area and  Great Lakes areas that are protected,   00:08:41.760 --> 00:08:47.600 are a little bit bigger than the state of Alaska. All right, so let's move on.  00:08:48.320 --> 00:08:50.720 And thanks for participating  in those polls, by the way.  00:08:52.480 --> 00:08:55.840 So, in tonight's talk, we're going  to take you from our hosts’ location   00:08:55.840 --> 00:09:00.240 of Olympic Coast National Marine Sanctuary. to Papahānaumokuākea Marine National Monument,   00:09:00.240 --> 00:09:05.680 which is further west into the Pacific and  is the largest contiguous, fully protected   00:09:05.680 --> 00:09:10.960 conservation area in the U S and currently the  largest marine conservation area in the world.  00:09:12.640 --> 00:09:16.800 And to Hawaiian Islands Humpback Whale  National Marine Sanctuary, which was created   00:09:16.800 --> 00:09:23.600 to protect humpback whales and their habitat. So, now, I'd like to turn the mic over to   00:09:23.600 --> 00:09:29.040 Rachel Brown, the Education Manager for Feiro  Marine Life Center to introduce our speaker.  00:09:34.160 --> 00:09:39.200 -Hello, I’m Rachele Brown, education  manager at Feiro Marine Life Center located in  00:09:39.200 --> 00:09:46.560 Port Angeles Washington. Feiro Marine Life  Center is a community aquarium that was founded in 1981  00:09:46.560 --> 00:09:55.680 by a community leader and educator, Arthur D Feiro. Feiro is committed to community on education and specializes in   00:09:55.680 --> 00:10:02.880 local marine invertebrates and fish. Thank all of you for joining the partnered 2021 00:10:02.880 --> 00:10:06.080 lecture series. Today, we're joined by Dr.   00:10:08.000 --> 00:10:13.200 Yvonne Barkley. Yvonne recently earned her PhD  from the University of Hawaii at Manoa   00:10:14.000 --> 00:10:20.160 in December, where she worked on developing new  analytical methods for passive acoustic data   00:10:21.840 --> 00:10:27.680 and identifying patterns of false killer whales and  sperm whale populations in Hawaiian waters.  00:10:29.360 --> 00:10:33.040 She learned about cetacean  acoustics after graduating college,   00:10:34.800 --> 00:10:41.200 working as an acoustic analyst. She was intrigued by the entire process,   00:10:41.200 --> 00:10:44.880 from designing hydrophones for data collection   00:10:44.880 --> 00:10:51.120 to the computer programming for analysis. At the time, she was just clicking boxes,   00:10:51.120 --> 00:10:56.320 around squiggly lines on a computer screen. but  she appreciated her role in the bigger picture   00:10:56.320 --> 00:11:01.760 where science and technology are combined  to learn more about the patient populations   00:11:02.320 --> 00:11:08.640 and the impact that humans have on them. This concept has driven Yvonne's interest  00:11:09.440 --> 00:11:15.360 through her career and lead to her work with NOAA Fisheries. She's started as an intern in 2009, leading to 00:11:17.600 --> 00:11:20.320 Technician position, and  then on to graduate school, intermittently 00:11:21.520 --> 00:11:27.840 going to sea for several NOAA cetacean research  cruises. Over the years, Yvonne worked with others   00:11:30.160 --> 00:11:33.280 sharing her same motivation, which  continues to inspire her.  00:11:34.240 --> 00:11:41.760 She's currently in a new role as a postdoctoral researcher  under the Joint Institute for Marine and   00:11:41.760 --> 00:11:51.840 Atmospheric Research. When Yvonne isn't analyzing acoustic data, you can find her mostly outside enjoying  the weather in Hawaii 00:11:51.840 --> 00:11:57.040 either surfing, hiking, or simply lounging. It is  my pleasure to introduce to you,   00:11:57.040 --> 00:11:59.310 Doctor Yvonne Barkley. 00:11:59.959 --> 00:12:01.920 Well, Cheyenne, maybe you can take a look and see if   00:12:01.920 --> 00:12:12.240 there are any questions in the question box right now? - OK I was just about to let you know...  00:12:14.240 --> 00:12:18.640 So just as a little update, to people who  are joining the call right now. We seem to   00:12:18.640 --> 00:12:23.680 have lost Yvonne Barkely, who's the presenter for today.  00:12:23.680 --> 00:12:30.240 I can only imagine that Internet is not  the strongest where she is calling in from,   00:12:30.800 --> 00:12:37.840 but a few people were asking Chris: If  and where this recording will be available   00:12:38.800 --> 00:12:42.000 after being closed caption? -Oh, great.  00:12:43.040 --> 00:12:49.200 So it's going to be on the National Marine  Sanctuaries Webinar Series page, so you can go to   00:12:49.200 --> 00:12:56.400 sanctuaries.noaa.gov and [Oops, let me  make sure that's the right address].   00:12:56.400 --> 00:13:02.640 And if you look for webinar series, or even  if you type in sanctuary webinar series,   00:13:02.640 --> 00:13:10.160 you should be able to find it. And it's actually  got a wonderful library of past presentations.  00:13:11.840 --> 00:13:19.840 Oh, yay! Yvonne has made it  back from the internet abyss.  00:13:21.040 --> 00:13:25.520 -Alright, I just made you the presenter,  Yvonne. So you should be able to   00:13:25.520 --> 00:13:30.240 pull up your presentation. Thank you everyone for your patience.  00:13:33.040 --> 00:13:39.920 -Let's see? OK, turn myself off.  00:13:42.080 --> 00:13:45.600 Great, I'm assuming I was  introduced. Sorry about that.  00:13:46.400 --> 00:13:53.680 Slight glitch, the internet connectivity. Uh, OK, I will present.  00:13:58.080 --> 00:14:04.480 and Close that. Sorry, one moment.  00:14:04.480 --> 00:14:10.080 I am now working on one monitor versus  two. It's a little bit different.  00:14:14.880 --> 00:14:15.840 Mmm hmm hmm.  00:14:21.120 --> 00:14:24.040 OK, we're just gonna go with it. 00:14:27.680 --> 00:14:29.840 Great.  00:14:30.800 --> 00:14:35.600 Move that down here. OK, thanks,  everybody, thanks for bearing with me here,   00:14:35.600 --> 00:14:41.840 and tuning in to my talk, where I will be  discussing a couple of projects that I recently   00:14:41.840 --> 00:14:46.240 completed for my dissertation research,  and thanks, Rachele for introducing me.  00:14:49.360 --> 00:14:53.040 OK, things can work, that  would be great. Excellent.  00:14:53.040 --> 00:14:57.680 So we'll start with a little bit of an  outline just to show you a roadmap of   00:14:57.680 --> 00:15:02.960 where we're headed today and starting  with some background information of   00:15:03.840 --> 00:15:08.080 the Cetacean Research Program at NOAA's  Pacific Island Fisheries Science Center   00:15:08.640 --> 00:15:13.600 where I've worked with them throughout my  dissertation as well as in my current position.  00:15:17.200 --> 00:15:21.280 Let's see, along with, OK, there we go.  00:15:23.440 --> 00:15:26.720 Sorry about this. Along with the   00:15:26.720 --> 00:15:32.000 study area where my data were collected and  where all these projects are based off of,   00:15:33.280 --> 00:15:38.640 just going over, also some of the main research  methods that we use to study cetaceans in general.  00:15:39.280 --> 00:15:45.120 And then of course talking about sperm whales,  some of the relevant biological and ecological   00:15:45.120 --> 00:15:51.120 information that helped to drive my studies. Then moving into the two different projects that   00:15:51.120 --> 00:15:57.760 I worked on, including localization of sperm whales,  and then also incorporating, incorporating that   00:15:57.760 --> 00:16:04.000 acoustic data into species distribution models. And then talking a little bit about some   00:16:04.000 --> 00:16:07.840 future and current directions that  we're going with these types of data.  00:16:09.520 --> 00:16:15.440 So, to start with, the mission of the Cetacean  Research Program is generally to study the   00:16:15.440 --> 00:16:21.040 whale and dolphin populations or cetacean  populations within the Central and Western   00:16:21.040 --> 00:16:26.560 Pacific Ocean. And that is done in order to  support conservation and management efforts.  