Biologist Explains How Drones Catching Whale "Snot" Helps Research

Whale research is difficult, but a conservation organization is trying to make it a bit easier with drones.

WIRED's Matt Simon talks with Ocean Alliance CEO Iain Kerr about the benefits of flying the "SnotBot" into a whale's blow spout.

- [Matt] Whales are majesticbut very difficult to study creatures.- Taking off, three, two, one.- [Matt] But drones are starting to makewhale research easier, if a bit messier.This is Snot Bot, yesSnot, as in whales snot.- So what's so cool about whale snotis they're blowing out not onlythe actual sort of whale tissue,at a very small level butwe're getting whale cellsthat will have DNA, we'regetting the microbiomeand we're also gettinghormones, you know the hormonesthat can tell us is the animal pregnant,is it lactating, is it sexuallyactive, is it stressed?- That's Iain Kerr,director of Ocean Alliancewhich developed the Snot Bot.I spoke with him about how dronesare transforming whale research.Before the drone camealong, how did you doa biopsy on an animalthe size of a school bus?- So in our case we'dhave quite a large boatwhich was expensive and we actually hada little boom that wentout and I would siton the end of this boomwith a crossbow, okay.And we'd try to getclose enough to the whaleto them biopsy it, wherewe shoot this crossbowand the piece of tissuewe get would be aboutthe end of your pinky finger or maybethe size of an eraser on a penciland that little piece of tissuewould have all of this biological data.But what's interesting here is thatthat biological datawas more often than not,legacy data, you know,this is what had happened.So we're all most interestedin what's going on right nowso when you get an exhalation,that's really speakingto the condition of theanimal right then and therewhich is pretty exciting.Certainly the exhaled breath condensateor whale snot, we're notat that point right nowwhere we're getting all of the datathat we're getting from the biopsy samplebut the very fact that it'snon-invasive is importantand also it's more cost effective.You know, we can run adrone from a small boat.When I was sitting on the bow of a boatwith my biopsy crossbow, I almost feltlike I was playing whack-a-molebecause a whale would appear over thereand I'd race over thereand then, oh, it would divebut the whale would appear over thereand I'd race over thereand then it would diveand you know I almostfelt like I was sittingon the bow of the boatripping up $100 billsbecause it was costing a lot of moneyand I wasn't collecting the dataand I'm like, god ifonly I could just like,fly over there and grab that dataand in fact it was one ofthese very frustrating days,we were down in the gulf of Mexicoworking with sperm whales afterthe Deep Water Horizon disasterand the whale had dived andit was the end of the dayand I'm like, ah I hadn't got any samplesand this cloud of snot engulfed meand it was wonderfully horrible,you know, it was sticky and smellyand all of those things that technicallya biologist is meant to like, I'm not surehow much I was liking it at the time.But that's when I realized, wow,maybe, maybe I smell an ideahere, I smell a solution.- I gotta ask for you to elaborate there,what does it smell likein a cloud of whale snot?- Well you know guess what,it's almost what you'd thinkand it's somewhat dependent onwhat the animals are eating.Do you know what I mean, like blue whalesthat are eating krill arenot as bad as humpback whalesthat are eating sand eels, you know,there's a volume issue here, I mean,a blue whales lungs areabout the size of a VW.So imagine they're exhaling this,I mean you are immersed in this cloud of,you know of the worst badbreath you've ever smelled.- The things you do forscience I guess, huh?- That's right, that'sright and the thingswe're excited about, Imean who woulda thought?- Can you walk us through thedevelopment of the Snot Bot?Maybe some design challengesthat you came acrossin the beginning and howyou surmounted those.- When I started withthis, it was a little bitcounterintuitive, 'causeif you think about it,the whale is blowing up andthe drone is blowing downso the early prototypes, weactually had poles sticking downyou know, the drone washere and the pole was down,it's like two feet long andat the bottom of the poleI had sponges and wedding veil and meshesand what I realized quite quickly,some of these things that aregreat at absorbing the blow,you then had to get them out againand suddenly now, how do Iknow that that wedding veilis leeching, you know,plastic sort of toxicant,you know what I mean, thatwould actually affect my data.But we noticed when we'reflying into the blow,we were getting snot onthe top of the drone.And I think what was going on was thatthe drone's sucking all this air inso if we sort of came up behind the whale,the snot was already on that trajectorybut it was sucking it back onto our droneso incredibly counterintuitive.We just started putting Petri disheson the top of the drone.- You collect the snot,you get back on land,what do you then do to analyze that sampleand what exactly are you looking for?