DUCKS, DRONES AND DATA: a new ap­proach to sea­son set­ting

As­so­ci­ate Pro­fes­sor Graham Hall was out in the field again dur­ing the 2016 open­ing week­end, col­lect­ing head and wing sam­ples. it is a frac­tion of the work be­ing un­der­taken on Aus­tralian water­fowl with the aim of im­ple­ment­ing a sci­en­tific based adap­tive ma

Field and Game - - Ducks, drones & data -

When news of the sud­den de­ci­sion to close Lake El­iz­a­beth be­cause of the pres­ence of about 155 threat­ened blue-billed ducks came through, As­so­ci­ate Pro­fes­sor Graham Hall was stand­ing on the shore­line sur­vey­ing the water­fowl.

On the eve of the sea­son open­ing, hunters were al­ready in the water pre­par­ing hides and set­ting de­coys.

Within an hour they would be in­formed by au­thor­i­ties they had to leave.

Pro­fes­sor Hall had to leave too, and find a new lo­ca­tion to col­lect head and wing sam­ples from hunters. The next morn­ing he was set up along Gun­bower Creek near Koon­drook. “It's been a chal­leng­ing week­end; the is­sues around clo­sures and so on has prob­a­bly meant we are down a few sam­ples but over­all it has gone well,” he said af­ter two days of col­lect­ing sam­ples. “The hunters are cer­tainly much more re­cep­tive to the pro­gram now than when we started, which was in 2009. “When I fi­nally do the num­bers on it we will have notched up more than 3000 sam­ples over those eight years.”

The sam­ples are used in con­junc­tion with meat ant sam­ples from across Aus­tralia. A sta­ble iso­tope sig­na­ture present in the duck sam­ples and the ants pro­vides ev­i­dence of the move­ment of birds.

The wing mea­sure­ments can also be used to pro­duce age data once base­line com­par­i­son mea­sure­ments from know­nage ducks are es­tab­lished. “We started the pro­gram be­cause up un­til then we might have had a num­ber of ducks taken in Vic­to­ria ev­ery year but we didn't re­ally know any­thing about that num­ber, so the whole pro­gram was de­signed around try­ing to get some bi­o­log­i­cal knowl­edge of the birds that made up that num­ber,” Pro­fes­sor Hall said. “We didn't know any­thing about the num­bers of each in­di­vid­ual species that got taken, we didn't know what sex ra­tios got taken and re­ally, where we want to end up, is that we have got all this in­for­ma­tion on un­known-age ducks and we would like to put that to­gether with a pro­gram on known-age ducks so that we can say, right, a wing length of X in an un­known age duck trans­lates to an age of X based on a study of known age ducks.”

While the sud­den clo­sure of Lake El­iz­a­beth to pro­tect a threat­ened species dis­rupted Pro­fes­sor Hall's plans, he's more concerned about what he says is a “hap­haz­ard” and un­sci­en­tific ap­proach to sea­son set­ting in Vic­to­ria.

He takes is­sue with the Eastern Aus­tralian Water­fowl Count and the key role it plays in sea­son de­ter­mi­na­tions. “The whole is­sue about har­vest of any­thing, whether it is but­ter­flies or ele­phants, is that is has to be sci­en­tif­i­cally ev­i­dence based and we re­ally don't have that for Aus­tralian game birds,” he said. “We have a num­ber, and a num­ber is just a num­ber; it is not telling you any­thing else. If you are go­ing to have ev­i­dence-based pro­grams you have got to know a lot more than just a num­ber.”

Pro­fes­sor Hall un­picks the Eastern Aus­tralian Water­fowl Count in this way: it started in 1983 and was de­signed to count Wood ducks in a small area of New South Wales. The method­ol­ogy out­lined in the orig­i­nal Com­mon­wealth Sci­en­tific and In­dus­trial Re­search Or­gan­i­sa­tion (CSIRO) tech­ni­cal bul­letins was never de­signed to be used to con­duct sur­veys from the Gulf of Car­pen­taria to Bass Strait. “They were also never de­signed to count all of the water­fowl species; they were never de­signed to count broods of duck­lings and all this sort of stuff,” Pro­fes­sor Hall said.

“There is a lot of sci­en­tific con­cern about that method­ol­ogy and yet it is re­ally the only piece of in­for­ma­tion used by the var­i­ous states. They say they use the South­ern Os­cil­la­tion In­dex and all th­ese other things but there is no cor­re­la­tion ever talked about be­tween all of those bits of in­for­ma­tion.

“It is re­ally a lick the fin­ger and stick it out the win­dow and say, ‘yeah that seems 'bout right, we'll do this'.”

Pro­fes­sor Hall calls it “sur­vey by rote”: fly­ing lines ev­ery two de­grees of lat­i­tude go­ing from north to south even if that means miss­ing sig­nif­i­cant water­fowl habi­tat.

