No Sign Anglers Threaten Glacier Bay’s Bartlett River Coho, Sonar Counts Show

The Bartlett River is a small wilderness stream in Glacier Bay National Park and Preserve, yet it draws anglers from far and wide and is one of the park’s most heavily fished rivers. When improved ferry service promised more fishers, park managers wanted to know whether its coho salmon could take it. A National Park Service study that counted the fish with sonar in 2012-2014 found a productive run and little angler harvest, and says there is no indication that recreational fishing is jeopardizing a natural and healthy coho population.

Field staff take a quick length measurement of a coho salmon captured during tangle netting. NPS / Craig Murdoch.

Why the count mattered

Pacific salmon are a symbol of wildness and a critical part of Glacier Bay’s ecosystem. They feed bears, wolves, harbor seals, bald eagles and people, and their decomposing bodies fertilize streams and nearby forests. Each summer the Bartlett has runs of sockeye, pink, chum and coho salmon.

Managers prioritized estimating coho escapement (the number passing upstream to spawn) and harvest because more fishers were expected, and set out to establish a baseline of abundance and harvest to ensure that current and future harvest is sustainable. Recreational fishing is a treasured pastime, and many parks were set aside partly to encourage it, and in 1918 the Interior secretary wrote to the Park Service’s first director that fishing was among the sports that would always be favored in national parks. Park Service policy adds that harvest is allowed only when monitoring is done and the Service has determined that it will not unacceptably affect park resources and natural processes. Yet in the past hundred years little new information on Bartlett River salmon has been gathered, and few harvest estimates exist.

A river with a salmon history

The Bartlett follows a lateral moraine left by the Glacier Bay ice sheet several hundred years ago, and in its roughly 250-year history it has changed with the ice sheet’s retreat, isostatic rebound and successional processes. Summer flows are typically between 100 and 300 cubic feet per second (2.8-8.5 m3/sec), small lakes in the middle of the basin buffer them and serve as sockeye spawning and rearing habitat, and the headwaters on Excursion Ridge are 22 river miles (35.4 km) from the mouth near Bartlett Cove. The ridge’s karst features have been linked to better fish productivity and invertebrate diversity and abundance.

The Bartlett River watershed drains Excursion Ridge and many small lakes into Bartlett Cove and is in many ways a relic of the Glacier Bay ice sheet. Map from Glacier Bay National Park and Preserve, National Park Service.

Early commerce suggests the river once teemed with salmon. A saltery at Bartlett Cove packed 120,000 pounds of sockeye fillets in the summer of 1899 and the next year salted about 24,000 pounds, or 3,500 coho, most believed to come from the Bartlett. A cannery operator later estimated the river held 75,000-100,000 large red salmon in 1901. The saltery went bankrupt in 1903, and by 1924 Congress had passed the White Act to limit commercial fishing; Glacier Bay, one of 15 areas in Southeast Alaska considered depleted of salmon, was closed to all commercial salmon fishing.

Alaska’s salmon fisheries are far better managed today. The state estimates that monitored Southeast Alaska coho stocks have substantial but moderate exploitation rates, 38-64% (average 52%) of annual returns from 2006 through 2010, and that they have stayed productive and healthy. Commercial harvest of Bartlett coho is likely average or below because the river is far from other intercept fisheries, and apart from commercial trolling, mostly outside Glacier Bay, probably only recreational anglers take its salmon.

How the fish were counted

Sonar suited a wilderness river because it has low impact. The Dual-frequency IDentification SONar (DIDSON) uses 96 high-frequency sound pulses (1.1 or 1.8 MHz) to make near video-quality images of migrating fish. The system recorded on its own, continuously, at 8 frames a second or roughly a gigabyte of data an hour, day and night, for each survey season. It needs only 45 watts, so a quiet, clean methanol fuel cell could power it instead of a gasoline generator.

The Bartlett River sonar site upstream of the partial weir; the green triangle marks the area the sonar images. Figure from the National Park Service article.

The biggest problem was that DIDSON cannot tell fish species apart. It measures length accurately and precisely, but length alone is a poor predictor of species when several species with overlapping lengths migrate together. A 2011 pilot study found that pink and chum salmon and, to a lesser extent, Dolly Varden overlapped with the coho in length and run timing. So the team sampled the migrating fish with a tangle net 40-50 feet (12-15 m) long, with mesh sized to entangle the head, jaws and teeth, to estimate the share of each species and so apportion the sonar counts to coho.

A raw sonar frame showing salmon-sized fish passing through the imaged area (left) and an echogram with eight measured fish (right). Figure from the National Park Service article.

What the counts showed

Sonar, tangle netting and creel (angler) surveys together revealed a productive coho stream with little angler harvest. The team estimates minimum escapement of between 4,600 and 13,900 coho over the three years, in line with the Alaska Department of Fish and Game’s finding that Southeast Alaska streams with lake systems typically produce escapements of 1,000-8,000 coho a year.

In 2012 too little net sampling meant fish could not be sorted by species, so that estimate counts all salmonids greater than 40 cm. The 2013 estimate is truncated because large numbers of pink salmon spawned near the sonar during the early coho run. In all three years the start of the run, which likely begins in early August, was missed.

Weekly estimates of upstream-migrating coho, 2012-2014. Peak timing and numbers varied from year to year. Chart from the National Park Service article.

Recreational effort and harvest varied widely, and the estimates have wide margins (1 SE is the standard error):

The 43% is biased high, because the escapement estimate for 2013 was truncated; had it started in the third week of August, as in 2012 and 2014, the harvest would be a much smaller share. The estimates are relatively consistent with the state’s Statewide Harvest Survey, which also suggests relatively low and variable harvest and effort compared with other streams of similar size in Southeast Alaska.

The team points to several likely causes of the swings: environmental conditions, how fast word of good catches spreads, and the size of the run. September precipitation at Bartlett Cove averages 10.02 inches (25.5 cm); 2012 was average and 2013 nearly 2 inches (5 cm) below. Wet fall weather is often good for salmon migration but makes fish harder for anglers to target and degrades the trail on the 2-mile (3.2 km) hike in, while dry, low water holds salmon in the lower river where anglers can reach them. When catch rates are high, word travels fast and more people come; poor success has the opposite effect. The relatively high harvest in 2013 also reflects high returns, since the state reported record salmon returns across Southeast Alaska that year.

What it means

The authors conclude that, despite concerns about more recreational harvest from new ferry service to Gustavus, there is no indication that it is jeopardizing a natural and healthy coho population. Effort and harvest are low and variable, and although the largest part of the harvest, the commercial catch, is unknown, the escapement estimates indicate the river exceeded the biological escapement goals the state set for other similar Southeast Alaska drainages (3,300-9,800 for the Situk River and 1,300-2,900 for Ford Arm Creek). There is no current threat of impairment, they say, but the work shows the value of a baseline as a reference for judging future change, and managers need clear objectives and biological thresholds. With climate change, ocean acidification and growing human demands on the horizon, baseline information is the first step in avoiding diminished, less resilient ecosystems.

Sources

• Craig Murdoch, National Park Service, SONAR Project Provides Coho Salmon Baseline for Glacier Bay National Park and Preserve, Alaska Park Science 19(1), 2020 (Below the Surface: Fish and Our Changing Underwater World); a source estimate by a naval commander is left out because its citation year disagrees with the year it is given for, and so is the source’s year for Congress creating the Park Service, which looks wrong; rewritten in hubnx's own words.