- The drone comes in and wejust close the Petri dish.We take it back to our hotel that nightand there's a wonderfulsound of one of the teamwith a mask on, theyliterally have like a car,a mini car squeegee and theygo like, ee, ee, ee, ee, ee, eeand they suck up the snotinto a little cryovalvethat we then send todifferent laboratories.- You're also able to identify the whalesby the shape of their flukes,can you talk about howthat technology works?- When you're looking at these animals,it's long been an issue ofhow can we identify them?And what's cool is a lot ofwhales lift their tails upand basically the patternson the back of their flukesare very unique and they tend tostay with the animal from birth.Now the flukes can actually get bite scarsand scratches on them but generallylots of whale flukes havebeen a very useful wayto track an animal throughout its life.Which is very exciting.I do think with dronesthough, we'll be usingmore looking down, maybepatterns on their back,maybe their sort of pectoral finsand maybe even their blowhole.So I think that's gonnachange but the fact thatwhen we fly our dronesout to collect snot,we're recording from thetime we leave the boat'til the time we get back.- Is there any artificial intelligenceinvolved there in identifying the flukeor are you doing that all still by sight?- We're doing it bysight, we actually workedon a project with Intel,they came on boardand used some pretty high computing powersto help us and we wereflying out over a whaleand they actually took the live feedand interrogated this database.And while I was still flying, so withina 20 minute window, they said Iain,you last saw that whale 10years ago, 20 miles from here.So imagine if we knewthat animal had been sickor that animal had been hit by a shipor that animal had beencaught in a net, you know,or the animal had acalf, knowing what animalwe were working with orknowing the life historyof the whale while we're with that animalwould certainly make thework more affordable,we'd get a larger data set.People like me have lived in a worldof sort of data deficit, you know,when you're studying a rhino,it doesn't dive beneath the Serengetiafter throwing a bucket of water on youand appear five miles away,you can sort of stay with that rhino.Whales are reallydifficult animals to study,you know they disappear under the water,and maybe spend more oftheir time underwaterdoing the things thatreally make up their lives.And that's why it's soexciting that we can lookat these hormone, hormonalclues when the animals exhaleas to what's been goingon over the last hour.- And are there anyindications that the animalsare at all bothered by Snot Bot,obviously it's not as invasiveas the previous methodbut do they panic at allwhen you fly over them?- Yeah you know what, I thinkthat's a really good questionand actually you need to put me to work,we have a publication thatI'm hoping will be outwithin the next month andbasically in 370 flights,we had 14 reactions, toget the permit for the workactually there was greatconcern about sort ofacoustic disturbance butwhat's actually interestingand we wrote a paper on this and a coupleof other people havewritten a paper on it.The high frequency sounddoesn't get through the water.So a low frequency sound like a helicopteror an airplane, youknow where you're like,whop, whop, whop, whop,does get through the waterbut the high frequency sound that might beincredibly irritatingto us, that sort of wah,doesn't seem to get through the water.Generally speaking, Ithink really good questionbut data to date suggestsit's very minimally invasive.- So what comes next here withyour projects with Snot Bot?- I'm still trying to pushwhat other technologiescan we use, we used a thermal camerato see if we could measurethe body temperatureof a whale by looking down the blowhole.So we're really excited to see if we canmeasure body temperature and also thenif the body temperature is elevated,we'd say okay, what elsecan we do with the droneor should we dig deeperon this hormone work.I'm also very interested inpotentially fleets of drones,can we take almost a threedimensional picture of a whalewith multiple drones, can wedo sort of morpho-metrics,look at body condition to help estimatethe health of a whale and you know,ultimately, there mighteven be intervation programswhere the whale has been, had a scar,it's been cut, it's been infected,maybe a drone will comealong and shoot antibioticsinto the whale to help itdeal with its infection.Using other drones, we have one dronewe affectionately call EarBot where we drop the dronein the water with a hydrophoneand record the vocalizations.And what'll be exciting is if we can landmultiple drones in thewater, we might actuallybe able to say, wait a minute,this animal's talking to that animaland this animal's replying,better understandinghow they communicate with each other.You know could provide some justamazing insights into their lives.- [Matt] Thank you for being with usand good luck out there.- Thank you very much.[upbeat music]