One ben­e­fit of an arid land­scape is that it is well known where water is and Pro­fes­sor Hall has been part of a team in­ves­ti­gat­ing a bet­ter model that reg­u­larly sur­veys and counts spe­cific wet­lands, pro­vid­ing not only ac­cu­rate count data but trends over time. “We've been work­ing in NSW for the past 18 months or so look­ing at us­ing drones to count water­fowl and that is a much more rig­or­ous and sci­en­tif­i­cally re­pro­ducible way of do­ing it. We're tak­ing pho­tos on th­ese wet­lands so if I say there are 100 ducks there and some­one else says there are only 10, we can ac­tu­ally dig out the frames,” he said. “With the Eastern Aus­tralian Water­fowl Count all we are pre­sented with is a re­port each year that says there are a cer­tain num­ber of ducks on that tran­sect line.”

The drone data is com­bined with new com­puter soft­ware that uses satel­lites to show the lo­ca­tion and quan­tity of water. “It tells us how much free water, how much edge and how much dry land is around that edge,” he said. “We can work all that into a sen­si­ble, ac­cu­rate in­dex of abun­dance of water­fowl, and that is the other is­sue with the Eastern Aus­tralian Water­fowl Count: they will give you an ab­so­lute num­ber and we re­ally don't need an ab­so­lute num­ber. We need to know whether the pop­u­la­tions are chang­ing, go­ing up or down and whether they are chang­ing in their dis­tri­bu­tion; an ab­so­lute num­ber is pretty mean­ing­less.”

Pro­fes­sor Hall cites a re­view of Field & Game Aus­tralia's an­nual on-ground water­fowl counts that con­cluded that some years are bet­ter than oth­ers, some re­gions are bet­ter or worse than oth­ers and that within those re­gions, some wet­lands are bet­ter or worse.

>> “The ques­tion then be­comes why some wet­lands are good and some aren't,” Pro­fes­sor Hall said. “There wasn't time to an­swer that in the study but it shows us that in­stead of count­ing hun­dreds of wet­lands, which in an ideal world we would do but we're all time poor, let's fo­cus on where we can get the big­gest bang for the buck. “There are cer­tain wet­lands that could be sur­veyed each year, which would give us that rel­a­tive num­ber of how water­fowl are go­ing up or go­ing down. “By then plug­ging in the new soft­ware show­ing where the water is and whether it is go­ing up or down and ty­ing it all to­gether rather than re­ly­ing on one sin­gle piece of in­for­ma­tion, we can start to say, for ex­am­ple, that Gipp­s­land has huge num­bers of water­fowl on lots of ar­eas, the north-west of the state has large con­cen­tra­tions but on smaller num­bers of wet­lands and we can start to pro­por­tion the har­vest.”

In a pa­per on adap­tive man­age­ment Pro­fes­sor Hall high­lighted the cur­rent gaps in knowl­edge con­cern­ing the re­la­tion­ships be­tween duck pop­u­la­tion den­sity through­out the land­scape in space and time, and the abil­ity of hunted pop­u­la­tions to re­cover to pre-hunted lev­els.

He con­cluded that long-term mon­i­tor­ing was re­quired in the re­la­tion­ship be­tween habi­tat, den­sity and hunt­ing.

A new re­search pa­per pub­lished a few weeks ago de­tail­ing the un­manned drone trial in NSW pro­vides the plat­form for a shift to tar­geted mon­i­tor­ing of key water­fowl habi­tat.

The team, which in­cluded Pro­fes­sor Hall, found a fixed-wing drone could be flown as low as 60 m above the water level with­out dis­turb­ing water­fowl and multi-ro­tor mod­els could get as low as 40 m.

Us­ing flight paths that did not cause dis­tur­bance, com­mer­cially avail­able on-board cam­eras were able to cap­ture im­ages of suf­fi­cient qual­ity to iden­tify water­fowl and even much smaller birds such as swal­lows. “Our re­sults show that with proper plan­ning of take-off and land­ing sites, flight paths and care­ful UAV (un­manned aerial ve­hi­cle) model se­lec­tion, UAVS can pro­vide an ex­cel­lent tool for ac­cu­rately sur­vey­ing wild water­fowl pop­u­la­tions and pro­vide archival data with fewer lo­gis­ti­cal is­sues than tra­di­tional meth­ods such as manned aerial sur­veys,” the re­searchers con­cluded.

The drone counts, satel­lite data on water avail­abil­ity and the ev­i­dence from on­go­ing re­search on har­vested birds can build a much big­ger and clearer pic­ture for man­age­ment au­thor­i­ties. “It gets you into a modern par­a­digm of man­age­ment that is done all around the world called adap­tive man­age­ment,” Pro­fes­sor Hall said. “You are col­lect­ing in­for­ma­tion on the pre­vi­ous sea­son and you are mak­ing de­ci­sions based on that and ad­di­tional sci­ence and adapt­ing your man­age­ment for the next sea­son. “That is be­ing done all over the world, ex­cept for Vic­to­ria it seems.”

Graham Hall wants to see tar­geted mon­i­tor­ing of key wet­lands like John­son Swamp us­ing un­manned drones and satel­lites to build a more com­plete pic­ture

Graham Hall record­ing head and wind data from sam­ples col­lected from hunters along the Gun­bower Creek dur­ing the open­ing week­end of the 2016 sea­son

Newspapers in English

Newspapers from Australia

© PressReader. All rights reserved.