Ya Ha Tinda summer ranges split encounters from elk in the scat

STEP 1 · Three ranges

East migrants, residents, and west migrants ranked different predators

This partially migratory herd summers along the eastern slopes of the Rocky Mountains in and next to Banff National Park. Historically more than 90 percent of the elk migrated about 31 miles (50 km) west from the Ya Ha Tinda winter range into the park. Earlier herd studies cited here put the migrant to resident ratio at three migrants per resident and later at one migrant per resident over the last two decades. More elk were also moving east into country affected by forestry.

The ranks in this section come from their scat-based index. Encounter and elk in the scat are separate parts of that index. Neither part is a count of every kill.

On that index, east migrants carried the highest risk from cougars. Cougar encounter values rose from low in the west to high in the east where forest-managed land had more edge. Cougar scats were more common where conifer cover was lower and forest edge was higher. That selection model used 42 cougar scats. The chance that a cougar scat held elk was highest in the east. Our reading is to hold the cougar diet rank more loosely than the wolf and coyote ranks. The content models used 33 cougar scats. Uncertainty was greatest there because the sample was small, and five models shared the top level of support.

Elk that stayed on the Ya Ha Tinda winter range through summer carried the highest risk from wolves and coyotes on that index. Wolf encounter values were highest on that resident range compared with the other summer ranges, and bear encounter values there were well below the western range. Elk in wolf scats and elk in coyote scats averaged highest on the resident summer range. Coyote encounter values stayed fairly even across the study area. Our reading is that the resident coyote rank came mainly from elk in coyote scats.

Elk were more likely to meet bears along western river drainages than in the east. MacAulay and colleagues scaled encounter, elk in the scat, and combined risk from 0 to 1. Mean bear encounter was 0.43 in the west, 0.24 on the resident range, and 0.19 in the east. Mean elk in bear scats was 0.14 in the west and 0.25 on both the resident range and the east. Mean combined bear risk was 0.21 in the west, 0.20 on the resident range, and 0.16 in the east. Our reading is that the lower diet share in the west still left combined bear risk near the other two ranges. The paper does not say bears killed few western elk. Encounter rate, predator numbers, and calf vulnerability are separate questions.

Combined risk from bears, cougars, coyotes, and wolves was highest on Ya Ha Tinda. That total came from moderate to high risk from all of those predators except cougars.

STEP 2 · Scat contents

Bear scats held elk less often than wolf, coyote, or cougar scats

In the MacAulay and colleagues survey, detection dogs searched from 1 July through 30 September in 2013 through 2016. Scats judged to have been deposited before 1 May were dropped. The July through September walks still included deposits from the May and June calving season. Elk in these scats can include calves. The day a scat was deposited is unknown, and the authors expected little resolution inside the summer.

The content check covered 130 bear scats, 114 coyote scats, 199 wolf scats, and 33 cougar scats. Elk was in 19 percent of the bear scats, 36 percent of the coyote scats, 38 percent of the wolf scats, and 46 percent of the cougar scats. Pairwise tests among coyotes, wolves, and cougars all had p at or above .27. A missed difference is not proof the true rates match. Each test of bears against those three had p below .001. Grizzly and black bear scats were pooled because visual species calls did not hold up against DNA in 24 checks.

Elk in a scat means the animal was consumed, and consumption can include a scavenged carcass. The authors treat the close match with kill sites as a sign of limited summer scavenging for this predator set. That is their interpretation rather than a count of scavenged carcasses. A scat can be left away from the carcass after travel and digestion. The analysis windows were about 1.9 miles (3 km) for wolves, 0.9 miles (1.5 km) for bears, and 1.2 miles (2 km) for cougars and coyotes.

Models of elk in a scat most consistently included higher herbaceous biomass, with cougars outside that pattern. The top bear model paired higher herbaceous biomass with elk in the scat and paired open cover with a lower chance of elk in the scat. The authors say these models can reflect where elk were more vulnerable or where more elk were available.

STEP 3 · Kill check

Combined scat risk tracked the kill-site model at broad scale

In the same MacAulay and colleagues comparison, the Spearman correlation was .98 with p below .0001 on 1,000 random points. The kill-site model used 104 summer elk deaths from 2002 through 2016 that were classed as predator kills rather than scavenging. The authors note that sample took 14 years and more than 650 collared-elk years to accumulate.

Scat-based selection and collar-based selection are a separate comparison. Wolf values at single points correlated at .18 on 1,000 points with p below .001. Bear point values correlated from .17 to .25 by season with p below .001, and those relationships were not strongly linear. In 10 bins the wolf correlation was .92 and the grizzly bear correlation was .68, both with p below .0001. The .68 figure uses grizzly telemetry, while bear scats in this study pooled grizzly and black bears. Our reading is that the scat-based and collar-based selection maps agreed weakly point by point but much more closely when binned, so the scat map suits broad patterns better than single locations.

The authors' check of the spatial pattern rests mainly on predation known for female elk. The authors know of no method that sexes elk hair in these scats, and the elk consumed can be male. Agreement with that cow pattern does not establish the same map for calves or bulls, and it does not establish hunting success.

STEP 4 · Field choice

Rank elk consumption only from a validated diet or kill

Our rule is to class the evidence before ranking which predator matters on summer elk ground. Presence is a track, a sighting, or a scat that has not been identified with a validated method. Presence shows that a predator uses the place. It does not show that elk were eaten. A lab hair call, a DNA diet call, or a kill classed by an explicit method can support a consumption rank. This study needed those checks. Visual separation of the two bear species failed against DNA, and deer-family hairs are hard to tell apart. Our test does not ask you to name elk hair by eye or to assign a mark to one predator from a guess.

What your notes can establish is whether you hold presence only or a validated diet or kill record on that ground and in that season. They can also show whether those classes point at the same predator. They cannot reproduce this scaled risk index, name an individual elk, or stand in for another herd.

If the summer notes are presence only, leave elk consumption unranked. If a validated diet or kill record exists, rank from that local record and keep unmarked sign as encounter information. Change the choice when the classes disagree, which is what the western bear result did here. Carry the Ya Ha Tinda directions to other country only when your own validated records show the same arrangement. The western result lies largely in Banff National Park. Read it as research on encounters and diet, and leave harvest planning off that park ground. Season dates and legal methods are outside this paper. Check the current agency page for the ground you hunt.