Cumulative effects of human footprint, natural features and predation risk best predict seasonal resource selection by white-tailed deer

Abstract

Land modified for human use alters matrix shape and composition and is a leading contributor to global biodiversity loss. It can also play a key role in facilitating range expansion and ecosystem invasion by anthrophilic species, as it can alter food abundance and distribution while also influencing predation risk; the relative roles of these processes are key to habitat selection theory. We researched these relative influences by examining human footprint, natural habitat, and predator occurrence on seasonal habitat selection by range-expanding boreal white-tailed deer (Odocoileus virginianus) in the oil sands of western Canada. We hypothesized that polygonal industrial features (e.g. cutblocks, well sites) drive deer distributions as sources of early seral forage, while linear features (e.g. roads, trails, and seismic lines) and habitat associated with predators are avoided by deer. We developed seasonal 2nd -order resource selection models from three years of deer GPS-telemetry data, a camera-trap-based model of predator occurrence, and landscape spatial data to weigh evidence for six competing hypotheses. Deer habitat selection was best explained by the combination of polygonal and linear features, intact deciduous forest, and wolf (Canis lupus) occurrence. Deer strongly selected for linear features such as roads and trails, despite a potential increased risk of wolf encounters. Linear features may attract deer by providing high density forage opportunity in heavily exploited landscapes, facilitating expansion into the boreal north.

Predator–prey co‐occurrence in harvest blocks: Implications for caribou and forestry

“…Mule deer summer occupancy of harvest blocks increased with increasing forest age in the surrounding area, and both white-tailed deer and grizzly bear occupancy increased as the amount of young forest (i.e., harvest blocks <50 years old) in the surrounding area decreased. These results contrast with research in northern Alberta, where the total amount of harvest blocks increased the probability of white-tailed deer presence (Dawe, 2011) and white-tailed deer selected areas closer to harvest blocks (Darlington et al, 2022), but complement research in west-central Alberta, where there was a negative relationship between white-tailed deer abundance and harvest blocks (Nielsen et al, 2017). While we found occupancy of harvest blocks was influenced by the age of the forest surrounding the harvest block, our results did not suggest that occupancy was directly related to the site-specific age of the harvest block.…”

“…Successional stages and timing of succession can vary with local conditions like soil moisture, acidity, and topography (Brulisauer et al, 1996; Hart & Chen, 2006), and harvest block age alone may not be an accurate indicator of available ungulate forage in our study area. While it is frequently suggested that deer select for early seral stands (Darlington et al, 2022; Fisher & Wilkinson, 2005; Toews et al, 2018), our results indicate that use of harvest blocks may be dependent on the availability of forest across a range of successional stages. Others have reported that deer select for uneven‐aged mature forest and large‐scale habitat heterogeneity (Kie et al, 2002; Nielsen et al, 2017; Wallmo & Schoen, 1980), and grizzly bear use of harvest blocks depends on the landscape‐level forest composition (Kearney et al, 2019; Nielsen et al, 2004; Stewart et al, 2012).…”

Integration of aerial surveys and resource selection analysis indicates human land use supports boreal deer expansion

“…Finally, we note that, unlike locations from satellite collars, helicopter‐based observations are made only in daylight hours during good flying weather, so helicopter samples might be biased in favor of energy‐intake behaviors—moving about and seeking browse—and against energy‐conservation behaviors, bedding down and seeking shelter, as expected at night and in cold, inclement weather. Notwithstanding this possibility, we note that results and conclusions about white‐tailed deer selection are in agreement with those from satellite‐telemetry‐based RSFs studying deer behavior (Darlington et al, 2022) and camera‐trap studies studying landscape‐scale deer distribution (Fisher et al, 2020, 2021; Fisher & Burton, 2020; Fisher & Ladle, 2022) from the same region. Both sampling modes are agnostic to diel period and weather, suggesting that any bias from helicopter observations do not alter conclusions.…”

“…Certainly, linkages between winter severity (snow depth, snow hardness, and temperature lows and temperature variability) and response to landscape cover have been observed for white‐tailed deer (Dawe et al, 2014; Dawe & Boutin, 2016; Kennedy‐Slaney et al, 2018), so a region‐specific relationship to conifer is not unexpected. The fact that avoidance of conifer for western boreal deer has been demonstrated via satellite‐collar studies (Darlington et al, 2022) as well as multiple camera‐trap studies (Fisher et al, 2020, 2021; Fisher & Burton, 2020; Fisher & Ladle, 2022) lends credence to our conclusions.…”

Behavioral “bycatch” from camera trap surveys yields insights on prey responses to human‐mediated predation risk

“…Understanding trade‐offs between attraction to forage and avoidance of predators, and the spatial scales at which they occur, is fundamental to predicting species responses to caribou management actions, such as habitat restoration and predator control. Further investigation of direct measures of forage availability, resource selection, and mortality from predation is warranted for ungulates in this system (Darlington et al, 2022 ; Finnegan et al, 2018 ; McKay et al, 2021 ).…”