By Jenny Neyman
Images courtesy of Debby Burwen, Alaska Department of Fish and Game. The dual-frequency identification sonar system being tested on the Kenai River creates images of fish on a screen, making it easier to count and size them, and differentiate kings from sockeye salmon.
Redoubt Reporter
Judging by the in-river sonar estimate numbers, this summer’s late run of Kenai River king salmon had the potential to be a smoking hot fishing season. The count started July 1 with 843, hit a high daily total of 2,654 on July 21, and finished strong with a 1,277 daily count Aug. 4, for a healthy cumulative total of 48,343.
But those numbers did not translate into bent rods, wet nets and happy fishermen. Far from it. Anglers on the river, harvest reports, creel surveys and test netting all tell a far different tale of Kenai king fishing this July. Fishery managers are estimating the run to be at the lower end of the escapement range, around 17,800 fish, with a weak run total in the 32,000 range, rather than the robust 48,000-plus indicated by the sonar estimate.
Kristy and Evertt McCullough, of Alaska’s Last Frontier Fishing Lodge, were two of many guides and private anglers left scratching their heads over the dichotomy between high expectations for a good season and the poor results that were actually netted. Sonar counts were good, water conditions were clear and favorable for king fishing, and boat traffic seemed a little lower than usual, Kristy McCullough said. But all that accounted for a big fat nothing of a season, as the kings simply didn’t show up.
“They weren’t there,” McCullough said. “River conditions were good. That is the scary thing. Last year we had flooding and very dirty water for a while there, but this year we had beautiful, clear-water days and still had very low catch rates. And that’s just the double whammy when fishing’s bad even with fewer boats out there. Even the drift-boat days. The very last drift-boat-only Monday, it’s usually just a blowout day, and this year we actually got one (king) on that Monday, but we only saw two others caught and we were out there that whole day.”
Low catch rates weren’t in the McCulloughs’ or anyone else’s imagination.
“It was a poor run. All indicators are it was a weak run,” said Robert Begich, area management biologist with the Alaska Department of Fish and Game Division of Sport Fish.
The older split-beam sonar system, right, results in dots on a screen.
All indicators, that is, except the sonar count, which clocked a healthy late-run king return. Debby Burwen, regional sonar biologist for Fish and Game, said the sonar was likely fooled by a combination of factors. The king run unexpectedly consisted of a lot of younger, smaller kings, which can be difficult for sonar to distinguish from sockeye salmon. And sockeyes, for some unknown reason, swam in the middle of the river more often than they usually do. Sockeyes usually tend to stick closer to shore, whereas kings are more often found midriver.
“We believe (the late-run king sonar counts) were biased high, and that’s no secret. We put it in the paper and on our website, everything like that. And the reason we think they were so biased high was because there was an extraordinary number of sockeye out in the middle of the river,” Burwen said. “Both the split-beam (sonar) and netting program (which is only operated midriver) showed we had more sockeye out in the middle of the river than prior years. We always get some sockeye out there, but this year was an extraordinary high number.”
Along with the daily and cumulative totals, Fish and Game posted a disclaimer noting that the sonar estimate was biased high by an unknown amount, essentially warning anglers that the counts should be taken with a grain of salt. Still, that didn’t stop some fishermen from feeling like the high estimates were salt in the wounds of low king harvest.
“I get it. We’re well aware of it. That’s why we put that note on the sonar page and also added all the (run estimate) models to our summary page,” Begich said.
But if the numbers are wrong, and Fish and Game knows the numbers are wrong, why release them to the public?
“I honestly feel that if those are incorrect numbers they shouldn’t even be published if they know they’re incorrect, because those numbers are quoted then by policymakers and agencies, they’re out there in a newspaper every day, on the phone line every day. If they’re really incorrect by an unknown quantity, they need to just stop publishing them,” McCullough said.
The department has tried doing just that when biologists know the sonar count is off, Burwen said, but it resulted in even more blowback than posting incorrect numbers with a disclaimer.
“At one point we said, ‘We’ve got to stop releasing these numbers each day, because people take them literally.’ And when we did do that there was a huge outcry and people said, ‘Please put them out there, just let us know when you think they’re biased so we don’t put too much weight in them.’ And that’s what we have done, but that doesn’t seem to help,” Burwen said.
Anglers are hungry for data to help them increase their odds of a successful king trip, and sonar counts are seen as a good way to gauge when to hit the river.
“We’re in a difficult position because the public wants as much information as possible, and they want it by noon the next day. And that may not be realistic if they really want to give the biologists time to put together the picture they want,” Burwen said. “People may not even know what we mean by ‘positively biased.’ I think it’s part of the new era where people want to be more involved and more knowledgeable in what you’re doing, and that puts the responsibility on you to be able to explain it to them.”
The problem seems to be a lack of explanation from the department and a lack of understanding by the public that the sonar counts are just one tool for estimating run strength, Burwen said. Certain conditions can skew the sonar counts, and when that happens, biologists employ other methods to estimate and manage the run, and the public shouldn’t rely on biased sonar numbers to manage their fishing plans or expectations, either.