00:16:27.440 --> 00:16:33.200 All of this research is really guided by  mandates of the Marine Mammal Protection Act,   00:16:33.200 --> 00:16:36.320 which is what all cetacean  species are protected under that  00:16:36.320 --> 00:16:40.800 as well as some of the endangered species  protected under the Endangered Species Act.  00:16:42.960 --> 00:16:46.000 So, the study area includes  the Hawaiian archipelago,   00:16:46.800 --> 00:16:52.960 shown here, which is actually home, to at least  25 different cetacean species that we know of.  00:16:53.600 --> 00:16:58.560 And this includes the north-west Hawaiian  Islands, the Hawaiian Islands up in the   00:16:58.560 --> 00:17:01.920 north-west. I wish I had my little  walk, we can hopefully see my little   00:17:01.920 --> 00:17:08.080 arrow here. So this area which encompasses the Papahānaumokuākea Marine National Monument,   00:17:08.720 --> 00:17:13.760 as well as the Main Hawaiian Islands, which  I am located on here, hopefully with the   00:17:13.760 --> 00:17:19.600 good Internet connection, on Oahu in the middle  of the Pacific. And this encompasses the five   00:17:19.600 --> 00:17:24.640 separate areas of NOAA’s Humpback Marine  or National Marine Sanctuary, as well.  00:17:27.440 --> 00:17:32.400 So, to just chat a little bit about the  different tools that we use to study these   00:17:32.400 --> 00:17:34.480 species that primarily live underwater.  00:17:35.680 --> 00:17:43.040 A lot of these are complimentary to each other,  and most of them are have really established   00:17:43.040 --> 00:17:46.480 the current knowledge of what we know  about Cetacean species worldwide.   00:17:48.000 --> 00:17:53.680 OK. Starting with visual observations. This is kinda the most basic tool that we   00:17:53.680 --> 00:17:58.560 have, to just see animals at the surface. And that can tell us a lot of information   00:17:58.560 --> 00:18:04.160 about what species are present, how many there  are, and a number of other pieces of information   00:18:06.320 --> 00:18:11.040 that includes photos and being able to take  photos of these animals at the surface. And   00:18:11.040 --> 00:18:18.480 we perform then photo identification analysis  which can take the images of their dorsal fins  00:18:18.480 --> 00:18:24.240 and flukes, which have sometimes have different  scratches and markings and niches in them,   00:18:24.240 --> 00:18:28.880 and notches rather, to use them as  sort of fingerprints to identify   00:18:28.880 --> 00:18:34.880 individuals and even estimate how many  animals are in a population, using the photos.  00:18:35.680 --> 00:18:41.920 We can also collect tissue samples, which allows  us to run genetic analysis to look more at the   00:18:41.920 --> 00:18:48.560 population structure of the species and tell  what animals are related to each other between,   00:18:49.680 --> 00:18:52.960 you know, even different geographic  regions or within the same region.  00:18:55.200 --> 00:19:01.040 We can also deploy tags and track animal  movements using satellite telemetry data.  00:19:01.040 --> 00:19:04.960 This is an example of a tag within  a dorsal fin of a pilot whale.  00:19:04.960 --> 00:19:07.200 And this really gives us a good idea of   00:19:08.480 --> 00:19:15.680 their habitat range and where they are within this  either study area or just where they tend to go off,   00:19:15.680 --> 00:19:20.240 you know, hundreds of kilometers away from where  you're thinking they normally would reside.  00:19:21.600 --> 00:19:29.760 And lastly, a newer approach, at least to me, I have  not ever used an unmanned aerial vehicle or a UAV.  00:19:30.320 --> 00:19:36.800 And this is your, in other words, it's a drone  that is used to look at cetaceans from above.   00:19:36.800 --> 00:19:42.160 And we can tell things like their body condition,  sort of the composition of the groups themselves,   00:19:42.160 --> 00:19:46.000 whether mom and calf pairs, or  what kind of animals or in a group.   00:19:47.360 --> 00:19:51.200 And all of these tools are very  complimentary, as I said before.  00:19:51.760 --> 00:19:57.760 But these ones, in particular, in particular,  really rely on seeing animals at surface.   00:19:57.760 --> 00:20:02.720 And for a species that primarily live  underwater, sometimes that can be problematic.  00:20:03.360 --> 00:20:07.360 So that's where passive acoustic  data collection comes in.  00:20:08.320 --> 00:20:12.560 Is because animals, cetaceans,  make a lot of sounds underwater.  00:20:13.200 --> 00:20:19.920 So we can passively listen to them using  hydrophones, or underwater microphones.  00:20:19.920 --> 00:20:25.280 And so, these come in a lot of different,  a lot of different configurations.  00:20:25.280 --> 00:20:32.400 The one in this diagram, like this ship is towing  a towed array of hydrophones that are housed in   00:20:32.400 --> 00:20:38.480 this case within tubing. These little yellow rectangles here, with the black dots, are kind   00:20:38.480 --> 00:20:44.640 of a diagram of what a towed hydrophone array  is and that's towed behind the research vessel   00:20:44.640 --> 00:20:51.440 so that we can hear everything that is within  the water column making noise at this point.  00:20:52.160 --> 00:20:57.760 Oh, I forgot, this is where I sit on the ship when  we are collecting these data in the field during   00:20:58.880 --> 00:21:01.360 the research surveys. And this is,   00:21:01.360 --> 00:21:06.240 the data comes in, and we are analyzing and  processing them using a lot of computers.  00:21:06.800 --> 00:21:11.360 And at this point, we know that we can  detect all the known cetacean species.   00:21:13.680 --> 00:21:19.520 And those include both baleen and toothed  whales, because both of these types of   00:21:19.520 --> 00:21:24.160 whales rely on sound to survive. Baleen whales in particular,   00:21:24.880 --> 00:21:29.840 communicate or vocalize in lower frequency  ranges, less than usually one kilohertz.  00:21:30.480 --> 00:21:36.160 Just for your reference, my voice is about  150 hertz. They typically produce moans,   00:21:36.160 --> 00:21:41.600 groans, and song, and are primarily used  for long range communication and mating.  00:21:43.200 --> 00:21:48.720 Then we have the toothed Whales, which is where  sperms also fall under and so those are higher   00:21:48.720 --> 00:21:54.800 in frequency. We call it the mid to high frequency  range, which is 1 to 140 Kilohertz, thereabouts.  00:21:55.360 --> 00:21:59.040 And those include whistles,  echolocation clicks, burst pulses,   00:21:59.840 --> 00:22:04.880 and are mainly used for communication,  foraging, and navigation. And so yes,  00:22:04.880 --> 00:22:08.640 this is where sperm whales  fall under as a toothed whale.  00:22:10.800 --> 00:22:15.520 And I want to talk a little bit about some of  the applications of passive acoustic data so that   00:22:16.240 --> 00:22:22.560 I can put my work into context for you. So, we can really use these data to detect animals   00:22:22.560 --> 00:22:27.840 and to see what's there. Is it even biological  or not? I guess, when we're out in the field,   00:22:27.840 --> 00:22:33.120 sometimes we hear all kinds of sounds and we're  able to know if they are a cetacean species.  