Isle Royale Wolves Used Beaver-Lodge Areas Most in Fall

STEP 1 · Lodge counts

October surveys recorded a 57 percent drop

Isle Royale National Park is an archipelago of 207 square miles (535 square km) in Lake Superior. The National Park Service flew October lodge surveys in 2018, 2020, 2021 and 2022. Active lodges in the analysis fell from 395 in 2018 to 168 in 2022, a 57 percent decline since wolf restoration. The intermediate counts were 386 in 2020 and 344 in 2021, so most of the drop came in the final year. The paper places that decline in the years since wolf restoration and does not test what caused each lost lodge.

The analysis used 834 unique active lodges. Passage Island was left out because wolves do not occur there. The authors drew a circle of 328 feet (100 m) around each remaining active lodge and treated that circle as habitat meant to cover lodges, dams and trails. That circle was mostly forest. Mapped cover inside the buffers was 28 percent dry mixed conifer, 22 percent mesic mixed conifer, 20 percent deciduous forest, 20 percent wet meadow or marsh, 9 percent water and 2 percent other. Crews in a fixed-wing aircraft on floats called a lodge active from a food cache, freshly peeled sticks, new cuttings, canals, fresh mud or a repaired dam.

STEP 2 · October peak

Fall fixes landed in lodge buffers most often

Founding wolves came from four mainland populations and were moved to the island from September 2018 to September 2019. Wolves were also captured on the island in April 2022. The habitat model used 18,932 GPS locations from 23 wolves in 2018, 2020, 2021 and 2022. The paper split the year on average interior-lake ice dates and on moose calving and calf timing. Spring was 15 April to 14 June, summer 15 June to 14 August, fall 15 August to 14 November and winter 14 November to 14 April.

Wolves selected beaver sites in summer, fall and winter. Absolute use was greatest in fall. They did not select beaver sites in spring. In the raw counts 134 of 2,254 fall fixes fell in lodge buffers, about 6 percent. Spring was 48 of 2,800, about 2 percent. Summer was 114 of 2,481, about 5 percent. Winter held 413 of 11,397 fixes in lodge buffers, about 4 percent. The winter count is larger because the winter sample is larger. The fall share is the largest of the four seasons. Across the study 709 of 18,932 wolf fixes fell in lodge buffers, about 4 percent. Paired random points put 319 of 18,932 in that class, about 2 percent. The table does not split available lodge habitat by season.

The day-of-year results show avoidance of lodge buffers from January through April and a change to selection in June. The discussion places lodge selection from July through November. Our read is to call that change early summer. Selection peaked in October and remained in selection into December. Winter selection in the season model was not what the authors expected. They had assumed beavers were mostly in lodges or under ice. They tie December use to ice they call likely broken and to daytime highs above freezing in records they cite. This study did not measure beaver activity above the ice.

STEP 3 · Marsh or lodge

Marsh use peaked in summer and lodge ground pulled ahead in late fall

Lodges sat disproportionately in or beside wet meadows and marshes, so the authors warned the two signals could blur. Wolves used lodge ground less than marshes in early spring and more strongly in late fall. The season model shows selection for wet meadows and marshes in every season, with greatest use in summer. The daily curve moves from marsh avoidance in January to selection in May and rises from January through July. Our read for this island is that marsh and wet meadow selection runs highest in summer, while lodge selection overtakes marsh selection in late fall even though marshes still held more raw fall fixes. The season model still shows marsh selection in winter. The daily curve still shows January avoidance.

The authors treat moose as the primary prey and beavers as secondary prey. They expected lodge selection to intensify through the ice-free season as beavers spent more time away from lodges and as moose became less vulnerable.

STEP 4 · Kill gap

The collars could not tell a pass from a hunt

The authors say wolves may have used the ground around lodges as travel routes or water sources. They may also have hunted moose, muskrats or fish there. Collars attempted a location every 4 to 5 hours and could not separate a passing wolf from a foraging wolf. Lodge selection in this paper rose through the ice-free season. The authors contrast that rise with a 2018 to 2019 summer diet study they cite, in which beaver hair in wolf scats stayed high and steady from May through October. They offer two explanations. The studies may differ in method and in power to detect change. The other explanation is that lodge selection can occur without a higher predation rate. They still read the timing as wolves hunting prey whose availability shifts through the year. Our read is that hair in scat shows consumption and can include scavenging. Tracks at a lodge do not identify a kill.

The authors also say their lodge-selection estimate may be biased low because part of each buffer was open water that wolves are unlikely to use. Random points could fall in that water. Water was 9 percent of the mapped buffer.

STEP 5 · Your water

Match a lodge against a meadow on the same days

Our read is a paired look on beaver water where wolves live and where access rules allow it. Pick one active lodge and a nearby wet meadow with no lodge. Check both for fresh wolf sign on the same days and with the same effort from early summer into December. Track each site against itself through the season rather than ranking raw totals, because mud at a lodge and meadow cover hold tracks differently. If lodge sign builds to a fall high while meadow sign peaks earlier in summer, the timing resembles the island. If lodge sign shows no fall rise, do not borrow the October timing. Matched sign can show where tracks turned up on the days you looked. It cannot recover the model's selection strength, a cause, which wolf left the tracks, or a population-wide result. It cannot show a kill or whether a colony survives.

Whether the lodge still looks active is a separate observation. Use the same signs the October crews used. If those signs are gone on a later October visit, the lodge would not meet the survey standard for active. Wolf sign alone cannot explain the change. On Isle Royale the value of this paper is observation and research only. The study cannot settle waters where moose are not the main prey, where lodges are scarce, or where hard ice arrives well before mid-November. Using the October timing on other wolf and beaver ground is our inference from one restored island population, four survey years and 23 wolves.