Wyoming pronghorn crossed turbines and paused farther in spring

STEP 1 · Fall routes

Fall paths ran closer to turbines than habitat predicted

Distance to existing turbines mattered for whole-route selection only in fall. After habitat was included, females selected routes closer to turbines than expected. Distance to turbines under construction also mattered for route selection only in fall. Females selected those construction-period routes closer than expected. Fall stopovers were not selected farther from turbines. Fall steps were not selected farther from turbines once habitat was included. In the fall step model the credible intervals on the turbine terms overlapped zero. Our read is that fall routes in this sample leaned toward the towers once habitat was accounted for. One fall crossing in the glass only shows that animals used that view.

Near turbines, females moved faster in spring and slower in fall. Spring green-up and fall sagebrush cover were the strongest habitat predictors of speed, and turbine distance stayed in the model after them. Winter speed had no strong predictor among the habitat or turbine terms. The authors suggested animals slowed in fall because turbines sat in high-quality habitat they still used. They also raised untested hunting pressure, with facilities as possible refuges. That suggestion is not evidence turbine ground hunts better. No speed difference was detected before construction versus during construction at Ekola Flats or TB Flats. Samples were too small to split by season. Our read is that the miss does not prove speed was unchanged. The authors pointed to snow correlated with distance to turbines under construction as one account of that null result. Construction of Ekola Flats (63 turbines) and TB Flats (132 turbines) ran from April 2019 to July 2021. Second-window monitoring ended in summer 2020.

STEP 2 · Spring pauses

Stopovers sat farther from turbines in spring and winter

Females selected stopover sites farther from turbines in spring and winter. Fall did not follow that pattern. A stopover here means the top 10 percent of each route's Brownian bridge use area. The model compared those slices with other polygons inside the same individual's route. Moving from the tower out to 0.62 miles (1 km) corresponded to a 33 percent increase in relative stopover selection in spring and a 300 percent increase in winter. Those figures are relative probabilities from the fitted stopover model.

On small-scale steps the same distance contrast corresponded to a 180 percent increase in relative selection in spring and a 68 percent increase in winter. The authors reported a 0.95 to 0.97 probability that spring and winter step selection favored ground farther from turbines, with large error around the estimates. Spring and winter steps were farther from turbines after habitat was included. Our read treats the 0 to 0.62 mile contrast as a step inside the model. It does not mark a ring where avoidance begins.

The Results say time at stopovers increased farther from turbines only during spring. Time was the count of GPS points in the polygon. The Discussion says fall showed the opposite time pattern. The Results text gives no fall estimate. Distance to turbines under construction was not an important predictor of stopover selection in spring or fall. Winter was not included in the construction models because the sample was small. A missed effect is not evidence the effect is zero.

Fixes in 2010 to 2012 were on an 11 hour or 7 hour schedule. In 2018 to 2020 collars used a 2 hour fix and transmitted every fourth location. Full 2 hour data came from 90 animals that died. Our read is that multi-hour fixes make point counts a coarse clock.

STEP 3 · Loose trails

Stopover overlap was low, and route overlap fell near construction

Mean route overlap for females tracked across two years was 0.29 plus or minus 0.04 in spring and 0.30 plus or minus 0.03 in fall. The spring sample was 25 migrations and the fall sample was 32. Overlap is the share of the second route that sat on the first. Stopover overlap averaged 0.06 plus or minus 0.03 in spring and 0.04 plus or minus 0.01 in fall. Our read is that one pause spot is a weak bet the next year.

In 2018 to 2020 route overlap rose farther from facilities under construction in both spring and fall. Samples were 21 spring routes and 30 fall routes. Spring stopover overlap also rose farther from turbines under construction. Fall stopover overlap had no strong tie to that distance. Our read is that last year's line is a weaker bet close to active construction.

STEP 4 · Open glass

A pause in the glass only describes that view

Our observation is optional and the paper did not run it. If you already glass a crossing beside a wind facility, note where groups pause and how much of the ground you can see sits close to the towers. Our reading of the spring result is that groups would pause close to the towers less often than the share of visible close ground suggests. Fall stopovers in this study were not selected farther from turbines. Our reading is that the spring pause pattern does not carry into fall on this evidence. The tally can establish only the pauses and the ground in that view. It cannot recover the model percentages or name individuals or describe the population. It cannot stand in for the GPS selection model. Visibility and the share of the route you can see will change the count.

Our rule drops a crossing only after absence repeats across years. Most migratory females here were conditional migrants, and that strategy can change between years, so one empty season is weak evidence the animals have abandoned it. Our rule drops the spring expectation when pauses near the towers are about as common as the visible ground near them. This paper cannot settle males or places beyond south-central Wyoming. It cannot settle forage intake, fawn survival or a filled tag. Where these females used habitat is not a forecast of hunting success.