“I can see where the problem is,” Burwen said. “It says they’re ‘biased high by an unknown amount,’ but what does that mean? That’s supposed to be our signal to the public that, yes, we know the sonar isn’t right and we’re not taking it literally and don’t think we’re being fooled by these high numbers. But the problem is people are using them in ways maybe we didn’t anticipate, like in planning their trips. The important thing for people to know, and this is where we need to do a better job of explaining to people, is that when the sonar is biased high, they have other tools to use and that they did use this year to manage the fishery. They weren’t managing off of the sonar numbers.”
Fish and Game has different methods for estimating run strength that are employed when the traditional method isn’t accurate. The indices also were mentioned in the late-run in-season data summary for the public to be aware that there are other methods for assessing run strength.
“We use them all to benchmark the fishery,” Begich said. “It’s when we’re dealing with uncertainty (in the sonar) that we present them all to the public, but they’re always available to us to use.”
The problem for fishermen is estimates generated through other methods aren’t generally available to the public until the season is over, whereas sonar is the most immediate way to make information available during the actual run.
“The managers are very good there, very experienced. If all they had to do is get a (sonar) number and just go, ‘OK let’s do something based on one number,’ anybody could do it. A computer could manage the runs. But for both the sport and commercial fisheries, they’re realists. They know that no one index they look at is perfect and that they all have their strengths during certain times, and they look at all available information then evaluate the run strength and what they think they should do,” Burwen said. “The thing we’re concerned mostly about is protecting the resource itself, so that’s why we’ll manage using the best tools, but the public is always seeing the sonar, so we have to address that through better communication with the public.”
Despite the high bias from the late-run Kenai king sonar count, Burwen said this summer was actually a success for the Kenai River sonar program overall, since technicians operated the old, split-beam sonar system side by side with a newer, dual-frequency identification sonar system that is being tested to replace the split-beam technology, used for counting kings, and the Bendix sonar used for sockeye.
Since 2002, biologists have been making a slow, studied switch to DIDSON, which is touted as being an easier, more accurate method for counting and sizing fish. DIDSON creates images of fish on a screen that can be easily measured, recorded, archived and played back. Split-beam sizes fish, too, but much less accurately and the data is harder to interpret.
“We were able to keep both systems in place, positioned side by side so we could compare and contrast the estimates from each sonar,” Burwen said. “We just wrapped it up and the DIDSON provided improvements in both detecting fish and sizing fish.”
So far, tests are showing the DIDSON to be less susceptible to bias in misclassifying fish than the split-beam. Split-beam sonar accuracy relies more heavily on range separation between kings and sockeye, with sockeye swimming nearer to shore and the split-beam sonar focusing on fish in the midriver, where kings swim. But if sockeye are in the middle of the river, as they were this year, split-beam can have a tough time distinguishing them from kings, especially when kings are small.
When split-beam sonar hits a fish broadside, it gets a strong signal return that is used to count and size the fish. But if a fish swims through the beam at an angle, the return signal is weaker, making accurate sizing difficult. And sockeye tend to swim through the beam perpendicularly, presenting a nice broad side for a strong signal return, whereas kings tend to change angles more as they swim through the beam. So if sockeyes are in the middle of the stream, as test netting showed they were this year, they can easily create error in split-beam sizing, or at least the potential for error.
“Whether it gets it right or not, split-beam sonar estimates have a high level of variability based on the fact that fish don’t swim exactly true, and based on what we refer to as high levels of background noise (confounding noise the current and river bottom and surface). The higher the level of variability, the less reliable the size estimates are,” Burwen said.
Biologists still have more analysis to do, but initial results show DIDSON isn’t as susceptible to inaccuracy in counting and sizing kings.
“What we’ve seen so far is (DIDSON) doesn’t tend to be biased in the way it sizes fish, and it is far more accurate at sizing fish than the split beam,” Burwen said.
DIDSON works at a much higher frequency than split-beam sonar, and produces images of fish from which technicians can take size measurements. DIDSON’s main limitation is its relatively shorter range. Split-beam can be used up to 100 to 200 meters, whereas DIDSON’s range is more around 50 to 60 meters, but that’s plenty far enough to count fish in the Kenai.
Burwen said the sonar transition won’t be complete for a few more years but that Fish and Game is actively working on how to implement it, and it may occur in several phases.
“We’re still working up all that data, and we’re being very careful because if we switch to something new we want to make sure it’s been thoroughly tested so we don’t find ourselves in this position again. Eventually, we will have a report out there with DIDSON numbers,” she said.
Until the switch from split-beam is complete, anglers won’t get the full benefit of DIDSON’s promising higher accuracy for counting Kenai kings. In the meantime, what was a good news, bad news year for the sonar program, showing DIDSON can work better in challenging sonar conditions, is still just bad news for anglers expecting to use sonar counts as a gauge for fishing.
But even a poor fishing season can result in some good, Burwen said, if it results in the public being more knowledgeable about the sonar program and fishery management.
“That’s one good thing about a bad year like this, it gets people’s attention. And it’s also a Board (of Fish) cycle. Every few years the Board of Fish reviews the management plan and it’s our opportunity to say, ‘Hey, this is where we are, and these are our plans for moving forward.’”
