Drone surveys provide fresh data on Unalaska’s fluctuating salmon stocks
New data from drone surveys flown over Unalaska’s three road-system lakes last summer show low sockeye salmon counts.
The counts total less than half of what they were in summer of 2020, according to data released in April by the Alaska Department of Fish and Game.
But Fish and Game biologist Tyler Lawson said the one-year drop isn’t too concerning. Escapement numbers often fluctuate and there’s more room for error in aerial surveys, he said.
“We call them a ‘high error survey,’ which kind of sounds bad, but it's just because in comparison to the weir — which is a very precise tool — there's variability whenever you're up in the air, looking down and trying to count salmon,” he said.
While the technology is still relatively new when it comes to counting salmon in Unalaska, Lawson said he’s hopeful that drones will play a key role in helping assess broader trends among salmon stocks in the region.
He ticked off their benefits. For one, he said drones are a lot cheaper than intensive studies like fish weirs — where a manmade gate is set up across the mouth of a lake or stream, allowing biologists to count and observe the fish that pass through.
Also, it’s easier to maneuver a drone into tight spaces than a fixed-wing plane. Plus, using the drone data alongside other types of research, helps create a more comprehensive understanding of the health of local salmon populations, he said.
“[Drones] are a newer kind of proprietary tool that don't have this more established background to them,” Lawson said. “So we're hoping to couple established ‘mark and recapture — tagging’ methods with the drone surveys just to kind of further validate them as a tool for estimating population sizes.”
The Federal Aviation Administration authorized the first commercial unmanned aircraft flight over land in 2014, for aerial surveys in Alaska. In 2016, the FAA then introduced rules for non-recreational use of drones, which detail operational, aircraft and licensing requirements. Just two years later, the unmanned aircraft were introduced locally for salmon surveying.
Prior to that, the island’s only documented salmon streams were at McLees Lake in Reese Bay. The weir at McLees was established in 2001, when local residents and Fish and Game were concerned that lack of reliable escapement estimates for salmon in the lake could jeopardize the health of the run, as well as future opportunities for subsistence fishing. Scientists continued monitoring the sockeye runs at McLees for 17 years until Fish and Game lost funding in 2018 and closed the weir.
As news about the closure circulated, the Unalaska Native Fishermen’s Association teamed up with local Andy Dietrick and provided funding for his drone business, Aleutian Aerial, to perform surveys on Summer Bay Lake, Iliuliuk Lake and Morris Cove Lake.
Eventually, other organizations like the Qawalangin Tribe, the City of Unalaska and Fish and Game also stepped in to help fund the surveys. For the past four years, Fish and Game has been analyzing data provided by Dietrick’s business at the three road-system lakes. With funding help from the Qawalangin Tribe, the weir reopened in 2020, and Dietrick started surveying McLees Lake as well.
That year was the worst on record for the island’s weir. The drone surveys in 2020 showed a similarly grim count, with a total indexed escapement of around 2,400 sockeye at McLees.
Last summer, though, the number ballooned for both the weir and drone surveys, which jumped to more than 13,000 salmon.
Lawson said the main reason for that big increase in the drone survey, and why it was more comparable to weir counts, is because Dietrick surveyed inlet streams where sockeye spawn — which he hadn’t done the previous year.
“There are literally thousands of sockeye that go up those little inlet streams to spawn,” he said. “So it was great that [Dietrick] was able to get way back there with the drone. Whereas if you have a [Cessna] 185 and you're trying to get back there, you don't have nearly as much wiggle room. So there are definitely salmon back there. And it was definitely a benefit to the surveys.”
Scientists counted less than half the number of salmon using fixed-wing surveys than they did with the drone surveys at McLees last summer.
Drone data is also useful in Unalaska because it’s not always feasible for Fish and Game to get resources like piloted aircraft so far west, Lawson said. Even when they can, temperamental Aleutian weather often makes it a challenge to fly the fixed-wing aircraft.
“The weather is very volatile there,” he said. “And it's either not safe or just not practical to actually fly around in an aircraft at all times, and actually try to get a good count of the salmon.”
In the past, Lawson said, that meant Fish and Game could only get aerial data once every couple years.
“We really haven't been able to get as much data as we would like to have in order to manage fisheries,” he said.
But with Dietrick on the ground, living on the island, Lawson said he’s able to get three to four surveys on each lake over the summer, which gives the scientists a more comprehensive picture of the peak run numbers. He said it’s also a lot easier to work with video footage as opposed to doing the counts in real time from a plane.
“To be able to sit at your computer and review this footage and pause it and rewind it and put it in slow motion — it's just this really awesome tool, as a biologist to have it at your hands, to estimate those salmon numbers and then make it available to the community so that they have an idea of what's going on with their salmon,” Lawson said.
But before the footage arrives at the biologists’ desks, Dietrick has to find the right weather conditions during the runs to get useful video. He said the process involves a lot of patience and attention to detail.
“You need your light winds, so the water surface isn't stirred up,” Dietrick said. “You need some sunlight, so you can't just have cloudiness, because then the water surface sort of has this cloud glare, and you can't see into the water very well.”
Dietrick has a background in weather forecasting and working with salmon, which give him a leg up in this type of drone work. But even with that, he said a key part of the job is just being ready to go when good weather hits.
“You’ve got to be able to pounce on it and go execute when that comes,” he said. “So I'll be up till the night before watching that weather forecast model run for the next day sometimes, and make a decision, ‘Hey, tomorrow morning at 6 a.m., we're headed to McLees.’”
On top of all that, Lawson said it gets particularly tricky counting sockeye from the air, which is the fish most commonly harvested by Unalaska subsistence fishers and makes up most of the data Fish and Game collects.
“Sockeye salmon, especially, kind of do this thing where they're huddling right where the water starts to get deeper, darker and murkier,” he said. “So it can be challenging to get a really accurate count of them, even on the best weather days.”
But in combination with weir counts, he said the drone surveys provide a more comprehensive picture of how the stocks are doing.
Over the past four years, he said they’ve been able to make major improvements in how they carry out the surveys. And if they’re unable to secure funding for the weir again, Lawson said they will hopefully have the drone surveys to fall back on.
“We’re trying to hone them in and make them as accurate as possible to serve as kind of a backup tool,” he said.
Weather depending, Lawson said they’ll be getting the count started out at the McLees Lake weir by the second week of June.