Plains sharp-tailed grouse leks tracked surrounding grassland

STEP 1 · Neighborhood scale

The 5 km square outranked the pixel under the lek

The study area is the prairie ecozone of southern Saskatchewan, in the northeastern range of plains sharp-tailed grouse (Tympanuchus phasianellus jamesi). Burda and colleagues fit a MaxEnt model to leks kept from 1990 onward and to a 2010 landcover map from Agriculture and Agri-Food Canada. Records came from ministry lek surveys, their 2018 and 2019 routes, Conservation Data Centre locations from 1995 to 2016, and morning dancing reports from 2017 to 2019. Only leks with three or more displaying males entered. Citizen-science points kept the analysis presence-only. The authors call the result an index of lek occurrence that does not necessarily reflect brood-rearing or winter habitat.

Contemporary crews located 147 leks. They found 87 in 2018 and added 60 in 2019. Only 28 of those 147 were already in the historical provincial survey records. Only 25 routes repeated historical routes. Our reading is that this overlap is a record match on a partial repeat of old routes. After points within 5 km of each other were collapsed to one record, 185 leks remained for the habitat model. Where a predictor was fit at both scales, the 5 km square outranked the 1 km pixel. The abstract gives grassland a permutation importance of 34.8 at 5 km and 26.9 at 1 km. The results give those two grassland terms a combined permutation importance of 61.7. Those values rank predictors inside the model and say nothing about the share of leks or birds. Terrain roughness came next at 9.7. Built-up cover in the 5 km square followed at 5.2.

They chose 1 km pixels to match provincial mapping and because some older lek records were known only to a quarter-section. That unit is 800 m on a side (about half a mile). The 2010 landcover came from a 30 m product that leaves out small features such as power poles. Neighborhood shares were measured inside the 5 km square centered on the lek. Held-out leks gave a testing AUC of 0.748. A separate set of 77 April and May reports from 2017 to 2021 with at least three grouse gave a CBI of 0.714. The authors read those scores as good predictive ability.

STEP 2 · Grass curve

Suitability rose with grass and bent near three-quarters

Habitat suitability rose with the grassland share of the 5 km square and started to fall near a proportion of about 0.75. At the 1 km pixel the grassland relationship stayed positive across the curve they show. The text gives no size for the drop past 0.75. The management statement for a landscape run from this model keeps grassland near 0.80 and keeps anthropogenic cover under 5 percent. The 0.80 figure sits above the fitted bend, and the text leaves the two unreconciled.

Our read is to favor a mostly grass block and to stop expecting a higher score once grass passes about three-quarters of it. On a block of about 9.7 square miles that share is roughly 7.2 square miles of grass. Grassland stood in for cropland because the two ran opposite each other in this landscape, so a high grass share mostly means a low crop share. The methods define that layer only as cover dominated by graminoids and forbs and do not say how seeded hay or tame pasture was classed. The bend applies to this southern Saskatchewan lek set on the 2010 landcover.

STEP 3 · Gentle country

Flatter ground ranked above rough country

Roughness was used only at the 1 km scale and was taken from the eight pixels around the lek pixel. Areas with lower roughness were more favorable for lek occurrence. The small rises are the authors' impression from the 2018 and 2019 surveys. They suggest future work finer than the 1 km grid could test whether males pick hills for sight lines or for some other trait. Our read is to start in gentle country and then check the low rises there. Repeated mornings with qualifying leks on the rough roads and quiet gentle roads would lead us to drop this ranking on that prairie.

STEP 4 · Built cover

Suitability dropped between 0 and 0.05 built-up cover

The fitted curve dropped sharply between a built-up share of 0 and 0.05 in the 5 km square. Shares above 0.05 had low suitability. The class covers roads, lots, buildings, and other ground where people have cleared vegetation or put up structures. On a block of about 9.7 square miles a 0.05 share is roughly 309 acres. The 30 m product leaves out small features such as power poles. The authors say a finer layer would be needed before small roads, wells, and power lines can be judged.

Tree cover at the 5 km scale rose with lek suitability. The display still needs open ground. Tree cover was under 5 percent of the study area and was mostly limited to river valleys and small forested patches. The authors place high tree cover in leftover natural ground that typically sits inside protected areas. Our read is that the tree signal may partly mark that remnant ground. Inside a park, use it as context for observation. Open grassland is still where this lek model places the display.

Wetland suitability was highest from none to a 0.1 share of the 5 km square, and high wetland cover ranked poorly. Shrub cover was a weak predictor, and the authors suggest low shrub patches missing from the landcover may explain it. Vegetation height was left out because a province-wide layer was impractical and height shifts with land use and weather.

The authors cite earlier work that sharp-tailed grouse may keep a disturbed lek as recruitment falls. They also cite a competing idea that development can aid nest success by displacing predators. This paper did not count recruits, predators, or nests. Our caution is that an active lek beside built-up ground leaves the condition of the surrounding block unproved.

STEP 5 · Road check

Several mornings can show which road holds leks

Our test is optional and local. Choose two roads in the same prairie. One crosses gentle grassland in a block about 3 miles on a side with little obvious built-up cover. The other crosses rougher ground or a block with a plainly larger share of buildings and pavement. Run the pair on several display mornings.

Historical crews stopped every mile (1.6 km) for at least two minutes. They used dry mornings and wind below about 15.5 mph (25 km/h). Borrow that habit on a calm morning if you want a similar listen. If you want their lek cutoff, count three or more displaying males. The 2018 and 2019 crews dropped the flush counts that the older surveys sometimes ran, to avoid disturbing the birds. If you watch a lek, give it at least ten minutes from the stop and then leave it alone.

What the mornings can establish is which of the two roads produced more qualifying leks on the days you were there. The MaxEnt weights, the size of any grass bend, nest survival, brood performance, and fall hunting outcomes stay outside that tally. This paper did not compare how easily leks are heard on rough ground and on gentle ground. A quiet rough road is weak evidence. A tie across a few mornings is weak evidence that the pattern is absent. Repeat the pair before you drop the ranking on that prairie. If the gentle grass road keeps the leks, you have repeated presence on those mornings.

Lek records from 1990 onward were matched to one 2010 map. Newer landcover, another subspecies, another region, and the sight line from one hill sit outside what this fit can settle. Season dates and method rules were outside it too. Check the current provincial or state agency page before you hunt.