00:22:33.120 --> 00:22:35.680 Sometimes not, though. There  are some, some instances.  00:22:36.720 --> 00:22:39.520 There's also a classification  of acoustic data which helps us,   00:22:39.520 --> 00:22:45.280 tells us, who or what species are present. Then there's localization, which is what we use to   00:22:45.280 --> 00:22:51.040 estimate where the animals are using their sounds. The distribution of the animals can be studied   00:22:51.040 --> 00:22:54.800 using passive acoustic data to show  that, tell us where and perhaps   00:22:54.800 --> 00:23:00.880 why animals are located where they are. Then the Holy Grail of passive acoustic data   00:23:00.880 --> 00:23:05.360 is to be able to estimate density or  estimate the number of whales in an area  00:23:06.640 --> 00:23:10.240 because that's a really important parameter  when it comes to conservation management.  00:23:11.120 --> 00:23:16.640 And so my research falls within, that I'm  talking about today at least is, falls within   00:23:16.640 --> 00:23:28.880 the localization and distribution categories. OK, let's see, so the new developments. I've, I   00:23:28.880 --> 00:23:33.680 will say that a lot of work has been done in these  two fields, especially when it comes to sperm   00:23:33.680 --> 00:23:37.760 whales. And I learned a lot of these techniques  when I was first starting out as an intern even.  00:23:38.480 --> 00:23:46.000 And along the way, I gradually learned about some  of the limitations that established methods have.  00:23:46.000 --> 00:23:51.040 Which, you know in any case there, there's  always limitations, which is what really drove my   00:23:52.160 --> 00:23:54.800 path into grad / graduate school  to try to address some of them.  00:23:55.360 --> 00:24:00.160 And so, really this work is improving  upon the existing methods that we have   00:24:00.160 --> 00:24:06.480 to study localization distributions from whales. Then, including the towed array data, which I'll   00:24:07.600 --> 00:24:14.800 be using by my projects, to use primarily and  using that data, in those data in different   00:24:14.800 --> 00:24:21.840 analyses to then better understand the ecology  of sperm whales, which helps us improve and make   00:24:21.840 --> 00:24:23.920 conservation management efforts more effective.  00:24:26.160 --> 00:24:32.640 So, getting into whales themselves, they are found  everywhere. So, this map, you see, just with all   00:24:32.640 --> 00:24:38.080 the blue that's, basically, their range. Sperm whales reside in areas high and high   00:24:38.080 --> 00:24:42.800 latitudes within ice free waters, all the  way to the warmer waters along the equator,   00:24:43.840 --> 00:24:50.000 including the Olympic Coast National  Marine Sanctuary, about here the pink star,   00:24:50.800 --> 00:24:55.200 and the Hawaiian archipelago.  It is within the orange circle.  00:24:56.080 --> 00:24:59.520 And in Hawaii, we have sperm  whales here year round,   00:24:59.520 --> 00:25:05.200 from all the different demographic groups,  including females, males, and juveniles.  00:25:07.440 --> 00:25:11.760 And sperm whales themselves  are really fascinating species.  00:25:11.760 --> 00:25:16.400 They, there's books about these  animals that I can't even touch upon,   00:25:16.960 --> 00:25:21.760 all of the details that are interesting to me. So, I'll just be telling you some of the more   00:25:21.760 --> 00:25:26.240 relevant biological information  that is, that my work relies on.  00:25:27.040 --> 00:25:32.640 So part of that is just to know that these animals  are the largest toothed whale species that we have  00:25:33.920 --> 00:25:37.600 on the planet, as well as they  are the largest hunting predator,   00:25:37.600 --> 00:25:40.320 which makes them really easy to see. They're one of the easiest   00:25:41.600 --> 00:25:44.720 cetaceans to see at the  surface and identify visually.  00:25:46.640 --> 00:25:49.680 They prey upon different  species of squid and fish,   00:25:50.560 --> 00:25:53.840 which reside lower in the water  column, really, really deep.   00:25:54.400 --> 00:26:00.560 And so that forces these whales to dive to  forage and search for their food at depths,   00:26:00.560 --> 00:26:06.800 sometimes over 6,000 feet and they can dive for  up to an hour, specifically about 45 minutes.  00:26:07.760 --> 00:26:10.960 And so they are spending  most of their time at depth.  00:26:13.440 --> 00:26:18.240 Luckily, they make a lot of sound when they're  doing that. So that's where passive acoustic data   00:26:18.240 --> 00:26:22.320 and monitoring comes in. We can  detect their echolocation clicks.  00:26:22.880 --> 00:26:26.720 And the characteristics of these clicks  makes them the loudest animal on earth.  00:26:27.520 --> 00:26:34.160 And I'll play a few examples in a moment. But, let's see if they, all these attributes,   00:26:34.160 --> 00:26:38.960 that the fact that they're really large, helps us  see them visually, the fact that they're really   00:26:38.960 --> 00:26:45.520 loud and helps us study them acoustically. So, this was a prime species for me to try and   00:26:45.520 --> 00:26:52.000 develop new techniques for using towed array data. All right. So talking a little bit about their   00:26:52.000 --> 00:26:55.200 echolocation clicks. Like I  said, they're really loud.  00:26:56.160 --> 00:26:59.520 They're lower frequency  than other cetacean species.  00:26:59.520 --> 00:27:04.000 Let me just march through here  because we have a spectrogram.   00:27:04.000 --> 00:27:10.000 This is how we are visualizing the sound. And I want to say that the dark lines   00:27:10.000 --> 00:27:15.120 along this plot, with frequency on the Y  axis and time in seconds on the X axis,   00:27:15.840 --> 00:27:23.520 shows that the lines or the lines themselves,  shows longer duration clicks than any other   00:27:23.520 --> 00:27:28.400 species that clicks. All the other toothed  whales have very short, echolocation clicks.  00:27:29.040 --> 00:27:33.920 And these different characteristics  allow us to detect these whales at great   00:27:33.920 --> 00:27:40.960 distances of usually over six miles, sometimes. And there are also four types of echolocation   00:27:40.960 --> 00:27:45.920 clicks, so that when we're viewing in this  picture around, includes the regular foraging   00:27:45.920 --> 00:28:01.840 or regular clicks that are used for foraging. And I will pull up an example of that. Now.  #snap or click# 00:28:06.160 --> 00:28:11.120 So, hopefully, you could hear that. But, if you had trouble, they, each   00:28:11.120 --> 00:28:15.840 of those lines on each of those snapping sounds  is basically what you can do with your fingers.  00:28:17.440 --> 00:28:21.200 And those are, you know, every click per  a second helps them search for food.  00:28:25.040 --> 00:28:28.880 They also make creaks, which are  also for foraging and those sound   00:28:28.880 --> 00:28:30.840 like. #rapid, louder repeating clicking# Yes.  00:28:46.240 --> 00:28:49.440 So those are thought to be when the whale is   00:28:49.440 --> 00:28:54.400 sort of zeroing in on their prey. They're,  they're producing these more rapid clicks.  00:28:56.560 --> 00:29:01.600 They also use clicks to socialize in  sort of a Morse code type of pattern.  00:29:02.240 --> 00:29:12.840 I will play an example here in this spectrogram. #rapid clicks# OK.  00:29:19.280 --> 00:29:23.120 So, there are different patterns of  codas that exist all over the world.   00:29:23.120 --> 00:29:25.840 Then there are people dedicated to just trying to   00:29:25.840 --> 00:29:29.840 identify them and compare them between  regions. That's pretty interesting.  00:29:30.880 --> 00:29:36.080 And last, but not least, we have the Slow  Clicks, which are primarily produced by males,   00:29:36.080 --> 00:29:38.400 as far as we know. Those sound like   00:29:38.960 --> 00:29:53.840 they're slow, obviously a few seconds between  the clicks, compared to the other click types.  00:29:58.720 --> 00:30:00.240 There we go. #intermittent clicks# So those are   00:30:00.240 --> 00:30:07.840 the four main types of clicks that sperm whales  produce and makes them, also just adds to their,   00:30:10.160 --> 00:30:14.960 as to why I think they're so interesting. So  to study these animals to study sperm whales   00:30:14.960 --> 00:30:18.800 along with others, all cetacean  species here in the Hawaiian Islands  00:30:20.000 --> 00:30:27.120 we conduct line transect surveys. And these  can last from one month to several months,   00:30:27.760 --> 00:30:32.480 And they're primarily used to estimate the  total number of animals in a population   00:30:32.480 --> 00:30:36.480 for each species. And we call that  estimating density and abundance.  00:30:37.760 --> 00:30:46.560 For this type of analysis, we rely on getting  perpendicular distances from / of the groups of   00:30:46.560 --> 00:30:53.600 animals or individuals to the track line, which is  shown here in this black line. Here is our little   00:30:53.600 --> 00:30:59.120 ship. And here are sperm whales at the surface  that are able to be seen by the visual observers.  00:31:00.480 --> 00:31:06.000 And as these analyses are conducted,  they're usually using just the visual   00:31:06.000 --> 00:31:08.960 observation data at this point. So if you watch the boat:   00:31:10.560 --> 00:31:16.480 that cruises down the track line. We're able  to get instantaneous distances from / of   00:31:16.480 --> 00:31:21.600 the groups from the track line and then that's  put into these analyses to estimate abundance.  00:31:21.600 --> 00:31:28.000 However, as we know now, sperm whales are  at depth a lot of the time echolocating.  00:31:28.000 --> 00:31:34.480 So, if we could get the passive acoustic data  included into these abundance estimations,   00:31:35.040 --> 00:31:40.240 in theory, we could produce more precise  estimates of how many animals are in a population.  00:31:41.440 --> 00:31:44.560 And, so, with towed arrays, we can  actually do that at this point.   00:31:45.760 --> 00:31:52.240 We do this on the ship in real time to also help  track animals, to help visual observers find them.  00:31:52.880 --> 00:31:55.920 And so, it starts by detecting  the clicks themselves.  00:31:56.480 --> 00:32:04.000 In this plot here, it's showing the bearings that  we can calculate from the array to the whale,   00:32:04.000 --> 00:32:10.640 given some software and different  calculations, that and as the whale   00:32:10.640 --> 00:32:14.240 passes the ship or 90 degrees of the  ship and out in that red dotted line,   00:32:14.800 --> 00:32:21.840 we can, um, calculate multiple bearings over  time. Because we have a moving ship and a moving whale  00:32:21.840 --> 00:32:27.360 it's a little more complicated than  visual observations getting the distances   00:32:27.360 --> 00:32:31.120 to then estimate the distance to  the whale using multiple bearings.  00:32:34.560 --> 00:32:40.000 And so, with that, there, you can picture a lot  more bearings happening for each of those dots.  00:32:40.560 --> 00:32:44.720 For example, you could get a bearing in theory  but, for this example, just showing you how   00:32:44.720 --> 00:32:50.000 this works, we can then use the software  to estimate that perpendicular distance.  00:32:50.800 --> 00:32:57.280 So, what I just described is really, can be  thought about as two dimensional localization,   00:32:57.280 --> 00:33:02.000 where you have the array at one depth, and you're  assuming that the animals are at the same depth.  00:33:03.200 --> 00:33:06.240 With sperm whales, that gets a  little trickier since they're usually   00:33:06.240 --> 00:33:11.040 way much deeper than the array. So, these assumptions   00:33:12.320 --> 00:33:15.680 that two-D, localization follow...  00:33:15.680 --> 00:33:18.800 let's see, so the assumptions include  that the animals are near the surface,  00:33:19.920 --> 00:33:23.520 we have a constant sound speed as we're  detecting them and localizing them,   00:33:23.520 --> 00:33:29.440 which is not necessarily the case all the time, and then that we know the exact position of the   00:33:29.440 --> 00:33:35.120 hydrophones as the towed array is being dragged  through the water, and that we're assuming the   00:33:35.120 --> 00:33:41.120 animals are stationary relative to the array. So really, and when it comes to Sperm whales, a   00:33:41.120 --> 00:33:46.880 lot of these assumptions are violated. Again, since Sperm whales are vocalizing at depth,   00:33:46.880 --> 00:33:53.360 and sound changes with depth, the first  two assumptions don't really hold and even   00:33:54.160 --> 00:33:57.360 knowing the hydrophones positions at all times is hard, because   00:33:58.880 --> 00:34:03.120 in theory, these hydrophones are actually moving  a little bit with the ocean currents and all that.  00:34:04.080 --> 00:34:08.560 And so with violations and assumptions,  we also don't include errors with this   00:34:08.560 --> 00:34:13.360 type of localization, which can lead to  biased distance estimates and perhaps,   00:34:13.360 --> 00:34:18.400 not the most accurate way of getting these  distances that we need to estimate local abundances. 00:34:19.600 --> 00:34:24.800 So, that's where I developed this, what we're  calling a model-based localization tool, for the   00:34:24.800 --> 00:34:30.480 towed array data. And this approach I'll describe,  accounts for the depth of the sperm whales.  00:34:30.480 --> 00:34:36.240 It also is able to incorporate errors that are  from, that occur due to the sound propagation or   00:34:36.960 --> 00:34:44.320 the change in sound speed over with depth,  ah the hydrophone positions, and then time   00:34:44.320 --> 00:34:50.160 difference of arrival measurements or TDOA measurements, which I'll describe a little later.  00:34:50.160 --> 00:34:55.680 And so the model-based approach, this a  term, and this is a term that helps, or that   00:34:55.680 --> 00:34:59.920 really, accounts for the effects of the ocean  environment on how sound travels through it.  00:35:00.720 --> 00:35:05.280 So this diagram, is just is a way  to show how sound travels. It can,   00:35:05.280 --> 00:35:10.480 due to differences in temperature and pressure  with depth. For example, this for a whale hanging   00:35:10.480 --> 00:35:15.760 out in the shadow zone wouldn't necessarily  be heard by a towed array near the surface.  00:35:15.760 --> 00:35:18.400 So, that could cause problems  when it comes to localization.   00:35:19.120 --> 00:35:24.320 And, at this point, there had been, or  at the time of my dissertation research,   00:35:24.320 --> 00:35:30.640 no, sorry, there had been no model-based  localization methods for shallow towed arrays,   00:35:30.640 --> 00:35:38.960 which is the datasets that I've been focused on. And so, TDOA stands for time difference of   00:35:38.960 --> 00:35:44.560 arrival. This is a really important parameter  in this approach, which is basically mean the   00:35:44.560 --> 00:35:48.880 clicks that we're hearing from the sperm whales  reached the hydrophones at different times.  00:35:48.880 --> 00:35:53.840 And that's how we're using that information  to estimate the direction to the whale   00:35:53.840 --> 00:35:59.760 and then ultimately establish a location  using multiple measurements of TDOA.  00:36:00.560 --> 00:36:06.080 So stepping through some of the process that was  used to develop this approach (I'm happy to answer   00:36:06.080 --> 00:36:10.960 more technical questions afterwards as well) But we started off with the raw data or the   00:36:10.960 --> 00:36:15.680 recordings that we were listening to, and  the spectragrams using the spectragrams.  00:36:16.880 --> 00:36:19.840 And then taking those recordings  and cleaning them up a little bit,   00:36:19.840 --> 00:36:25.600 reducing the noise in the background and  narrowing down the data itself to this,   00:36:25.600 --> 00:36:28.640 the frequency range of the  sperm whale echolocation clicks.  00:36:30.560 --> 00:36:35.120 Then taking the clicks themselves and detecting,  when we detect them, we can also extract them.  00:36:36.880 --> 00:36:42.400 And then use that information to measure the  TDOA. And this is really an automated approach,   00:36:42.400 --> 00:36:46.560 which is handy because that  really limits the operator error.  00:36:47.520 --> 00:36:55.360 And so we have our measured TDOAs from the data. And we're using those along with some of the   00:36:55.360 --> 00:37:03.200 modeled side to produce these what we're calling  "volumes of ambiguity." And it's a kind of ambiguous   00:37:03.200 --> 00:37:11.280 term, but really it just means it's our way  of, um, deriving 3-D, a three-dimensional   00:37:11.280 --> 00:37:16.240 shape to represent all possible locations where sperm whales could be located because   00:37:17.040 --> 00:37:21.600 those two-dimensional bearings don't  really show us the whole picture. And   00:37:21.600 --> 00:37:26.640 this approach also allows us to incorporate the  differences in sound speed and different errors.  00:37:27.360 --> 00:37:32.160 And so, if you can picture a  very empty grid of grid space in,   00:37:33.840 --> 00:37:39.040 which is where we, we can model the TDOAs themselves, and that are generated based   00:37:39.040 --> 00:37:44.560 on the different, the environment, and  the changes in sound speed that do occur.  00:37:46.560 --> 00:37:50.960 Then, taking the, the measured  TDOAs and these modeled TDOAs,  00:37:50.960 --> 00:37:56.960 we compute these ambiguity volumes using this  equation here, which then can incorporate the   00:37:56.960 --> 00:38:03.200 errors, and everything is it's really a  nice way of being able to incorporate all   00:38:03.200 --> 00:38:10.720 these pieces into one, one ambiguity volume. And then from that ambiguity volume, we can   00:38:11.520 --> 00:38:17.920 estimate a distance and depth, which is  what we need for the abundance estimations.  00:38:20.800 --> 00:38:24.400 So, to just give an example of what this really  looks like, because I know it's probably hard   00:38:24.400 --> 00:38:30.960 to visualize it, because it was for me when  I first started getting into this. I ran some   00:38:30.960 --> 00:38:36.880 simulations or really just some experiments  that I can program in all the different   00:38:38.240 --> 00:38:44.160 parameters, such as how deep the whale is, how  far it is, from the track line, how the ship   00:38:44.160 --> 00:38:48.960 is moving, whether, it's either going straight  down, or if it's making a turn of some sort.  00:38:49.520 --> 00:38:55.600 And when you calculate the measured TDOAs,  Model TDOAs, and produce ambiguity volumes  00:38:55.600 --> 00:39:01.920 they look like this. And so we have our whale,  which is located, we know that it's located here,   00:39:01.920 --> 00:39:07.760 and when the volumes are calculated,  this yellow and orange shape is what's produced.  00:39:08.640 --> 00:39:12.960 With higher probability being in the color black,  which can't really see because it's actually in   00:39:12.960 --> 00:39:18.320 the middle of this, sort of horseshoe  shaped volume. And so this is our boat  00:39:18.320 --> 00:39:22.800 It's going this straight track  line and you can see that the   00:39:22.800 --> 00:39:29.840 whale is located directly below the ship. This is all the possible locations that it   00:39:29.840 --> 00:39:36.640 can technically be located. And if you're  assuming that they're at the surface this,   00:39:36.640 --> 00:39:41.360 these two points here on either side of the  U shape, would be where that two-dimensional   00:39:41.360 --> 00:39:47.920 localization approach would estimate the whale,  which is much farther away from the track line  00:39:47.920 --> 00:39:52.640 then the whale is truly located. So that is  where those biased distance estimates could occur.  00:39:53.920 --> 00:39:56.480 So I was curious to what  happens when you turn the boat.  00:39:57.200 --> 00:40:03.440 And, as the turn itself can  then cancel out some of that, um,   00:40:04.640 --> 00:40:10.640 some of the shape of the volume and really  make it more precise in that little blob, here.  00:40:11.280 --> 00:40:20.720 So, this, turning the ship is very important, when  to make these localization estimates more precise,   00:40:20.720 --> 00:40:26.480 and those distance estimates more precise. OK,  so some conclusions, just to wrap this up.   00:40:28.160 --> 00:40:32.960 We've developed a more automated approach  for localizing sperm whales in general.   00:40:33.520 --> 00:40:36.400 So that's part of this, too, is really important.  00:40:36.400 --> 00:40:42.080 because tackling, re-localizing  data, it was not an easy thing   00:40:42.080 --> 00:40:48.560 to do before I develop this, this method. So this takes out a lot of the user error,   00:40:48.560 --> 00:40:52.640 it helps us able to control a lot  of the parameters more easily.  00:40:54.080 --> 00:40:59.040 Then that includes incorporating the effects of  the environment, the different errors themselves,   00:40:59.040 --> 00:41:05.200 and even understand how the whale's position  can affect these distance estimates that are so   00:41:05.200 --> 00:41:11.760 important for getting good abundance estimates. And these locations themselves, also can be,   00:41:13.280 --> 00:41:18.960 they're now given position bounds, sort of an  error range around the best distance estimate.  00:41:18.960 --> 00:41:26.960 Now, we can also measure sort of a plus, plus  or minus range around that best distance.  00:41:28.720 --> 00:41:35.440 And, as we knew before, with even two-dimensional  localization, we could turn the ship to improve   00:41:35.440 --> 00:41:40.480 the location, the location estimates. And it holds  true for this method, which, just, in general,   00:41:40.480 --> 00:41:47.280 will give us more precise information. At the  moment this is, this paper is in review. So,   00:41:47.280 --> 00:41:52.240 hopefully, that will be something I can  share in the future, ah the published paper.  00:41:54.320 --> 00:41:59.040 OK, good job, everybody. We got  through the localization part.  00:42:01.200 --> 00:42:08.160 And, we will now move into how we can use this  data, these data and this location information,   00:42:08.720 --> 00:42:12.640 and look at the distribution patterns as  from animals in the Hawaiian archipelago.  00:42:14.720 --> 00:42:19.040 All right. So, we're calling this the  Species Distribution Models, or SDMs.  00:42:19.600 --> 00:42:23.520 This is basically a tool, a  statistical tool ,that helps us   00:42:23.520 --> 00:42:29.840 predict the different patterns of distribution  that over time and space for species.  00:42:31.520 --> 00:42:35.920 And we can take these, there are field  observations which can be either, either visual   00:42:35.920 --> 00:42:41.920 observations or passive acoustic detections  and relate that to environmental variables.  00:42:43.520 --> 00:42:47.040 And that helps us better understand  the ecological relationships   00:42:47.040 --> 00:42:49.040 between a species and their environment,   00:42:50.080 --> 00:42:56.480 and even just predict where the animals may occur  based on different environmental measurements.  00:42:57.600 --> 00:43:03.840 And so for cetaceans, we primarily use sighting  data at this point to produce cetacean SDMs.   00:43:04.640 --> 00:43:08.240 And then a sort of a rough  overview of how that's done,   00:43:08.240 --> 00:43:12.800 so we have our ship and our track line again. And the ship moves down the track line.   00:43:12.800 --> 00:43:16.080 Its detecting animals, which  are denoted with the stars.  00:43:16.960 --> 00:43:23.600 And when these types of analyses, they chop  the data up to kind of get, put it in uniform   00:43:23.600 --> 00:43:28.560 units, to be put into the model. And that's done by segmenting the track line,   00:43:29.120 --> 00:43:33.840 finding the midpoint with the red dot there,  and then associating all the different   00:43:33.840 --> 00:43:42.000 sightings along each segment to that midpoint with  environmental data, to make it more co, cohesive.  00:43:43.520 --> 00:43:50.240 And so the spatial models for cetaceans,  can also just with the number of   00:43:50.240 --> 00:43:56.160 cetaceans estimated in each of these groups,  we can model the species density over space.  00:43:56.720 --> 00:44:04.240 So for example, a paper by Elith and co-authors estimated sperm whale density to   00:44:04.240 --> 00:44:11.120 have this broad pattern of increasing up towards  the north-west Hawaiian islands, which is shown in   00:44:11.120 --> 00:44:18.320 the orange and red colors in that, in that plot. So that's one way that we can use SDM to estimate   00:44:18.320 --> 00:44:25.120 and predict where these animals can be. And I was part of this project after that  00:44:25.760 --> 00:44:32.640 to try to incorporate pathway acoustic data  into these SDMs because we hear so many whales   00:44:32.640 --> 00:44:36.080 when we're out in the field, that they're  not really accounted for in some of these   00:44:36.080 --> 00:44:41.200 analyses. So it would be interesting to know  how that would change, these, the results.  00:44:41.840 --> 00:44:45.680 So using that same track line  segment method I just described,  00:44:45.680 --> 00:44:53.120 we incorporated both sighting and acoustic data  into these models, to try and test that theory.  00:44:53.120 --> 00:45:00.400 And so unfortunately, with this sort  of exploratory study, we weren't really   00:45:00.400 --> 00:45:05.920 sure of the accuracy of the models themselves. And you can see with the visual model on the top   00:45:05.920 --> 00:45:09.680 panel compared to the a model,  they're really different.  00:45:09.680 --> 00:45:12.960 And it's, it's hard to know if one  is better than the other without   00:45:13.520 --> 00:45:22.240 further analyses, which is where my PhD came in. So part of this, we thought, would be due   00:45:22.240 --> 00:45:25.120 to the fact that sperm whales  can be acoustically detected   00:45:25.120 --> 00:45:30.800 really far away. And perhaps, the environmental  data wasn't being correctly associated with   00:45:30.800 --> 00:45:35.360 those farther or more distant whales. And that's where the localization part   00:45:35.360 --> 00:45:41.040 of this came in and really drove  that study as well as this one.  00:45:42.320 --> 00:45:47.280 And so, the data that I were was  able to use included survey data   00:45:47.280 --> 00:45:56.720 from four years from 2010, 13, 2016, and 2017. And it's really spans the entire archipelago.   00:45:56.720 --> 00:46:03.200 So it was giving us a really good sample size  of over, just over 200 sperm whale encounters.  00:46:03.200 --> 00:46:09.040 And this includes all sighted and acoustic,  acoustically detected sperm whale groups.  00:46:12.080 --> 00:46:18.080 And so when it comes to putting them all together,  it's really important to have location data for   00:46:18.080 --> 00:46:23.680 the sperm whales to, to input into the model. And so, I broke them down into different   00:46:23.680 --> 00:46:29.600 categories including: sperm whales that  were sighted only; sperm whales that were   00:46:29.600 --> 00:46:35.040 acoustically detected and we were able to  localize them using the model based approach.  00:46:36.480 --> 00:46:42.240 Some sperm whales were not able to be localized  when we acoustically detected them so, but   00:46:42.240 --> 00:46:48.720 we could still include them in the model. Then, we had the combination, where we both,   00:46:48.720 --> 00:46:51.760 we sighted or we saw and heard  the animal, the sperm whales.  00:46:53.200 --> 00:47:00.560 And with acoustic data, we were or, I  categorized the sprawl counters down to   00:47:01.120 --> 00:47:07.600 foraging animals and non-forging animals. And based on a type of click that we,   00:47:07.600 --> 00:47:14.800 that I identified within the acoustic detection. So that is kind of a really important point.  00:47:15.600 --> 00:47:22.720 In the, for one of the reasons why acoustic data  can be important and more informative sometimes or   00:47:22.720 --> 00:47:28.400 at least provide a more complete picture. And so using these categories,   00:47:29.280 --> 00:47:34.320 from sightings of the non-foraging whales I  create a different models for each scenario.  00:47:35.440 --> 00:47:40.160 So oh first, the environmental variables  that were included in these models  00:47:40.800 --> 00:47:45.520 were all biologically relevant to sperm  whales, including what I'm calling   00:47:45.520 --> 00:47:50.640 static variables that are environmental  variables that don't change over time.  00:47:50.640 --> 00:47:59.280 So that means bathymetry, distances to land or  seamounts, and other, um, static variables that   00:47:59.280 --> 00:48:05.840 really help, or at least are acting as  a proxy for a suitable prey habitat or   00:48:06.960 --> 00:48:10.160 features that might influence the  water column, to then influence   00:48:10.880 --> 00:48:16.640 prey and and presuming that sperm whales  will be located where there is more food.  00:48:18.000 --> 00:48:21.760 Also, I included dynamic variables  that do change over time.  00:48:21.760 --> 00:48:27.280 A lot of these are remotely sensed from  satellites. They are also all really acting   00:48:27.280 --> 00:48:32.400 as proxy for primary productivity  or related to the prey in some way.  00:48:33.360 --> 00:48:39.200 The last one there at temperature, 584  meters was a, a new variable that I found,   00:48:40.480 --> 00:48:45.520 that for to me, new to me, and it  was modeled from a model dataset.  00:48:45.520 --> 00:48:52.560 And that depth in particular was representative  of one of the main depth ranges for   00:48:52.560 --> 00:48:57.120 a type of squid that sperm whales tend  to eat, a species of Cockeyed Squid.  00:48:58.400 --> 00:49:01.840 So with all this environmental variable  on the location of the whales themselves,   00:49:02.960 --> 00:49:07.040 we can then create the models. So this is again that   00:49:07.040 --> 00:49:10.800 tract line segment method for configuring a model.  00:49:11.520 --> 00:49:16.400 And if you want to just picture the sperm  whales, instead, that's great. Sometimes they’re,   00:49:17.360 --> 00:49:20.800 then they can be associated with the  environmental data at the midpoint.  00:49:20.800 --> 00:49:24.000 But what about those whales  that are really far away? Those   00:49:24.000 --> 00:49:28.800 are the ones that we think were problematic  before in the models using acoustic data.  00:49:29.600 --> 00:49:34.560 So part of this was to develop a different  way of configuring the model itself.  00:49:35.200 --> 00:49:45.520 And by using a grid format, we can, the point  was to more closely associate that environmental,   00:49:45.520 --> 00:49:50.400 at the red dots which are the Grid's  centroids, to the whales that are located   00:49:50.400 --> 00:49:57.840 farther from the tract line, in hopes that  this would provide us more accurate results.  00:50:00.320 --> 00:50:04.960 So starting with model results, for the  sighting base models, that just used   00:50:06.640 --> 00:50:11.520 data from whales that we're only seeing, so we have a performance metric here just   00:50:11.520 --> 00:50:18.560 to give us an idea or a relative measure  of performance, comparing the models   00:50:18.560 --> 00:50:24.480 themselves between the models. We have the output and these   00:50:24.480 --> 00:50:28.320 are using generalized additive models. And for those of you who are familiar with those,   00:50:28.960 --> 00:50:36.800 they're pretty common in the cetacean SDM world. And so, the important environmental predictors   00:50:36.800 --> 00:50:40.800 that created the best model included sea  surface temperature with a decreasing   00:50:41.760 --> 00:50:47.680 relationship and a standard deviation of sea  surface height, which is one of those variables  00:50:47.680 --> 00:50:54.160 that's a good proxy for primary productivity. Then, we have this spatial term here, which is a   00:50:54.880 --> 00:51:00.480 basically using the location of the whale to help  predict the distribution that was found important,   00:51:01.360 --> 00:51:04.800 predicting more whales with yellow,  which is a little counter-intuitive.  00:51:06.160 --> 00:51:09.120 But, yes, more whales predicted  in the yellow than the red.  00:51:10.400 --> 00:51:15.840 And we compare that to an acoustic-based model  that used only acoustically detected whales.  00:51:15.840 --> 00:51:20.240 The performance metric of the percent of  deviance explained, is a little bit lower   00:51:20.240 --> 00:51:26.560 even with larger sample size. And it predicted, the environmental predictors   00:51:26.560 --> 00:51:30.640 were selected to be depths and sea surface  height, which are slightly different than before.  00:51:31.360 --> 00:51:36.560 But this spatial terrain shows a little bit more  of a detail on the contours of these lines here,   00:51:37.520 --> 00:51:43.440 showing higher probability of sperm whales  groups being detected in yellow, again.  00:51:44.560 --> 00:51:49.120 But, in more of a range in the north-west and  a little bit north of the Hawaiian Islands,   00:51:49.120 --> 00:51:53.600 the main Hawaiian islands. We  combined all of these data together.  00:51:55.440 --> 00:52:02.800 We have a slightly better performance metric. Again, with depth as being important,   00:52:02.800 --> 00:52:05.120 standard deviation of sea surface height,   00:52:05.120 --> 00:52:08.960 and a little bit more detail and the  contouring of the spatial terrain here.  00:52:11.680 --> 00:52:16.560 Moving into the foraging category... So these are  all whales that were producing either the regular,   00:52:17.120 --> 00:52:23.440 regular clicks or the creaks foraging for prey. And it had a very, much higher   00:52:24.880 --> 00:52:30.080 performance metric, not as high still though  as a sighting based model, surprisingly.  00:52:30.640 --> 00:52:37.200 But it did predict or did select the  temperature at depth of 584 meters,  00:52:37.200 --> 00:52:43.520 also standard deviation of sea surface height,  and even chlorophyll for this model. But you can   00:52:43.520 --> 00:52:46.640 see here the spatial terrain is a little  more interesting. It kind of shows the   00:52:47.520 --> 00:52:53.280 foraging males to be predicted here with  between, this is Pearl and Hermes, sorry,   00:52:53.280 --> 00:52:59.440 Pearl and Hermes Atoll, Lisianski and Laysan area of the north-west and also a little bit   00:53:00.000 --> 00:53:04.640 of a higher probability north  of Maui and the Big Island.  00:53:05.760 --> 00:53:10.640 When we look at the non-foraging whales, the ones  that are probably producing, I think the codas   00:53:10.640 --> 00:53:16.960 and the slow clicks primarily, and those look  very similar to the first sighting-based model   00:53:16.960 --> 00:53:21.680 as far as a spatial terrain. But it also  selected depth as an important variable   00:53:21.680 --> 00:53:24.240 for predicting where non-foraging whales would be.  00:53:27.040 --> 00:53:34.000 So we can take these models and then predict onto  new data, to sort of see how well they work still.  00:53:34.640 --> 00:53:41.920 And when we look at the models that using  just the sightings, it does fairly well.  00:53:41.920 --> 00:53:49.760 It's still predicts the whale's just to be  similar to the previous models done by Forney and 00:53:49.760 --> 00:53:55.920 co-authors up in the north-west region. But then you look at the   00:53:55.920 --> 00:54:02.080 foraging and non-foraging models. You see these sort of hotspots in red.   00:54:02.080 --> 00:54:07.920 These are like higher density areas for  these types of whale groups to be located.  00:54:07.920 --> 00:54:11.840 So it's interesting that the sighted whales  and the non-foraging models are very similar,   00:54:12.800 --> 00:54:17.360 that, not exactly sure why, but  it was an interesting comparison.  00:54:17.920 --> 00:54:25.440 But really, it's the main takeaway is that  the passive acoustic data allows us to really   00:54:25.440 --> 00:54:33.040 look at these different types of groups of sperm  whales instead of just viewing them all as 1 type.  00:54:34.800 --> 00:54:42.880 So in conclusion, we found that depth, sea surface  height, the temperature at depth of 584 meters,   00:54:42.880 --> 00:54:47.920 and even just the location of the whale groups were very important when building   00:54:47.920 --> 00:54:52.800 the models. And that including the  acoustic data, help distinguish two  00:54:54.000 --> 00:54:56.400 foraging areas that we didn't  really know about before.  00:54:57.040 --> 00:55:02.640 Of course, that will take further surveys  to really confirm that. But it would be   00:55:02.640 --> 00:55:08.880 interesting to maybe focus in on those areas  in the future and see if there were more,   00:55:09.600 --> 00:55:15.680 more whales in that area. Then, all together these two analyses,   00:55:15.680 --> 00:55:22.000 these two projects I just described really help us improve how we can utilize the acoustic data   00:55:22.000 --> 00:55:25.760 to help us better understand sperm whale populations here in Hawaii.  00:55:25.760 --> 00:55:31.040 And yes, this, this can be applied  to other populations, the sperm whales in   00:55:31.040 --> 00:55:36.640 other parts of the world as well. And so, for that, at this point,   00:55:37.440 --> 00:55:43.920 I wrap those projects up. And I'm sort of working  on a new piece to this whole puzzle of acoustic   00:55:43.920 --> 00:55:49.520 abundance estimation in a postdoctoral research  position with this Cetacean Research Program,   00:55:50.160 --> 00:55:56.080 looking at the relationship between sperm whale click  rates and their group size, to try to get at the   00:55:56.080 --> 00:56:01.840 question of what proportion of time are these  whales making sound so that we have a better,   00:56:02.720 --> 00:56:07.520 so that we can get better acoustic abundance  estimates for these for this population.  00:56:10.560 --> 00:56:16.640 And with that, I want to just acknowledge that  all of the data that I analyzed for my PhD   00:56:18.400 --> 00:56:24.800 was collected during these really large scale  surveys. And it took a ton of crew and scientists,  00:56:24.800 --> 00:56:30.400 the board, the Oscar Elton Study NOAA ship during these these long line   00:56:30.400 --> 00:56:34.240 transect series to collect the data. So, I'm very thankful for that.   00:56:34.240 --> 00:56:38.880 And just acknowledging my funding  from the NOAA fisheries, NSF 00:56:38.880 --> 00:56:43.840 Graduate Research Fellowship Fellowship Program,  and the Hawaii Institute of Marine Biology.  00:56:44.960 --> 00:56:50.240 And then my advisor, Doctor Franklin, as, as well as my committee members,   00:56:50.240 --> 00:56:55.280 Doctor Eva Nosal, and Doctor Erin Oleson,  that helps really get these projects going and   00:56:56.080 --> 00:57:02.080 helped me through the process. And also, Taiki Sakai from the south-west Fishery Science Center.  00:57:02.080 --> 00:57:07.680 He was instrumental in helping me with the  species distribution modeling part with   00:57:07.680 --> 00:57:09.840 figuring out how to grid out my,   00:57:11.360 --> 00:57:16.240 my data and an associate, the experiment,  was with environmental data in that format.  00:57:17.120 --> 00:57:21.280 With that, I'd like to thank you all  for listening today, and bearing with   00:57:21.280 --> 00:57:25.840 me with the technical difficulties, and  I will, I'm glad to take any questions.  00:57:31.840 --> 00:57:37.120 -Thank you, Yvonne. That was really  interesting and great photos.  00:57:37.120 --> 00:57:41.920 I appreciate that. Well, what we'll do,   00:57:41.920 --> 00:57:48.880 if it's OK with you, since we just have a couple  of minutes, is to just maybe pick two questions   00:57:50.480 --> 00:57:58.640 and, and then assure everyone that all  of your questions have been collected   00:57:58.640 --> 00:58:06.000 in the question box. And so we'll be sending them  off to Yvonne to get answers to all the questions.  00:58:07.360 --> 00:58:14.800 Um so one question that was asked is,  can you identify and distinguish between   00:58:14.800 --> 00:58:18.080 individuals' sperm whales  based solely on their clicks?  00:58:19.760 --> 00:58:22.800 -Hmm, I'm going to click  characteristics of themselves.  00:58:24.800 --> 00:58:30.480 At this point, not necessarily out there. We can identify between   00:58:31.120 --> 00:58:37.440 individual sperm whales based on the localization  data because that plot I showed with the   00:58:37.440 --> 00:58:43.040 black dots as the whale moved past the ship. If there were several lines of those bearings   00:58:43.760 --> 00:58:49.280 we can denote individual whales  clicking and diving in the same area,   00:58:49.280 --> 00:58:55.680 but using just their click characteristics, as  far as I know, that's not quite possible yet.  00:58:58.160 --> 00:59:06.800 -OK, sort of a somewhat related question, have you  tried to use passive acoustic data to estimate the   00:59:06.800 --> 00:59:11.360 size of sperm whales using this technique? -Yeah.  00:59:12.240 --> 00:59:16.480 Yeah. So one another factoid about sperm whales is that   00:59:16.480 --> 00:59:21.520 their heads are about a third of their body  size and that's where the clicks are produced.  00:59:21.520 --> 00:59:25.600 And so there's the way that  the clicks are emitted, it's   00:59:26.640 --> 00:59:32.320 they are bounced inside and out or sorry, they're  bounced within their head as they are emitted.  00:59:32.880 --> 00:59:38.640 And people have used that information and the  timing of the clicks, because they're not what I   00:59:38.640 --> 00:59:42.560 showed in the spectragrams were these dark lines. And in reality,   00:59:42.560 --> 00:59:48.640 there's multiple pulses within a click. And people have used those measurements, to  00:59:49.360 --> 00:59:52.640 estimate the size of the whale. And that's has been validated by   00:59:53.200 --> 00:59:58.400 actually going in the field, recording clicks  from whales, sort of as head on as they can,   00:59:59.120 --> 01:00:04.720 and then comparing that to the length of the  world that they measure from the field, also.  01:00:04.720 --> 01:00:11.920 So that is a really interesting aspect of sperm whale acoustics that is possible to measure or at   01:00:11.920 --> 01:00:18.320 least estimate their body size using their clicks. -Great. Oh, that's really interesting. I   01:00:19.120 --> 01:00:24.320 hadn't really thought about that before. Well, as I mentioned, we're getting really   01:00:24.320 --> 01:00:30.240 close to time, so I'm not going to keep everyone,  but I am going to assure you that Yvonne will   01:00:30.240 --> 01:00:35.680 answer the questions you've posed to her. And  we'll get those off to her fairly quickly,   01:00:35.680 --> 01:00:40.000 and hopefully she couldn't get them turned  around for you all in the next week or so.  01:00:41.600 --> 01:00:52.640 Not too much pressure there, I hope. Then also, I wanted to thank you for, not only   01:00:53.760 --> 01:00:57.440 Yvonne, but all of the attendees. And I wanted to give you a little   01:00:58.720 --> 01:01:04.800 information about some other options,  um, that for events that we have.  01:01:05.520 --> 01:01:16.880 So, um, I am going to change, there,  and, so, if I can try to, wanted to show   01:01:16.880 --> 01:01:21.760 you something but, I'm having some challenges  here switching over to myself, there we go.  01:01:22.960 --> 01:01:31.520 OK, so, let's do this from this current slide. So, um, some of the resources that the sanctuary   01:01:31.520 --> 01:01:37.840 offers, sanctuary system offers as, so we  have some national marine sanctuaries live   01:01:37.840 --> 01:01:43.040 interactions and these are they're all in  archive right now but there's some amazing   01:01:43.040 --> 01:01:49.520 topics that are available for viewing in  on the national marine sanctuary website.  01:01:50.720 --> 01:01:56.240 There are ongoing events “Exploring by the  Seat of your Pants." These are really great for   01:01:57.760 --> 01:02:00.960 school age children, but adults  would find them interesting,   01:02:00.960 --> 01:02:05.840 as well. And as you can see, it, there are a  lot of different topics coming up very soon.  01:02:07.200 --> 01:02:14.240 And also, there are more National Marine  Sanctuary Webinar Series and the next one,   01:02:14.240 --> 01:02:20.160 which is tomorrow, is “Understanding Marine  Biodiversity in the Papahānaumokuākea   01:02:20.160 --> 01:02:25.200 Marine National Monument” and you'll find  that there are several others, as well.  01:02:27.120 --> 01:02:35.840 And then just lastly, put out a little Save  the Date for you, March 17th. So a month away,   01:02:35.840 --> 01:02:42.720 also at 4:00 PM Pacific Standard, will have  Jennifer Hagen, who is the Marine Policy  01:02:45.520 --> 01:02:50.240 Advisor at the Quileute Indian Tribe with  our Natural Resources Department. And she'll   01:02:50.240 --> 01:02:56.720 be talking to us about some research that she's  doing with larval crab and Harmful Algal Blooms.  01:02:59.760 --> 01:03:04.960 So again, we'll get back to you with all of  your questions. I hope you've enjoyed this.   01:03:05.600 --> 01:03:13.200 And then just a little reminder to please  take a few moments to complete the survey,   01:03:13.200 --> 01:03:29.840 and have a very good evening.  We'll see you next time.